April 2014 Volume 93, Number 4

S R Y C T I E L N U C O E

A 11908908 S N S O O C I AT I

ISSN 0032-5791

Official Journal of the Poultry Science Association Inc. EDITOR-IN-CHIEF ® T. E. Porter (2016) POULTRY SCIENCE SECTION EDITORS ASSOCIATE EDITORS (2013–2014) Environment, Well-Being, ǯȱȱǻŘŖŗśǼ I. Hanning (2016) M. Pines (2014) Ȳȱ ǯȱȱǻŘŖŗśǼ ǯȱ ěȱǻŘŖŗŚǼ T. Poole (2016) I. Estevez (2014) H. Ahmadi (2013) R. M. Hulet (2014) ǯȱȱǻŘŖŗśǼ M. M. Beck (2016) W. Alali (2016) ǯȱ ȱǻŘŖŗśǼ A. Pradhan (2014) ǯȱ ǯȱȱǻŘŖŗśǼ A. Jackson-Davis (2016) ǯȬ ǯȱȱǻŘŖŗśǼ Genetics C. Ashwell (2016) D. Jackwood (2014) ǯȱȱǻŘŖŗŚǼ J. Dodgson (2016) ǯȱȱǻŘŖŗśǼ P. A. Johnson (2016) ǯȱȱǻŘŖŗśǼ ¢ǰȱ ǰ ǯȱȱǻŘŖŗŜǼ C. Jones (2016) K. Reed (2016) Ȳȱ M. R. Bakst (2014) P. Kaiser (2014) T. B. Rodenburg (2014) R. L. Taylor Jr. (2016) R. Beckstead (2016) ǯȱ ȱǻŘŖŗśǼ G. J. M. Rosa (2016) B. R. Behrends (2016) N. Kansaku (2016) W. B. Roush (2013) ȱȱ L. Berghman (2014) E. Kebreab (2013) I. Rozenboim (2014) G. Cherian (2014) W. Berry (2014) E. J. Kim (2016) C. Ruiz-Feria (2013) M. Rodehutscord (2016) D. Biswas (2014) W. Kim (2016) ǯȱǯȱȱǻŘŖŗśǼ E. Esteve-Garcia (2014) J. Brake (2013) ǯȱ ȱǻŘŖŗśǼ ǯȱȱǻŘŖŗśǼ K. Bregendahl (2014) K. W. Koelkebeck (2013) C. Schmidt (2016) ǰȱǰ B. Brehm-Stecher (2016) M. H. Kogut (2014) P. Selle (2013) ȱȱȱ¢ J. Buyse (2014) A. Kollanoor-Johny (2016) ǯȱ ǯȱȱǻŘŖŗśǼ G. Bedecarrats (2016) D. Caldwell (2014) B.-W. Kong (2013) D. H. Shah (2016) ¢¢ǰȱ¢ǰȱ F. Carbonero (2016) ǯȬ ǯȱ ȱǻŘŖŗśǼ ǯȱȱǻŘŖŗśǼ Ȳȱ H. Cervantes (2014) ǯȱ ȱǻŘŖŗśǼ ǯȱǯȱȱǻŘŖŗśǼ S. C. Chapman (2016) W. Kuenzel (2014) K. Shimada (2016) ǯȱĴȱǻŘŖŗśǼ H.-W. Cheng (2016) ǯȱ ¢£ȱǻŘŖŗśǼ F. Siewerdt (2016) A. L. Johnson (2014) K. K. Chousalkar (2013) S. J. Lamont (2014) M. Singh (2014) ǰȱǰ S. Cloutier (2016) ǯȱǯȱ ȱǻŘŖŗśǼ B. Svihus (2014) Ȳȱȱ¢ A. Collin (2014) J. Lee (2013) C. Swaggerty (2016) C. M. Owens (2014) E. Collisson (2014) E. Leone (2016) S. Tesseraud (2014) S. C. Ricke (2016) ǯȱȱǻŘŖŗśǼ F. C. Leung (2016) L. Thompson (2016) M. E. Cook (2014) ǯȱȱǻŘŖŗśǼ K. Tona (2014) ǰȱǰ ǯȱĴȱǻŘŖŗŚǼ ǯȱȱǻŘŖŗśǼ K. Turner (2014) Ȳȱ ǯȱȱǻŘŖŗřǼ H.-C. Liu (2016) H. van den Brand (2014) A. McElroy (2016) P. A. Curtis (2013) K. Maciorowski (2014) ǯȱȱȱȱǻŘŖŗśǼ E. Delezie (2014) R. H. Marin (2016) ǯȱȱǻŘŖŗśǼ ¢ȱ R. Dennis (2014) A. McElroy (2016) R. Walzem (2014) T. E. Porter (2016) ǯȱȱǻŘŖŗśǼ ǯȱǯȱ ȱǻŘŖŗśǼ ǯȱȱǻŘŖŗśǼ OFFICERS ǯȱǯȱ£ȱ ȱǻŘŖŗśǼ ǯȱ££ȱǻŘŖŗśǼ B. Werner (2014) C. Dunkley (2016) K. Mertens (2013) H. Willemsen (2014) K. D. Dunkley (2014) S. Mignon-Grasteau (2014) ǯȱ ǯȱȱǻŘŖŗśǼ M. O. Smith ǯȱ ǯȱȱǻŘŖŗśǼ S. Milillo (2014) ǯȱĴȱǻŘŖŗŚǼ ȱȱ ǯȱȱǻŘŖŗśǼ M. T. Musgrove (2014) E. A. Wong (2016) T. J. Applegate ǯȱȱǻŘŖŗśǼ I. A. Naas (2016) M. Yegani (2016) J. Evans (2016) R. Nannapaneni (2016) A. Yersin (2014) ȱȱ N. Everaert (2014) ǯȱ¢ȱǻŘŖŗśǼ D. Zadworny (2016) D. J. Caldwell M. Farnell (2016) K. J. Navara (2014) M. Zeman (2014) ¢Ȭ S. Foley (2016) R. Nayak (2016) J. Zentek (2016) R. M. Hulet ǯȱȱǻŘŖŗśǼ S. L. Noll (2016) ǯȱȱǻŘŖŗśǼ N. French (2014) ǯȱǯȱȂ ȱǻŘŖŗŜǼ H. Zhang (2016) ȱ J. Fulton (2014) R. Okimoto (2016) X. Zhang (2013) M. T. Kidd R. Gast (2014) O. Onagbesan (2014) ǯȱȱǻŘŖŗśǼ ǯȱǯȱ ȱǻŘŖŗśǼ H. Parmentier (2014) H. Zhou (2016) A. P. McElroy (2014) ǯȱ ȱǻŘŖŗśǼ ǯȱǯȱ¢ȱǻŘŖŗśǼ ǯȱȱǻŘŖŗśǼ I. Gimeno (2016) A. J. Pescatore (2014) ǯȱĚȱǻŘŖŗŚǼ D. R. McIntyre (2014) H. Guo (2016) T. Pharr (2016) ǯȱǯȱȱǻŘŖŗśǼ ǯȱ ȱǻŘŖŗśǼ ȱȱě Chris Davies Ron Keller J. B. Hess (2016) ȓǯ Sharon Frick Lisa Krohn S. R. Fernandez (2016) Susan Pollock, Managing Editor Gayle Gleichman Shauna Miller Louise Adam Armgard Haken ȱ Jorge Cazares Christine Horger K. A. Byrne (2014) ǯȱ ǯȱȱǻŘŖŗśǼ ¢ȱ®ȱǻ ȱŖŖřŘȬśŝşŗǼȱȱȱŗŘȱȱȱ¢ȱǻ¢Ǽǯȱȱȱȱȱ¢ǰȱǰȱ ȱȱȱĜǯȱǱȱȱȱȱȱ¢ȱ, Journals Customer Service ǰȱ¡ȱ¢ȱǰȱŘŖŖŗȱȱǰȱ¢ǰȱȱŘŝśŗřȬŘŖŖşǯ S R Y C Subscription informationȱ¢ȱȱȱǰȱǞśŞşȱǻǼǰȱǞŜŚŝȱǻǼǰȱȱǞŜŞŗȱǻȱȱ T I E L N U Ǽǯȱ¢ȱȱȱȱǰȱǡřŞŖȦǧŚŜŝȱǻǼǰȱǡŚřŞȦǧśřşȱǻǼǰȱȱǡŚŜŗȦǧśŜŝȱǻȱȱǼǲȱ C O E

P ȱȱȱǞŚśǯȱȱȱȱȱȱȱȱȱ ȱȱȱǰȱ ¡ȱ ǰȱ ȱȱǰȱ¡ȱŘȱŜǰȱ ǰȱǱȱƸŚŚȱǻŖǼȱŗŞŜśȬřśȬřşŖŝǰȱȬǱȱǯǯȓǯ ǯȱ ȱȱǰȱȱȱ ȱȱȱǰȱ¡ȱ ǰȱŘŖŖŗȱȱǰȱ

A 11908908 ¢ǰȱȱŘŝśŗřǯȱǱȱŞŖŖȬŞśŘȬŝřŘřȱǻȬȱȱȦǼȱȱşŗşȬŜŝŝȬŖşŝŝǰȱȬǱȱȓǯǯȱ S N S O O C I AT I ȱȱȱȱȱȱȱǞŗŘŖȱȱȱȱȱȱ¢ȱ. A print subscrip- ȱȱȱȱȱȱȱǞŞŖȱǻǼȱȱǞŗŖŝȱǻȱȱǼǯȱȱȱȱȱȱȱ POULTRY SCIENCE ȱȱ¢ȱȱȱ ǯǰȱŗŞŖŖȱǯȱȱǰȱȱŗŖŖǰȱǰȱ ȱŜŗŞŘŖȬŜşŝŚǰȱǱȱŘŗŝȬřśŜȬśŘŞśǰȱ ASSOCIATION INC. ȬǱȱȓǯǯ Organized 1908 Claimsȱȱȱȱęȱȱȱ ȱȱȱȱȦȱȱǻ ȱȱǼǯȱ- ĴǱȦȦ ǯ¢ǯ ȱȱȱȱ¢ȱȱȱȱȱǯȱȱȱȱȱȱ¢ȱ, Journals Customer ĴǱȦȦǯǯ ȱǰȱ¡ȱ¢ȱǰȱŘŖŖŗȱȱǰȱ¢ǰȱȱŘŝśŗřǰȱǱȱŞŖŖȬŞśŘȬŝřŘřȱǻȬȱȱȦ ǼȱȱşŗşȬŜŝŝȬŖşŝŝǰȱȬǱȱȓǯǯ

ȚŘŖŗŚȱ¢ȱȱȱ ǯȱȱȱȱȱ¢ȱȱȱȱ¢ȱ¡ȱ¢ȱǯ POULTRY SCIENCE®

Volume 93 April 2014 Number 4

CONTENTS ENVIRONMENT, WELL-BEING, AND BEHAVIOR Modelling of feather pecking behavior in beak-trimmed and non-beak-trimmed crossbred laying hens: Variance component and trait-based approach Y. Sun, E. D. Ellen, J. J. van der Poel, H. K. Parmentier, and P. Bijma ...... 773 Comparative proteomic analysis of the breast muscle response to chronic corticosterone administration in broiler chickens showing long or short tonic immobility Wenyan Fu, Yujing Duan, Song Wang, Yingdong Ni, R. Grossmann, and Ruqian Zhao ...... 784

Research Note Determination of space use by laying hens using kinematic analysis Joy A. Mench and Richard A. Blatchford ...... 794

GENETICS The potential for archiving and reconstituting valuable strains of turkey (Meleagris gallopavo) using primordial germ cells Alexander J. Wade, Nick A. French, and Grenham W. Ireland ...... 799

Genetic parameters for feather pecking and aggressive behavior in a large F2-cross of laying hens using generalized linear mixed models J. Bennewitz, S. Bögelein, P. Stratz, M. Rodehutscord, H. P. Piepho, J. B. Kjaer, and W. Bessei ...... 810

IMMUNOLOGY, HEALTH, AND DISEASE In ovo inoculation of chicken embryos with probiotic bacteria and its effect on posthatch Salmonella susceptibility J. E. de Oliveira, E. van der Hoeven-Hangoor, I. B. van de Linde, R. C. Montijn, and J. M. B. M. van der Vossen ...... 818 Dramatic differences in the response of macrophages from B2 and B19 MHC-defined haplotypes to interferon gamma and polyinosinic:polycytidylic acid stimulation Maisie E. Dawes, Lisa M. Griggs, Ellen W. Collisson, Worthie E. Briles, and Yvonne Drechsler ...... 830 Effects of dietary clays on performance and intestinal mucus barrier of broiler chicks challenged with Salmonella enterica serovar Typhimurium and on goblet cell function in vitro J. A. S. Almeida, N. P. Ponnuraj, J. J. Lee, P. Utterback, H. R. Gaskins, R. N. Dilger, and J. E. Pettigrew ...... 839 Transcriptome analysis reveals an activation of major histocompatibility complex 1 and 2 pathways in chicken trachea immunized with infectious laryngotracheitis virus vaccine Juan Luo, José A. Carrillo, Kimberly R. Menendez, Nathaniel L. Tablante, and Jiuzhou Song ...... 848 Fluoroquinolone resistance and molecular characterization of gyrA and parC quinolone resistance- determining regions in Escherichia coli isolated from poultry M. Vanni, V. Meucci, R. Tognetti, P. Cagnardi, C. Montesissa, A. Piccirillo, A. M. Rossi, D. Di Bello, and L. Intorre ...... 856 Chicken NK-lysin is an alpha-helical cationic peptide that exerts its antibacterial activity through damage of bacterial cell membranes Mi Ok Lee, Hyun-Jun Jang, Jae Yong Han, and James E. Womack...... 864 Contribution of Salmonella Enteritidis virulence factors to intestinal colonization and systemic dissemination in 1-day-old chickens Tarek M. Addwebi, Douglas R. Call, and Devendra H. Shah ...... 871 Arginine and vitamin E improve the immune response after a Salmonella challenge in broiler chicks X. Liu, J. A. Byrd, M. Farnell, and C. A. Ruiz-Feria ...... 882

Research Note Exposure of juvenile Leghorn chickens to lead acetate enhances antibiotic resistance in enteric bacterial flora M. Nisanian, S. D. Holladay, E. Karpuzoglu, R. P. Kerr, S. M. Williams, L. Stabler, J. Vaun McArthur, R. Cary Tuckfield, and R. M. Gogal Jr...... 891

METABOLISM AND NUTRITION Effects of low ambient temperatures and dietary vitamin C supplement on growth performance, blood parameters, and antioxidant capacity of 21-day-old broilers X. Yang, Y. H. Luo, Q. F. Zeng, K. Y. Zhang, X. M. Ding, S. P. Bai, and J. P. Wang ...... 898 Effect of calcium level and phytase addition on ileal phytate degradation and amino acid digestibility of broilers fed corn-based diets A. M. Amerah, P. W. Plumstead, L. P. Barnard, and A. Kumar ...... 906 Lactobacillus sakei modulates mule duck microbiota in ileum and ceca during overfeeding F. Vasaï, K. Brugirard Ricaud, L. Cauquil, P. Daniel, C. Peillod, K. Gontier, A. Tizaoui, O. Bouchez, S. Combes, and S. Davail ...... 916 Understanding carbohydrate structures fermented or resistant to fermentation in broilers fed rapeseed (Brassica napus) meal to evaluate the effect of acid treatment and enzyme addition Annemieke M. Pustjens, Sonja de Vries, Henk A. Schols, Harry Gruppen, Walter J. J. Gerrits, and Mirjam A. Kabel ...... 926 Dietary electrolyte balance influences ileal endogenous amino acid losses in broiler chickens S. A. Adedokun and T. J. Applegate ...... 935

MOLECULAR, CELLULAR, AND DEVELOPMENTAL BIOLOGY Bone characteristics and femoral strength in commercial toms: The effect of protein and energy restriction R. C. Van Wyhe, P. Regmi, B. J. Powell, R. C. Haut, M. W. Orth, and D. M. Karcher ...... 943 Spatiotemporal expression patterns of doublesex and mab-3 related transcription factor 1 in the chicken developing gonads and Müllerian ducts T. Omotehara, C. A. Smith, Y. Mantani, Y. Kobayashi, A. Tatsumi, D. Nagahara, R. Hashimoto, T. Hirano, Y. Umemura, T. Yokoyama, H. Kitagawa, and N. Hoshi ...... 953

PHYSIOLOGY, ENDOCRINOLOGY, AND REPRODUCTION Effects of inorganic and organic manganese supplementation on gonadotropin-releasing hormone-I and follicle-stimulating hormone expression and reproductive performance of broiler breeder hens Jingjing Xie, Chuanhuan Tian, Yongwen Zhu, Liyang Zhang, Lin Lu, and Xugang Luo ...... 959 Effects of genetic strain and light intensity on blood physiological variables of broilers grown to heavy weights H. A. Olanrewaju, J. L. Purswell, S. D. Collier, and S. L. Branton ...... 970 Supplemental l-arginine and vitamins E and C preserve xanthine oxidase activity in the lung of broiler chickens grown under hypobaric hypoxia J. Bautista-Ortega, A. Cortes-Cuevas, E. A. Ellis, and C. A. Ruiz-Feria ...... 979

PROCESSING, PRODUCTS, AND FOOD SAFETY Preparation and characterization of gelatin from collagen biomass obtained through a pH-shifting process of mechanically separated turkey meat L. Du, L. Keplová, Z. Khiari, and M. Betti ...... 989 Sequential separation of lysozyme, ovomucin, ovotransferrin, and ovalbumin from egg white E. D. N. S. Abeyrathne, H. Y. Lee, and D. U. Ahn ...... 1001 Separation of ovotransferrin and ovomucoid from chicken egg white E. D. N. S. Abeyrathne, H. Y. Lee, and D. U. Ahn ...... 1010 Bioactive amines and microbiological quality in pasteurized and refrigerated liquid whole egg I. O. P. Rêgo, L. D. M. Menezes, T. C. Figueiredo, D. D. Oliveira, J. S. R. Rocha, L. J. C. Lara, A. L. Lima, M. R. Souza, and S. V. Cançado ...... 1018 Prevalence and antimicrobial resistance of Listeria, Salmonella, and Yersinia species isolates in ducks and geese Hossein Jamali, Behrad Radmehr, and Salmah Ismail ...... 1023

PRODUCTION, MODELING, AND EDUCATION Alternative growth functions for predicting body, carcass, and breast weight in ducks: Lomolino equation and extreme value function A. Faridi, D. Murawska, A. Golian, M. Mottaghitalab, A. Gitoee, S. Lopez, and J. France ...... 1031

ASSOCIATION NOTES Nunc Dimittis ...... 1043 Positions Available ...... 1044 ENVIRONMENT, WELL-BEING, AND BEHAVIOR

Modelling of feather pecking behavior in beak-trimmed and non-beak-trimmed crossbred laying hens: Variance component and trait-based approach

Y. Sun ,*1 E. D. Ellen ,* J. J. van der Poel,* H. K. Parmentier ,† and P. Bijma *

* Animal Breeding and Genomics Centre, Wageningen University, PO Box 338, 6700 AH Wageningen, the Netherlands; and † Adaptation Physiology Group, Wageningen University, PO Box 338, 6700 AH Wageningen, the Netherlands Downloaded from

ABSTRACT Because of a ban on the use of beak trim- netic parameters for FCS with 2 variance components ming in some European countries, feather pecking is models: a traditional linear animal model and a model becoming a substantial problem in the layer industry, combining direct and associative genetic effects. In the both from animal welfare and economic points of view. second part of the present study, a trait-based analy- http://ps.oxfordjournals.org/ The feather condition score (FCS) as a measure of sis for FCS was conducted to investigate whether NAb feather damage has been shown to be closely related isotype titers can explain variation in FCS among in- to feather pecking behavior in laying hens housed in dividuals, by fitting a linear mixed model. Though the groups. To obtain a better understanding of genetic estimated associative genetic variance was substantial, and other biological mechanisms underlying feather associative effects for FCS were not statistically signifi- pecking behavior, data on FCS of a population of 2,724 cant in both populations (P = 0.09 in beak-trimmed

female offspring from crossing 50 male W1 and 907 birds, and P = 0.08 in non-beak-trimmed birds). This at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 female WB purebred lines were used. The offspring of suggests an insufficient number of records on FCS. In- 25 sires were beak-trimmed, and the offspring of an- dividual’s NAb isotypes titers did not show direct effect other 25 sires were non-beak-trimmed. Titers of plasma for FCS of itself, but individual’s IgG titers showed a natural antibody (NAb) isotypes IgM and IgG binding suggestive effect on the FCS of cage mates (associative keyhole limpet hemocyanin at 24 wk of age were mea- effect) in beak-trimmed laying hens, which need further

sured. Feather condition was scored at 53 wk of age. confirmation. In the first part of the present study, we estimated ge- Key words: laying hen, beak trimming, feather pecking, feather condition score, natural antibody 2014 Poultry Science 93 :773–783 http://dx.doi.org/ 10.3382/ps.2013-03367

INTRODUCTION birds, and can even lead to death (Gentle and Hunter, 1991). Feather pecking is not only a welfare but also a During the past decades, poultry breeders have suc- serious economic problem (Rodenburg et al., 2008). De- cessfully improved the production performance, either creased egg production caused by feather pecking was egg production in layers or meat production in broil- observed (Johnsen et al., 1998). Feather loss because ers. However, considering the ever-increasing social of feather pecking can lead to heat loss, which results concern, future animal husbandry is required to pay in higher maintenance energy requirements (Blokhuis more attention to enhancing animal welfare. In laying and Wiepkema, 1998). Mortalities due to cannibalism, hens, animal welfare is particularly focused on feather which is considered the ultimate phase of severe feather pecking behavior. Feather pecking is defined as pecking pecking, can be substantial. Hill (1986), for example, toward the plumage of other birds. Two major forms of found up to 15% mortality in laying hens housed in feather pecking can be distinguished: gentle and severe aviaries, whereas Peeters et al. (2012) and Ellen et al. feather pecking (Keeling, 1995). Severe feather pecking (2008) found around 32 and 48% mortality, respective- causes damage to the birds, results in bald patches, ly, due to cannibalism in cage-housed birds. Prohibi- denuded area, hemorrhage, wounds, is painful for the tion of both cage housing system and beak trimming because of animal welfare concern in many European Union member countries increases the risk of feather © 2014 Poultry Science Association Inc. pecking and cannibalism. Received May 31, 2013. Accepted November 3, 2013. Better understanding of the genetic and biological 1 Corresponding author: [email protected] mechanisms of feather pecking is needed to find alter-

773 774 SUN ET AL. native ways of preventing this unfavorable behavior. effect of the individual’s own genotype on its FCS). Feather condition score (FCS) is a measure of feather This suggests that the NAb isotypes titers may not damage, which has been shown to be closely related to only be related with the susceptibility to be pecked at, feather pecking behavior in hens housed in groups (Bil- but also particularly with the propensity to perform cík and Keeling, 1999; Uitdehaag et al., 2008). Differ- feather pecking. In our previous study, NAb isotype ent from ordinary traits, FCS is a so-called interacting titers were reported to be associated with survival of phenotype, a trait whose value is also affected by the laying hens (Sun et al., 2011). As mentioned previously, behavior of an individual’s conspecifics (the cage mates severe feather pecking and cannibalism may also induce that are kept with the focal individual in the same cage mortality. It is possible that the NAb isotype is associ- in case of laying hens; Moore et al., 1997). In contrast ated with survival by regulating the feather pecking to the direct genetic effect of an individual on its own behavior. Therefore, in the second part of the present Downloaded from phenotype, the heritable effect of an individual on the study, we model an individual’s FCS as a function of phenotype of a conspecific is known as associative effect the NAb isotype titers of the individual and those of or indirect genetic effect (Griffing, 1967; Wolf, 2003; Bi- its cage mates to investigate the possible relationship jma et al., 2007). Associative effects influence a trait’s between feather pecking behavior and levels of NAb inherence and contribute to heritable variation (Moore isotype IgM and IgG in beak-trimmed and non-beak- http://ps.oxfordjournals.org/ et al., 1997; Bijma, 2011). For the genetic parameters trimmed laying hens. estimation of survival days in non-beak-trimmed laying hens, the inclusion of associative effects in the model MATERIALS AND METHODS gave higher heritable variation than a traditional linear animal model (Muir, 2005; Ellen et al., 2008; Peeters et Study Population al., 2012). Consequences of feather pecking behavior in

beak-trimmed and non-beak-trimmed laying hens are Female crossbred offspring of 2 commercial purebred at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 different, and therefore FCS can be a different trait White Leghorn layer lines (male W1 and female WB) in both types of birds. In the first part of the pres- were provided by Institut de Sélection Animale (ISA) ent study, we estimated genetic parameters for FCS B.V., the layer breeding division of Hendrix Genetics in beak-trimmed and non-beak-trimmed laying hens, (Boxmeer, the Netherlands). The 2 purebred lines and respectively, with 2 variance components models: a tra- the crossbred show high mortality with intact beaks ditional linear model and a linear animal model com- (Ellen et al., 2008; Peeters et al., 2012). The W1 and bining direct and associative effects. WB lines were verified as “high and low natural an- Besides modeling the FCS using the variance com- tibody isotype” lines, respectively (Star et al., 2007; ponent approach, a trait-based approach can help to Sun et al., 2011). Uitdehaag et al. (2008) showed that understand the biological mechanism of social interac- birds of the W1 lines have more severe feather pecking tions (Moore et al., 1997; Wolf et al., 1998). Knowledge behavior and feather damage than birds of the WB line of the traits that underlie the interacting phenotype (Uitdehaag et al., 2008), whereas Ellen et al. (2008) is, however, needed for the trait-based approach (Kirk- found a higher mortality in the WB line. Fifty sires patrick and Lande, 1989). Brain serotonergic levels of line W1 were randomly chosen and mated with 908 (Chaouloff, 2000) and some neurotransmitters such as dams of line WB, where dams were nested within sires. dopamine and hormones (Cheng et al., 2003) were cor- Sires and dams were housed individually in cages. Each related with feather pecking. El-Lethey et al. (2003) sire was mated to approximately 18 dams. The eggs found that feather pecking was related with corticoste- from the same dam were collected and incubated in rone levels, which also reduced immune responses. Re- a separate cell of the incubation plate, discriminated cently, the effects of immunity on feather pecking be- from the eggs of the other dams. After hatch, the fe- havior were suggested by several studies. Buitenhuis et male chicks from the same cell (dam) were assigned al. (2004) reported a significant genetic and phenotypic unique barcode wing bands, allowing for identification correlation between feather pecking and primary anti- of individuals. Each dam contributed on average 3 fe- body response to keyhole limpet hemocyanin (KLH). male offspring, resulting in 2,724 offspring. Parmentier et al. (2009) found that when chickens were challenged intratracheallly and repeatedly at a young Housing and Management age with different doses of the endotoxin lipopolysaccharide and the protein human serum albumin, they All chickens were hatched, sexed, and wing-banded in showed different levels of feather damage at an older the right wing for individual identification at the same age. In addition, some SNP or QTL that were associ- time. Only female chicks were kept for this study. Off- ated with FCS (Biscarini et al., 2010b) were also sig- spring of 25 sires were beak trimmed, whereas offspring nificantly associated with levels of natural antibody of another 25 sires were kept with intact beaks. Chicks (NAb) isotypes IgM and IgG binding KLH (Sun et were trimmed manually at 1 d old using a hot blade al., 2013a). These variations were mostly reported to to remove and cauterize the tip of the beak. Chicks be associated with the associative genetic effects on were allocated to rearing cages randomly with respect FCS, and few with the direct genetic effect (the genetic to beak trimming, 60 individuals per cage. From 5 wk FEATHER PECKING AND NATURAL ANTIBODIES IN HENS 775 of age onward, the hens were housed with 20 individu- dead (0) or alive (1) at the end of the study. From these als per cage. The cage number for each hen was not data, survival rate was calculated as the percentage of recorded. At 17 wk of age, all hens were transported laying hens still alive at the end of the study. Survival to a high-light intensity laying house with conventional days were defined as the number of days from the start 5-bird cages (44 cm height × 40 cm depth × 55 cm of the observation until either death or termination of width). Each pair of back-to-back cages shared 2 drink- the present study, with a maximum of 457 d. At 24 wk ing nipples. A feeding trough was in front of the cages, of age, 2-mL blood samples of all birds were taken from with a length of 55 cm per cage. After placing the birds the wing vein using the plastic vacuum blood collection in the cages, hen were wing-banded in the left wing as tubes containing sodium heparin. The bleeding proce- well, to avoid loss of data. There were 6 rows (3 double dure for each bird was 15 to 30 s. The plasma samples rows) of cages in the laying house, with corridors in were collected after the centrifugation of the blood and Downloaded from between to allow employees to have access to the cages used to measure NAb isotype IgM and IgG titers bind- (Figure 1). The outer 2 double rows consisted of 3 lev- ing KLH. At 53 wk of age, the individual feather condi- els (top; middle, closest to the light; and bottom). The tion of neck, back, rump, and belly areas was scored. middle double rows consisted of 4 levels (super top; http://ps.oxfordjournals.org/ top; middle, closest to the light; and bottom). Hens NAb Isotype IgM and IgG Titers were only placed in the top and middle levels. Five Binding KLH hens that were a mix of half sibs and full sibs were allocated to the same cage. The hens in each pair of There are no antigens in the environment of laying back-to-back cages had received the same treatment hens that show immunological cross-reactivity with regarding beak trimming and could contact their back KLH based on the literature and pilot experiments. neighbors through the wire mesh. Contact with hens Thus, prior exposure or sensitization to this protein in adjacent cages was impossible because of the closed is considered unlikely. According to definitions, NAb at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 wall in between. Water and standard commercial layer are immunoglobulins present in animals in the absence diet was provided ad libitum. Rearing started with a of earlier (deliberate) immunization, vaccination, or 9L:15D light scheme and increased 1 h per week until infection (Avrameas, 1991). Therefore, the antibodies 16L:8D was reached when the hens were 26 wk of age. detected in the plasma binding KLH were regarded as The hens received routine vaccinations for Marek’s dis- NAb. Indirect ELISA as described earlier (Sun et al., ease (d 1), infectious bronchitis (d 1, wk 2, 10, 12, and 2011) was performed to measure levels of plasma NAb 15), Newcastle disease (wk 2, 6, 12, and 15), infectious isotypes IgM and IgG binding KLH at 24 wk of age bursal disease (wk 3 and 15), turkey rhinotracheitis (wk by the same person on different days. The NAb iso- 8 and 18), fowl pox (wk 15), chicken anemia virus (wk types titers were only measured once for each sample. 15), and avian encephalomyelitis (wk 15). However, each plate was run with 2 duplicated positive plasma samples of 8 step-wise dilutions. The interassay Study Design CV and intraassay CV was calculated as 5.1 and 4.2%, respectively. All hens were observed daily from 17 until 83 wk of age for survival. Hens that died were removed from FCS the cages without replacement. Wing-band number and date of death were recorded. Cause of death was not In the present study, feather damage of the laying determined. For each hen, information was collected hens was assessed by evaluating the individual’s feather on survival and survival days. Survival was defined as condition at 53 wk of age of 4 body areas: neck, back, rump, and belly, which are the frequent targets of feather pecking. The scoring was performed by 4 persons, following the classification of Bilcík and Keeling (1999), as modified by Uitdehaag et al. (2008). In a pilot study, the average correlation between persons performing the scoring was estimated 0.82 for neck, back, and rump, and 0.72 for belly by E. D. Ellen (Wageningen Univer- sity, Wageningen, the Netherlands, personal communication). There were 6 classes for FCS, ranging from 0 (intact feathers) to 5 (almost all feathers missing), with higher score indicating more damage. The sum of scores of 4 areas was used as an overall parameter of feather condition. Sum = individual neck score + individual back score + individual rump score + individual Figure 1. The division of the stable, showing the light arrangement belly score (ranging from 0 to 20). Birds that died be- and numbers of the cages per row and level (Sun et al., 2013a). fore 53 wk of age (252 out of 1,169 non-beak-trimmed, 776 SUN ET AL. and 31 out of 1,555 beak-trimmed laying hens) did not ⎡ 2 ⎤ ⎡ ⎤ ⎢ σσ⎥ ⎢a1_D ⎥ A1_DA 12_D receive their FCS. var ⎢ ⎥ = ⎢ ⎥ ⊗⊗ A, ⎢ 2 ⎥ ⎣⎢a2_D ⎦⎥ ⎢σσ⎥ ⎣ AA12_D 2_D ⎦ Data Analysis where σ2 is the direct genetic variance for beak- Descriptive statistical analyses were performed using A1_D SAS 9.1.2 (SAS Institute Inc., Cary, NC). Effects were trimmed laying hens, σ2 is the direct genetic vari- considered significant at the level of P < 0.05. A GLM A2_D ance for non-beak-trimmed laying hens, σ is the was used to study the differences in FCS between beak- A12_D trimmed and nontrimmed birds, and the differences in direct genetic covariance between beak-trimmed and IgM and IgG titers binding KLH between both groups. non-beak-trimmed laying hens. The ⊗ indicates the Downloaded from The correlations between FCS of 4 different body areas Kronecker product of matrices. The A is the additive were estimated by Pearson product-moment correla- genetic relationship matrix generated from a 5-generation. The average IgM and IgG titer of cage mates of tion pedigree. Phenotype variance was calculated as σ22=+σσ 2 + σ 2. every individual laying hen was also calculated. P A cage e Heritabilities were calculated as D http://ps.oxfordjournals.org/ 222 Variance Components Estimation of FCS. A tra- h = σσ/. ditional linear animal model and a direct-associative APD (2) Direct-associative effect model: To estimate ge- effect model were used to estimate the variance compo- netic parameters for both direct and associative effects, nent of FCS. The FCS of beak-trimmed and non-beak- the following model extended from Muir (2005) was trimmed laying hens were analyzed separately using used: the GLM procedure of the SAS program (SAS Institute Inc.). To correct for systematic nongenetic differences ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ y X0b ⎢Z01_D ⎥ ⎢a1_D ⎥ among observations, factors with P < 0.10 from the ⎢ 1 ⎥ ⎢ 1 ⎥ ⎢ 1 ⎥ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 ⎢ ⎥ = ⎢ ⎥ ⎢ ⎥ + ⎢ ⎥⎥ ⎢ ⎥ GLM were included as fixed effects in the model for ⎣y2 ⎦ ⎣ 0X2 ⎦ ⎣b2 ⎦ ⎣⎢ 0Z2_D ⎦⎥ ⎣⎢a2_D ⎦⎥ [2] estimating genetic parameters. Fixed effects for FCS ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ Z01_S a1_S V0cage e in beak-trimmed laying hens were (1) row of cages, (2) ⎢ ⎥ ⎢ ⎥ + ⎢ 1 ⎥ ⎢ 1 ⎥ ⎢ 1 ⎥ , + ⎢ ⎥ ⎢ ⎥⎥ ⎢ ⎥ ⎢ ⎥ + ⎢ ⎥ level of the cages where the laying hens were located to ⎣⎢ 0Z2_S ⎦⎥ ⎣⎢a2_S ⎦⎥ ⎣ 0V2 ⎦ ⎣cage2 ⎦ ⎣e2 ⎦ account for infrastructural effects such as light intensity difference (Kjaer and Vestergaard, 1999), and (3) person who scored the feather condition. In non-beak- where the vectors and incidence matrices correspond trimmed laying hens, only the person who scored the to those in the traditional linear animal model; ZS is feather condition was included as a fixed effect. an incidence matrix linking an individual’s FCS to its To compare genetic parameters for FCS between cage mates’ breeding value vector; and aS is a vector beak-trimmed and non-beak-trimmed laying hens, a of social breeding values for all cage mates. The direct- bivariate model was used by treating FCS as different associative genetic (co)variance structure was traits for both populations. ⎡ ⎤ (1) Traditional linear animal model: Genetic param- ⎢a1_D ⎥ ⎢ ⎥ eters of FCS were first estimated using a traditional ⎢a ⎥ var ⎢ 2_D ⎥ linear animal model as implemented in the ASReml ⎢ ⎥ = software package (Gilmour et al., 2006): ⎢a1_S ⎥ ⎢ ⎥ ⎣⎢a2_S ⎦⎥ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡Z0⎤ ⎡a ⎤ y1 X01 b1 ⎢ 1_D ⎥ ⎢ 1_D ⎥ ⎡ ⎤ ⎢ ⎥ = ⎢ ⎥ ⎢ ⎥ + ⎢ ⎥⎥ ⎢ ⎥ σσ2 σσσ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ A1_D AA12_D 1_DS A1_D_2_S ⎥ ⎣y2 ⎦ ⎣ 0X2 ⎦ ⎣b2 ⎦ ⎣⎢ 0Z2_D ⎦⎥ ⎣⎢a2_D ⎦⎥ ⎢ ⎥ [1] ⎢ ⎥ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎢ σσσσ2 ⎥ V0cage e AA12_D 2_DAA 2_D_1_S 2_DS + ⎢ 1 ⎥ ⎢ 1 ⎥ + ⎢ 1 ⎥ , ⎢ ⎥ ⊗⊗ A, ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ 2 ⎥ ⎣ 0V2 ⎦ ⎣cage2 ⎦ ⎣e22 ⎦ ⎢ σσ σ σσ ⎥ ⎢ AA1_DS 2_D_1_S A 1_S A12_S ⎥ ⎢ ⎥ ⎢σσσσ2 ⎥ where subscript 1 indicates beak-trimmed laying hens ⎣ AA1_D_2_S 2_DSAA 12_S 2_S ⎦ and subscript 2 indicates non-beak-trimmed laying y hens; is a vector of individual sum of FCS at 53 wk of where σ2 , σ2 , and σ are the same as in the age; b is a vector of fixed effects, with incidence matrix A1_D A2_D A12_D traditional linear animal model. The σ2 is the asso- X linking FCS to fixed effects; a is a vector of usual A1_S breeding values, with incidence matrix ZD linking FCS ciative genetic variance for beak-trimmed laying hens; to the breeding value; cage is a vector of independent σ2 is the associative genetic variance for non-beak- random cage effects; V is an incidence matrix linking A2_S trimmed laying hens; σ is the associative genetic observations to random cage effects; and e is vector A12_S of random residuals. The direct genetic (co)variance covariance between beak-trimmed and non-beak- trimmed laying hens; σ is the direct-associative ge- structure was A1_DS FEATHER PECKING AND NATURAL ANTIBODIES IN HENS 777 netic covariance in beak-trimmed laying hens; σ is of individual IgG or individual IgM or average IgM ti- A2_DS the direct-associative genetic covariance in non-beak- ters of the cage mates or average IgG titers of the cage trimmed laying hens; σ is the genetic covariance mates, b is the estimated parameter for the covariable A1_D_2_S x. The model for non-beak-trimmed laying hens was between the direct effect of beak-trimmed and the associative effect of non-beak-trimmed laying hens; y=+μ row + person + bx + cage , [4] σ is the genetic covariance between the associa- ikl i k l A2_D_1_S tive effect of non-beak-trimmed laying hens and the where all the terms are the same as those specified in direct effect of beak-trimmed laying hens. The total model [3]. The sum FCS for beak-trimmed and non- heritable variance for response to selection was 22 22 beak-trimmed laying hens was tested for normality be- Downloaded from σσ=+−21()nn σ +− () 1 σ (Bijma et al., TBV ADDA SSA fore model [3] and [4] were run with a MIXED proce- σ2 2007). The P is the phenotypic variance, dure of the SAS program (SAS Institute Inc.). 22 2 2 2 σσ=+−+()n 1 σ σ + σ . The n is the number PADS A cagee of laying hens kept in the same cage, and n = 5 in the RESULTS

present study. The T 2 expresses the total heritable http://ps.oxfordjournals.org/ variance relative to the phenotypic variance: FCS of Beak-Trimmed and Non-Beak- 22= σσ/. 2 T TBV P Trimmed Crossbred Laying Hens Likelihood ratio tests were used to test the signifi- The average individual FCS of the 4 body areas cance of the random associative effect in a univariate (neck, back, rump, and belly) at 53 wk of age for beak- model in beak-trimmed and non-beak-trimmed laying trimmed and non-beak-trimmed laying hens is shown hens, respectively. in Table 1. In both populations, the score for belly was Trait-Based Approach. To investigate whether the lowest among the 4 areas, indicating that the belly at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 NAb can explain variation in FCS among individuals, area was less pecked at. In contrast, the neck and rump a trait-based analysis for FCS was conducted by fitting were areas with the highest scores, indicating more a linear mixed model, following Moore et al. (1997). damage in these areas. The CV ranged from 23 to 48% The fixed effects were the same as those detected in the for non-beak-trimmed, and ranged from 31 to 70% for variance components approach. Therefore, the model beak-trimmed laying hens. This indicated considerable for beak-trimmed laying hens was variations for the FCS of both populations. The GLM analysis showed that the FCS for different body areas y=+μ row + level + person + bx +cage , [3] ijkl i j k l and the sum of the FCS in beak-trimmed laying hens was significantly lower than that in non-beak-trimmed where yijkl is individual sum FCS at 53 wk of age; μ is laying hens, indicating that non-beak-trimmed laying the overall mean; rowi is the fixed effect of row of the hens had more feather damage (Table 1). cage (i = 1, 2, 3, 4, 5, 6); levelj is the fixed effect of level Both in beak-trimmed and non-beak-trimmed laying of the cage (j = 1, 2); personk is the effect of the kth (k hens, as expected, the correlation coefficients between = 1, 2, 3, 4) person who scored the feather condition; the scores of different body areas were positive (Table cagel is the random effect of cage l, x is the fixed effect 2). The correlations between the areas that were close

Table 1. Number of observations (n), and average feather condition score (FCS; ±SD) of 4 body areas and sum FCS for surviving and nonsurviving beak-trimmed and non-beak-trimmed crossbred laying hens at 53 wk of age1,2

Population n Neck*** Back*** Rump*** Belly*** Sum*** Non-beak-trimmed Survival3 813 3.78 ± 0.91 3.54 ± 1.04 3.76 ± 1.03 2.77 ± 1.31 13.77 ± 3.48 Nonsurvival4 104 3.69 ± 0.85 3.70 ± 1.16 3.92 ± 1.12 2.92 ± 1.44 14.32 ± 4.10 Total 917 3.70 ± 0.85 3.56 ± 1.05 3.78 ± 1.04 2.79 ± 1.33 13.83 ± 3.55 CV5 (%) 23 29 28 48 26 Beak-trimmed Survival 1,447 2.97 ± 0.91 2.64 ± 1.20 2.84 ± 1.35 1.79 ± 1.25 10.23 ± 3.80 Nonsurvival 77 2.94 ± 1.13 2.78 ± 1.17 3.13 ± 1.28 1.86 ± 1.31 10.70 ± 4.02 Total 1,524 2.96 ± 0.92 2.65 ± 1.20 2.85 ± 1.35 1.79 ± 1.25 10.25 ± 3.81 CV (%) 31 45 47 70 37 1There are 6 classes for FCS, ranging from 0 (intact feathers) to 5 (almost all feathers missing), with higher score indicating more damage. 2Sum FCS = individual neck score + individual back score + individual rump score + individual belly score (ranging from 0 to 20). 3Survival indicated the birds survived until the end of the observation period (83 wk of age). 4Nonsurvival indicated the bird died between 53 wk of age (when feather condition scoring was performed) and the end of the observation period (83 wk of age) because the birds that died before were not scored. 5CV (%) = (SD/mean) × 100%. ***P < 0.0001, which indicates that FCS for beak-trimmed and non-beak-trimmed laying hens is significantly different. 778 SUN ET AL.

Table 2. Pearson correlation coefficients between the feather condition scores of 4 body areas (neck, back, rump, and belly) in beak-trimmed (above the diagonal) and non-beak-trimmed crossbred laying hens (below the diagonal)1

Item Neck Back Rump Belly Sum2

Neck — 0.66 0.44 0.37 0.73 Back 0.65 — 0.71 0.46 0.88 Rump 0.50 0.71 — 0.52 0.85 Belly 0.40 0.60 0.62 — 0.75 Sum2 0.73 0.89 0.86 0.83 — 1All the correlations were significantly different from 0 (P < 0.0001). 2 Sum = sum of the individual feather condition scores of neck, back, rump, and belly areas. Downloaded from to each other, such as back and rump, back and neck, models. This observation agrees with the SE of the es- were higher than those between the areas that were timated associative genetic variance and the direct-as- further away from each other, such as neck and belly. sociative genetic covariance, which were not significant- The sum of scores of these 4 body areas was used as the ly different from zero (Table 3). http://ps.oxfordjournals.org/ aggregated FCS. NAb Isotype Titers in Beak-Trimmed Genetic Parameters of FCS and Non-Beak-Trimmed Laying Hens

The estimated genetic parameters for FCS in beak- Table 4 shows the average NAb isotypes IgM and IgG trimmed and non-beak-trimmed laying hens, using ei- titers binding KLH in beak-trimmed and non-beak- ther a bivariate traditional linear animal model or a trimmed laying hens. For IgG binding KLH, the titers in at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 bivariate direct-associative effect model, are given in non-beak-trimmed hens were significantly higher than Table 3. Using a traditional linear animal model, simi- in beak-trimmed hens. Furthermore, in beak-trimmed ()σ2 lar and significant additive genetic variance A were laying hens, IgG titers in the nonsurviving birds were found in both populations. The proportion of pheno- significantly higher than in the surviving birds. For typic variance explained by direct genetic variance was 2 2 IgM binding KLH, there was no significant difference denoted as h . The estimated h of FCS was slightly between beak-trimmed and non-beak-trimmed lay- higher in non-beak-trimmed laying hens (0.20, SE = ing hens. There was no significant difference for IgM 0.06) than in beak-trimmed laying hens (0.17, SE = between the surviving and nonsurviving birds within 0.05). 2 beak-trimmed nor non-beak-trimmed group. Overall, Using a direct-associative effect model, direct ()σ AD the IgM and IgG titers were higher in laying hens with 2 and associative genetic variance ()σ for FCS were es- higher FCS, although the difference was not significant. AS timated in beak-trimmed and non-beak-trimmed laying hens, respectively. Total heritable variance relative to 2 Direct and Associative Effect of NAb the phenotypic variance (T ), and genetic correlations Isotypes on FCS between the direct and associative effect (rDS) for FCS were also calculated based on those estimations. In both populations, the direct effects of IgM and Likelihood ratio tests were used to statistically IgG on FCS were not significantly different from zero, compare the traditional linear animal model and the which indicated that an individual’s FCS was not sig- direct-associative effect model. In the beak- nificantly affect by its own isotype titers (Table 5). In trimmed laying hens, final log-likelihoods as reported non-beak-trimmed laying hens, the estimated parame- from traditional linear animal model and from the ters for average titers of IgM and IgG titers of the focal direct-associative effect model were −1,334.52 and individual’s cage mates were not significantly different −1,332.12, respectively. The test statistics were from zero. In beak-trimmed laying hens, the average χ2 =×−2[, 1 332 . 12 −− (, 1 334 .)]., 52 = 4 8 2df which corre- IgG titers of cage mates was a significant factor for the sponds to P = 0.09. In the non-beak-trimmed laying individual’s FCS (P = 0.03). The estimated parameter hens, final log-likelihoods as reported from the tradi- for average IgG was 0.36 (SE = 0.16), which indicated tional linear model and direct-associative effect model that when its cage mates had higher IgG titers, the were −2,258.99 and −2,261.48, respectively. The test individual may have worse feather condition. Average χ2 =×−2[, 2 258 . 99 −− (, 2 261 .)]., 48 = 4 98 statistics were 2df IgM titers of cage mates were not significant (P = 0.83). which corresponds to P = 0.08. Therefore, using the common criterion of P < 0.05, the associative effect for DISCUSSION FCS was not a significant random effect in both beak- trimmed or non-beak-trimmed laying hens, and includ- In the present study, we compared FCS in beak- ing this random effect did not significantly improve the trimmed and non-beak-trimmed crossbred laying hens. FEATHER PECKING AND NATURAL ANTIBODIES IN HENS 779

Table 3. Estimated parameters with SE from traditional and direct-associative animal model for beak-trimmed and non-beak- trimmed laying hens

Traditional linear animal model Direct-associative effect model

Item Non-beak-trimmed Beak-trimmed Non-beak-trimmed Beak-trimmed Log-likelihood1 — — 2.40 2.49 σ2 cage 6.95 ± 0.78 7.18 ± 0.68 6.14 ± 0.88 6.50 ± 0.75 2 σ 2.31 ± 0.79 2.16 ± 0.60 1.46 ± 0.85 1.99 ± 0.77 AD σ −0.30 ± 0.30 0.24 ± 0.24 ADS 2

σ 0.36 ± 0.20 0.07 ± 0.15 Downloaded from AS σ2 e 2.40 ± 0.52 3.11 ± 0.39 2.34 ± 0.60 3.44 ± 0.38 σ2 P 10.76 ± 1.08 12.46 ± 0.71 11.39 ± 0.81 12.22 ± 0.71 σ2 TBV 4.84 ± 2.73 5.03 ± 2.15 h2 or T2 0.20 ± 0.06 0.17 ± 0.05 0.42 ± 0.24 0.41 ± 0.17 rDS −0.41 ± 0.40 0.63 ± 1.06 http://ps.oxfordjournals.org/ σ 1.41 ± 0.75 1.24 ± 0.89 A12_D σ −0.05 ± 0.32 A12_S σ −0.006 ± 0.56 A2_D_1_S σ 0.21 ± 0.59 A1_D_2_S rD 0.63 ± 0.29 0.73 ± 0.48 rS −0.33 ± 2.02 r 0.24 ± 1.41 T at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 1Log-likelihoods for the direct-associative model are expressed as a deviation from those of the traditional linear animal model. In the traditional 2 2 22 2 2 linear animal model, σ is the direct additive genetic variance, σ is the phenotypic variance, σ =+σσ + σ, h2 is the direct heritability, AD P P AD cage e 222= σσ/. σ2 σ h AP The A is the associative genetic variance, and A is the direct-associative genetic covariance. In the direct-associative effect model, D S 22 DS 22 2 the total heritable variance is σ =+−σσ21()nn +− () 1 σ (Bijma et al., 2007). The σ is the phenotypic variance, TBV ADDA SSA P σσ22=+−+()1 σ 2 σ 2 + σ 2; 2 PAn A cagee n is the number of laying hens kept in the same cage, n = 5 in the present study. The T is the total heritable DS 22 2 variance relative to the phenotypic variance: T = σσ/. The r is the genetic correlation between the direct and associative effect. The σ is TBV P D A12_D direct genetic covariance between beak-trimmed and non-beak-trimmed laying hens, and σ is the associative genetic covariance between beak- A12_S trimmed and non-beak-trimmed laying hens. The σ is the genetic covariance between the direct effect of beak-trimmed and the associative A1_D_2_S effect of non-beak-trimmed laying hens. The rD is the genetic correlation between direct effect of beak-trimmed and non-beak-trimmed laying hens, 22 r = σσσ/ . The r is the genetic correlation between associative effect of beak-trimmed and non-beak-trimmed laying hens, DA12___DDD AA1 2 S 22 r = σσσ/ . The r is the genetic correlation between total heritable variance beak-trimmed and non-beak-trimmed laying hens, SA12___SSS AA1 2 T σσ+−()111 +− ()σσ +− ()2 AA12__nnn12___AA21___12 r = DDSDSS (Peeters et al., 2012). T 2 22 2 222 σ ++−21()nnσσσσσ +− () 1 +−21() nn +− () 1 A1_D AA11__DS SDAAA222___DS S

Variance component estimation indicated that there trimmed laying hens was, however, larger than that was relevant heritable variation for FCS in both popu- in non-beak-trimmed laying hens. This indicated that lations using a traditional linear animal model. Using beak-trimmed laying hens still have feather damage a linear animal model combining the direct and asso- problems due to different extent of feather pecking be- ciative effects, there were no significant social genetic havior. Beak trimming only reduces the mortality in- effects. A possible link between the NAb isotype titers stead of preventing the feather pecking propensity. binding KLH and feather pecking behavior was also Among the 4 body areas, the belly received the least investigated. feather damage, whereas the neck and rump received the most. Similar patterns were also found in non-beak- FCS in Beak-Trimmed and Non-Beak- trimmed purebred laying hens (Biscarini et al., 2010b). Trimmed Laying Hens Bright et al. (2006) also found that the rump area of free- ranged laying hens was most damaged. This could be Beak trimming is the removal of the tip of the beak because the back and rump are the areas most exposed of a bird. This treatment is performed as part of an to other cage mates. However, in laying hens raised in overall strategy to reduce feather pecking and canni- floor pens, Bilcík and Keeling (1999) observed that the balism, especially in laying hens. In the present study, belly region became denuded first. The difference may the feather condition in beak-trimmed birds was sig- be caused by the difference in housing systems. Both nificantly better than that in non-beak-trimmed lay- in beak-trimmed and non-beak-trimmed laying hens, ing hens (Table 1). Thus, as expected, beak-trimming the Pearson correlations between the scores of different reduces feather damage. The variation of FCS in beak- body areas were positive and high, especially between 780 SUN ET AL.

Table 4. Number (n), and average titers (±SD) of natural antibody isotypes IgM and IgG binding keyhole limpet hemocyanin of beak-trimmed and non-beak-trimmed laying hens

Population n IgM titers IgG titers Non-beak-trimmed Total 1,169 8.17 ± 1.07 6.71 ± 1.18a Survival1 Surviving 813 8.17 ± 1.06 6.69 ± 1.16 Nonsurviving 356 8.18 ± 1.10 6.77 ± 1.24 Sum FCS2 0–4 6 8.38 ± 1.30 7.92 ± 1.60 5–8 60 7.93 ± 0.99 6.46 ± 1.13 9–12 251 8.06 ± 1.10 6.53 ± 1.19 13–16 377 8.18 ± 1.04 6.81 ± 1.16 17–20 223 8.40 ± 1.05 6.76 ± 1.07

NA 252 8.18 ± 1.10 6.77 ± 1.27 Downloaded from Beak-trimmed Total 1,555 8.24 ± 1.05 6.61 ± 1.22b Survival1 Surviving 1,447 8.24 ± 1.05 6.59 ± 1.22A Nonsurviving 108 8.30 ± 1.03 6.86 ± 1.16B Sum FCS2 0–4 89 8.30 ± 1.06 6.35 ± 1.42 5–8 436 8.23 ± 0.97 6.58 ± 1.24 9–12 594 8.20 ± 1.09 6.62 ± 1.19 13–16 308 8.30 ± 1.10 6.64 ± 1.22 http://ps.oxfordjournals.org/ 17–20 97 8.32 ± 0.98 6.80 ± 1.19 NA 31 8.19 ± 1.00 6.59 ± 1.13 a,bThe total IgG titers in beak-trimmed and non-beak-trimmed laying hens is significantly different with P < 0.05. A,BThe IgG titers for the survived and nonsurviving beak-trimmed laying hens is significantly different with P < 0.05. 1Surviving indicated the birds survived until the end of the observation period (83 wk of age), and nonsurviving indicated the birds died between 24 wk of age and the end of the observation period. 2Sum FCS = the sum of feather condition score of 4 body areas. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 the areas that were close to each other, back and rump, cannibalism frequencies in the brooding period of 4 to for example (Table 2). Feather condition scoring as a 11 wk of age (Hughes and Duncan, 1972), and around behavioral measurement at the individual level is labor- the onset of egg laying at 20 wk of age (McKeegan and intensive and time-consuming work. Given the extent Savory, 1998). It could be speculated that the birds of damage and correlations between the FCS for dif- that died before the scoring had poor feather condition. ferent body areas, it may be efficient and sufficient to Peeters et al. (2012) detected a substantial associative only score one representative body area, such as the effect for survival days in non-beak-trimmed laying back or rump. hens. This together with our analysis on FCS supports The survival of beak-trimmed laying hens was signifi- the argument that severe feather pecking contributed cantly higher than non-beak-trimmed laying hens (Sun to the mortality in non-beak-trimmed laying hens. et al., 2013b). This suggests that severe feather damage maybe a causative factor for the mortality afterward. Direct and Associative Effect for FCS However, neither in the beak-trimmed nor in the non- beak-trimmed laying hens, a significant difference was Similar to other interaction phenotypes, such as sur- detected for the FCS between nonsurviving and survival days in laying hens (Ellen et al., 2008; Peeters et viving birds. This might rest not only on the limited al., 2012), FCS of an individual hen is affected by both number of nonsurviving hens after 53 wk of age (when the individual and its conspecifics when kept in groups. the feather scoring was performed), but also on the fact Ignoring the social interaction among individuals that that the laying hens that died before 53 wk did not generate additional heritable variation may result in receive a FCS [in non-beak-trimmed laying hens, a high biased genetic parameter estimation. In the present death rate was observed from 18 to 26 wk of age, and study, we estimated the genetic parameters for FCS in around 35 wk of age (Sun et al., 2013b)]. Feather dam- beak-trimmed and non-beak-trimmed laying hens, usage due to feather pecking cumulates over time. Feather ing the traditional linear animal model, and direct-asso- pecking has been observed as early as 1 d after hatching ciative effect model, respectively. Using the traditional (Roden and Wechsler, 1998) with sudden increases of linear model, the heritability (h2) for FCS in nonbeak-

Table 5. Parameter estimates with SE and the significance level (P-value) of the fixed effects of individual IgG titers, individual IgM titers, average IgG titers of cage mates, and average IgM titers cage mates, on individual feather condition score

Non-beak-trimmed Beak-trimmed

Fixed effect Estimate (SE) P-value Estimate (SE) P-value IgG 0.09 (0.07) 0.20 −0.07 (0.06) 0.21 IgM 0.09 (0.10) 0.38 −0.08 (0.08) 0.32 IgG of cage mates −0.15 (0.19) 0.43 0.36 (0.16) 0.03 IgM of cage mates 0.05 (0.19) 0.80 0.04 (0.18) 0.83 FEATHER PECKING AND NATURAL ANTIBODIES IN HENS 781 trimmed laying hens was estimated to be 0.20 (SE = effect of individual NAb on individual FCS), and the 0.06), and in beak-trimmed laying hens was estimated relationship between performing feather pecking and to be 0.17 (SE = 0.05). Using a direct-associative effect NAb isotype titers (associative effect of individual NAb model, the estimated associative genetic variance was on cage mates’ FCS), a mixed model with either the substantial, causing the total heritable variation (T2) focal individual’s or cage mates’ average isotype titers to be 2-fold greater than ordinary heritability (Table as fixed effects was fitted for FCS of the individual in 3). Nevertheless, the estimated associative genetic vari- beak-trimmed (model 3) and non-beak-trimmed laying ance was not significantly different from zero. This also hens (model 4), respectively. agrees with the results from the likelihood ratio test for In both populations, the direct effects of IgM and the significance of random associative genetic effect (P IgG for individual FCS were not significant (Table 5). = 0.09 in beak-trimmed birds, and P = 0.08 in non- This indicated that the individual’s own NAb isotype Downloaded from beak-trimmed birds). Hence, our results suggest that titers may not affect its FCS, although Biscarini et al. associative genetic effects may be important in those (2010a) detected a link between receiving feather peck- populations, but lacked the statistical power to accu- ing and the individual’s innate and adaptive immune rately estimate those effects, probably due to a limited parameters. number of records (Table 1). T. Brinker (Wageningen A link between performing feather pecking and the http://ps.oxfordjournals.org/ University, Wageningen, the Netherlands, personal immune parameters was detected (Buitenhuis et al., communication) showed that social effects had a sub- 2004; Parmentier et al., 2009; Biscarini et al., 2010b; stantial effect on the total heritable variation of FCS in Hughes and Buitenhuis, 2010; Brunberg et al., 2011). purebred non-beak-trimmed laying hens (n = 6,276 and In non-beak-trimmed laying hens, a significant associa- 6,916 for 2 purebred layer lines). Using large data sets, tive effect of NAb titers on feather damage was not Ellen et al. (2008; n = 3,988 to 6,916 for different pure- detected. However, in beak-trimmed laying hens, the

bred layer lines) and Peeters et al. (2012; n = 15,012 parameter estimate for average titers of IgG of the cage at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 for crossbred laying hens) also found large and strongly mates was 0.34 (SE = 0.16, P = 0.03). This indicated significant social effects on survival time in non-beak- that when the cage mates have higher IgG titers, the trimmed laying hens. However, a suggestive heritable individual may have higher suffer from more feather variation from the associative effect for FCS was still damage. However, multiple hypothesis testing will in- indicated from the comparison of analysis with 2 mod- crease the false-positive results. As a statistical method els in the present study. Ignoring the associative effect used to correct for multiple comparisons, the false dis- and the genetic correlation between direct and asso- covery rate adjusted P-value was 0.43. This suggested ciative effect may induce underestimation of heritable that the relationship between the individual FCS and variation for FCS and inappropriate breeding strategies the average IgG titers of its cage mate may need to be for less feather damage in laying hens. Behavioral mea- replicated for further confirmation. The relationships surement on an individual level is a real effort. Hence, were also fitted for the 2 populations together, by add- our results also illustrate the difficulty of collecting ing beak treatment as an extra fixed effect. Still, the sufficient data to accurately estimate genetic param- relationships were not significant (data not shown). eters for behavioral traits. As we discussed before, to In our previous study about the relationship between get proper estimation for genetic parameters for FCS, NAb isotype titers and survival in laying hens, NAb enlarging the numbers of birds involved maybe more isotypes especially IgM was shown to be a protective valuable than scoring for multiple body areas, because factor for health-related survival (Sun et al., 2011), and of the high and positive correlation between the FCS therefore a promising trait to be bred for higher sur- of closely-located body areas. Further improvement of vival of the population. The nonsignificant relationship statistical power may come from optimizing the cage between NAb isotype titers (both IgG and IgM) and composition, as results in Bijma (2010) indicate that individual FCS and nonsignificant relationship between the SE of the estimated associative genetic variance isotype IgM titers of the cage mates’ and the individual is minimized when each cage consists of members of 2 FCS as shown in the present study suggest that the families, each family contributing half. improvement of individual IgM levels does not result in more feather damage. However, the suggested rela- Direct and Associative Effect of NAb tionship between individual FCS and the cage mates’ Isotype IgM and IgG Titers Binding average IgG isotype titers needs further confirmation to determine the potential greater feather damage caused KLH on FCS by the improvement of IgG levels. Several studies showed links between feather pecking and the immune system (Buitenhuis et al., 2004; Conclusions Biscarini et al., 2010b). The NAb is claimed to be an important parameter of the immune system (Star et al., To the best of our knowledge, this is the first time 2007). To investigate the possible relationship between that FCS of the laying hens was modeled by a direct receiving feather pecking and NAb isotype titers (direct and associative effect model, and the first time that the 782 SUN ET AL. direct and associative effects of NAb isotype titers on Blokhuis, H. J., and P. R. Wiepkema. 1998. Studies of feather peck- individual FCS were investigated. The estimated as- ing in poultry. Vet. Q. 20:6–9. Bright, A., T. A. Jones, and M. S. Dawkins. 2006. A non-intrusive sociative genetic variance for FCS was substantial, but method of assessing plumage condition in commercial flocks of not significantly different from zero, probably due to laying hens. Anim. Welf. 15:113–118. the limited number of records. Results suggested, how- Brunberg, E., P. Jensen, A. Isaksson, and L. Keeling. 2011. Feather ever, that including associative effects in the model, pecking behavior in laying hens: hypothalamic gene expression in birds performing and receiving pecks. Poult. Sci. 90:1145–1152. both in the beak-trimmed and non-beak-trimmed lay- Buitenhuis, A. J., T. B. Rodenburg, P. H. Wissink, J. Visscher, P. ing hens, is important to estimate genetic parameters Koene, H. Bovenhuis, B. J. Ducro, and J. J. van der Poel. 2004. for FCS. Although the effects of immunity on feather Genetic and phenotypic correlations between feather pecking be- pecking behavior were suggested by several studies. havior, stress response, immune response, and egg quality traits

in laying hens. Poult. Sci. 83:1077–1082. Downloaded from The NAb isotypes titers did not show significant direct Chaouloff, F. 2000. Serotonin, stress and corticoids. J. Psychophar- effects or an associative effect for individual FCS in the macol. 14:139–151. present study. However, further studies are needed to Cheng, H. W., P. Singleton, and W. M. Muir. 2003. Social stress differentially regulates neuroendocrine responses in laying hens: I. confirm the suggestive relationship between IgG titers Genetic basis of dopamine responses under three different social and feather pecking. conditions. Psychoneuroendocrinology 28:597–611. El-Lethey, H., B. Huber-Eicher, and T. W. Jungi. 2003. Exploration http://ps.oxfordjournals.org/ of stress-induced immunosuppression in chickens reveals both ACKNOWLEDGMENTS stress-resistant and stress-susceptible antigen responses. Vet. Immunol. Immunopathol. 95:91–101. Ellen, E. D., J. Visscher, J. A. van Arendonk, and P. Bijma. 2008. This research was part of a joint project of ISA and Survival of laying hens: genetic parameters for direct and associa- Wageningen University on “The genetics of robustness tive effects in three purebred layer lines. Poult. Sci. 87:233–239. in laying hens,” which was financially supported by Gentle, M. J., and L. N. Hunter. 1991. Physiological and behavioural SenterNovem (Den Haag, the Netherlands). The author responses associated with feather removal in Gallus gallus var

domesticus. Res. Vet. Sci. 50:95–101. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Yanyan Sun was financially supported by China Schol- Gilmour, A. R., B. J. Gogel, B. R. Cullis, and R. Thompson. 2006. arship Council (Beijing, China). The contribution of Pi- ASReml User Guide Release 2.0. VSN International Ltd., Hemel ter Bijma was supported by the foundation for applied Hempstead, UK. research of the Dutch science council (the Netherlands Griffing, B. 1967. Selection in reference to biological groups. I. Indi- vidual and group selection applied to populations of unordered Organization for Scientific Research, NWO-STW). The groups. Aust. J. Biol. Sci. 20:127–139. authors thank ISA B.V., the layer breeding division Hill, J. A. 1986. Egg production in alternative systems—A review of Hendrix Genetics, for providing birds and facilities of recent research in the UK. Research and Development in and the animal caretakers for taking good care of the Agriculture. 3:6. Hughes, A. L., and A. J. Buitenhuis. 2010. Reduced variance of gene laying hens. Mike Nieuwland and Laura Star from Ad- expression at numerous loci in a population of chickens selected aptation Physiology Group, Wageningen University for high feather pecking. Poult. Sci. 89:1858–1869. (Wageningen, the Netherlands) are acknowledged for Hughes, B. O., and I. J. Duncan. 1972. The influence of strain and blood sampling and technical assistance in ELISA. Bas environmental factors upon feather pecking and cannibalism in fowls. Br. Poult. Sci. 13:525–547. Rodenburg, Patrick Wissink, and Koen Uitdehaag from Johnsen, P. F., K. S. Vestergaard, and G. Nørgaard-Nielsen. 1998. Animal Breeding and Genomics Centre, Wageningen Influence of early rearing conditions on the development of feath- University (Wageningen, the Netherlands) are acknowl- er pecking and cannibalism in domestic fowl. Appl. Anim. Be- edged for their effort in feather scoring. hav. Sci. 60:25–41. Keeling, L. J. 1995. Feather pecking and cannibalism in layers. Int. J. Poult. Sci. 6:46–50. Kirkpatrick, M., and R. Lande. 1989. The evolution of maternal REFERENCES characters. Evolution 43:485–503. Kjaer, J. B., and K. S. Vestergaard. 1999. Development of feather Avrameas, S. 1991. Natural autoantibodies: From ‘horror autotoxi- pecking in relation to light intensity. Appl. Anim. Behav. Sci. cus’ to ‘gnothi seauton’. Immunol. Today 12:154–159. 62:243–254. Bijma, P. 2010. Estimating indirect genetic effects: Precision of esti- McKeegan, D. E., and C. J. Savory. 1998. Behavioural and hormonal mates and optimum designs. Genetics 186:1013–1028. changes associated with sexual maturity in layer pullets. Br. Bijma, P. 2011. A general definition of the heritable variation that Poult. Sci. 39(Suppl.):S6–S7. determines the potential of a population to respond to selection. Moore, A. J., E. D. Brodie III, and B. W. Jason. 1997. Interacting Genetics 189:1347–1359. phenotypes and the evolutionary process: I. Direct and indirect Bijma, P., W. M. Muir, and J. A. Van Arendonk. 2007. Multilevel genetic effects of social interactions. Evolution 51:1352–1362. selection 1: Quantitative genetics of inheritance and response to Muir, W. M. 2005. Incorporation of competitive effects in forest tree selection. Genetics 175:277–288. or animal breeding programs. Genetics 170:1247–1259. Bilcík, B., and L. J. Keeling. 1999. Changes in feather condition in Parmentier, H. K., T. B. Rodenburg, G. De Vries Reilingh, B. Beer- relation to feather pecking and aggressive behaviour in laying da, and B. Kemp. 2009. Does enhancement of specific immune hens. Br. Poult. Sci. 40:444–451. responses predispose laying hens for feather pecking? Poult. Sci. Biscarini, F., H. Bovenhuis, J. A. van Arendonk, H. K. Parmentier, 88:536–542. A. P. Jungerius, and J. J. van der Poel. 2010a. Across-line SNP Peeters, K., T. T. Eppink, E. D. Ellen, J. Visscher, and P. Bijma. association study of innate and adaptive immune response in lay- 2012. Indirect genetic effects for survival in domestic chickens ing hens. Anim. Genet. 41:26–38. (Gallus gallus) are magnified in crossbred genotypes and show a Biscarini, F., H. Bovenhuis, J. van der Poel, T. B. Rodenburg, A. P. parent-of-origin effect. Genetics 192:705–713. Jungerius, and J. A. van Arendonk. 2010b. Across-line SNP asso- Roden, C., and B. Wechsler. 1998. A comparison of the behaviour ciation study for direct and associative effects on feather damage of domestic chicks reared with or without a hen in enriched pens. in laying hens. Behav. Genet. 40:715–727. Appl. Anim. Behav. Sci. 55:317–326. FEATHER PECKING AND NATURAL ANTIBODIES IN HENS 783

Rodenburg, T. B., H. Komen, E. D. Ellen, K. A. Uitdehaag, and Sun, Y., H. K. Parmentier, K. Frankena, and J. J. van der Poel. J. A. M. van Arendonk. 2008. Selection method and early-life 2011. Natural antibody isotypes as predictors of survival in lay- history affect behavioural development, feather pecking and ing hens. Poult. Sci. 90:2263–2274. cannibalism in laying hens: A review. Appl. Anim. Behav. Sci. Uitdehaag, K., H. Kornen, T. B. Rodenburg, B. Kemp, and J. van 110:217–228. Arendonk. 2008. The novel object test as predictor of feather Star, L., K. Frankena, B. Kemp, M. G. Nieuwland, and H. K. Par- damage in cage-housed Rhode Island Red and White Leghorn mentier. 2007. Natural humoral immune competence and survival laying hens. Appl. Anim. Behav. Sci. 109:292–305. in layers. Poult. Sci. 86:1090–1099. Wolf, J. B. 2003. Genetic architecture and evolutionary constraint Sun, Y., F. Biscarini, H. Bovenhuis, H. K. Parmentier, and J. J. van when the environment contains genes. Proc. Natl. Acad. Sci. der Poel. 2013a. Genetic parameters and across-line SNP associa- USA 100:4655–4660. tions differ for natural antibody isotypes IgM and IgG in laying Wolf, J. B., E. D. Brodie Iii, J. M. Cheverud, A. J. Moore, and M. hens. Anim. Genet. 44:413–424. J. Wade. 1998. Evolutionary consequences of indirect genetic ef-

Sun, Y., E. D. Ellen, H. K. Parmentier, and J. J. van der Poel. fects. Trends Ecol. Evol. 13:64–69. Downloaded from 2013b. Genetic parameters of natural antibody isotypes and survival analysis in beak trimmed and non-beak trimmed crossbred laying hens. Poult. Sci. 92:2024–2033. http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Comparative proteomic analysis of the breast muscle response to chronic corticosterone administration in broiler chickens showing long or short tonic immobility

Wenyan Fu,* Yujing Duan ,* Song Wang,* Yingdong Ni ,*1 R. Grossmann ,† and Ruqian Zhao *

* Key Laboratory of Animal Physiology and Biochemistry, Ministry of Agriculture, Nanjing Agricultural University, Downloaded from Nanjing 210095, China; and † Department of Functional Genomics and Bioregulation, Institute of Animal Genetics, FLI, Mariensee, 31535 Neustadt, Germany

ABSTRACT Broilers of the same genetic origin were muscle damage. Real-time PCR results indicated that classified as short or long tonic immobility duration (STI expression of these proteins is transcriptionally and http://ps.oxfordjournals.org/ and LTI, respectively) phenotypes and treated chroni- posttranscriptionally regulated. Protein synthesis ca- cally with vehicle (control) or corticosterone (CORT) pacity, estimated by the RNA-to-protein ratio, was sig- dissolved in drinking water between 27 and 42 d of age. nificantly lower in the breast muscle of CORT-treated Differential expression of proteins and mRNA was ex- broilers than in untreated control broilers. The level of amined using 2-dimensional gel electrophoresis and re- Leu, Gly, and Ser in serum was significantly higher in al-time PCR to elucidate the mechanism behind the se- CORT-treated broilers than in the control birds. There- vere retardation of broiler breast muscle growth caused fore, we conclude that CORT treatment retards the by LTI and CORT. The majority of the 13 proteins growth of skeletal muscle by suppressing protein syn- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 found to be differentially expressed in breast muscle of thesis and augmenting protein catabolism, paralleling STI and LTI broilers are involved in either glycolysis the response to severe stress and muscle damage, and (5 proteins) or myofilament formation (5 proteins). Of the negative effect of LTI on muscle growth is likely the 16 proteins differentially expressed in breast muscle mediated through glucose metabolism. No interaction following CORT treatment, 6 are structural proteins, 5 was observed between CORT and tonic immobility af- are categorized as cellular defense and stress proteins, fecting growth performance or any parameter examined and 3 (pyruvate kinase, l-lactate dehydrogenase, and in the current study. creatine kinase) are involved in responses to stress and Key words: broiler , corticosterone , proteomics , skeletal muscle, tonic immobility 2014 Poultry Science 93 :784–793 http://dx.doi.org/ 10.3382/ps.2013-03456

INTRODUCTION coping capability and well-being in response to stressors (Leonard, 2006). Corticosterone (CORT), a ter- Tonic immobility (TI) is a valuable biological pa- minal hypothalamus-pituitary-adrenal gland axis hor- rameter for investigating whether emotional reactivity mone, has been used to mimic the response to stressors modulates chronic stress-related behavioral and physi- (Malheiros et al., 2003). Oral administration of CORT ological dysfunction in animals (Campo and Redondo, through drinking water for 2 wk has been used to es- 1996; Campo et al., 2000; Calandreau et al., 2011). The tablish an experimental model of chronic stress in broil- duration of the TI reaction is a reliable indicator of er chickens (Post et al., 2003; Shini et al., 2009). fearfulness (Jones et al., 1991). Due to its strong and It is generally accepted that long TI duration (LTI) clearly distinguishable reaction, poultry is the most and chronic stress negatively affect growth in verte- commonly used animal model for studying TI induc- brates. The BW of LTI line birds is lower than that of tion (Gallup et al., 1977). Alteration of the stress axis short TI duration (STI) birds (Minvielle et al., 2002). (hypothalamus-pituitary-adrenal gland) is thought to Chronically elevated CORT levels retard the growth of be the most common means of controlling an animal’s young animals (Morici et al., 1997; Glennemeier and Denver, 2002; Hayward and Wingfield, 2004). Lin et al. (2004a,b, 2006) showed that both short- and long-term © 2014 Poultry Science Association Inc. Received July 1, 2013. CORT administration severely retards skeletal muscle Accepted December 16, 2013. growth in birds in a dose- and age-dependent manner 1 Corresponding author: [email protected] through suppression of protein synthesis and augmen-

784 COMPARATIVE PROTEOMICS IN BROILERS 785 tation of protein catabolism. However, the mechanism STI and LTI phenotypes (Wang et al., 2013). Chick- underlying the negative effect of LTI on growth remains ens demonstrating 2 extremes (shortest and longest TI unclear. duration) were registered. Eighty chickens showing the It is unclear how nonphysical entities, such as the du- shortest TI duration (29.6 ± 2.3 s) and 80 scoring the ration of TI and level of CORT, could engage in cross- longest duration (246.2 ± 26.8 s) were classified into talk. Feeding CORT to laying hens has been shown to STI and LTI groups, respectively. induce stress behavior, as indicated by a longer TI reac- Chicken of the STI and LTI groups were allocated tion and reduction in BW gain (El-Lethey et al., 2003). into control (CON) and CORT-treated subgroups to Embryonic exposure to CORT leads to behavioral and four 2 × 2.7 m2 pens (10 birds in each pen). From growth deficits in chicken, and posthatch handling 27 d to 42 d, chickens in the CORT groups of both increases the duration of TI in CORT-treated birds STI and LTI phenotypes (40 per phenotype) were sup- Downloaded from (Janczak et al., 2007). Long tonic immobility broilers plied water supplemented with 5 mg/L of corticoste- reportedly show a greater CORT response to crating rone (C2505, Sigma, St. Louis, MO), whereas those in and heat challenge than STI birds (Zulkifli et al., 2009). CON groups (40 from LTI and 40 from STI) were sup- A consistent trend toward greater adrenocortical acti- plied water supplemented with an equivalent volume vation was observed in high-fear compared with low- of the solvent (absolute ethanol). Each broiler chicken http://ps.oxfordjournals.org/ fear hens (Beuving et al., 1989). In contrast, Hazard consumed approximately 0.2 to 0.3 L of water per day et al. (2005, 2008) reported that Japanese quail with on average during the experimental period (d 27–41). the LTI phenotype demonstrate lower CORT responses Therefore, the daily intake of corticosterone and etha- under restraint stress. Nevertheless, these results are nol was about 1.0 to 1.5 mg and 0.2 to 0.3 mL, respec- suggestive of cross-talk between the TI response and tively. endogenous CORT level. Our previous results (Wang et On d 42, all birds were fasted overnight and weighed

al., 2013) showed that the growth performance and bas- before slaughter. The blood was collected from the at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 al protein synthesis of broiler chickens with inborn long jugular vein and serum samples were stored at −20°C or short TI duration differ, but that chronic CORT until analyzed. The breast muscles was removed from administration neither highlights nor masks such in- the body, quickly put into liquid nitrogen, and stored nate differences. Studies on the mechanism underlying at −80°C for further analysis. the growth suppression of LTI and chronic stress will benefit welfare and health of chickens as well as pro- Serum Free Amino Acids Analysis duction efficiency. The current study is the first report demonstrating differences in the protein synthesis ca- The serum concentration of free amino acids was pacity and protein expression profile in breast muscle determined according to Sunde et al. (2003). In brief, that suggest the negative effects of LTI and chronic the serum supernatant was deproteinized with a Milli- stress on muscle development in birds are mediated by pore Ultrafree-MC 10,000 NMWL filter unit (Millipore, different mechanisms. Bedford, MA) at 5,000 × g and 4°C for 30 min. The filtrate of the plasma supernatant was derivatized with Waters ACCQ·Taq reagent, and analyzed with a buf- MATERIALS AND METHODS fer gradient system using the Alliance HPLC system. Separation of the free amino acids was performed with The experiment was conducted following the guide- a Waters Nova-Pak C18 column (60 Å, 4 μm, 3.9 × 300 lines of the Animal Ethics Committee at Nanjing Agri- mm, Waters, Milford, MA). cultural University. The study was approved by Animal Ethics of Nanjing Agricultural University. The sampling procedures complied with the Guidelines on Ethi- Protein Analysis cal Treatment of Experimental Animals (2006) no. 398 Protein Extraction. A 0.1-g muscle sample was set by the Ministry of Science and Technology, China, placed in liquid nitrogen and ground thoroughly to a and the Regulations regarding the Management and very fine powder with a mortar and pestle. The powder Treatment of Experimental Animals (2008) no. 45 set (about 100 mg) was transferred to sterile tubes con- by the Jiangsu Provincial People’s Government. taining a 1-mL sample preparation buffer [7 M urea, 2 M thiourea, 4% (wt/vol) 3-[(3-cholamidopropyl)- Birds and Management imethylammonio]-1-propanesulfonate, 1% (vol/vol) dithiothreitol, 0.5% (vol/vol) immobilized pH gradient Tonic immobility test is commonly used to assess the buffer pH 4–7, protease inhibitor cocktail purchased fear, and a long duration of TI is an indication for high from Roche, Mannheim, Germany]. The mixture was levels of fearfulness. A TI phenotype may serve as a then incubated for 60 min at room temperature with selection target for a low-stress response broiler line. In occasional vortexing, and centrifuged at 12,000 × g for the current study, broiler breeder chickens (Ross 308) 60 min at 4°C. The supernatants were collected and were tested for TI to establish 2 segregated groups of stored at −80°C until analysis. The protein concentra- 786 FU ET AL. tion was measured by the Bradford assay using BSA RNA Isolation, cDNA Synthesis, and Real-Time (A7030, Sigma) as the standard. PCR. Total RNA was isolated from muscle tissues Two-Dimensional Gel Electrophoresis. For iso- using Trizol Reagent (catalog no. 15596–026, Invit- electric focusing, a sample volume equivalent to 300 rogen, Carlsbad, CA) according to manufacturer’s in- μg of protein extract was added to rehydration buf- structions. Concentration and quality of the extract- fer (BioRad, Hercules, CA). The samples were loaded ed RNA were measured using a NanoDrop ND-1000 onto 17-cm, pH 4 to 7 IPG strips (BioRad) and each Spectrophotometer (Thermo Scientific). Prime-Script protein sample was conducted in triplicate. The IPG 1st Strand cDNA Synthesis Kit (catalog no. D6110A, strips were passively rehydrated for 13 h and isoelectric Takara, Dalian, China) was used to synthesize cDNA focusing was performed at 250 V for 1 h, 500 V for from 2 μg of total RNA from each sample according 1 h, 1,000 V for 5 h, followed by linearly ramping to to the manufacturer’s instructions. Two microliters of Downloaded from 10,000 V over 3 h and then holding at 10,000 V until diluted cDNA (1:20, vol/vol) was used in each real-time 60,000 V-h had been accumulated. Prior to the second PCR. Primer sequences are shown in Table 1. Real- dimension, the strips were equilibrated for 15 min in 6 time PCR was performed with Mx3000P (Stratagene, M urea, 30% glycerol, 2% SDS, 50 mM Tris pH 8.8, 1% La Jolla, CA). Expression of the housekeeping gene (vol/vol) DTT, and then for an additional 15 min in β-actin mRNA in breast muscle was not changed by http://ps.oxfordjournals.org/ the same buffer except that DTT was replaced by 4% the TI or CORT treatment. Thus, β-actin gene expres- iodoacetamide. After equilibration, the second dimension was used as the reference gene in relative real-time sion was run on a 12% polyacrylamide SDS gel using PCR. The specificity of amplification was determined the Ettan DALTSix (GE Healthcare, Pittsburgh, PA). by melting curve analysis and agarose gel electrophore- The analytical gels were stained with colloidal Coo- sis. The PCR products were sequenced to validate the massie Blue G-250. identity of the amplicons.

Image and Protein Identification. Stained gels were at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 scanned and analyzed using PD Quest v710 software Statistical Analyses (BioRad). After alignment, spots between gels were first automatically matched. The matched spots were All data are presented as means ± SEM and were then reexamined manually to ensure accuracy. Gener- analyzed using 2-way ANOVA with SPSS 18.0 software ally, only those spots with expression differences >2, (StatSoft Inc., Tulsa, OK). We evaluated the effects of were chosen for further analysis (Doherty et al., 2004). CORT and TI, as well as their interactions. Differences Protein identification was analyzed using a fuzzy log- were considered significant when P < 0.05. The method ic feedback control system (Ultraflex II MALDI TOF/ of 2−ΔΔCt was used to analyze the real-time PCR data TOF system Bruker, Bremen, Germany). Mass spec- (Livak and Schmittgen, 2001). The mRNA expression trophotometry fingerprinting data searches were per- was expressed as the fold change relative to that of STI formed by search engines of MS-fit1 against the NCBInr group. database (http://blast.ncbi.nlm.nih.gov/Blast.cgi) in the taxa of Gallus gallus with the parameter sets of RESULTS trypsin digestion, 2 missed cleavages, complete modification of iodoacetamide, partial modification of Met Protein Synthesis Capacity and Levels oxidation, protein mass (±20% of the observed protein of Serum Amino Acids mass), and a mass tolerance for mono-isotopic data of 100 ppm. Protein identification was assigned when Protein content in breast muscle was not affected by there were at least 4 matching peptides and >20% se- the TI phenotype or CORT administration. However, quence coverage. RNA content in breast muscle was significantly reduced by CORT (P < 0.001) but not TI. The reduction in Validation of Protein and Gene Expression RNA content in turn resulted in a significant suppression (P < 0.05) of the apparent capacity for protein Western Blot Analysis. Western blotting was used synthesis in CORT-treated broilers as estimated by the to validate the main differentially expressed proteins RNA-to-protein ratio (Table 2). and carried out using a tank system (Bio-Rad). The As shown in Table 3, of the 8 essential amino acids, protein concentration was determined with Pierce BCA only the concentration of Leu was significantly elevated Protein Assay Kit (catalog no. 23225, Thermo Scientif- (P < 0.05) in serum following CORT administration. In ic, Hudson, NH). Western blot analysis for heat shock addition, CORT administration resulted in a significant cognate 70 (HSC70; catalog no. ab19136, Abcam, decrease of nonessential amino acid Gly (P < 0.05) and Cambridge, UK; diluted 1:10,000), desmin (catalog no. a significant increase of Ser (P < 0.05) in serum. Of the ab8976, Abcam; diluted 1:1,000), and glyceraldehyde 18 amino acids detected in serum, only Cys content dif- 3-phosphate dehydrogenase (GAPDH; catalog no. fered significantly between the TI phenotypes, and was AP0066, Bioworlde, St. Louis, MO; diluted 1:500) was significantly higher in LTI broilers (P < 0.01). Amino conducted according to the recommended protocols acid levels in serum were not affected by any interac- provided by the manufacturers. tions between TI and CORT (P > 0.05). COMPARATIVE PROTEOMICS IN BROILERS 787

Table 1. Primer sequences for the PCR amplification of specific genes

GenBank PCR product Target gene accession no. Primer sequence1 (5c–3c) (bp) Myosin light chain 1f NM_001044632.1 F: TGACAGGACTGGTGATGCC 108 R: GGGTTGCCCAGGATCTTGT Actin, α 1, skeletal muscle NM_204127.1 F: ACTTTGCCAGATGCCGACA 123 R: TGGGGTGATTGTAGTGTAAGGG Pyruvate kinase muscle isozyme NM_205469.1 F: CGACTCCGAGCCAACCATT 147 R: TGATACTCGTGGGTGCCGT l-Lactate dehydrogenase A chain X53828.1 F: GGGTGGATTGTTGGAGAGCA 148 R: CCACCACCTGCTTGTGAACCT

Creatine kinase M-type NM_205507.1 F: CACCGACCTCAACCACG 173 Downloaded from R: GGCTGTTCAGGGCTTCC Heat shock protein 70 GU980869.1 F: TAACACCACCATTCCCACCA 96 R: GCCCTCTCACCTTCATACACCT Desmin AB011672.1 F: CCGCTTCGCCAACTACATT 212 R: GCAGGTCATCTAGCAGGTTGTC Apolipoprotein A-I precursor M25559.1 F: TGGACCGCATTCGGGATA 137 R: CAGCGTGTCCAGGTTGTCA http://ps.oxfordjournals.org/ Serum albumin precursor NM_205261.1 F: GTGATGTCGGTGCTTGCCT 173 R: CAGCGTGTCCAGGTTGTCA Fatty acid-binding protein, heart NM_001030889.1 F: ATGGTGGAAGCGTTCGTGG 124 R: TGGTGGTGGGTTTGGTCAG α-Actin NM_205323.1 F: AGAGAGGGAAAGGAGGAT 138 R: GATGGGTTGAAGATAGCAG Actinin α 2 NM_001031229.1 F: AGCAGTTTTCCCTTCCAT 249 R: TTCAGTGGTGCTTCAGTCA Glucose-6-phosphate isomerase NM_001006128.1 F: AATGCTGATTGAACTGGC 133

R: TGGGGTATTGGAACGATT at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Glyceraldehyde-3-phosphate dehydrogenase NM_204305.1 F: AAGGCGAGATGGTGAAAG 241 R: CGCTCCTGGAAGATAGTGAT 6-Phosphofructokinase AB061205.1 F: CGATACCCCTGCCTGTGT 222 R: GCCAATGTGTGACCGCTCT Calsequestrin M58048.1 F: TGATGGGAAAGACCGAGTG 138 R: TCCGTCATCTGGAACTGC Pyruvate kinase muscle isozyme NM_205469.1 F: CGACTCCGAGCCAACCATT 147 R: TGATACTCGTGGGTGCCGT 1F = forward primer; R = reverse primer.

Proteomic Analysis (Table 4). The differentially expressed proteins were grouped into 4 categories: (1) myofilament (5 proteins), Complete separation of protein spots was apparent in (2) metabolism (5 proteins), (3) transport (1 protein), the two-dimensional gel electrophoresis (2-DE) analy- and (4) miscellaneous (2 proteins). Of the 20 protein ses of the breast muscle of STI and LTI broiler chick- spots analyzed from the gels of CON and CORT-treat- ens (Figure 1), as well as control and CORT chickens ed broilers, 16 proteins were identified by matching (Figure 2). Coomassie blue staining of the gels revealed the matrix-assisted laser desorption ionization–time of about 400 spots, with the molecular weight of most flight mass spectrometry peptide data to chicken pro- proteins falling between 14.4 and 200 kDa. tein sequences in the NCBInr database (Table 5). The A total of 15 protein spots exhibiting differential differentially expressed proteins were grouped into 5 abundance in the breast muscle of LTI and STI broil- categories: (1) myofilament (4 proteins), (2) cytoskel- ers were subjected to subsequent matrix-assisted laser eton (1 protein), (3) cellular defense and stress response desorption ionization-time of flight mass spectrometry (5 proteins), (4) metabolism (3 proteins), and (5) trans- analysis, resulting in the identification of 13 proteins port (3 proteins).

Table 2. Effect of tonic immobility (TI) and corticosterone (CORT) on RNA, protein contents, and the capacity of protein synthesis in muscle1

Control2 CORT P-value

Item STI LTI STI LTI TI CORT TI × CORT RNA (mg/g) 0.67 ± 0.02 0.70 ± 0.04 0.45 ± 0.07 0.30 ± 0.07 0.298 0.000 0.149 Protein content (100 mg/g) 1.18 ± 0.19 0.89 ± 0.08 0.92 ± 0.11 0.95 ± 0.14 0.367 0.469 0.262 CS3 (mg of RNA/g of protein) 6.10 ± 1.07 7.65 ± 0.49 5.21 ± 1.02 3.40 ± 0.92 0.886 0.012 0.079 1Data are expressed as means ± SEM and were analyzed using the general linear model (univariate) to evaluate the effects of CORT and TI, as well as their interactions. Differences were considered significant when P < 0.05, n = 6/group. 2STI = short TI; LTI = long TI. 3CS = capacity of protein synthesis. 788 FU ET AL.

Table 3. Effect of tonic immobility (TI) and corticosterone (CORT) on the level of amino acids in serum (arbitrary units)

Control1 CORT P-value

Item STI LTI STI LTI TI CORT TI × CORT Essential amino acid Lys 1.00 ± 0.14 0.59 ± 0.19 0.77 ± 0.12 0.58 ± 0.13 0.067 0.447 0.497 Trp 1.00 ± 0.05 1.00 ± 0.08 1.11 ± 0.09 0.96 ± 0.11 0.370 0.684 0.368 Phe 1.00 ± 0.03 1.04 ± 0.04 1.14 ± 0.08 1.10 ± 0.05 0.979 0.081 0.493 Met 1.00 ± 0.07 0.85 ± 0.03 0.83 ± 0.06 0.84 ± 0.03 0.220 0.136 0.173 Thr 1.00 ± 0.07 1.03 ± 0.07 0.98 ± 0.09 1.08 ± 0.15 0.506 0.875 0.716 Ile 1.00 ± 0.12 0.95 ± 0.07 1.27 ± 0.13 1.01 ± 0.05 0.165 0.139 0.323

Leu 1.00 ± 0.10 1.07 ± 0.09 1.34 ± 0.08 1.20 ± 0.05 0.891 0.011 0.367 Downloaded from Val 1.00 ± 0.12 1.19 ± 0.15 1.48 ± 0.10 1.18 ± 0.10 0.673 0.063 0.056 Nonessential amino acid Pro 1.00 ± 0.11 0.90 ± 0.06 1.38 ± 0.26 0.99 ± 0.20 0.178 0.198 0.408 Gly 1.00 ± 0.04 1.02 ± 0.09 0.80 ± 0.04 0.88 ± 0.06 0.409 0.012 0.647 Ser 1.00 ± 0.09 1.11 ± 0.13 1.19 ± 0.07 1.36 ± 0.06 0.152 0.034 0.715 His 1.00 ± 0.09 1.03 ± 0.11 1.16 ± 0.11 1.14 ± 0.10 0.943 0.215 0.837 Ala 1.00 ± 0.13 0.72 ± 0.08 0.59 ± 0.06 0.92 ± 0.14 0.607 0.105 0.054 http://ps.oxfordjournals.org/ Cys 1.00 ± 0.74 3.43 ± 0.78 0.34 ± 0.05 2.42 ± 1.11 0.008 0.272 0.812 Orn 1.00 ± 0.11 0.87 ± 0.09 1.17 ± 0.08 1.06 ± 0.17 0.313 0.132 0.917 Glu 1.00 ± 0.12 1.18 ± 0.06 1.11 ± 0.07 1.15 ± 0.05 0.205 0.636 0.445 Tyr 1.00 ± 0.03 1.00 ± 0.03 1.04 ± 0.01 0.99 ± 0.02 0.360 0.496 0.345 Asp 1.00 ± 0.16 0.83 ± 0.08 1.03 ± 0.16 0.91 ± 0.18 0.340 0.723 0.862 1STI = short TI; LTI = long TI.

Quantitative Reverse-Transcription lation had occurred. Actin (α 1) mRNA expression was at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 PCR Analysis markedly increased by CORT administration, which was consistent with the change in protein expression The abundance of the mRNA transcripts encoding in breast muscle. In LTI broilers, expression of mRNA the 16 differentially expressed proteins identified in the encoding actin (α 1) was significantly upregulated. Ex- breast muscle of CORT-treated and TI broilers was depression of mRNA encoding 6-phosphofructokinase was termined using quantitative reverse-transcription PCR downregulated by CORT administration (P < 0.05). (Table 6). Expression of the mRNA encoding α-actin and actin-α 2 was significantly lower (P < 0.05) in LTI Western Blot Analysis broilers than STI broilers, which was in agreement with the level of α-actin and actin-α 2 protein expression in Western blot analysis showed a significant increase the breast muscle, indicating that transcriptional regu- of HSC70 (P < 0.05) and with a tendency for an in-

Figure 1. Differentially expressed muscle proteins in breast muscle of the short tonic immobility (STI) and long tonic immobility (LTI) groups by two-dimensional gel electrophoresis analysis. The differentially expressed proteins between STI and LTI chickens were spotted and numbered. Neither the STI nor the LTI group was exposed to corticosterone. pI, isoelectric point; Mr, molecular mass. COMPARATIVE PROTEOMICS IN BROILERS 789

Table 4. Differentially expressed muscle proteins of the short and long tonic immobility groups identified by two-dimensional gel electrophoresis analysis and MALDI-TOF-MS1,2

Accession Moleculur Protein Spot number Protein name number weight (KDa) pI expression Myofilament 2 α-Actin gi|178027 42.28 5.23 p 6 α-Actinin-2 gi|46048687 104.78 5.26 p 7 Myosin heavy chain gi|238274 223.8 5.61 p 8 Myosin regulatory light chain 2 gi|311314962 18.92 4.77 n 9 Myosin L2 gi|223047 18.74 4.77 p Metabolism

3 Glucose-6-phosphate isomerase gi|57524920 62.41 7.4 p Downloaded from 5 6-Phosphofructokinase gi|16610202 84.78 8.35 p 1 Glyceraldehyde-3-phosphate dehydrogenase gi|46048961 35.91 8.71 p 12 Phosphoglycerate mutase 1 gi|71895985 29.05 7.03 p 13 Pyruvate kinase muscle isozyme gi|45382651 58.43 7.29 n Transport 10 Apolipoprotein A-I precursor gi|211146 30.67 5.97 n Miscellaneous http://ps.oxfordjournals.org/ 4 Calsequestrin gi|211497 45.09 4 p 11 Skeletal muscle C-protein gi|212659 112.18 6.05 n 1MALDI-TOF-MS = matrix-assisted laser desorption ionization-time of flight mass spectrometry. pI, isoelectric point; gi, GenInfo Identifier. 2Protein name and accession numbers were derived from NCBI database. crease for desmin (0.05 < P < 0.1) protein in the breast with CORT administration has been well documented

muscle of CORT-treated chickens compared with con- (Malheiros et al., 2003; Dong et al., 2007). Dong et at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 trol, which was consistent with the results in 2-DE pro- al. (2007) reported that chronic CORT exposure sig- teomic analyses. However, Western blot analysis did nificantly decreases the capacity of protein synthesis in not detect a significant change of GAPDH protein in the pectoralis major muscle of broilers as estimated by breast muscle of LTI and STI chickens as observed in the RNA-to-protein ratio. In our study, the capacity of 2-DE proteomic analyses (Figure 3). protein synthesis as estimated by RNA-to-protein ratio was consistently and significantly suppressed by CORT DISCUSSION administration, but not LTI. Our results also showed that the levels of amino acids in serum generally in- It is generally accepted that LTI and chronic stress creased in CORT-treated chickens, and that the serum negatively affect growth and development. Severe sup- concentrations of Ser and Leu in particular increased pression of BW gain in broilers and quail associated significantly. The increase in amino acid concentrations

Figure 2. Differentially expressed muscle proteins of control (CON) and corticosterone (CORT) group in breast muscle by two-dimensional gel electrophoresis analysis. The differentially expressed proteins between control and CORT-treated chickens were spotted and numbered. pI, isoelectric point; Mr, molecular mass. 790 FU ET AL.

Table 5. Differentially expressed muscle proteins of the control and corticosterone groups identified by two-dimensional gel electrophoresis analysis and MALDI-TOF-MS1,2

Accession Molecular Protein Spot number Protein name number weight (KDa) pI expression Myofilament 2 Actin, α skeletal muscle gi|27819614 42.45 5.31 p 3 Myosin L2 gi|223047 18.74 4.77 p 4 Myosin light chain 1f gi|212330 20.96 4.79 n 13 Actin, α 1, skeletal muscle gi|55741890 42.37 5.23 n Cytoskeleton 6 Desmin gi|2959450 51.69 5.30 n

Cellular defense and stress Downloaded from 7 Heat shock protein 70B gi|161408079 71.03 5.37 n 9 Heat shock protein 70 gi|30962014 70.10 5.66 n 11 Heat shock protein β-1 gi|45384222 21.72 5.77 n 14 Heat shock protein β-1-like gi|326931250 21.89 6.23 n 15 Heat shock cognate 71 kDa protein gi|45384370 71.07 5.47 n Metabolism 1 Pyruvate kinase muscle isozyme gi|45382651 58.43 7.29 n http://ps.oxfordjournals.org/ 12 l-Lactate dehydrogenase A chain gi|45384208 36.78 7.75 p 16 Creatine kinase M-type gi|45382875 43.53 6.5 p Transport 5 Apolipoprotein A-I precursor gi|211146 30.67 5.97 n 8 Serum albumin precursor gi|45383974 71.87 5.51 n 10 Fatty acid-binding protein, heart gi|71894843 14.81 5.92 n 1Corticosterone (C 2505, Sigma, St. Louis, MO); MALDI-TOF-MS = matrix-assisted laser desorption ionization-time of flight mass spectrometry. pI, isoelectric point; gi, GenInfo Identifier.

2Protein name and accession numbers were derived from NCBI database. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 in the blood is indicative of augmented protein catabo- cle (Nakashima et al., 2005) and to play a role in regu- lism and protein breakdown in the skeletal muscle fol- lation of the mechanistic target of rapamycin signaling lowing CORT treatment, consistent with previous re- pathway (Lynch, 2001). However, of the 18 amino acids ports (Hayashi et al., 1994; Dong et al., 2007). we detected, only Cys content in serum differed signifi- Increases in the concentrations of circulating amino cantly between the TI phenotypes, as the Cys level was acids might act as signals regulating biological process- significantly higher in LTI broilers. The discrepancies es and organ development. Leucine has been shown to between LTI and CORT broilers, with respect to the suppress myofibrillar proteolysis in chick skeletal mus- protein synthesis capacity and serum amino acid profile

Table 6. Differentially expressed at the mRNA level1

Control2 CORT2 P-value2

Item STI LTI STI LTI TI CORT TI × CORT Myofilament Myosin light chain 1f 1.00 ± 0.08 0.86 ± 0.07 0.78 ± 0.11 0.92 ± 0.10 0.993 0.409 0.140 Actin, α 1, skeletal muscle 1.00 ± 0.53 2.84 ± 1.06 3.93 ± 0.91 6.47 ± 0.99 0.024 0.002 0.702 α-Actin 1.00 ± 0.22 0.63 ± 0.09 0.85 ± 0.13 0.58 ± 0.11 0.041 0.496 0.732 Actin α 2 1.00 ± 0.11 0.81 ± 0.14 0.83 ± 0.10 0.56 ± 0.05 0.041 0.496 0.732 Metabolism Pyruvate kinase muscle isozyme 1.00 ± 0.06 0.99 ± 0.24 0.89 ± 0.21 0.87 ± 0.06 0.903 0.373 0.986 Creatine kinase M-type 1.00 ± 0.06 0.84 ± 0.06 0.93 ± 0.20 0.63 ± 0.10 0.067 0.243 0.565 Glucose-6-phosphate isomerase 1.00 ± 0.06 0.82 ± 0.12 0.98 ± 0.16 0.87 ± 0.04 0.155 0.894 0.713 Glyceraldehyde-3-phosphate 1.00 ± 0.05 1.07 ± 0.06 0.97 ± 0.12 0.66 ± 0.17 0.399 0.138 0.205 dehydrogenase 6-Phosphofructokinase 1.00 ± 0.08 1.00 ± 0.14 0.87 ± 0.17 0.89 ± 0.09 0.291 0.043 0.284 l-Lactate dehydrogenase A chain 1.00 ± 0.12 1.21 ± 0.16 0.91 ± 0.18 0.96 ± 0.04 0.817 0.065 0.199 Pyruvate kinase muscle isozyme 1.00 ± 0.06 0.99 ± 0.24 0.89 ± 0.21 0.87 ± 0.06 0.903 0.373 0.986 Transport Apolipoprotein A-I precursor 1.00 ± 0.14 0.80 ± 0.12 0.77 ± 0.06 0.75 ± 0.13 0.336 0.246 0.448 Fatty acid-binding protein, heart 1.00 ± 0.10 0.96 ± 0.11 0.95 ± 0.14 0.77 ± 0.09 0.310 0.290 0.536 Cytoskeleton Desmin 1.00 ± 0.53 2.49 ± 1.07 2.83 ± 0.60 2.58 ± 0.36 0.351 0.159 0.198 Cellular defense and stress Heat shock protein 70 1.00 ± 0.17 0.83 ± 0.07 0.68 ± 0.08 0.77 ± 0.14 0.735 0.133 0.311 Miscellaneous Calsequestrin 1.00 ± 0.21 0.85 ± 0.10 0.89 ± 0.12 0.91 ± 0.07 0.353 0.649 0.275 1Values are means ± SEM, n = 6/group. 2CORT = corticosterone (C 2505, Sigma, St. Louis, MO); STI = short tonic immobility duration; LTI = long tonic immobility duration; TI = tonic immobility. COMPARATIVE PROTEOMICS IN BROILERS 791 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 3. Western blot analysis of differentially expressed protein in two-dimensional gel electrophoresis (2-DE) gels. (A) Western blot analysis of desmin, heat shock cognate 70 (HSC70), and glyceraldehyde 3-phosphate dehydrogenase (GADPH) in total protein extracts from skeletal muscle. (B) The proteins from representative 2-DE gels are displayed. CON = control group; CORT = corticosterone-treated group; STI = short tonic immobility group not exposed to CORT; LTI = long tonic immobility group not exposed to CORT. Means with asterisks (*) are significantly different from those of the corresponding controls (P < 0.05). data, indicate that different mechanisms are respon- (Ohlendieck, 2010). Actin is a major component of sible for the negative effects on muscle development the contractile apparatus of vertebrate skeletal, cardi- associated with these variables. The potential effect of ac, and smooth muscle (Kabsch and Vandekerckhove, significantly higher concentrations of Cys in the blood 1992). In our study, 2-DE indicated that desmin was of LTI broilers remains unknown. highly expressed in breast muscle after CORT admin- Myofibrillar filaments are responsible for generat- istration, and this result was confirmed by Western ing the physical movement of skeletal muscles. Myo- blot analysis with an antibody specific for desmin. The sin heavy chain and myosin L2 were more abundant in modifications that we observed in regulatory proteins STI than LTI chickens, whereas myosin regulatory light were likely related to the characteristic disorganization chain 2 was more abundant in LTI chickens. Compared of myofibers that occurs in response to fear and chronic with control chickens, the abundance of myosin L2 was stress mimicked by chronic CORT administration. higher, but that of myosin light chain 1f was lower af- Skeletal muscle tissue is unique in its ability to handle ter CORT administration. The significant down- or up- very rapid and coordinated changes in energy supply regulation of these myosin light chain proteins is the and oxygen flux during contraction (Ge et al., 2004). primary determinant of force and velocity in muscle Metabolic proteins are heavily involved in the processes fibers, which is coincident with the change in skeletal that occur in skeletal muscle during myogenesis and muscle fiber types during skeletal muscle development development. In the pectoralis muscle, glycolysis is one 792 FU ET AL. of the main pathways through which birds derive en- expression of glycolytic enzymes is dramatically altered ergy for muscle contraction and fuel energy demands in STI but not CORT-treated broilers. The observed for growth (Doherty et al., 2004). In the present study, differences in biochemical parameters, gene expression, the content of glycolytic enzymes in breast muscle was and protein expression profiles in breast muscle imply greatly affected by TI, as LTI broilers had lower levels that the stress response mechanism differs in CORT- of glucose-6-phosphate isomerase, 6-phosphofructoki- treated and TI broilers. This is the first report demon- nase, GAPDH, and phosphoglycerate mutase 1 than strating that in parallel with the absence of interaction STI broilers. Although the expression of pyruvate ki- of LTI and CORT on BW and muscle weight, TI and nase muscle isozyme was upregulated in breast muscle CORT do not interact on the protein synthesis capac- of both CORT-treated and LTI broilers, no differences ity, serum amino acid levels, protein expression profile, were observed in the expression of 4 other glycolytic or gene transcription in the breast muscle of broilers. Downloaded from enzymes in the breast muscle of control and CORT- treated broilers. Our results indicated that the differences in glucose metabolism in muscle might contribute REFERENCES to the divergent growth performance between STI and

Awerman, J. L., and L. M. Romero. 2010. Chronic psychological http://ps.oxfordjournals.org/ LTI birds. When a quantitative trait loci study was stress alters body weight and blood chemistry in European star- undertaken to identify genome regions involved in the lings (Sturnus vulgaris). Comp. Biochem. Physiol. A Mol. Integr. response of fearfulness between LTI and STI, the key Physiol. 156:136–142. Beuving, G., R. B. Jones, and H. J. Blokhuis. 1989. Adrenocorti- enzymes controlling glucose metabolism can be taken cal and heterophil/lymphocyte responses to challenge in hens into consideration in birds. Moreover, the expression of showing short or long tonic immobility reactions. Br. Poult. Sci. both l-lactate dehydrogenase A chain (LDH) and cre- 30:175–184. atine kinase M-type (CK-M) was significantly down- Calandreau, L., A. Favrea-Peigne, A. Bertin, P. Constantin, C. Ar- nould, A. Laurence, S. Lumineau, C. Houdelier, M. A. Richard- regulated by CORT administration. The observed de- Yris, A. Boissy, and C. Leterrier. 2011. Higher inherent fear- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 creases in the expression of LDH and CK-M in breast fulness potentiates the effects of chronic stress in the Japanese muscle caused by CORT were consistent with the re- quail. Behav. Brain Res. 225:505–510. ported lower activities of LDH and CK-M in serum Campo, J. L., M. G. Gil, I. Munoz, and M. Alonso. 2000. Relation- ships between bilateral asymmetry and tonic immobility reac- after CORT administration (Wang et al., 2013). In tion or heterophil to lymphocyte ratio in five breeds of chickens. contrast to other studies demonstrating that plasma Poult. Sci. 79:453–459. CK-M and LDH levels increase in response to stress Campo, J. L., and A. Redondo. 1996. Tonic immobility reaction and (Awerman and Romero, 2010), we found that the de- heterophil to lymphocyte ratio in hens from three Spanish breeds laying pink eggshells. Poult. Sci. 75:155–159. creased activity of CK-M and LDH induced by chronic Doherty, M. K., L. McLean, J. R. Hayter, J. M. Pratt, D. H. Rob- CORT exposure is likely related to significant decreases ertson, A. El-Shafei, S. J. Gaskell, and R. J. Beynon. 2004. The in growth rate and muscle mass. proteome of chicken skeletal muscle: changes in soluble protein expression during growth in a layer strain. Proteomics 4:2082– Heat shock proteins (HSP) are a family of highly 2093. conserved stress proteins found in almost all cells. We Dong, H., H. Lin, H. C. Jiao, Z. G. Song, J. P. Zhao, and K. J. found that the expression of 5 HSP increased as a result Jiang. 2007. Altered development and protein metabolism in of CORT treatment. The increase in the abundance skeletal muscles of broiler chickens (Gallus gallus domesticus) by corticosterone. Comp. Biochem. Physiol. A Mol. Integr. Physiol. of HSC70 in breast muscle following CORT treatment 147:189–195. was confirmed by Western blotting. Heat shock protein El-Lethey, H., B. Huber-Eicher, and T. W. Jungi. 2003. Exploration 70 is one of the most highly inducible stress response of stress-induced immunosuppression in chickens reveals both proteins (Welch, 1992). Heat shock protein B1 and stress-resistant and stress-susceptible antigen responses. Vet. Immunol. Immunopathol. 95:91–101. heat shock protein B1-like, which are small HSP, are Gallup, G. G., L. B. Jr. Wallnau, J. L. Boren, G. J. Gagliardi, J. D. very important for proper myogenesis (Sugiyama et al., Maser, and P. H. Edson. 1977. Tryptophan and tonic immobility 2000; Middleton and Shelden, 2013) and are required in chickens: effects of dietary and systemic manipulations. J. to sustain muscle function under physiological stress. Comp. Physiol. Psychol. 91:642–648. Ge, Y., M. P. Molloy, J. S. Chamberlain, and P. C. Andrews. 2004. We report herein the first demonstration that chronic Differential expression of the skeletal muscle proteome in mdx administration of CORT induces a substantial increase mice at different ages. Electrophoresis 25:2576–2585. in HSP expression in breast muscle, which contributes Glennemeier, K. A., and R. J. Denver. 2002. Small changes in whole- to retarded muscle development via the stress response. body corticosterone content affect larval Rana pipiens fitness components. Gen. Comp. Endocrinol. 127:16–25. However, no differential HSP expression was observed Hayashi, K., Y. Nagai, A. Ohtsuka, and Y. Tomita. 1994. Effects in LTI phenotype broilers, indicating that there are of dietary corticosterone and trilostane on growth and skeletal divergent pathways of muscle development in breast muscle protein turnover in broiler cockerels. Br. Poult. Sci. 35:789–798. muscle. Hayward, L. S., and J. C. Wingfield. 2004. Maternal corticosterone In conclusion, we found that the growth of breast is transferred to avian yolk and may alter offspring growth and muscle is severely retarded in broilers of the LTI innate adult phenotype. Gen. Comp. Endocrinol. 135:365–371. phenotype and broilers subjected to chronic CORT ad- Hazard, D., M. Couty, J. M. Faure, and D. Guémené. 2005. Relation- ship between hypothalamic-pituitary-adrenal axis responsiveness ministration. Proteomic analyses showed that HSP ex- and age, sexual maturity status, and sex in Japanese quail se- pression is significantly upregulated in the breast mus- lected for long or short duration of tonic immobility. Poult. Sci. cle of CORT-treated but not LTI broilers, whereas the 84:1913–1919. COMPARATIVE PROTEOMICS IN BROILERS 793

Hazard, D., M. Couty, S. Richard, and D. Guémené. 2008. Intensity Minvielle, F., A. D. Mills, J. M. Faure, J. L. Monvoisin, and D. and duration of corticosterone response to stressful situations Gourichon. 2002. Fearfulness and performance related traits in in Japanese quail divergently selected for tonic immobility. Gen. selected lines of Japanese quail (Coturnix japonica). Poult. Sci. Comp. Endocrinol. 155:288–297. 81:321–326. Janczak, A. M., M. Heikkilä, A. Valros, P. Torjesen, I. L. Andersen, Morici, L. A., R. M. Elsey, and V. A. Lance. 1997. Effects of long- and M. Bakken. 2007. Effects of embryonic corticosterone expo- term corticosterone implants on growth and immune function sure and post-hatch handling on tonic immobility and willingness in juvenile alligators, Alligator mississippiensis. J. Exp. Zool. to compete in chicks. Appl. Anim. Behav. Sci. 107:275–286. 279:156–162. Jones, R. B., A. D. Mills, and J. M. Faure. 1991. Genetic and expe- Nakashima, K., A. Ishida, M. Yamazaki, and H. Abe. 2005. Leucine riential manipulation of fear-related behavior in Japanese quail suppresses myofibrillar proteolysis by down-regulating ubiquitin- chicks (Coturnix coturnix japonica). J. Comp. Psychol. 105:15– proteasome pathway in chick skeletal muscles. Biochem. Bio- 24. phys. Res. Commun. 336:660–666.

Kabsch, W., and J. Vandekerckhove. 1992. Structure and function of Ohlendieck, K. 2010. Proteomics of skeletal muscle glycolysis. Bio- Downloaded from actin. Annu. Rev. Biophys. Biomol. Struct. 21:49–76. chim. Biophys. Acta 1804:2089–2101. Leonard, B. E. 2006. HPA and immune axes in stress: Involvement of Post, J., J. M. Rebel, and A. A. ter Huurne. 2003. Physiological ef- the serotonergic system. Neuroimmunomodulation 13:268–276. fects of elevated plasma corticosterone concentrations in broiler Lin, H., E. Decuypere, and J. Buyse. 2004a. Oxidative stress induced chickens. An alternative means by which to assess the physiologi- by corticosterone administration in broiler chickens (Gallus gal- cal effects of stress. Poult. Sci. 82:1313–1318. lus domesticus) 1. Chronic exposure. Comp. Biochem. Physiol. Shini, S., A. Shini, and G. R. Huff. 2009. Effects of chronic and B Biochem. Mol. Biol. 139:737–744. repeated corticosterone administration in rearing chickens on http://ps.oxfordjournals.org/ Lin, H., E. Decuypere, and J. Buyse. 2004b. Oxidative stress induced physiology, the onset of lay and egg production of hens. Physiol. by corticosterone administration in broiler chickens (Gallus gal- Behav. 98:73–77. lus domesticus) 2. Short-term effect. Comp. Biochem. Physiol. B Sugiyama, Y., A. Suzuki, M. Kishikawa, R. Akutsu, T. Hirose, M. Biochem. Mol. Biol. 139:745–751. M. Waye, S. K. Tsui, S. Yoshida, and S. Ohno. 2000. Muscle Lin, H., S. J. Sui, H. C. Jiao, J. Buyse, and E. Decuypere. 2006. develops a specific form of small heat shock protein complex com- Impaired development of broiler chickens by stress mimicked posed of MKBP/HSPB2 and HSPB3 during myogenic differen- by corticosterone exposure. Comp. Biochem. Physiol. A Mol. tiation. J. Biol. Chem. 275:1095–1104. Integr. Physiol. 143:400–405. Sunde, J., A. Kiessling, D. Higgs, and J. Opstvedt. 2003. Evaluation

Livak, K. J., and T. D. Schmittgen. 2001. Analysis of relative of feed protein quality by measuring plasma free amino acids in at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 gene expression data using real-time quantitative PCR and the Atlantic salmon (Salmo salar L.) after dorsal aorta cannulation. 2(-Delta Delta C(T)) . Methods 25:402–408. Aquacult. Nutr. 9:351–360. Lynch, C. J. 2001. Role of leucine in the regulation of mTOR by Wang, S., Y. D. Ni, F. Guo, W. Y. Fu, R. Grossmann, and R. Q. amino acids: Revelations from structure-activity studies. J. Zhao. 2013. Effect of corticosterone on growth and welfare of Nutr. 131:861S–865S. broiler chickens showing long or short tonic immobility. Comp. Malheiros, R. D., V. M. Moraes, A. Collin, E. Decuypere, and J. Biochem. Physiol. A Mol. Integr. Physiol. 164:537–543. Buyse. 2003. Free diet selection by broilers as influenced by di- Welch, W. J. 1992. Mammalian stress response: Cell physiology, etary macronutrient ratio and corticosterone supplementation. structure/function of stress proteins, and implications for medi- 1. Diet selection, organ weights, and plasma metabolites. Poult. cine and disease. Physiol. Rev. 72:1063–1081. Sci. 82:123–131. Zulkifli, I., A. Al-Aqil, A. R. Omar, A. Q. Sazili, and M. A. Rajion. Middleton, R. C., and E. A. Shelden. 2013. Small heat shock pro- 2009. Crating and heat stress influence blood parameters and tein HSPB1 regulates growth of embryonic zebrafish craniofacial heat shock protein 70 expression in broiler chickens showing short muscles. Exp. Cell Res. 319:860–874. or long tonic immobility reactions. Poult. Sci. 88:471–476. Research Note

Determination of space use by laying hens using kinematic analysis

Joy A. Mench1 and Richard A. Blatchford

Department of Animal Science and Center for Animal Welfare, University of California, Davis 95616

ABSTRACT Two states in the United States now have by superimposing a grid over videos of the hens lying legislation requiring that laying hens be provided with down in their home cages. On average, hens required a Downloaded from sufficient space to perform particular behaviors. To mean area of 563 (± 8) cm2 to stand, 1,316 (± 23) cm2 provide a framework for translating these performance to turn around, 318 (± 6) cm2 to lie down, and 1,693 (± standards into a space requirement, kinematic analysis 136) cm2 to wing flap. The mean heights used were 34.8 was used to measure the amount of space needed for (± 1.3) cm for standing, 38.6 (± 2.3) cm for turning,

White Leghorn hens to stand, turn around 180°, lie and 49.5 (± 1.8) cm for wing flapping. However, space http://ps.oxfordjournals.org/ down, and wing flap. Hyline W-36 hens (n = 9) were requirements for hens housed in multiple-hen groups in marked on the tops of their heads and the tips of both cage or noncage systems cannot be based simply on in- wings and 3 toes with black livestock marker. Each formation about the space required for local movement hen was then placed in a floor pen (91.4 × 91.4 cm) by a single hen. It must also incorporate consideration and filmed using 2 high-speed cameras. The resulting of the tendency of hens in a flock to synchronize their images were processed using a software program that behaviors. In addition, it must include not just local generated 3-dimensional space use for each behavior. movement space but also the space that hens may need Because none of the hens lay down in the test pen, the to use for longer-distance movements to access resourc- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 2-dimensional space required for lying was determined es such as food, water, perches, and nest boxes. Key words: laying hen, welfare , behavior , movement , kinematics 2014 Poultry Science 93 :794–798 http://dx.doi.org/ 10.3382/ps.2013-03549

INTRODUCTION explicitly mandates or outlaws the use of particular hen housing systems. In 1965, the Brambell Committee in the United These laws pose challenges in terms of translating the Kingdom issued a report advocating that farm animals performance-based requirement for freedom of move- be ensured Five Freedoms: to turn around, lie down, ment into an engineering-based standard applicable stand up, stretch, and groom, without restriction of to all hen housing systems. Engineering standards for movement (Brambell, 1965). Two US states have re- minimum floor area for both cage and noncage systems cently enacted legislation that echoes this language. for commercial egg-laying hens are already in use in For example, California’s Prevention of Farm Animal many countries and regions, including the European Cruelty Act (2008; effective 2015) stipulates that con- Union (EU; CEC, 1999) and the United States (United fined farm animals must not be prevented for most of Egg Producers (UEP), 2010; Table 1). The EU stan- the day from standing up, lying down, fully extending dards were based upon scientific research, practical ex- their limbs, and turning around without touching the perience, and discussion occurring during comments on side of an enclosure or another animal. For egg-laying proposed regulations (Appleby, 2003). The UEP (2010) hens, “extending limbs” is defined as spreading both standards for conventional cages were based on a litera- wings without touching the sides of an enclosure or ture review focused mainly on mortality, feather qual- other hens. Michigan has a similarly worded statute, ity, stress, and egg production data (Bell et al., 2004), but which also requires that each hen have “access to whereas those for noncage systems were also based on 2 at least 1.0 square feet of usable space” (929 cm ; Ani- practical experience because there were less available mal Industry Act, 1988, amended 2009). Neither law published data. Although there has been significant research on the effects of stocking density in different hen housing systems on health, production, and availability of resourc- © 2014 Poultry Science Association Inc. es, there has been less on how hens use space while Received August 9, 2013. Accepted November 29, 2013. performing particular behaviors. Most research has in- 1 Corresponding author: [email protected] volved evaluating how providing different amounts of

794 RESEARCH NOTE 795

Table 1. European Union (EU) and United Egg Producers (UEP) minimum space requirements for different hen housing systems

System and aspects European Union United Egg Producers1 Conventional cages2 Floor area per hen 550 cm2 (85.3 in2) of unrestricted space 432–555 cm2 (67–86 in2) usable space, depending on hen size and cage configuration Enclosure height No less than 35 cm (13.8 in) at any point; No numerical requirement: sufficient for at least 40 cm (15.7 in) over 65% of cage hens to “stand comfortably upright” area Enriched cages Floor area per hen 750 cm2 (116.3 in2) per hen,3 600 cm2 (93 No standards 2 in ) of which must be of usable height Downloaded from Enclosure height Must be at least 45 cm (17.7 in) to be considered “usable”; cage height outside usable area no less than 20 cm (7.9 in) Minimum enclosure size 2,000 cm2 (310 in2) Noncage systems Floor/barn 9 hens/m2 (1.2 ft2)3 usable area 1,394 cm2 (216 in2)3 Floor/barn with roosting area over droppings pit Same 1,116 cm2 (173 in2)3 for brown hens, 929 http://ps.oxfordjournals.org/ cm2 (144 in2)3 for white hens Multilevel (aviary) Same 929 cm2 (144 in2)3 1The UEP standards are in US rather than metric units, so the metric units represent a conversion of the original standard. 2Banned in the EU effective 2012. 3Excluding the nest area.

space affects the expression of various behaviors (Sa- Research Facility flock at the University of California, at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 vory et al., 2006), particularly in the context of flock Davis, were used. They were housed individually in social interactions (reviewed by Weeks and Nicol, cages (45.7 × 45.7 cm) before the start of the study. 2006). An approach more directly relevant to the afore- One hen was euthanized for flock management reasons mentioned state regulations involves measuring the before the recording of lying behavior; all other hens amount of space used during the performance of dif- were returned to the flock after being filmed. Hous- ferent behaviors. Dawkins and Hardie (1989) measured ing and experimental procedures were approved by the video images to determine the space required for hens Institutional Animal Care and Use Committee of the to stand, ground scratch, turn, wing stretch, wing flap, University of California, Davis. feather ruffle, and preen. They used Ross Brown hens, Each hen was marked with black livestock marker which are larger than the white birds that make up the on the top of her head, tip of her tail, tips of 3 of her majority of US egg production. toes, and tips of both of her wings and placed in a test The development of kinematic technology provides pen constructed of polyvinyl chloride (PVC) pipe and the opportunity for more accurate determination of the chicken wire (91.4 × 91.4 cm) for up to 1 h while being space used by animals to perform behaviors than man- video recorded. A round PVC perch (2.54 cm diameter, ual measurement from video recordings and has the 91.4 cm long) was placed in the pen approximately 12.7 additional advantage that space use can be evaluated cm from the floor to stimulate the hen to jump up and in 3 dimensions. Kinematic analysis is most often used down and thus to wing flap. to assess gait patterns of animals (e.g., Nielsen et al., A control object of known measurement was placed 2003), but also been employed to assess the amount of in the pen and the space calibrated by filming and digi- space that dairy cows use when lying down in order to tizing the object. Two high-speed cameras (Fastcam make recommendations about improving stall design PCI, Photron, San Diego, CA) were placed at the front (Ceballos et al., 2004). The goal of the current study of the testing pen. Hens were recorded 1) standing in a was to use kinematic analysis to evaluate the space relaxed posture, 2) turning 180°, and 3) wing flapping. required for Hy-Line W-36 hens, the strain most com- The 2-dimensional files from these cameras were commonly used in US egg production, to stand up, lie down, bined to create one 3-dimensional (3-D) event file using fully extend both wings (wing flap), and turn around Vicon Motus 9.2 software (Vicon, Los Angeles, CA). freely. This information was then used to derive a for- Standing and turning were filmed at 60 Hz, and wing mula for calculation of space required per hen that can flapping was filmed at 500 Hz. The average floor space be used to determine a minimum space requirement, in 3-D used by the hens was calculated for each behav- taking into consideration different interpretations of ior using the maximum length and width of the hens, as regulatory language related to freedom of movement. determined by the software. Each hen was filmed until she had performed all of the behaviors once. Because MATERIALS AND METHODS none of the hens lay down during the test period, space required for lying was determined by recording lying Nine mature (approximately 1.5 yr of age) well- hens from above in their home cages, superimposing a feathered Hy-Line W-36 hens from the Hopkins Avian premeasured grid over the video recording, and deter- 796 MENCH AND BLATCHFORD

Table 2. The floor area [cm2 (in2)] used by hens when performing particular behaviors, as well as the height [cm (in)], wingspan [cm (in)], and wing flap floor area [cm2 (in2)] with 2.54 cm (1 in) added to the length and width of the hen

Area mean Height mean Behavior (± SEM) Range (± SEM) Height range Stand 563 ± 8 391.0 to 716.8 34.8 ± 1.3 30.5 to 41.1 (87.3 ± 1.2) (60.6 to 111.1) (13.7 ± 0.5) (12.0 to 16.2) Turn (180°) 1,316 ± 23 925.8 to 2,191 38.6 ± 2.3 37.7 to 52.3 (204 ± 4) (143.5 to 339.6) (15.2 ± 0.9) (12.1 to 20.6) Lie down 318 ± 6 290.3 to 335.5 — — (49.3 ± 1) (45 to 52) Wingspan 50.6 ± 3.1 37.8 to 61.7 — —

(20 ± 1) (18.8 to 24.3) Downloaded from Wing flap 1,693 ± 136 1,085.6 to 2,446.6 49.5 ± 1.8 37.6 to 57.2 (263 ± 21) (168.3 to 379.2) (19.1 ± 0.7) (14.8 to 22.5) Wing flap + 2.54 cm 1,913 ± 145 1,261 to 2,711 52 ± 1.8 40.1 to 59.8 (297 ± 22) (195.5 to 420.1) (19.4 ± 0.7) (16.3 to 23.1) http://ps.oxfordjournals.org/ mining the hen’s length from head to tail and her width the videotapes showed that hens do not fully extend at the widest cross-section. Hens were also weighed us- their wings when flapping to the extent that a person ing a digital scale (Ohaus, Pinebrook, NJ). Mean values would extend a hen’s wings when taking a wingtip-to- and their SE were calculated for each behavior and for wingtip measurement. Although Dawkins and Hardie BW for the 9 hens used in the study. (1989) did not provide a measure of wingspan for the Ross Browns they used, the current data demonstrate that using body measurements such as wingspan to

RESULTS AND DISCUSSION at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 calculate space may not provide an accurate measure- The means and SE for the space measurements ob- ment of the actual space required to perform particular tained from the kinematic recording and video mea- behaviors. surements are presented in Table 2. The data are simi- Table 2 also shows values for the behavior that on lar to those reported by Dawkins and Hardie (1989) for average used the most space, wing flapping, with an larger-bodied Ross Brown hens weighing between 1.9 additional 2.54 cm added to the mean hen length and and 2.6 kg (mean = 2.2), compared with the 1.3 to 1.9 width. This value was arbitrarily chosen to provide ad- kg (mean = 1.6 ± 0.7) weights of the W-36 hens in the ditional space for hens to wing flap without touching current study. This indicates that space for movement the enclosure or other hens. With this addition, enclo- is only minimally influenced by BW, at least within sures that provide less than 1,913 cm2 total floor area this range of weights. (the mean area needed for wing flapping), do not have However, the results of the current study do differ in a length or depth of at least 53.1 cm (mean wingspan some respects from those of Dawkins and Hardie (1989). plus 2.54 cm), or both, would not meet the require- The mean value for standing for the W-36 hens (563 ments even for a singly housed hen. cm2) was larger than that for the Ross Browns (475.3 In addition, enclosures that do not have a height of cm2), and the range of values was also wider (W-36: at least 49.5 cm over the minimum floor area might 391–716.8 cm2; Ross: 428–592 cm2). This was probably prevent wing flapping. Nicol (1987) and Albentosa and because the hen’s head position during standing can be Cooper (2004) evaluated behavior in cages of differ- either erect or extended, which plays an important role ent heights. In the former study, which was carried out in determining the floor area when kinematic analysis in conventional cages, wing flapping/raising was very is used, because the area is measured using the “drop” rare and was observed only in the tallest cage (55 cm), from the beak tip and tail. The mean area for turn- whereas in the latter study, which used small (10-hen) ing (1,316 cm2) was similar to that for Ross Browns furnished cages, full wing flapping was not observed (1,271.8 cm2), although the range was wider (W-36: even in the tallest cage (45 cm). This could be either 925.8–2,191 cm2; Ross: 978–1,628 cm2). Hens can make because cages physically restrict wing flapping, or be- turns that vary in the area covered. In our study we cause they restrict long-distance movement and wing defined a turn as a 180° movement, whereas Dawkins flapping is usually observed mainly when hens are run- and Hardie (1989) did not provide a turning definition. ning or flying. Wing flapping is often referred to as For wing flapping, the W-36 mean value (1,693 cm2) “comfort” (stretching) behavior, and hens also show was somewhat less than that for Ross Browns (1,876.3 other comfort-type behaviors that involve wing rais- cm2), and the range was also slightly narrower (W-36: ing (slight elevation of both wings) and wing stretching 1,085.6–2,446.6 cm2; Ross: 1,085–2,606 cm2). The max- (one wing stretched downward). Because these behav- imum value we recorded for wingspan (61.7 cm) was iors require less space than wing flapping, they were smaller than Hyline’s reference value (71.9 cm) for the not measured in this study. Although wing flapping is W-36 wingspan (Neil O’Sullivan, Hyline, Des Moines, often associated with running or flying it is also seen IA, personal communication). However, examination of when hens are stationary, making its function unclear RESEARCH NOTE 797 Downloaded from http://ps.oxfordjournals.org/

Figure 1. Change in the amount of floor space needed per hen as group size (n) increases. The numbers are based on the assumptions and calculations provided in the text: 2,711 cm + [(n − 1) × 563.2 cm]/n.

(Duncan, 1981). Regardless, a recent study of US com- per hen is required as group size increases, although the

mercial hen facilities indicates that the majority of A- overall size of the enclosure obviously also increases. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 frame cages are not only insufficient in height for hens The floor space required per hen essentially levels off to wing flap, but to stand (Kiess et al., 2012), based on at approximately 599 cm2 (93 in2) when the group size the current data. reaches 60, although very large groups (1,000 or more) Extrapolating these data to derive a minimum space require slightly less space per hen [approximately 564 recommendation for group-housed hens requires in- cm2 (87 in2)]. The Michigan (Animal Industry Act, formation about how many hens need to be able to 1988, amended 2009) engineering standard of “at least perform the stipulated behavior(s) simultaneously. Al- 1 square feet of usable space” (0.093 m2) would there- though there are several studies evaluating behavioral fore provide sufficient space under this scenario in a synchrony in hens kept in pairs in cages (Mench et al., 5-hen enclosure that measures at least 0.47 m2 (5 ft2), 1986; Webster and Hurnik, 1994) or in small groups but not enough space if more than one hen must be al- in pens (Collins et al., 2010), the data are insufficient lowed to wing flap simultaneously. to determine how motivated hens are to perform be- These results agree with theoretical predictions about haviors simultaneously in different commercial housing relationships between group size and freedom of move- systems under different conditions of group size and re- ment within enclosures. Appleby (2004) modeled free source provision. Using the Dawkins and Hardie (1989) movement in furnished cages and showed that larger data, Broom and Fraser (2007) provided an estimation enclosures provided more free movement space even if of the per-hen space that would be needed in 5-hen con- each hen was only provided with slightly more space ventional cages under several scenarios, with the least than her body size. Free space opens up in larger en- space required for 4 hens crowded together and one closures because hens do not use all of the space avail- wing flapping (544 cm2), and the most space for 2 hens able to them but instead cluster (Keeling and Duncan, standing, 2 turning, and 1 wing flapping (917 cm2). 1989; Collins et al., 2010). This free space permits lo- Because hens in all housing systems spend a signifi- cal freedom of movement for behaviors such as feeding, cant proportion of their time standing or lying down scratching, stretching, preening, and sitting (Appleby, (e.g., Keeling, 1994; Webster and Hurnik, 1994; Chan- 2004). ning et al., 2001; Savory et al., 2006), one reasonable A variety of factors, some static and some dynamic, interpretation of the US regulatory language might be affect how much space hens need and how that space that space needs to be provided for one hen to wing is used (Keeling, 1995). These factors include genetics, flap while the remaining hens are standing or lying. group size, environmental variables (e.g., ambient tem- Using the maximum value for wing flapping (plus 2.5 perature), hen age, social effects (e.g., Keeling, 1994; cm; 1 in) and the average value for standing (which re- Lindberg and Nicol, 1996), and resource provision and quired more space than lying), the per-hen floor space distribution. The floor space allocation calculated here (in cm2) required is estimated for different group sizes for a 60-hen group (599 cm2 or 93 in2) is virtually iden- (n) as follows: 2,711 + [(n − 1) × 563.2]/n. Using this tical to the minimum EU usable floor space require- equation, Figure 1 shows the amount of space needed ment for furnished cages (600 cm2; Table 1); 60 hens is per hen for different group sizes in both centimeters a common group size in newer versions of such cages. and inches. As can be seen from this figure, less space It is also very similar to the minimum floor space value 798 MENCH AND BLATCHFORD (excluding the nesting and foraging area) calculated by Channing, C. E., B. O. Hughes, and A. W. Walker. 2001. Spatial Appleby (2004) using his model for groups of 8 or more distribution and behaviour of laying hens housed in an alternative system. Appl. Anim. Behav. Sci. 72:335–345. hens in furnished cages. However, the calculations for Collins, L. M., L. Asher, D. U. Pfeiffer, W. J. Browne, and C. J. Ni- the large groups typical of noncage systems are below col. 2010. Clustering and synchrony in laying hens: The effect of both the EU and UEP standards. Providing this little environmental resources on social dynamics. Appl. Anim. Behav. per-hen space in noncage systems could potentially Sci. 129:43–53. Commission of the European Communities (CEC). 1999. Council obstruct the longer-distance movements necessary for Directive 1999/111/EEC laying down minimum standards for hens to reach resources such as feed, water, nests, and the protection of laying hens. Official Journal of the European perches, stressing the importance of basing space stan- Community L203:53–57. dards on multiple welfare criteria. Dawkins, M. S., and S. Hardie. 1989. Space needs of laying hens. Br.

Poult. Sci. 30:413–416. Downloaded from Duncan, I. J. H. 1981. Animal rights–animal welfare: A scientist’s ACKNOWLEDGMENTS assessment. Poult. Sci. 60:489–499. Keeling, L. J. 1994. Inter-bird distances and behavioural priorities in laying hens: The effect of spatial restriction. Appl. Anim. Behav. This research was supported by California Depart- Sci. 39:131–140. ment of Food and Agriculture Contract 09-0854. We Keeling, L. J. 1995. Spacing behaviour and an ethological approach thank Sue Stover and Tanya Garcia-Nolen of the School to assessing optimal space allocations for groups of laying hens. http://ps.oxfordjournals.org/ of Veterinary Medicine, University of California, Davis, Appl. Anim. Behav. Sci. 44:171–186. Keeling, L. J., and I. J. H. Duncan. 1989. Inter-individual distances for their generous assistance with recording and coding and orientation in laying hens housed in groups of three in two the 3-D kinematic data. We also gratefully acknowledge different-sized enclosures. Appl. Anim. Behav. Sci. 24:325–342. the infrastructure support of the Department of Animal Kiess, A. S., P. Y. Hester, J. A. Mench, R. C. Newberry, and J. Science, the College of Agricultural and Environmental P. Garner. 2012. A standardized cage measurement system: A versatile tool for calculating usable cage space. J. Appl. Poult. Sciences, and the California Agricultural Experiment Res. 21:657–668.

Station of the University of California–Davis. Lindberg, A. C., and C. J. Nicol. 1996. Effects of social and environ- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 mental familiarity on group preferences and spacing behavior in laying hens. Appl. Anim. Behav. Sci. 49:109–123. REFERENCES Mench, J. A., A. van Tienhoven, J. A. Marsh, C. C. McCormick, D. L. Cunningham, and R. C. Baker. 1986. Effects of cage and Albentosa, M. J., and J. J. Cooper. 2004. Effects of cage height floor pen management on behavior, production, and physiological and stocking density on the frequency of comfort behaviour per- stress responses of laying hens. Poult. Sci. 65:1058–1069. formed by laying hens housed in furnished cages. Anim. Welf. Nicol, C. J. 1987. Effect of cage height and area on the behaviour of 13:419–424. hens housed in battery cages. Br. Poult. Sci. 28:327–335. Animal Industry Act. 1988. PA 466. § 46 (Amended 2009). Michi- Nielsen, C., S. M. Stover, K. S. Schultz, M. Hubbard, and K. S. gan. Schulz. 2003. Two-dimensional link-segment model of the fore- Appleby, M. C. 2003. The European Union ban on conventional limb of dogs at a walk. Am. J. Vet. Res. 64:609–617. cages for laying hens: History and prospects. J. Appl. Anim. Prevention of Farm Animal Cruelty Act. 2008. CA hsc. §§ 25990– Welf. Sci. 6:103–121. 25994 (2008). Appleby, M. C. 2004. What causes crowding? Effects of space, fa- Savory, C. J., M. C. Jack, and V. Sandilands. 2006. Behavioural cilities and group size on behaviour, with particular reference to responses to different floor space allowances in small groups of furnished cages for hens. Anim. Welf. 13:313–320. laying hens. Br. Poult. Sci. 47:120–124. Bell, D., B. Chase, A. Douglass, P. Hester, J. Mench, R. Newber- United Egg Producers (UEP). 2010. Husbandry Guidelines for US ry, M. Shea-Moore, L. Stanker, J. Swanson, and J. Armstrong. Egg Laying Flocks. Accessed Jan. 18, 2014. http://www.unit- 2004. UEP uses scientific approach in its establishment of welfare edegg.org/. guidelines. Feedstuffs 76 (11; March 15):1–10. Webster, A. B., and J. F. Hurnik. 1994. Social synchronization of Brambell, F. W. R. 1965. Command Paper 2836. Her Majesty’s behaviour amongst laying hens in battery cages. Appl. Anim. Stationery Office, London, UK. Behav. Sci. 40:153–165. Broom, D. M., and A. F. Fraser. 2007. Domestic Animal Behaviour Weeks, C. A., and C. J. Nicol. 2006. Behavioural needs, priorities and Welfare. 4th ed. CABI, Wallingford, UK. and preferences of laying hens. World’s Poult. Sci. J. 62:297–306. Ceballos, A., D. Sanderson, J. Rushen, and D. M. Weary. 2004. Improving stall design: Use of 3-D kinematics to measure space use by dairy cows when lying down. J. Dairy Sci. 87:2042–2050. GENETICS

The potential for archiving and reconstituting valuable strains of turkey (Meleagris gallopavo) using primordial germ cells

Alexander J. Wade ,* 1 Nick A. French ,† and Grenham W. Ireland ‡

* EW Group, Platts Lane, Old Moss, Stapleford, Tarvin, Chester, CH3 8HR, United Kingdom; †Aviagen, 11 Lochend Rd., Edinburgh, Midlothian, EH28 8SZ, United Kingdom; and ‡ University of Manchester, Faculty of Life Sciences, Oxford Rd., M13 9PT, United Kingdom Downloaded from

ABSTRACT Diseases such as avian inﬂuenza can de- 5.4 PGC recovered, respectively. Primordial germ cells stroy turkey ﬂocks, potentially resulting in the loss of were frozen using Dulbecco’s modified Eagle medium, valuable or rare genetic material. Consequently, there 20% fetal calf serum, and 10% dimethylsulfoxide and is an urgent need to develop a means to archive such demonstrated 90 ± 1.7% viability after 3 mo frozen http://ps.oxfordjournals.org/ germplasm. Germline chimeras produced by intravas- in liquid nitrogen. Freshly isolated and frozen thawed cular transfer of primordial germ cells (PGC) have been DiI- and Q-Tracker-labeled PGC repopulated stage reported in other avian species but not turkeys. This 30 gonads after vascular transfer into ex ovo cultured study examined the feasibility of both establishing an embryos. The DiI-labeled cells repopulated gonads less archive of frozen PGC, and producing germline chime- frequently, with 36 ± 13.2% of gonads containing the ras by injecting the thawed PGC into host embryos. To DiI-labeled PGC, and 7 ± 3.8% of reinjected PGC

meet these aims, the following experiments were per- reaching the gonads of positive embryos. The Q-track- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 formed: (1) PGC identification within turkey embryos; er-labeled cells were detected more frequently in embry- (2) development of an efficient method for isolation os, with 67 ± 21.1% having positive signals, and 44 ± of turkey PGC; (3) demonstration that PGC can be 4.9% of reinjected Q-tracker-labeled PGC colonized the cryopreserved, recovered, and retain viability; (4) re- gonads of positive embryos. This study demonstrated injection into embryos and detection of injected PGC. the feasibility of using turkey PGC to archive turkey

Primordial germ cells were identified using periodic ac- germplasm from different strains because frozen PGC id-Schiff reagent and the immunological marker OLP-1. reintroduced into host embryos can colonize the host Bloodstream PGC were isolated using Ficoll density gonads, suggesting the possibility of producing turkey gradient centrifugation with PGC recovery peaking at germline chimeras. stages 13, 14, and 15 with 32 ± 4.9, 33 ± 6.4, and 26 ± Key words: primordial germ cell, turkey , germline chimera, cryopreservation , genetic conservation 2014 Poultry Science 93 :799–809 http://dx.doi.org/10.3382/ps.2013-03629

INTRODUCTION loss of different strains of poultry, and as a consequence a vast reduction of valuable/rare genetic material from Highly selected, genetically improved, populations of the gene pool. Given such a scenario, it is imperative birds have been a major contribution to the success of that poultry germplasm be preserved to maintain its the modern poultry industry (McKay, 2009). However, current genetic variation (Moore et al., 2006). there are also many rare and specialized lines of poul- Primordial germ cells (PGC) are stem cell precursors try that are threatened with extinction primarily due to the gametes. Unlike somatic cells PGC contain all to economic reasons, thereby reducing the biodiversity the genetic information that is capable of being trans- of poultry species (Fulton and Delany, 2003). Poul- ferred to the next generation upon their maturation try stocks are at risk from potential disease epidem- into mature gametes (D’Costa et al., 2001). Primordial ics (Blackburn, 2006). Highly virulent strains of avian germ cells demonstrate a complex migratory pathway influenza, if established, would inevitably result in the throughout the developing avian embryo. Initially PGC are identified on the ventral surface of the stage X– XIV chicken or equivalent stage VII–XI turkey epiblast © 2014 Poultry Science Association Inc. [Roman numerals refer to the staging systems of Eyal- Received September 17, 2013. Accepted December 3, 2013. Giladi and Kochav (1976) for chicken, and Gupta and 1 Corresponding author: [email protected] Bakst (1993) and Bakst et al. (1997) for turkey em-

799 800 Wade et al. bryos]. The PGC then gradually translocate into the for germline chimera production through the detection underlying hypoblast (Karagenç et al., 1996; D’Costa of reinjected PGC in the gonads of host embryos. and Petitte, 1999) and enter the germinal crescent phase of their migration during the gastrulation process (Ginsburg and Eyal-Giladi, 1987; Wentworth and MATERIALS AND METHODS Wentworth, 2000). During this period the hypoblast is displaced anteriorly, and the PGC is caught and carried Sources of Eggs and Egg Incubation along with it to the anterior region of the embryo. From Turkey (Meleagris gallopavo) eggs of strain BUT BIG the germinal crescent, the PGC enter the embryonic 5FLX were obtained from British United Turkeys Ltd. circulation associated with the forming blood vascular

(Chester, UK). Eggs were stored at room temperature, Downloaded from system at approximately stages 8 to 10 (Arabic nu- with 55 to 60% RH. All eggs were used within 7 d to merals refer to the staging system of Hamburger and ensure the best embryonic viability. Eggs were incu- Hamilton, 1951), and enter the vascular phase of their bated in forced-draft incubators (Brinsea Multihatch, migratory process. By stage 12, the extra- and intra- Mark II, Sandford, UK) at 38°C with 65% RH. Eggs embryonic blood system has been fully established and

were automatically rotated throughout incubation by a http://ps.oxfordjournals.org/ PGC can be found circulating throughout this system tray sliding mechanism at 1-h intervals. (Fujimoto et al., 1975, 1976). Chick PGC appear in the blood in increasing numbers from stage 12 up to stage 16 where numbers peak and then begin to decline as Collection of Embryo Stages PGC gradually extravasate toward the gonadal anlagen 12 to 18 (2.5–3.5 d) (Nakamura et al., 2007), where, based on the sex of the embryo, they will either populate the ovary or testes as Eggs were secured in a 50-mL glass beaker and forceps were used to open the blunt end of the egg, expos- oogonia or spermatogonia at the time of hatch. These at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 unique features of avian PGC migration have helped ing the underlying air space. Any remaining eggshell or facilitate their subsequent isolation and transfer (Naka- shell membrane was removed, exposing the blastoderm, mura et al., 2010). located on the surface of the yolk, and recognized as a The potential for using avian PGC as a means of pre- pale to red rounded patch, depending upon the stage serving and reconstituting avian genetic material has of development. Embryos (stages 12–18) were removed been extensively demonstrated within chickens via the from the eggs by placing filter paper disks (2.3 cm, formation of germline chimeras. This process utilizes Whatman, Fisher Scientific, Loughborough, UK) with the potential for PGC transfer between individual em- a 1-cm hole cut from the center, over the embryo. Dis- bryos, and the ability of the donor cells to contribute to section scissors were used to cut around the circumfer- the germline of the recipient. Germline chimeric birds ence of the disk through the perivitelline layer. The have been produced previously by various means; us- disk with embryo attached was removed and inverted ing blastodermal cells (Petitte et al., 1990), germinal using forceps, and cleaned in warm PBS (pH 7.4) to crescent PGC (Vick et al., 1993), bloodstream PGC remove any adherent yolk. (Tajima et al., 1993; Naito et al., 1994a; van de Lavoir et al., 2006; Tagami et al., 2007; Yamamoto et al., Collection of Embryo Stages 2007), and frozen thawed bloodstream PGC (Naito et 29 to 30 (8–8.5 d) al., 1994b; Tajima et al., 2003; Kuwana et al., 2006), and gonadal PGC (Chang et al., 1995b, 1997; Park et Eggs were opened as for stages 12 to 18. A pair of al., 2003; Mozdziak et al., 2006). curved forceps was used to locate the embryo, which This previous research in chickens provides evidence was hooked under the neck and lifted from the eggshell. that PGC can be harvested, stored, and reintroduced The adherent yolk sac was removed and the embryos into the germline of recipient embryos. However, there is rinsed using PBS and placed under a dissection micro- relatively little information in the published literature, scope (WILD M3Z, Heerbrugg, Switzerland) in a Petri if any, about such methods to produce germline chime- dish containing black dissection wax and PBS. ras in turkeys. Early research by Reynaud (1969, 1976) demonstrated the incorporation of turkey PGC into the Embryonic Blood Collection gonads of chicken embryos via intravascular germinal crescent PGC transfer. However, these methods only Embryos were removed from the yolk as described produced interspecific embryonic chimeras, which dem- previously for stages 12 to 18. The embryos were placed onstrated no potential for producing donor cell-derived into Petri dishes and observed using a stereodissection offspring. Therefore, the aims of this study were to in- microscope. Blood was collected from the embryos by vestigate the feasibility of establishing a method for using a glass microinjection needle (Borosilicate glass, identifying and isolating PGC from turkey lines for the 15 cm length: 1 mm outside diameter, 0.75 mm inside purpose of producing a cryopreserved archive of valu- diameter), pulled by a Flaming Brown horizontal-P-97, able genetic material, and to demonstrate the potential Sutter, Novato, CA) puller. The needle was inserted TURKEY PRIMORDIAL GERM CELLS 801 into the embryonic dorsal aorta, and gentle suction ap- ation and after treatment with PA and Schiff’s reagent. plied to facilitate blood collection. Collected blood (1 All procedures were performed at room temperature, to 5 µL) was dispensed into individual wells of a 96-well and the stained PGC were observed under an inverted plate (Corning Inc., London, UK). Each well contained microscope (TE2000-U, Nikon, Tokyo, Japan). a 100 μL of Medium 199 (M199) supplemented with 10% fetal bovine serum (FBS; M199-FBS). Immunological Staining of Isolated PGC Embryonic Gonad Isolation Anti-OLP-1 (1B3, IgM; supernatant) antibody was obtained from Willi Halfter (University of Pittsburgh, Embryos at stages 29 to 30 were obtained using the Pittsburgh, PA). Both anti-EMA-1 (IgM; superna- method described above and placed in PBS. The em- tant) and anti-SSEA-1 (MC-480, IgM; ascites) were Downloaded from bryos were manipulated under a dissecting microscope purchased from the Developmental Studies Hybridoma using epi-illumination. Each embryo was positioned Bank (Iowa City, IA). Putative PGC were isolated from ventral side upwards, and extremely fine dissection for- embryonic blood as described previously. Primordial ceps (Dumont No 5, Sigma-Aldrich, Gillingham, UK) germ cells were first fixed in 4% PFA for 15 min and were used to open the abdomen along the ventral mid- washed in PBS before use. Primordial germ cells were http://ps.oxfordjournals.org/ line. Visceral organs were removed revealing the embry- incubated in an appropriate dilution of primary anti- onic mesenephroi and gonads. Each gonad was identifi- body (OLP-1, EMA-1; neat; SSEA-1; 1/100) for 1 h at able as an associated piece of whitish tissue attached 4°C. Primordial germ cells were then washed 3 times to the lateral edges of each mesenephros. Gonads were in PBS. Primordial germ cells were transferred into a removed from the mesenephroi using the forceps and 1/20 dilution of goat anti-mouse IgM FITC-conjugated transferred to a 1.5-mL Eppendorf tube (TreffLab, De- secondary antibody (F9259, Sigma-Aldrich). The cells

gersheim, Switzerland) containing 500 μL of M199-FBS were then incubated at 4°C overnight. After incuba- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 for storage before fixation for further analysis. tion the cells were washed 3 times in PBS. Cells were placed into bisbenzimide (Hoechst 33342, Sigma-Al- Isolation of Bloodstream PGC drich), 1/100 dilution, for nuclear counterstaining. A final wash in PBS was performed, and the cells trans- Ficoll density gradient centrifugation was performed ferred to glass slides for observation under a fluores- immediately after blood collection. Blood samples cence microscope (Olympus Vanox-AHBS3, Olympus, were collected into a single 1.5-mL Eppendorf tube. Southend-on-Sea, UK). No more than 10 to 20 embryos worth of blood was placed into each Eppendorf tube. Collected blood was Staining of Gonad Tissue Sections pelleted in a cooled centrifuge spun at 20°C (Sigma laboratory centrifuges, Philip Harris, model 3K18C, Embryonic gonads were isolated as described above, rotor 12154) for 4 min at 400 × g. Both 16 and 6.3% and frozen in OCT embedding medium (Tissue-Tek, Ficoll 400 (Sigma-Aldrich) solutions were prepared Raymond A Lamb, Eastbourne, UK). Frozen sections and used for the isolation procedure as per the method were obtained by cutting using a cryostat (Kryostat of Yasuda et al. (1992). Briefly, the blood pellet was 1720, Leica, Milton Keynes, UK) at a thickness of 7 to resuspended in 90 μL of M199-FBS. The resuspended 10 μm, and collecting onto poly-l-lysine-coated slides. pellet was dispersed into 900 μL of the 16% Ficoll A 1/10 dilution of normal goat serum/PBS with 5% and mixed thoroughly. Two hundred microliters of the BSA (Sigma-Aldrich) blocking solution was applied at 6.3% Ficoll was overlaid on top of the 16% Ficoll solu- 50 μL per sample. Samples were placed in a humidified tion, and an interface line became visible between the box for 1 h and incubated at room temperature. Slides 2 concentrations and was marked. The tube was spun were washed 3 times in PBS on an orbital shaker (Stu- for a further 30 min at 800 × g. After this spin, 200 μL art Scientific, Stone, UK) at 5-min intervals. The PBS was removed from the interface. The removed solution was removed and primary antibody added (50 μL) per was dispersed in M199-FCS (900 μL) and centrifuged section at the concentrations described above. Samples again 3 times for 4 min at 400 × g to remove residual were incubated at room temperature for 1 h or 4°C Ficoll. overnight, in a humidified chamber. The samples were again washed 3 times with PBS. Secondary antibody Periodic Acid-Schiff Staining was added (50 μL), and incubated in the humidified chamber at room temperature for 1 h. Secondary anti- Isolated cells were fixed in 4% paraformaldehyde body was removed, and samples washed as above, then (PFA) for 10 min. The cells were then immersed in incubated with 1/1,000 bisbenzimide/PBS for 15 min. 0.1% periodic acid (PA) solution (Sigma-Aldrich) for After 3 washes in PBS, samples were mounted in Gel- 5 min and subsequently treated with Schiff’s reagent vatol (pH 7.4). Slides were placed into slide wallets, (Sigma-Aldrich) for 15 min. Cells were washed in PBS wrapped in aluminum foil, left to dry for 24 h, and by centrifugation at 400 × g for 3 min at 20°C after fix- stored at –20°C until analysis. 802 Wade et al. PGC Cryopreservation and Recovery Detection of Reinjected PGC The PGC were isolated from embryos between stages Gonads were removed from stage 29 to 30 embryos as 12 to 18 of development. Isolated cells were placed into described above. Whole mounts of gonadal tissue were 1 mL of freezing mix [Dulbecco’s modified Eagle medi- produced by fixing in 4% PFA for 10 min at 4°C. After um (DMEM)] supplemented with 20% FBS and 10% fixation, the tissue was washed in PBS for 1 h. Gonads dimethlysulfoxide contained within a cryotube (NUNC were placed into 1 mL of bisbenzimide diluted 1/100 in Cryotube Vials, Hvidovre, Denmark). Cryotubes were 1% Triton X-100 (Sigma-Aldrich) diluted in PBS for 2 placed in a freezing vessel (Nalgene cryo 1°C, Nalgene h at room temperature. The tissue was washed in PBS Nunc International, Loughborough, UK) supplemented for 1 h at room temperature. Gonads were dehydrated with 250 mL of isopropyl alcohol (Sigma-Aldrich) pro- through a series of graded alcohol solutions −50, 70, Downloaded from viding a controlled rate of cooling at −1°C/min. The 90%, and a 50/50 alcohol glycerol solution over a 1-h freezing vessel was placed at –80°C and left overnight. period. Gonads were placed into a 100% glycerol so- The cryotubes were removed from the freezing vessels, lution (BDH Lab Supplies, UK) and allowed to clear placed into liquid nitrogen at –196°C within a liquid ni- overnight and stored at 4°C. trogen cell storage unit (34HC, Taylor-Wharton, Theo- The whole mount gonads and number of remigrated http://ps.oxfordjournals.org/ dore, AL), and stored for 3 mo. To recover the frozen PGC were analyzed using a Nikon C1 Eclipse 90 up- PGC, the cells were removed from liquid nitrogen and right confocal microscope. The nuclear counterstain placed into water at 37°C for 5 min. After thawing, the bisbenzimide was excited using a blue 408 nm laser, cell suspension was immediately diluted in 10 mL of and the emitted fluorescent light was collected using a DMEM and centrifuged at 200 × g at 20°C for 15 min. 450/35 emission filter. Vybrant CM-DiI- and Q-track- The supernatant was removed and cells resuspended er-labeled samples were detected using a red 568 nm in 100 µL of M199. The viability of the frozen thawed laser and emitted fluorescence was collected using a at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 PGC was determined by Trypan blue (T8154, Sigma- 600/30 emission filter. Analysis of gonads in the z-axis Aldrich) dye exclusion. An equal volume of cell suspen- was performed by taking a z-series of individual optical sion (5 μL) was mixed with an equal volume of Trypan image slices through the tissue at different focal lev- blue and left to incubate at room temperature for 1 to els, between 100 and 150 steps at 2 to 5 µm per step. 2 min. Projections of the z-series images were produced using ImageJ v 1.34 image analysis software (National Insti- Preparation of Recipient (Ex Ovo) Embryos tutes of Health, Bethesda, MD). The numbers of remigrated PGC were assessed from the stacked projections Turkey embryos were incubated, as above, until they by extracting individual optical slices from the different reached stage 15. The eggs were then broken into a 100 focal planes that contained labeled PGC. The numbers × 20 mm tissue culture dish so that the embryo was of cells were then counted in these associated regions. oriented on the upper surface of the yolk once broken into the dish. The embryonic stage was then assessed and the plates placed into an incubator (LEEC, Not- Statistical Analysis tingham, UK) at 38°C until they were ready for PGC injection. Statistical analysis of data was carried out using GraphPad Prism (version 3.02, GraphPad Software Injection of Frozen-Thawed DiI/Q-Tracker- Inc., La Jolla, CA). This program helps organize, analyze, and graph repeated experiments and apply appro- Labeled PGC priate statistical tests. To determine if there was one or The PGC were isolated as described previously, and more significant differences located within a data set fluorescently labeled using either Vybrant CM-DiI (when comparing 2 or more groups) the 1-way ANOVA (V22888, Life Technologies, Molecular Probes, Paisley, test was used. If ANOVA analysis indicated significant UK) or Q-tracker 605 cell labeling kit, components A differences existed between the means of the groups and B (Q25001, Life Technologies, Molecular Probes) within the data, post hoc test analysis was applied. as per the manufacturer’s instructions. Labeled PGC Post hoc multiple comparison test analysis (Tukey’s were picked up into a custom-designed microinjec- honestly significant difference) identified exactly where tion capillary (Origio Ltd., Reigate, UK) produced to the significant differences lay within the data. This test the following specifications: borosilicate glass, 15 cm was considered the most appropriate as it was consid- length: 1 mm outside diameter, 0.75 mm inside diam- ered to be the most powerful test when analyzing all eter. The bore of the needle was produced at 40 µm and possible pair-wise comparisons. In cases where only 2 a 20° bevel was applied. The labeled PGC were then columns of data were to be compared, the t-test is the introduced into the heart of stage 15 embryos cultured corresponding ANOVA test. The unpaired t-test was ex ovo. The embryos were then placed into an incuba- used to analyze data to determine if a difference in tor (LEEC) at 38°C to incubate for the desired period the mean values existed when a particular variable was of time. being compared between 2 groups. Differences were TURKEY PRIMORDIAL GERM CELLS 803 regarded as significant at P < 0.05 unless otherwise stage 15, the numbers of PGC/embryo that could be stated. recovered off the Ficoll system dropped significantly (P < 0.05) between stages 15 and 16. By stage 18, RESULTS very few PGC could be recovered at all. This trend was also observed when results for the PGC/embryo data PGC Identification: Stages 12 to 18 were being expressed as PGC recovered per microliter of blood/embryo (Figure 2B). The periodic acid-Schiff (PAS) reagent positively identified the PGC, differentiating them from blood cells (Figure 1A). The PGC because of their charac- Viability of Frozen-Thawed PGC Downloaded from teristic high glycogen content stained a darker purple/ The PGC were removed from liquid nitrogen and magenta color than the surrounding smaller blood cells. thawed in a water bath. The recovered PGC were iden- These stained a lighter shade of pink because they con- tified on the basis of their morphological criteria, and tain smaller amounts of endogenous glycogen. Of the viability was found to be 90 ± 2% (mean ± SEM; Fig- antibodies tested (anti-OLP-1, anti-SSEA-1, and anti- ure 3) as assessed by Trypan blue dye exclusion after http://ps.oxfordjournals.org/ EMA-1) only anti-OLP-1 antibody showed any strong 3 mo stored in liquid nitrogen. The viability of thawed positive staining for the isolated PGC (Figure 1B). PGC was less than freshly isolated PGC whose viabil- Positive cells possessed characteristics of PGC that in- ity was 96 ± 3% (mean ± SEM; Figure 3). However, cluded; larger sized cells (15–20 µm), a granular ap- the freezing treatment was determined not to have had pearance, and numerous highly refractive lipid droplets a significant (P = 0.18) effect on cell viability when (Figure 1E, 1F). The anti-OLP-1 antibody showed pre- compared with the freshly isolated cells. This result dominantly strong staining on the cell surface (Figures would suggest that turkey PGC are capable of being

1B). Both the anti-EMA-1 and anti-SSEA-1 antibodies cryopreserved for prolonged periods in liquid nitrogen. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 showed very little to no staining at all on the isolated cells (results not shown). PGC Colonization of Host Gonad Tissue PGC Identification in Gonad Detection of Frozen-Thawed DiI-Labeled PGC. Serial cryosections of 8.5-d gonad tissue were labeled Host embryos were injected with 30 to 400 DiI-labeled with the antibodies anti-OLP-1, anti-SSEA-1, and PGC at stage 15 and had a 74% (n = 19) survival rate anti-EMA1. Results indicated that anti-EMA-1 and 6 d postinjection when cultured ex ovo. Uninjected con- anti-SSEA-1 failed to label gonadal PGC (results not trol embryos had an 80% survival rate (n = 15). The shown). However, the anti-OLP-1 antibody strongly la- removed gonads were examined by confocal microscopy. beled PGC within the gonad tissue (Figure 1C and Optical sectioning of whole mount gonads revealed DiI- 1D). The cells were identified as PGC on the basis of labeled PGC (Figure 4A, 4B). However, the numbers their large size (15–20 µm) and positive labeling. The of PGC that were detected in the gonads of these stage majority of the staining was restricted to the cell sur- 30 embryos was very low. The frequency of detection face, with some cytoplasmic labeling evident (Figure of the PGC was also low with 36 ± 13.2% (mean ± 1C and 1D). SEM) of embryonic gonads containing DiI-labeled PGC (Figure 5A). The colonization of the gonads with DiI- labeled PGC was also very low with 7 ± 3.8% (mean PGC Recovered off Ficoll at Stages 12 to 18 ± SEM) of reinjected PGC reaching the gonadal tissue After Ficoll centrifugation, a drop of media contain- of positive embryos (Figure 5B). It was also found that ing both blood cells and PGC was placed in the center 74% of PGC detected were located in the left gonad, of a tissue culture plate (100 × 20 mm). The PGC with the remaining 26% located within the right gonad were separated from the blood cells using a micropi- (data not shown). pette under a dissection microscope and counted. The PGC numbers per embryo were increased significantly Detection of Frozen-Thawed Q-Tracker- (P < 0.05) from stage 12 to stage 13. However the Labeled PGC greatest numbers of PGC per embryo were obtained from stages 13 to 15, with 32 ± 4.9, 33 ± 6.4, and 26 ± The PGC labeled with Q-tracker nanocrystals were 5.4 (mean ± SEM) PGC recovered respectively (Figure injected into the heart of stage 15 embryos, and were left 2A). However, no significant differences were observed to incubate until they reached stage 29/30. These in- between these groups (P > 0.05). The highest concen- jected embryos had a survival rate of 47% (n = 17). The tration of PGC on a per microliter of blood/embryo gonads were removed and analyzed using confocal mi- basis was obtained from stage 13 embryos (Figure 2B). croscopy. Optical sectioning revealed Q-tracker-labeled This number was highly significantly different from the PGC in the gonads (Figure 4C, 4D). The frequency of numbers of PGC recovered from stage 12 embryos (P < detection of PGC within the surviving embryos was 67 0.001) and stage 14 and 15 embryos (P < 0.01). After ± 21.1% (mean ± SEM; Figure 5A), which, although 804 Wade et al. Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 1. Putative turkey primordial germ cells (PGC) were identified in isolation by using the histochemical periodic acid-Schiff (PAS) reagent (A) and the immunological marker anti-OLP-1 (B). Primordial germ cells were also identified within the gonads of stage 30 embryos by using anti-OLP-1 antibody (C and D). Characteristic morphology of PGC isolated from blood of stage 12 to 18 embryos (E and F). Arrows indicate PGC; B = blood cells. Bars = 20 µm (E), 50 µm (A, B, D, F), 100 µm (C). higher than that of the DiI-labeled PGC (36 ± 13.2%; DISCUSSION mean ± SEM), was not deemed significantly different (P = 0.25). The colonization of gonads with Q-tracker- In the present experiment, PGC were efficiently iso- labeled PGC was greater than DiI-labeled PGC, with lated, cryopreserved, and reinjected/detected in the go- 44 ± 4.9% (mean ± SEM) of reinjected PGC colonizing nads of recipient turkey embryos. These results indicate the embryonic gonads in positive embryos (Figure 5B). that PGC could be used to preserve turkey germplasm, This was significantly different (P = 0.001) from that and potentially to produce germline chimeras. Such a of DiI-labeled PGC, where 7 ± 3.8% (mean ± SEM) method would provide a valuable strategy for recover- PGC colonized the gonads (Figure 5B). Of the PGC ing any turkey stocks that could inevitably be lost as detected, 70% were located in the left gonad and 30% a result of serious disease outbreaks. The principal aim in the right gonad (data not shown). of this study was to determine if PGC could be used TURKEY PRIMORDIAL GERM CELLS 805 Downloaded from http://ps.oxfordjournals.org/ Figure 3. Indicates the percentage of viable primordial germ cells (PGC) after a prolonged period (3 mo) stored in liquid nitrogen, compared with PGC freshly isolated after Ficoll density gradient centrifugation. Viability determined by Trypan blue dye exclusion. Values are the means ± SEM (n = 4).

efficient marker of chicken and rat PGC, where OLP was shown to label the surface of migrating chick PGC at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 from the embryonic circulation through to their colonization of the gonads (Halfter et al., 1996). Ovomucin- like protein (OLP) is defined as a subclass of glycoproteins with multiple O-linked carbohydrate side chains on a protein core with tandem repeat peptide units (Strous and Dekker, 1992). Ovomucin-like protein is a major constituent of the oocyte perivitelline membrane and has been suggested to possess an antiadhesive property (Halfter et al., 1996). It is thought that the Figure 2. Indicates the numbers of primordial germ cells (PGC) OLP confers this property to the surface of the PGC isolated from the blood of stage 12 to 18 embryos by using Ficoll den- (Halfter et al., 1996). Therefore, OLP may facilitate sity gradient centrifugation. The data were expressed as PGC isolated the migration of the PGC to the gonads by preventing on a per embryo basis (A), and per microliter of embryonic blood (B). Values are the means ± SEM. H&H = Hamburger and Hamilton. a precocious adhesion of the cells to blood vessel walls and to the mesenchyme. Two other immunological markers were tested, anti- to preserve turkey germplasm, and to demonstrate via EMA-1 and anti-SSEA-1. Both of these markers have their intravascular transfer to host embryos, the po- conventionally been used to identify chicken PGC tential to produce germline chimeras. Such a method (Urven et al., 1988; Jung et al., 2005). Anti-SSEA-1 has yet to be demonstrated for turkeys, and if feasible, has also been used to identify PGC from the germinal could provide a means to capture the entire genetics crescent and blood of stage 4 to 18 turkey embryos of the stock and recover a given line within only a few (D’Costa and Petitte, 1999). However, these markers generations (Moore et al., 2006). did not result in any positive staining of the putative In this study the immunological marker anti-OLP-1 turkey PGC in our study. D’Costa and Petitte (1999) identified isolated putative turkey PGC, both in isola- reported that turkey gonadal PGC lose the SSEA- tion and within the gonads of turkey embryos. This is 1 marker at approximately stage 20. This finding is the first report of the use of this marker to help char- contradictory to observations made by Halfter et al. acterize turkey PGC. The histochemical marker PAS (1996), who demonstrated the presence of the SSEA- also identified isolated putative turkey PGC within the 1 epitope on the surface of chick gonadal PGC. The blood of embryos. These markers provided a means of former observation is consistent with the findings in identifying PGC in turkey embryos within both the our study, although the observation in this report that blood and gonads at particular time points. The ob- the anti-SSEA-1 antibody does not bind to putative served PAS positive staining of PGC supports previous PGC derived from the blood is in contradiction with observations by Meyer (1960) for chicken PGC, and the previous study by D’Costa and Petitte (1999). The D’Costa and Petitte (1999) for circulating turkey PGC. failure of anti-EMA-1 to identify turkey circulating or The OLP-1 antibody was first demonstrated to be an gonadal PGC also contradicts the previous study by 806 Wade et al. Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 4. Donor primordial germ cells (PGC) labeled with Vybrant CM-DiI and Q-tracker nanocrystals were detected in chimeric stage 30 turkey gonads after injection into the hearts of stage 15 embryos. Confocal projections revealed Q-tracker-labeled PGC within gonads of stage 30 turkey embryos (A and B). Single confocal image slices show DiI-labeled PGC in gonads of stage 30 turkey embryos (C and D). Arrows indicate PGC, Go = gonads. Bars = 100 µm (A, D), 50 µm (B, C).

Urven et al. (1988). The reasons behind the negative were the focus of the isolation study, due to the relative results obtained for both the anti-SSEA-1 and anti- ease of access to the blood, fewer contaminating cell EMA-1 identification of turkey PGC are unclear. Be- types (in comparison with gonadal tissue), and knowl- cause both EMA-1 and SSEA-1 antibodies recognize edge that the PGC were in an active state of migration. different parts of the same epitope, this could indicate The technique was quick to perform and yielded many the possibility that the epitope has been lost or masked viable PGC. It was demonstrated that PGC numbers in some way. It may also suggest that chicken PGC peaked within embryonic turkey blood at stages 13 to have different characteristics than turkey PGC. 15. It was found that fewer PGC could be isolated from Putative PGC were identified after Ficoll isolation turkey embryos than chick using Ficoll. The highest on the basis of their large size and characteristic mor- concentration of turkey PGC within this study was at phology in comparison with blood cells. All of the fea- stage 13. These results are similar to those found by tures of the positively labeled cells were characteristic Tajima et al. (1999) for chick embryos, where the peak of avian PGC, along with the anatomical locations in concentration of PGC was at stage 14. Comparisons of which they were identified. Therefore, it was concluded PGC obtained from turkey with that of chicken indi- that the anti-OLP-1 and PAS-positive cells were PGC. cated that turkey embryos contain fewer PGC. Within This study reports on the use of Ficoll density gradi- chicken embryos PGC numbers rose to almost 70 PGC/ ent centrifugation to purify the PGC from the blood of µL of blood (Tajima et al., 1999), whereas in quail stage 12 to 18 turkey embryos a technique previously numbers peaked at around 100 PGC/µL of blood (Ono used for chicken PGC (Chang et al., 1992; Yasuda et and Machida, 1999). al., 1992; Tajima et al., 1993; Naito et al., 1994a,b, It was demonstrated that turkey PGC could be fro- 1998; Liu et al., 2007; Tagami et al., 2007). There is zen and thawed after 3 mo with high viability. Previ- no previous report on the use of this method to isolate ous work on cryopreserved chick PGC found them to turkey PGC. Circulating PGC from embryonic blood have similar viability (94.2%) after thawing (Naito et TURKEY PRIMORDIAL GERM CELLS 807 this study because, unlike PKH26, CM-DiI is retained in cells throughout fixation and permeabilization procedures, thus making it more suitable for examining the fixed/permeabilized whole mount gonad tissue, which was analyzed in this study by confocal microscopy. The incorporation of chicken donor PGC into recipient gonads has been demonstrated previously by labeling with the fluorescent dye PKH26 (Chang et al., 1995b; Park et al., 2003). The number of PKH26-labeled PGC detected in chicken gonads ranged from 487 to 1,160 after 150 were initially injected (Chang et al., 1995b), and Downloaded from the frequency of detection indicated 88.9% of surviving embryos had positive signals (Park et al., 2003). These numbers were higher than what was detected in turkey gonads when CM-DiI-labeled PGC were reinjected. It was noted that the majority of reinjected PGC were http://ps.oxfordjournals.org/ detected in the left gonad of turkey embryos, a finding that was corroborated in chick embryos, where 79.6 and 20.3% of reinjected PGC were found within the left and right gonads, respectively (Chang et al., 1995b). The poor detection of CM-DiI-labeled PGC could be attrib- utable to several factors. Studies using DiI-labeled mes-

enchymal stem cells indicated that labeled cells may at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 be rendered undetectable after several mitotic divisions as the fluorescence intensity of these cells decreased exponentially by 50% after each cell division (Ferrari Figure 5. A comparison between CM-DiI- and Q-tracker-labeled et al., 2001). The CM-DiI label could also influence primordial germ cells (PGC) detected within the gonads of host stage PGC survivability. In a previous study where CM-DiI 30 turkey embryos. Gonads removed from stage 30 embryos that had was tested for labeling human preadipocytes, it was received injections of CM-DiI and Q-tracker PGC at stage 15 were analyzed. The percentage number of surviving injected embryonic go- demonstrated that CM-DiI-labeled cells were up to 55 nads containing fluorescently labeled PGC (CM-DiI and Q-tracker) ± 21% viable 24 h after labeling, thus demonstrating was calculated and compared (A). The percentage number of labeled substantial toxicity (Hemmrich et al., 2006). It was due PGC colonizing the gonads compared with the numbers originally injected was also determined for both PGC labeling methods (B). N val- to this poor detection of CM-DiI-labeled PGC, that a ues for (A) represent total numbers of embryos analyzed for presence relatively new cell labeling technology (Q-tracker nano- of injected PGC; (B) represents total number of embryos with positive crystals) was used to better assess donor PGC incorpo- signals in which PGC were counted. Values are the means ± SEM. ration into host gonads. This method of fluorescently Significance is denoted by **P < 0.01. labeling avian PGC has not previously been reported. Encouragingly Q-tracker-labeled PGC were detected al., 1994a). Studies which used a range of commercial in the gonads of stage 30 embryos, with a higher fre- cryoprotectants to freeze chick PGC found PGC to be quency of detection and in higher numbers than CM- up to 86% viable after thawing (Setioko et al., 2007). DiI-labeled PGC. This result suggests that Q-tracker- A similar study in chick (Chen et al., 2007) used 3 dif- labeled cells offer a more efficient means of assessing ferent cryoprotectants and protocols for freezing chick donor PGC incorporation into host turkey gonad tissue PGC, and the viability was found to be 85.9%, which than CM-DiI-labeled cells. It is likely that PGC reis comparable to the study by Setioko et al. (2007). migration using Q-tracker was more effective because For turkey PGC the viability was higher than previous Q-tracker nanocrystals are thought to offer a greater studies using chick PGC (Chen et al., 2007; Setioko et photostability (Alvisatos, 2004), making them resistant al., 2007), but not as high as that found by Naito et to photobleaching, thus enabling long-term imaging ex- al. (1994a). This study suggests that it is feasible to periments to be performed under conditions that may produce a frozen archive of turkey PGC, thus making it lead to the photoinduced deterioration of other more possible to conserve rare genetic resources from turkey commonly used fluorophores such as CM-DiI. species. Overall the results from this study have helped elu- The reinjection of frozen-thawed fluorescently la- cidate some important questions with regards to the beled turkey PGC indicated that these cells retained archiving of turkey PGC and the potential for germ- the potential to repopulate the host gonads. Initial ex- line chimera production in this species. Although no periments using CM-DiI-labeled PGC revealed donor live chimeras or donor-derived offspring were generated PGC in host gonads by confocal microscopy. However, from this study, these were important first steps to be the number and proportion of reinjected turkey PGC addressed. It was demonstrated that it is possible to detected was low. The CM-DiI was initially chosen in identify, isolate, cryopreserve, and recover turkey PGC 808 Wade et al. with a high viability, demonstrating that it is possible the area pellucida irrespective of the embryo-forming process. to produce a frozen genomic archive of turkey genet- Development 101:209–219. Gupta, S. K., and M. R. Bakst. 1993. Turkey embryo staging from ic material. The reinjection and detection of frozen- cleavage through hypoblast formation. J. Morphol. 217:313–325. thawed fluorescently labeled PGC in host gonads using Halfter, W., B. Schurer, H. M. Hasselhorn, B. Christ, E. Gimpel, confocal microscopy demonstrates the potential for es- and H. H. Epperlein. 1996. An ovomucin like protein on the sur- tablishing donor embryo genetics into the host germ- face of migrating primordial germ cells of the chick and rat. De- velopment 122:915–923. line. Hamburger, V., and H. L. Hamilton. 1951. A series of normal stages in the development of the chick embryo. J. Morphol. 88:49–92. ACKNOWLEDGMENTS Hemmrich, K., M. Meersch, D. von Heimburg, and N. Pallua. 2006. Applicability of the dyes CFSE, CM-DiI and PKH26 for tracking

Many thanks are given to the Biotechnology and Bio- of human preadipocytes to evaluate adipose tissue engineering. Downloaded from Cells Tissues Organs 184:117–127. logical Sciences Research Council (Swindon, UK) and Jung, J. G., D. K. Kim, T. S. Park, S. D. Lee, J. M. Lim, and J. Y. British United Turkeys (Chester, UK) for their spon- Han. 2005. Development of novel markers for the characterization sorship of the work. Much gratitude is also extended to of chicken primordial germ cells. Stem Cells 23:689–698. Willi M. Halfter (University of Pittsburgh, Pittsburgh, Karagenç, L., Y. Cinnamon, M. Ginsburg, and J. N. Petitte. 1996. Origin of primordial germ cells in the prestreak chick embryo. PA) for his kind donation of the anti-OLP-1 antibody. Dev. Genet. 19:290–301. http://ps.oxfordjournals.org/ Kuwana, T., T. Kawashima, M. Naito, H. Yamashita, M. Matsuzaki, and T. Takano. 2006. Conservation of a threatened indigenous REFERENCES fowl (Kureko Dori) using the germline chimeras transplanted Alivisatos, P. 2004. The use of nanocrystals in biological detection. from Primordial Germ Cells. J. Poult. Sci. 43:60–66. Nat. Biotechnol. 22:47–52. Liu, C. H., I. K. Chang, J. Sasse, C. J. Dumatol, J. V. Basker, and Bakst, M. R., S. K. Gupta, and V. Akuffo. 1997. Comparative devel- U. Wernery. 2007. Xenogenic oogenesis of chicken (Gallus do- opment of the turkey and chicken embryo from cleavage through mesticus) female primordial germ cells in germline chimeric quail hypoblast formation. Poult. Sci. 76:83–90. (Coturnix japonica) ovary. Anim. Reprod. Sci. 101:344–350. Blackburn, H. D. 2006. The National Animal Germplasm Program: McKay, J. C. 2009. The Genetics of Modern Commercial Poultry. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Challenges and opportunities for poultry genetic resources. Pages 3–9 in Biology of Breeding Poultry. Poultry Science Sym- Poult. Sci. 85:210–215. posium Series. Vol. 29. P. M. Hocking, ed. Chang, I. K., D. K. Jeong, Y. H. Hong, T. S. Park, Y. K. Moon, Meyer, D., B. 1960. Application of the periodic acid-Schiff technique T. Ohno, and J. Y. Han. 1997. Production of germline chimeric to whole chick embryos. Stain Technol. 35:83–89. chickens by transfer of cultured primordial germ cells. Cell Biol. Moore, D., P. H. Purdy, and H. D. Blackburn. 2006. A method Int. 21:495–499. for cryopreserving chicken primordial germ cells. Poult. Sci. Chang, I. K., A. Tajima, Y. Yasuda, T. Chikamune, and T. Ohno. 85:1784–1790. 1992. Simple method for isolation of primordial germ cells from Mozdziak, P. E., R. Wysocki, J. Angerman-Stewart, S. L. Pardue, chick embryos. Cell Biol. Int. Rep. 16:853–857. and J. N. Petitte. 2006. Production of chick germline chimeras Chang, I. K., A. Yoshiki, M. Kusakabe, A. Tajima, T. Chikamune, from fluorescence-activated cell-sorted gonocytes. Poult. Sci. M. Naito, and T. Ohno. 1995b. Germ line chimera produced by 85:1764–1768. transfer of cultured chick primordial germ cells. Cell Biol. Int. Naito, M., A. Tajima, T. Tagami, Y. Yasuda, and T. Kuwana. 19:569–576. 1994a. Preservation of chick primordial germ cells in liquid nitro- Chen, H. F., H. C. Kuo, C. L. Chien, C. T. Shun, Y. L. Yao, C. Y. gen and subsequent production of viable offspring. J. Reprod. Chuang, C. C. Wang, Y. S. Yang, and H. N. Ho. 2007. Derivation, Fertil. 102:321–325. characterization and differentiation of human embryonic stem Naito, M., A. Tajima, Y. Yasuda, and T. Kuwana. 1994b. Produc- cells: Comparing serum-containing versus serum-free media and tion of germline chimeric chickens, with high transmission rate of evidence of germ cell differentiation. Hum. Reprod. 22:567–577. donor-derived gametes, produced by transfer of primordial germ D’Costa, S., and J. N. Petitte. 1999. Characterization of stage-spe- cells. Mol. Reprod. Dev. 39:153–161. cific embryonic antigen-1 (SSEA-1) expression during early de- Naito, M., A. Tajima, Y. Yasuda, and T. Kuwana. 1998. Donor velopment of the turkey embryo. Int. J. Dev. Biol. 43:349–356. primordial germ cell-derived offspring from recipient germline D’Costa, S., S. L. Pardue, and J. N. Petitte. 2001. Comparative de- chimaeric chickens: Absence of long-term immune rejection and velopment of avian primordial germ cells and production of germ effects on sex ratios. Br. Poult. Sci. 39:20–23. line chimeras. Avian Poult. Biol. Rev. 12:151–168. Nakamura, Y., F. Usui, T. Ono, K. Takeda, K. Nirasawa, H. Ka- Eyal-Giladi, H., and K. Kochav. 1976. From cleavage to primitive gami, and T. Tagami. 2010. Germline replacement by transfer streak formation: A complementary normal table and a new look of primordial germ cells into partially sterilized embryos in the at the first stages of the development of the chick. Dev. Biol. chicken. Biol. Reprod. 83:130–137. 49:321–337. Nakamura, Y., Y. Yamamoto, F. Usui, T. Mushika, T. Ono, A. R. Ferrari, A., D. Hannouche, K. Oudina, M. Bourguignon, A. Meunier, Setioko, K. Takeda, K. Nirasawa, H. Kagami, and T. Tagami. L. Sedel, and H. Petite. 2001. In vivo tracking of bone marrow 2007. Migration and proliferation of primordial germ cells in ear- fibroblasts with fluorescent carbocyanine dye. J. Biomed. Mater. ly chicken embryo. Poult. Sci. 86:2182–2193. Res. 56:361–367. Ono, T., and Y. Machida. 1999. Immunomagnetic purification of vi- Fujimoto, T., T. Ninomiya, and A. Ukeshima. 1976. Observations able primordial germ cells of Japanese quail (Coturnix japonica). of the primordial germ cells in blood samples from the chick em- Comp. Biochem. Physiol. A Mol. Integr. Physiol. 122:255–259. bryo. Dev. Biol. 49:278–282. Park, T. S., D. K. Jeong, J. N. Kim, G. H. Song, Y. H. Hong, J. M. Fujimoto, T., A. Ukeshima, and R. Kiyofuji. 1975. (Origin and mi- Lim, and J. Y. Han. 2003. Improved germline transmission in gration of the primordial germ cells in the chick–light and elec- chicken chimeras produced by transplantation of gonadal primor- tron microscopic studies).Kaibogaku Zasshi 50:22–40. (in Japa- dial germ cells into recipient embryos. Biol. Reprod. 68:1657– nese) 1662. Fulton, J. E., and M. E. Delany. 2003. Poultry genetic resources- Petitte, J. N., M. E. Clark, G. Liu, A. M. Verrinder Gibbins, and R. operation rescue needed. Science 300:1667–1668. J. Etches. 1990. Production of somatic and germline chimeras in Ginsburg, M., and H. Eyal-Giladi. 1987. Primordial germ cells of the chicken by transfer of early blastodermal cells. Development the young chick blastoderm originate from the central zone of 108:185–189. TURKEY PRIMORDIAL GERM CELLS 809

Reynaud, G. 1969. Transfert de cellules germinales primordiales de Tajima, A., M. Naito, Y. Yasuda, and T. Kuwana. 1993. Production dindon a l’embryon de poulet par injection intravasculaire. J. of germ line chimera by transfer of primordial germ cells in the Embryol. Exp. Morphol. 21:485–507. domestic chicken. Theriogenology 40:509–519. Reynaud, G. 1976. Capacities reproductrices et descendence de pou- Urven, L. E., C. A. Erickson, U. K. Abbott, and J. R. McCarrey. let ayant subi un transfert de cellules germinales primordiales 1988. Analysis of germ line development in the chick embryo us- durant la vie embryonnaire. Archives of Developmental Biology. ing an anti-mouse EC cell antibody. Development 103:299–304. 179:85–110. van de Lavoir, M. C., J. H. Diamond, P. A. Leighton, C. Mather- Setioko, R. A., T. Tagami, H. Tase, Y. Nakamura, K. Takeda, and Love, B. S. Heyer, R. Bradshaw, A. Kerchner, L. T. Hooi, T. M. K. Nirasawa. 2007. Cryopreservation of Primordial Germ Cells Gessaro, S. E. Swanberg, M. E. Delany, and R. J. Etches. 2006. (PGCs) from White Leghorn embryos using commercial cryopro- Germline transmission of genetically modified primordial germ tectants. J. Poult. Sci. 44:73–77. cells. Nature 441:766–769. Strous, G. J., and J. Dekker. 1992. Mucin-type glycoproteins. Crit. Vick, L., Y. Li, and K. Simkiss. 1993. Transgenic birds from trans-

Rev. Biochem. Mol. Biol. 27:57–92. formed primordial germ cells. Proc. R. Soc. Lond. B Biol. Sci. Downloaded from Tagami, T., H. Kagami, Y. Matsubara, T. Harumi, M. Naito, K. 251:179–182. Takeda, H. Hanada, and K. Nirasawa. 2007. Differentiation of Wentworth, A. L., and B. C. Wentworth. 2000. The avian primordial female primordial germ cells in the male testes of chicken (Gallus germ cell, an enigma of germline evolution, development and im- gallus domesticus). Mol. Reprod. Dev. 74:68–75. mortality. Avian Poult. Biol. Rev. 11:173–282. Tajima, A., G. F. Barbato, T. Kuwana, and R. H. Hammerstedt. Yamamoto, Y., T. Ono, and H. Kagami. 2007. Dynamic analysis of 2003. Conservation of a genetically selected broiler line (42L) the developmental fate of cells in the center of the area pellucida using cryopreserved circulating primordial germ cells (PGCs) iso- of the blastoderm in chicken. J. Poult. Sci. 44:85–91. http://ps.oxfordjournals.org/ lated by filtration method. J. Poult. Sci. 40:53–61. Yasuda, Y., A. Tajima, T. Fujimoto, and T. Kuwana. 1992. A meth- Tajima, A., H. Hayashi, A. Kamizumi, J. Ogura, T. Kuwana, and od to obtain avian germ-line chimaeras using isolated primordial T. Chikamune. 1999. Study on the concentration of circulating germ cells. J. Reprod. Fertil. 96:521–528. primordial germ cells (cPGCs) in early chick embryos. J. Exp. Zool. 284:759–764. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Genetic parameters for feather pecking and aggressive behavior in a large F2-cross of laying hens using generalized linear mixed models

J. Bennewitz,* 1 S. Bögelein ,* P. Stratz ,* M. Rodehutscord ,† H. P. Piepho ,‡ J. B. Kjaer ,§ and W. Bessei *

* Institute of Animal Husbandry and Breeding, † Institute of Animal Nutrition, and ‡ Institute of Crop Science, University of Hohenheim, 70593 Stuttgart, Germany; and § Institute for Animal Welfare and Animal Husbandry, Downloaded from Friedrich-Loeffler-Institute, Dörnbergstr. 25-27, 29223 Celle, Germany

ABSTRACT Feather pecking and aggressive pecking is approximated on the link scale by the delta method. a well-known problem in egg production. In the present The results showed a heritability around 0.10 on the study, genetic parameters for 4 feather-pecking-related link scale for FPD and APD and of 0.04 for APR. The http://ps.oxfordjournals.org/ traits were estimated using generalized linear mixed heritability of FPR was zero. For all behavior traits, models. The traits were bouts of feather pecking deliv- substantial permanent environmental effects were ob- ered (FPD), bouts of feather pecking received (FPR), served. The approximate genetic correlation between bouts of aggressive pecking delivered (APD), and bouts FPD and APD (FPD and APR) was 0.81 (0.54). Egg of aggressive pecking received (APR). An F2-design production and feather eating records were collected was established from 2 divergent selected founder lines. on the same hens as well and were analyzed with a

The lines were selected for low or high feather pecking generalized linear mixed model, assuming a binomial at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 for 10 generations. The number of F2 hens was 910. distribution and using a probit link function. The heri- They were housed in pens with around 40 birds. Each tability on the link scale for egg production was 0.40 pen was observed in 21 sessions of 20 min, distributed and for feather eating 0.57. The approximate genetic over 3 consecutive days. An animal model was applied correlation between FPD and egg production was 0.50 that treated the bouts observed within 20 min as re- and between FPD and feather eating 0.73. Selection peated observations. An over-dispersed Poisson distri- might help to reduce feather pecking, but this might bution was assumed for observed counts and the link result in an unfavorable correlated selection response function was a log link. The model included a random reducing egg production. Feather eating and feather animal effect, a random permanent environment effect, pecking are genetically correlated and this needs fur- and a random day-by-hen effect. Residual variance was ther investigation. Key words: feather pecking, laying hen, heritability, generalized linear mixed model, overdispersion 2014 Poultry Science 93 :810–817 http://dx.doi.org/ 10.3382/ps.2013-03638

INTRODUCTION Quantitative genetic analyses in pullets have revealed low heritabilities with estimates typically around 0.06 to Feather pecking in laying hens is characterized by 0.12 (Bessei, 1984; Kjaer and Soerensen, 1997; Roden- nonaggressive pecks directed toward the plumage of burg et al., 2003). Higher heritabilities in the range of other hens (Kjaer et al., 2001). It is a well known, but 0.14 to 0.33 were reported when the trait was recorded yet unsolved problem in egg production. Feather peck- in the laying period (Kjaer and Soerensen, 1997; Kjaer ing causes economic losses due to higher feeding costs et al., 2001; Rodenburg et al., 2003). when larger parts of the body are denuded, and higher Aggressive pecking is clearly distinguishable from mortality rates when feather pecking is followed by feather pecking. Aggressive pecks are delivered in an cannibalism. The underlying mechanisms are not well upright body posture and are mainly directed toward understood, but physiological, nutritional as well as ge- the head of the recipient birds (Bilcík and Keeling, netic factors are known to influence this trait (Su et 1999; Kjaer et al., 2001). Like feather pecking, aggres- al., 2005; Wysocki et al., 2010; Kjaer and Bessei, 2013). sive pecking causes stress and welfare problems (Gross and Siegel, 1985). The interrelation between feather pecking and aggressive pecking is not clear. Early stud- © 2014 Poultry Science Association Inc. ies reported no (Blokhuis and Arkes, 1984) or a very Received September 20, 2013. Accepted December 16, 2013. low positive correlation between feather pecking and 1 Corresponding author: [email protected] aggressive pecking (Hughes and Duncan, 1972). Kjaer

810 GENETIC PARAMETERS FOR BEHAVIOR TRAITS IN LAYING HENS 811 et al. (2001) found no differences in aggressive pecks generations at the Danish Institute of Animal Science delivered in lines selected for high (HFP) and low (Kjaer et al., 2001) and then for 5 additional genera- feather pecking (LFP) for 3 generations. Bessei et al. tions at the Institute of Animal Husbandry and Breed- (2013a) compared the aggressive behavior of hens from ing, University Hohenheim, Germany. From the 10th the same line in the 10th generation of divergent selec- generation, an F1-cross was generated using an equal tion. The aggression level was significantly higher in number of birds from each line. These were mated re- the HFP than in the LFP. In addition, aggressive pecks ciprocally with males or females from the other line received and delivered were mostly observed between to generate 10 F1 families. From these families an F2- hens of the HFP line. Hens from the LFP line not only cross was established. One male from each of 10 F1 revealed a lower level of active aggression but also suc- families was mated with hens from the other families Downloaded from cessfully avoided aggressive encounters in general. In in 4 hatches of 3-wk intervals. This resulted in 10 F2 a subsequent study, Bessei et al. (2013b) studied the paternal half-sib families with an average number of 91 relationship between feather pecking and aggression female offspring, and in 910 hens in total. One-day-old in a large F2-cross established from 2 lines divergent chicks were neck banded for individual identification. selected on feather pecking. After categorizing the F2 Feeding, lighting program, and management were car- individuals in a high or low group with respect to the ried out under conventional pullet rearing conditions. http://ps.oxfordjournals.org/ number of feather pecks delivered, a close and nonlin- At 26 wk of age, the pullets were transferred to 1-tier ear relationship between aggressive pecks delivered and deep-litter pens with perches, nests, feeders, and drink- feather pecks delivered was found. ers. The group size varied between 36 and 42 birds. Recent investigations have shown that one motiva- The birds were marked with plastic tags on their back. tion for feather pecking is the consumption of feathers The behavioral data collection started at 27 wk of age. (McKeegan and Savory, 1999, 2001; Harlander-Mataus- Seven experienced observers recorded feather pecking

chek and Häusler, 2009; Kjaer and Bessei, 2013). McK- and aggressive pecking within each pen during ses- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 eegan and Savory (2001) found that birds identified as sions of 20 min each. Each pen was observed for 140 feather peckers ate more feathers from plastic dishes min per day and over 3 consecutive days making a to- when kept in individual cages. Harlander-Matauschek tal of 21 sessions of 20 min. All incidences (bouts) of and Bessei (2005) detected a significantly higher num- feather pecking and aggressive pecking were recorded. ber of feathers eaten in a high feather pecking line com- For each bout, the identity of the actor and receiver pared with a low line. It has been speculated that feath- was recorded. This resulted in 4 behavior traits: bouts ers may have a specific dietary effect. Feather keratin of feather pecking delivered (FPD), bouts of feather might not contribute much to the protein supply of the pecking received (FPR), bouts of aggressive pecking birds due to its low digestibility, but it might influence delivered (APD), and bouts of aggressive pecking re- the function of the gastrointestinal tract via its struc- ceived (APR). tural components and the microbial activity following For the feather eating test and the recording of egg degradation (Meyer et al., 2013). Furthermore, feather production the hens were transferred to individual cag- keratin is high in sulfur-containing amino acid concen- es at 29 wk of age. The cages were kept in a ventilated tration. To investigate if the interrelationship between windowless room with constant temperature. Over a feather pecking and feather eating is genetically moti- period of 10 consecutive days, a transparent plastic vated, the estimation of genetic correlations is needed. sheet containing 10 feathers of about 4 cm length was The aim of the present study was to estimate genetic fixed next to the feeder. The number of feathers eaten and phenotypic variance components for feather peck- was counted once a day and missing feathers were re- ing and aggressive behavior traits and feather eating placed. For analysis, the number of feathers eaten (FE) in a large F2-cross established from the high and low was the sum of feathers eaten over all 10 d. The number feather pecking selection lines mentioned above. Be- of eggs laid (LE) in a test period of 16 d was recorded cause the traits did not follow a normal distribution, for each hen. generalized linear mixed models were used. To investigate a putative correlated effect on egg production, this Statistical Analysis trait was included in the analysis as well. For the traits FPD, FPR, APD, and APR, the number of recorded pecks and received pecks, respectively, MATERIALS AND METHODS in each 20 min observation period were modeled as repeated observations in the statistical analysis. This Experimental Design and Data Collection resulted in 21 repeated observations per hen and trait across 3 d. Because the observations were count val- The research protocol was approved by the Ethi- ues they were initially assumed to be Poisson distrib- cal Commission of Animal Welfare of the Provincial uted and were analyzed with a generalized linear mixed Government of Baden-Wuerttemberg. Starting from a model (GLMM, e.g., McCullagh and Nelder, 1989; founder generation, 2 lines were divergently selected for Tempelman, 1998). The following linear predictor was low or high feather pecking. Selection took place for 5 used: 812 BENNEWITZ ET AL. η =+ + + +, Xb Zpen pen Zpe pe Z de de Z a a [1] Additionally, a within test-day repeatability was calculated as

where b is a vector of fixed effects including the ob- 22 2 2 22 2− 1 t =++()σσ σ/( σ ++++ σσ σφ λ). server, test-day, and observer-by-test-day interaction; w a pe de pen a pe de pen is a vector with random pen effects; pe is a vector with random permanent environment effects of the The traits LE and FE were analyzed with a general- hens; de is a vector of random test-day-by-hen effects; ized linear model as well. For LE the number of laid a is a vector with the random additive-genetic effects; eggs within a period of 16 d was modeled as a binomial and X, Zpen, Zpe, Zde, and Za are known design matri- trait with n = 16. The following linear predictor was

ces. The covariance structure of the random effects used: Downloaded from var( )= *σ2 , var( )= *σ2 , were pen I pen pe I pe 2 2 2 2 η =+, var( )= *σ , var( )= *σ , σ , σ , Xb Za a [3] de I de and aAa where pen pe σ2 , σ2 de and a are pen variance, permanent environmental variance, test-day-by-hen variance, and additive genet- where the vector b contains the fixed effect of the hatch ic variance, respectively; and A (I) is the numerator and the vector a contains the random additive-genetic http://ps.oxfordjournals.org/ var( )= *σ2 . effects with aAa The expectations of the ob- relationship (identity) matrix. The expectation of the −1 random effects was zero. The expectations of the obser- servations (vector y) were πη==Ey(|)()./ n a g The −1 vations (vector y) were λη==E(| y pen ,, a pe , de ) g (), link function was a probit link. The pedigree structure where g is the link function, in this case a log link, i.e., was the same as above. The models were fitted using = log (η ). ASReml 3.0 (Gilmour et al., 2009). A dispersion pa- g e As we detected over- and underdispersion relative to the Poisson model, a dispersion parameter rameter (ϕ) was added to the model via the variance

(ϕ) was added to the models on the observed scale by function on the observed scale; that is, at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 var( | , , , )= λφ . φπ()1− π assuming the variance function ypenapede var()yn/ = . The same statistical model was Overdispersion (ϕ > 1) or underdispersion (ϕ < 1) may n occur when the observations are more or less variable, applied to analyze the FE records. The observations respectively, than expected given the underlying Pois- were the number of eaten feathers from a total of n = son model. The pedigree included in total 1,059 indi- 100 supplied feathers. For the same reason described viduals starting from the selected founder animals to for the Poisson model, the estimation of the heritability generate the F1- and subsequently the F2-cross. The was not straightforward. The following approximation models were fitted for each trait separately using AS- was used, which is derived in the Appendix:

Reml 3.0 (Gilmour et al., 2009). Unlike in linear mixed 2 models, calculating the repeatability and heritability 2 σ h = a , [4] from the variance components is not straightforward σσ22+ ae for GLMM. This is because it is not obvious how the

residual variance in the heritability can be obtained 2 φπ()1− π where σ = , π =Φ()η is the expectation of the under a Poisson model, especially if ϕ ≠ 1. Foulley et e [()]ϕη 2 al. (1987) suggested approximating the heritability on n the linear predictor scale. This was applied by, for ex- binomial distribution, Φ is the cumulative probability ample, Vazquez et al. (2009), in analyzing mastitis data density function, and φ is the probability density func- in dairy cattle. However, the approach assumes ϕ = 1. tion of the normal distribution. The values of π and φ were estimated for each subject and then averaged over In the Appendix, an extension of the approach of Foul- φπ()1− π , ley et al. (1987) toward accounting for ϕ ≠ 1 is given. all subjects to obtain an average 2 which was Following this, the heritability on the η scale was com- [()]ϕη n puted according to used in equation [4]. Due to software restrictions, bivariate analyses con- 2 sidering 2 different nonnormal distributions of traits 2 σ h = a , [2] was not possible. Therefore, approximate genetic cor- σσσσλ2222++ + +φ − 1 pen a pe de relations (rg) between 2 traits were estimated according to Calo et al. (1973) as where λ is the average Poisson parameter. The Poisson parameter was estimated for each subject and then av- = rEBV , rg eraged over all subjects to obtain λ. Note that for ϕ = rel12* rel 1, the expression reduces to the suggestion of Foulley et al. (1987). Similarly, the repeatability was calculated as where rEBV is the Pearson correlation between the estimated breeding values (EBV) for the traits and rel1 =+()(σσ22/ σσ 22 + +++ σ 2 σ 2φ λ− 1). t apeapenpede (rel2) is the average reliability of the EBV for trait 1 GENETIC PARAMETERS FOR BEHAVIOR TRAITS IN LAYING HENS 813

Table 1. Trait, mean, SD, minimum (Min), and maximum (Max) of the observed trait values within the considered time period

Trait Period Mean SD Min Max Feather pecks delivered 20 min 0.64 1.97 0 34 Feather pecks received 20 min 0.63 1.25 0 16 Aggressive pecks delivered 20 min 0.26 0.81 0 14 Aggressive pecks received 20 min 0.25 0.73 0 10 Feathers eaten — 63.17 33.53 0 100 Eggs laid 16 d 11.10 4.39 0 16 Downloaded from (trait 2). The solutions for the hens with observations for LE and FE. The effects of the observer, test-day, from the GLMM [1] and [3] were taken as EBV. The and observer by test-day interaction were significant reliabilities were approximated from the correspond- for all 4 behavior traits (not shown). The heritability ing SE of prediction and the estimated additive genetic of APD and FPD are similar (0.10–0.11) and a little

variance (Mrode, 1996). higher than for APR (0.04). For these traits the perma- http://ps.oxfordjournals.org/ nent environment variance is substantial, especially for RESULTS FPD where it is more than 3 times larger than the additive genetic variance. This leads to a remarkably high The histograms of the behavior data are presented repeatability for FPD. For APD and APR, the repeat- in Figure 1. They show that the traits are not nor- ability is also substantial and more than 2 times larger mally distributed and there are large proportions of than the heritability. The FPR shows a heritability of zero counts. Summary statistics are given in Table 1. zero, but it still shows significant permanent environ- The average number of FPD and FPR is the same, but ment variance and thus a repeatability above zero. The at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 for FPD the SD is substantially higher, which can also test-day-by-hen variance component was large for all 4 be seen by comparing the 2 histograms. The same pat- behavior traits, especially for FPD. This lead to high tern can be observed when comparing the observations values for the within test-day repeatability. The disper- for APD and APR. The incidence of feather pecking is sion parameter indicated overdispersion for FPD and higher than that of aggressive pecking. FPR and underdispersion for APD and APR (Table 2). The results of the variance component analysis are The LE and FE showed a moderate to high heritability shown in Table 2 for behavior traits and in Table 3 (Table 3). The pen variance was small for all 4 traits.

Figure 1. Histogram of feather pecks delivered (FPD, top left), feather pecks received (FPR, top right), aggressive pecks delivered (APD, bottom left), and aggressive pecks received (APR, bottom right). Hen trait counts within observation periods of 20 min are shown. Trait values above 8 are shown in class 8. 814 BENNEWITZ ET AL.

(σˆ2), (σˆ2 ), Table 2. Estimated additive genetic variance a permanent environment variance pe test-day- (σˆ2 ), (σˆ2 ), (φˆ), (λ−1φˆ), by-hen variance de pen variance pen dispersion parameter residual variance heritabil- (ˆ2), (ˆ), (ˆ ) ity h repeatability t and within test-day repeatability tw for the behavior traits (SE shown in parentheses)1

Item FPD FPR APD APR σˆ2 a 0.46 (0.18) <0.01 (0.01) 0.42 (0.13) 0.17 (0.06) σˆ2 pe 1.40 (0.15) 0.20 (0.02) 0.53 (0.09) 0.30 (0.05) σˆ2 de 0.62 (0.03) 0.25 (0.02) 0.35 (0.03) 0.25 (0.03) σˆ2 pen <0.01 (0.01) 0.02 (0.01) 0.03 (0.02) 0.05 (0.02) Downloaded from ˆ φ 1.13 (0.01) 1.45 (0.01) 0.79 (0.01) 0.85 (0.01) λ−1φˆ 1.85 2.36 3.04 3.40 ˆ2 h 0.11 <0.01 0.10 0.04 tˆ 0.43 0.08 0.22 0.12 tˆ 0.57 0.17 0.30 0.18

w http://ps.oxfordjournals.org/ 1FPD = feather pecks delivered; FPR = feather pecks received; APD = aggressive pecks delivered; APR = aggressive pecks received.

The Pearson correlation coefficients between the sis of behavioral traits is generally problematic due to EBV of the traits and the approximate genetic correla- differing experimental populations and differing trait tions are shown in Table 4. Note that for FPR no cor- definitions and recording schemes. Especially if differ-

relations were estimated because heritability was esti- ent statistical models are applied, it is difficult to com- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 mated to be zero for this trait. The FPD was positively pare results across studies. In the present study, Pois- correlated with APD and APR. The 2 aggressive peck- son models were applied. The genetic parameters were ing traits (APD and APR) were moderately negatively estimated on the η scale (log scale), which was, except correlated. The FE was positively correlated with FPD, for the study by Kjaer and Soerensen (1997), not the FPR, and APD. The LE was positively correlated with case in the studies mentioned above. FPD and, to a lesser extent, with APD and APR. The SE of the variance component estimated (Table 2) are in general small given the number of hens with DISCUSSION observations. This might be due to the thorough observations of the traits using a standardized protocol. Ad- The 3 behavior traits FPD, APD, and APR were ditionally, the model accounted for the nonnormal dis- estimated to have medium to low heritability. The heri- tribution of the data and separated additive genetic tability of FPD was lower than those found by Craig effects from other hen-specific effects (permanent envi- and Muir (1993; 0.65), Kjaer et al. (2001; 0.19), and ronment and test-hen-by-day effects). The latter could Rodenburg et al. (2003; 0.12 to 0.15). However, they be done because the observations from each test period agreed well with those found by Kjaer and Soerensen were treated as repeated observations with an appro- (1997; 0.06, 0.14, and 0.38 at 6, 38, and 69 wk of age, priate covariance structure. The calculation of the heri- respectively) when taking age into consideration, which tability was done using an approximated residual vari- was around 27 wk in the present study. Rodenburg et ance on the η scale, allowing for ϕ ≠ 1 (see equation [2] al. (2004) found a zero heritability of APD and APR, and Appendix). No SE were reported for the heritabil- which is in contrast to the present results. In this study ity estimates because this would require a SE of the FPR showed a heritability of zero. This was also report- average Poisson parameter λ. The same arguments hold ed by Kjaer and Soerensen (1997) in 38-wk-old hens for the heritability estimates for LE and FE, where the and by Rodenburg et al. (2003) in 6- and 30-wk-old SE of several parameters forming the residual variance birds. However, comparing results from genetic analy- are unknown (equation [4]). The histograms of the behavior traits, defined as count values within a 20-min observation period, reveal an excess of zero values (Fig- ure 1). This resulted in relatively small average λ, which (σˆ2), Table 3. Estimated additive genetic variance a dispersion (φˆ), (σˆ2), (ˆ2) possibly yielded downward biased heritability esti- parameter residual variance e and heritability h for laid eggs (LE) and feathers eaten (FE; SE shown in parentheses) mates. A distribution that is able to model this excess explicitly might be more appropriate (e.g., a zero-in- Item LE FE flated Poisson distribution). However, the dispersion σˆ2 0.51 (0.17) 0.63 (0.21) parameter is close to 1 (Table 2), indicating that the a φˆ 2.24 (0.78) 24.90 (6.24) Poisson model fits the data well. σˆ2 0.31 0.47 Data from a large F2-cross, set up from lines select- e ed divergent on feather pecking, were used. Both lines ˆ2 0.40 0.57 h originated from the same founder population of laying GENETIC PARAMETERS FOR BEHAVIOR TRAITS IN LAYING HENS 815

Table 4. Approximate genetic correlation (above the diagonal) and Pearson correlation coefficient of estimated breeding values (EBV; below the diagonal; error probability in parentheses)

Trait1 FPD APD APR FE LE FPD 0.81 0.46 0.73 0.50 APD 0.34 (0.00) −0.54 0.23 0.29 APR 0.18 (0.00) −0.23 (0.00) 0.38 0.07 FE 0.36 (0.00) 0.12 (0.00) 0.19 (0.00) 0.04 LE 0.25 (0.00) 0.14 (0.00) 0.03 (0.16) 0.03 (0.40) 1FPD = feather pecks delivered; APD = aggressive pecks delivered; APR = aggressive pecks received; FE = feathers eaten; LE = eggs laid. Downloaded from hens. In the cross, the distribution of gene frequencies ing the molt has been reported by McKeegan and Sa- is not expected to be strongly U-shaped but closer to vory (1999, 2001) and Ramadan and von Borell (2008). intermediate values, which increases additive genetic Groups that showed extensive feather eating during the variance. This might be true especially for genes in- time of rearing tended to develop feather pecking dur- volved in feather pecking, the trait the lines were select- ing the laying period. Feathers decrease the passage http://ps.oxfordjournals.org/ ed for, but also for correlated traits due to correlated time of the digesta similarly to insoluble crude fiber selection response as well as for uncorrelated traits due (Benda, 2008). Recent experiments have shown that to genetic drift that operated within the lines. On the feeding feathers leads to significant changes in the gut other hand, it can be expected that some genes might microbiota (Meyer et al., 2013). It has been speculated be lost due to selection and drift within the selection that these changes in the digestive tract may be one lines before the cross was established. In any case, it factor for the persistency of feather pecking in indi- is difficult to compare the genetic variance and heri- vidual hens (Kjaer and Bessei, 2013). at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 tabilities reported in the present study based on an The way of modeling the data provided new insights. F2 experiment with those that would be expected or A remarkable result is the high permanent environment are reported in other populations. If selection against effect for the behavior traits, especially for FPD and feather pecking or aggression is to be performed, this APD, which was not reported so far. This indicates might result in a reduced egg production because of the that if hens delivered feather pecks or aggressive pecks positive genetic correlations, especially for FPD and once, they might continue in doing this. A similar result LE, but also for APD and LE (Table 4). The posi- was reported for behavior traits in pigs. Gäde et al. tive correlation between FPD and egg production is (2008) reported for sows a small additive genetic vari- in contrast, however, to results found during the early ance but a large permanent environment variance for generations of selection when comparing the pure lines. the trait savaging own piglets. But also for the traits In the 5th generation, egg number and egg mass in the APR and FPR there is a substantial permanent envi- low feather pecking line were higher than those in the ronment effect. This implies that once hens were identi- high feather pecking line during a 4-wk trial (24.4 vs. fied as “victims,” they stayed within this category. This 18.3 and 1,223 vs. 1,132 g, respectively; Su et al., 2006). could also be due to plumage damage, caused by re- The positive correlation between FPD and APD (Ta- ceived pecks, drawing attention from other hens, and ble 4) supports the finding of Bessei et al. (2013b), who stimulation of feather pecking (McAdie and Keeling, postulated a close relationship between these 2 traits. 2000). On the other hand, there are hens that success- However, it is in contrast to various previous studies. fully avoid receiving pecks. There seems to be a signifi- Blokhuis and Arkes (1984) found no difference in ag- cant day-to-day variation for the behavior traits. This gression in experimental groups that differed in feather is especially obvious for FPD and leads to the higher pecking. McKeegan and Savory (1999) found only a within test-day repeatability compared with the re- ˆˆ> low nonsignificant positive correlation between feather peatability in general, i.e., tw t (Table 2). One expla- pecks/pulls and aggressive pecks, but significant corre- nation might be a different response of the hens on lations of aggressive pecking with feather damages and changes in the day-to-day environment (e.g., temporary feather eating. fluctuations in temperature or light intensity). Feathers eating showed a moderate to high heritabil- Behavior traits depend on the interactions among ity (Table 3). This result might be facilitated by the individuals. Models were developed that include in- method implemented to record FE. The test was car- teraction or associated effects (Bijma et al., 2007a,b; ried out with individually housed adult layers under Bijma, 2013), and it was shown that these effects can standardized conditions and within a short time, which contribute substantially to the heritable variation of minimized environmental effects. The positive corre- survival of hens related to feather pecking and can- lation between FPD and FE (Table 4) supports the nibalism (Ellen et al., 2008). As suggested by Bijma hypothesis that FE and FPD have a common genetic (2013), we chose the simplest form to capture shared basis. There are several possible explanations. The con- environment effects and associated effects by fitting a sumption of cast feathers from the floor by pullets dur- random pen effect to the GLMM. The pen variances 816 BENNEWITZ ET AL. were small, but interpreting the magnitude of associ- and associative effects in three purebred layer lines. Poult. Sci. ated effects is not trivial (Bijma, 2013). Extending the 87:233–239. Foulley, J. L., D. Gianola, and S. Im. 1987. Genetic evaluation of GLMM toward a more sophisticated modeling of as- traits distributed as Poisson-binomial with reference to reproduc- sociated effects would be a logical next step. When de- tion characters. Theor. Appl. Genet. 73:870–877. signing selection schemes, multilevel selection (Bijma et Gäde, S., J. Bennewitz, K. Kirchner, H. Looft, P. W. Knap, G. al., 2007a,b) and also group selection (Muir and Craig, Thaller, and E. Kalm. 2008. Genetic parameters for maternal behavior traits in sows. Livest. Sci. 114:31–41. 1998; Ellen et al., 2007) should be considered. Gilmour, A. R., B. J. Gogel, B. R. Cullis, and R. Thompson. 2009. In conclusion, the behavior traits feather pecking ASReml User Guide 3.0. VSN International Ltd., Hemel Hemp- and aggressive pecking delivered and aggressive peck- stead, UK. ing received (FPD, APD, and APR) are heritable, and Gross, W. B., and P. B. Siegel. 1985. Selective breeding of chickens

for corticosterone response to social stress. Poult. Sci. 64:2230– Downloaded from selection might help to reduce feather pecking and ag- 2233. gressive pecking. Selection against these traits might Harlander-Matauschek, A., and W. Bessei. 2005. Feather eating and result in an unfavorable correlated selection response crop filling in laying hens. Arch. Geflügelk. 69:241–244. Harlander-Matauschek, A., and A. K. Häusler. 2009. Understanding reducing egg production. Feather eating and feather feather eating behavior in laying hens. Appl. Anim. Behav. Sci. pecking are genetically correlated, and this needs fur- 117:35–41. ther investigation. Hughes, B. O., and I. J. H. Duncan. 1972. The influence of strain http://ps.oxfordjournals.org/ and environmental factors upon feather picking and cannibalism in fowls. Br. Poult. Sci. 13:525–547. ACKNOWLEDGMENTS Kjaer, J., P. Soerensen, and G. Su. 2001. Divergent selection on feather pecking behavior in laying hens (Gallus gallus domesticus). Appl. Anim. Behav. Sci. 71:229–239. This study was supported by a grant from the Ger- Kjaer, J. B., and P. Soerensen. 1997. Feather pecking behavior in man Research Foundation (DFG, Bonn, Germany). J. White Leghorns, a genetic study. Br. Poult. Sci. 38:333–341. B. thanks Vanessa Grams (University of Hohenheim) Kjaer, J. P., and W. Bessei. 2013. The interrelationship of nutri-

tion and feather pecking in the domestic fowl–A review. Arch. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 for her support in editing the data sets. Geflügelk. 77:1–9. Lynch, M., and B. Walsh. 1998. Genetics and Analysis of Quantita- tive Traits. Sinauer Associates Inc., Publishers, Sunderland, MA. REFERENCES McAdie, T. M., and L. J. Keeling. 2000. Effect of manipulating feathers of laying hens on the incidence of feather pecking and Benda, I. 2008. Untersuchungen zu den Beziehungen von Feder- cannibalism. Appl. Anim. Behav. Sci. 68:215–229. picken, Exploration und Nahrungsaufnahme bei Legehennen. McCullagh, P., and J. A. Nelder. 1989. Generalised Linear Models. Dissertation der Universität Hohenheim, Institut für Tierhaltung Chapman and Hall, London, UK. und Tierzucht, Fachgebiet Nutztieretholgie und Kleintierzucht, McKeegan, D. E. F., and C. J. Savory. 1999. Feather eating in layer Stuttgart, Germany. pullets and its possible role in the aetiology of feather pecking Bessei, W. 1984. Untersuchungen zur Heritabilität des Federpickver- damage. Appl. Anim. Behav. Sci. 65:73–85. haltens bei Junghennen. I. Mitteilung. Arch. Geflügelk. 48:224– McKeegan, D. E. F., and C. J. Savory. 2001. Feather eating in in- 231. dividually caged hens which differ in their propensity to feather Bessei, W., H. Bauhaus, and S. Bögelein. 2013a. The effect of se- peck. Appl. Anim. Behav. Sci. 73:131–140. lection for high and low feather pecking on aggression-related Meyer, B., J. Zentek, and A. Harlander-Matauschek. 2013. Differ- behaviors of laying hens. Arch. Geflügelk. 77:10–14. ences in intestinal microbial metabolites in laying hens with high Bessei, W., S. Bögelein, J. Bennewitz, and J. B. Kjaer. 2013b. The and low levels of repetitive feather-pecking behavior. Physiol. relationships between social behavior and feather pecking in lay- Behav. 110–111:96–101. ing hens. Pages 58–64 in Proc. XIth Eur. Symp. Poultry Welfare, Mrode, R. A. 1996. Linear Models for the Prediction of Animal Uppsala, Sweden. Breeding Values. CAB International, Wallingford, Oxon, UK. Bijma, P. 2013. The quantitative genetics of indirect genetic effects: Muir, W. M., and J. V. Craig. 1998. Improving animal well-being A selective review of modeling issues. Heredity 112:1–9. through genetic selection. Poult. Sci. 77:1781–1788. Bijma, P., W. M. Muir, E. D. Ellen, J. B. Wolf, and J. A. M. Van Ramadan, S. G. A., and E. von Borell. 2008. Role of loose feathers Arendonk. 2007a. Multilevel selection 2: Estimating the genetic on the development of feather pecking in laying hens. Br. Poult. parameters determining inheritance and response to selection. Sci. 49:250–256. Genetics 175:289–299. Rodenburg, T. B., A. J. Buitenhuis, B. Ask, K. A. Uitdehaag, P. Bijma, P., W. M. Muir, and J. A. M. Van Arendonk. 2007b. Mul- Koene, J. J. van der Poel, and H. Bovenhuis. 2003. Heritability tilevel selection 1: Quantitative genetics of inheritance and re- of feather pecking and open-field response of laying hens at two sponse to selection. Genetics 175:277–288. different ages. Poult. Sci. 82:861–867. Bilcík, B., and L. J. Keeling. 1999. Changes in feather condition Rodenburg, T. B., A. J. Buitenhuis, B. Ask, K. A. Uitdehaag, P. in relation to feather pecking and aggressive behavior in laying Koene, J. J. van der Poel, J. A. van Arendonk, and H. Bovenhuis. hens. Br. Poult. Sci. 40:444–451. 2004. Genetic and phenotypic correlations between feather peck- Blokhuis, H. J., and J. G. Arkes. 1984. Some observations on the ing and open-field response in laying hens at two different ages. development of feather pecking in poultry. Appl. Anim. Behav. Behav. Genet. 34:407–415. Sci. 12:145–147. Su, G., J. B. Kjaer, and P. Søerensen. 2005. Variance components Calo, L. L., R. E. McDowell, L. D. VanVleck, and P. D. Miller. and selection response for feather-pecking behavior in laying 1973. Genetic aspects of beef production among Holstein-Friesian hens. Poult. Sci. 84:14–21. pedigree selected for milk production. J. Anim. Sci. 37:676–682. Su, G., J. B. Kjaer, and P. Søerensen. 2006. Divergent selection on Craig, J. V., and W. M. Muir. 1993. Selection for reduction in beak- feather pecking behavior in laying hens has caused differences inflicted injuries among caged hens. Poult. Sci. 72:411–420. between lines in egg production, egg quality, and feed efficiency. Ellen, E. D., W. M. Muir, F. Teuscher, and P. Bijma. 2007. Genetic Poult. Sci. 85:191–197. improvement of traits affected by interactions among individuals: Tempelman, R. 1998. Generalized linear mixed models in dairy cat- Sib selection schemes. Genetics 176:489–499. tle breeding. J. Dairy Sci. 81:1428–1444. Ellen, E. D., J. Visscher, J. A. M. Van Arendonk, and P. Bijma. Vazquez, A. I., D. Gianola, D. Bates, K. A. Weigel, and B. Hering- 2008. Survival of laying hens: Genetic parameters for direct stad. 2009. Assessment of Poisson, logit, and linear models for GENETIC PARAMETERS FOR BEHAVIOR TRAITS IN LAYING HENS 817

genetic analysis of clinical mastitis in Norwegian red cows. J. var⎡ −1 ()ηη** |⎤ = var⎡ exp() ηη | ⎤ ≈ Dairy Sci. 92:739–748. ⎣⎢g ⎦⎥ ⎣⎢ ⎦⎥ Wysocki, M., W. Bessei, J. B. Kjaer, and J. Bennewitz. 2010. Ge- 2 ⎛ * ⎞ 2 netic and physiological factors influencing feather pecking in ⎜∂λ ⎟ 2 ⎡ * ⎤ 222 chickens. World’s Poult. Sci. J. 66:659–671. ⎜ ⎟ ×=σησλσλ⎢exp()⎥ ×= =φ . ⎜∂η* ⎟ eee⎣ ⎦ηη*= ⎝ ⎠η**=η APPENDIX σλ21= φ − . This results in e The Poisson parameter was The challenge in defining heritability using our estimated for each subject and then averaged over all GLMM is that part of the nongenetic variation occurs subjects to obtain λ, which was used in equation [2] of on the observed scale, whereas the genetic effects occur

the main text. For the special case of ϕ = 1, this leads Downloaded from on the link scale. A natural idea, therefore, is to ask to the result of Foulley et al. (1987). what would be a suitable model for an error term e if it Over/underdispersed binomial distribution: could be specified entirely on the link scale as well. η* =+η . 2 Thus, we define e We may ask, what is a suit- ⎡∂Φ η ⎤ 2 ⎡ −1 **⎤ ⎡ ⎤ ⎢ ()⎥ 2 able definition for the variance of e ()σ on the link var⎢g ()ηη |⎥ = var⎢Φ() ηη | ⎥ ≈ ××=σ e ⎣ ⎦ ⎣ ⎦ ⎢ ⎥ e http://ps.oxfordjournals.org/ −1 * ⎢ ∂η ⎥ * scale, such that the conditional variance of g (η ) giv- ⎣ ⎦ηη= ⎡ −1 * ⎤ 2 ππ− var()ηη | , ⎡ * ⎤ 2 2 2 φ ()1 en η, ⎣⎢g ⎦⎥ approximately equals that of the ϕη() ×= σ⎡ ϕη()⎤ σ = . ⎣⎢ ⎦⎥ * ee⎣⎢ ⎦⎥ observed data y. To derive the conditional variance ηη= n var⎡ −1 ()ηη* | ⎤ ⎣⎢g ⎦⎥ we used the Delta method (Lynch and 2 φπ()1− π Walsh, 1998; p. 810–811) and approximated the vari- This results in σ = . The values of π and φ e ⎡ ⎤2 ance by a Taylor expansion around the mean η. This ⎣⎢ϕη()⎦⎥ n idea was already applied by Foulley et al. (1987) to the were estimated for each subject and then averaged over at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 φπ()1− π Poisson case, without considering a dispersion param- , all subjects to obtain an average 2 which was eter ϕ. We here extend the idea to over/underdispersed [()]ϕη n Poisson and binomial data. used in equation [4] of the main text. Over/underdispersed Poisson distribution: IMMUNOLOGY, HEALTH, AND DISEASE

In ovo inoculation of chicken embryos with probiotic bacteria and its effect on posthatch Salmonella susceptibility

J. E. de Oliveira,* 1 E. van der Hoeven-Hangoor,† I. B. van de Linde,† R. C. Montijn ,‡ and J. M. B. M. van der Vossen‡

* Cargill R&D Centre Europe, Animal Science, Havenstraat 84 Vilvoorde Flanders 1800, Belgium; † Cargill Animal Nutrition, Veilingweg 23, NL-5334 LD Velddriel, the Netherlands; Downloaded from and ‡ TNO, Microbiology and Systems Biology, PO Box 360, NL-3700 AJ, Zeist, Utrecht, the Netherlands

ABSTRACT The feasibility of establishing probiotic with challenged chicks fed antibiotic, but no significant bacteria in the intestine of broiler chickens by in ovo differences in mortality was observed. Probiotics of- inoculation was investigated, followed by verifying fered in ovo or through the diet could only partially re- http://ps.oxfordjournals.org/ possible subsequent protection against Salmonella En- cover performance compared with antibiotic-fed chicks. teriditis infection. In a first study, 7 commercially avail- A significant reduction in the number of Salmonella able probiotics were screened for compatibility with in Enteritidis positive chicks was observed when chicks ovo inoculation. Two of these probiotics, one being a were in ovo inoculated with E. faecium and continued Enterococcus faecium and the other a Bacillus subtilis, receiving it in the diet. This work establishes standards were selected for colonizing the chick gut without com- for future in ovo colonization research and emphasizes

promising hatchability. In a second study, these 2 prod- its value as a promising method to deliver individu- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 ucts were administered in ovo and in the feed to chicks al precise dose of probiotics to poultry in mass scale reared until 18 d in comparison with noninoculated at the earliest possible age based on the competitive chicks and with chicks fed an antibiotic. All chicks were exclusion concept. In ovo colonization with probiotic orally challenged with Salmonella Enteritidis at 4 d of can therefore become an important ally in combination age. Results showed reduced performance of Salmonella with other approaches to combat Salmonella and other

Enteritidis challenged chicks fed no additives compared intestinal bacterial infections in poultry. Key words: bacteria colonization, in ovo , Salmonella Enteritidis, probiotic 2014 Poultry Science 93 :818–829 http://dx.doi.org/ 10.3382/ps.2013-03409

INTRODUCTION to chance is an unnecessary risk if you can intention- ally inoculate them to probiotic bacteria (Cukrowska At birth, animals receive a natural inoculation of et al., 2002). Human studies have shown that there are microbes, which establish themselves into the intes- marked differences in timing of colonization and bacte- tine. The colonization of beneficial bacteria allows the rial populations between infants born vaginally com- animal to resist potential environmental challenges. In pared with those born by cesarean. These differences mammals, this inoculation happens during parturition, persist for months or even throughout the whole life of with bacteria living in the vaginal mucus, and through those individuals (Schultz et al., 2004; Rao et al., 2009), breast feeding (Penders et al., 2006; Flint and Gar- with the down side that cesarean born babies are usual- ner, 2009). In birds, colonization happens in the nest ly more susceptible to intestinal disorders (Grönlund et through contact with the hen and nest materials (Mills al., 1999; Penders et al., 2006; Cochetiere et al., 2007). et al., 1999). Modern poultry production excludes the According to Hashemzadeh et al. (2010), artificial incu- contact between chick and the hen, and depends mostly bation in poultry also delays enteric tract colonization on bacteria present in the hatchery and barn environ- by desirable microorganisms due to lack of contact with ment (Nisbet et al., 1994; Sterzo et al., 2005). Know- adult birds. Initial microbe colonization is not only im- ing that the first contact with microbes may include portant to prevent pathogenic bacteria to colonize by pathogens, leaving neonatal intestine colonization open competitive exclusion, but it is also very important to stimulate the development and maturation of the immune system (Cukrowska et al., 2002; Lu et al., 2003). © 2014 Poultry Science Association Inc. Gut health and function are key factors in the supply Received June 12, 2013. Accepted December 19, 2013. of all necessary nutrients for growth and maintenance, 1 Corresponding author: [email protected] especially in newborn animals with high growth rates

818 IN OVO INOCULATION OF CHICKS WITH PROBIOTICS 819 (Amit-Romach et al., 2004; Foye and Black, 2006). Un- tion of probiotic (Meijerhof and Hulet, 1997). With the wanted intestinal conditions caused by dysbacteriosis advances of in ovo inoculation techniques, it is possible or parasites can disrupt the crucial uptake of nutri- to present the poultry embryo’s digestive tract with ex- ents. Also, some bacteria present in the intestine of live- ternal material before hatching. This research explored stock can pose an important food safety risk to human the effects of inoculating chicken eggs with probiotic health upon consumption of animal products (Jain et bacteria before hatch and its impact on posthatch Sal- al., 2009). Unwanted intestinal conditions were usually monella colonization susceptibility. controlled by adding low doses of antibiotic growth promoters (AGP) and coccidiostats to commercial poul- MATERIALS AND METHODS try diets (Craven, 1995; Mateos et al., 2002; Buchanan Downloaded from et al., 2008; Flint and Garner, 2009). The agriculture Experiment 1 industry is reducing or even eliminating drug use as growth promoters in animal diets (Casewell et al., 2003; A total of 960 Aviagen Ross 308 broiler chicken eggs Barug et al., 2006; Buchanan et al., 2008; Menconi et were purchased from a commercial hatchery (Morren al., 2011). Therefore, many alternatives to AGP have BV, Lunteren, the Netherlands) and incubated under appeared in the market (Buchanan et al., 2008; Kim standard conditions (37.5°C and 54% RH). At arrival, http://ps.oxfordjournals.org/ et al., 2011). Ideally, the best approach to avoid drug the eggs were individually weighed and divided into 32 use would be to avoid pathogen colonization in the first flats of 30 eggs each, with similar average egg weights place. One way to achieve this can be intentional colo- per flat (67 g). Groups made of 2 flats of eggs were as- nization with known nonpathogenic bacteria, to create signed to 1 of 16 treatments, resulting in 60 eggs per an environment that prevents pathogen establishment treatment. The treatments consisted of 7 different com- (Cukrowska et al., 2002). In birds, colonization is con- mercially available probiotics inoculated at 2 doses, a

sidered to take place after hatching (Amit-Romach et negative (noninjected), and a positive (carrier-injected) at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 al., 2004), although there is scientific evidence indicat- control group. The list of probiotic bacteria in each ing that live bacteria can be found in small numbers product and their inoculated dose per egg is given in in the chick intestine before hatch (Pedroso, 2009; Bo- Table 1. horquez, 2010). Still, probiotics are routinely fed in pre- In Ovo Inoculation. At 17.5 d of incubation (E), starter and starter diets. Unfortunately, chicks can be all eggs except the negative control were injected with exposed to pathogens while still at the hatchery, dur- 500 μL of their respective solution into the amniotic ing hatching, sexing, vaccination, and transport, even fluid according to the procedure described by Tako et before they consume their first feed. Salmonella is still al. (2004). All eggs (including the negative control) re- one of the most prevalent food safety risks and has mained outside the incubator for the same amount of always been associated with poultry products (Craven, time (approximately 15 min). The eggs were disinfected 1995; Jain et al., 2009). Previous attempts to establish with 70% ethanol and then a hole was punched into intestinal microflora through in ovo inoculation failed the shell at the blunt end of the egg (air-cell chamber) due to high embryonic mortality and lower level of pro- with a 22-ga needle. The eggs were then injected using tection against Salmonella compared with oral applica- a repeating pipette (Eppendorf Repeater Plus Pipettor,

Table 1. List of treatments of experiment 1 with inoculated probiotics and dose

Treatment code Probiotic1 Dose NCON Negative control (noninjected) — PCON Positive control (carrier2-injected) — PRB1-1 Bacillus subtilis, Bacillus licheniformis 15 × 109 cfu/egg PRB1-2 B. subtilis, B. licheniformis 30 × 109 cfu/egg PRB2-1 B. subtilis, B. licheniformis 5 × 1012 cfu/egg PRB2-2 B. subtilis, B. licheniformis 10 × 1012 cfu/egg PRB3-1 Lactobacillus 37 mg/egg PRB3-2 Lactobacillus 75 mg/egg PRB4-1 Bacillus amyloliquefaciens 5 × 109 cfu/egg PRB4-2 B. amyloliquefaciens 10 × 109 cfu/egg PRB5-1 Enterococcus faecium 5 × 109 cfu/egg PRB5-2 E. faecium 10 × 109 cfu/egg PRB6-1 Lactobacillus fermentum, Lactobacillus helveticus, Lactobacillus paracasei, Lactobacillus salivarius, 3 mg/egg Pediococcus parvulus PRB6-2 L. fermentum, L. helveticus, L. paracasei, L. salivarius, Pediococcus parvulus 6 mg/egg PRB7-1 B. subtilis 16 × 109 cfu/egg PRB7-2 B. subtilis 32 × 109 cfu/egg 1All prebiotics are commercially available products suitable for posthatch feeding of broilers. Dose is preferentially presented in cfu, but alternatively in milligrams when product did not have cfu described in the label. The PRB1 and PRB2 were 2 different sources of the same bacteria with different cfu per gram. 2Saline solution was used as carrier for all probiotic solutions. 820 DE OLIVEIRA ET AL. Eppendorf, Nijmegen, the Netherlands) attached to a lands) for 2 min with zirconium beads (0.1 mm) and 23-ga needle. All eggs were injected by the same indi- phenol. After cooling on ice, the samples were centri- vidual to reduce variation. The shell holes were sealed fuged and the DNA containing the upper phase was with adhesive tape before eggs were returned to the transferred to new tubes for DNA purification with a incubator. commercial DNA isolation kit (Agowa, LGC Genom- Embryo Sampling. Twelve eggs per treatment were ics, Berlin, Germany), following the manufacturer’s in- randomly chosen for sampling at 19E (48 h after instructions (van der Hoeven-Hangoor et al., 2013). Se- oculation). The eggs were weighed and then opened rial dilutions of bacterial cells and their colony counts through the blunt end. Embryos were euthanized by were used as controls for the qPCR assay. To calculate cervical dislocation and dissected to weigh both yolk- CE, the amount of DNA per cell was assumed to be ap- free body and yolk sac. These weights were used to proximately 3 fg (Christensen et al., 1995; Button and Downloaded from calculate embryo yolk-free body mass (YFBM) and Robertson, 2001; Nadkarni et al., 2002), so CE/mL = yolk-sac size, both expressed as percentage of original (DNA Qty qPCR/3fg) × (1,000 μGl/150 μGl). This egg weight. The gizzards of sampled embryos were re- method was used to determine the bacterial count in moved, opened, and had their contents snapped-frozen the gizzard of embryos at 19E (48 h after inoculation) for PCR analysis of total bacterial DNA. Treatment and in the ceca of chicks at hatch. These organs were http://ps.oxfordjournals.org/ sampling order was rigorously followed, starting with chosen at these specific ages because they are filled noninoculated controls (NCON), carrier-injected posi- with digesta contents at these respective times. The tive control (PCON), and then bacteria inoculated second method was plate counting of 10-, 1,000-, and eggs (always from lower to higher dose) with full dis- 100,000-fold diluted samples. The bacteria were allowed infection of material and surfaces between treatments. to grow on differential agar media for 48 h to analyze Hatch Sampling. At hatch, the number of live the number of colony-forming units (cfu/mL). Based

hatched chicks and nonhatched chicks was counted to on our own laboratory tests previously performed with at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 determine hatchability of fertile eggs (%). Nonhatched these probiotic products, 2 media and incubation con- eggs were opened to determine cause of death. After ditions were used: tryptic soy agar under aerobic condi- all hatched chicks were weighed, 12 chicks from each tions (PRB1, PRB2, PRB4, and PRB7) and de Man, treatment were randomly selected for sampling. Chick Rogosa and Sharpe medium under anaerobic conditions length was measured (Hill, 2001; Lourens et al., 2006) (PRB3, PRB5, and PRB7). This method was used to and corrected for initial egg weight. The other external determine the number of cfu per milliliter in fresh cecal biometric parameter measured was chick leg shank rela- samples from chicks at hatch. Swabs were taken from tive asymmetry (RA) measured according to Yalcin et the incubator hatching trolley, table surfaces, and walls al. (2003). The chicks were then weighed, euthanized by in the hatchery after sampling to determine the number cervical dislocation, and had both their ceca removed. of bacteria cfu present in the environment. One cecum was used for PCR analysis and the other for plate culturing. Remaining yolk sacs were also weighed. Experiment 2 As for 19E sampling, chick YFBM and remaining yolk- sac weights were expressed as a percentage of the initial Production of Chicks Colonized with Probiotics. egg weight. A total of 450 Aviagen Ross 308 broiler chicken eggs Determination of Bacterial Counts. The bacterial from a 40-wk-old breeder flock were purchased from presence was determined at 2 time points, 2 d after the same commercial hatchery as in experiment 1, and in ovo inoculation (19E) and at hatch. Measurements incubated under standard conditions (37.5°C and 54% were done using 2 methods. The first method was de- RH). At arrival, the eggs were individually weighed and veloped using a universal quantitative PCR (qPCR), divided into 2 groups of 180 eggs, and subdivided into which universally targeted the 16S rRNA encoding groups of 10 eggs (replicates) with similar average egg bacterial DNA. The qPCR was performed using the weights (57 g). Each group was incubated in 1 of 2 Applied Biosystems 7500 Fast Real-Time PCR system identical incubators (NMC 2340, NatureForm Europe (Life Technologies Europe BV, Bleiswijk, the Nether- Limited, Oxfordshire, UK). Additionally, 2 groups of 45 lands) on optical grade 96-well plates. The reaction eggs were set as NCON, one in each incubator (NCON mixture contained Diagenode Master Mix (Diagenode incubator 1 and NCON incubator 2). The inoculated S.A., Liège, Belgium) and qPCR primers based on treatments consisted of 2 probiotics chosen from experi- the 16S ribosomal DNA sequence, being 16S-UniF1: ment 1 (PRB5 Enterococcus faecium and PRB7 Bacil- CGAAAGCGTGGGGAGCAAA (764–782), 16S- lus subtilis). Inoculations were done at one dose (PRB5- UniR1: GTTCGTACTCCCCAGGCGG (894–879), 1 at 5 × 109 cfu/egg and PRB7-1 at 16 × 109 cfu/egg) and 16S-UniP, FAM/MGB: ATTAGATACCCTGG- using the same method described in experiment 1. All TAGTCCA (787–807). eggs belonging to the same treatment were set in the The qPCR results were converted to cell equivalents same incubator to avoid cross contamination between (CE). In short, 150 μL of 10% dilution of the intestinal probiotics. Because it was not possible to avoid cross samples was homogenized in a BeadBeater (BioSpec contamination inside each incubator, there was no rea- Products Inc., Lab Services B.V., Breda, the Nether- son to produce groups of control chicks for the growth- IN OVO INOCULATION OF CHICKS WITH PROBIOTICS 821

Table 2. List of treatments of experiment 2 (grow-out study)1

Treatment In ovo probiotic Probiotic Antibiotic code Description inoculated in the feed in the feed NC Negative control — — — PC Positive control — — Apralan PRB5 PRB5 in diet — PRB5 — PRB5_IO PRB5 in ovo PRB5 PRB5 — PRB7 PRB7 in diet — PRB7 — PRB7_IO PRB7 in ovo PRB7 PRB7 — 1PRB5 = probiotic 5; PRB7 = probiotic 7. Downloaded from out trial, so all eggs in each machine were inoculated before Salmonella inoculation. The grow-out period except the 45 eggs used to verify cross-contamination. lasted 17 d, and the performance parameters calculated The objective was to produce 96 male chicks colonized at the end of the period were BW gain (WG), feed prior to hatch with each probiotic bacteria for a growth intake (FI), and feed conversion rate (FCR). Chick challenge study. performance data were also used to calculate the Euro- http://ps.oxfordjournals.org/ Hatch Evaluation. As done in experiment 1, the pean Performance Index (EPI; De Herdt et al., 1999). number of live hatched chicks and nonhatched chicks Prevalence of Salmonella was assessed at the end of were counted to determine hatchability (%). Non- the grow-out period by collecting cecal swabs from all hatched eggs were opened to determine cause of death. birds at 17 d and analyzing them for Salmonella pres- After all chicks were weighed and sexed, 10 female ence. Salmonella Enteritidis inoculum and subsequent chicks from each treatment were euthanized by cervical sample analysis were provided by Nutricontrol (Veghel,

dislocation and were dissected to measure YFBM and the Netherlands). at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 remaining yolk-sac, as done in experiment 1. Sampled All statistical analysis were performed using the Fit chicks also had their ceca removed to determine the Model procedure of JMP (SAS Institute Inc., 2008). number of bacteria present by plate culture the same Means were separated using least squares means Stu- way as described in experiment 1. Environmental sam- dent’s t-test (P < 0.05). ples from hatchery air, walls, and surfaces were also taken (swabs) to determine the bacterial load in the RESULTS room during sampling. Grow-Out Challenge Study. To test the posthatch Experiment 1 effects of chicks colonized with probiotics in ovo, a grow-out Salmonella challenge study was carried out. A In this experiment, several different probiotics was total of 288 Aviagen Ross 308 male chicks were distrib- tested for their dose and compatibility with in ovo in- uted into 6 treatments, with 9 replicates (cages) and oculation. Hatchability of chicks inoculated with differ- 6 chicks per replicate. Each cage was equipped with 2 ent probiotics in experiment 1 can be found in Figure nipple drinkers and 1 feed trough (outside the cage). 1. Great variability was found for hatchability between The wire cage floor did not allow the birds have contact different sources and doses of probiotics (Figure 1). In with excreta, and plastic partitions prevented birds ovo administration of carrier (PCON) or probiotic re- from one cage to contact birds in neighboring cages. duced hatchability by about 10% compared with non- At 3 d of age, all chicks were individually challenged injected NCON (98% hatch), with exception of PRB7-1 with 0.4 mL of 106 cfu of Salmonella Enteritidis by oral (95%; Figure 1). The results of biometric evaluation gavage. The treatments of the grow-out study are de- showed that chick length increased with PRB1, PRB2, scribed in Table 2. Chicks used in treatments PRB5_IO and PRB7 compared with PCON (P < 0.05). The RA and PRB7_IO were the in ovo inoculated chicks previ- also increased (P < 0.01) in chicks inoculated with ously described. Chicks for treatments NC, PC, PRB5, PRB2 and PRB6 compared with both control groups and PRB7 came from the same batch of eggs as treat- (Table 4). ments PBR5_IO and PRB7_IO, but were incubated at a commercial hatchery (no inoculation). All chicks had Experiment 2 similar average BW (41 g). Chicks were not vaccinated in ovo or posthatch, but tested negative for Salmonella Hatchability and Gut Colonization. The different presence. The experimental diet compositions can be characteristics of the 2 bacteria inoculated were clearly found in Table 3. A wheat basal diet containing NSP demonstrated by recovery of bacteria in the hatchery enzyme (isocaloric and isoproteic) was formulated to environment and the presence and viability of the inoc- fulfill or exceed the chicks’ requirements according to ulated bacteria in the chicks’ ceca (Figure 4). Bacillus CVB tables (CVB, 2006), to which the testing products (PRB7) was found in small amounts (<102 cfu/mL) in were added to subbatches at the expense of wheat. The the hatchery environment, and concentrated in incuba- diets were fed as fine mash. Feed and water were of- tor 1, which received the eggs inoculated with PRB7. fered ad libitum, with exception of 2 h feed restriction Enterococcus faecium (PRB5) was found in the incuba- 822 DE OLIVEIRA ET AL. Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 1. Hatchability (% of fertile) of eggs inoculated with different probiotics. NCON = negative control (noninjected); PCON = positive control (carrier-injected); PRB1-1 = probiotic 1, dose 1; PRB1-2 = probiotic 1, dose 2; PRB2-1 = probiotic 2, dose 1; PRB2-2 = probiotic 2, dose 2; PRB3-1 = probiotic 3, dose 1; PRB3-2 = probiotic 3, dose 2; PRB4-1 = probiotic 4, dose 1; PRB4-2 = probiotic 4, dose 2; PRB5-1 = probiotic 5, dose 1; PRB5-2 = probiotic 5, dose 2; PRB6-1 = probiotic 6, dose 1; PRB6-2 = probiotic 6, dose 2; PRB7-1 = probiotic 7, dose 1; PRB7-2 = probiotic 7, dose 2.

Table 3. Percentage composition of experimental diets

Negative Item control Antibiotic Probiotic Ingredient (%) Wheat 40.90 40.88 40.80 Corn 12.50 12.50 12.50 Soybean meal 22.58 22.58 22.58 Soybeans, heat treated 5.00 5.00 5.00 Isolated soy protein 1.00 1.00 1.00 Corn gluten meal 3.50 3.50 3.50 Potato protein 3.20 3.20 3.20 Soybean oil 3.35 3.35 3.35 Animal fat 3.00 3.00 3.00 Limestone 1.75 1.75 1.75 Monocalcium phosphate 1.32 1.32 1.32 Vitamin-mineral premix1 1.00 1.00 1.00 Sodium bicarbonate 0.22 0.22 0.22 Salt 0.19 0.19 0.19 dl-Methionine 0.24 0.24 0.24 l-Lysine HCl 0.24 0.24 0.24 l-Threonine 0.02 0.02 0.02 Enzyme (xylanase, β-glucanase) 0.005 0.005 0.005 Apralan (antibiotic) — 0.02 — Probiotic (PRB7 or PRB5) — — 0.10 Calculated nutrient composition CP (%) 24.3 24.3 24.3 Total P (%) 0.67 0.67 0.67 Available P (%) 0.43 0.43 0.43 Ca (%) 0.97 0.97 0.97 ME (kcal/kg) 2,950 2,950 2,950 1Supplied per kilogram of diet: riboflavin, 4.5 mg; niacinamide, 40 mg; d-pantothenic acid, 9 mg; choline chloride, 500 mg; dl-α-tocopherol, 30 mg; menadione, 2.3 mg; retinyl-acetate, 4.3 mg; cholecalciferol, 125 μg; biotin, 100 μg; folic acid, 0.5 mg; FeSO4·H2O, 147 mg; MnO2, 100 mg; CuSO4·5H2O, 40 mg; ZnSo4·H2O, 143 mg; Se(org), 0.5 mg; KI, 2 mg; antioxidant (ethoxyquin), 125 mg. IN OVO INOCULATION OF CHICKS WITH PROBIOTICS 823

Table 4. Biometric measurements of perinatal chicks inoculated in ovo with different probiotic bacteria1,2

% Embryo % Yolk-sac % Chick % Yolk-sac Chick length Chick length Relative Item YFBM E19 E19 YFBM E21 E21 (cm) (corr.) asymmetry NCON 50.9 17.6 71.5 8.02 17.9ab 26.5abcd 1.41bcd PCON 50.8 19.1 71.1 8.77 17.5bcde 25.5d 1.39bcd PRB1-1 52.2 16.2 69.9 8.02 17.1e 25.7cd 0.96cd PRB1-2 53.4 18.7 69.9 7.59 17.6abcd 27.3a 0.73d PRB2-1 49.5 19.0 70.5 8.02 17.6abcd 26.8abcd 2.93a PRB2-2 52.0 19.0 70.1 6.24 17.8abc 27.6ab 1.09cd PRB4-1 52.2 18.5 69.9 9.29 17.6abcde 26.7abcd 1.43bcd PRB4-2 53.0 18.6 71.1 8.97 17.4cde 25.7cd 2.37abc de cd ab PRB5-1 54.3 16.5 70.1 8.13 17.3 25.8 2.55 Downloaded from PRB5-2 52.6 19.3 69.6 8.52 17.7abcd 25.8cd 2.13abc PRB6-2 52.7 18.2 70.8 7.11 17.9abc 27.0abc 2.97a PRB7-1 52.9 17.5 69.5 8.81 17.8abc 27.2abc 2.37abc PRB7-2 50.5 18.0 70.3 8.64 18.0a 26.9abc 2.37abc P-value 0.342 0.431 0.682 0.400 0.017 0.049 0.006 SEM 2.195 1.507 1.015 1.047 0.187 0.584 0.591 http://ps.oxfordjournals.org/ a–eMeans with different superscripts within a column differ significantly (P ≤ 0.05). 1n = 12. 2YFBM = yolk-free body mass; E = day of incubation; corr. = corrected for initial egg weight; NCON = negative control (noninjected); PCON = positive control (carrier-injected); PRB1-1 = probiotic 1, dose 1; PRB1-2 = probiotic 1, dose 2; PRB2-1 = probiotic 2, dose 1; PRB2-2 = probiotic 2, dose 2; PRB4-1 = probiotic 4, dose 1; PRB4-2 = probiotic 4, dose 2; PRB5-1 = probiotic 5, dose 1; PRB5-2 = probiotic 5, dose 2; PRB6-2 = probiotic 6, dose 2; PRB7-1 = probiotic 7, dose 1; PRB7-2 = probiotic 7, dose 2.

tor that received eggs inoculated with these specific nella colonization reduced broiler performance (−10% at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 bacteria (incubator 2) and in open air, but at much WG, −5 points FCR, and −41 points EPI, P < 0.05) as higher numbers compared with Bacillus (PRB7; 2.5 × indicated by the significant difference between NC and 102 and 5 × 102 cfu/mL, respectively). This demon- PC (antibiotic, Table 5). In this trial, the performance strates that, first, the bacteria inoculated were viable of chicks that received probiotics, in ovo or in the feed, and correctly administrated because we could recover remained statistically intermediary. This was the case them in their respective incubators, and second, that E. for PRB5 (E. faecium) in the feed (only), which had the faecium (PRB5) is a much more versatile bacteria able best general performance not differing from PC (anti- to grow in both media and spread in the environment biotic) for any performance parameter (Table 5). Al- at higher numbers compared with Bacillus (PRB7; Fig- though, when chicks were inoculated in ovo with PRB5, ure 4A). their performance was not improved after Salmonella The NCON chicks hatched in incubator 1 had basi- infection. Chicks fed PRB7 (Bacillus) still showed low- cally no bacteria in their ceca (Figure 4B); this would er WG but similar FCR and EPI compared with PC be the expected result of a regular chick hatched in a (Table 5). In general, the performance of all chicks that noncontaminated and noninoculated environment (Nis- received probiotic, independent of bacteria or method, bet et al., 1994; Amit-Romach et al., 2004; Sterzo et al., was numerically superior to NC for all measured pa- 2005). Unexpectedly, NCON chicks hatched in incuba- rameters (Table 5). tor 2, where their counterparts were inoculated with PRB5, had the exact same bacteria profile and number DISCUSSION as the inoculated chicks (Figure 4B). Salmonella Challenge. The posthatch performance Oral inoculation with probiotic bacteria at birth has of the chicks in this present study showed that Salmo- been proposed since the beginning of the 1900s, and

Table 5. Performance (weight gain, WG; feed intake, FI; feed conversion rate, FCR; European performance index, EPI) of chicks in the period from 1 to 18 d, colonized with probiotics (PRB5, PRB7) in ovo (IO) or posthatch, when challenged with Salmonella Enteritidis

Item WG FI FCR EPI1 Negative control 494b 592 1.199a 248b Positive control 555a 636 1.146b 289a PRB7 512b 597 1.165ab 264ab PRB7_IO 520b 606 1.166ab 268ab PRB5 525ab 609 1.163ab 271ab PRB5_IO 498b 596 1.197a 251b P-value 0.0072 0.1080 0.0023 0.0026 SEM 13.33 12.38 0.0106 7.56 a,bMeans with different superscripts within a column differ significantly (P ≤ 0.01). 1EPI = [daily weight gain (g) × livability (%)/feed conversion ratio (g/g)] × 100. 824 DE OLIVEIRA ET AL. it is now recognized as a way to reduce risk of intes- al., 1992; Hashemzadeh et al., 2010; Hosseini-Mansoub tinal infection by pathogenic bacteria and allergic re- et al., 2011). Studies designed to compare the efficiency sponses in premature and newborn babies (Cukrowska of these different methods concluded that all of them et al., 2002; Lodinová-Zádníková et al., 2003). Feed- have some effect against Salmonella, but vent lip was ing beneficial bacteria directly (probiotic) or dietary the most effective in reducing the number of Salmo- components that favor these beneficial microbial popu- nella positive birds up to 7 d, followed by in ovo in- lations (prebiotic) or a combination of both (symbi- oculation into the air cell, spraying, and oral gavage otic) in livestock diets is becoming common practice (Hashemzadeh et al., 2010; Hosseini-Mansoub et al., (Mateos et al., 2002). Previous research showed that 2011). Another study concluded that in ovo administra- time of initial gut colonization by bacteria and sub- tion and immersion in broth did not prevent Salmonella strate of first diet may both play an important role on colonization (Yamawaki et al., 2013). There was there- Downloaded from future susceptibility to pathogen colonization (Barrow fore evidence that in ovo colonization was possible, but and Tucker, 1986; Amit-Romach et al., 2004). In a re- their commercial application stumbled on practical is- search review, Flint and Garner (2009) discussed how sues such as reduced hatchability and higher posthatch direct-fed microbials were able to significantly increase mortality (Cox et al., 1992; Meijerhof and Hulet, 1997). performance, and reduce morbidity and the incidence In this trial, some reduction on hatchability was caused http://ps.oxfordjournals.org/ of pathogens in poultry fed no AGP. Probiotic benefit by inoculation technique rather than by probiotic in- claims have been associated with boosting the immune oculation per se. Therefore, it shows that there is room status, reducing incidence and duration of diarrhea for improvement of the technique to a point where and other gastrointestinal disorders, protection against there will be no difference in hatchability between in- carcinogenic substances, reducing development of aller- jected and control chicks. Lower hatchability seems to gies, and reducing risk of infections of respiratory and be mostly associated with contamination followed by

genitourinary systems (Senok et al., 2005). Artificial uncontrolled bacteria growth. Although, Edens et al. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 incubation prevents contact of chicks with adult birds. (1997) proved that it was possible to successfully in- This causes a delay in gut colonization with desirable oculate Lactobacillus reuteri in ovo through the air cell microorganisms and exposes them to risk of coloniza- and amnion without reducing hatchability. All reported tion with pathogenic bacteria (Hashemzadeh et al., work since then has concentrated in air cell inoculation 2010). Salmonella is commonly found in the intestinal of different bacteria, where hatchability was affected population of chicken and other birds (Stern, 2008). (Meijerhof and Hulet, 1997; Hashemzadeh et al., 2010; The serotypes that cause foodborne illness are not nec- Hosseini-Mansoub et al., 2011; Yamawaki et al., 2013). essary pathogenic to birds. After Salmonella establish- The reasons for these differences can be attributed to ment occurs, primarily in the ceca (Nisbet et al., 1994), factors such as delivery technique, site of injection (air it is very difficult to eliminate and reduce shedding cell vs. amnion), type of probiotic bacterium, and inoc- without using specific antibiotics. Therefore, reducing ulated dose. The great variability found in experiment Salmonella colonization and shedding to the environ- 1 between different sources and doses of probiotics on ment is a very important food safety measure (Craven, hatchability (Figure 1) demonstrates the importance of 1995). Although some studies have shown that perina- considering several parameters to successfully inoculate tal treatment of chickens with probiotic bacteria could bacteria in ovo. Among those, time of inoculation, site significantly reduce Salmonella colonization and fecal of injection, bacteria species and strain (as exempli- shedding (Barrow and Tucker, 1986; Hashemzadeh et fied by lethality of PRB3 regardless of dose), and dose al., 2010; Hosseini-Mansoub et al., 2011), others have (bacterial load) can be mentioned. Dose was especially found no effect (Yamawaki et al., 2013). Because con- important for some of the probiotics tested in experi- tact between the chick and adult birds during hatch re- ment 1. The PRB6 and PRB7 had hatchability reduced mains an impossibility, it has been proposed that chicks by half or more as the dose of bacteria doubled. It was should be administered with probiotic bacteria as early also demonstrated that probiotic PRB3 seems not to be in life as possible (Schneitz et al., 1992). The idea of suitable for in ovo inoculation because it was lethal to early inoculation of the gut with probiotic bacteria as embryos at all tested doses (Figure 1). a way to reduce the risk of colonization by pathogenic Some key distinctions between the previously men- bacteria is the basic concept of competitive exclusion tioned studies and this current study are that they all being tested since the early 1970s (Nurmi and Rantala, injected probiotic solutions at lower volumes (0.1–0.2 1973; Cox et al., 1992; Meijerhof and Hulet, 1997). Pre- mL) into the air cell of incubating eggs around 18E, viously proposed methods for early delivery of probiot- whereas in this study 0.5 mL of probiotic solutions was ics include in ovo injection (Edens et al., 1997; Hossei- injected into the amniotic fluid at 17.5E. The main ni-Mansoub et al., 2011; Hashemzadeh et al., 2010), egg reasons for concentrating efforts on air cell injections immersion in probiotic broth (Yamawaki et al., 2013), were the work of Cox et al. (1992), which showed that oral gavage (Schneitz et al., 1992; Hashemzadeh et al., amnion injections of an undefined cecal culture of bac- 2010; Hosseini-Mansoub et al., 2011), vent lip (Hash- teria severely reduced chick hatchability, and the work emzadeh et al., 2010; Hosseini-Mansoub et al., 2011), of Edens et al. (1997), which showed no difference be- and chick spraying with probiotic solution (Schneitz et tween air cell and amnion inoculations. The results IN OVO INOCULATION OF CHICKS WITH PROBIOTICS 825 here presented confirmed the results of Edens et al. interest in having these 2 bacteria represented because (1997) that based on hatchability there is no reason to they are quite different from each other. The PRB5 (E. avoid amnion inoculation. In fact, amnion inoculations faecium) is a facultative anaerobic bacterium, whereas should be preferred because they overcome limitations PRB7 (B. subtilis) is an aerobic bacterium. Both Bacil- of other techniques such as imprecise dose reaching the lus and Enterococcus have been identified as part of a chick by air cell and spraying delivery, required toler- healthy cecal microflora in chickens fed diets without ance to low pH and digestive enzymes in oral gavage animal products, antibiotics, or anticoccidials (Lu et (embryonic digestion is not yet fully functional), and al., 2003). The differences observed in the recovery of practical issues of individual bird inoculation of oral these bacteria in the incubator environment and in the gavage and vent lip application (Hashemzadeh et al., ceca of their respective hatched chicks exemplify the 2010). Furthermore, in ovo amnion administration of great potential for fast colonization of the ceca when Downloaded from vaccines, nutrients, prebiotics, probiotics, and other chicks hatch in a contaminated environment. It also components should be combined into a single proce- demonstrates higher flexibility, survival, and growth dure to guarantee individual precise delivery of an in of the Enterococcus strain compared with the Bacillus ovo package of goods. strain in hatcher conditions in this study. Moreover, No other reports were found linking in ovo inocula- it shows that if you want to control and contain chick http://ps.oxfordjournals.org/ tion with probiotics and chick length or RA as found probiotic colonization, PRB5 is the best option, but if in experiment 1. Longer chicks have been associated you want to take advantage of competitive exclusion with better use of egg nutrients and higher posthatch by higher presence of a known bacterium and colonize growth, whereas higher RA has been associated with chicks without having to inoculate every egg, PRB7 poor skeletal development (Yalcin et al., 2003; Mei- would be more appropriate. These probiotics and their jerhof, 2006; Molenaar et al., 2008; Willemsen et al., inoculation conditions were then confirmed in in this

2008). Combining hatchability and biometrics results, second experiment, where hatchability of inoculated at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 the best 2 probiotics chosen were PRB5 and PRB7 at chicks was not statistically different from the controls the lowest dose. This choice was further confirmed by (Table 6). However, when comparing the 2 probiotics, the results of bacteria present in the gizzard 48 h after PRB5 numerically reduced hatchability compared with inoculation (Figure 2) and in the ceca at hatch (Fig- PRB7 (96.11 vs. 81.67%). It also yielded chicks 3.6% ure 3). These 2 treatments showed the highest num- lighter (P < 0.05) than PRB7 chicks expressed as a ber of bacteria in the contents of these organs (P < percentage of initial egg weight (Table 6). 0.05). Bacteria presence is a critical measurement to A lower level of protection against Salmonella of in prove that inoculated probiotic bacteria was alive and ovo air cell inoculations compared with vent lip has viable in the gut as it passed through the digestive been attributed to bacteria destruction in the upper di- tract during the hatching process. There was also an gestive tract (Hashemzadeh et al., 2010; Hosseini-Man-

Figure 2. Number of bacteria present in the gizzard of chicken embryos 48 h after in ovo inoculation with probiotics (PRB; all dose 1). NCON = negative control (noninjected); TSA = tryptic soy agar; MRS = DeMan, Rogosa, and Sharpe medium. Columns with different letters (a–c) differ significantly (P ≤ 0.05). 826 DE OLIVEIRA ET AL.

Table 6. Hatchability, BW, and yolk weight of chicks colonized with probiotics in ovo

All chicks2 Sampled chicks3

Item1 Egg wt Chick wt Hatchability (%) % Chick Chick wt Yolk wt % Yolk YFBM

NCON 57.9 43.4 100 74.89a 43.05 2.537 5.819 40.51 PRB5 57.3 40.4 96.11 70.37b 41.66 2.405 5.751 39.25 PRB7 57.1 42.6 81.67 73.97a 42.18 2.257 5.352 39.93 P-value 0.874 0.105 0.116 0.003 0.457 0.627 0.735 0.459 SEM 0.0375 0.039 36.34 3.8924 3.6443 0.8677 1.7349 3.483 a,bMeans with different superscripts within a column differ significantly (P ≤ 0.05). 1

NCON = negative control (noninjected); PRB5 = probiotic 5 (Enterococcus faecium); PRB7 = probiotic 7 (Bacillus subtilis). Downloaded from 2n = 18. 3n = 10. All sample chicks were female. YFBM = yolk-free body mass. soub et al., 2011), but the current findings showed that of infected chicks in the flock, but also on other gut, probiotic bacteria inoculated in ovo into the amnion immune, or metabolic conditions that depend on the http://ps.oxfordjournals.org/ were not killed in the upper digestive tract. microflora developed in the bird. In the case of PRB7 In this study, 2 methods to evaluate bacterial popula- (Bacillus), inoculated chicks in ovo or only after hatch tion were evaluated, which we believe to be complemen- (in feed) resulted in the same level of protection against tary. The qPCR method targeting 16S ribosomal RNA Salmonella colonization (Table 7), therefore not justify- (experiment 1) is a method able to detect and quantify ing the in ovo procedure. Edens et al. (1997) showed all bacteria populations representing the microflora, that the combination of in ovo and in feed supplemen-

whereas the traditional culture-based method (experi- tation of Lactobacillus reuteri significantly reduced at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 ments 1 and 2) can give an indication of how viable chick mortality under severe Salmonella Typhimirium the bacteria are once given favorable conditions (Amit- infection. As for the responses to E. faecium given in Romach et al., 2004). Similar comparison of qPCR and ovo (PRB5_IO) or only posthatch (PRB5), the better culturing methods concluded that qPCR could identify response with posthatch administration may be related bacteria populations faster and more precisely than to the environment created by these specific bacteria in culturing (Hashemzadeh et al., 2010). the gut when given at different times. Earlier exposure The results of microflora presence at hatch and Sal- to PRB5 (in ovo) resulted in a higher bacterial load monella presence at the end of the grow-out trial dem- in the ceca (Figure 4B), and lower Salmonella preva- onstrated that, in these conditions, even though Salmo- lence at 17 d (competitive exclusion), but it did not nella infection caused little mortality (data not shown), improve bird performance. Among the possible expla- it significantly reduced performance. This reduction in nations, this could mean creation of a microflora that performance may be dependent not only on the number was competing for nutrients, overstimulating the im-

Figure 3. Number of bacteria present in the gut after in ovo inoculation of probiotics (PRB; measured by quantitative PCR and converted to cell equivalents, CE/mL). (A) Number of bacteria cells in the gizzard of chicken embryos 48 h after inoculation (19E). (B) Number of bacteria cells in the ceca of chicks at hatch (all dose 1). NCON = negative control (noninjected). Columns with different letters (a,b) within the same chart differ significantly (P ≤ 0.05). IN OVO INOCULATION OF CHICKS WITH PROBIOTICS 827

Table 7. Salmonella presence in cecal swabs collected at 17 d from chicks colonized with probiotics (PRB5, PRB7) in ovo (IO) or posthatch, when challenged with Salmonella Enteritidis1

Positive Negative P-value

Item n % n % Compared with NC Compared with PC Negative control 29 81 7 19 — — Positive control 1 3 35 97 <0.0001 — PRB7 26 81 6 19 0.942 — PRB7_IO 31 86 5 14 0.527 — PRB5 28 78 8 22 0.772 — PRB5_IO 19 53 17 47 0.012 <0.0001 Downloaded from 1n = number of birds out of 36 total. NC = negative control; PC = positive control. mune system, or overproducing/lacking some intrinsic mice, prefeeding with dahi containing Lactobacillus for compound. 7 d before Salmonella Enteritidis inoculation reduced

These results can be better explained after consider- Salmonella colonization in liver and spleen by enhanc- http://ps.oxfordjournals.org/ ing resilience against Salmonella colonization presented ing the host immune system (Jain et al., 2009). Accord- in Table 7. As expected, the difference between NC and ing to Taghavi et al. (2008), Salmonella Typhimurium PC chicks was significant, so it was really due to higher infections early in life can cause high mortality, whereas Salmonella infection that performance was reduced in in older birds no clinical signs are seen. In this research NC chicks. This also shows that the antibiotic chosen chicks were challenged at 4 d, so we have intermedi- (Apralan) was very efficient (97%) against the Salmo- ate effects where mortality was not high but perfor- nella serovar tested (Tables 5 and 7). The intermediary mance was still compromised. Probiotic administration at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 performance results of PRB7, PRB7_IO, and PRB5 is not the only method to prevent Salmonella infection. can be explained by the fact that these treatments did It is also not probable that one single approach will not reduce cecal presence of Salmonella (still above be able to resolve Salmonella issues, especially in an 80%), but they could have reduced systemic invasion AGP-free environment, but a combined approach with of other organs as observed by Jain et al. (2009). And other practices such as farm biosecurity, feed strate- finally, the observation that PRB5_IO did not change gies (particle size, pelleting, fiber, mineral chelators, performance but increased resistance against Salmo- acidification, optimal amino acid balance, and addi- nella infection by 50% (highest among probiotic treat- tives), immune enhancers, and administration of other ments) indicates that factors other than competitive compounds such as DNA containing CpG motifs and exclusion can be of importance when it comes to bird bacteriophages (Taghavi et al., 2008; MacKinnon et al., performance. 2009; Vandeplas et al., 2010). Salmonella invades several organs affecting many This current research shows that the best approach physiological functions. Probiotic bacteria can poten- for colonization of chicks as early as possible can be tially reduce systemic invasion of Salmonella, therefore summarized as follows: inoculate embryos around improving performance. In an experiment done with 17.5E, into the amnion fluid, at a dose up to 109 cfu/

Figure 4. (A) Number of bacteria (cfu/mL) measured in the hatchery environment, and (B) in the ceca of chicks at hatch after in ovo inoculation of probiotics (PRB) at 17E (average, n = 10). Inc = incubator; NCON = negative control (noninjected); BSUB and BAC = Bacillus subtilis; ENTC and Entero = Enterococcus faecium; TSA = tryptic soy agar; MRS = de Man, Rogosa and Sharpe medium. 828 DE OLIVEIRA ET AL. egg of a previously screened source of probiotic bac- CVB. 2006. Veevoedertabel (Livestock Feed Table). Central Bureau teria (Edens et al., 1997; Tako et al., 2004; Uni et al., for Livestock Feeding, Lelystad, the Netherlands. De Herdt, P., R. Ducatelle, E. Uyttebroek, J. Hermans, A. Sneep, 2005; Smirnov et al., 2006). This research represents and R. Torbeyns. 1999. Reovirus serology in broiler parents and a further step in defining delivery method and condi- their progeny and its correlation with performance. Avian Dis. tions as well as a procedure to screen for the effects 43:271–278. of potential probiotics. The probiotic combination able Edens, F. W., C. R. Parkhurst, I. A. Casas, and W. J. Dobrogosz. 1997. Principles of ex ovo competitive exclusion and in ovo ad- not only to reduce but also to eliminate the presence ministration of Lactobacillus reuteri. Poult. Sci. 76:179–196. of Salmonella in broilers by market age still remains to Flint, J. F., and M. R. Garner. 2009. Feeding beneficial bacteria: A be determined. More work is also necessary to optimize natural solution for increasing efficiency and decreasing patho- conditions under a larger scale before moving to com- gens in animal agriculture. J. Appl. Poult. Res. 18:367–378.

http://dx.doi.org/10.3382/japr.2008-00133. Downloaded from mercial application. Foye, O. T., and B. L. Black. 2006. Intestinal adaptation to diet in the young domestic and wild turkey (Meleagris gallopavo). Comp. Biochem. Physiol. A Mol. Integr. Physiol. 143:184–192. ACKNOWLEDGMENTS Grönlund, M., O. Lehtonen, E. Eerola, and P. Kero. 1999. Fecal microflora in healthy infants born by different methods of delivery: The authors gratefully acknowledge the staff of the Permanent changes in intestinal flora after cesarean delivery. J. http://ps.oxfordjournals.org/ Cargill Animal Nutrition Innovation Center (Velddriel, Pediatr. Gastroenterol. Nutr. 28:19–25. Hashemzadeh, Z., M. A. Karimi Torshizi, S. Rahimi, V. Razban, and the Netherlands) for conducting of the animal study, T. Zahraei Salehi. 2010. Prevention of Salmonella colonization and C. Brouwer and A. M. T. Ouwens of TNO (Zeist, in neonatal broiler chicks by using different routes of probiotic the Netherlands) for the microbial analysis. This study administration in hatchery evaluated by culture and PCR tech- was partially supported by the Dutch government niques. J. Agric. Sci. Tech. 12:425–432. Hill, D. 2001. Chick length uniformity profiles as a field measure- through the Food Nutrition Delta initiative. ments of chick quality? Avian Poult. Biol. Rev. 12:188. Hosseini-Mansoub, N., T. Vahdatpour, M. Arjomandi, and S. Vah-

datpour. 2011. Comparison of different methods of probiotic pre- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 REFERENCES scription against Salmonella infection in hatchery broiler chicken. Adv. Environ. Biol. 5:1857–1860. Amit-Romach, E., D. Sklan, and Z. Uni. 2004. Microflora ecology of Jain, S., H. Yadav, and P. R. Sinha. 2009. Probiotic Dahi containing the chicken intestine using 16S ribosomal DNA primers. Poult. Lactobacillus casei protects against Salmonella Enteritidis infec- Sci. 83:1093–1098. tion and modulates immune response in mice. J. Med. Food Barrow, P. A., and J. F. Tucker. 1986. Inhibition os colonization of 12:576–583. the chicken caecum with Salmonella Typhimurium by pre-treat- Kim, G.-B., Y. M. Seo, C. H. Kim, and I. K. Paik. 2011. Effect ment with strains of Escherichia coli. J. Hyg. Camb. 96:161–169. of dietary prebiotic supplementation on the performance, intes- Barug, D., M. W. A. Verstegen, J. D. Jong, and A. K. Kies. 2006. tinal microflora, and immune response of broilers. Poult. Sci. Antimicrobial growth promoters: Where do we go from here? 90:75–82. Wageningen Academic Publishers, Wageningen, the Netherlands. Lodinová-Zádníková, R., B. Cukrowska, and H. Tlaskalova-Hogeno- Bohorquez, D. 2010. Nutritional influences on the ultra-structural va. 2003. Oral administration of probiotic Escherichia coli af- development of the small instestinal epithelium of the perinatal ter birth reduces frequency of allergies and repeated infections turkey embryo and poult. PhD Diss. North Carolina State Uni- later in life (after 10 and 20 years). Int. Arch. Allergy Immunol. versity, Raleigh. 131:209–211. Buchanan, N. P., J. M. Hott, S. E. Cutlip, A. L. Rack, A. Asamer, Lourens, A., R. Molenaar, H. van den Brand, M. Heetkamp, R. and J. S. Moritz. 2008. The effects of a natural antibiotic alterna- Meijerhof, and B. Kemp. 2006. Effect of egg size on heat productive and a natural growth promoter feed additive on broiler per- tion and the transition of energy from egg to hatchling. Poult. formance and carcass quality. J. Appl. Poult. Res. 17:202–210. Sci. 85:770–776. Button, D. K., and B. R. Robertson. 2001. Determination of DNA Lu, J., U. Idris, B. Harmon, C. Hofacre, J. J. Maurer, and M. D. Lee. content of aquatic bacteria by flow cytometry. Appl. Environ. 2003. Diversity and succession of the intestinal bacterial commu- Microbiol. 67:1636–1645. nity of the maturing broiler chicken. Appl. Environ. Microbiol. Casewell, M., C. Friis, E. Marco, P. McMullin, and I. Phillips. 2003. 69:6816–6824. The European ban on growth-promoting antibiotics and emerg- MacKinnnon, K. M., H. He, C. L. Swaggerty, J. L. McReynolds, K. ing consequences for human and animal health. J. Antimicrob. J. Genovese, S. E. Duke, J. R. Nerren, and M. H. Hogut. 2009. In Chemother. 52:159–161. ovo treatment with CpG oligoxynucleotides decreases coloniza- Christensen, H., R. A. Olsen, and L. R. Bakken. 1995. Flow cyto- tion of Salmonella Enteritidis in broiler chickens. Vet. Immunol. metric measurements of cell volumes and DNA contents during Immunopathol. 127:371–375. culture of indigenous soil bacteria. Microb. Ecol. 29:49–62. Mateos, G. G., R. Lazaro, and M. I. Gracia. 2002. The feasibility of Cochetiere, M. F. L., C. Rouge, D. Darmaun, J. C. Roze, G. Potel, using nutritional modifications to replace drugs in poultry feeds. and C. G. Leguen. 2007. Intestinal microbiota and preterm in- J. Appl. Poult. Res. 11:437–452. fants: A review. Curr. Pediatr. Rev. 3:21–34. Meijerhof, R. 2006. Chick size matters. World Poult. 22:30–31. Cox, N. A., J. S. Bailey, L. C. Blankenship, and R. P. Gilder- Meijerhof, R., and R. M. Hulet. 1997. In ovo injection of competitive sleeve. 1992. Research note: In ovo administration of a competi- exclusion culture in broiler hatching eggs. J. Appl. Poult. Res. tive exclusion culture treatment to broiler embryos. Poult. Sci. 6:260–266. 71:1781–1784. Menconi, A., A. D. Wolfenden, S. Shivaramaiah, J. C. Terraes, T. Craven, S. E. 1995. Salmonella Typhimurium, hydrophobic antibi- Urbano, J. Kuttel, C. Kremer, B. M. Hargis, and G. Tellez. 2011. otics, and the intestinal colonization of broiler chicks. J. Appl. Effect of lactic acid bacteria probiotic culture for the treatment Poult. Res. 4:333–340. of Salmonella enterica serovar Heidelberg in neonatal broiler Cukrowska, B., R. Lodínová-Žádníková, C. Enders, U. Sonnenborn, chickens and turkey poults. Poult. Sci. 90:561–565. J. Schulze, and H. Tlaskalová-Hogenová. 2002. Specific prolif- Mills, T. K., M. P. Lombardo, and P. A. Thorpe. 1999. Microbi- erative and antibody responses of premature infants to intes- al colonization of the cloacae of nestling tree swallows. Auk tinal colonization with nonpathogenic probiotic E. coli strain 116:947–956. Nissle 1917. Scand. J. Immunol. 55:204–209. http://dx.doi. Molenaar, R., I. A. Reijrink, R. Meijerhof, and H. van den Brand. org/10.1046/j.1365-3083.2002.01005.x. 2008. Relationship between hatching length and weight on lat- IN OVO INOCULATION OF CHICKS WITH PROBIOTICS 829

er productive performance in broilers. World’s Poult. Sci. J. Stern, N. J. 2008. Salmonella species and Campylobacter jejuni cecal 64:599–603. colonization model in broilers. Poult. Sci. 87:2399–2403. Nadkarni, M. A., F. E. Martin, N. A. Jacques, and N. Hunter. Sterzo, E., J. Paiva, R. Penha Filho, and A. Berchieri Junior. 2005. 2002. Determinqtion of bacterial load by real-time PCR using Time required to protect the intestinal tract of chicks against a broad-range (universal) probe and primers set. Microbiology Salmonella enterica serovar Enteritidis using competitive exclu- 148:257–266. sion. Braz. J. Poult. Sci. 7:119–122. Nisbet, D. J., S. C. Ricke, C. M. Scanlan, D. E. Corrier, A. G. Taghavi, A., B. Allan, G. Mutwiri, A. van Kessel, P. Willson, L. Holllster, and J. R. Deloach. 1994. Inoculation of broiler chicks Babiuk, A. Potter, and S. Gomis. 2008. Protection of neonatal with a continuous-flow derived bacterial culture facilitates early broiler chicks against Salmonella Typhimurium septicemia by cecal bacterial colonization and increases resistance to Salmonella DNA containing CpG motifs. Avian Dis. 52:398–406. Typhimurium. J. Food Prot. 57:12–15. Tako, E., P. R. Ferket, and Z. Uni. 2004. Effects of in ovo feeding Nurmi, E., and M. Rantala. 1973. New aspects of Salmonella infec- of carbohydrates and beta-hydroxy-beta-methylbutyrate on the

tion in broiler production. Nature 241:210–211. development of chicken intestine. Poult. Sci. 83:2023–2028. Downloaded from Pedroso, A. 2009. Which came first: The egg or its microbiota? Uni, Z., P. R. Ferket, E. Tako, and O. Kedar. 2005. In ovo feeding Pages 1–4 in Poultry Informed Professional. University of Geor- improves energy status of late-term chicken embryos. Poult. Sci. gia, Athens. 84:764–770. Penders, J., C. Thijs, C. Vink, F. F. Stelma, B. Snijders, I. Kummel- van der Hoeven-Hangoor, E., J. M. B. M. van der Vossen, F. H. ing, P. A. van den Brandt, and E. E. Stobberingh. 2006. Factors J. Schuren, M. W. A. Verstegen, J. E. de Oliveira, R. C. Mon- influencing the composition of the intestinal microbiota in early tijn, and W. H. Hendriks. 2013. Ileal microbial composition of infancy. Pediatrics 118:511–521. broilers fed various commercial diet compositions. Poult. Sci. http://ps.oxfordjournals.org/ Rao, S., R. Srinivasjois, and S. Patole. 2009. Prebiotic supplementa- 92:2713–2723. tion in full-term neonates: A systematic review of randomized Vandeplas, S., R. Dubois Dauphin, Y. Beckers, P. Thonart, and controlled trials. Arch. Pediatr. Adolesc. Med. 163:755–764. A. Thewis. 2010. Salmonella in chicken: Current and developing SAS Institute Inc. 2008. JMP User’s Guide. SAS Institute Inc., strategies to reduce contamination at farm level. J. Food Prot. Cary, NC. 73:774–785. Schneitz, C., L. Nuotio, G. Mead, and E. Nurmi. 1992. Competitive Willemsen, H., N. Everaert, L. de Smith, M. Debonne, F. Verschuere, exclusion in the young bird: Challenge models, administration O. Garain, D. Berckmans, E. Decuypere, and V. Bruggeman. and reciprocal protection. Int. J. Food Microbiol. 15:241–244. 2008. Critical assessment of chick quality measurements as an

Schultz, M., C. Göttl, R. J. Young, P. Iwen, and J. A. Vanderhoof. indicator of posthatch performance. Poult. Sci. 87:2358–2366. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 2004. Administration of oral probiotic bacteria to pregnant wom- Yalcin, S., S. Ozkan, M. Cabuk, and P. B. Siegel. 2003. Criteria for en causes temporary infantile colonization. J. Pediatr. Gastroen- evaluating husbandry practices to alleviate heat stress in broilers. terol. Nutr. 38:293–297. J. Appl. Poult. Res. 12:382–388. Senok, A. C., A. Y. Ismaeel, and G. A. Botta. 2005. Probiotics: Yamawaki, R. A., E. L. Milbradt, M. P. Coppola, J. C. Z. Rodrigues, Facts and myths. Clin. Microbiol. Infect. 11:958–966. R. L. Andreatti Filho, C. R. Padovani, and A. S. Okamoto. 2013. Smirnov, A., E. Tako, P. R. Ferket, and Z. Uni. 2006. Mucin gene Effect of immersion and enoculation in ovo of Lactobacillus spp. expression and mucin content in the chicken intestinal goblet in embryonated chicken eggs in the prevention of Salmonella En- cells are affected by in ovo feeding of carbohydrates. Poult. Sci. teritidis after hatch. Poult. Sci. 92:1560–1563. 85:669–673.

Dramatic differences in the response of macrophages from B2 and B19 MHC-defined haplotypes to interferon gamma and polyinosinic:polycytidylic acid stimulation

Maisie E. Dawes ,* Lisa M. Griggs ,* Ellen W. Collisson ,* Worthie E. Briles ,† and Yvonne Drechsler *1

* College of Veterinary Medicine, Western University of Health Sciences, 309 E. Second St., Pomona, Downloaded from CA 91766-1854; and †Department of Biological Sciences, Northern Illinois University, 415 Montgomery Hall, DeKalb 60115-2861

ABSTRACT The chicken MHC has been associated IFNγ, a cytokine at the interface of innate and adaptive with disease resistance, though the mechanisms are not immunity. Not only did B2-derived peripheral mono- http://ps.oxfordjournals.org/ understood. The functions of macrophages, critical to cytes differentiate into macrophages more readily than both innate and acquired immunity, were compared the B19 monocytes, but as determined by NO produc- between the more infectious bronchitis virus-resistant tion, macrophages from B2 and B2 on B19 genetic B2 and the more infectious bronchitis virus-susceptible background chicks were also significantly more respon- B19 lines. In vivo peripheral blood concentrations of sive to either stimulant. In conclusion, the correlation monocytes were similar in B2 or B19 homozygous hap- with resistance to illness following viral infection may lotypes. Peripheral blood-derived macrophages were be directly linked to a more vigorous innate immune stimulated with poly I:C, simulating an RNA virus, or response. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Key words: macrophage , monocyte , innate immunity, disease resistance, infectious bronchitis virus 2014 Poultry Science 93 :830–838 http://dx.doi.org/ 10.3382/ps.2013-03511

INTRODUCTION morphism, and SNP (Juul-Madsen et al., 1993; Fulton et al., 2006; Juul-Madsen et al., 2006). Selective breeding of poultry has resulted in improved Retrospective studies have correlated susceptibil- productivity in broilers at the cost of decreased im- ity and resistance to avian infectious bronchitis virus mune competence and greater susceptibility to disease (IBV)-induced clinical illness, with the B21 and B15 (Han and Smyth, 1972; Nestor et al., 1996; Bayyari et haplotype lines, respectively (Bacon et al., 2004). More al., 1997). On the other hand, some lines of birds have recently, infectivity studies identified B2, B5, and B8 been shown to be more resistant to a variety of patho- chicken lines as being more resistant to IBV-induced ill- gens (Bearse et al., 1939; Hutt and Scholes, 1941). Sev- ness than their B12 and B19 counterparts (Banat et al., eral studies have demonstrated an association between 2013). Given that the differences identified regarding the chicken MHC-B haplotype and resistance to vari- disease resistance were observed early after infection, ous pathogens including Marek’s disease virus, avian we propose that innate immunity plays a major role leukosis virus, Newcastle disease virus, Rous sarcoma in B haplotype-associated disease resistance, with the virus, and Salmonella (Heinzelmann et al., 1981; Briles macrophage being a key player. et al., 1982; Lamont, 1989; Dunnington et al., 1992; Monocytes/macrophages not only play a vital role Yoo and Sheldon, 1992; Mays et al., 2005; Kim et al., within the innate immune system, but they are also 2008; Banat et al., 2013). Described in 1950 (Briles et important in the activation of specific adaptive im- al., 1950), at least 51 B haplotypes of the chicken B mune responses acting as antigen-presenting cells. As complex have been identified using traditional serologi- such, these cells can affect the progression of clinical cal techniques (Briles and Briles, 1987; Landesman et disease. Innate cellular activity depends solely on the al., 1993; Sung et al., 1993; Miller et al., 2004) or gene expression of recognition molecules, such as Toll-like sequencing, RFLP, single-strand conformational poly- receptors (TLR) that bind the conserved pathogen- associated molecular patterns of common infectious organisms. The significance of the innate immune func- © 2014 Poultry Science Association Inc. Received July 22, 2013. tion of the macrophage in promoting disease resistance Accepted November 24, 2013. or susceptibility is not well understood. Macrophages 1 Corresponding author: [email protected] are activated or primed upon encountering a pathogen

830 MAJOR HISTOCOMPATIBILITY COMPLEX DIFFERENCES 831 or inflammatory cytokines, such as interferon (IFN) γ brooder (35°C). In addition to daily health monitoring, and subsequently undergo functional maturation, in- fresh food and water were provided ad libitum. At 3 wk creasing expression of Fc-receptors on their cell sur- of age, room temperature was adjusted to and main- face (Dietert et al., 1991; Qureshi, 2003). Macrophages tained at 25°C. Chicks were housed in standard poultry of various chicken lines congenic for MHC have been cages. To minimize the risk of pecking disorders, chicks shown to differ in their chemotactic activity (Qureshi were kept under restricted lighting conditions through- et al., 1988) and recruitment and activation (Qureshi out the study. Experimental birds were euthanized by et al., 1986). Like mammalian macrophages, avian mac- insufflation of isoflurane gas (Butler, Dublin, OH). rophages produce nitric oxide (NO) upon activation, killing bacteria or protozoa and inhibiting viral rep-

Peripheral Blood Collection and Cell Counts Downloaded from lication (Boockvar et al., 1994; Kreil and Eibl, 1996; MacMicking et al., 1997). Expression of inducible nitric Whole blood samples were collected via jugular ve- oxide synthase (iNOS) has been shown to differ among nipuncture in EDTA tubes from age-matched chicks 5 chickens of several genotypes, as a result of enhanced to 12 wk old. At no time did the amount of blood har- transcriptional activity in iNOS hyper-responder geno- vested from each bird exceed 3% of total blood volume types (Hussain and Qureshi, 1997, 1998). http://ps.oxfordjournals.org/ [8% × BW in kilograms (BWkg)]. A 2-wk window was Another indicator for the importance of macrophages allowed between subsequent blood samplings. Samples in disease susceptibility is the fact that White Leghorn were sent to Antech Diagnostics (Irvine, CA) for com- (Cornell K-Strain) chicken bone marrow produces more plete blood count and differentials. macrophage colonies than those from broiler bone marrow cells, leading to the speculation that White Leg- horn chickens may be more disease resistant. It has also Peripheral Blood Mononuclear Cell Isolation

been observed that broiler mortality losses are double at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Peripheral blood mononuclear cells (PBMC) from the losses of layer-type chickens through the first 6 wk individual birds were isolated using the differential cen- of age (Liljequist et al., 1993). Taken together, these trifugation as previously described (Seo and Collisson, results suggest a genetic basis of macrophage activa- 1997; Dawes et al., 2008; Drechsler et al., 2009, 2013). tion playing an important role in the innate immune Briefly, 3 mL of blood was mixed with an equal volume response and resistance to disease. of Alsever’s solution (Sigma-Aldrich, St. Louis, MO) Using an in vitro model, we investigated functional after which 3 mL of this mixture were layered over an differences in the activation of differentiated chicken equal volume of Ficoll-Hypaque (density 1.077 or 1.083; macrophages from chicks with distinct haplotypes. Sigma-Aldrich). Samples were centrifuged for 35 min Type II IFNγ, a potent activator of macrophages, typi- (400 × g, 23°C; brake off) for retrieval of mononuclear cally secreted by natural killer cells and other cells of cells. Isolated cells were washed twice in 5 mL of PBS the immune system upon encountering various patho- (400 × g, 10 min, 23°C), counted, and viability con- gens, was used to stimulate cultured macrophages in- firmed based on the exclusion of 0.1% trypan blue dye cluding B2/B2, B19/B19, and B2/B2 bred on the B19 (≥90%). The PBMC were resuspended in PBS to a genetic background (B19bgB2MHC). In addition, in an final concentration of 5 × 107 cells/mL. effort to emulate viral infection, the TLR3 agonist and the synthetic analog of dsRNA, poly (I:C), were used to stimulate cultured macrophages. The results show Macrophage Cell Culture differences in activation of macrophages measured by 7 NO release among several B haplotypes. One milliliter of PBMC suspension (5 × 10 cells/ mL) was incubated (37°C/5% CO2) for 3 h in each well of a 12-well plate containing RPMI w/o Phenol Red MATERIALS AND METHODS supplemented with 10% heat-inactivated fetal bovine Experimental Birds serum; nonessential amino acids (0.1 mM/mL; Invit- rogen, Carlsbad, CA), l-glutamine (2 mM; Sigma-Al- Bird protocols were performed under the approval drich), 2-mercaptoethanol (55 μM/mL; Sigma-Aldrich), of the Institutional Animal Care and Use Committee penicillin (50 U/mL; Invitrogen), and streptomycin (50 at Western University of Health Sciences, Pomona, μg/mL; Invitrogen). Following removal of nonadherent California. Fertilized eggs (from B2/B2, B15/B15, cells with warm PBS, medium was replenished and cells B21/B21, B19/B19, and B2/B2 MHC on B19 back- were incubated for 24 to 48 h to facilitate monocyte ground (F3, B19bgB2) parents, descended from Modi- adherence and complete removal of thrombocytes, non- fied Wisconsin Line 3) were obtained from W. Elwood adherent lymphocytes and other semi-adherent cells. Briles, Northern Illinois University, and incubated and Prior to the replacement of medium, adherent cell cul- hatched under standard conditions at (38°C, 50 to 65% tures were washed in warm PBS. Monocytes were cul- humidity; Banat et al., 2013) at Western University of tured for 6 d to allow maturation and differentiation of Health Sciences. Posthatch, chicks were maintained in cells, with medium changes occurring every 3 to 4 d, the incubator for 24 h, then transferred to a preheated thus ensuring that optimal nutrient requirements were 832 DAWES ET AL. met. The morphology of adherent cells was observed /ipofectaminePoly IC 6timulation daily under bright field microscopy (20× objective). Lipofectamine-mediated transfection of poly (I:C) into cells was used as a model of RNA virus infection. Confirming Cell Purity Using Five microliters each of poly (I:C) (10 mg/mL) and an Immunofluorescence Assay lipofectamine solution (Calbiochem, San Diego, CA) were incubated with 195 μL of OptiMEM medium (In- Purity of monocyte cultures, as well as 6-d-old prima- vitrogen) for 5 min at room temperature. Both solu- ry macrophages, was confirmed by immunofluorescence tions were then combined and incubated at room tem- assay (IFA). The monoclonal antibody, KUL01, has perature for 45 min before diluting the mixture 1:10 been used for flow cytometric identification of peripher- with OptimMEM to obtain a final concentration of 25 Downloaded from ally circulating chicken cells of the mononuclear phago- μg/mL of poly (I:C). Macrophages were incubated in cyte system throughout their ontogeny (monoblasts, 500 μL of the working solution for 5 to 6 h. Transfec- promonocytes, monocytes, and tissue macrophages), tion was terminated by the addition of 500 μL of RPMI whereas CV1-chNL68 and the monocyte monoclonal to each well. antibody K1 cross-react with chicken lymphocytes and http://ps.oxfordjournals.org/ thrombocytes, respectively (Mast et al., 1998). Following the removal of medium, cells were fixed in NO Assays a 1:1 (vol/vol) methanol:acetone mixture. Cells were rehydrated in PBS for 15 min before incubating with Nitrite, a stable, reactive nitrogen intermediate, blocking buffer (5% nonfat dry milk/PBS) for 20 min. served as an indirect measure of NO production rep- Staining with the mouse anti-chicken monocyte/macro- resenting macrophage/monocyte stimulation (Crippen phage antibody KUL01-FITC (Southern BioTech, Bir- et al., 2003; Singh et al., 2010). Briefly, cultured mac- mingham, AL; 1:250) was performed for 45 min to 1 h rophages were centrifuged at 200 × g for 10 min at at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 in the dark. Cells were then washed 3 times in PBS and room temperature to remove cell debris. One hundred viewed under a Nikon Eclipse Ti fluorescence micro- microliters of each culture supernatant were incubated scope (10× objective, Nikon Instruments Inc., Melville, in the dark with an equal volume (1:1) of Griess re- NY). agent (1% sulfanilamide/0.1 naphthylenediamine/2.5% phosphoric acid) in individual wells of a 96-well plate for 15 min. Absorbance was measured at a wavelength Cell Counts and Flow Cytometry of 590 nM using an ELISA plate reader (Cambrex, East Rutherford, NJ). Values were compared with concen- To optimize procedures and establish macrophage trations derived from a standard curve prepared by se- counts after adherence, macrophages were counted in rial dilution (in RPMI w/o Phenol Red) of a 2 mM wells after adherence via the NIS Elements program sodium nitrite stock solution (Sigma-Aldrich). Assays (Nikon Instruments Inc.), with cells being counted per were performed in triplicate at designated times. mm2 and calculated per well, and flow cytometry was performed to compare yields from 1.077 versus 1.083 histopaque gradients. 6tatistics Briefly, mononuclear cells were washed twice in cold Nitric oxide concentrations were expressed as aver- PBS and suspended in blocking buffer (1% NFDM, 0.1% ages with SEM. Paired t-test, 2-tailed, was used to ana- sodium azide, and 5% heat-inactivated fetal bovine lyze differences in the kinetics between B2 and B19. serum) for 20 min. After blocking, cells were washed Analysis of variance one-way test was used to determine and stained in the dark for 45 min with (1:250) FITC- statistical differences between NO groups; significance labeled mouse anti-chicken monocyte/macrophage an- was set at P < 0.05. Differential blood cell counts of tibody KUL01-FITC (Southern BioTech, Birmingham, B2 and B19 haplotypes (monocytes, lymphocytes, and AL), again washed in cold PBS 3 times, resuspended in heterophils) were analyzed with t-test, paired, 2-tailed. 1 mL of cold PBS, and analyzed on Beckman Coulter cytomics FC 500 flow cytometer (10,000 events were obtained for each sample). RESULTS

IF1Ȗ 6timulation Concentrations of B19 and B2 Monocytes in :hole Blood :ere 6imilar A 50 ρg/mL of ch-IFNγ solution (Invitrogen) was prepared in RPMI w/o Phenol Red culture medium Studies were designed to identify potential differenc- (Invitrogen). After washing the cells twice with warm es in the function of monocytes/macrophages purified PBS, macrophage cultures were stimulated with 1 mL from birds of defined haplotypes. Based on noted dif- of RPMI-ch-IFNγ mixture. ferences in susceptibility to avian coronavirus infection MAJOR HISTOCOMPATIBILITY COMPLEX DIFFERENCES 833

Table 1. Peripheral blood cell counts1

B haplotype Monocytes Lymphocytes Heterophils Eosinophils Basophils B2/B2 455 (±72) 6,826 (±496) 2,422 (±286) 0 (±0) 64 (±21) B19/B19 405 (±75) 5,736 (±541) 2,191 (±397) 43 (±31) 60 (±31) 1Peripheral blood cell counts (mean cells/μL ± SD) performed by Antech Diagnostics (Irvine, CA) for B2/B2 (n = 6) and B19/B19 (n = 6) chickens. Values showed no statistically significant differences between haplotypes by a paired t-test.

(Banat et al., 2013), potential differences in the concen- Greater NO Production of B2-Derived trations of circulating cell populations and specifically Macrophages in Response to 6timulation Downloaded from that of monocytes in birds with distinct haplotypes had to be established. Cell counts from peripheral blood of Potential differences in macrophage functions were monocytes, lymphocytes, and heterophils were not sig- examined following exposure to either IFNγ or poly nificantly different (t-test, paired, 2-tailed) between the (I:C). All stimulations were performed on d 6 of cul-

more resistant B2 and more susceptible B19 homozy- ture when monocytes/macrophages from both B2 and http://ps.oxfordjournals.org/ gous chickens (Table 1). Yields of PBMC prepared from B19 birds were mature (Figure 2). Nitric oxide produc- either Ficoll-Hypaque 1.077, previously used for lym- tion served as the indicator of macrophage activation. phocyte studies, or 1.083, which has been recommended Timed exposure (24, 48, and 72 h) of macrophages cul- for mononuclear cell enrichment in mammals (Feldman tured 6 d to either IFNγ or poly (I:C) demonstrated and Mogelesky, 1987), were compared. Whereas the the existence of functional differences between B2/B2 concentrations from birds of either haplotype were sim- and B19/B19 lines of birds aged 5 to 6 wk. Beginning ilar, the higher density gradient (1.083) yielded higher at 24 h postexposure to IFNγ (Figure 4A) and poly amounts of avian monocytes from PBMC derived from (I:C) (Figure 4B), induced NO production by macro- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 either haplotype (Table 2). All PBMC isolations were, phages derived from B19/B19 birds was significantly therefore, performed using the higher density solution. (P < 0.01, t = test, paired) lower compared with that by B2/B2-derived macrophages. Additionally, the responsiveness of B2/B2 cells to either stimulant was sig- Ex Vivo Monocytes from B2 Birds nificantly greater compared with that of unstimulated Differentiate More Readily than Those control B2/B2 cells (P < 0.01). At 48 h poststimulation of B19 Birds with poly (I:C), there was a dramatic difference in the NO concentration of B2/B2 culture (92.76 μg/mL) su- By d 4 of culture, B2 cells demonstrated distinct pernatants than that of B19/B19 cells (17.90 μg/mL; macrophage-like morphologies: stellate cytoplasmic Figure 5). Because background NO concentration, as processes, vacuolation, and decreased nuclear to cyto- well as cell death, increased at 72 h, especially in the plasmic ratio (N/C; Figure 1, panel A). In contrast, B19 cultures, stimulation of cells in subsequent experi- differences in N/C and number of cytoplasmic process- ments with poly (I:C) (Figure 5A) or IFNγ (Figure 5B) es were not evident in B19-derived cultures until 6 d was only characterized up to the 48 h time point using postinitiation of culture (Figure 1, panel B). Similarly, age-matched birds ages 6 to 12 wk. In addition, macro- during the kinetic studies, the number of cytoplasmic phages from B8/B8, B15/B15, and B21/21 haplotypes, processes displayed by B2 macrophages and the degree as well as macrophages from B2/B2 on B19 background of vacuolation observed at 48 h poststimulation with haplotypes (B19bgB2MHC), were stimulated with IFNγ IFNγ was more marked than the changes noted in B19- (Table 3). Nitric oxide release from B19bgB2MHC mac- derived cells (Figure 2). To verify purity of macrophage rophages was very similar (86.2 μg/mL) to that from cultures, B2/B2 and B19/B19 cells were positively B2/B2 haplotypes, and significantly different from NO stained with monoclonal antibody KUL01 after 6 d of release of B19/B19 macrophages (P < 0.01). However, culture (Figure 3, panels A and B). stimulation of B8/B8 macrophages was very variable

Table 2. Cell concentrations

PBMC1 PBMC MØ/well2 B haplotype 1.077 gradient 1.083 gradient after adherence B2/B2 6.4 × 107 7.3 × 107 2.70 (±0.33) × 105 B19/B19 7.8 × 107 7.8 × 107 2.62 (±0.35) × 105 1Total cell numbers after peripheral blood mononuclear cells (PBMC) isolation via gradient 1.077 and 1.083 are given for 3 mL of whole blood drawn counted by fluorescence-activated cell sorting. 2Macrophage numbers given per each in well after both 6 d of adherence (n = 5/haplotype) and stimulation with IFNγ counted via the NIS Elements program, with cells being counted per mm2 and calculated per well. Macrophage counts are reported as mean (±SD). 834 DAWES ET AL. Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 1. Ex vivo differentiation of monocytes prestimulation. Images demonstrate the rate of differentiation of chicken monocytes from homozygous B2 (panel A) and B19 (panel B) chicks at d 2, 4, and 6 d postadherence. Distinct vacuoles, increased numbers of stellate cytoplasmic processes, and decreased nuclear to cytoplasmic ratio are visible on d 4 with B2 cells and on d 6 with B19 cells. Pictures were taken at 20× magnification, inverted, with Nikon Eclipse Ti fluorescence microscope (Nikon Instruments Inc., Melville, NY). with repeated experiments (17.7 μg/mL). Although chickens, including an in vivo IBV infectivity study of NO release was less than that of B2/B2 cells, B21/B21 B haplotype-defined chicks (Banat et al., 2013), have (39.8 μg/mL) and B15/B15 (45.9 μg/mL) macrophages shown an association between disease susceptibility and yielded higher levels of NO than B19/B19 cells. the MHC (Briles et al., 1983; Taylor, 2004; Goto et al., 2009). Because differences in IBV resistance of B2/B2 DISCUSSION chicks to clinical illness were observed early after initial infection, innate immunity was targeted in the current Inbred B haplotype chicken lines provide an excel- study. To our knowledge, this is the first study in which lent resource for studying the genetic bases of disease differences in activation of monocytes/macrophages of resistance and susceptibility. Several clinical studies in B haplotype-defined birds have been investigated in the context of disease resistance and susceptibility. The rate of differentiation of macrophages derived from the more resistant B2/B2 chickens and their subsequent activation in response to stimuli was dramatically greater than that of macrophages from the more IBV-susceptible B19 haplotype chickens. The consistent, significantly greater response of the B2 macrophages could potentially explain the differences observed early after IBV infection. A role of TLR3, which binds to dsRNA, is implicated in the macrophage response and ultimately in in vivo resistance (Malmgaard et al., 2004; Whitmore et al., 2004; Kogut et al., 2005; Nang et al., 2011). Double-stranded RNA is a replica- Figure 2. Differentiation of 48 h interferon (IFN)γ-stimulated tion intermediate during infection with RNA viruses, macrophages. Macrophages cultured for 6 d from B2 and B19 haplotypes were stimulated with IFNγ for 48 h, and pictures were taken such as IBV, avian influenza virus, and Newcastle dis- with a Nikon Eclipse Ti fluorescence microscope (Nikon Instruments ease virus (Malmgaard et al., 2004; Nang et al., 2011). Inc., Melville, NY), inverted at 10× magnification. Macrophage-like Potential functional differences related to macrophage features (stellate cytoplasmic processes, vacuolation, and decreased communication with adaptive immunity were evaluated nuclear to cytoplasmic ratio) are more pronounced in B2 cells (A) compared with their B19 counterparts (B). Color version available in through stimulation by the lymphocyte-derived cyto- the online PDF. kine, IFNγ. The demonstrated significant differences in MAJOR HISTOCOMPATIBILITY COMPLEX DIFFERENCES 835 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 3. Immunofluorescent labeling of adherent macrophages. Panel A depicts the brightfield image of macrophages. The green fluorescence noted in panel B depicts positive labeling of B2 and B19 chicken cells by monoclonal antibody KUL01-FITC, confirming their monocyte/macrophage cell lineage. Pictures were taken with Nikon Eclipse Ti fluorescence microscope (Nikon Instruments Inc., Melville, NY), inverted at 10× magnification. Color version available in the online PDF.

NO production that both poly (I:C) and IFNγ induced iNOS and TLR expression on macrophages from B2 in the B2/B2 and the B19/B19 haplotypes may explain, versus B19 haplotypes to investigate if similar path- at least in part, the greater resistance of birds of the ho- ways are responsible for the difference in activation. mozygous B2 line to IBV-associated early onset clinical In addition to differences in activation, even when illness (Banat et al., 2013). As Dil and Qureshi (2002) cultured ex vivo in the absence of poly (I:C) and IFNγ, showed, expression of iNOS varies among macrophages B2/B2 monocytes differentiated more readily than from chickens of different genetic lines and is associated B19/B19 monocytes. Based on changes in morphology, with differential TLR4 expression in response to lipo- B2-derived cells obtain classical features of the macro- polysaccharide. Further research is necessary to analyze phage, such as increased cytoplasm to nuclear ratio,

Figure 4. Kinetics. Nitric oxide release is shown as mean (μg/mL) ± SD by B2 and B19 macrophages at 24, 48, and 72 h poststimulation with poly (I:C); n = 8 (A) and interferon (IFN) γ; n = 6 (B). In the presence of either stimulant, NO production at all 3 time points was significantly lower in B19 than in B2 supernatants (P < 0.01, t-test, paired). Samples were taken at 6 wk of age. Color version available in the online PDF. 836 DAWES ET AL.

Table 3. Interferon (IFN) γ stimulation of different B haplotype macrophages1

B haplotype Control IFNγ B2 (n = 21) 8.7 (±2.69) 92.76 (±9.67)a,b B19 (n = 21) 11.88 (±1.94) 17.90 (±1.08) B19bgB2MHC (n = 7) 15.5 (±0.6) 86.2 (±19.0)a,b B8 (n = 12) 4.8 (±1.2) 17.7 (±16.0) B15 (n = 2) 13.6 (±2.0) 45.9 (±9.4) B21 (n = 13) 10.0 (±1.8) 39.8 (±4.8)a aSignificantly different compared with unstimulated sample of same haplotype by ANOVA one-way test (P < 0.01). Downloaded from bSignificantly different compared with B19 stimulated by ANOVA one-way test (P < 0.01). 1Nitric oxide release given as mean (μg/mL) ± SD from macrophages cultured 6 d from chicks with the B2, B19, B19bgB2MHC, B8, B15, B21 haplotypes (5–12 wk of age) at 48 h poststimulation with IFNγ. The B2 and B19bgB2MHC exhibit the largest nitric oxide release compared with other haplotypes. http://ps.oxfordjournals.org/ vacuolation, and cytoplasmic pseudopod-like projections, within 4 d as compared with the 6 d required by B19-derived cells. Although the significance of the higher rate of activation is not known, this could reflect an overall lower threshold of response of the B2 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 cells compared with that of the B19 homozygous macrophages. Participation of pure cultures of monocytes/macrophages in these studies was confirmed by the 100% labeling by the KUL01 macrophage monoclonal antibody marker (Mast et al., 1998). This negates any influence by contaminating lymphocytes and their regulatory Figure 5. A: Nitric oxide release by poly(I:C)-stimulated B2 and B19 macrophages. Cells were stimulated for 48 h with 25 μg/mL poly cytokines. Furthermore, the cultures were plated with (I:C). Although not significantly different, basal NO concentration equivalent concentrations of monocytes/macrophages. was consistently greater in B19 culture supernatants (10.88 μg/mL) Hence, these studies reflect differences in cell activity than in B2 supernatants (9.75 μg/mL). Following stimulation, NO and not increased B2 cell numbers. Additionally, the production by B2 cells (98.12 μg/mL) was significantly greater than that by B19 macrophages (17.19 μg/mL; P < 0.01, one-way ANOVA). concentration of peripherally circulating monocytes in Values are reported as mean with SE, n = 28. B: Nitric oxide release B19 and B2 birds was not significantly different. by interferon (IFN)γ-stimulated B2 and B19 macrophages. Cells were The MHC-B locus, which has been associated with stimulated for 48 h with 50 pg/mL of ch-IFNγ. Although not significantly different, basal NO concentration was consistently greater resistance to tumor-associated viruses in chickens, has in B19 culture supernatants (11.88 μg/mL) than in B2 supernatants also been implicated in observed differences in mac- (8.7 μg/mL). Following stimulation, NO production was significantly rophage function (Briles et al., 1983; Taylor, 2004). greater in B2 culture supernatants (92.76 μg/mL) than in B19 culture supernatants (17.9 μg/mL; P < 0.01, one-way ANOVA). Values are Qureshi et al. (1986, 1988) demonstrated that macro- reported as mean with SE, n = 21. phages from chickens congenic for MHC also showed differences in chemotactic activity, recruitment, and activation. The congenic B2 on the B19 background michi et al., 2008). Likewise, genomic regions lying out- demonstrated the B2 phenotype suggesting at least for side the B region could also be implicated in monocyte B2 and B19, the genetic difference may lie within the B differentiation and activation. locus. Monocytes derived from additional homozygous Future studies identifying differences in protein and MHC haplotypes were also differentially activated by mRNA expression following macrophage stimulation IFNγ. The cells purified from B15/B15 birds, which will provide important clues as to the genes and path- were shown by Bacon et al. (2004) to be more resistant ways responsible for greater resistance to illness. Our to IBV infection also demonstrated greater macrophage preliminary results from sequencing the RNA of stim- stimulation. It is possible that genes lying outside this ulated macrophages of B2 and B19 haplotype chick- region could also affect macrophage activation as dem- ens indicate differential regulation of at least the TLR onstrated in other lines. Genes lying outside the B2 ap- pathway (manuscript in preparation), which is a criti- parently can influence resistance to highly pathogenic cal pathway in innate immunity and vital to the host avian influenza (Hunt et al., 2010). immune response to pathogens (Medzhitov and Jane- Sequence differences between the MHC-B region of way, 1997; Medzhitov, 2001). B2 and B19 haplotypes have indicated several differ- In conclusion, the potential for using macrophage ences in a 14-gene region between BG and BF (Hoso- function to screen for resistance could be a valuable MAJOR HISTOCOMPATIBILITY COMPLEX DIFFERENCES 837 asset in breeding programs. Furthermore, the chicken Briles, W. E., R. W. Briles, R. E. Taffs, and H. A. Stone. 1983. is an exceptional resource for deciphering MHC-related Resistance to a malignant lymphoma in chickens is mapped to subregion of major histocompatibility (B) complex. Science mechanisms behind disease resistance. Although chick- 219:977–979. ens with various defined B haplotypes have previously Briles, W. E., G. W. Mc, and M. R. Irwin. 1950. On multiple alleles been shown to differ in their responses to pathogens, effecting cellular antigens in the chicken. Genetics 35:633–652. the cellular differences described in these studies pro- Crippen, T. L., C. L. Sheffield, H. He, V. K. Lowry, and M. H. Kogut. 2003. Differential nitric oxide production by chicken im- vide a basis for future work characterizing the molecu- mune cells. Dev. Comp. Immunol. 27:603–610. lar and genetic basis of immune-mediated resistance. Dawes, M. E., J. W. Tyler, A. E. Marsh, R. L. Larson, B. J. Steevens, These studies provide a background for eventually es- and J. Lakritz. 2008. In vitro effects of lactoferrin on lipopoly- tablishing lines with resistance to infection or with en- saccharide-induced proliferation, gene expression, and prostanoid

production by bovine peripheral blood mononuclear cells. Am. Downloaded from hanced immunity to vaccines. J. Vet. Res. 69:1164–1170. Dietert, R. R., K. A. Golembroski, S. E. Bloom, and M. A. Qureshi. 1991. The avian macrophage in cellular immunity. Pages 71–95 ACKNOWLEDGMENTS in Avian Cellular Immunology. J. M. Sharma, ed. CRC Press, Boca Raton, FL. This work was funded by grants from the USDA- Dil, N., and M. A. Qureshi. 2002. Differential expression of induc- National Research Institute #2006-35204-18810, US- ible nitric oxide synthase is associated with differential Toll-like http://ps.oxfordjournals.org/ receptor-4 expression in chicken macrophages from different ge- DA-National Institute of Food and Agriculture #2008- netic backgrounds. Vet. Immunol. Immunopathol. 84:191–207. 00875, USDA-Cooperative State Research Service Drechsler, Y., R. L. Bohls, R. Smith, N. Silvy, H. Lillehoj, and E. W. Competitive Grants Program (2006-35204-16560), and Collisson. 2009. An avian, oncogenic retrovirus replicates in vivo the Avian Influenza Cooperative Agriculture Program. in more than 50% of CD4+ and CD8+ T lymphocytes from an endangered grouse. Virology 386:380–386. The authors thank Suzana Tkalcic (veterinary patholo- Drechsler, Y., S. Tkalcic, M. D. Saggese, H. L. Shivaprasad, D. K. gist, College of Veterinary Medicine, Western Universi- Ajithdoss, and E. W. Collisson. 2013. A DNA vaccine expressing ty of Health Sciences, Pomona, CA), Miguel D. Saggese ENV and GAG offers partial protection against reticuloendothe- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 (veterinary microbiologist, College of Veterinary Med- liosis virus in the prairie chicken (Tympanicus cupido). J. Zoo Wildl. Med. 44:251–261. icine, Western University of Health Sciences), Renee Dunnington, E. A., C. T. Larsen, W. B. Gross, and P. B. Siegel. Kopulos (Department of Biological Sciences, North- 1992. Antibody responses to combinations of antigens in White ern Illinois University, DeKalb, IL), and Linda Yates Leghorn chickens of different background genomes and major his- (Department of Biological Sciences, Northern Illinois tocompatibility complex genotypes. Poult. Sci. 71:1801–1806. Feldman, D. L., and T. C. Mogelesky. 1987. Use of Histopaque for University) for the invaluable expertise and technical isolating mononuclear cells from rabbit blood. J. Immunol. support they provided. We are very grateful to Ghida Methods 102:243–249. Banat (College of Veterinary Medicine, Western Uni- Fulton, J. E., H. R. Juul-Madsen, C. M. Ashwell, A. M. McCarron, J. A. Arthur, N. P. O’Sullivan, and R. L. Taylor Jr.. 2006. Mo- versity of Health Sciences) and Chris Petro (California lecular genotype identification of the Gallus gallus major histo- Polytechnic Institute, Pomona, CA) for their care of compatibility complex. Immunogenetics 58:407–421. the chickens. Goto, R. M., Y. Wang, R. L. Taylor Jr., P. S. Wakenell, K. Ho- somichi, T. Shiina, C. S. Blackmore, W. E. Briles, and M. M. Miller. 2009. BG1 has a major role in MHC-linked resistance to REFERENCES malignant lymphoma in the chicken. Proc. Natl. Acad. Sci. USA 106:16740–16745. Bacon, L. D., D. B. Hunter, H. M. Zhang, K. Brand, and R. Etches. Han, P. F., and J. R. Smyth Jr.. 1972. The influence of growth rate 2004. Retrospective evidence that the MHC (B haplotype) of on the development of Marek’s disease in chickens. Poult. Sci. chickens influences genetic resistance to attenuated infectious 51:975–985. bronchitis vaccine strains in chickens. Avian Pathol. 33:605–609. Heinzelmann, E. W., K. K. Clark, W. M. Collins, and W. E. Briles. Banat, G. R., S. Tkalcic, J. A. Dzielawa, M. W. Jackwood, M. D. 1981. Host age and major histocompatibility genotype influence Saggese, L. Yates, R. Kopulos, W. E. Briles, and E. W. Collisson. on Rous sarcoma regression in chickens. Poult. Sci. 60:2171– 2013. Association of the chicken MHC B haplotypes with resis- 2175. tance to avian coronavirus. Dev. Comp. Immunol. 39:430–437. Hosomichi, K., M. M. Miller, R. M. Goto, Y. Wang, S. Suzuki, J. Bayyari, G. R., W. E. Huff, N. C. Rath, J. M. Balog, L. A. Newber- K. Kulski, M. Nishibori, H. Inoko, K. Hanzawa, and T. Shiina. ry, J. D. Villines, J. K. Skeeles, N. B. Anthony, and K. E. Nestor. 2008. Contribution of mutation, recombination, and gene con- 1997. Effect of the genetic selection of turkeys for increased body version to chicken MHC-B haplotype diversity. J. Immunol. weight and egg production on immune and physiological respons- 181:3393–3399. es. Poult. Sci. 76:289–296. Hunt, H. D., S. Jadhao, and D. E. Swayne. 2010. Major histocom- Bearse, G. E., C. F. McClary, and M. W. Miller. 1939. The results patibility complex and background genes in chickens influence of eight years’ selection for disease resistance and susceptibility susceptibility to high pathogenicity avian influenza virus. Avian in White Leghorns. Poult. Sci. 18:400–401. Dis. 54:572–575. Boockvar, K. S., D. L. Granger, R. M. Poston, M. Maybodi, M. K. Hussain, I., and M. A. Qureshi. 1997. Nitric oxide synthase activ- Washington, J. B. Hibbs Jr., and R. L. Kurlander. 1994. Nitric ity and mRNA expression in chicken macrophages. Poult. Sci. oxide produced during murine listeriosis is protective. Infect. 76:1524–1530. Immun. 62:1089–1100. Hussain, I., and M. A. Qureshi. 1998. The expression and regulation Briles, W. E., and R. W. Briles. 1987. Genetics and classification of of inducible nitric oxide synthase gene differ in macrophages from major histocompatibility complex antigens of the chicken. Poult. chickens of different genetic background. Vet. Immunol. Immu- Sci. 66:776–781. nopathol. 61:317–329. Briles, W. E., R. W. Briles, D. L. Pollock, and M. Pattison. 1982. Hutt, F. B., and J. C. Scholes. 1941. Genetics of the fowl XIII breed Marek’s disease resistance of B (MHC) heterozygotes in a cross differences in susceptibility to Salmonella pullorum. Poult. Sci. of purebred Leghorn lines. Poult. Sci. 61:205–211. 20:342–352. 838 DAWES ET AL.

Juul-Madsen, H. R., T. S. Dalgaard, C. M. Rontved, K. H. Jensen, Medzhitov, R., and C. A. Janeway Jr. 1997. Innate immunity: Im- and N. Bumstead. 2006. Immune response to a killed infectious pact on the adaptive immune response. Curr. Opin. Immunol. bursal disease virus vaccine in inbred chicken lines with differ- 9:4–9. ent major histocompatibility complex haplotypes. Poult. Sci. Miller, M. M., L. D. Bacon, K. Hala, H. D. Hunt, S. J. Ewald, J. 85:986–998. Kaufman, R. Zoorob, and W. E. Briles. 2004. 2004 Nomenclature Juul-Madsen, H. R., J. E. Hedemand, J. Salomonsen, and M. Simon- for the chicken major histocompatibility (B and Y) complex. Im- sen. 1993. Restriction fragment length polymorphism analysis of munogenetics 56:261–279. the chicken B-F and B-L genes and their association with sero- Nang, N. T., J. S. Lee, B. M. Song, Y. M. Kang, H. S. Kim, and S. logically defined B haplotypes. Anim. Genet. 24:243–247. H. Seo. 2011. Induction of inflammatory cytokines and toll-like Kim, D. K., H. S. Lillehoj, Y. H. Hong, D. W. Park, S. J. Lamont, receptors in chickens infected with avian H9N2 influenza virus. J. Y. Han, and E. P. Lillehoj. 2008. Immune-related gene ex- Vet. Res. 42:64. pression in two B-complex disparate genetically inbred Fayoumi Nestor, K. E., Y. M. Saif, J. Zhu, and D. O. Noble. 1996. Influence

chicken lines following Eimeria maxima infection. Poult. Sci. of growth selection in turkeys on resistance to Pasteurella multo- Downloaded from 87:433–443. cida. Poult. Sci. 75:1161–1163. Kogut, M. H., M. Iqbal, H. He, V. Philbin, P. Kaiser, and A. Smith. Qureshi, M. A. 2003. Avian macrophage and immune response: An 2005. Expression and function of Toll-like receptors in chicken overview. Poult. Sci. 82:691–698. heterophils. Dev. Comp. Immunol. 29:791–807. Qureshi, M. A., R. R. Dietert, and L. D. Bacon. 1986. Genetic varia- Kreil, T. R., and M. M. Eibl. 1996. Nitric oxide and viral infection: tion in the recruitment and activation of chicken peritoneal mac- NO antiviral activity against a flavivirus in vitro, and evidence rophages. Proceedings of the Society for Experimental Biology for contribution to pathogenesis in experimental infection in vivo. and Medicine Soc. Exp. Biol. Med. 181:560–568. http://ps.oxfordjournals.org/ Virology 219:304–306. Qureshi, M. A., R. R. Dietert, and L. D. Bacon. 1988. Chemotactic Lamont, S. J. 1989. The chicken major histocompatibility com- activity of chicken blood mononuclear leukocytes from 15I5-B- plex in disease resistance and poultry breeding. J. Dairy Sci. congenic lines to bacterially-derived chemoattractants. Vet. Im- 72:1328–1333. munol. Immunopathol. 19:351–360. Landesman, E., Z. Uni, and E. D. Heller. 1993. Designation by re- Seo, S. H., and E. W. Collisson. 1997. Specific cytotoxic T lympho- striction fragment length polymorphism of major histocompat- cytes are involved in in vivo clearance of infectious bronchitis ibility complex class IV haplotypes in meat-type chickens. Anim. virus. J. Virol. 71:5173–5177. Genet. 24:349–354. Singh, S., H. Toro, D. C. Tang, W. E. Briles, L. M. Yates, R. T.

Liljequist, B., D. Weinstock, M. Qureshi, and J. Brake. 1993. Natu- Kopulos, and E. W. Collisson. 2010. Non-replicating adenovirus at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 ral-killer-cell activity in male Single Comb White Leghorns and vectors expressing avian influenza virus hemagglutinin and nu- commercial broilers. Colloq. INRA 62:57–62. cleocapsid proteins induce chicken specific effector, memory and MacMicking, J., Q. W. Xie, and C. Nathan. 1997. Nitric oxide and effector memory CD8(+) T lymphocytes. Virology 405:62–69. macrophage function. Annu. Rev. Immunol. 15:323–350. Sung, A. M., A. W. Nordskog, S. J. Lamont, and C. M. Warner. Malmgaard, L., J. Melchjorsen, A. G. Bowie, S. C. Mogensen, and S. 1993. Isolation and characterization of cDNA clones for chick- R. Paludan. 2004. Viral activation of macrophages through TLR- en major histocompatibility complex class II molecules. Anim. dependent and -independent pathways. J. Immunol. 173:6890– Genet. 24:227–233. 6898. Taylor, R. L. Jr. 2004. Major histocompatibility (B) complex control Mast, J., B. M. Goddeeris, K. Peeters, F. Vandesande, and L. R. of responses against Rous sarcomas. Poult. Sci. 83:638–649. Berghman. 1998. Characterisation of chicken monocytes, mac- Whitmore, M. M., M. J. DeVeer, A. Edling, R. K. Oates, B. Simons, rophages and interdigitating cells by the monoclonal antibody D. Lindner, and B. R. Williams. 2004. Synergistic activation of KUL01. Vet. Immunol. Immunopathol. 61:343–357. innate immunity by double-stranded RNA and CpG DNA pro- Mays, J. K., L. D. Bacon, A. R. Pandiri, and A. M. Fadly. 2005. motes enhanced antitumor activity. Cancer Res. 64:5850–5860. Response of White Leghorn chickens of various B haplotypes to Yoo, B. H., and B. L. Sheldon. 1992. Association of the major his- infection at hatch with subgroup J avian leukosis virus. Avian tocompatibility complex with avian leukosis virus infection in Dis. 49:214–219. chickens. Br. Poult. Sci. 33:613–620. Medzhitov, R. 2001. Toll-like receptors and innate immunity. Nat. Rev. Immunol. 1:135–145.

Effects of dietary clays on performance and intestinal mucus barrier of broiler chicks challenged with Salmonella enterica serovar Typhimurium and on goblet cell function in vitro

J. A. S. Almeida, N. P. Ponnuraj, J. J. Lee , P. Utterback , H. R. Gaskins , R. N. Dilger , and J. E. Pettigrew 1 Downloaded from Department of Animal Sciences, University of Illinois, Urbana 61801

ABSTRACT In vivo and in vitro experiments were adenocarcinoma cell line (LS174T) was cultured in vi- conducted to test for beneficial effects of dietary clays tro in 3 separate experiments in the absence or presence

on broiler chicks challenged with Salmonella enterica of 3 concentrations (0.05, 0.10, and 0.50%) of SMA. http://ps.oxfordjournals.org/ serovar Typhimurium and to explore potential mecha- Expression of mucin 2 (MUC2), resistin-like molecule β nisms. First, two hundred forty 1-d-old male broilers (RELMß), and trefoil factor 3 (TFF3) were determined (initial BW: 41.6 ± 0.4 g) were allotted in a 2 × 4 by real-time reverse-transcription PCR. The expression factorial arrangement in a randomized complete block of RELMβ was increased and expression of MUC2 was design. There were 2 infection treatments (with or reduced (P < 0.05) by 0.10% SMA. Also, LS174T cells without Salmonella) and 4 diets: basal (BAS), 0.3% were cultured without or with SMA (0.05 and 0.10%) smectite A (SMA), 0.3% smectite B, and 0.3% zeolite. and the medium and cell lysate were analyzed for The Salmonella reduced (P < 0.05) the growth rate of RELMβ using an immunoblot assay. Protein expression at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 chicks fed the BAS, and feeding clay largely restored of RELMß in the cell lysate was reduced (P < 0.05) it (challenge × diet interaction, P < 0.05). Goblet cell by SMA addition but increased in the medium, indi- number and size were increased (P < 0.05) by Salmo- cating that SMA increased secretion of RELMß, thus nella in chicks fed the BAS and were reduced (P < depleting the cell and concentrating this protein in the 0.05) in Salmonella-challenged chicks by feeding SMA. medium. In conclusion, the dietary clays restored the Villus height was reduced by the Salmonella challenge growth depression caused by Salmonella, and changes in the chicks fed dietary clays (P < 0.01) but not in in goblet cell function may contribute to the benefits of chicks fed the BAS (interaction P < 0.05). A human one of the clays, specifically SMA. Key words: goblet cell , chick, clay , Salmonella enterica serovar Typhimurium, LS174T cell line 2014 Poultry Science 93 :839–847 http://dx.doi.org/ 10.3382/ps.2013-03587

INTRODUCTION in humans (Carretero, 2002; Szajewska et al., 2006), but their efficacy in chicks challenged with enteric dis- Economic loss due to enteric diseases is an impor- ease has not been reported. Their current use in diets tant problem in the food animal industry, in spite for poultry and pigs is largely for adsorption of myco- of powerful health technologies such as all-in/all-out toxins (Ledoux and Rottinghaus, 2000). movements, sanitation, biosecurity, vaccines, and oth- Several mechanisms have been proposed for the ben- ers (Hardy, 2002). Growing evidence shows that several eficial effects of clays on enteric health, including toxin dietary factors can help maintain health and growth binding (Phillips et al., 1988; Schell et al., 1993; Harp- performance of disease-challenged animals (Perez et al., er et al., 2010), enhancement of the immune response 2011; Che et al., 2012). Also, we found that feeding a (Gonzales et al., 2004), strengthened intestinal mucus low dietary concentration of specific clays (smectite, barrier (Reichardt et al., 2009; Trckova et al., 2009), zeolite, and kaolinite) to pigs challenged with a patho- mineral binding (Katsumata et al., 2003), and others genic Escherichia coli reduced diarrhea (Song et al., (Droy-Lefaix et al., 1985; Fioramonti et al., 1987; Ward 2012). Clays have been used to reduce enteric disorders et al., 1991). All of these mechanisms are supported by controlled observations, but their relative importance under specific conditions is unknown. Potential impacts © 2014 Poultry Science Association Inc. on barrier function and immune response appear espe- Received August 30, 2013. Accepted December 16, 2013. cially strong and relevant to health (Gonzales et al., 1 Corresponding author: [email protected] 2004; Trckova et al., 2009).

839 840 ALMEIDA ET AL. Goblet cells are specialized epithelial cells that secrete (Salmonella challenge versus sham challenge) and 1 of cysteine-rich products (Moncada and Chadee, 2002) 4 dietary treatments: basal diet, smectite A (SMA), such as mucin 2 (MUC2), trefoil factor 3 (TTF3), and smectite B (SMB), and zeolite (ZEO) at 0.3% inclu- resistin-like molecule β (RELMβ). These products are sion in the basal diet. All the experimental diets were involved in maintaining integrity of the gastrointestinal made from the basal diet with the addition of 0.3% of mucosal surface (Specian and Oliver, 1991). The genes each dietary clay. of the MUC family encode the peptide backbones of A primary poultry isolate of Salmonella enterica se- mucins. There are secreted (MUC2, MUC5A, MUC5B, rovar Typhimurium (ST-10, Southern Plains Agricul- MUC6, and MUC7) and membrane-bound (MUC3, tural Research Center, USDA, College Station, TX) MUC4) mucins (Corfield et al., 2000). The MUC2 is was used as previously reported by Faber et al. (2012). the major secreted mucin in the small and large intes- On d 0 (the day of inoculation, 10 d of age) inocu- Downloaded from tine (Kim and Ho, 2010). lum was diluted to 4 × 108 cfu/mL using sterile PBS. The TFF are secreted in the gastrointestinal tract Chicks were administered a single 0.5-mL oral dose of (Thim et al., 2002) and usually associated with the either 2 × 108 cfu of Salmonella Typhimurium/dose mucin layer where they have a healing function (Thim, or PBS as sham control using 1-mL syringes without 1997). The TFF3 is expressed mainly in the intestine needles as previously described by Faber et al. (2012). http://ps.oxfordjournals.org/ (Thim, 1997) and is almost exclusively expressed by goblet cells. The TFF co-localizes with MUC2 (Taupin Ingredients, Diets, and Feeding and Podolsky, 2003). The expression of RELMβ is induced upon bacterial colonization; thus, it is associated Clays were obtained from Milwhite Inc. (Brownsville, with mucosal injury (McVay et al., 2006) in certain TX). All diets were formulated to meet or exceed NRC pathological conditions and also triggers protective and (1994) recommendations and did not include antibiot-

pro-inflammatory mechanisms in animals exposed to ics or coccidiostatics (Table 1). Feed and fresh water at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 chemically induced colitis (Kim and Ho, 2010). were offered to the chicks ad libitum. Our objectives were first to test for beneficial effects of dietary clays on young broiler chicks challenged with Data Recording and Sample Collection pathogenic Salmonella and to explore potential mechanisms through which clays may produce benefits, with Pens of chicks and feeders were weighed on d 0 (in- primary emphasis on barrier function. oculation day), 3, 7, and 10 postinfection (PI) to determine ADG, ADFI, and G:F. These weighings delineate MATERIALS AND METHODS the approximate periods expected to be associated with pathogen incubation, active disease, and recovery. Two In Vivo Experiment pens were eliminated after d 3 PI because the chicks were unable to access feed for a day due to defective The Institutional Animal Care and Use Committee attachment of the feeders to these pens. Due to this un- and Institutional Biosafety Committee of the Univer- sity of Illinois reviewed and approved the animal care procedures for this experiment. Table 1. Ingredient composition of basal diet (as-fed basis)

Item % Birds, Housing, and Experimental Design Ingredient Corn, ground 52.85 Two hundred forty 1-d-old male broiler chicks (initial Dehulled soybean meal, 47% CP 37.50 BW: 41.6 ± 0.4 g; Ross × Ross, line 308) were allotted Pork meal, 50% CP 2.00 in a randomized complete block design with level on Dicalcium phosphate 1.50 Limestone 1.10 the battery as the block and pen as the experimental Salt 0.40 unit. Six replicates of 5 chicks were assigned to each Poultry trace-mineral premix1 0.15 treatment. At hatch, chicks were weighed and identified dl-Methionine 0.20 Choline chloride, 60% 0.10 with wing bands. They were assigned to pens in a man- Poultry vitamin premix2 0.20 ner to equalize mean BW across all 48 pens. The chicks Soybean oil 4.00 were housed in starter batteries with raised wire floors Analyzed composition CP 23.65 inside disease-containment chambers where the temper- Ca 0.79 ature was controlled and there was continuous lighting. P 0.41 The Salmonella-challenged chicks were in one chamber, 1Provided the following per kilogram of diet: Mn, 75 mg (MnO); Fe, 75 the sham-challenged ones in a separate chamber to re- mg (FeSO4·H2O); Zn, 75 mg (ZnO); Cu, 5 mg (CuSO4·H2O); I, 0.75 mg duce the possibility of cross-contamination. Feeding (ethylene diamine dihydroiodide); Se, 0.1 mg (Na2SeO3). 2 and water troughs were attached to the side of each Provided the following per kilogram of diet: retinyl acetate, 1,514 μg; cholecalciferol, 25 μg; dl-α-tocopheryl acetate, 11 mg; niacin, 22 pen. There were a feeder and drinker in each pen. Pens mg; d-Ca-pantothenate, 10 mg; riboflavin, 4.4 mg; vitamin B12, 11 μg; were randomly assigned to 1 of 2 infection treatments menadione sodium bisulfite, 2.33 mg. DIETARY CLAYS AND ENTERIC INFECTION 841 wanted feed restriction, the data for performance after Statistical Analyses d 3 PI for these 2 pens were not used. After the problem was resolved, chicks were able to access the feeders Data were analyzed by ANOVA with procedures ap- normally so the data for all the other parameters were propriate for a randomized complete block design. Data used. On 20 d of age (d 10 PI), chicks were euthanized were analyzed using the MIXED procedure (SAS Insti- by CO2 inhalation as previously described by Faber et tute Inc., Cary, NC) with blocks considered random. al., 2012, and 1 chick was randomly selected from each The statistical model included the fixed main effects of pen. Chicks were then dissected, and ileum, cecal tis- infection and diet, and their interaction. To test for the sue, and serum of 1 chick/pen were collected. presence of outliers, the PROC Univariate procedure of On d 1 through 9 PI, excreta were collected from SAS (SAS Institute Inc.) was used. Differences among each pen daily. Just before the collection, excreta pans the clay treatments within each challenge group were Downloaded from were cleaned and fresh butcher paper was replaced. Ap- tested by pair-wise comparisons when the overall main proximately 1 h later, all excreta from each pen was effect of diet or the diet × challenge interaction was collected, and weighed, and stored at −20°C until DM significant and in one case (ADFI d 3–7 PI) when P = analysis. 0.06. Differences with a probability of P ≤ 0.05 were The ileum was opened longitudinally and washed with accepted as statistically significant, whereas mean dif- http://ps.oxfordjournals.org/ PBS. Tissues were gently washed in buffered saline then ferences with P-values ranging from >0.05 to 0.10 were fixed in Carnoy’s solution, stained with high iron di- accepted as trends. amine and Alcian blue, pH 2.5, as previously described (Deplancke et al., 2000). The slides were scanned by In Vitro Experiment NanoZoomer Digital Pathology (NDP) System (Ham- amatsu Co., Bridgewater, NJ), and the total number of Treatments. One kind of smectite (SMA) was pro-

goblet cells per villus in the ileum was viewed, counted, vided by Milwhite Inc. (Brownsville, TX) for use in this at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 and the measure of the cross-sectional area (μm2) of research. The clay treatment was prepared at 3 differ- individual goblet cells (i.e., goblet cell sized) was per- ent concentrations (0.05, 0.10, and 0.50%) for the gene formed in the associated NDP.view software. The mea- expression data. The photos of the cell culture with surements were performed in 7 well-oriented villi at 40x the 0.50% concentration showed a considerable layer of resolution. The villus height (VH) was measured from clay covering the cells, leading to concern that the cells the tip of the villus to the valley between individual may not function normally. For that reason, the data villi, and crypt depth (CD) measurements were taken from that treatment are not presented and that treat- from the valley between individual villi to the basolat- ment was eliminated from the measurement of the gene eral membrane and the ratio VH:CD was calculated product. The lower concentrations of clay also partially (Fasina et al., 2010). These measurements were made covered the cells but to a more modest extent. The 2 in the same villi where the goblet cell size and number lower clay concentrations appear to be in the range were measured plus in 3 other well-oriented villi result- that might be encountered in vivo. The SMA was auto- ing in a total of 10 well-oriented villi per chick. claved before use to avoid contamination and minimum The mRNA of cecal tissue was assayed using re- essential medium (MEM) was added to achieve the verse-transcription PCR. The primers and probes desired concentrations of 0.05, 0.10, and 0.50%. were obtained from Applied Biosystems, Foster City, Cell Culture. The human colorectal adenocarcinoma CA, which included expression assays for interferon cell line (LS174T) passage 114 was obtained from (IFN)-γ (gene of interest; Gg03348616_m1) and glyc- the American Type Culture Collection. The cells were eraldehyde 3-phosphate dehydrogenase (housekeeping maintained in the culture flasks with MEM including gene, Gg03346982_m1). Quantitative real-time PCR 100 IU of penicillin/mL and 100 μg of streptomycin/ was performed using using TaqMAN PCR Master Mix mL and 10% fetal bovine serum (HyClone Laborato- (Applied Biosystems, Foster City, CA). Reactions were ries Inc., Logan, UT) at 37°C in 5% CO2 as previously run in triplicates or quadruplicates in a 384-well plate described (Zheng et al., 2011). All treatments were in- using the ABI PRISM 7900 Sequence Detection System cluded on each of 3 separate occasions, each time with (Applied Biosystems). Thermal cycling conditions were 3 replicate wells. Each well was an experimental unit. 50°C for 2 min and 95°C for 10 min, followed by 40 Cells were harvested after 24 h incubation for gene ex- cycles with 15 s at 95°C and 1 min at 60°C. The cycle pression measurements and after 96 h of incubation for threshold (Ct) values were calculated, and fold change protein measurements and stored at −80°C for RNA was calculated as 2−ΔΔCt as previously described (Li- isolation. The medium was harvested after 96 h incu- vak and Schmittgen, 2001). bation for protein measurements, but the medium was The concentration of the acute phase protein α-1-acid changed after 48 h incubation so the medium that was glycoprotein (α-1-AGP) in the serum was measured collected reflects only the last 48 h of incubation. by ELISA according to the manufacturer’s recommen- Quantitative Real-Time Reverse-Transcription dation (Life Diagnostics Inc., West Chester, PA). All PCR. The β-glucuronidase (GUSß) enzyme catalyzes samples were analyzed in duplicate. breakdown of complex carbohydrates and was used as 842 ALMEIDA ET AL. a housekeeping gene. All data obtained from the target RESULTS genes were normalized according to the expression of the housekeeping gene. In Vivo Experiment and Growth Total RNA was isolated from cells using the RNeasy Performance mini kit (Qiagen Inc., Valencia, CA) according to the manufacturer’s instruction for animal cells. The Dry matter analysis of excreta was conducted, but probes and primers were obtained from Applied Bio- there were no effects of challenge or dietary treatments systems, which included expression assays for MUC2 so data are not presented. (Hs00159374_m1), TFF3 (Hs00173625_m1), RELMß Several measures of growth performance were reduced (Hs00395669_m1), and GUSß (Hs99999908_m1). by the Salmonella challenge in chicks fed the basal diet, Quantitative real-time PCR was performed as de- but those measures in chicks fed clays were similar to Downloaded from scribed above. those of unchallenged chicks with no clear differences Immunoblot Analysis. For the cell lysate analysis, among the clays [Table 2; interaction (P < 0.05)]. The at the end of the incubation period (96 h), cells were Salmonella challenge reduced (P < 0.05) the growth washed with Hanks’ buffered salt solution, then lysed rate of chicks fed the basal diet by 11% during d 3 to with 400 μL of 1X Cell Lysis Buffer (Cell Signaling 7 PI, approximately the period of active infection. The http://ps.oxfordjournals.org/ Technology Inc., Danvers, MA) with protease inhibitor interaction between challenge and diet occurred (P < cocktail (Sigma Aldrich Co., St. Louis, MO). Follow- 0.05) for ADFI d 7 to 10 and ADFI and ADG during ing incubation for 5 min on ice, cells were collected the overall period, and the pattern was similar but not by scraping and then sonicated. The supernatant of significant for other measures. There were no effects of the extract was removed for use. An aliquot contain- clays on G:F (Table 2). ing 40 μg of protein was electrophoresed on 10 to 15% polyacrylamide gel and transferred to a polyvinylidene Intestinal Barrier at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 fluoride membrane (Fisher Scientific Inc., Pittsburgh, PA) as previously described (Fujino et al., 2003). Mem- Goblet Cell Number and Size. The pairwise com- branes were incubated with skim milk reconstituted in parison indicated that Salmonella challenge increased Tris-buffered saline (8–10% final concentration) with both the number (P < 0.05; interaction) and the size 0.1% Tween 20 (PBST) for 1 h at room temperature. (P < 0.05; diet) of ileal goblet cells in chicks fed the The bound proteins were probed with the primary an- basal diet (Table 3). Feeding one of the clays (SMA) tibodies, RELMβ (rabbit anti-human at a dilution 0.2 to the Salmonella-challenged chicks reduced (P < 0.05) μg/mL) and β-actin (mouse monoclonal antibody at both the number and size of those cells compared with a dilution of 1:800) overnight at 4°C. Following incu- challenged chicks fed the basal diet. The goblet cell bation, the blots were extensively washed PBST and number was 132.0 in the chicks fed the basal diet in incubated with respective (goat anti-rabbit and rabbit the Salmonella-challenged group, and 99.8 in the chicks anti-mouse) horseradish peroxidase-labeled secondary fed the SMA diet in the Salmonella-challenged group. antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) The goblet cell size was 29.7 μm2 in the chicks fed the at a dilution of 1:2,000 for 1 h at room temperature. basal diet in the Salmonella-challenged group, and 23.4 After thorough washing in PBST, blots were developed μm2 in the chicks fed the SMA diet in the Salmonella- using the Super Signal West Femto System (Thermo challenged group. The SMB and ZEO improved growth Scientific, Rockford, IL) 1:1 and incubated for 3 min in performance of Salmonella-challenged birds, but they the dark. Membranes were photographed using Image did not have the same effect on goblet cell number and Quanto LAS 4000 (GE, Uppsala, Sweden) at −25°C size as SMA did. The ZEO treatment increased (P < and the densitometry of the bands was performed using 0.05) the number of goblet cells in the sham-challenged Image J analysis software (NIH). birds. For the media analysis, the procedure was as de- Ileal Morphometry. Villus height was reduced (P < scribed above except that, at the end of the incubation 0.05) by the Salmonella challenge in chicks fed dietary period (96 h), cells were washed with Hanks’ buffered clays (Table 4). The SMA and ZEO reduced (P < 0.05) salt solution and the medium was collected and lyophi- VH:CD in the Salmonella-challenged chicks compared lized, then resuspended in 500 μL of PBS 1× solution, with chicks fed the basal diet (Table 4). then quantified for total protein before electrophoresis. Immune System Statistical Analysis mRNA Expression. The Ct values for IFN-γ were Data were analyzed using the GLM procedure (SAS too high (>33) to be meaningful, reflecting very low Institute Inc., Cary, NC). The model included treat- mRNA concentrations, so the data are not reported ments only, and each level of clay was compared with here. the control. Each well was considered an experimental Acute Phase Protein Į--AGP . The Salmonella unit. An α value of 0.05 was used to assess significance challenge did not change the concentrations of the among means. acute phase protein (α-1-AGP). Clays did not alter DIETARY CLAYS AND ENTERIC INFECTION 843

Table 2. Effect of clays on growth performance of broiler chicks experimentally infected with Salmonella enterica serovar Typhimuri- um1

Treatment2

Sham SALM P-value,3 main effect

Item BAS SMA SMB ZEO BAS SMA SMB ZEO SEM S D S × D Before challenge ADFI (g) 19.3 19.0 18.9 18.8 18.7 20.4 19.4 19.1 1.41 0.32 0.56 0.39 ADG (g) 14.8 14.9 13.7 14.0 14.3 15.9 15.1 14.4 1.54 0.21 0.24 0.43 G:F 0.77 0.78 0.73 0.74 0.76 0.78 0.78 0.76 0.044 0.25 0.25 0.43 d 0 to 3 Downloaded from ADFI (g) 48.1 48.0 49.2 48.6 45.4 51.1 46.1 46.5 3.41 0.21 0.22 0.10 ADG (g) 43.7 44.0 44.7 43.3 39.4 43.3 41.4 43.2 3.14 0.03 0.39 0.32 G:F 0.91 0.92 0.91 0.89 0.87 0.86 0.90 0.93 0.051 0.28 0.67 0.12 d 3 to 7 ADFI (g) 64.2ab 66.1a 66.1a 64.9ab 58.8b 64.0ab 62.6ab 65.5ab 3.56 0.02 0.06 0.23 ADG (g) 54.8 55.8 55.8 54.5 48.8 52.6 52.7 54.4 2.91 0.01 0.13 0.14 G:F 0.85 0.85 0.84 0.84 0.83 0.82 0.84 0.83 0.021 0.02 0.62 0.55 http://ps.oxfordjournals.org/ d 7 to 10 ADFI (g) 84.6ab 81.6ab 87.0a 82.3ab 71.4b 85.4ab 84.4ab 87.0a 7.49 0.41 0.09 0.03 ADG (g) 62.4 57.4 66.0 59.9 56.6 65.5 65.0 66.7 9.26 0.45 0.44 0.22 G:F 0.74 0.70 0.76 0.73 0.78 0.77 0.77 0.77 0.095 0.14 0.85 0.92 Overall, after challenge ADFI (g) 65.5a 65.3a 67.3a 65.2a 57.1b 66.6a 64.2ab 66.2a 3.94 0.06 0.02 0.02 ADG (g) 53.7ab 52.7ab 55.5a 52.8ab 47.1b 53.7ab 53.2ab 54.7a 3.89 0.18 0.10 0.05 G:F 0.82 0.81 0.83 0.81 0.83 0.81 0.83 0.83 0.036 0.57 0.58 0.94

a,bValues within a row lacking a common superscript letter are different (P < 0.05). at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 1n = 6 pens/treatment except n = 5 pens/treatment for BAS and SMB in challenged group for all periods except before challenge and 0 to 3 postinfection. Initial BW: 41.6 ± 0.4 g. 2Sham = unchallenged; SALM = Salmonella challenged; BAS = basal diet; SMA = 0.3% smectite A; SMB = 0.3% smectite B; ZEO = 0.3% zeolite. 3S = Salmonella challenge effect; D = diet effect; S × D = interaction between Salmonella Typhimurium and diet effects.

Table 3. Effect of clays on goblet cell number and size in ileum of chicks experimentally infected with Salmonella enterica serovar Typhimurium1

Treatment2

Sham SALM P-value,3 main effect

Ileum BAS SMA SMB ZEO BAS SMA SMB ZEO SEM S D S × D Number4 99.5b 107.5b 113.3ab 121.4a 132.0a 99.8b 116.9a 118.2a 5.27 0.26 0.16 0.04 Size5 (μm2) 23.3b 22.2b 28.0ab 27.1ab 29.7a 23.4b 28.6a 30.1a 1.52 0.07 0.04 0.56 a,bValues within a row lacking a common superscript letter are different (P < 0.05). 1n = 6 chicks/treatment. 2Sham = unchallenged; E. coli = Escherichia coli challenged; BAS = basal diet; SMA = 0.3% smectite A; SMB = 0.3% smectite B; ZEO = 0.3% zeolite. 3S = Salmonella challenge effect; D = diet effect; S × D = interaction between Salmonella and diet effects. 4Goblet cell number; total number of goblet cells per villus, average of 7 villi. 5Goblet cell size, cross-sectional area.

Table 4. Effects of clays on villus height (VH) and crypt depth (CD) of chicks experimentally infected with Salmonella enterica serovar Typhimurium1

Treatment2

Sham SALM P-value,3 main effect

Item BAS SMA SMB ZEO BAS SMA SMB ZEO SEM S D S × D VH 391 399 415 426 409 353 384 355 15.7 0.05 0.67 0.24 CD 143.9 169.3 156.5 160.8 145.3 151.0 147.5 147.2 4.39 0.03 0.11 0.44 VH:CD4 2.75ab 2.39b 2.76ab 2.79ab 2.93a 2.40b 2.68ab 2.50b 0.099 0.68 0.02 0.41 a,bValues within a row lacking a common superscript letter are different (P < 0.05). 1n = 6 chicks/treatment. 2Sham = unchallenged; SALM = Salmonella challenged; BAS = basal diet; SMA = 0.3% smectite A; SMB = 0.3% smectite B; ZEO = 0.3% zeolite. 3S = Salmonella challenge effect; D = diet effect; S × D = interaction between SALM and diet effects. 4VH:CD = villus height to crypt depth ratio. 844 ALMEIDA ET AL.

Table 5. Effects of clays on serum α-1-acid glycoprotein (α-1-AGP) levels of chicks experimentally infected with Salmonella enterica serovar Typhimurium1

Treatment2

Sham SALM P-value,3 main effect

Item BAS SMA SMB ZEO BAS SMA SMB ZEO SEM S D S × D α-1-AGP (μg/mL) 182b 266ab 151b 210ab 188b 241ab 194b 333a 21.55 0.08 <0.01 0.08 a,bValues within a row lacking a common superscript letter are different (P < 0.05). 1n = 6 chicks/treatment.

2Sham = unchallenged; SALM = Salmonella challenged; BAS = basal diet; SMA = 0.3% smectite A; SMB = 0.3% smectite B; ZEO = 0.3% zeolite. Downloaded from 3S = Salmonella challenge effect; D = diet effect; S × D = interaction between SALM and diet effects. the concentration of α-1-AGP in the sham-challenged an increase (P < 0.05) of RELMβ secretion into the

group but did in the Salmonella-challenged group (Ta- medium (Figure 1C). http://ps.oxfordjournals.org/ ble 5). Specifically, ZEO increased (P < 0.05) α-1-AGP concentration in the Salmonella-challenged group com- DISCUSSION pared with the basal group. Our results show that the clays tested herein inter- In Vitro Experiment acted with the Salmonella challenge, reducing the impact of Salmonella on the growth performance of broiler The mRNA expression of RELMβ (Table 6) was in- chicks, with no clear differences in responses among the creased by SMA in 2 of the 3 cultures (P < 0.04) and clays. To our knowledge, this is the first demonstra- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 in the overall data (P < 0.01). In the third culture, tion that clays in poultry diets alleviate detrimental the trend was in the same direction. The expression effects of enteric disease, although corresponding ben- of MUC2 was reduced (P = 0.02 overall) by a smaller efits have been reported for other species. Song et al. margin in the presence of 0.10% clay. These data show (2012) reported that, when pigs were challenged with no indication that the clay altered the expression of a pathogenic E. coli, feeding dietary clays including TFF3. smectite, zeolite, kaolinite, or combinations of them at The relative abundance of the gene product, RELMβ, 0.3% of the diet reduced diarrhea. Children with acute in control and clay-treated cells was assessed by immu- gastroenteritis when treated with smectite along with noblot analysis. A predominant band was detected in rehydration had shorter duration of diarrhea compared the cell lysate of LS174T cells exposed to MEM for with children treated without smectite (Szajewska et 96h (Figure 1A) but not in the cell lysate exposed to al., 2006). Clays’ effects on growth performance are clay. When LS174T cells were grown in the presence of variable. Some studies have shown no effect (Shurson et 0.05 or 0.10% SMA for 96 h, there was a reduction (P al., 1984; Poulsen and Oksbjerg, 1995), whereas others < 0.05) in RELMβ in the cell lysate (Figure 1B) and have shown a positive effect (Papaioannou et al., 2004;

Table 6. Effect of different concentrations of smectite A (SMA) on gene expression in human adenocarcinoma cell line (LS174T) cells, in fold change1

Clay (%) P-value

Cell Con2 versus Con versus Con versus cultures Gene 0 0.05 0.10 SEM 0.05% 0.1% SMA 1 MUC2 3 1.04 0.85 0.60 0.186 0.33 0.04 0.08 1 TFF3 4 1.01 1.10 1.04 0.117 0.50 0.83 0.54 1 RELMβ 5 1.03 1.56 1.27 0.164 0.01 0.18 0.03 2 MUC2 1.01 0.84 0.75 0.121 0.21 0.06 0.07 2 TFF3 1.19 1.07 1.15 0.131 0.36 0.76 0.49 2 RELMβ 1.03 1.12 1.68 0.125 0.47 <0.01 <0.01 3 MUC2 1.01 1.10 0.91 0.171 0.60 0.58 0.99 3 TFF3 1.10 1.04 1.13 0.269 0.83 0.91 0.95 3 RELMβ 1.00 1.12 1.38 0.240 0.63 0.14 0.26 All MUC2 1.02 0.94 0.77 0.100 0.44 0.02 0.06 All TFF3 1.10 1.07 1.11 0.110 0.33 0.90 0.64 All RELMβ 1.02 1.24 1.44 0.120 0.07 <0.01 <0.01 1Data are means of 11 observations; the first culture was conducted in triplicates and the last 2 in quadruplicates. 2Con = control. 3MUC2 = mucin 2. 4TFF3 = trefoil factor 3. 5RELMβ = resistin-like molecule β. DIETARY CLAYS AND ENTERIC INFECTION 845 and Gaskins, 2001; Collier et al., 2008; Fasina et al., 2010). Previous authors interpreted increases in size of goblet cells in response to pathogenic or commensal bacteria as an indication of increased mucin production (Deplancke and Gaskins, 2001; Collier et al., 2008; Fa- sina et al., 2010) and therefore as a protective response. The size of goblet cells increases during the acute phase of infection, but diminishes during the chronic phase as the secretion of the cell products is not matched by new synthesis (Kim and Ho, 2010). Further, RELMβ triggers secretion of mucins, depleting the cells and re- Downloaded from ducing cell size (Krimi et al., 2008). The in vitro experiment was conducted to determine if SMA alters the intestinal barrier by increasing expression of any of 3 genes involved in maintaining that barrier. When LS174T cells were exposed to SMA at http://ps.oxfordjournals.org/ 0.10%, there was an increase in RELMβ and a decrease in MUC2 mRNA expression. The magnitude of the response was greater for the upregulation of RELMβ (41%) than for the downregulation of MUC2 (25%) so we focused measurement of gene products on RELMβ only. Moreover, the upregulation of RELMβ is consis-

tent with the in vivo results in goblet cell size data. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Because RELMβ is involved in maintenance of colonic epithelial cell barrier function, we can conclude that SMA may have a protective effect in maintaining the mucosal defense barrier by upregulating RELMB. The RELMβ is a secreted protein and the protein measured in the western blot was collected from the cell lysate (96 h incubation) as well as media (48 h incubation). The cells treated with SMA secreted more RELMβ so there was more RELMβ in media collected in the last 48 h of incubation and less in the cell lysate. These results of increased signal (mRNA expression) Figure 1. Effects of smectite A (SMA) at different concentrations and increased secretion of RELMβ demonstrate that in the production and secretion of resistin-like molecule β (RELMβ). SMA may have protective effects on the intestinal bar- (A) Human adenocarcinoma cell line (LS174T) cells exposed to mini- rier. mum essential medium (MEM) only (control treatment) for 96 h of incubation. Densitometry analysis of blots in cell lysate (B) and Differentiation of stem cells in the crypts into goblet medium (C). A predominant band was detected in cell lysate when cells is increased by inflammation, via stimulation of LS174T cells were treated with control treatment (CON; A). There Krüppel-like factor 4, a transcription factor that is ex- was a decrease in RELMβ in the cell lysate when LS174T cells were pressed in a variety of tissues including the epithelium exposed to 0.05 and 0.10% SMA and incubated for 96 h (B). There was an increase in RELMβ secretion by LS174T cells when exposed of the intestine and plays a role in cell differentiation to 0.05% SMA and incubated for 96 h in the medium (C). *P < 0.01; but inhibits cell proliferation (Evans and Liu, 2008). **P < 0.001. More precisely, the number of goblet cells increases during the acute phase of infection (Kim and Ho, 2010). Alexopoulos et al., 2007), or even a negative effect on Reasons for the reduction in goblet cell number when performance (Shurson et al., 1984) when fed to pigs. SMA was fed in the present research are not clear, but The present data showed no effect of clays on growth perhaps the protective effects of the clay reduce the in- performance when chicks were sham-challenged. fectious challenge, which reduces inflammation, which Goblet cells are specialized epithelial cells which reduces differentiation into goblet cells. secrete cysteine-rich proteins that become the core The relationship of the apparent downregulation of components of the mucus barrier, including MUC2, MUC2 to other observations, including reduced diar- RELMβ, and trefoil factors (Specian and Oliver, 1991). rhea (Song et al., 2012) is unclear. Diosmectite 0.1% The size of these cells is determined largely by the bal- has been shown to upregulate MUC2 levels in HT29- ance between the synthesis and secretion of these pro- MTX cells (Gonzales et al., 2004), in contrast to the teins, with MUC2 being quantitatively the most impor- current results. The simple fact that different clays are tant (Kim and Ho, 2010). This finding may explain the used in different experiments may explain the appar- increase in size of goblet cells in infected chicks in the ently conflicting results. It was shown in the in vivo present study, as well as in previous studies (Deplancke experiment that different clays have different effects 846 ALMEIDA ET AL. even though they sometimes are classified in a common concentration of SMA and the secretion of RELMβ was category, such as “smectites.” increased. There was no indication of similar effects of Overall, the present data on goblet cell number and the other clays, even though all clays appeared to pro- size support the notion that one dietary clay (SMA) vide beneficial effects by restoring the performance of has effects consistent with strengthening the intestinal the Salmonella-challenged birds, so they may function mucus barrier, but do not show that the other clays through different mechanisms. Combined, these data tested (SMB and ZEO) exert positive effects. The in- suggest that some, but not all, clays help improve bar- teraction between clays and Salmonella was significant rier function of chicks during an enteric infection in vi- for ADG (P < 0.05) and ADFI (P < 0.02) for the tro. To the best of our knowledge, this is the first time overall period where the Salmonella challenge reduced that expression of RELMβ was measured in LS174T ADG and ADFI but clay provided a beneficial effect cells exposed to SMA or even to clays. Downloaded from by restoring these parameters in the Salmonella-challenged group. The pattern of response leads to the suggestion, which requires further testing, that different REFERENCES clays produce benefits through different mechanisms,

Alexopoulos, C., D. S. Papaioannou, P. Fortomaris, C. S. Kyriakis, http://ps.oxfordjournals.org/ even among the smectites. Different clays have different A. Tserveni-Goussi, A. Yannakopoulos, and S. C. Kyriakis. 2007. chemical structures, which may provide different prop- Experimental study on the effect of in-feed administration of a cl- erties, but specific relationships between structures and inoptilolite-rich tuff on certain biochemical and hematological pa- the physiological effects reported here are unknown. rameters of growing and fattening pigs. Livest. Sci. 111:230–241. Carretero, M. I. 2002. Clay minerals and their beneficial effects upon The VH, CD, and VH:CD are used as indicators of in- human health. A review. Appl. Clay Sci. 21:155–163. testinal epithelial integrity. An overall reduction in VH Che, T. M., V. G. Perez, M. Song, and J. E. Pettigrew. 2012. Ef- after Salmonella challenge indicates damage to the bar- fect of rice and other cereal grains on growth performance, pig removal, and antibiotic treatment of weaned pigs under com- rier and reduction in nutrient absorption that are con- mercial conditions. J. Anim. Sci. http://dx.doi.org/10.2527/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 sistent with the reduction in performance and increase jas.2011-4916. in goblet cell size and number discussed earlier. How- Collier, C. T., C. L. Hofacre, A. M. Payne, D. B. Anderson, P. Kai- ever, a deepening of the crypt is also expected during ser, R. I. Mackie, and H. R. Gaskins. 2008. Coccidia-induced mu- cogenesis promotes the onset of necrotic enteritis by supporting infection and we did not observe it in this experiment. Clostridium perfringens growth. Vet. Immunol. Immunopathol. We also observed an unexpected reduction in VH:CD 122:104–115. when SMA was fed in the Salmonella-challenged group. Corfield, A. P., N. Myerscough, R. Longman, P. Sylvester, S. Arul, The values observed for VH, CD, and VH:CD ratio are and M. Pignatelli. 2000. Mucins and mucosal protection in the gastro-intestinal tract: New prospects for mucins in the pathol- in the range reported by other authors (Tsirtsikos et ogy of gastro intestinal disease. Gut 47:589–594. al., 2012; Yitbarek et al., 2012; Zhang et al., 2012). Deplancke, B., and H. R. Gaskins. 2001. Microbial modulation of The timing of the various components of the immune innate defense: Goblet cells and the intestinal mucus layer. Am. response after infection of chicks with the Salmonella J. Clin. Nutr. 73:1131S–1141S. Deplancke, B., K. R. Hristova, H. A. Oakley, V. J. McCracken, R. used in this experiment is not well described, and may Aminov, R. J. Mackie, and H. R. Gaskins. 2000. Molecular eco- explain the absence of effects of the challenge on mea- logical analysis of the succession and diversity of sulfate-reducing sures of immunity. bacteria in the mouse gastrointestinal tract. Appl. Environ. Mi- The lack of response of the acute phase protein to crobiol. 66:2166–2174. Droy-Lefaix, M. T., Y. Drouet, and B. Schatz. 1985. Sodium glyco- the challenge agrees with the lack of response of IFN-γ deoxycholate and spinability of gastrointestinal mucus: Protec- mRNA expression to the challenge at 10 d PI. Holt and tive effect of smectite. Gastroenterology 88:1369. (Abstr.) Gast (2002) reported an increase in α-1-AGP 3 at 9 d Evans, P. M., and C. Liu. 2008. Roles of krüppel-like factor 4 in normal homeostatis, cancer and stem cells. Acta Biochim. Biophys. after hens were challenged with Salmonella enteriditis; Sin. (Shanghai) 40:554–564. however, the timing of the immune response may differ Faber, T. A., R. N. Dilger, M. Iakiviak, A. C. Hopkins, N. P. Price, between the mature hens in that experiment and the and G. C. Fahey Jr. 2012. Ingestion of a novel galactoglucoman- young chicks in the present experiment. Young chicks nan oligosaccharide-arabinoxylan (GGMO-AX) complex affected growth performance and fermentative and immunological charac- produce α-1-AGP after LPS injection (Takahashi et al., teristics of broiler chicks challenged with Salmonella typhimuri- 1994), but that does not provide insight into the time um. Poult. Sci. 91:2241–2254. of response after Salmonella infection. Fasina, Y. O., F. J. Hoerr, S. R. McKee, and D. E. Conner. 2010. In conclusion, challenging chicks with Salmonella re- Influence of Salmonella enterica serovar Typhimurium infection on intestinal goblet cells and villous morphology in broiler chicks. duced performance, but feeding a low dietary concen- Avian Dis. 54:841–847. tration of clay moderated or prevented that reduction, Fioramonti, J., M. T. Droy-Lefaix, and L. Bueno. 1987. Changes in with no clear differences among the 3 clays tested. This gastro-intestinal motility induced by cholera toxin and experi- response confirms that small amounts of these clays mental osmotic diarrhea in dogs: Effects of treatment with an argillaceous compound. Digestion 36:230–237. provide benefits to chicks challenged with enteric dis- Fujino, H., W. Xu, and J. W. Regan. 2003. Prostaglandin E2 in- ease, as occurs in other species. One of the clays (SMA) duced functional expression of early growth response factor-1 by reduced goblet cell size and number, suggesting its ben- EP4, but not EP2, prostanoid receptors via the phosphatidylino- efit may be mediated through strengthening the mu- sitol 3-kinase and extracellular signal-regulated kinases. J. Biol. Chem. 278:12151–12156. cus barrier. When LS174T cells were exposed to SMA, Gonzales, R., F. S. de Medina, O. Martinez-Augustin, A. Nieto, J. the expression of RELMβ was upregulated at the lower Galvez, S. Risco, and A. Zarzuelo. 2004. Anti-inflammatory ef- DIETARY CLAYS AND ENTERIC INFECTION 847

fect of diosmectite in hapten-induced colitis in the rat. Br. J. 2009. Interactions between ingested kaolinite and the intestinal Pharmacol. 141:951–960. mucosa in rat: Proteomic and cellular evidences. Fundam. Clin. Hardy, B. 2002. The issue of antibiotic use in the livestock industry: Pharmacol. 23:69–79. What have we learned? Anim. Biotechnol. 13:129–147. Schell, T. C., M. D. Lindemann, E. T. Kornegay, D. J. Blodgett, Harper, A. F., M. J. Estienne, J. B. Meldrum, R. J. Harrell, and D. and J. A. Doerr. 1993. Effectiveness of different types of clay for E. Diaz. 2010. Assessment of a hydrated sodium calcium alumi- reducing the detrimental effects of aflatoxin-contaminated diets nosilicate agent and antioxidant blend for mitigation of alfatoxin- on performance and serum profiles of weanling pigs. J. Anim. induced physiological alterations in pigs. J. Swine Health Prod. Sci. 71:1226–1231. 18:282–289. Shurson, G. C., P. K. Ku, E. R. Miller, and M. T. Yokoyama. 1984. Holt, P. S., and R. K. Gast. 2002. Comparison of the effects of infec- Effects of zeolite A or clinoptilolite in diets of growing swine. J. tion with Salmonella enteritidis, in combination with an induced Anim. Sci. 59:1536–1545. molt, on serum levels of the acute phase protein, alpha 1 acid Song, M., Y. Liu, J. A. Soares, T. M. Che, O. Osuna, C. W. Mad-

glycoprotein, in hens. Poult. Sci. 81:1295–1300. dox, and J. E. Pettigrew. 2012. Dietary clays alleviate diarrhea of Downloaded from Katsumata, H., S. Kaneco, K. Inomata, K. Itoh, K. Funasaka, K. weaned pigs. J. Anim. Sci. 90:345–360. Masuyama, T. Suzuki, and K. Ohta. 2003. Removal of heavy met- Specian, R. D., and M. G. Oliver. 1991. Functional biology of intes- als in rinsing wastewater from plating factory by adsorption with tinal goblet cells. Am. J. Physiol. 260:C183–C193. economical viable materials. J. Environ. Manage. 69:187–191. Szajewska, H., P. Dziechciarz, and J. Mrukowicz. 2006. Meta-anal- Kim, Y. S., and S. B. Ho. 2010. Intestinal goblet cells and mucins in ysis: Smectite in the treatment of acute infectious diarrhea in health and disease: Recent insights and progress. Curr. Gastro- children. Aliment. Pharmacol. Ther. 23:217–227. enterol. Rep. 12:319–330. Takahashi, K., N. Kaji, Y. Akiba, and K. Tamura. 1994. Plasma http://ps.oxfordjournals.org/ Krimi, R. B., L. Kotelevets, L. Dubuquoy, P. Plainsancié, F. Walker, alpha 1-acid glycoprotein concentration in broilers: Influence of T. Lehy, P. Desreumaux, I. Van Seuningen, E. Chastres, M. E. age, sex and injection of Escherichia coli lipopolysaccharide. Br. Forgue-Lafitte, and J. C. Marie. 2008. Restistin-like molecule Poult. Sci. 35:427–432. beta regulates intestinal mucous secretion and curtails TNBS- Taupin, D., and D. K. Podolsky. 2003. Trefoil factors: Initiators of induced colitis in mice. Inflamm. Bowel Dis. 14:931–941. mucosal healing. Nat. Rev. Mol. Cell Biol. 4:721–732. Ledoux, D. R., and G. E. Rottinghaus. 2000. Animal model for test- Thim, L. 1997. Trefoil peptides: From structure to function. Cell. ing adsorbents to detoxify mycotoxins. Feed Mix 8:18–20. Mol. Life Sci. 53:888–903. Livak, K. J., and T. D. Schmittgen. 2001. Analysis of relative gene Thim, L., F. Madsen, and S. S. Poulsen. 2002. Effect of trefoil fac- − expression data using real-time quantitative PCR and the 2 tors on the viscoelastic properties of mucus gels. Eur. J. Clin. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 ΔΔCT method. Methods 25:402–408. Invest. 32:519–527. McVay, L. D., S. A. Keilbaugh, T. M. Wong, S. Kierstein, M. E. Trckova, M., H. Vondruskova, Z. Zraly, P. Alexa, J. Hamrik, V. Shin, M. Lehrke, M. I. Lefterova, D. E. Shifflett, S. L. Barnes, Kummer, J. Maskova, V. Mrlik, K. Krizova, I. Slana, L. Leva, F. Cominelli, S. M. Cohn, G. Hecht, M. A. Lazar, A. Haczku, and I. Pavlik. 2009. The effect of kaolin feeding on efficiency, and G. D. Wu. 2006. Absence of bacterially induced RELMbeta health status and course of diarrhoeal infections caused by en- reduces injury in the dextran sodium sulfate model of colitis. J. terotoxigenic Escherichia coli strains in weaned pigs. Vet. Med. Clin. Invest. 116:2914–2923. 54:47–63. Moncada, D., and K. Chadee. 2002. Production, structure, and Tsirtsikos, P., K. Fegeros, C. Balaskas, A. Kominakis, and K. C. biologic relevance of gastrointestinal mucins. Pages 57–79 in In- Mountzouris. 2012. Dietary probiotic inclusion level modulates fections of the Gastrointestinal Tract. Lippincott Williams and intestinal mucin composition and mucosal morphology in broil- Wilkins, Philadelphia, PA. ers. Poult. Sci. 91:1860–1868. NRC. 1994. Nutrient Requirements of Poultry. 9th rev. ed. National Ward, T. L., K. L. Watkins, L. L. Southern, P. G. Hoyt, and D. D. Academy Press, Washington DC. French. 1991. Interactive effects of sodium zeolite-A and copper Papaioannou, D. S., C. S. Kyriakis, C. Alexopoulos, E. D. Tzika, in growing swine: Growth, and bone and tissue mineral concen- Z. S. Polizopoulou, and S. C. Kyriakis. 2004. A field study on trations. J. Anim. Sci. 69:726–733. the effect of the dietary use of a clinoptilolite-rich tuff, alone or Yitbarek, A., H. Echeverry, J. Brady, J. Hernandez-Doria, G. Came- in combination with certain antimicrobials, on the health status lo-Jaimes, S. Sharif, W. Guenter, J. D. House, and J. C. Rodri- and performance of weaned, growing and finishing pigs. Res. guez-Lecompte. 2012. Innate immune response to yeast-derived Vet. Sci. 76:19–29. carbohydrates in broiler chickens fed organic diets and challenged Perez, V. G., A. M. Waguespack, T. D. Binder, L. L. Southern, T. with Clorstridium perfringes. Poult. Sci. 91:1105–1112. M. Fakler, T. L. Ward, M. Steidinger, and J. E. Pettigrew. 2011. Zhang, B., Y. Shao, D. Liu, P. Yin, Y. Guo, and J. Yuan. 2012. Zinc Additivity of effects from dietary copper and zinc on growth per- prevents Salmonella enterica serovar Typhimurium-induced loss formance and fecal microbiota of pigs after weaning. J. Anim. of intestinal mucosal barrier function in broiler chickens. Avian Sci. 89:414–425. Pathol. 41:361–367. Phillips, T. D., L. F. Kubena, R. B. Harvey, D. R. Taylor, and N. D. Zheng, X., K. Tsuchiya, R. Okamoto, M. Iwasaki, Y. Kano, N. Saka- Heidelbaugh. 1988. Hydrated sodium calcium aluminosilicate: A moto, T. Nakamura, and M. Watanabe. 2011. Suppression of high affinity sorbent for aflatoxin. Poult. Sci. 67:243–247. hath1 gene expression directly regulated by hes1 via notch signal- Poulsen, H. D., and N. Oksbjerg. 1995. Effects of dietary inclusion of ing is associated with goblet cell depletion in ulcerative colitis. a zeolite (clinoptilolite) on performance and protein metabolism Inflamm. Bowel Dis. 17:2251–2260. of young growing pigs. Anim. Feed Sci. Technol. 53:297–303. Reichardt, F., C. Habold, B. Chaumande, A. Ackermann, L. Ehret- Sabatier, Y. Le Maho, F. Angel, N. Liewig, and J.-H. Lignot.

Transcriptome analysis reveals an activation of major histocompatibility complex 1 and 2 pathways in chicken trachea immunized with infectious laryngotracheitis virus vaccine

Juan Luo ,*1 José A. Carrillo ,*1 Kimberly R. Menendez ,† Nathaniel L. Tablante ,† and Jiuzhou Song *2

* Department of Animal and Avian Sciences, and † Virginia-Maryland Regional College of Veterinary Medicine, Downloaded from University of Maryland, College Park 20742

ABSTRACT Infectious laryngotracheitis is an acute, immune genes, such as Granzyme A (GZMA), CD4 contagious, upper respiratory disease of chickens caused molecule (CD4), CD8a molecule (CD8A), and CD8b by gallid herpes virus 1. Due to mortality rates that can molecule (CD8B), that were upregulated upon vacci- http://ps.oxfordjournals.org/ reach up to 70% depending on the virulence of the vi- nation. The gene ontology analysis shows that genes rus, the disease is of great economic importance to the included in the biological process cluster were related poultry industry. In this study, 15-d-old specific patho- to antigen processing and presentation, positive regula- gen-free White Leghorn chickens were used to perform tion of immune system processes, T cell selection, and transcriptome analysis of chicken trachea immunized positive regulation of T cell activation. In conclusion, with infectious laryngotracheitis virus vaccine. Myo- chicken embryo origin vaccine activation of the major sin and several collagen-related genes were downregu- histocompatibility complex 1 and 2 pathways provides lated in the immunized group, suggesting that normal insight for evaluation and design of infectious laryngo- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 function and structure may be compromised. In addi- tracheitis vaccines. tion, we identified some cytokine receptors and several Key words: infectious laryngotracheitis, poultry , gene expression, vaccine , major histocompatibility complex 1 and 2 pathways

2014 Poultry Science 93 :848–855 http://dx.doi.org/ 10.3382/ps.2013-03624

INTRODUCTION by ILT-infected birds, and mortality rates can range from 0 to 70% depending on the virulence of the virus, Infectious laryngotracheitis (ILT) is an acute, con- highlighting the economic importance of the disease to tagious, upper respiratory disease of chickens caused the poultry industry (Fuchs et al., 2007). by gallid herpes virus 1 (Fuchs et al., 2007; Guy and Prevention of ILT is currently achieved through vac- Garcia, 2008). Gallid herpes virus 1 is an alpha herpes cination, typically including either of 2 live-attenuated virus commonly recognized as infectious laryngotrache- virus vaccines. The first vaccine, chicken embryo origin itis virus (ILTV), which comprises a genome size of (CEO) vaccine, is generated by sequential passage in approximately 150 kb and integrates the Iltovirus genus chicken embryos; the second, the tissue culture origin (Lee et al., 2011). As with other herpes viruses, the vaccine, is generated by sequential passage in tissue genome is composed of a unique long, unique short, in- culture (Saif, 2003; Rodríguez-Avila et al., 2007). How- verted repeat, and terminal repeat and encodes a total ever, due to inadequate attenuation of some vaccines of 80 predicted open reading frames (Lee et al., 2011). or nonuniform vaccination, these vaccine viruses can Birds infected with the virus display clinical signs, such be transmitted from bird-to-bird and revert to viru- as oculo-nasal discharge, conjunctivitis, coughing and lence, causing vaccinal laryngotracheitis (VLT). To sneezing, decreased production efficiency, and, in severe better understand the molecular basis of ILT pathogen- cases, expectoration of bloody mucus and death (Guy esis and to better control the disease, experiments were and Garcia, 2008). The spread of ILT occurs via direct conducted focusing on both the virus itself and host contact with infectious respiratory exudates secreted responses. Genome sequences of several virulent and vaccine ILT strains have been decoded in recent years. © 2014 Poultry Science Association Inc. These studies provide adequate reference sequences to Received September 14, 2013. identify new mutations of other strains that may be Accepted December 16, 2013. 1 These authors contributed equally to this work. involved in inducing new clinical signs (Thureen and 2 Corresponding author: [email protected] Keeler, 2006; Lee et al., 2011; Chandra et al., 2012;

848 MAJOR HISTOCOMPATIBILITY COMPLEX IN CHICKEN TRACHEA 849 Spatz et al., 2012). The expression of micro RNA in gen, Grand Island, NY), and 1% l-glutamine (Sigma- ILTV have also been identified, with one of the micro Aldrich, St. Louis, MO). Cells were incubated at 37°C RNA, ILTV-miR-I5, confirmed to be involved in down- for 96 h and plaque-forming units were then counted to regulation and cleavage of ICP4 mRNA (Rachamadugu quantify virus. et al., 2009; Waidner et al., 2011). Genome-wide host responses to virulent and vaccine ILT strains have been Experimental Design studied in vitro by using primary cultured chicken embryo lung cells (Lee et al., 2010). Different host tran- Ten 15-d-old specific pathogen-free White Leghorn scriptional profiles have also been identified between chickens were obtained (Charles River, Wilmington, the virulent and vaccine ILT strain-infected cells, with MA) and divided into 2 groups of 5 birds each. At 21 Downloaded from the expression of the genes bone morphogenetic pro- d of age, one group of 5 birds was inoculated with ster- tein 2 (BMP2), chromosome 8 open reading frame 79 ile vaccine diluent without addition of the lyophilized (C8orf79), coagulation factor X (F10), and neuropep- Trachivax vaccine (50 μL per nostril, 50 μL per eye; tide Y (NPY) being the most distinct during vaccine 200 μL total) and placed in a separate animal room as infection (Lee et al., 2012). Understanding the host re- a control group. The vaccine was then prepared accord- http://ps.oxfordjournals.org/ sponse after vaccination is essential for designing novel ing to the manufacturer’s recommendation for the eye and effective vaccines, and development of these vac- drop administration method. The second group of birds cines is crucial to the future of ILT control. However, no was then inoculated with Trachivax vaccine (50 μL per experiments have been done to study the host response nostril, 50 μL per eye; 200 μL total at a concentration to vaccination in vivo. of 3.3 × 103 plaque-forming units) and housed as a In the current study, we hypothesized that the iden- group in an animal isolator. All animal experiments tification of differential expressed genes induced by and procedures were performed under University of

ILTV would lead us to recognize the main pathways Maryland Animal Care and Use Committee guidelines at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 involved in the immunological response to ILT vacci- and approval (R-11–26). nation. Here we report our findings on a pilot study on the genome-wide transcriptional profile of in vivo host response to CEO vaccine infection by the RNA Clinical Scoring sequencing method (RNA-Seq). We identified a total Birds were monitored daily for clinical signs starting of 158 genes that are differentially expressed after ILT each day postinoculation and scored based on respira- CEO vaccination. Gene ontology (GO) term analysis tory signs and conjunctivitis on a 0 to 4 scale, where 0 of the genes revealed functions including antigen pro- = normal, 1 = mild eye inflammation, 2 = oculo-nasal cessing and presentation, positive regulation of immune discharge, 3 = coughing and sneezing, and 4 = expecto- system process, T cell selection, and positive regulation ration of bloody mucus and respiratory distress. of T cell activation. In addition, the CEO vaccine was found to induce various biological pathways, including the antigen processing and presentation pathways for Sample Collection both class I and class II major histocompatibility com- At 6 d postinoculation, all birds were humanely eu- plex (MHC-I and MHC-II, respectively). thanized and tracheal scrapings were obtained from each bird. Briefly, whole tracheas were dissected imme- MATERIALS AND METHODS diately after euthanasia. Using sterile scissors for each sample, each trachea was opened along its entire length Vaccine and the mucosa was scraped using a sterile disposable The vaccine used for this study was the modified live scalpel from the larynx to the syrinx. Scrapings were chick embryo origin Trachivax vaccine (Merck Animal placed in individual tubes. Health, Millsboro, DE). The vaccine was prepared according to the manufacturer’s recommendations for the mRNA Extraction and RNA-Seq eye drop administration method. The vaccine was ti- Library Preparation trated in chicken embryo kidney cells prepared from 20-d-old specific pathogen-free chicken embryos. Vac- The mRNA was extracted directly from tracheal cine was diluted from 10−1 to 10−5, and 4 replicates of samples using the Oligotex Direct mRNA Mini Kit each dilution were inoculated onto 12-well plates with (Qiagen, Valencia, CA) according to the manufac- 100% chicken-embryo kidney cell confluence for 1 h. turer’s instructions. First strand cDNA was generated After 1 h, cells were washed with sterile PBS and over- using the SuperScript II Reverse Transcriptase Kit laid with 0.8% methylcellulose in MEM d-valine media (Invitrogen, Carlsbad, CA) with random hexamers (Promo Cell, Heidelberg, Germany) containing 2% fe- (Invitrogen) following the manufacturer’s instructions tal bovine serum (HyClone Laboratories, Logan, UT), in a final volume of 40 μL. Second strand cDNA was 1% penicillin, streptomycin, amphotericin solution (MP synthesized using DNA polymerase I (NEB, Ipswich, Biomedicals, Solon, OH), 1% sodium pyruvate (Invitro- MA) supplemented with 30 nmol of deoxyribonucleo- 850 LUO ET AL. tides and 2 U of RNase H (Invitrogen), incubated at generate significantly enriched pathways that overlap 25° C for 2.5 h, then purified with QIAquick PCR pu- with our differentially expressed genes. rification kit (Qiagen). For cost savings, index libraries were constructed using Multiplexing Sample Prepara- Validation of RNA-Seq Results tion Oligonucleotide Kit (Illumina, San Diego, CA) by Quantitative Real-Time PCR following the manufacturer’s instructions. Briefly, the double-stranded cDNA was sonicated to approximately The RNA-Seq results were validated by quantita- 200 to 500 bp using a Bioruptor sonicator (Diagenode, tive real-time PCR (q-PCR) using an iCycler iQ PCR Denville, NJ) in a final volume of 50 μL for 40 min. system (Bio-Rad, Hercules, CA). The q-PCR reactions The resulting fragments were repaired using the end were performed in a final volume of 20 μL with the Downloaded from repair module (NEB) and 3cA was added by Klenow QuantiTect SYBR Green PCR Kit (Qiagen) accord- Fragment (NEB). The Solexa adaptors (Illumina) were ing to the manufacturer’s instructions. The primers then ligated to the double-stranded cDNA fragments were designed on NCBI (http://www.ncbi.nlm.nih.gov/ by T4 DNA ligase (NEB). Polymerase chain reaction tools/primer-blast/index.cgi) using the criteria that all with 18 cycles was performed using the adaptor primers amplicons would encompass at least one intron. Primer http://ps.oxfordjournals.org/ and index primers. The PCR products with a length of sequences are shown in Supplemental File 4 (http:// 200 to 500 bp were isolated by QIAquick Gel purifica- dx.doi.org/10.3382/ps.2013-03624). tion kit (Qiagen). Clustering and sequencing was then performed on the Hi-Seq 2000 (Illumina) following the RESULTS manufacturers protocols. RNA-Seq Analysis of Tracheal RNA-Seq Data Analysis Transcriptome After ILT CEO Vaccination at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 The sequenced files were aligned to the reference ge- To study the host response to CEO vaccine infec- nome galGal3 (WUGSC 2.1) obtained from the UCSC tion, mRNA from tracheal samples were extracted to (http://genome.ucsc.edu/) using Bowtie (ultrafast, construct 3 libraries for RNA-Seq analysis on Illumina memory-efficient short read aligner). For reading the Hi-Seq 2000 (Table 1). Prior to euthanasia and sample output in R, SAMtools was employed to transform the collection, all birds within the vaccinated group showed sam data to an R-compatible format (bam) (R Devel- moderate to acute clinical signs of VLT infection, rang- opment Core Team, 2008; Li et al., 2009). Account- ing from conjunctivitis (clinical score = 1) to coughing for differential expression levels was obtained using ing and sneezing (clinical score = 3), and received no DEseq (R package from Bioconductor, http://www. higher than a clinical score of 3 for the duration of the bioconductor.org/), which employs a negative binomial experiment. Additionally, all birds within the mock im- distribution model for the count data. This model is munization control group received clinical scores of 0, similar to previous tools, such as edgeR, but the vari- showing no signs of infection for the duration of the ex- ance estimation was determined locally using different periment. Two biological replicates from the vaccinat- variation coefficients for dissimilar expression strength ed samples and the age-matched mock immunization and demonstrates better fitness compared with other control sample were used to extract tracheal mRNA models that assume Poisson distribution for the data. for RNA-Seq analysis (Table 1). In the control sample, 37,206,542 raw reads were obtained with 23,295,641 GO Term and Pathway Analysis reads mapped to the chicken genome, which account of Differentially Expressed Genes for 62.61% of the raw reads. In the ILTV vaccinated samples, 65,553,649 and 31,250,060 raw reads were gen- The GO term analysis used as the functional annota- erated from the 2 replicates respectively. The mapabili- tion of the differentially expressed genes was performed ties of the vaccinated samples were 56.53 and 61.73%, on an online bioinformatics system [Database for An- respectively. notation, Visualization and Integrated Discovery (DA- VID); http://david.abcc.ncifcrf.gov/]. Genes over- and Gene Expression Profile of the Trachea underexpressed by a log fold change more than 1.5- After ILT CEO Vaccination fold in the ILT-immunized group compared with the mock-immunized group were input as the gene list and By using the DESeq package from R with the crite- chicken genes were set as background list. The default ria of log fold change greater than 1.5, a P-value less parameters in DAVID were used to perform the enrich- than 0.05, and a false discovery rate less than 0.1, a ment of GO biological processes of genes. Gene pathway total of 158 genes that were significantly differentially enrichment was obtained from a web-based pathway expressed following vaccination were found. Among analysis program [Kyoto Encyclopedia of Genes and these, 113 genes were upregulated in the vaccinated Genomes (KEGG); http://www.kegg.jp]. The chicken samples, compared with the mock-immunized samples, and human KEGG genes were used as a reference to and 45 downregulated (Supplemental File 1; http:// MAJOR HISTOCOMPATIBILITY COMPLEX IN CHICKEN TRACHEA 851

Table 1. Sequencing and mapping details of the infectious laryngotracheitis virus vaccinated and mock-immunized samples

Reads Aligned Unaligned Sample processed reads reads Aligned (%) Control 37,206,542 23,295,641 13,910,901 62.61 Vac-11 65,553,649 37,060,417 28,493,232 56.53 Vac-22 31,250,060 19,286,773 11,963,287 61.72 1Infectious laryngotracheitis virus vaccinated sample 1. 2Infectious laryngotracheitis virus vaccinated sample 2. Downloaded from dx.doi.org/10.3382/ps.2013-03624). The MA plot (M = of biological processes, cellular components, and mo- log ratio and A = mean average) was used to illustrate lecular functions in total (Supplemental File 2; http:// the magnitude distribution of the genes significantly dx.doi.org/10.3382/ps.2013-03624). The genes in the altered by ILTV vaccination (Figure 1). The differen- cluster of biological processes were related to functions tially expressed genes and annotation of the Ensembl such as, but not limited to, antigen processing and pre- http://ps.oxfordjournals.org/ genes by DAVID are shown in Supplemental File 1. sentation, positive regulation of immune system pro- In the significant genes list, several collagen family cesses, T cell selection, and positive regulation of T cell genes were downregulated, such as collagen, type I, α 2 activation (Figure 2). (COL1A2), collagen, type II, α 1 (COL2A1), collagen, type III, α 1 (COL3A1), and collagen, type XXI, α 1 ILT CEO Vaccination Activates Antigen (COL21A1). In addition, some cytokine receptors, such Processing and Presentation Pathway as IL-2 receptor, gamma (IL2RG) and IL-13 receptor at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 subunit α-2 (IL13RA2), and several immune genes, Following GO term analysis, investigation into the such as granzyme A (GZMA), CD4 molecule (CD4), pathways involved was performed. The list of differ- CD8a molecule (CD8A), and CD8b molecule (CD8B), entially expressed genes was overlapped with KEGG were upregulated upon vaccination. pathways, and 6 pathways were found to be significantly enriched (P < 0.05; Table 2), including primary GO Term Analysis of the Differentially immunodeficiency, antigen processing and presenta- Expressed Genes tion, viral myocarditis, T cell receptor signaling pathway, and hematopoietic cell lineage and cell adhesion To annotate the functions of the differentially ex- molecules. All genes involved in the antigen processing pressed genes during vaccination, we performed GO and presentation pathway were upregulated for both analysis by using the online bioinformatics tool DA- the MHC-I and MHC-II antigen processing pathways. VID. The GO term annotation contains 3 functional For the MHC-I pathway, 4 genes were induced by vacci- clusters including biological processes, cellular compo- nation, including β-2-microglobulin (B2M), transporter nents, and molecular functions. Among the 158 differ- 2, ATP-binding cassette, sub-family B (TAP), CD8A, entially expressed genes, 156 genes were functionally and CD8B (Figure 3). Three genes, B locus M α chain annotated and assigned (P < 0.05) into 91 categories 1(BMA1/MHC-II), CD74 molecule (CD74), and CD4, which encode 6 proteins, were induced for the MHC-II pathway (Figure 3).

q-PCR Validation of the RNA-Seq Results and Activated Antigen Processing and Presentation Pathway To further clarify the upregulation of cell adhesion molecules pathway by ILT CEO vaccination and the quality of RNA-Seq, q-PCR was performed on genes involved in the antigen processing and presentation pathway, including CD4, CD8A, and CD8B, in addition to 13 other genes that were randomly chosen from the significant gene list. As shown in Table 3, the Figure 1. MA plot (M = log ratio and A = mean average) of majority of the results were consistent with the RNA- significantly differentially expressed genes in infectious laryngotrache- Seq results. Additionally, the linear regression analysis itis virus vaccination-immunized and mock-immunized trachea. The revealed a good correlation (rho = 0.62) between the MA plot represents the log2 fold changes in y, against the normalized means in the x-axis. The black dots are the differentially expressed q-PCR and RNA-Seq (Supplemental File 3; http:// genes with a P-value less than 0.05. dx.doi.org/10.3382/ps.2013-03624). 852 LUO ET AL.

Table 2. Significantly enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways in differentially expressed genes induced by infectious laryngotracheitis virus vaccination

KEGG pathway n % P-value Ensembl gene Primary immunodeficiency 8 5.59 4.16E-08 ENSGALG00000004958, ENSGALG00000007418, ENSGALG00000014477, ENSGALG00000015816, ENSGALG00000015902, ENSGALG00000021139, ENSGALG00000005638, ENSGALG00000000181 Antigen processing and presentation 7 4.89 1.76E-04 ENSGALG00000014477, ENSGALG00000004594, ENSGALG00000015816, ENSGALG00000015902, ENSGALG00000002160, ENSGALG00000000158, ENSGALG00000000181

Viral myocarditis 5 4.19 7.19E-04 ENSGALG00000007511, ENSGALG00000000158, Downloaded from ENSGALG00000023705, ENSGALG00000023800, ENSGALG00000021211 T cell receptor-signaling pathway 6 4.89 7.35E-04 ENSGALG00000015441, ENSGALG00000007418, ENSGALG00000014477, ENSGALG00000015816, ENSGALG00000015902, ENSGALG00000010870 Hematopoietic cell lineage 4 3.49 0.011 ENSGALG00000007418, ENSGALG00000014477, ENSGALG00000015816, ENSGALG00000015902 http://ps.oxfordjournals.org/ Cell adhesion molecules 5 3.49 0.045 ENSGALG00000014477, ENSGALG00000015816, ENSGALG00000015902, ENSGALG00000007511, ENSGALG00000000158 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 2. Gene ontology (GO) term analysis of the differentially expressed genes induced by infectious laryngotracheitis virus vaccination. The GO term analysis was performed on Database for Annotation, Visualization and Integrated Discovery by using chicken genes as background. Only the biological process terms with an enrichment of P-value < 0.01 were shown in this figure. MAJOR HISTOCOMPATIBILITY COMPLEX IN CHICKEN TRACHEA 853

Table 3. Validation of the RNA sequencing results

Quantitative RNA Ensembl identification Gene symbol PCR sequencing ENSGALG00000014477 CD4 18.84 49.15 ENSGALG00000015816 CD8A 17.67 43.81 ENSGALG00000015902 CD8B 5.03 93.37 ENSGALG00000007511 ITGB2 4.64 17.89 ENSGALG00000023705 MYH3 0.06 0.04 ENSGALG00000005693 NOS2 0.14 0.29 ENSGALG00000009963 LYZ 0.14 0.13 ENSGALG00000010299 RSPH9 0.31 0.25

ENSGALG00000000184 SLC27A6 0.36 0.14 Downloaded from ENSGALG00000008997 STEAP4 0.36 0.07 ENSGALG00000009641 COL1A2 0.38 0.19 ENSGALG00000011696 JAG2 0.77 0.16 ENSGALG00000013575 ISG12–2 2.01 10.76 ENSGALG00000015441 CD247 3.75 19.09 ENSGALG00000007713 CD300A 5.71 9.82 ENSGALG00000013548 GZMA >3,650 94.97 http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 3. Infectious laryngotracheitis virus vaccine significantly activated genes enriched in antigen processing and presentation pathway (Kanehisa and Goto, 2000). Gray rectangle = significantly activated genes (log fold change >1.5, P-value < 0.05, false discovery rate <0.1); white rectangle = expression unchanged genes. IFN = interferon; TNF = tumor necrosis factor; TCR = T cell receptor; NK = natural killer; PA28 = proteasome activator subunit 3; HSP70 = heat shock 70 kDa protein 1A; HSP90=heat shock protein 90 kDa alpha; BiP = binding immunoglobin protein; CANX = calnexin; BRp57=protein disulfide isomerase family A, member 3; CALR = calreticulin; MHCI = MHC, class I; TAPBP = TAP binding protein (tapasin); GILT = gamma-interferon-inducible lysosomal thiol reductase; AEP=alkaline extracellular protease (legumain); CTSB = cathepsin B; li = CD74 molecule, MHC, class II invariant chain; SLIP = speckled-like pattern; CLIP = cytoplasmic linker protein; HLA-DM = MHC, class II, DM alpha; CIITA = class II, MHC, transactivator; RFX = regulatory factor X 5; CREB = cAMP responsive element binding protein 1; NFY = nuclear transcription factor Y, alpha; TAP = transporter 2, ATP-binding cassette, subfamily B; B2M = β-2-microglobulin. 854 LUO ET AL. DISCUSSION 2010). Following vaccination, the vaccine virus is sensed by dendritic cells or macrophages and internalized into Live attenuated ILTV vaccines have been widely used phagosomes, subsequently undergoing a proteolytic to control ILT for many years, with the CEO vaccine processing to produce antigenic peptides delivered by representing a dominant choice of the industry (Bagust MHC-I and MHC-II molecules to the cell surface. The et al., 2000; Saif, 2003; Rodríguez-Avila et al., 2007). MHC-I molecules present antigenic peptides by an in- However, the ability of these vaccines to revert to viru- tracellular route to CD8 + T cells, whereas MHC-II lence remains a worldwide problem, resulting in out- molecules present peptides from an endogenous route breaks of ILT in major poultry-producing regions, such to CD4+ T cells (Neefjes et al., 2011). In the MHC-I as North America (Oldoni and Garcia, 2007), Europe intracellular presentation pathway, one specifically tar- (Neff et al., 2008), and Australia (Blacker et al., 2011). geted by herpes viruses, the cleaved antigenic peptides Downloaded from To better understand host responses to vaccination at are translocated into the endoplasmic reticulum lumen the molecular level and pave the way for the design of by the transporter associated with antigen processing novel and effective vaccines, RNA-Seq was used to iden- to then be loaded on to MHC-I molecules (Neefjes et tify the genome-wide transcriptome induced by one of al., 2011). Before they are capable of carrying peptides, the ILT CEO vaccines in the trachea. The CEO vaccine MHC-I molecules are assembled from a polymorphic http://ps.oxfordjournals.org/ used in this experiment, though attenuated, remains heavy chain and light chain, called B2M, within the capable of eliciting a humoral as well as local (cell- endoplasmic reticulum (Neefjes et al., 2011). This path- mediated) immune response similar to that caused by a way is targeted by herpes virus proteins, such as the virulent virus. Several collagen family genes, COL1A2, ICP47 protein in herpes simplex virus, specifically in- COL2A1, COL3A1, and COL21A, were downregulat- hibiting the function of TAP, leading to repression of ed by ILT CEO immunization, indicating that normal MHC-I pathway, thus allowing these herpes viruses to function and structure may be compromised by the replicate freely in infected cells (Hill et al., 1995). In at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 vaccine, as collagen proteins are important for molecu- our experiment, we identified upregulation of both TAP lar architecture, shape, and mechanical properties of and B2M following immunization, indicating activa- tissues (Ricard-Blum, 2011). In addition, myosin genes, tion of the MHC-I antigen processing pathway by ILT such as myosin heavy chain 3 (MYH3), were downregu- vaccine virus, which is essential to vaccine stimulation lated and represent a group of actin-dependent motor of antigen-specific CD8+ T cells for the formation of proteins known for their function in muscle contraction. immunological memory (Harty and Badovinac, 2008). This loss of myosin expression as a result of vaccination In addition to the MHC-I antigen processing pathway, may be responsible for the various clinical signs, such activation of MHC-II antigen processing pathway in as gasping and dyspnea associated with VLT. immunized samples was found. Stimulation of specific Aside from the genes involved in the physiological helper T cell subsets and other antigen-specific lym- reaction to ILT CEO vaccination, many genes, such phocytes is a critical factor involved in the activation of as GZMA, CD4, CD8A, and CD8B, related to the im- the MHC-II pathway and recognition and activation of mune system were upregulated. This induction of im- CD4+ T cells that promote the generation of memory mune genes indicates that the host immune response CD8+ T cells and other antiviral functions (Neefjes et to vaccine infection was activated. However, little is al., 2011; Pulendran and Ahmed, 2011; Swain et al., known about the mechanism of host response to vac- 2012). We are designing a systematic temporal-spatial cine by which they provide protection, despite the study, including longitudinal infection experiment, in fact that vaccination against ILT began in the 1930s which multiple samples, including bursa, will be exam- (Gibbs, 1934). Recent research involving the innate im- ined. We believe that the longitudinal experiments in mune system revealed that several subsets of T cells ILTV will give insight into the biopathways related to are involved in effective protection of the host against vaccine response. different pathogens (Pulendran and Ahmed, 2011), in- In conclusion, RNA-Seq analysis revealed the ge- dicating that different subsets of immunocytes may be nome-wide transcriptome in chicken trachea upon ILT induced following infection depending on the pathogen. CEO vaccination. Specifically, CEO vaccine activation In mice, T helper 1 cells are induced by virus through of the antigen presentation and processing pathway, in- + splenic CD8α cells and monocyte-derived dendritic cluding both the MHC-I and MHC-II pathways, pro- cells (Maldonado-López et al., 1999; Pulendran et al., vides preliminary insight into the molecular mechanism 1999). Conversely, T helper 17 cells are induced by fun- of immune response in chickens and holds potential for + gi through intestinal CD11b dendritic cells (Denning evaluation and design of new ILT vaccines in a manner et al., 2007). adapted to the host immune response to the virus. Although it is still not known which T cell subsets are involved in reacting to ILT vaccination, activation of the antigen processing and presentation pathway and REFERENCES the T cell receptor-signaling pathway is consistent with + + the increased number of tracheal CD4 and CD8 cells Bagust, T. J., R. C. Jones, and J. S. Guy. 2000. Avian infectious typically seen in ILT-vaccinated chickens (Devlin et al., laryngotracheitis. Rev. Sci. Tech. 19:483–492. MAJOR HISTOCOMPATIBILITY COMPLEX IN CHICKEN TRACHEA 855

Blacker, H. P., N. C. Kirkpatrick, A. Rubite, D. O’Rourke, and A. Moser. 1999. CD8alpha+ and CD8alpha− subclasses of dendritic H. Noormohammadi. 2011. Epidemiology of recent outbreaks of cells direct the development of distinct T helper cells in vivo. J. infectious laryngotracheitis in poultry in Australia. Aust. Vet. Exp. Med. 189:587–592. J. 89:89–94. Neefjes, J., M. L. Jongsma, P. Paul, and O. Bakke. 2011. Towards a Chandra, Y. G., J. Lee, and B. W. Kong. 2012. Genome sequence systems understanding of MHC class I and MHC class II antigen comparison of two United States live attenuated vaccines of infec- presentation. Nat. Rev. Immunol. 11:823–836. tious laryngotracheitis virus (ILTV). Virus Genes 44:470–474. Neff, C., C. Sudler, and R. K. Hoop. 2008. Characterization of west- Denning, T. L., Y. C. Wang, S. R. Patel, I. R. Williams, and B. Pu- ern European field isolates and vaccine strains of avian infectious lendran. 2007. Lamina propria macrophages and dendritic cells laryngotracheitis virus by restriction fragment length polymor- differentially induce regulatory and interleukin 17-producing T phism and sequence analysis. Avian Dis. 52:278–283. cell responses. Nat. Immunol. 8:1086–1094. Oldoni, I., and M. Garcia. 2007. Characterization of infectious laryn- Devlin, J. M., A. Viejo-Borbolla, G. F. Browning, A. H. Noormo- gotracheitis virus isolates from the US by polymerase chain reac-

hammadi, J. R. Gilkerson, A. Alcami, and C. A. Hartley. 2010. tion and restriction fragment length polymorphism of multiple Downloaded from Evaluation of immunological responses to a glycoprotein G defi- genome regions. Avian Pathol. 36:167–176. cient candidate vaccine strain of infectious laryngotracheitis vi- Pulendran, B., and R. Ahmed. 2011. Immunological mechanisms of rus. Vaccine 28:1325–1332. vaccination. Nat. Immunol. 12:509–517. Fuchs, W., J. Veits, D. Helferich, H. Granzow, J. P. Teifke, and T. C. Pulendran, B., J. L. Smith, G. Caspary, K. Brasel, D. Pettit, E. Mettenleiter. 2007. Molecular biology of avian infectious laryngo- Maraskovsky, and C. R. Maliszewski. 1999. Distinct dendritic tracheitis virus. Vet. Res. 38:261–279. cell subsets differentially regulate the class of immune response Gibbs, C. S. 1934. Infectious laryngotracheitis vaccination. Mass. in vivo. Proc. Natl. Acad. Sci. USA 96:1036–1041. http://ps.oxfordjournals.org/ Agric. Exp. Stn. Bulletin 311:1–20. Rachamadugu, R., J. Y. Lee, A. Wooming, and B. W. Kong. 2009. Guy, J. S., and M. Garcia. 2008. Laryngotracheitis. Pages 137–152 Identification and expression analysis of infectious laryngotrache- in Diseases of Poultry. Y. M. Saif, A. M. Fadly, J. R. Glisson, itis virus encoding microRNAs. Virus Genes 39:301–308. L. R. McDougald, L. K. Nolan, and D. E. Swayne, ed. Blackwell R Development Core Team. 2008. R: A language and environment Publishing, Ames, IA. for statistical computing. R Foundation for Statistical Comput- Harty, J. T., and V. P. Badovinac. 2008. Shaping and reshaping ing, Vienna, Austria. http://www.R-project.org. CD8+ T-cell memory. Nat. Rev. Immunol. 8:107–119. Ricard-Blum, S. 2011. The collagen family. Cold Spring Harb. Per- Hill, A., P. Jugovic, I. York, G. Russ, J. Bennink, J. Yewdell, H. spect. Biol. 3:a004978.

Ploegh, and D. Johnson. 1995. Herpes simplex virus turns off the Rodríguez-Avila, A., I. Oldoni, S. Riblet, and M. Garcia. 2007. Rep- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 TAP to evade host immunity. Nature 375:411–415. lication and transmission of live attenuated infectious laryngotra- Kanehisa, M., and S. Goto. 2000. KEGG: Kyoto encyclopedia of cheitis virus (ILTV) vaccines. Avian Dis. 51:905–911. genes and genomes. Nucleic Acids Res. 28:27–30. Saif, Y. M. 2003. Diseases of Poultry. Iowa State Press, Ames. Kanehisa, M., S. Goto, Y. Sato, M. Kawashima, M. Furumichi, and Spatz, S. J., J. D. Volkening, C. L. Keeler, G. F. Kutish, S. M. M. Tanabe. 2014. Data, information, knowledge and principle: Riblet, C. M. Boettger, K. F. Clark, L. Zsak, C. L. Afonso, E. Back to metabolism in KEGG. Nucleic Acids Res. 42:D199– S. Mundt, D. L. Rock, and M. Garcia. 2012. Comparative full D205. genome analysis of four infectious laryngotracheitis virus (Gallid Lee, J., W. G. Bottje, and B. W. Kong. 2012. Genome-wide host herpesvirus-1) virulent isolates from the United States. Virus responses against infectious laryngotracheitis virus vaccine infec- Genes 44:273–285. tion in chicken embryo lung cells. BMC Genomics 13:143. Swain, S. L., K. K. McKinstry, and T. M. Strutt. 2012. Expanding Lee, J. Y., J. J. Song, A. Wooming, X. Li, H. Zhou, W. G. Bottje, roles for CD4(+) T cells in immunity to viruses. Nat. Rev. Im- and B. W. Kong. 2010. Transcriptional profiling of host gene munol. 12:136–148. expression in chicken embryo lung cells infected with laryngotra- Thureen, D. R., and C. L. Keeler Jr.. 2006. Psittacid herpes virus cheitis virus. BMC Genomics 11:445. 1 and infectious laryngotracheitis virus: Comparative genome Lee, S. W., P. F. Markham, J. F. Markham, I. Petermann, A. H. sequence analysis of two avian alphaherpesviruses. J. Virol. Noormohammadi, G. F. Browning, N. P. Ficorilli, C. A. Hartley, 80:7863–7872. and J. M. Devlin. 2011. First complete genome sequence of infec- Waidner, L. A., J. Burnside, A. S. Anderson, E. L. Bernberg, M. A. tious laryngotracheitis virus. BMC Genomics 12:197. German, B. C. Meyers, P. J. Green, and R. W. Morgan. 2011. A Li, H., B. Handsaker, A. Wysoker, T. Fennell, J. Ruan, N. Homer, G. microRNA of infectious laryngotracheitis virus can downregulate Marth, G. Abecasis, and R. Durbin. 2009. The Sequence Align- and direct cleavage of ICP4 mRNA. Virology 411:25–31. ment/Map format and SAMtools. Bioinformatics 25:2078–2079. Maldonado-López, R., T. De Smedt, P. Michel, J. Godfroid, B. Pajak, C. Heirman, K. Thielemans, O. Leo, J. Urbain, and M.

Fluoroquinolone resistance and molecular characterization of gyrA and parC quinolone resistance-determining regions in Escherichia coli isolated from poultry

M. Vanni ,* V. Meucci ,* R. Tognetti ,* P. Cagnardi ,† C. Montesissa ,‡ A. Piccirillo,‡ A. M. Rossi ,§ D. Di Bello ,§ and L. Intorre *1 Downloaded from * Department of Veterinary Sciences, University of Pisa, 56122 San Piero a Grado (PI), Italy; † Department of Health, Animal Science and Food Safety, University of Milano, 20133 Milano, Italy; ‡ Department of Comparative Biomedicine and Food Science, University of Padova, 35020 Legnaro (PD), Italy; and § Department of Biology, University of Pisa, 56126 Pisa, Italy http://ps.oxfordjournals.org/ ABSTRACT Escherichia coli are a common inhabit- reported (<40%). Sixty-four isolates (27.2%) showed ant of the gastrointestinal tract of mammals and birds; full susceptibility toward the tested FQ, but 57 iso- nevertheless, they may be associated with a variety of lates (24.2%) were resistant to all tested FQ. The re- severe and invasive infections. Whereas fluoroquino- maining 114 E. coli isolates (48.5%) were grouped in 5 lones (FQ) have been banned in the United States for different resistance patterns. Isolates resistant only to use in poultry production, the use of these antimicrobi- flumequine or nalidixic acid or both possessed 1 gyrA als in poultry husbandry is still possible in the Euro- mutation, whereas isolates with further resistance to pean Union, although with some restrictions. The aim enrofloxacin, difloxacin, danofloxacin, and sarafloxacin at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 of this study was to investigate the FQ resistance of had in addition 1 or 2 parC substitutions. Two gyrA 235 E. coli isolates recovered from chickens and turkeys. mutations coupled with 1 substitution in parC were de- Minimum inhibitory concentrations were determined tected in isolates resistant to all tested FQ. The num- by a microdilution method, whereas mutations in the ber of mutations and their correlation with the in vitro quinolone resistance-determining regions of the target activity of FQ reflected the currently accepted model, genes, gyrA and parC, were detected by a PCR-based according to which a single gyrA substitution is associ- method. High resistance rates (>60%) were observed ated with resistance or decreased susceptibility to older for nalidixic acid, flumequine, and difloxacin, whereas quinolones, whereas further gyrA or parC substitutions resistance to ciprofloxacin, danofloxacin, enrofloxacin, are needed for a higher level of resistance. marbofloxacin, and sarafloxacin was less frequently Key words: Escherichia coli , poultry , antimicrobial resistance, fluoroquinolone 2014 Poultry Science 93 :856–863 http://dx.doi.org/ 10.3382/ps.2013-03627

INTRODUCTION intestinal environment has been considered as a reservoir of potentially zoonotic E. coli, with the subsequent Escherichia coli are generally considered as a com- possible contamination of poultry products with such mon inhabitant of the gastrointestinal tract of mam- bacteria during slaughter (Wasteson, 2001). mals and birds; nevertheless, strains possessing specific Treatment strategies include attempts to control virulence factors may be associated with a variety of se- predisposing infections or environmental factors and vere and invasive infections (Kaper et al., 2004). Avian early use of antibacterials indicated by susceptibility pathogenic E. coli is responsible for different local and tests. Potentiated sulfonamides are considered as first systemic diseases of poultry, such as colibacillosis, a se- choice antimicrobial agents for the treatment of avian vere systemic disease that is considered a major cause colibacillosis, whereas tetracyclines and aminopenicil- of morbidity and mortality in the poultry industry lins (ampicillin, amoxicillin) should be used as second worldwide (Ahmed et al., 2013). Moreover, the avian choice antimicrobials (Löhren et al., 2008). Fluoroqui- nolones (FQ) have been widely used in poultry since their introduction in the early 1990s (Prescott and Bag- © 2014 Poultry Science Association Inc. Received September 17, 2013. got, 1993). The advantage of oral administration, bac- Accepted December 7, 2013. tericidal activity at low tissue concentrations and high 1 Corresponding author: [email protected] potency against many gram-negative microorganisms,

856 FLUOROQUINOLONE RESISTANCE IN ESCHERICHIA COLI 857

Table 1. Fluoroquinolone (FQ) resistance rates and minimum inhibitory concentrations (MIC) of Escherichia coli isolated from poultry in Europe, United States (US), and Asian countries

Resistant FQ isolates (%) MIC range Country (source) Reference

Ciprofloxacin 75.2 nr1 China (chickens) Liu et al., 2012 73 >162 China (chickens) Yang et al., 2004 60.2 3–≥32 South Korea (chickens) Lee et al., 2005 50 nr UK (turkeys) Gosling et al., 2012 43 nr Bolivia (chickens) Riccobono et al., 2012 5 0.12–8 US (chickens) Zhao et al., 2005 nr 0.03–64 Spain (broilers) Sáenz et al., 2003 2 Danofloxacin 23.6 0.25 Japan (chickens) Ozawa et al., 2010 Downloaded from 16 0.25–16 US (chickens) Zhao et al., 2005 Difloxacin 91 >162 China (chickens) Yang et al., 2004 57 1–>16 US (chickens) Zhao et al., 2005 Enrofloxacin 90 >162 China (chickens) Yang et al., 2004 81 nr China (chickens) Liu et al., 2012 73.4 2–≥32 South Korea (chickens) Lee et al., 2005 40 0.015–8 France (turkeys) Giraud et al., 2001 http://ps.oxfordjournals.org/ 24.4 0.252 Japan (chickens) Ozawa et al., 2010 16 0.25–16 US (chickens) Zhao et al., 2005 Flumequine 48 0.5–256 France (turkeys) Giraud et al., 2001 Nalidixic acid 100 >2562 China (chickens) Yang et al., 2004 100 64–>256 US (chickens) Zhao et al., 2005 97.4 nr China (chickens) Liu et al., 2012 81 nr Bolivia (chickens) Riccobono et al., 2012 49 4–>1,024 France (turkeys) Giraud et al., 2001 nr 32–>256 Spain (broilers) Sáenz et al., 2003 Norfloxacin 60.2 16–≥32 South Korea (chickens) Lee et al., 2005 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 22.8 0.52 Japan (chickens) Ozawa et al., 2010 Sarafloxacin 100 >162 China (chickens) Yang et al., 2004 1nr = not reported. 2Mean MIC value. made FQ potentially useful in the treatment of colibac- Improved management of the use of FQ in food illosis and other infections caused by E. coli in chickens animals, particularly reducing those critically impor- and turkeys. Nevertheless, the widespread use of FQ in tant for human medicine, is an important step toward the poultry industry, and the subsequent selective pres- preserving the benefits of these antimicrobials. In 2005, sure on bacteria of animal origin, contributed to the the US Food and Drug Administration banned the acquisition of resistance toward this antimicrobial class use of FQ in chickens and turkeys because of concerns and reduction in their therapeutic efficacy (Harada and about increasing resistance in E. coli and other zoo- Asai, 2010). notic bacteria (Goetting et al., 2011). In the European A growing number of studies have reported a high Union, a reflection paper on the prudent use of FQ in level of FQ-resistant E. coli isolated from poultry, most- food producing animals was adopted in 2006 by the Eu- ly to nalidixic acid and ciprofloxacin, with minimum ropean Medicines Agency (EMEA, 2006) to critically inhibitory concentration (MIC) values ranging from 4 review its effect on the development of FQ resistance to >1,024 and from 0.003 to ≥32 μg/mL, respectively in zoonotic bacteria and the potential effect on hu- (Table 1). Moreover, isolates displaying resistance to man and animal health. On the basis of the European nalidixic acid are often co-resistant to difloxacin, enro- Union-published directives, FQ have been approved for floxacin, and norfloxacin. Nevertheless, the use of FQ, the treatment of bacterial infections in poultry. As a such as enrofloxacin, danofloxacin, and difloxacin, is of- consequence, FQ continue to be used in the poultry ten inevitable due to the increasing level of resistance to industry in Europe, thereby contributing to maintain first or second choice antimicrobials observed in recent a relatively high resistance level toward this antimicro- years (Haritova et al., 2006; Löhren et al., 2008; Furtula bial class (EMEA, 2012). et al., 2010; de Jong et al., 2012). To face these prob- The most frequent mechanism of FQ resistance in lems, international bodies and health authorities have E. coli is represented by alterations occurring in genes concerns on development of FQ resistance in human encoding for DNA gyrase and topoisomerase IV, the and animal pathogens, including E. coli. The World FQ primary and secondary target enzymes in gram- Health Organization has classified these antimicrobials negative bacteria, respectively, both playing vital roles as critically important in human medicine based on the in DNA replication and chromosome segregation (Fà- evidence that antimicrobial resistance in E. coli infec- brega et al., 2008). Alterations involve mainly chromo- tions in humans may derive from farm antibiotic use somal mutations located in the quinolone resistance- and that resistant E. coli can spread from farm animals determining regions (QRDR) of the gyrA and gyrB to humans (WHO, 2011). genes, which encode for the 2 DNA gyrase subunits, 858 VANNI ET AL. and its homologous region of the parC and parE genes, mL and kept at −20°C until needed. The tests were encoding for the 2 topoisomerase IV subunits. Other done with freshly prepared dilutions of the stock solu- mechanisms affecting the bacterial accumulation of FQ, tions. For each FQ, 16 concentrations in 2-fold dilution such as efflux pump systems activity or modifications series, ranging from 0.0078 to 256 μg/mL, were tested. of porins, as well as target protection and drug en- The inoculum was prepared by suspending bacterial zymatic modification, can contribute to enhance the colonies from an overnight sheep blood agar culture in degree of resistance in E. coli (Fàbrega et al., 2008). sterile 0.9% saline and adjusting the turbidity to a 0.5 The aim of the present study was to investigate the McFarland standard (ca. 108 cfu/mL). The suspension phenotypic resistance toward FQ and the presence of was further diluted to provide a final inoculum of 5 × QRDR mutations in the gyrA and parC genes of E. 104 cfu/well. All plates were incubated aerobically at coli isolates isolated from poultry samples from 2010 to 37°C for 16 to 18 h. Each test was performed in tripli- Downloaded from 2012 in Italy. cate. Regular quality assurance was performed by using the American Type Culture Collection reference strain MATERIALS AND METHODS E. coli ATCC 25922 (Oxoid). The MIC of each FQ was defined as the lowest concentration of antimicrobial Bacterial Isolates that completely inhibits growth of the tested isolate http://ps.oxfordjournals.org/ in the microdilution wells as detected by the unaided A total of 235 E. coli isolates recovered from fecal eye (absence of turbidity) compared with the positive samples and viscera (brain, pericardial sac, liver, lungs, control (antibiotic-free well). and joints) from chickens and turkeys were included in Minimum inhibitory concentration of FQ at which this study. The isolates were collected from 2010 to 2012 50% of the isolates were inhibited (MIC50) were also from different colibacillosis outbreaks that occurred in calculated. Isolates were further classified as resistant

80 industrial poultry farms located in a densely popu- or susceptible to the tested FQ according to breakpoint at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 lated poultry area of northern Italy. Poultry affected by proposed by the Committee for Antibiogram of the clinical disease with suspected colisepticemia (at least French Society of Microbiology for Enterobacteriaceae 3 wk of age) were euthanized and necropsy was carried of animal origin (for ENR, FLU, MAR, NAL) and hu- out immediately after to avoid postmortem contamina- man origin (CIP; CASFM, 2011); whereas for DAN, tion of the carcass. No more than 1 isolate of E. coli DIF, and SAR, breakpoints adopted by the Clinical from the same farm per year was included in the study. and Laboratory Standards Institute, formerly National The animals had not been treated with antimicrobial Committee for Clinical Laboratory Standards for bac- agents in the 3 wk before sample collection. Samples teria isolated from animals were applied (Table 2; NC- were processed for bacterial isolation on sheep blood CLS, 2003a,b; CLSI, 2008). Isolates which showed in- agar (Oxoid, Milano, Italy) and Eosin Methylene Blue termediate susceptibility to FQ were included with the agar (Oxoid) and incubated aerobically at 37°C for 24 resistant ones. h. Presumptive E. coli isolates were confirmed using the RapID E 20 kit (Biomérieux, Mercy l’Etoile, France) Amplification and DNA Sequencing of gyrA following the manufacturer’s instructions. All isolates were immediately subcultured on sheep blood agar and and parC QRDR subjected to antimicrobial susceptibility testing. Four to 5 bacterial colonies, depending on their size, of a sheep blood agar overnight culture were suspended Antimicrobial Susceptibility Testing in 900 μL of 20 mM Tris-HCl buffer (pH 7.6), incubated at room temperature for 10 min and centrifuged CIP The MIC of ciprofloxacin ( ), danofloxacin for 2 min at 9,000 × g and 4°C. Pellets were then re- DAN DIF ENR ( ), difloxacin ( ), enrofloxacin ( ), flume- suspended in 600 μL of lysis buffer (10 mM Tris-HCl, quine (FLU), marbofloxacin (MAR), nalidixic acid (NAL), and sarafloxacin (SAR) were determined by the broth microdilution method with Mueller-Hinton Table 2. Fluoroquinolone (FQ) tested and minimum inhibitory broth (Oxoid) in a 96-well plate (Steroglass, Perugia, concentrations (MIC) breakpoint Italy), following the guidelines proposed by the Com- 1 mittee for Antibiogram of the French Society of Mi- MIC breakpoint crobiology (for ENR, FLU, MAR, NAL, and CIP) and FQ R I S those proposed by the Clinical and Laboratory Stan- Ciprofloxacin >1 1 ≤0.5 dards Institute, formerly National Committee for Clini- Danofloxacin >0.25 — ≤0.25 cal Laboratory Standards (for DAN, DIF, and SAR; Difloxacin >2 1–2 ≤0.5 NCCLS, 2003a,b; CLSI, 2008; CASFM, 2011). Antimi- Enrofloxacin >2 1–2 ≤0.5 Flumequine >8 8 ≤4 crobials were purchased as powders suitable for suscep- Marbofloxacin >2 2 ≤1 tibility testing (Sigma-Aldrich, Milano, Italy) and stock Nalidixic acid >16 16 ≤8 solutions were prepared following the manufacturer’s Sarafloxacin ≥0.25 0.125 ≤0.06 instructions, divided into aliquots of approximately 1 1R = resistant; I = intermediate; S = susceptible. FLUOROQUINOLONE RESISTANCE IN ESCHERICHIA COLI 859 400 mM NaCl, 2 mM sodium EDTA, pH 7.6) containing 50 μL of 10% SDS and 6 μL of proteinase K (100 (%) μg/mL), vortexed, and incubated at 55°C. After 4 h, 4 samples were cooled, added to 3 μL of RNAase (4 mg/ rates mL), incubated at 37°C for 1 h, supplemented with 200 Resistance μL of potassium acetate, vortexed, and centrifuged for

3 min at 9,000 × g and 4°C. Finally, supernatants were 3 50 40.5 65.1 38.7 0.250.060.06 26.8 26.8 24.2 0.5 24.2 removed and DNA was purified by addition of isopro- 16 70.2 MIC ≥256 72.8 panol-alcohol and precipitation with ethanol. The DNA (μg/mL) was resuspended in 50 μL of sterile water and stored at 4°C until use. Downloaded from A 347-bp fragment (from nucleotides 108–454) con- ≥256 taining the QRDR of gyrA was amplified using primers P1F (5c-TGTCCGAGATGGCCTGAAGC-3c) and P2R

(5c-TACCGTCATAGTTATCCACG-3c; Griggs et al., 128 1996). A 964-bp fragment (from nucleotides 17–981) http://ps.oxfordjournals.org/ containing the QRDR region of parC was amplified 2 14 155 with primers ParCF43 (5c-AGCGCCTTGCGTACAT- 64 GAAT-3c) and ParCR981 (5c-GTGGTAGCGAAGAG- GTGGTT-3c; Komp Lindgren et al., 2003). Amplifica- tion was performed in a final volume of 50 μL containing 32 each primer at a concentration of 0.2 μM, 1 × PCR buf- 16 fer, 200 μM (each) deoxynucleotide phosphate, 3 mM at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 MgCl2, 2.5 U of Taq polymerase (Promega, Madison, WI), and approximately 100 ng of chromosomal DNA. 7145 60 475174 The PCR reactions were performed using an iCycler iQ 8 Real-Time PCR system (Bio-Rad, Hercules, CA) with an initial denaturing cycle at 94°C for 5 min followed 4 by 33 cycles of 94°C for 1 min, 54.5°C for 1 min (gyrA) (μg/mL) or 54°C for 1 min (parC), and 72°C for 1 min, with a 2 7 6 24 12 6 2 final extension step at 72°C for 5 min. The amplified 2 PCR products for gyrA and parC were sequenced on

both strands by using the BigDye Terminator Cycle 2143612 6 8 22 11 4 3 3 1 487 6 4 9 37 438478 Sequencing Kit v3.1 in the ABI PRISM 3100 Genetic 26 2 12 4 11 20 7 5 4 Analyzer (Applied Biosystems, Foster City, CA), ac-

cording to the manufacturer’s instructions. Editing of 1 3 4 6 18 20 3 4 2 2 chromatograms and assembly of nucleotide sequences 0.5 were performed by using the software ChromasPro v. Number of isolates with MIC Number

1.42 (Technelysium Pty Ltd., Tewantin, Australia). 1 4 9 4 24 13 2 0.25 The resulting DNA sequences were compared with the E. coli gyrA and parC genes from GenBank at the Na- tional Center for Biotechnology Information website (http://www.ncbi.nlm.nih.gov/blast). 0.125

RESULTS 0.06 FQ Resistance Phenotypes of E. coli Isolates 0.03

The MIC distributions, MIC50 values, and FQ resistance rates of the 235 E. coli isolates are shown in 0.016 Table 3. The large majority of isolates showed high resistance rates to NAL (72.8%, n = 171), FLU (70.2%, n = 165), and DIF (65.1%, n = 153). Resistance was less frequently observed to ENR (38.7%, n = 91), DAN ≤0.008 (26.8%, n = 63), and SAR (26.8%, n = 63). The lowest = at given concentration of antimicrobials, 50% of tested isolates were inhibited. 50% of tested isolates were of antimicrobials, concentration = at given resistance rates were observed for CIP (24.2%, n = 57) 50 1 and MAR (24.2%, n = 57). Sixty-four isolates (27.2%) MAR = marbofloxacin. CIP = ciprofloxacin, SAR = sarafloxacin; = danofloxacin; DAN ENR = enrofloxacin; NAL = nalidixic acid; FLU flumequine; DIF difloxacin; isolates. indicate resistant Bold numbers MIC ones. grouped with the resistant were susceptibility Isolates with intermediate 1 2 3 4 A41283074 ENRDAN44142 10 22 16 10 5 19 62 DIF 4262527 FLUDIF 1 1 7 39 12 5 5 NAL SARCIPMAR 46 57 5 12 6 39 21 13 15 93 44 1 6 43 4 10 30 5 63 21 showed full susceptibility to all tested FQ, whereas 57 isolated from poultry coli Escherichia resistance rates of 235 (MIC) distributions and fluoroquinolone (FQ) inhibitory concentration 3. Minimum Table FQ 860 VANNI ET AL. isolates (24.2%) were resistant to all tested FQ (Table infections in chicken and turkey. However, the preva- 4). The remaining E. coli isolates were grouped in 5 dif- lence and dissemination of resistance toward this class ferent resistance patterns: 62 isolates (26.4%) showed of antimicrobial agents in avian E. coli increased sig- resistance to DIF, FLU and NAL; 28 isolates (11.9%) nificantly in recent years, which is an alarming public- showed a DIF-ENR-FLU-NAL resistance pattern; 12 health concern, especially considering that FQ-resistant (5.1%) were resistant to FLU-NAL; 6 isolates (2.6%) E. coli often exhibits a multidrug-resistant phenotype showed a DAN-DIF-ENR-FLU-NAL-SAR resistance; (Ozawa et al., 2010; Gosling et al., 2012; Liu et al., and 6 isolates (2.6%) were only NAL-resistant. 2012; Riccobono et al., 2012).

Analysis of gyrA and parC Mutations Phenotypical FQ Resistance Downloaded from The amino acid changes in the QRDR of gyrA and Data from the present study are consistent with those parC and the FQ MIC50 values of E. coli isolates are of recent reports, with high resistance rates observed reported in Table 5. Isolates that were resistant only toward older quinolones, such as NAL and FLU, as well to NAL or to NAL and FLU possessed a gyrA muta- as DIF, which is currently approved for use in poul- tion at codon 83, substituting a serine with a leucine try husbandry in many European countries (Andreu http://ps.oxfordjournals.org/ (Ser83Leu). Isolates with further resistance to DIF, or et al., 2007). Lower resistance rates, similar to those to DIF and ENR, had, in addition to the gyrA Ser- recently reported, were observed for second and third 83Leu mutation, 1 or 2 parC substitutions, replacing an generation FQ DAN, ENR, and SAR (26.8%), and for alanine with a valine at codon 108 (Ala108Val) or an CIP and MAR (24.2%), whose MIC50 values ranged be- asparagine with a serine at codon 91 (Asn91Ser). Simi- tween 0.06 and 0.5 μg/mL (Gosling et al., 2012; Ozawa larly, isolates with the DAN-DIF-ENR-FLU-NAL-SAR et al., 2010; Russo et al., 2012).

resistance pattern had, in addition to 1 gyrA muta- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 tion at codon 83, 1 or 2 parC substitutions, replacing a Resistance to Older Quinolones: Single serine with an arginine (Ser80Arg) or an aspartic acid gyrA Mutation (Ser80Asp) at codon 80 or an asparagine with a serine at codon 91 (Asn91Ser). Two gyrA mutations at codons Bacterial resistance to FQ develops mainly by accu- 83 (Ser83Leu) and 87 (Asp87Asn) coupled with a serine mulation of target enzymes point mutations. Topoisom- to isoleucine substitution in parC at codon 80 (Ser80Ile) erase IV tends to be the primary target of FQ in gram- were detected in isolates resistant to all tested FQ. positive bacteria; whereas in gram-negative bacteria, including E. coli, DNA gyrase is the primary target and DISCUSSION topoisomerase IV is less sensitive to FQ action, thus representing a secondary target compared with DNA Escherichia coli isolated from poultry of different gyrase (Zhao et al., 2005; Intorre et al., 2007). Muta- geographic origins have been frequently reported to tion replacing a serine with a leucine at codon 83 (Ser- be resistant to several antimicrobial classes, includ- 83Leu) is the most frequently identified in E. coli from ing β-lactams, tetracyclines, potentiated sulfonamides, chicken and turkey with resistance to older quinolones and aminoglycosides (Vandemaele et al., 2002; Yang et (nalidixic acid, flumequine) and reduced susceptibility al., 2004; Zhao et al., 2005; Egea et al., 2012; Zhao et to second and third generation FQ, such as ciprofloxa- al., 2012). Owing to their broad spectrum of activity cin, enrofloxacin, norfloxacin, ofloxacin, and gatifloxa- against a wide range of bacteria and due to their favor- cin (Giraud et al., 2001; Sáenz et al., 2003; Yang et able pharmacokinetic profile, FQ are frequently used al., 2004; Lee et al., 2005). In our study, gyrA analysis in the treatment of colibacillosis and other bacterial among avian E. coli isolates confirmed Ser83Leu as a

Table 4. Phenotypical fluoroquinolone (FQ) resistance patterns of 235 Escherichia coli isolated from poultry

FQ resistance Number of (n) FQ resistance pattern1 isolates2 (%) 0 No resistance 64 (27.2) 1 NAL 6 (2.6) 2 FLU-NAL 12 (5.1) 3 DIF-FLU-NAL 62 (26.4) 4 DIF-ENR-FLU-NAL 28 (11.9) 6 DAN-DIF-ENR-FLU-NAL-SAR 6 (2.6) 8 CIP-DAN-DIF-ENR-FLU-MAR-NAL-SAR 57 (24.2) 1CIP = ciprofloxacin; DAN = danofloxacin; DIF = difloxacin; ENR = enrofloxacin; FLU = flumequine; MAR = marbofloxacin; NAL = nalidixic acid; SAR = sarafloxacin. 2Isolates with intermediate susceptibility were grouped with the resistant ones. FLUOROQUINOLONE RESISTANCE IN ESCHERICHIA COLI 861 hot spot mutation, being present in 37 out of 39 (95%) of the sequenced isolates with gyrA alterations. More- over, isolates showing only the gyrA Ser83Leu substitution resulted resistant to NAL and FLU but they were still susceptible to CIP, DAN, DIF, ENR, MAR, and SAR, although with an average 3.6 doubling dilution difference in MIC50 values compared with wild-type isolates.

Low Level of Resistance to FQ: 1 gyrA Downloaded from or 1 or 2 parC Mutations Resistance pattern DIF-FLU-NAL DAN-DIF-ENR-FLU-NAL-SAR SAR Once a first-step mutation has reduced the sensitivity of DNA gyrase, alteration of secondary target

topoisomerase IV can further reduce the activity of FQ, SAR extending resistance to enrofloxacin, ciprofloxacin, le- http://ps.oxfordjournals.org/ vofloxacin, and gatifloxacin. Isolates with single gyrA mutation coupled with mutations in parC have been frequently observed in poultry isolates of E. coli, mostly NAL at codon 80, resulting in amino acid substitutions of a serine with an arginine (Ser80Arg) or an isoleucine (Ser80Ile) (Giraud et al., 2001; Sáenz et al., 2003; Yang et al., 2004; Lee et al., 2005; Liu et al., 2012). As previ- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 MAR ously reported, parC mutations are usually associated with a variable level of FQ resistance. Our results are consistent with those of previous reports because E. coli isolates with 1 gyrA mutation coupled with 1 or FLU 2 mutations in parC were resistant to FLU, NAL, and DIF, as well as ENR, DAN, and SAR. A major finding of the present study was the observation that parC substitutions at codon 80 replacing a serine with an arginine (Ser80Arg) or an aspartic acid (Ser80Asp), ENR compared with substitution at codon 108 replacing an alanine with a valine (Ala108Val), were generally as-

sociated with higher MIC50 values and resistance level. DIF The Ala108Val mutation resulted in resistance only to FLU, NAL, and DIF. On the contrary, Ser80Arg or Se- r80Asp substitutions led to a marked increase in MIC50 values for DAN, DIF, ENR, FLU, MAR, and SAR isolates) (μg/mL; doubling dilution difference compared with gyrA -wild type DAN 2,3 compared with parC wild-type isolates. To the author’s 50

knowledge, the parC mutation Asn91Ser has not been MIC reported previously in avian E. coli, although the pres-

ence of this substitution in isolates that already had CIP 0.06 (3) 0.25 (5) 4 (4) 0.5 (5) 16 (5) 0.5 (5) 256 (6) 0.06 (3) one mutation in parC has not determined a higher level 0.25 (5) 2 (8) 32 (7) 4 (8) >256 (>9) 1 (6) >256 (>6) 2 (8) of resistance, both in terms of MIC50 values and in the number of FQ toward which the isolates were resistant.

High Level of Resistance to FQ: 2 gyrA 1 or 1 or 2 parC Mutations parC Asn91Ser Asn91Ser A single mutation in gyrA, coupled with one or more parC mutations, is sufficient to confer a moderate level of resistance to FQ widely used in poultry husbandry, such as nalidixic acid, danofloxacin, difloxacin, and enrofloxacin, whereas recent reports have shown that = at given concentration of antimicrobials, 50% of tested isolates were inhibited. 50% of tested isolates were of antimicrobials, concentration = at given additional alterations of DNA gyrase are necessary to Amino acid substitution(s) reach a high level of resistance. The relationship be- 50 Substituted amino acids and position number (e.g., Ser83Leu indicates substitution of a serine with leucine at position 83). (e.g., Ser83Leu indicates substitution amino acids and position number Substituted MIC acid; SAR = sarafloxacin. NAL = nalidixic FLU = flumequine; MAR marbofloxacin; ENR = enrofloxacin; DIF = difloxacin; = danofloxacin; DAN CIP = ciprofloxacin; tween a high level of FQ resistance and a second gyrA 1 2 3 Wild-typeSer83Leu Wild-type Wild-type 0.008 0.06 (3) 0.008 0.125 (4) 0.5 (1) 0.25 0.5 (5) 0.016 4 (3) 0.5 0.5 (5) 128 (5) 0.016 0.06 (3) NAL 4 0.008 No resistance isolated from poultry coli resistance patterns in Escherichia and corresponding fluoroquinolone (FQ) 5. Mutations in gyrA and parC Table gyrA Ser83Leu Wild-type 0.06 (3) 0.125 (4) 0.5 (1) 0.25 (4) 16 (5) 0.25 (4) 256 (6) 0.06 (3) FLU-NAL Ser83Leu Ala108Val 0.125 (4) 0.25 (5) 8 (5) 1 (6) 16 (5) 0.5 (5) >256 (>6) 0.06 (3) DIF-ENR-FLU-NAL Ser83Leu Ala108Val 0.06 (3) 0.25 (5) 4 (4) 0.5 (5) 16 (5) 0.5 (5) 256 (6) 0.06 (3) DIF-FLU-NAL Ser83Leu Ala108Val, Ser83LeuSer83Val Ser80Arg Ser80Asp 0.5 (6) 0.5 (6) 1 (7) 1 (7) 32 (7) 32 (7) 4 (8) 4 (8) 256 (9) 256 (9) 1 (6) 1 (6) >256 (>6) >256 (>6) 4 (9) 4 (9) DAN-DIF-ENR-FLU-NAL-SAR DAN-DIF-ENR-FLU-NAL-SAR Ser83Leu Ser80Arg, Ser83Leu, Asp87Asn Ser80Ile 4 (9) 16 (11) 256 (10) 32 (11) >256 (>9) 8 (9) >256 (>6) 16 (11) CIP-DAN-DIF-ENR-FLU-MAR-NAL- 862 VANNI ET AL. mutation at codon 87, replacing an aspartic acid with EMEA. 2006. Reflection paper on the use of fluoroquinolones in an asparagine (Asp87Asn) or a tyrosine (Asp87Tyr), food producing animals. Precautions for use in the SPC regarding prudent use guidance. European Medicines Agency. Accessed has been well characterized in E. coli isolated from September 2013. http://www.ema.europa.eu/docs/en_GB/docu- poultry (Sáenz et al., 2003; Yang et al., 2004; Zhao et ment_library/Other/2009/10/WC500005173.pdf. al., 2005; Gosling et al., 2012; Zhao et al., 2012). Ac- EMEA. 2012. Sales of veterinary antimicrobial agents in 19 EU/EEA cordingly, in the present study, 2 gyrA mutations (Ser- countries in 2010. Accessed November 2013. http://www.ema. europa.eu/docs/en_GB/document_library/Report/2012/10/ 83Leu and Asp87Asn) coupled with a Ser80Ile substitu- WC500133532.pdf. tion in parC were observed only in isolates that were Fàbrega, A., J. Sánchez-Céspedes, S. Soto, and J. Vila. 2008. Qui- fully resistant to all tested FQ, resulting in an average nolone resistance in the food chain. Int. J. Antimicrob. Agents 10 doubling dilution difference with wild-type isolates. 31:307–315.

Furtula, V., E. G. Farrell, F. Diarrassouba, H. Rempel, J. Pritchard, Downloaded from and M. S. Diarra. 2010. Veterinary pharmaceuticals and anti- Conclusions biotic resistance of Escherichia coli isolates in poultry litter from commercial farms and controlled feeding trials. Poult. Sci. 89:180–188. In conclusion, the high prevalence of FQ-resistant E. Giraud, E., S. Leroy-Sétrin, G. Flaujac, A. Cloeckaert, M. Dho-Mou- coli of poultry origin found in our study, mostly toward lin, and E. Chaslus-Dancla. 2001. Characterization of high-level nalidixic acid, flumequine, and difloxacin, stresses the fluoroquinolone resistance in Escherichia coli O78:K80 isolated http://ps.oxfordjournals.org/ importance of prudent use of these antimicrobial agents from turkeys. J. Antimicrob. Chemother. 47:341–343. Goetting, V., K. A. Lee, and L. A. Tell. 2011. Pharmacokinetics of and the need for further surveillance programs. Where- veterinary drugs in laying hens and residues in eggs: a review of as FQ have been banned in the United States in chicken the literature. J. Vet. Pharmacol. Ther. 34:521–556. and turkey, these antimicrobials are currently approved Gosling, R. J., C. S. Clouting, L. P. Randall, R. A. Horton, and R. in poultry production in European countries. Despite H. Davies. 2012. Ciprofloxacin resistance in E. coli isolated from turkeys in Great Britain. Avian Pathol. 41:83–89. control strategies based on antimicrobial restriction Griggs, D. J., K. Gensberg, and L. J. Piddock. 1996. Mutations in policies that have been adopted by several health au- gyrA gene of quinolone-resistant Salmonella serotypes isolated at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 thorities, FQ continue to be widely used in poultry pro- from humans and animals. Antimicrob. Agents Chemother. duction in European countries, possibly contributing 40:1009–1013. Harada, K., and T. Asai. 2010. Role of antimicrobial selective to a high level of resistance. Therefore, the presence of pressure and secondary factors on antimicrobial resistance FQ-resistant isolates should be routinely screened to prevalence in Escherichia coli from food-producing animals preserve the efficacy of these important antimicrobial in Japan. J. Biomed. Biotechnol. 2010:1–12. http://dx.doi. org/10.1155/2010/180682. agents. The present study confirms alterations of target Haritova, A. M., N. V. Rusenova, P. R. Parvanov, L. D. Lashev, enzymes DNA gyrase and topoisomerase IV as a deter- and J. Fink-Gremmels. 2006. Pharmacokinetic-pharmacody- mining factor for the development of FQ resistance in namic modelling of danofloxacin in turkeys. Vet. Res. Commun. E. coli. The correlation between the number and type 30:775–789. Intorre, L., M. Vanni, D. Di Bello, C. Pretti, V. Meucci, R. Tognetti, of amino acid substitutions and the in vitro activity of G. Soldani, G. Cardini, and O. Jousson. 2007. Antimicrobial sus- FQ showed that multiple gyrA and parC mutations are ceptibility and mechanism of resistance to fluoroquinolones in necessary for a high level of resistance, often reflecting Staphylococcus intermedius and Staphylococcus schleiferi. J. Vet. in different FQ resistance patterns depending on the Pharmacol. Ther. 30:464–469. Kaper, J. B., J. P. Nataro, and H. L. Mobley. 2004. Pathogenic Esch- affected codon. erichia coli. Nat. Rev. Microbiol. 2:123–140. Komp Lindgren, P., A. Karlsson, and D. Hughes. 2003. Mutation rate and evolution of fluoroquinolone resistance in Escherichia REFERENCES coli isolates from patients with urinary tract infections. Antimi- Ahmed, A. M., T. Shimamoto, and T. Shimamoto. 2013. Molecular crob. Agents Chemother. 47:3222–3232. characterization of multidrug-resistant avian pathogenic Esch- Lee, Y. J., J. K. Cho, K. S. Kim, R. B. Tak, A. R. Kim, J. W. erichia coli isolated from septicemic broilers. Int. J. Med. Micro- Kim, S. K. Im, and B. H. Kim. 2005. Fluoroquinolone resistance biol. http://dx.doi.org/10.1016/j.ijmm.2013.06.009. and gyrA and parC mutations of Escherichia coli isolated from Andreu, V., C. Blasco, and Y. Pico. 2007. Analytical strategies chicken. J. Microbiol. 43:391–397. to determine quinolone residues in food and the environment. Liu, B. T., X. P. Liao, S. S. Yang, X. M. Wang, L. L. Li, J. Sun, Trends Analyt. Chem. 26:534–556. Y. R. Yang, L. X. Fang, L. Li, D. H. Zhao, and Y. H. Liu. 2012. CASFM. 2011. Recommandations 2011. Comité de l’Antibiogramme Detection of mutations in the gyrA and parC genes in Escherich- de la Société Française de Microbiologie, Paris, France. ia coli isolates carrying plasmid-mediated quinolone resistance CLSI. 2008. Performance Standards for Antimicrobial Disk and Di- genes from diseased food-producing animals. J. Med. Microbiol. lution Susceptibility Tests for Bacteria Isolated from Animals: 61:1591–1599. Approved Standard M31–A2. Clinical and Laboratory Standards Löhren, U., A. Ricci, and T. S. Cummings. 2008. Guidelines for Institute, Wayne, PA. antimicrobial use in poultry. Pages 126–142 in Guide to Anti- de Jong, A., V. Thomas, S. Simjee, K. Godinho, B. Schiessl, U. microbial Use in Animals. L. Guardabassi, L. B. Jensen, and H. Klein, P. Butty, M. Vallé, H. Marion, and T. R. Shryock. 2012. Kruse, ed. Blackwell Publishing, Oxford, UK. Pan-European monitoring of susceptibility to human-use anti- NCCLS. 2003a. Methods for dilution antimicrobial susceptibility microbial agents in enteric bacteria isolated from healthy food- tests for bacteria that grow aerobically: Approved Standard M7– producing animals. J. Antimicrob. Chemother. 67:638–651. A6. National Committee for Clinical and Laboratory Standards, Egea, P., L. López-Cerero, E. Torres, C. Gómez-Sánchez Mdel, L. Wayne, PA. Serrano, M. D. Navarro Sánchez-Ortiz, J. Rodriguez-Baño, and NCCLS. 2003b. Performance Standards for Antimicrobial Disk and A. Pascual. 2012. Increased raw poultry meat colonization by Dilution Susceptibility Tests for Bacteria Isolated from Animals: extended spectrum beta-lactamase-producing Escherichia coli in Approved Standard M31–A2. National Committee for Clinical the south of Spain. Int. J. Food Microbiol. 159:69–73. and Laboratory Standards, Wayne, PA. FLUOROQUINOLONE RESISTANCE IN ESCHERICHIA COLI 863

Ozawa, M., K. Baba, Y. Shimizu, and T. Asai. 2010. Comparison of Wasteson, Y. 2001. Zoonotic Escherichia coli. Acta Vet. Scand. in vitro activities and pharmacokinetics/pharmacodynamics esti- Suppl. 95:79–84. mations of veterinary fluoroquinolones against avian pathogenic WHO. 2011. Critically important antimicrobials for human medicine. Escherichia coli isolates. Microb. Drug Resist. 16:327–332. World Health Organization. Accessed September 2013. http:// Prescott, J. F., and J. D. Baggot. 1993. Fluoroquinolones. Pages apps.who.int/iris/bitstream/10665/77376/1/9789241504485_ 252–262 in Antimicrobial Therapy in Veterinary Medicine. J. F. eng.pdf. Prescott, J. D. Baggot, ed. Iowa State University Press, Ames. Yang, H., S. Chen, D. G. White, S. Zhao, P. McDermott, R. Walker, Riccobono, E., L. Pallecchi, A. Mantella, F. Bartalesi, I. C. Zebal- and J. Meng. 2004. Characterization of multiple-antimicrobial-re- los, C. Trigoso, A. L. Villagran, A. Bartoloni, and G. M. Rosso- sistant Escherichia coli isolates from diseased chickens and swine lini. 2012. Carriage of antibiotic-resistant Escherichia coli among in China. J. Clin. Microbiol. 42:3483–3489. healthy children and home-raised chickens: A household study in Zhao, S., K. Blickenstaff, S. Bodeis-Jones, S. A. Gaines, E. Tong, a resource-limited setting. Microb. Drug Resist. 18:83–87. and P. F. McDermott. 2012. Comparison of the prevalences and

Russo, E., L. Lucatello, D. Giovanardi, P. Cagnardi, G. Ortali, V. antimicrobial resistances of Escherichia coli isolates from differ- Downloaded from Di Leva, and C. Montesissa. 2012. Approved medication of water ent retail meats in the United States, 2002 to 2008. Appl. Envi- with enrofloxacin to treat turkey colibacillosis: Assessment of ef- ron. Microbiol. 78:1701–1707. ficacy using a PK/PD approach. Vet. Microbiol. 161:206–212. Zhao, S., J. J. Maurer, S. Hubert, J. F. De Villena, P. F. McDermott, Sáenz, Y., M. Zarazaga, L. Briñas, F. Ruiz-Larrea, and C. Tor- J. Meng, S. Ayers, L. English, and D. G. White. 2005. Antimicro- res. 2003. Mutations in gyrA and parC genes in nalidixic acid- bial susceptibility and molecular characterization of avian patho- resistant Escherichia coli strains from food products, humans and genic Escherichia coli isolates. Vet. Microbiol. 107:215–224. animals. J. Antimicrob. Chemother. 51:1001–1005. http://ps.oxfordjournals.org/ Vandemaele, F., M. Vereecken, J. Derijcke, and B. M. Goddeeris. 2002. Incidence and antibiotic resistance of pathogenic Esche- richia coli among poultry in Belgium. Vet. Rec. 151:355–356. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Chicken NK-lysin is an alpha-helical cationic peptide that exerts its antibacterial activity through damage of bacterial cell membranes

Mi Ok Lee,*† Hyun-Jun Jang,* Jae Yong Han ,* and James E. Womack *†1

* World Class University Biomodulation Major, Department of Agricultural Biotechnology, Seoul National University, Seoul, Korea 151-742; and † Department of Veterinary Pathobiology, Texas A & M University, College Station 77843-4467 Downloaded from

ABSTRACT The antimicrobial peptides (AMP) are peptides in negatively charged liposomes demonstrat- important elements of the first line of defense against ed transition in the conformation of α-helical peptides pathogens in animals, and an important constituent of relative to the charged environment. Chicken NK-lysin innate immunity. Antimicrobial peptides act on a broad inhibits the growth of a representative gram-negative http://ps.oxfordjournals.org/ spectrum of microbial organisms. NK-Lysin is a cat- bacterium, Escherichia coli. The antimicrobial activity ionic antibacterial peptide that was originally isolated of 2 peptides designated H23 and H34 was similar to from porcine intestinal tissue based on its antibacte- that of mature NK-lysin, cNK-78. Microscopic analyses rial activity. We synthesized peptides corresponding to revealed the death of bacterium with disrupted mem- each helical region of chicken NK-lysin and analyzed branes after peptide treatment, suggesting that chicken their secondary structures in addition to their anti- NK-lysin, an alpha-helical cationic peptide, exerts its

microbial activity. Circular dichroism spectroscopy of antimicrobial activity by damaging the bacterial cell at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 the synthetic chicken NK-lysin (cNK-78) and 4 small membrane. Key words: NK-lysin, antimicrobial peptide, innate immunity, chicken 2014 Poultry Science 93 :864–870 http://dx.doi.org/ 10.3382/ps.2013-03670

INTRODUCTION influence their interactions with membranes, and the relationship of these structures to biological activities The antimicrobial peptides (AMP) are important (Jin et al., 2000; Koczulla and Bals, 2003; Mader and elements of the first line of defense in animals against Hoskin, 2006; Iwasaki et al., 2009; Gross and Andra, pathogens, and an important constituent of innate im- 2012; Lin et al., 2013; Park et al., 2013). munity. Antimicrobial peptides are an evolutionarily Although the cathelicidin and defensin families are conserved component of the innate immune response, well-studied AMP, much less is known about NK-lysin act on a wide range of microorganisms, and are widely and its human orthologue, granulysin. Antimicrobial distributed among multicellular organisms (Boman, peptides are generally small amphipathic molecules, 1994). Antimicrobial peptides act as effector molecules between 6 and 100 amino acid residues, that vary wide- of the innate immune system and play roles in cytokine ly in amino acid sequence and structure (Jenssen et al., release, cell proliferation, angiogenesis, wound healing, 2006). Compared with the other antimicrobial peptides, chemotaxis, and the induction of adaptive immunity NK-lysin is much larger (78 amino acid residues) with a (Bals and Wilson, 2003; Beisswenger and Bals, 2005). globular structure. NK-Lysin is a cationic peptide with The emergence of antibiotic-resistant bacteria and the antibacterial activity that was originally isolated from scarcity of new classes of useful antibiotics have led to porcine intestinal tissue (Andersson et al., 1995b). It a focus on natural antimicrobial agents to replace or is also present in the granules of T-lymphocytes and enhance available therapies (Joerger, 2003). Antimicro- natural killer (NK) cells. The structures of NK-lysin bial peptides may serve as drugs for the treatment of genes characterized to date are highly conserved. Their microbial infections that are resistant to antibiotics. A primary sequences are rich in positively charged amino large amount of experimental work has been conducted acids and include conserved cysteine residues that form aimed at understanding how the structures of AMP intrachain disulfide bonds (Wang et al., 2000, 2006b). NK-Lysin is a member of the saposin-like protein family, and exhibits potent antitumor activity. Extensive © 2014 Poultry Science Association Inc. research has been conducted on the structure and anti- Received October 4, 2013. Accepted November 29, 2013. microbial activities of NK-lysin isolated from different 1 Corresponding author: [email protected] species (Stenger et al., 1998; Gansert et al., 2003; Ja-

864 CHICKEN NK-LYSIN ANTIBACTERIAL ACTIVITY 865 cobs et al., 2003; Davis et al., 2005; Wang et al., 2006a; added, incubated for 5 min after which bioluminescence Andra et al., 2007). Evidence indicates that the mi- (in relative light units) was measured with a GLOMAX crobiocidal and tumor cytolytic activities of NK-lysin, multi detection system (Promega). similar to those of other saposin-like proteins, result from the ability to form pores in the cell membrane due Liposome Preparation to its α-helical structure (Zhang et al., 2000). We have identified a polymorphism in the chicken Dioleoylphosphatidylcholine (DOPC) and dio- NK-lysin gene and found that the encoded protein vari- leoylphosphatidylglycerol (DOPG) were obtained from ants are cytotoxic for bacteria and cell lines derived Avanti Polar Lipids (Alabaster, AL). The large unila- from human cancers (Lee et al., 2012). The SNP giv- mellar vesicle (LUV) solution was prepared by dissolv- Downloaded from ing rise to the variants is located in the region that ing the appropriate amounts of lipids in chloroform and encodes helix 2. Here, we report the analysis of syn- methanol (2:1) mixture. The lipid solution was evapo- thetic peptides representing the various helical regions rated under a stream of N2 gas, and further dried under to determine the contribution of each helical region of vacuum. The dried mixture was resuspended in 10 mM chicken NK-lysin to its antibacterial activity against sodium phosphate buffer (pH 7.2) and hydrated over- http://ps.oxfordjournals.org/ a representative gram-negative bacterium, Escherichia night at 4°C. The total lipid concentration was 2 mg/ coli. We also conducted fluorescence microscopy to as- mL. The solution was extruded through a polycarbon- sess the antimicrobial effects of synthetic peptides rep- ate membrane with a pore diameter of 100 nm. The resenting the helical regions of NK-lysin and used scan- resulting LUV were equilibrated overnight at 4°C. ning electron microscopy to observe the morphological effects of chicken NK-lysin on bacterial cell membranes. Circular Dichroism Spectrum Analysis

MATERIALS AND METHODS Peptide secondary structures were examined by JAS- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 CO J-715 circular dichroism (CD) spectroscopy. The Peptide Synthesis light path of the cell was 0.2 cm for wavelengths ranging Five synthetic peptides (Table 1), H12, H23, H34, from 190 to 260 nm. Spectra were baseline corrected by H345, corresponding to each helix region, and cNK-78, subtracting a blank spectrum containing only buffer. corresponding to mature intact chicken NK-lysin, were Then the spectra of peptides were measured at a con- synthesized and purified to >95% purity grade through centration of 25 μM in 10 mM sodium phosphate buffer reverse-phase high-pressure liquid chromatography by in the presence of liposomes. All data were expressed as Abclon (Seoul, Korea). Lyophilized peptides (1 mg molar ellipticity. The CD spectra results were obtained by averaging 3 scans and expressed as ellipticity in (deg each) were stored in desiccant at −20°C and dissolved 2 −1 3 in phosphate buffer (pH 7.2) before use. cm dmol ) × 10 .

Antibacterial Assay Confocal Laser-Scanning Microscopic Analysis Gram-negative bacteria, Escherichia coli ATCC25922, were purchased from Korean Collection for Type Cul- Escherichia coli cells in mid-logarithmic phase were ture and used for testing antimicrobial activity of each prepared as described in the antibacterial assay section. peptide. Overnight cultures of bacteria were subcul- Escherichia coli cells (6.5 × 106 cfu) in 0.85% NaCl were tured for an additional 2 to 3 h at 37°C to a mid- incubated with 5 μM peptide. After incubation, cells logarithmic phase and suspended to a 5 × 106 cfu/mL were washed with and stained with LIVE/DEAD Ba- in the same buffer. Bacteria suspensions (90 μL) were cLight Bacterial Viability Kits (Invitrogen, Carlsbad, placed into 96-well plates, followed by the addition of CA) according to the manufacturer’s instructions. Cells 10 μL of serial diluted peptide (final 0, 1, 5, 10 μM) in were then mounted and analyzed by a Carl Zeiss LSM triplicate. After 2 h incubation at 37°C, equal volume 700 laser-scanning confocal microscope (Carl Zeiss Inc., of BacTiter-Glo Reagent (Promega, Madison, WI) was Thornwood, NY).

Table 1. Sequence of peptides used in this study

Size (number of Peptide residues) Sequence Mw (g/mol) H12 35 KCRFCVSLVKKVQKIVGDDPDEDAINNALNKVCST 3,851.49 H23 30 PDEDAINNALNKVCSTG RQRSICKQLLKK 3,399.95 H34 27 RRQRSICKQLLKKLRQQLSDALQNNDD 3,268.76 H345 36 RRQRSICKQLLKKLRQQLSDALQNNDDPRDVCTTLG 4,211.85 cNK-78 78 GIKCRFCVSLVKKVQKIVGDDPDEDAINNALNKVCSTG 8,674.12 RRQRSICKQLLKKLRQQLSDALQNNDDPRDVCTTLGLCKG 866 LEE ET AL. Scanning Electron Microscopic Analysis α-helical structure that became readily apparent in the presence of anionic liposomes. To visualize membrane damages of E. coli, 6.5 × 106 cfu of E. coli were incubated with 5 μM H23 at 37°C for 5 min. After incubation, the bacteria were fixed with Fluorescence Microscopy of E. coli 2% glutaraldehyde, washed, mounted, and damaged Exposed to NK-Lysin membranes were observed by a Field-Emission Scan- ning Electron Microscope (Carl Zeiss Inc.). We next determined whether NK-lysin altered the morphology and viability of E. coli. The effects of NK- Statistical Analysis lysin peptide treatment on E. coli membrane were determined with confocal laser scanning microscope. We Downloaded from GraphPad Prism software (GraphPad Software Inc., exposed E. coli ATCC 25922 to peptides for 2 h before La Jolla, CA) was used for cell viability data analyses staining. As expected, untreated cells fluorescenced and data were expressed as mean ± SD. Statistical sig- green, and the cell membranes were intact (Figure 3A). nificance between groups or conditions was analyzed In contrast, most but not all cells treated with peptides by 2-way or 1-way ANOVA followed by Bonferroni’s were stained red, indicating damaged cell membranes http://ps.oxfordjournals.org/ post hoc test unless stated otherwise. Differences were and death of the bacteria. (Figure 3B–F). The H12 considered to be statistically significant when P < 0.05. treatment resulted in less cell death than treatment with the other peptides (Figure 3B). RESULTS Morphological Observation Using Scanning Antibacterial Activity of Synthetic Electron Microscopy

Peptides Derived from the Chicken at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 NK-Lysin Sequence Microscopic techniques can be applied to study the effects of AMP on biomimetic model membranes (An- To address the contribution of each helix to the andra et al., 2004; Jelinek and Kolusheva, 2005; Willu- tibacterial activity of chicken NK-lysin, we applied a meit et al., 2005; Lee et al., 2013). Here, we examined luciferase-based screening system with the synthetic the effects of chicken NK-lysin peptides on the mor- NK-lysin (cNK-78) and the 4 small peptides using the phology of E. coli treated with H23. Untreated bacteria BacTiter-Glo Microbial Cell Viability Assay kit. Brief- showed intact, uniform surfaces (Figure 4A). However, 6 ly, 1 × 10 E. coli were incubated with different con- after treatment with H23 peptide, a significant differ- centrations of peptides for 2 h (Figure 1) and analyzed. ence in morphology was observed. The scanning elec- The H23 and H34 peptides exhibited the strongest an- tron microscope image shows different states of dam- timicrobial activity against E. coli, and H345 showed aged cells affected by NK-lysin treatment; some cells slightly lesser lytic activity. The H12 was much less active. The cNK-78, mature peptide, was as active as H23 and H34 against E. coli.

Secondary Structure of NK-Lysin in the Presence of Lipids The conformations in solution of cNK-78 and the 4 small peptides was examined by CD spectroscopic analysis in both aqueous and liposome environments. In phosphate buffer, the CD spectra of peptides H12 and H23 displayed a single negative minimum at ~203 nm, suggesting a less organized, periodic conformation in solution. However, upon the addition of anionic DOPG liposomes, the spectral changes indicated a conformational transition to an α-helical structure. Of interest, addition of zwitterionic DOPC liposomes did not induce a significant transition in the conformation of the peptide (Figure 2). This might indicate that Figure 1. The antibacterial properties of chicken NK-lysin and its the peptides did not bind to the zwitterionic liposomes peptides against Escherichia coli. Cell viability was analyzed by using or that they bound weakly to induce a conformational a BacTiter-Glo Microbial Cell Viability Assay (Promega, Madison, change. The conformation of both H34 and H345 was WI). Cells were incubated with peptide concentrations of 1, 5, and 10 alpha-helical in the absence of liposomes. The CD spec- μM, and cell viability was compared with control. The average was calculated from 3 independent experiments. Each bar represents the tra of cNK-78 in aqueous solution and in the presence mean ± SD value of 6 independent experiments. ***P ≤ 0.001; **P of zwitterionic liposomes indicated the presence of an ≤ 0.005. CHICKEN NK-LYSIN ANTIBACTERIAL ACTIVITY 867 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 2. Circular dichroism (CD) spectra of chicken NK-lysin peptides in 10 mM phosphate buffer (solid line), dioleoylphosphatidylcholine (gray dotted line), and dioleoylphosphatidylglycerol liposomes (dashed line). The figure is representative of at least 3 independent experiments. show visible damage of the membrane (Figure 4B, C, DISCUSSION and D, arrow), whereas others have almost normal cell shape with partial bumpy surface (Figure 4B, C, and Antimicrobial peptides are distributed throughout D, filled triangle). The cells appeared rumpled and lost the animal and plant kingdoms. Although the best their regularly arranged surface layer and the coarse- studied families of antimicrobial peptides are the defen- ness of their surfaces. Moreover, the cells were burst sins and cathelicidins, granulysins and NK-lysin have and appeared to be crushed upon exposure to peptide received considerable attention as well. NK-Lysin and and were surrounded by debris. These findings suggest granulysin have been discovered in pigs (Andersson et that the NK-lysin destroys bacterial cells via membrane al., 1995b), horses (Davis et al., 2005), fish (Hirono et damage. al., 2007), cows (Endsley et al., 2004), chickens (Hong 868 LEE ET AL. Downloaded from http://ps.oxfordjournals.org/ Figure 4. The scanning electron microscopic observation of Esch- erichia coli treated with H23 peptide. Treated cells (B–D) showed undulating deformation and partial collapse of the cell surface as compared with untreated, intact cells (A). The figure is representative of 3 replicates.

(Andra and Leippe, 1999; Andreu et al., 1999; Wang

et al., 2000; Li et al., 2005; Linde et al., 2005; Siano et at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 al., 2010). Very recent results indicate that cNK-2, corresponding to helix 3 to part of helix 5 of chicken NK- lysin, exhibited the strongest cytotoxic activity against E. acervulina (Lee et al., 2013). Our results confirmed that the peptides corresponding to helix 2, helix 3, and helix 4 are most important for the antibacterial activity Figure 3. Membrane damage in Escherichia coli treated with of NK-lysin against bacteria. chicken NK-lysin. Escherichia coli membrane damage treated by The CD was applied to predict the secondary struc- chicken NK-lysin peptide was detected with a LIVE/DEAD BacLight Bacterial Viability Kit (Invitrogen, Carlsbad, CA). The representative tures of cNK-78 and the 4 small peptides in both aque- fluorescence images of E. coli after 2 h incubation in the absence (A) ous solution and liposome environments. NK-Lysin and presence of 5 μM peptide (B–F). Green fluorescence indicates live changes its structure in a matter of minutes after treat- bacteria with intact membranes; red fluorescence indicates dead bac- ment of anionic liposomes. The peptides H12 and H23 teria with damaged membranes. Bar = 10 μm. A, control; B, H12; C, H23; D, H34; E, H345; and F, cNK-78. showed a conformational transition into an α-helical structure only upon treatment with anionic DOPG liposomes. This result is consistent with many of the et al., 2006), and humans (Gansert et al., 2003), but α-helical peptides, which only become helical with an- were not found in mice (Endsley et al., 2004). ionic phospholipid membranes and provide the energet- In this study, we synthesized 4 peptides that con- ic source to drive the lipid interaction of amphipathic tained at least 2 of the 5 helical regions of chicken NK- α-helical AMP (Henzler Wildman et al., 2003; Mahalka lysin to examine the contribution of each helical region and Kinnunen, 2009). The peptides H34 and H345, in to synthetic, mature NK-lysin (cNK-78) structure and contrast, maintained alpha-helical conformation with- function. First, we examined antibacterial activity of out liposome incubation. The CD spectra of mature each peptide against E. coli by using the BacTiter-Glo NK-lysin, cNK-78 in aqueous solution and with zwit- Microbial Cell Viability Assay. The result revealed that terionic liposomes, shared similar α-helical structures H23 and H34 peptides showed the highest activity from that became more apparent in the presence of anionic 1 to 10 μM. We previously examined H23 antibacte- liposome. rial activity (annotated as N29N) against 4 different Cationic AMP can bind to bacterial membranes, dis- bacteria including E. coli to compare NK-lysin single rupt membrane potential, form pores in membranes, nucleotide polymorphism effects on its function. This and consequently cause the leakage of cell contents and is in good agreement with our previous results that cell death. To visualize structural changes of E. coli showed that 10% of E. coli cells were alive under a by chicken NK-lysin treatment we examined peptide concentration of 5 μM (Lee et al., 2012). Several syn- treated cells by using confocal fluorescence microscopy thetic peptide fragments derived from granulysin and and scanning electron microscopy. These studies in- NK-lysin have been investigated to determine the mini- dicate that E. coli cell membranes are disrupted by mally active portion required for antimicrobial activity peptide treatment. Moreover, a portion of the cells re- CHICKEN NK-LYSIN ANTIBACTERIAL ACTIVITY 869 mained green, indicating live bacteria with intact cell improved activity against Gram-negative pathogens. J. Biol. membranes under this condition (Figure 3). The scan- Chem. 282:14719–14728. Andreu, D., C. Carreno, C. Linde, H. G. Boman, and M. Andersson. ning electron microscopic image of this work did not 1999. Identification of an anti-mycobacterial domain in NK-lysin allow sequential processing of the NK-lysin effects on and granulysin. Biochem. J. 344:845–849. its membrane damage, but the visible difference be- Bals, R., and J. M. Wilson. 2003. Cathelicidins—A family of multi- tween E. coli cells exposed to NK-lysin and control cells functional antimicrobial peptides. Cell. Mol. Life Sci. 60:711– 720. was evident. The NK-lysin treatment destroyed cell Beisswenger, C., and R. Bals. 2005. Functions of antimicrobial pep- membranes. There are few reports on the morphologi- tides in host defense and immunity. Curr. Protein Pept. Sci. cal effects on cells resulting from treatment with AMP 6:255–264. compared with those focused on their antibacterial ac- Boman, H. G. 1994. Antimicrobial peptides. Chairman’s opening

remarks. Ciba Found. Symp. 186:1–4. Downloaded from tivities, including very recent chicken NK-lysin peptide, Davis, E. G., Y. Sang, B. Rush, G. Zhang, and F. Blecha. 2005. cNK-2 effects on E. tenella sporozoites (Jin et al., 2000; Molecular cloning and characterization of equine NK-lysin. Vet. Meincken et al., 2005; Hartmann et al., 2010; Joshi et Immunol. Immunopathol. 105:163–169. Endsley, J. J., J. L. Furrer, M. A. Endsley, M. A. McIntosh, A. C. al., 2010; Wang et al., 2012; Lee et al., 2013). Knowl- Maue, W. R. Waters, D. R. Lee, and D. M. Estes. 2004. Charac- edge about the cellular morphological change induced terization of bovine homologues of granulysin and NK-lysin. J. by AMP at the ultrastructural level will provide a basis Immunol. 173:2607–2614. http://ps.oxfordjournals.org/ for understanding the mechanisms of the antimicrobial Gansert, J. L., V. Kiessler, M. Engele, F. Wittke, M. Rollinghoff, A. M. Krensky, S. A. Porcelli, R. L. Modlin, and S. Stenger. 2003. actions of AMP. Our finding can also contribute to the Human NKT cells express granulysin and exhibit antimycobacte- exploration of new AMP therapeutic drugs for antibi- rial activity. J. Immunol. 170:3154–3161. otic-resistant bacterial infections. Gross, S., and J. Andra. 2012. Anticancer peptide NK-2 targets cell Here, we examined the contribution of chicken NK- surface sulphated glycans rather than sialic acids. Biol. Chem. 393:817–827. lysin helical regions to its structural status and anti- Hartmann, M., M. Berditsch, J. Hawecker, M. F. Ardakani, D. bacterial action. All peptides corresponding to each Gerthsen, and A. S. Ulrich. 2010. Damage of the bacterial cell at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 helical region of NK-lysin have antibacterial activity. envelope by antimicrobial peptides gramicidin S and PGLa as The H23 and H34 showed the strongest antibacterial revealed by transmission and scanning electron microscopy. An- timicrob. Agents Chemother. 54:3132–3142. effects against E. coli and H12 has weak activity com- Henzler Wildman, K. A., D. K. Lee, and A. Ramamoorthy. 2003. pared with the others. Chicken NK-lysin transits to Mechanism of lipid bilayer disruption by the human antimicro- an α-helical structure when it binds to bacterial mem- bial peptide, LL-37. Biochemistry 42:6545–6558. Hirono, I., H. Kondo, T. Koyama, N. R. Arma, J. Y. Hwang, R. branes. Bacteria cells killed by chicken NK-lysin and its Nozaki, N. Midorikawa, and T. Aoki. 2007. Characterization of peptide components showed damaged membranes. In Japanese flounder (Paralichthys olivaceus) NK-lysin, an antimi- conclusion, our findings reveal that chicken NK-lysin, crobial peptide. Fish Shellfish Immunol. 22:567–575. an α-helical cationic AMP, directly damages the bac- Hong, Y. H., H. S. Lillehoj, R. A. Dalloul, W. Min, K. B. Miska, W. Tuo, S. H. Lee, J. Y. Han, and E. P. Lillehoj. 2006. Molecular terial cell membrane and is an effective antibacterial cloning and characterization of chicken NK-lysin. Vet. Immunol. agent. Immunopathol. 110:339–347. Iwasaki, T., J. Ishibashi, H. Tanaka, M. Sato, A. Asaoka, D. Taylor, and M. Yamakawa. 2009. Selective cancer cell cytotoxicity of en- ACKNOWLEDGMENTS antiomeric 9-mer peptides derived from beetle defensins depends on negatively charged phosphatidylserine on the cell surface. This research was supported by WCU (World Class Peptides 30:660–668. University) program (R31-10056) through the National Jacobs, T., H. Bruhn, I. Gaworski, B. Fleischer, and M. Leippe. Research Foundation of Korea funded by the Ministry 2003. NK-Lysin and its shortened analog NK-2 exhibit potent activities against Trypanosoma cruzi. Antimicrob. Agents Che- of Education, Science and Technology. mother. 47:607–613. Jelinek, R., and S. Kolusheva. 2005. Membrane interactions of host- defense peptides studied in model systems. Curr. Protein Pept. REFERENCES Sci. 6:103–114. Andersson, M., H. Gunne, B. Agerberth, A. Boman, T. Bergman, R. Jenssen, H., P. Hamill, and R. E. Hancock. 2006. Peptide antimicro- Sillard, H. Jornvall, V. Mutt, B. Olsson, and H. Wigzell. 1995b. bial agents. Clin. Microbiol. Rev. 19:491–511. NK-lysin, a novel effector peptide of cytotoxic T and NK cells. Jin, Z., Y. P. Han, W. H. Zhang, F. Wang, S. Q. Zhang, and Y. M. Structure and cDNA cloning of the porcine form, induction by Mao. 2000. (Ultrastructural observation of E. coli K12 treated interleukin 2, antibacterial and antitumour activity. EMBO J. with antibacterial peptide CM4). Shi Yan Sheng Wu Xue Bao 14:1615–1625. 33:89–95. Andra, J., M. H. Koch, R. Bartels, and K. Brandenburg. 2004. Bio- Joerger, R. D. 2003. Alternatives to antibiotics: Bacteriocins, anti- physical characterization of endotoxin inactivation by NK-2, an microbial peptides and bacteriophages. Poult. Sci. 82:640–647. antimicrobial peptide derived from mammalian NK-lysin. Anti- Joshi, S., G. S. Bisht, D. S. Rawat, A. Kumar, R. Kumar, S. Maiti, microb. Agents Chemother. 48:1593–1599. and S. Pasha. 2010. Interaction studies of novel cell selective Andra, J., and M. Leippe. 1999. Candidacidal activity of shortened antimicrobial peptides with model membranes and E. coli ATCC synthetic analogs of amoebapores and NK-lysin. Med. Microbiol. 11775. Biochim. Biophys. Acta 1798:1864–1875. Immunol. (Berl.) 188:117–124. Koczulla, A. R., and R. Bals. 2003. Antimicrobial peptides: Current Andra, J., D. Monreal, G. Martinez de Tejada, C. Olak, G. Brezesin- status and therapeutic potential. Drugs 63:389–406. ski, S. S. Gomez, T. Goldmann, R. Bartels, K. Brandenburg, and Lee, M. O., E. H. Kim, H. J. Jang, M. N. Park, H. J. Woo, J. Y. I. Moriyon. 2007. Rationale for the design of shortened deriva- Han, and J. E. Womack. 2012. Effects of a single nucleotide poly- tives of the NK-lysin-derived antimicrobial peptide NK-2 with morphism in the chicken NK-lysin gene on antimicrobial activity 870 LEE ET AL.

and cytotoxicity of cancer cells. Proc. Natl. Acad. Sci. USA Siano, A., G. Tonarelli, M. S. Imaz, J. C. Perin, N. Ruggeri, M. 109:12087–12092. Lopez, M. N. Santi, and E. Zerbini. 2010. Bactericidal and he- Lee, S. H., H. S. Lillehoj, W. Tuo, C. A. Murphy, Y. H. Hong, and molytic activities of synthetic peptides derived from granulysin. E. P. Lillehoj. 2013. Parasiticidal activity of a novel synthetic Protein Pept. Lett. 17:517–521. peptide from the core α-helical region of NK-lysin. Vet. Parasi- Stenger, S., D. A. Hanson, R. Teitelbaum, P. Dewan, K. R. Niazi, C. tol. 197:113–121. J. Froelich, T. Ganz, S. Thoma-Uszynski, A. Melian, C. Bogdan, Li, Q., C. Dong, A. Deng, M. Katsumata, A. Nakadai, T. Kawada, S. A. Porcelli, B. R. Bloom, A. M. Krensky, and R. L. Modlin. S. Okada, C. Clayberger, and A. M. Krensky. 2005. Hemolysis of 1998. An antimicrobial activity of cytolytic T cells mediated by erythrocytes by granulysin-derived peptides but not by granuly- granulysin. Science 282:121–125. sin. Antimicrob. Agents Chemother. 49:388–397. Wang, Q., B. Bao, Y. Wang, E. Peatman, and Z. Liu. 2006a. Char- Lin, M. C., C. F. Hui, J. Y. Chen, and J. L. Wu. 2013. Truncated acterization of a NK-lysin antimicrobial peptide gene from chan- antimicrobial peptides from marine organisms retain anticancer nel catfish. Fish Shellfish Immunol. 20:419–426.

activity and antibacterial activity against multidrug-resistant Wang, Q., Y. Wang, P. Xu, and Z. Liu. 2006b. NK-Lysin of chan- Downloaded from Staphylococcus aureus. Peptides 44:139–148. nel catfish: Gene triplication, sequence variation, and expression Linde, C. M., S. Grundstrom, E. Nordling, E. Refai, P. J. Bren- analysis. Mol. Immunol. 43:1676–1686. nan, and M. Andersson. 2005. Conserved structure and function Wang, W., R. Tao, Z. Tong, Y. Ding, R. Kuang, S. Zhai, J. Liu, and in the granulysin and NK-lysin peptide family. Infect. Immun. L. Ni. 2012. Effect of a novel antimicrobial peptide chrysophsin-1 73:6332–6339. on oral pathogens and Streptococcus mutans biofilms. Peptides Mader, J. S., and D. W. Hoskin. 2006. Cationic antimicrobial pep- 33:212–219. tides as novel cytotoxic agents for cancer treatment. Expert Wang, Z., E. Choice, A. Kaspar, D. Hanson, S. Okada, S. C. Lyu, A. http://ps.oxfordjournals.org/ Opin. Investig. Drugs 15:933–946. M. Krensky, and C. Clayberger. 2000. Bactericidal and tumori- Mahalka, A. K., and P. K. Kinnunen. 2009. Binding of amphipathic cidal activities of synthetic peptides derived from granulysin. J. alpha-helical antimicrobial peptides to lipid membranes: Lessons Immunol. 165:1486–1490. from temporins B and L. Biochim. Biophys. Acta 1788:1600– Willumeit, R., M. Kumpugdee, S. S. Funari, K. Lohner, B. P. Navas, 1609. K. Brandenburg, S. Linser, and J. Andra. 2005. Structural rear- Meincken, M., D. L. Holroyd, and M. Rautenbach. 2005. Atomic rangement of model membranes by the peptide antibiotic NK-2. force microscopy study of the effect of antimicrobial peptides on Biochim. Biophys. Acta 1669:125–134. the cell envelope of Escherichia coli. Antimicrob. Agents Che- Zhang, G., C. R. Ross, and F. Blecha. 2000. Porcine antimicrobial

mother. 49:4085–4092. peptides: New prospects for ancient molecules of host defense. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Park, H. J., K. M. Kang, K. Dybvig, B. L. Lee, Y. W. Jung, and I. Vet. Res. 31:277–296. H. Lee. 2013. Interaction of cationic antimicrobial peptides with Mycoplasma pulmonis. FEBS Lett. 5793:00630–00633.

Contribution of Salmonella Enteritidis virulence factors to intestinal colonization and systemic dissemination in 1-day-old chickens

Tarek M. Addwebi ,* Douglas R. Call ,† and Devendra H. Shah *†1

* Department of Veterinary Microbiology and Pathology, and † Paul Allen School for Global Animal Health, Washington State University, Pullman 99164-7040 Downloaded from ABSTRACT Salmonella enterica serovar Enteritidis is factors and Salmonella Enteritidis-specific traits to es- one of the most common serovars associated with poul- tablish infection in chickens. Four Salmonella Enteriti- try and poultry product contamination in the United dis mutants (SEN0034::Tn5, fliH::Tn5, SEN1393::Tn5, States. We previously identified 14 mutant strains of and spvR::Tn5) were indistinguishable from the isogenic Salmonella Enteritidis phage type 4 (PT4) with signifi- wild-type strain when orally inoculated in 1-d-old chick- http://ps.oxfordjournals.org/ cantly reduced invasiveness in human intestinal epithe- ens, whereas 2 mutants (CsgB::Tn5 and PegD::Tn5) lial cells (Caco-2), chicken macrophages (HD-11), and were defective for intestinal colonization (P < 0.05) and chicken hepatocellular epithelial cells (LMH). These in- 8 mutants (hilA::Tn5, SEN3503::Tn5, SEN0803::Tn5, cluded Salmonella Enteritidis mutants with transposon SEN2278::Tn5, fljB::Tn5, rfbM::Tn5, rfbN::Tn5, and insertions in 6 newly identified Salmonella Enteritidis- pipA::Tn5) showed significant in vivo attenuation in specific genes (pegD and SEN1393), and genes or ge- more than one organ (P < 0.05). Complementation

nomic islands common to most other Salmonella se- studies confirmed the role of rfbN and SEN3503 during at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 rovars (SEN0803, SEN0034, SEN2278, and SEN3503) infection. This study should contribute to a better un- along with 8 genes previously known to contribute to derstanding of the mechanisms involved in Salmonella enteric infection (hilA, pipA, fliH, fljB, csgB, spvR, and Enteritidis pathogenesis, and the target genes identified rfbMN). We hypothesized that Salmonella Enteritidis here could potentially serve as targets for the develop- employs both common Salmonella enterica colonization ment of live-attenuated or subunit vaccine.

Key words: Salmonella Enteritidis, chicken , virulence 2014 Poultry Science 93 :871–881 http://dx.doi.org/ 10.3382/ps.2013-03710

INTRODUCTION 2010 because of Salmonella Enteritidis contamination (Kuehn, 2010). Contaminated poultry meat is another Salmonella enterica causes more than 1 million food- important source of Salmonella Enteritidis foodborne borne illnesses each year in the United States (Scallan infections (Altekruse et al., 2006). Of the 64 poultry et al., 2011; Jackson et al., 2013). Among all foodborne meat-associated outbreaks, Salmonella Enteritidis was outbreaks caused by Salmonella, the incidence has de- the most common serotype accounting for a total of creased for some serotypes but has remained an im- 28% outbreaks (Ong et al., 2010; Jackson et al., 2013). portant concern for several others. The latter pattern Upon oral infection, Salmonella Enteritidis primarily is typical for Salmonella Enteritidis where the major- colonizes the ceca (Popiel and Turnbull, 1985; Desmidt ity of outbreaks have been associated with consump- et al., 1997). Subsequent invasion of the intestinal epi- tion of contaminated poultry products (Patrick et al., thelium triggers infiltration of phagocytic cells to the 2004; Ong et al., 2010; Scallan et al., 2011; Silva et infected site where Salmonella Enteritidis are engulfed al., 2012; Jackson et al., 2013). Epidemiological studies by macrophages and heterophils (Lu et al., 1999; Van show that eggs are the major vehicle of transmission of Immerseel et al., 2002). The infected phagocytes then Salmonella Enteritidis. With approximately 65 billion disseminate Salmonella Enteritidis to the internal or- eggs produced every year in the United States, more gans including the liver, spleen, and reproductive tis- than 500 million eggs were recalled from the market in sues (Turnbull and Snoeyenbos, 1974). Colonization of the reproductive organs leads to contamination of internal contents of eggs (Gast and Beard, 1990a,b; © 2014 Poultry Science Association Inc. Gast et al., 2007). Salmonella Enteritidis may use dif- Received October 24, 2013. Accepted December 19, 2013. ferent genetic factors to complete different steps during 1 Corresponding author: [email protected] chicken infection including intestinal colonization and

871 872 ADDWEBI ET AL. systemic dissemination. Identification of these factors is MATERIALS AND METHODS essential to understand pathogenesis of Salmonella En- teritidis in chickens and to develop strategies to control Bacterial Strains this serotype in poultry. R Previous work using in vivo expression technology Salmonella Enteritidis G1 Nal used in this study and transposon-mediated mutagenesis have identified is a phage type 4 strain that was stored at −80°C in genes that contribute to avian reproductive tract colo- 15% (vol/vol) phosphate-buffered glycerol (Table 1). nization and survival of Salmonella Enteritidis within This strain is invasive in cultured intestinal epithelial egg albumen, respectively (Clavijo et al., 2006; Gan- cells and virulent in orally inoculated mice and chickens tois et al., 2008). Although intestinal colonization and (Shah et al., 2011, 2012a,b). Salmonella Enteritidis mu- systemic dissemination are primary mechanisms of tants were constructed as previously described (Shah Downloaded from Salmonella Enteritidis infection cycle in chickens, the et al., 2012b; Table 1). All isolates were confirmed as underlying genetic mechanisms are poorly understood. Salmonella Enteritidis using a plate agglutination test Recently, our laboratory employed a transposon muta- with Salmonella O antiserum Group D1 (Difco, Sparks, genesis approach to identify several genes that contrib- MD) according to the manufacturer’s instructions ute to invasion, survival, or both within human intesti- and also by serotype specific PCR targeting sdfI gene http://ps.oxfordjournals.org/ nal cells (Caco-2 cells), chicken macrophages (HD-11), (Agron et al., 2001; Alvarez et al., 2004; Clavijo et al., and hepatocellular epithelial (LMH) cells (Shah et 2006). All isolates were grown in Luria Bertani broth al., 2012b). Several invasion-attenuated mutants were (Difco) at 37°C for 16 h with shaking at 200 rpm. When identified with transposon insertions in genes encod- required, antibiotics were added at the following concentrations: nalidixic acid (Nal, 30 μg·mL−1), kanamy- ing fimbriae, motility, LPS biosynthesis, carbohydrate −1 −1 metabolism, DNA recombination and replication, Sal- cin (50 μg·mL ), and chloramphenicol (30 μg·mL ).

monella pathogenicity islands (SPI), and genes encod- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 ing hypothetical proteins. Genes previously known to Chicken Experiment contribute to enteric infection (e.g., SPI-1, SPI-5, csgB, fliH, fljB, spvR, and rfbMN) were identified, and several Specific-pathogen-free eggs were obtained from Sun- new genes or genomic islands common to most other rise Farm (New York, NY), incubated for 18 d in the Salmonella serovars (e.g., SPI-14, SEN0034, SEN2278, egg incubator (Ova-Easy 190 Advance Cabinet Incuba- and SEN3503) and additional Salmonella Enteritidis- tor, Brinsea Products Inc, Titusville, FL), and subse- specific genes (e.g., pegD and SEN1393) were also iden- quently transferred to the hatcher (1550 hatcher-GQF, tified. The objective of this study was to determine the GQF manufacturer Co., Savannah, GA) for 3 d fol- contribution of these invasion factors identified using lowing the manufacturer’s instructions. Newly hatched in vitro cultured cells to the virulence of Salmonella chicks were transferred to the chicken facility, randomly Enteritidis in chickens with specific emphasis on the assigned to different treatment groups of 16 chickens intestinal colonization and systemic dissemination. per group, and housed in HEPA-filtered isolation cages.

Table 1. Salmonella Enteritidis strains used in this study

Salmonella Enteritidis strain (laboratory number1) Function2

G1-NalR strain Wild-type strain Mutant strain SEN0034::Tn5 (2A5) Putative hypothetical protein fliH::Tn5 (1F8) Flagellar assembly protein H (motility) hilA::Tn5 (1B12) SPI-1, T3SS regulatory protein SEN3503::Tn5 (2D6) Putative sugar kinase (carbohydrate metabolism) SEN0803::Tn5 (2B4) SPI-14, putative acyl-CoA dehydrogenase SEN2278::Tn5 (1E1) Putative hypothetical protein (ais protein) SEN1393::Tn5 (2F10) ϕSE14 bacteriophage protein (DNA recombination) fljB::Tn5 (1E9) Flagellin (motility) csgB::Tn5 (1H3) Curli fimbriae minor subunit (fimbriae) pegD::Tn5 (1E5) Putative fimbrial-like adhesion protein (fimbriae) rfbM::Tn5 (2B5) Mannose-1 phosphate guanylytransferase (LPS biosynthesis) rfbN::Tn5 (1D10) Rhamnosyl transferase (LPS biosynthesis) spvR::Tn5 (1A9) Large plasmid-associated virulence regulatory protein pipA::Tn5 (2E3) SPI-5, pathogenicity island protein Complemented strain rfbN::Tn5-pACYC-rfbN (1D10) Plasmid pACYC184 carrying rfbN gene transformed into rfbN::Tn5 rfbM::Tn5-pACYC-rfbM (2B5) Plasmid pACYC184 carrying rfbM gene transformed into rfbM::Tn5 SEN0803::Tn5-pACYC-SEN0803 (2B4) Plasmid pACYC184 carrying SEN0803 gene transformed into SEN3503::Tn5 SEN3503::Tn5-pACYC-SEN3503 (2D6) Plasmid pACYC184 carrying SEN3503 gene transformed into SEN0803::Tn5 1Shows names of strains reported by Shah et al. (2012b). 2SPI = Salmonella pathogenicity islands; LPS = lipopolysaccharide. SALMONELLA ENTERITIDIS VIRULENCE FACTORS 873 Temperature and humidity inside the cages were moni- Statistical Analysis tored using hygrometers during the course of experiment. Antibiotic-free feed (Purina, St. Louis, MO) and Recovered bacterial counts were log10 transformed water were supplied ad libitum until the end of the ex- and analyzed using multivariate ANOVA followed by periment. To test for Salmonella at the beginning of the multiple comparisons using Tukey’s method to deter- experiment, cloacal swabs were taken from 10 chickens mine significant differences between treated groups. per group and placed in 1 mL of maximum recovery Differences with P-values less than 0.05 were consid- diluent (MRD, 1.0 g/L of peptone, 8.5 g/L of NaCl). ered significant. Comparison of bacterial counts from Cloacal samples were enriched by addition of 0.5 mL of spleen and liver was carried out using Pearson corre- MRD to 10 mL of tetrathionate broth (Difco). After in- lation test. Statistical analysis was carried out using cubation overnight at 37°C, 100 μL of resulting culture SPSS 19.0 software (IBM SPSS Statistics, Chicago, IL). Downloaded from was plated onto xylose-lysine deoxycholate (XLD) agar (HiMedia Laboratories, Mumbai, India) and incubated RESULTS AND DISCUSSION overnight at 37°C. To infect chickens bacterial isolates were grown overnight in Luria Bertani medium supple- The aim of this study was to identify Salmonella En- mented with appropriate concentrations of antibiotics teritidis genes that contribute to the intestinal coloni- http://ps.oxfordjournals.org/ at 37°C with shaking at 200 rpm for 20 h. Overnight zation and systemic dissemination in poultry. A total of cultures were serially diluted in MRD to obtain cell 14 Salmonella Enteritidis mutants carrying transposon counts. Chicks in the positive control group were orally insertions in 8 categorically different genes were tested. challenged at the age of 1 d with 1 × 109 to 2.5 × 109 These included genes related to SPI (hilA, pipA, and cfu of wild-type Salmonella Enteritidis G1-NalR in 200 SEN0803), LPS biosynthesis pathway (rfbMN), carbo- μL of MRD. Chicks in the negative control group were hydrate metabolism (SEN3503), DNA recombination

given 200 μL of MRD. Chicks in other groups were and replication (SEN1393), fimbriae (csgB and pegD), at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 orally challenged with 1 × 109 to 2.5 × 109 cfu of each motility (fliH and fljB), virulence plasmid (spvR), and Salmonella Enteritidis mutant. For bacteriological anal- genes encoding hypothetical proteins (SEN0034, and ysis, 3 chickens were euthanized from each group on d SEN2278). All of these transposon insertion mutants 1, 2, 4, and 8 postinfection (PI) and 4 chickens were had insertions in single gene (Shah et al., 2012b). Six out euthanized on d 12 PI. Internal organs including the of 14 Salmonella Enteritidis mutants (SEN0034::Tn5, small intestine, cecum, liver, and spleen from eutha- fliH::Tn5, SEN1393::Tn5, csgB::Tn5, pegD::Tn5, and nized chickens were aseptically collected in sterile bags, spvR::Tn5) were not attenuated in vivo or were only at- weighed, and homogenized in the stomacher (Seward tenuated in their ability to colonize the small intestine. Brinkmann Stomacher 80, USA) with MRD. The num- The remaining 8 mutants (hilA::Tn5, SEN3503::Tn5, ber of bacteria per gram in these samples was obtained SEN0803::Tn5, SEN2278::Tn5, fljB::Tn5, rfbM::Tn5, by direct plating of 10-fold dilutions onto XLD agar rfbN::Tn5, and pipA::Tn5) were in vivo attenuated in plates supplemented with appropriate concentrations their ability to colonize more than one organ (Table of antibiotics. Samples that were negative after direct 2). Overall, there was a significant correlation between plating were enriched in tetrathionate broth and incu- bacterial counts of virulence-attenuated mutants in bated at 37°C for 24 h followed by plating onto XLD spleen and liver (r = 0.81, P < 0.0001) consistent with agar. Challenge experiments were performed following reports from other studies (Bohez et al., 2006, 2008; the protocol approved by the Washington State Uni- Lawley et al., 2006). versity Institutional Animal Care and Use Committee. Flagellar Genes Complementation Study To explore the contribution of flagella to the viru- Genetic complementation was used to verify in vivo lence of Salmonella Enteritidis in poultry, 2 Salmonella virulence-attenuation of selected mutants. Comple- Enteritidis mutants with transposon insertions in fliH mented strains were generated as previously described (encoding a regulatory protein FliH) and fljB (encoding (Shah et al., 2012b; Table 1). Briefly, full-length genes flagellin subunit protein FljB) were tested in this study. were amplified using forward and reverse primers de- FliH binds the ATPase FliI and prevents it from wast- signed from the Salmonella Enteritidis P125109 se- ing ATP when the energy is not required for flagellar quence. Amplified genes were cloned into the low copy- biogenesis or assembly (Minamino and MacNab, 2000; number plasmid pACYC184 (New England Biolabs, Macnab, 2004; Apel and Surette, 2008). Salmonella USA) by insertion into the tetracycline resistance gene Enteritidis expresses 2 flagellin subunit proteins called of the plasmid using appropriate restriction enzymes, FliC and FljB, which make up the filament on the sur- allowing in trans constitutive expression from the pro- face of Salmonella Enteritidis, and these proteins are moter of the tetracycline resistance gene. The chicken synthesized asynchronously (Ikeda et al., 2001). The experiment using the complemented strains was carried fljBA promoter, located directly upstream of fljBA, in- out as described in the previous section. duces transcription of fljB and fljA. FljA is a transcrip- 874 ADDWEBI ET AL.

Table 2. Mean (log10) ± SE of the number of bacteria recovered from the internal organs of orally inoculated chickens over a 12-d experimental period1

Mean log10 cfu per gram of organs ± SEM Time Strain (DPI2) Small intestine Cecum Liver Spleen Wild-type strain 1 7.37 ± 0.25 9.35 ± 0.04 6.69 ± 0.13 6.47 ± 0.30 2 7.48 ± 0.16 9.32 ± 0.08 6.37 ± 0.39 6.47 ± 0.35 4 7 ± 0.13 9.1 ± 0.04 3.69 ± 0.19 3.99 ± 0.12 8 6.73 ± 0.07 8.75 ± 0.07 3.24 ± 0.7 3.74 ± 0.11 12 6.68 ± 0.47 8.25 ± 0.15 2.78 ± 0.24 3.54 ± 0.09 SEN0034::Tn5 1 7.34 ± 0.02 9.19 ± 0.02 6.94 ± 0.11 6.92 ± 0.06 2 5.21 ± 0.52** 9.30 ± 0.00 6.69 ± 0.021 6.88 ± 0.02 Downloaded from 4 7.13 ± 0.18 9.01 ± 0.10 2.99 ± 0.05 3.81 ± 1.10 8 6.97 ± 0.15 8.58 ± 0.04 3.28 ± 0.01 3.56 ± 0.15 12 5.89 ± 0.20 8.04 ± 0.05 2.81 ± 0.04 2.89 ± 0.25 fliH::Tn5 1 7.46 ± 0.03 9.21 ± 0.00 6.93 ± 0.11 6.92 ± 0.05 2 7.0 ± 0.16 9.29 ± 0.02 7.17 ± 0.04 6.97 ± 0.02

4 7.15 ± 0.06 8.90 ± 0.03 3.19 ± 0.06 4.00 ± 0.05 http://ps.oxfordjournals.org/ 8 6.8 ± 0.21 8.31 ± 0.11 3.25 ± 0.07 3.59 ± 0.10 12 5.8 ± 0.16 7.99 ± 0.07 2.84 ± 0.05 3.18 ± 0.02 SEN1393::Tn5 1 6.49 ± 0.20 9.28 ± 0.09 6.80 ± 0.03 6.16 ± 0.36 2 7.47 ± 0.06 9.02 ± 0.10 3.64 ± 1.99 5.69 ± 0.56 4 7.3 ± 0.16 8.58 ± 0.11 4.04 ± 0.57 3.40 ± 0.10 8 6.17 ± 0.20 8.65 ± 0.03 3.08 ± 0.04 4.12 ± 0.17 12 5.22 ± 0.05 7.29 ± 0.17 1.37 ± 0.05 3.57 ± 0.11 CsgB::Tn5 1 8.13 ± 0.029 9.36 ± 0.08 5.55 ± 0.24 5.69 ± 0.51 2 6.80 ± 0.22** 8.72 ± 0.32 5.37 ± 0.68* 6.3 ± 0.36

4 6.35 ± 0.09 9.10 ± 0.03 3.71 ± 0.37 4.33 ± 0.01 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 8 5.41 ± 0.21** 8.63 ± 0.03 3.06 ± 0.12 3.54 ± 0.03 12 4.56 ± 0.27* 8.24 ± 0.14 2.20 ± 0.03 3.75 ± 0.08 PegD::Tn5 1 7.29 ± 0.16 9.49 ± 0.14 5.26 ± 0.14 5.66 ± 0.20 2 7.29 ± 0.30* 8.72 ± 0.25 5.56 ± 0.43 6.09 ± 0.50 4 6.63 ± 0.25 9.16 ± 0.02 3.60 ± 0.01 4.32 ± 0.01 8 4.19 ± 0.20** 8.44 ± 0.11 2.67 ± 0.01 3.53 ± 0.03 12 5.51 ± 0.76 8.70 ± 0.19 2.23 ± 0.10 3.71 ± 0.16 spvR::Tn5 1 7.44 ± 0.03 9.20 ± 0.01 7.05 ± 0.07 6.94 ± 0.04 2 6 ± 0.01 9.28 ± 0.06 6.59 ± 0.04 6.77 ± 0.02 4 7.20 ± 0.16 9.51 ± 0.68 3.21 ± 0.01 4.00 ± 0.08 8 6.42 ± 0.02 8.50 ± 0.07 3.19 ± 0.08 3.59 ± 0.09 12 5.29 ± 0.07 7.81 ± 0.02 2.72 ± 0.02 3.17 ± 0.01 hilA::Tn5 1 7.19 ± 0.01 9.01 ± 0.01 2.15 ± 0.73** 0** 2 6.47 ± 0.47 8.37 ± 0.10* 6.77 ± 0.17 6.86 ± 0.02 4 5.66 ± 0.18** 8.78 ± 0.45 1.02 ± 0.71** 0.23 ± 0** 8 3.11 ± 0.06** 7.92 ± 0.03** 0.23 ± 0** 0.23 ± 0** 12 1.66 ± 0.07** 6.52 ± 0.08** 0.69 ± 0** 0.69 ± 0** SEN3503::Tn5 1 6.78 ± 0.40 9.44 ± 0.02 1.34 ± 0.87** 2.35 ± 0.41** 2 4.79 ± 0.20* 8.97 ± 0.14 0.65 ± 0.36** 1.13 ± 0.8** 4 3.53 ± 0.01* 7.56 ± 0.5 0.65 ± 0.32** 0.94 ± 0.94* 8 4.89 ± 0.06 8.60 ± 0.20 1.09 ± 0.21** 3.03 ± 0.04* 12 3.52 ± 0.06** 7.94 ± 0.08 1.00 ± 0.06* 2.50 ± 0.4* SEN0803::Tn5 1 6.51 ± 0.11 9.23 ± 0.03 1.40 ± 0.37** 0 ± 0** 2 6.85 ± 0.24 8.46 ± 0.15 2.20 ± 0.27 ** 2.23 ± 0.9** 4 4.20 ± 0.51* 8.35 ± 0.09 1.99 ± 0.14** 3.17 ± 0.26 8 4.19 ± 0.31* 8.13 ± 0.16 1.59 ± 0.03** 2.98 ± 0.05* 12 3.57 ± 0.06* 7.72 ± 0.07 0.94 ± 0.009* 2.87 ± 0.20 SEN2278::Tn5 1 6.92 ± 0.29 9.31 ± 0.04 0 ± 0** 0 ± 0** 2 6.21 ± 0.06 8.66 ± 0.11 0.23 ± 0.23** 0 ± 0** 4 3.70 ± 0.15* 7.99 ± 0.09 0.96 ± 0.49** 2.47 ± 0.50* 8 5.54 ± 0.20 8.13 ± 0.23 1.30 ± 0.32** 2.53 ± 0.29* 12 2.46 ± 0.31** 8.06 ± 0.13 0.80 ± 0.48* 1.56 ± 0.90* fljB::Tn5 1 6.75 ± 0.37 9.01 ± 0.14 2.32 ± 0.34** 2.21 ± 0.90** 2 5.75 ± 0.19* 8.89 ± 0.15 0.29 ± 0.29** 0 ± 0** 4 5.26 ± 0.46* 8.32 ± 0.32 0.65 ± 0.36** 2.05 ± 0.39* 8 5.33 ± 0.26* 8.45 ± 0.26 0.56 ± 0.29** 2.24 ± 0.21** 12 2.84 ± 0.32* 7.78 ± 0.24 0.74 ± 0.32* 0 ± 0** rfbM::Tn5 1 7.39 ± 0.23 8.79 ± 0.28* 0 ± 0** 1.46 ± 1.13** 2 5.91 ± 0.13** 8.49 ± 0.19* 1.35 ± 0.66** 2 ± 0** 4 5.93 ± 0.23* 8.27 ± 0.14** 1.98 ± 0.07** 4.10 ± 0.15 8 5.31 ± 0.30** 8.32 ± 0.1 2.89 ± 0.04 3.03 ± 0.23 12 6.16 ± 0.24 8.19 ± 0.13* 2.30 ± 0.01 3.62 ± 0.04 rfbN::Tn5 1 6.49 ± 0.10* 8.50 ± 0.25* 0.40 ± 0.40** 0 ± 0* 2 5.76 ± 0.33* 9.02 ± 0.09 1.12 ± 0.62** 1.23 ± 1.23** 4 5.72 ± 0.46* 8.41 ± 0.10** 3.29 ± 0.10 3.82 ± 0.24 8 3.07 ± 0.23** 7.98 ± 0.22** 0.46 ± 0.23** 1.33 ± 0.66** 12 3.82 ± 0.29** 7.29 ± 0.26** 0.345 ± 0.199** 0.57 ± 0.57** Continued SALMONELLA ENTERITIDIS VIRULENCE FACTORS 875

Table 2 (Continued). Mean (log10) ± SE of the number of bacteria recovered from the internal organs of orally inoculated chickens over a 12-d experimental period1

Mean log10 cfu per gram of organs ± SEM Time Strain (DPI2) Small intestine Cecum Liver Spleen pipA::Tn5 1 6.47 ± 0.24* 8.88 ± 0.08** 0 ± 0** 0.86 ± 0.86** 2 5.31 ± 0.95** 8.52 ± 0.17* 0 ± 0** 2.33 ± 0.20** 4 4.94 ± 0.32** 8.91 ± 0.03* 2.08 ± 0.38* 1.61 ± 0.92** 8 4.07 ± 0.19** 7.4 ± 0.17** 1.04 ± 0.72** 2.71 ± 0.05 12 4.39 ± 0.08** 7.13 ± 0.17** 2.71 ± 0.02 3.20 ± 0.02 rfbN::Tn5-pACYC-rfbN 1 4.49 ± 0.49** 9.46 ± 0.01 1.96 ± 1.03** 1.23 ± 0.90** 2 4.76 ± 0.62** 9.15 ± 0.05 0.23 ± 0.23** 0.23 ± 0.23** Downloaded from 4 6.30 ± 0.38 9.13 ± 0.11 1.12 ± 0.43** 0.69 ± 0** 8 4.88 ± 0.91* 8.09 ± 0.05** 1.03 ± 0.34** 0.69 ± 0** 12 4.35 ± 0.49** 7.16 ± 0.30** 0.69 ± 0** 2.11 ± 0.11 rfbM::Tn5-pACYC-rfbM 1 6.42 ± 0.28 9.46 ± 0.01 0.89 ± 0.89** 0.23 ± 0.23** 2 4.30 ± 0** 8.65 ± 0.26** 0.46 ± 0.23** 2.63 ± 0.20**

4 6.76 ± 0.06 8.45 ± 0.25* 3.2 ± 0.10 3.39 ± 0.07 http://ps.oxfordjournals.org/ 8 5.73 ± 0.28 8.39 ± 0.05 2.99 ± 0.05 3.70 ± 0.04 12 5.82 ± 0.15 7.89 ± 0.44* 3.13 ± 0.02 3.56 ± 0.02 SEN0803::Tn5-pACYC-SEN0803 1 4.66 ± 1.56** 9.41 ± 0.56 0.98 ± 0.00** 1.56 ± 0.87** 2 4.30 ± 0** 8.52 ± 0.02** 0.23 ± 0.23** 1.92 ± 1.02** 4 5.37 ± 0.53* 8.21 ± 0.28* 0.46 ± 0.23** 2.40 ± 0.10* 8 1.55 ± 0.07** 6.50 ± 0.17** 2.36 ± 0.08 2.94 ± 0.27 12 0.51 ± 0.17** 5.79 ± 0.05** 0.51 ± 0.17** 0.51 ± 0.17** SEN3503::Tn5-pACYC-SEN3503 1 4.92 ± 1.09** 9.45 ± 0.49 1.10 ± 0.01** 1.09 ± 0.77** 2 4.86 ± 0.56** 8.54 ± 0.03** 1.62 ± 1.29** 2.20 ± 1.12**

4 6.47 ± 0.37 8.70 ± 0.10* 2.38 ± 0.46* 2.49 ± 0.49* at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 8 4.47 ± 0.64** 7.37 ± 0.28** 2.08 ± 0.37* 1.18 ± 1.18** 12 4.68 ± 0.20** 5.86 ± 0.06** 1.28 ± 0.59** 1.72 ± 0.67** 1Averages were based on values from 3 independent chicks for each time point with exception of 4 chickens at 12 d postinfection. 2DPI, days postinfection. *One asterisk indicates significant differences with P-values between 0.01 and 0.005 between Salmonella Enteritidis mutants and wild-type strain. **Two asterisks show significant differences with P-values < 0.001 between Salmonella Enteritidis mutants and wild-type strain. tional inhibitor that binds to the fliC promoter, inhibit- len-Vercoe and Woodward, 1999). Allen-Vercoe et al. ing fliC transcription while expressing fljB. When fljA (1999) also reported that the number of fliC mutant is not transcribed, the FliC protein is produced. Our bacteria was significantly reduced in orally infected 1-d- results demonstrated that fliH::Tn5 mutant was not in old chickens in the liver and spleen, but not in the ceca, vivo attenuated because bacterial counts from chickens which is consistent with our results (Allen-Vercoe et infected with this mutant were not significantly differ- al., 1999). In contrast, Parker and Guard-Petter (2001) ent from chickens infected with the wild-type (WT) showed that fliC mutant was not significantly different strain in all organs during the entire experiment (Table from the WT strain in the spleen in chickens infected 2). Four chickens out of 16 (25%) died in the group in- at the age of 20 d. fected with fliH::Tn5 mutant and the mortality was observed up to 4 d PI, whereas 3 out of 16 (18%) chickens SPI Genes challenged with the WT strain died and mortality occurred by 48 h PI. Compared with its WT parent strain, Bacterial counts from all internal organs of chickens the bacterial counts in chickens infected with fljB::Tn5 infected with hilA::Tn5 mutant were significantly re- mutant were significantly reduced in the small intestine duced compared with the WT strain, starting from 4 d from d 2 onward, and in the liver and spleen from d PI in the small intestine, 2 d PI in the ceca, and during 1 onward (Table 2). In contrast, the trends over time the entire experiment in the liver and spleen with the were not significantly different in the ceca between the exception at 2 d PI. These results are in accordance WT strain and fljB::Tn5 mutant. Shah et al. (2012b) with the previous report showing that significant dif- showed that both fliH::Tn5 and fljB::Tn5 mutants were ferences between ΔhilA and WT Salmonella Enteritidis nonmotile and attenuated in their invasiveness in hu- strain in the ceca, liver, and spleen in 1-d-old chickens man intestinal epithelial (Caco-2) cells compared with after oral inoculation (Bohez et al., 2006). Others have the WT parent strain; the roles of FljB and FliH in Sal- reported that Salmonella Enteritidis mutants lacking monella Enteritidis in chickens have not been reported the entire SPI-1 are attenuated for intestinal coloni- previously. However, Allen-Vercoe et al. (1999) showed zation and systemic infection of the liver and spleen that fliC mutant was significantly less able to adhere to in chickens (Desin et al., 2009; Rychlik et al., 2009; chicken duodenal gut sections, suggesting that flagella Wisner et al., 2010). In Salmonella Typhimurium, an are involved in adherence of Salmonella Enteritidis to ΔSPI-I mutant showed significantly reduced bacterial intestinal epithelial cells (Allen-Vercoe et al., 1999; Al- counts in the ceca and spleen (Dieye et al., 2009). The 876 ADDWEBI ET AL. HilA is a regulatory protein that is known to bind to intestine, and during the entire experimental period for the inv-spa promoter and the prg-org promoter, activat- the liver and spleen compared with the wild-type strain ing the expression of SPI-I (Hacker and Kaper, 2000; (Table 2). Bacterial counts from the ceca of chickens Dieye et al., 2009). Activated SPI-I effector proteins challenged with this mutant were not significantly dif- are secreted into host cells that subsequently stimulate ferent from the wild-type strain. In agreement with Rho family GTPases. This leads to cytoskeleton po- our results, a mutant with transposon insertion in the lymerization that, in turn, facilitates Salmonella Enter- SEN0803 homolog was attenuated in the ability to col- itidis uptake into the host cells. In corroboration with onize the intestine in chickens, pigs, and cattle, but other studies that demonstrated that SPI-1 facilitates not in intravenously inoculated mice (Chaudhuri et al., systemic spread of Salmonella Enteritidis in chickens 2013). The SEN0803 encodes a putative acyl-CoA de- (Desin et al., 2009; Rychlik et al., 2009), our results hydrogenase, which uses long-chain fatty acids in star- Downloaded from also showed that the hilA gene is involved in the coloni- vation conditions, yielding carbon as a source of energy zation of the liver and spleen. The decrease in bacterial (Spector et al., 1999). It is unlikely that the defect in counts in the liver and spleen may result from a defect fatty acid degradation is responsible for attenuation in in intestinal colonization resulting in slower and lower this study. However, Shah et al. (2011) showed that the systemic dissemination. Another possible explanation is SEN0803::Tn5 mutant was more susceptible to anti- http://ps.oxfordjournals.org/ that hilA::Tn5 mutant is less fit in its ability to survive microbial components in egg albumen compared with within the macrophages that disseminate Salmonella to the wild-type strain (Shah et al., 2012b). Therefore, it internal organs (Bohez et al., 2006; Shah et al., 2012b). is likely that functionally deleted SEN0803 may render The SPI-5 encodes 7 genes, including, pipABC, sopB, the SEN0803::Tn5 mutant more vulnerable to the host SEN0957, and pipD. One mutant with a transposon in- antimicrobial peptides resulting in reduced infectivity. sertion in pipA gene was tested in this study. Our results demonstrated that pipA::Tn5 mutant was significantly Lipopolysaccharide O-Antigen at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 attenuated compared with the WT strain in all organs Biosynthesis Genes at all time points, with the exception at 12 d PI in the liver and spleen. The role of SPI-5 in virulence has been Somatic (O) antigen of Salmonella is a part of lipo- studied using different models, providing conflicting polysaccharide (LPS) and a recognized virulence de- data. In Salmonella Enteritidis, Rychlik et al. (2009) terminant that is encoded within the rfb operon. Two showed that the absence of SPI-5 in Salmonella Enter- Salmonella Enteritidis mutants carrying transposon itidis did not result in reduced cecal colonization and insertions in rfbM and rfbN, respectively, were tested systemic spread in chickens (Rychlik et al., 2009). In in this study. The number of bacteria isolated from contrast, Dhawi et al. (2011) demonstrated that SPI-5 chickens infected with the rfbN::Tn5 mutant was sig- genes were upregulated in the ceca of chickens infected nificantly lower in all internal organs compared with with WT Salmonella Enteritidis (Dhawi et al., 2011). the wild-type strain, with the exceptions including 2 d In mice, Karasova et al. (2010) found no differences in PI in the ceca, and 4 d PI in the liver and spleen (Table cecal colonization and systemic infection between WT 2). The rfbM::Tn5 mutant was also attenuated at all Salmonella Enteritidis and a ΔSPI-5 mutant (Karasova time points in the ceca, at 2, 4, and 8 d PI in the small et al., 2010). Screening of Salmonella Enteritidis trans- intestine, and at 1 and 2 d PI in the liver and spleen. poson mutants recovered from intraperitoneally inocu- Although the in vivo defects of rfbN::Tn5 mutant could lated mice indicated that SPI-5 genes were negatively be partially rescued by in trans complementation of selected (Silva et al., 2012). Although the exact roles rfbN at 8 d PI in the small intestine, 4 d PI in the ceca, of pipA and SPI-5 genes in the Salmonella pathogenesis and 12 d PI in the spleen (Figure 1), the in vivo defects are not clear, SPI-5 genes are coregulated with SPI-1 of rfbM::Tn5 mutant could not be restored by in trans and SPI-2. Both PipA and PipB are SPI-2 T3SS-se- complementation of rfbM gene (Table 2). Others have creted proteins. Furthermore, other genes of SPI-5 are reported that Salmonella Typhimurium mutants with also regulated by SPI-1 and SPI-2 (Ong et al., 2010). transposon insertions in rfb genes, including rfbN, are Finally, transposon-mediated mutations in SPI-5 genes attenuated in their ability to colonize the intestine in are attenuated in animal models (Morgan et al., 2004; cattle, pigs, and chickens (Craven et al., 1993; Craven, Silva et al., 2012); however, mutations generated by 1994; Turner et al., 1998; Morgan et al., 2004; Lawley complete deletion of a particular gene within an operon et al., 2006; Ilg et al., 2009; Kong et al., 2011; Chaud- do not show in vivo virulence attenuation (Rychlik et huri et al., 2013). Signature-tagged mutagenesis studies al., 2009; Karasova et al., 2010). This suggests that in Salmonella Enteritidis have also confirmed the role transposon insertions in SPI-5 may have a polar effect of rfb genes in systemic infection in mice and chickens on other genes that downregulates the expression of (Chang et al., 2008; Silva et al., 2012; Coward et al., SPI-1 and SPI-2 and may explain the in vivo virulence 2013). In addition, Salmonella Dublin mutant with a attenuation of pipA::Tn5 mutant observed in our study. transposon insertion in rfbM was attenuated in orally The SPI-14 harbors 6 genes from SEN0800-SEN0805. infected mice (Thomsen et al., 2003). Although the ex- The SEN0803::Tn5 mutant showed significant in vivo act role of Salmonella Enteritidis O antigen in virulence virulence attenuation, starting from 4 d PI in the small is not fully understood, it has been hypothesized that SALMONELLA ENTERITIDIS VIRULENCE FACTORS 877 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

(complemented strain rfbN::Tn5-pACYC-rfbN (o), and negative control (Δ ,(څ) Figure 1. Log10 cfu of wild-type strain (), rfbN::Tn5 mutant in the small intestine (a), cecum (b), liver (c), and spleen (D). Each time point represents the mean ± SE values from 3 chickens, with the exception of 4 chickens at 12 d postinfection. P-values show significant differences between rfbN::Tn5 mutant and its complemented strain. Significant differences between the wild-type strain, rfbN::Tn5 mutant and complemented strain are shown in Table 2. rfb genes may facilitate the epithelial cell entry of Sal- with the wild-type strain in the small intestine at 4, 8, monella Enteritidis as the first step to establish the and 12 d PI, and in the liver and spleen during the en- infection in chickens (Stone et al., 1992; Scallan et al., tire experiment (Table 2). The SEN2278 is homologous 2011). Indeed, rfbN::Tn5 and rfbM::Tn5 mutants were to the hypothetical protein in Salmonella Typhimurium both attenuated in the small intestine and ceca in oral- (SL1344_ 2265) that has recently been shown to play a ly infected chickens. This latter hypothesis needs to be role in intestinal colonization of Salmonella Typhimuri- further examined by testing these mutants in chickens um in chickens, pigs, and cattle, but not in intrave- using an alternative route of inoculation such as intra- nously inoculated mice (Chaudhuri et al., 2013). The venous challenge. In addition to the entry mechanism core region of LPS in Salmonella contains a negatively to epithelial cells, O antigen protects Salmonella En- charged hep (II) phosphate to which positively charged teritidis from certain antimicrobial peptides produced antimicrobial compounds can bind, thereby increasing by the host (Clavijo et al., 2006; Gantois et al., 2008, the susceptibility of Salmonella to these compounds 2009; Coward et al., 2013). Therefore, the virulence at- (Nishino et al., 2006). The SEN2278 protein appears tenuation of rfbM and rfbN mutants observed in this to dephosphorylate the hep (II) phosphate in the core study could also be due in part to their susceptibility region of LPS, preventing this region from binding to to antimicrobial peptides within the host. antimicrobial compounds that, in turn, protects Salmo- nella from positively charged antimicrobial compounds. Hypothetical Proteins Indeed, SEN2278::Tn5 was more susceptible to poly- myxin B when compared with the wild-type strain We also investigated SEN2278 and SEN0034, both of (Shah et al., 2012b). Presumably, the SEN2278::Tn5 which encode hypothetical proteins. The SEN2278::Tn5 mutant cannot dephosphorylate the hep(II) phosphate, mutant was significantly attenuated when compared rendering the SEN2278::Tn5 mutant more susceptible 878 ADDWEBI ET AL. to host-produced antimicrobial compounds in chickens. who found that 3 out of 5 Salmonella Enteritidis strains The SEN0034 gene encodes a putative outer membrane lacking the 55 kb plasmid did not cause any mortal- export protein belonging to a family of channel-forming ity in chickens after oral inoculation whereas 2 strains transmembrane proteins of nucleosides. Although the lacking this virulence plasmid killed only 1 out of 5 SEN0034::Tn5 mutant was attenuated in epithelial cell chickens. In contrast, 19 wild-type strains carrying the invasiveness (Shah et al., 2012b), this mutant was not 55 kb plasmid caused 40 to 100% mortality in chick- attenuated in chickens (Table 2). ens (Bakshi et al., 2003). Our results are consistent with Halavatkar and Barrow (1993) who demonstrated Virulence Plasmid-Associated Gene that bacterial counts were not significantly different between the wild-type strain and an isogenic, plasmid- and Fimbriae Downloaded from cured strain; however, these authors did not observe Most strains of Salmonella Enteritidis harbor a high- any significant difference in mortality between the wild- molecular mass plasmid that encodes a virulence gene type strain and Salmonella Enteritidis strain lacking cluster composed of 5 genes (spvRABCD). The SpvR this virulence plasmid (Halavatkar and Barrow, 1993). regulates the expression of spv genes and has been re- There are 14 predicated fimbrial operons in Salmo- ported as a virulence factor. Interestingly, spvR::Tn5 nella Enteritidis (Clayton et al., 2008; Thomson et al., http://ps.oxfordjournals.org/ mutant was not significantly compromised with respect 2008). Thirteen fimbrial loci are encoded on the ge- to bacterial counts in the internal organs of orally chal- nome, whereas one locus (pef operon) is plasmid encod- lenged chickens, but it did not cause any mortality ed. In this study, we assessed the role of pegD belonging compared with the wild-type strain in which 3 out of to the chaperone-usher class known as peg fimbriae and 16 chickens (18%) died during the experiment by 48 h the csgB belonging to the nucleator-dependent class PI. These results corroborate with Bakshi et al. (2003), known as csg fimbriae in virulence in chickens. For at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

-complemented strain SEN3503::Tn5-pACYC-SEN3503 (o) and nega ,(څ) Figure 2. Log10 cfu of wild-type strain (), SEN3503::Tn5 mutant tive control (Δ) in the small intestine (a), cecum (b), liver (c), and spleen (d). Each time point represents the mean ± SE values from 3 chickens, with the exception of 4 chickens at 12 d postinfection. P-values show significant differences between SEN3503::Tn5 mutant and its complemented strain. Significant differences between the wild-type strain, SEN3503::Tn5, and complemented strain are shown in Table 2. SALMONELLA ENTERITIDIS VIRULENCE FACTORS 879 both csgB::Tn5 and pegD::Tn5 mutants, we observed SEN3503::Tn5) could not be restored. It is important no significant differences in the ceca, liver, and spleen to note that rfbN, rfbN, and SEN0803 genes are a part compared with the wild-type strain. In contrast, the of a large operon and therefore in trans complementa- bacterial counts were significantly reduced in the small tion of genes may be less efficient or there could be intestine for both mutants (Table 2). Fimbriae are im- polar effects of mutations on genes upstream or down- portant for attachment to intestinal cells (Thiagarajan stream. In addition, in some cases, pACYC184 plasmid et al., 1996; Edelman et al., 2003; Li et al., 2003). In has been shown to reduce the ability of invasion in accordance with this observation, the reduced coloni- Hela cells in Salmonella Typhimurium (Knodler et al., zation of small intestine by pegD::Tn5 and csgB::Tn5 2005) or affect the heterologous gene expression due to could be due to the reduced adherence of mutants to the the chloramphenicol resistance cassette present on the small intestinal epithelial cells. The mutations in other plasmid (Clark et al., 2009). In cis complementation Downloaded from genes within these fimbrial operons have had variable may circumvent some of these issues. effects on virulence. For instance, Clayton et al. (2008) found no significant difference in cecal colonization of Conclusions Salmonella Enteritidis between ΔcsgA mutant and the wild-type strain, consistent with our results (Clayton In this study, 14 Salmonella Enteritidis mutants http://ps.oxfordjournals.org/ et al., 2008). In contrast to our results, however, they were screened for their in vivo virulence in orally in- showed a significant reduction in bacterial counts in the oculated 1-d-old chickens. All of these mutants were ceca between ΔpegA mutant and the wild-type strain. previously shown as invasion attenuated in cultured hu- In Salmonella Typhimurium, the fimbrial regulatory man intestinal epithelial cells (Caco-2 cells; Shah et al., CsgD is important for cecal colonization of chickens, 2012b). Here, we show that at least 10 of these Salmo- but not bovines (Morgan et al., 2004). nella Enteritidis mutants were also in vivo attenuated

in chickens. These findings illustrate how cell culture at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Genes Involved in DNA Recombination models address the principles of reduction, refinement, and replacement of animals in research, although it is and Carbohydrate Metabolism also evident that cell culture models may not identify The SEN1393 belongs to the region annotated as the all genes involved in virulence in chickens. Previous- Salmonella Enteritidis specific prophage ϕSE14 that ly identified virulence genes (e.g., SPI-1, SPI-5, pegD, harbors a total of 20 genes (Thomson et al., 2008). csgB, fljB, and rfbMN) with known functions were con- Bacterial counts from chickens orally infected with the firmed in this study. Furthermore, 3 genes with puta- SEN1393::Tn5 mutant were not significantly different tive or previously unknown functions (e.g., SEN2278, from the wild-type strain in all organs, suggesting that SEN0803, and SEN3503) were also characterized. The SEN1393 may not be important for virulence in chick- latter group needs to be further investigated to iden- ens. In contrast, infection with the SEN3503::Tn5 mu- tify the mechanisms underlying virulence attenuation tant (a putative transcription regulatory gene) was sig- in chickens. Some of the virulence-attenuated mutants nificantly attenuated in the small intestine, liver, and identified in this study could serve as potential candi- spleen compared with the WT strain during the entire dates for development of either live-attenuated or sub- experiment, with the exception at 1 d PI. Bacterial unit vaccines for prevention of Salmonella Enteritidis counts obtained from the ceca, however, were not sig- infection in chicken. nificantly different between chickens infected with this mutant and the wild-type strain. In vivo attenuation in ACKNOWLEDGMENTS the small intestine, liver, and spleen was only partially restored at 4 d after inoculation of the complemented We thank Thomas E. Besser, David Prieur and Mi- strain (Figure 2). To our knowledge, this is the first chael E. Konkel (professors, Department of Veterinary time that SEN3503 has been demonstrated as virulence Microbiology and Pathology at Washington State Uni- factor in chickens, although the mechanism underlying versity, Pullman) for the helpful technical discussions. this virulence attenuation remains to be determined. The funding for this work was provided by the College We attempted to rescue the virulence defects in 4 of Veterinary Medicine, Washington State University, mutants (rfbM::TN5, rfbN::Tn5, SEN3503::Tn5, and Pullman. SEN00803::Tn5) by in trans complementation on a low-copy number plasmid pACYC184 in which the het- REFERENCES erologous gene expression was controlled by the native Agron, P. G., R. L. Walker, H. Kinde, S. J. Sawyer, D. C. Hayes, J. promoter for tetracycline resistance gene. Although Wollard, and G. L. Andersen. 2001. Identification by subtractive this plasmid has been successfully used to comple- hybridization of sequences specific for Salmonella enterica serovar ment gene function (Deiwick et al., 1998), we could enteritidis. Appl. Environ. Microbiol. 67:4984–4991. only partially rescue the virulence defects for 2 mutants Allen-Vercoe, E., A. R. Sayers, and M. J. Woodward. 1999. Viru- lence of Salmonella enterica serotype Enteritidis aflagellate and (e.g., SEN0803::Tn5 and rfbM::Tn5), whereas the viru- afimbriate mutants in a day-old chick model. Epidemiol. Infect. lence defects in the other 2 mutants (rfbM::Tn5 and 122:395–402. 880 ADDWEBI ET AL.

Allen-Vercoe, E., and M. J. Woodward. 1999. The role of flagella, Desmidt, M., R. Ducatelle, and F. Haesebrouck. 1997. Pathogenesis but not fimbriae, in the adherence of Salmonella enterica serotype of Salmonella enteritidis phage type four after experimental infec- Enteritidis to chick gut explant. J. Med. Microbiol. 48:771–780. tion of young chickens. Vet. Microbiol. 56:99–109. Altekruse, S. F., N. Bauer, A. Chanlongbutra, R. DeSagun, A. Nau- Dhawi, A. A., A. Elazomi, M. A. Jones, M. A. Lovell, H. Li, R. D. gle, W. Schlosser, R. Umholtz, and P. White. 2006. Salmonella Emes, and P. A. Barrow. 2011. Adaptation to the chicken in- enteritidis in broiler chickens, United States, 2000–2005. Emerg. testine in Salmonella Enteritidis PT4 studied by transcriptional Infect. Dis. 12:1848–1852. analysis. Vet. Microbiol. 153:198–204. Alvarez, J., M. Sota, A. B. Vivanco, I. Perales, R. Cisterna, A. Dieye, Y., K. Ameiss, M. Mellata, and R. Curtiss 3rd.. 2009. The Rementeria, and J. Garaizar. 2004. Development of a multiplex Salmonella Pathogenicity Island (SPI) 1 contributes more than PCR technique for detection and epidemiological typing of Sal- SPI2 to the colonization of the chicken by Salmonella enterica monella in human clinical samples. J. Clin. Microbiol. 42:1734– serovar Typhimurium. BMC Microbiol. 9:3. 1738. Edelman, S., S. Leskela, E. Ron, J. Apajalahti, and T. K. Korhonen.

Apel, D., and M. G. Surette. 2008. Bringing order to a complex mo- 2003. In vitro adhesion of an avian pathogenic Escherichia coli Downloaded from lecular machine: The assembly of the bacterial flagella. Biochim. O78 strain to surfaces of the chicken intestinal tract and to ileal Biophys. Acta 1778:1851–1858. mucus. Vet. Microbiol. 91:41–56. Bakshi, C. S., V. P. Singh, M. Malik, R. K. Singh, and B. Sharma. Gantois, I., R. Ducatelle, F. Pasmans, F. Haesebrouck, and F. Van 2003. 55 kb plasmid and virulence-associated genes are positively Immerseel. 2008. Salmonella enterica serovar Enteritidis genes correlated with Salmonella enteritidis pathogenicity in mice and induced during oviduct colonization and egg contamination in chickens. Vet. Res. Commun. 27:425–432. laying hens. Appl. Environ. Microbiol. 74:6616–6622. Bohez, L., R. Ducatelle, F. Pasmans, N. Botteldoorn, F. Haeseb- Gantois, I., R. Ducatelle, F. Pasmans, F. Haesebrouck, and F. Van http://ps.oxfordjournals.org/ rouck, and F. Van Immerseel. 2006. Salmonella enterica serovar Immerseel. 2009. The Salmonella Enteritidis lipopolysaccharide Enteritidis colonization of the chicken caecum requires the HilA biosynthesis gene rfbH is required for survival in egg albumen. regulatory protein. Vet. Microbiol. 116:202–210. Zoonoses Public Health 56:145–149. Bohez, L., I. Gantois, R. Ducatelle, F. Pasmans, J. Dewulf, F. Haese- Gast, R. K., and C. W. Beard. 1990a. Isolation of Salmonella en- brouck, and F. Van Immerseel. 2008. The Salmonella Pathogenic- teritidis from internal organs of experimentally infected hens. ity Island 2 regulator ssrA promotes reproductive tract but not Avian Dis. 34:991–993. intestinal colonization in chickens. Vet. Microbiol. 126:216–224. Gast, R. K., and C. W. Beard. 1990b. Production of Salmonella Chang, J., E. Pang, H. He, and J. Kwang. 2008. Identification of enteritidis-contaminated eggs by experimentally infected hens.

novel attenuated Salmonella Enteritidis mutants. FEMS Immu- Avian Dis. 34:438–446. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 nol. Med. Microbiol. 53:26–34. Gast, R. K., R. Guraya, J. Guard-Bouldin, P. S. Holt, and R. W. Chaudhuri, R. R., E. Morgan, S. E. Peters, S. J. Pleasance, D. L. Moore. 2007. Colonization of specific regions of the reproductive Hudson, H. M. Davies, J. Wang, P. M. van Diemen, A. M. Buck- tract and deposition at different locations inside eggs laid by ley, A. J. Bowen, G. D. Pullinger, D. J. Turner, G. C. Langridge, hens infected with Salmonella enteritidis or Salmonella heidel- A. K. Turner, J. Parkhill, I. G. Charles, D. J. Maskell, and M. P. berg. Avian Dis. 51:40–44. Stevens. 2013. Comprehensive assignment of roles for Salmonella Hacker, J., and J. B. Kaper. 2000. Pathogenicity islands and the typhimurium genes in intestinal colonization of food-producing evolution of microbes. Annu. Rev. Microbiol. 54:641–679. animals. PLoS Genet. 9:e1003456. Halavatkar, H., and P. A. Barrow. 1993. The role of a 54-kb plasmid Clark, L., I. Martinez-Argudo, T. J. Humphrey, and M. A. Jepson. in the virulence of strains of Salmonella enteritidis of phage type 2009. GFP plasmid-induced defects in Salmonella invasion de- 4 for chickens and mice. J. Med. Microbiol. 38:171–176. pend on plasmid architecture, not protein expression. Microbiol- Ikeda, J. S., C. K. Schmitt, S. C. Darnell, P. R. Watson, J. Bi- ogy 155:461–467. spham, T. S. Wallis, D. L. Weinstein, E. S. Metcalf, P. Adams, Clavijo, R. I., C. Loui, G. L. Andersen, L. W. Riley, and S. Lu. 2006. C. D. O’Connor, and A. D. O’Brien. 2001. Flagellar phase varia- Identification of genes associated with survival of Salmonella en- tion of Salmonella enterica serovar Typhimurium contributes to terica serovar Enteritidis in chicken egg albumen. Appl. Environ. virulence in the murine typhoid infection model but does not Microbiol. 72:1055–1064. influence Salmonella-induced enteropathogenesis. Infect. Immun. Clayton, D. J., A. J. Bowen, S. D. Hulme, A. M. Buckley, V. L. 69:3021–3030. Deacon, N. R. Thomson, P. A. Barrow, E. Morgan, M. A. Jones, Ilg, K., K. Endt, B. Misselwitz, B. Stecher, M. Aebi, and W. D. M. Watson, and M. P. Stevens. 2008. Analysis of the role of 13 Hardt. 2009. O-antigen-negative Salmonella enterica serovar Ty- major fimbrial subunits in colonisation of the chicken intestines phimurium is attenuated in intestinal colonization but elicits coli- by Salmonella enterica serovar Enteritidis reveals a role for a tis in streptomycin-treated mice. Infect. Immun. 77:2568–2575. novel locus. BMC Microbiol. 8:228. Jackson, B. R., P. M. Griffin, D. Cole, K. A. Walsh, and S. J. Coward, C., L. Sait, T. Cogan, T. J. Humphrey, and D. J. Maskell. Chai. 2013. Outbreak-associated Salmonella enterica serotypes 2013. O-antigen repeat number in Salmonella enterica serovar and food commodities, United States, 1998–2008. Emerg. Infect. Enteritidis is important for egg contamination, colonisation of Dis. 19:1239–1244. the chicken reproductive tract and survival in egg albumen. Karasova, D., A. Sebkova, H. Havlickova, F. Sisak, J. Volf, M. Faldy- FEMS Microbiol. Lett. 343:169–176. na, P. Ondrackova, V. Kummer, and I. Rychlik. 2010. Influence Craven, S. E. 1994. Altered colonizing ability for the ceca of broiler of 5 major Salmonella pathogenicity islands on NK cell deple- chicks by lipopolysaccharide-deficient mutants of Salmonella ty- tion in mice infected with Salmonella enterica serovar Enteritidis. phimurium. Avian Dis. 38:401–408. BMC Microbiol. 10:75. Craven, S. E., N. A. Cox, J. S. Bailey, N. J. Stern, R. J. Meiners- Knodler, L. A., A. Bestor, C. Ma, I. Hansen-Wester, M. Hensel, B. mann, and L. C. Blankenship. 1993. Characterization of Salmo- A. Vallance, and O. Steele-Mortimer. 2005. Cloning vectors and nella california and S. typhimurium strains with reduced abil- fluorescent proteins can significantly inhibit Salmonella enterica ity to colonize the intestinal tract of broiler chicks. Avian Dis. virulence in both epithelial cells and macrophages: Implications 37:339–348. for bacterial pathogenesis studies. Infect. Immun. 73:7027–7031. Deiwick, J., T. Nikolaus, J. E. Shea, C. Gleeson, D. W. Holden, and Kong, Q. J., X. Y. Ren, N. Hu, C. R. Sun, and Y. J. Pan. 2011. M. Hensel. 1998. Mutations in Salmonella pathogenicity island 2 Identification of isomers of resveratrol dimer and their analogues (SPI2) genes affecting transcription of SPI1 genes and resistance from wine grapes by HPLC/MS(n) and HPLC/DAD-UV. Food to antimicrobial agents. J. Bacteriol. 180:4775–4780. Chem. 127:727–734. Desin, T. S., P. K. Lam, B. Koch, C. Mickael, E. Berberov, A. L. Kuehn, B. M. 2010. Salmonella cases traced to egg producers: Wisner, H. G. Townsend, A. A. Potter, and W. Koster. 2009. Findings trigger recall of more than 500 million eggs. JAMA Salmonella enterica serovar enteritidis pathogenicity island 1 is 304:1316. not essential for but facilitates rapid systemic spread in chickens. Lawley, T. D., K. Chan, L. J. Thompson, C. C. Kim, G. R. Govoni, Infect. Immun. 77:2866–2875. and D. M. Monack. 2006. Genome-wide screen for Salmonella SALMONELLA ENTERITIDIS VIRULENCE FACTORS 881

genes required for long-term systemic infection of the mouse. Shah, D. H., X. Zhou, H. Y. Kim, D. R. Call, and J. Guard. 2012b. PLoS Pathog. 2:e11. Transposon mutagenesis of Salmonella enterica serovar Enter- Li, W., S. Watarai, and H. Kodama. 2003. Identification of possible itidis identifies genes that contribute to invasiveness in human chicken intestinal mucosa receptors for SEF21-fimbriated Salmo- and chicken cells and survival in egg albumen. Infect. Immun. nella enterica serovar Enteritidis. Vet. Microbiol. 91:215–229. 80:4203–4215. Lu, S., A. R. Manges, Y. Xu, F. C. Fang, and L. W. Riley. 1999. Silva, C. A., C. J. Blondel, C. P. Quezada, S. Porwollik, H. L. Analysis of virulence of clinical isolates of Salmonella enteritidis Andrews-Polymenis, C. S. Toro, M. Zaldivar, I. Contreras, M. in vivo and in vitro. Infect. Immun. 67:5651–5657. McClelland, and C. A. Santiviago. 2012. Infection of mice by Macnab, R. M. 2004. Type III flagellar protein export and flagellar Salmonella enterica serovar Enteritidis involves additional genes assembly. Biochim. Biophys. Acta 1694:207–217. that are absent in the genome of serovar Typhimurium. Infect. Minamino, T., and R. M. MacNab. 2000. FliH, a soluble component Immun. 80:839–849. of the type III flagellar export apparatus of Salmonella, forms a Spector, M. P., C. C. DiRusso, M. J. Pallen, F. Garcia del Por-

complex with FliI and inhibits its ATPase activity. Mol. Micro- tillo, G. Dougan, and B. B. Finlay. 1999. The medium-/long- Downloaded from biol. 37:1494–1503. chain fatty acyl-CoA dehydrogenase (fadF) gene of Salmonella Morgan, E., J. D. Campbell, S. C. Rowe, J. Bispham, M. P. Stevens, typhimurium is a phase 1 starvation-stress response (SSR) locus. A. J. Bowen, P. A. Barrow, D. J. Maskell, and T. S. Wallis. 2004. Microbiology 145:15–31. Identification of host-specific colonization factors of Salmonella Stone, B. J., C. M. Garcia, J. L. Badger, T. Hassett, R. I. Smith, enterica serovar Typhimurium. Mol. Microbiol. 54:994–1010. and V. L. Miller. 1992. Identification of novel loci affecting en- Nishino, K., F. F. Hsu, J. Turk, M. J. Cromie, M. M. Wosten, and try of Salmonella enteritidis into eukaryotic cells. J. Bacteriol. E. A. Groisman. 2006. Identification of the lipopolysaccharide 174:3945–3952. http://ps.oxfordjournals.org/ modifications controlled by the Salmonella PmrA/PmrB system Thiagarajan, D., H. L. Thacker, and A. M. Saeed. 1996. Experimen- mediating resistance to Fe(III) and Al(III). Mol. Microbiol. tal infection of laying hens with Salmonella enteritidis strains that 61:645–654. express different types of fimbriae. Poult. Sci. 75:1365–1372. Ong, S. Y., F. L. Ng, S. S. Badai, A. Yuryev, and M. Alam. 2010. Thomsen, L. E., M. S. Chadfield, J. Bispham, T. S. Wallis, J. E. Analysis and construction of pathogenicity island regulatory Olsen, and H. Ingmer. 2003. Reduced amounts of LPS affect pathways in Salmonella enterica serovar Typhi. J. Integr. Bio- both stress tolerance and virulence of Salmonella enterica serovar inform. 7:145. Dublin. FEMS Microbiol. Lett. 228:225–231. Parker, C. T., and J. Guard-Petter. 2001. Contribution of flagella Thomson, N. R., D. J. Clayton, D. Windhorst, G. Vernikos, S. Da-

and invasion proteins to pathogenesis of Salmonella enterica se- vidson, C. Churcher, M. A. Quail, M. Stevens, M. A. Jones, M. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 rovar enteritidis in chicks. FEMS Microbiol. Lett. 204:287–291. Watson, A. Barron, A. Layton, D. Pickard, R. A. Kingsley, A. Patrick, M. E., P. M. Adcock, T. M. Gomez, S. F. Altekruse, B. Bignell, L. Clark, B. Harris, D. Ormond, Z. Abdellah, K. Brooks, H. Holland, R. V. Tauxe, and D. L. Swerdlow. 2004. Salmonella I. Cherevach, T. Chillingworth, J. Woodward, H. Norberczak, A. enteritidis infections, United States, 1985–1999. Emerg. Infect. Lord, C. Arrowsmith, K. Jagels, S. Moule, K. Mungall, M. Sand- Dis. 10:1–7. ers, S. Whitehead, J. A. Chabalgoity, D. Maskell, T. Humphrey, Popiel, I., and P. C. Turnbull. 1985. Passage of Salmonella enteritidis M. Roberts, P. A. Barrow, G. Dougan, and J. Parkhill. 2008. and Salmonella thompson through chick ileocecal mucosa. Infect. Comparative genome analysis of Salmonella Enteritidis PT4 and Immun. 47:786–792. Salmonella Gallinarum 287/91 provides insights into evolution- Rychlik, I., D. Karasova, A. Sebkova, J. Volf, F. Sisak, H. Hav- ary and host adaptation pathways. Genome Res. 18:1624–1637. lickova, V. Kummer, A. Imre, A. Szmolka, and B. Nagy. 2009. Turnbull, P. C., and G. H. Snoeyenbos. 1974. Experimental salmo- Virulence potential of five major pathogenicity islands (SPI-1 nellosis in the chicken. 2. Fate of a temperature-sensitive filamen- to SPI-5) of Salmonella enterica serovar Enteritidis for chickens. tous mutant. Avian Dis. 18:178–185. BMC Microbiol. 9:268. Turner, A. K., M. A. Lovell, S. D. Hulme, L. Zhang-Barber, and P. Scallan, E., R. M. Hoekstra, F. J. Angulo, R. V. Tauxe, M. A. Wid- A. Barrow. 1998. Identification of Salmonella typhimurium genes dowson, S. L. Roy, J. L. Jones, and P. M. Griffin. 2011. Food- required for colonization of the chicken alimentary tract and for borne illness acquired in the United States—Major pathogens. virulence in newly hatched chicks. Infect. Immun. 66:2099–2106. Emerg. Infect. Dis. 17:7–15. Van Immerseel, F., J. De Buck, I. De Smet, J. Mast, F. Haesebrouck, Shah, D. H., C. Casavant, Q. Hawley, T. Addwebi, D. R. Call, and J. and R. Ducatelle. 2002. Dynamics of immune cell infiltration Guard. 2012a. Salmonella Enteritidis strains from poultry exhibit in the caecal lamina propria of chickens after neonatal infec- differential responses to acid stress, oxidative stress, and survival tion with a Salmonella enteritidis strain. Dev. Comp. Immunol. in the egg albumen. Foodborne Pathog. Dis. 9:258–264. 26:355–364. Shah, D. H., X. Zhou, T. Addwebi, M. A. Davis, L. Orfe, D. R. Call, Wisner, A. L., T. S. Desin, B. Koch, P. K. Lam, E. M. Berberov, C. S. J. Guard, and T. E. Besser. 2011. Cell invasion of poultry-associ- Mickael, A. A. Potter, and W. Koster. 2010. Salmonella enterica ated Salmonella enterica serovar Enteritidis isolates is associated subspecies enterica serovar Enteritidis Salmonella pathogenicity with pathogenicity, motility and proteins secreted by the type III island 2 type III secretion system: Role in intestinal colonization secretion system. Microbiology 157:1428–1445. of chickens and systemic spread. Microbiology 156:2770–2781.

Arginine and vitamin E improve the immune response after a Salmonella challenge in broiler chicks

X. Liu ,* J. A. Byrd ,† M. Farnell,‡ and C. A. Ruiz-Feria *1

* Poultry Science Department, Texas A&M University, College Station 77843-2472; † USDA-ARS, Southern Plains Agricultural Research Center, College Station TX 77845; and ‡ Poultry Science Department, Mississippi State University, Mississippi State 39762 Downloaded from

ABSTRACT Two experiments were conducted to ment, birds fed the AVE diet had higher MOB than evaluate the effects of Arg, vitamin E (VE), and man- birds fed CTL+ or the AVM diet at 7 d PI, whereas 9 nanoligosaccharide (MOS) on the immune response d PI birds fed the AVM diet had the highest MOB. In and clearance of Salmonella in broiler chickens. In each experiment 2, birds fed the AVE diet had higher MOB, http://ps.oxfordjournals.org/ experiment, 1-d-old chicks (n = 160) were randomly HOB, and LPR than birds in the other treatments 7 distributed into 4 groups: antibiotic-free diet (negative and 14 d PI, except at 7 d PI, when MOB was not control, CTL−), antibiotic-supplemented diet (positive different among treatments. Birds fed the AVM diet control, CTL+), antibiotic free-diet plus Arg and VE had the highest IgA antibody titer, and a higher IgM (AVE), or antibiotic-free diet plus Arg, VE, and MOS antibody titer than the CTL+ birds. In experiment 1, (AVM). Birds were orally challenged with 106 cfu of Salmonella Typhimurium content in the ceca was lower

a novobiocyn and nalidixic acid-resistant Salmonella in birds fed the AVM diet compared with birds fed at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 enterica serovar Typhimurium strain at d 7 (experi- the CTL− diet 3 d PI, but later on (10 and 17 d PI), ment 1) or at d 3 (experiment 2). Heterophil- (HOB) and in experiment 2 (7, 14, and 21 d PI), Salmonella and monocyte- (MOB) oxidative burst and lymphocyte Typhimurium concentrations were not different among proliferation (LPR), antibody titers, and Salmonella treatments. Thus, Arg and VE improved immune re- content in the ceca were measured at several intervals sponse after a Salmonella Typhimurium challenge in

postinfection (PI). In experiment 1, both AVM and young chicks, and although they did not reduce Sal- AVE decreased HOB compared with the controls 5 and monella Typhimurium concentrations in the ceca, they 9 d PI, but increased LPR 9 d PI. In the same experi- may improve bacterial resistance against other pathogens in commercial growing conditions. Key words: immune response, Salmonella , arginine , vitamin E, mannanoligosaccharide 2014 Poultry Science 93 :882–890 http://dx.doi.org/ 10.3382/ps.2013-03723

INTRODUCTION of antimicrobial in food-producing animals for growth promotion and prophylaxis may lead to the develop- Consumption of contaminated poultry products is a ment of microbial strains resistant to antibiotic therapy major cause of human salmonellosis, with most of the (Bach Knudsen, 2001; Harrison et al., 2013). Therefore, cases being attributed to Salmonella enterica serovar it is essential to find alternatives to antibiotics in food- Enteritidis or Typhimurium (Beal et al., 2006). On the producing animals. other hand, intensive poultry production practices may One way to improve immune response and avian increase susceptibility to disease due to high stocking health is through the use of nutritional supplements densities, recycled litter, and environmental conditions (Klasing, 1998; Kidd, 2004). Arginine, an essential that are sometimes less than optimal. To reduce the amino acid for avian species, and vitamin E (VE), an incidence of disease and improve growth, antibiotics important antioxidant, have been shown to influence are used in commercial animal production; drug makers both the humoral- and cell-mediated immune responses sold 13.5 million kg of antibiotics for use in food-pro- of birds. In chickens, an Arg-deficient diet has been ducing animals in 2011 (FDA, 2013). However, there associated with poor development of the thymus and are increasing concerns that the subtherapeutic use spleen, which are key organs of the immune system (Kwak et al., 1999). Arginine supplementation modu- © 2014 Poultry Science Association Inc. lates or boosts humoral and cellular immune response, Received October 30, 2013. Accepted December 18, 2013. and improves nitric oxide (NO) production and the 1 Corresponding author: [email protected] acute phase inflammatory response following lipopoly-

882 IMMUNE FUNCTION AFTER SALMONELLA CHALLENGE 883 saccharide (LPS) injections, and NO is an important antibody titers, and Salmonella concentrations in the component of the macrophage defense against Salmo- ceca, after an experimental challenge in young broiler nella (Moncada et al., 1991; Xie et al., 1994). Vitamin chickens. E has been shown to improve the immune response in birds by enhancing macrophage phagocytic function, MATERIALS AND METHODS decreasing prostaglandin E2 production, increasing IL-1 secretion by macrophages, and enhancing IL-2 produc- Birds and Treatments tion and T-cell proliferation (Moriguchi et al., 1993). Erf et al. (1998) reported that the T-helper/T-cyto- Two experiments were conducted. One-day-old broil- toxic lymphocyte ratio and the percentages of T-helper er chicks (Cobb 500) were obtained from a local hatch- lymphocytes in the spleen and thymus were higher in ery and housed in a biosafety level 2 facility at the Downloaded from broilers fed diets with high levels of VE. Prebiotics are USDA-Southern Plains Agriculture Research Center in nondigestible feed ingredients that beneficially affect College Station, Texas. Chicks were grown on fresh pine the host by selectively stimulating the growth or meta- shavings and brooded following standard temperature bolic activity of a limited number of intestinal micro- regimens, which gradually decreased from 32 to 24°C, organisms (Gibson and Roberfroid, 1995; Macfarlane and under a 16L:8D cycle throughout the studies. All http://ps.oxfordjournals.org/ et al., 2006). Mannanoligosaccharides (MOS), a prebi- birds were fed a corn-soybean meal-based diet formu- otic derived from yeasts cells, have been shown to have lated to meet or exceed all of the NRC (1994) require- antimicrobial properties, modulate intestinal microbial ments. The basal diet contained 23% of CP, 3,200 kcal populations, and stimulate immune response in several of ME/kg, 1.54% Arg, and 40.3 IU of VE/kg. A com- animal species, including poultry (Spring et al., 2000; pletely randomized design with 4 dietary treatments Fairchild et al., 2001; Fernandez et al., 2002; Denev (40 birds/treatment) was used, as follows: basal diet

et al., 2005). Spring et al. (2000) suggested that MOS free of antibiotics (negative control diet, CTL−), basal at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 reduced Salmonella colonization in the ceca of chickens diet plus 40 mg of bacitracin/kg of feed (positive con- by adsorbing bacteria and keeping them from adher- trol, CTL+), basal diet free of antibiotics supplementing to the gut wall. We previously reported that MOS ed with Arg (0.8%, wt/wt) and 40 IU of VE/kg of feed supplementation increased the populations of beneficial (AVE; for a total content of 2.34% Arg and 80 IU of bacteria (bifidobacteria and lactobacilli), and reduced VE/kg of feed), or AVE diet plus 0.2% (wt/wt) MOS Escherichia coli shedding in the litter (Baurhoo et al., (AVM; BioMos, Alltech Co., Lexington, KY). All the 2007a), showing the capacity of MOS to modulate in- experimental procedures were approved by the Institu- testinal microbial populations. More recently, it was tional Animal Care Committee. reported that birds fed MOS had an increased innate immune response (upregulation of IL-3, and upregula- Bacterial Challenge and Cecal Counts tion of signal transducer and activator of transcription 2 gene, which increases macrophage phagocytic activ- A primary poultry isolate of Salmonella enterica se- ity) than birds fed virginiamycin after a challenge with rovar Enteritidis or Typhimurium was obtained from Salmonella LPS (Baurhoo et al., 2012). the National Veterinary Services Laboratory (Ames, Although the effects of Arg, VE, and prebiotics, on IA) and selected for resistance to novobiocin and nali- immune response have been investigated, their com- dixic acid, and maintained in media containing 25 μg of bined effects on the immune response and resistance to novobiocin and 20 μg of nalidixic acid/mL. Portions (1 bacterial challenges have not been evaluated. We have to 2 mL) of cultures grown overnight at 37°C in tryptic documented that the combination of high levels of Arg soy broth (Difco Laboratories, Detroit, MI) were used and VE have complementary or synergistic effects on as inocula for challenging broilers. At d 7 (experiment the immune response against several immune challeng- 1) or d 3 (experiment 2), all the chicks were orally es, suggesting that the combination of Arg and VE may challenged with 106 cfu of Salmonella Typhimurium (1 improve the health of broilers and potentially eliminate mL/chick). Ten birds per treatment were euthanized the need for antimicrobials (Abdukalykova and Ruiz- by cervical dislocation at 3, 10, and 14 d after chal- Feria, 2006; Abdukalykova et al., 2008; Ruiz-Feria and lenge (experiment 1) or 7, 14, and 21 d after challenge Abdukalykova, 2009; Perez-Carbajal et al., 2010; Chan- (experiment 2). Cecal contents (0.25 g) were asepti- Díaz et al., 2012). We have also reported that MOS cally collected and deposited in 2.25 mL of PBS (pH supplementation improved intestinal integrity and re- 6.5). Samples were thoroughly mixed, serially diluted duced E. coli colonization in the ceca of broiler chickens in PBS, and then spread plated onto XLT4 Agar (BD (Baurhoo et al., 2007b). Thus, we hypothesized that Diagnostics, Sparks, MD) containing 20 μg of nalix- the concurrent supplementation of Arg, VE, and pre- idic acid/mL and 25 μg of novobiocin/mL, at a final biotics (MOS) will improve the immune response and dilution of 1:100, 1:1,000, and 1:10,000. Plates were in- will reduce Salmonella levels in the ceca. The objectives cubated for at least 24 h at 37°C, and the number of of the present experiments were to evaluate the effects Salmonella cfu per gram was determined. Salmonella of Arg, VE, and MOS on monocyte and heterophil oxi- colony numbers were log-transformed before being ana- dative burst, lymphocyte proliferation in vitro, serum lyzed. For qualitative enrichment, cecal contents were 884 LIU ET AL. enriched in Rappaport-Vassiliadis Broth (Difco) for 24 (0.2 mg/mL) was added to monocytes and heterophils, h at 37°C, and were then struck on XLT4 agar plates. mixed thoroughly, and aliquotted into a clear 96-well After 24 h of incubation at 42°C, the plates were read flat-bottomed plate. Oxidative burst was then mea- as positive or negative. Cecal contents that were nega- sured at an excitation/emission wavelength of 485/535 tive at a 100-fold dilution on XLT4 agar plates but nm. were positive on the enriched solution were assigned a 1.50 log10 Salmonella/g of cecal content (Byrd et al., Lymphocyte Proliferation 1998; Kubena et al., 2001; Kogut et al., 2010). A modified cell proliferation assay was conducted to evaluate cell-mediated immunity (Pauly et al., 1973).

Blood Collection Downloaded from After isolation, heterophils and monocytes were depos- Blood samples were taken from a jugular vein of ited into the plates. Then equal volumes of the lympho- chickens 5 and 9 d (experiment 1) or 7 and 14 d (experi- cyte suspension and concanavalin A were added in each ment 2) after challenge. In each experiment, 10 birds of 6 wells of 96-well flat-bottomed tissue culture plates. were sampled before being euthanized for Salmonella Negative controls were maintained with RPMI only counting, until 180 mL of blood were collected from and cell suspension without concanavalin A. Plates http://ps.oxfordjournals.org/ each treatment using EDTA (EMD Chemicals Inc., were then incubated in a water-jacketed 5% CO2 in- Gibbstown, NJ) as an anticoagulant. Blood samples cubator at 42°C for 24 h. Fifteen microliters of stock were separated evenly into four 50-mL tubes and kept 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium on ice until use for oxidative burst and lymphocyte bromide (5 mg/mL) was added to each well and plates proliferation assays. were incubated for 4 h. The color change was measured at 450 nm by a colorimetric plate reader (Sunrise,

Isolation of Heterophils and Monocytes Tecan, Austria). at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 The polymorphonuclear and mononuclear cell frac- ELISA tions were isolated from peripheral blood as described previously (Kogut et al., 1995; Stringfellow et al., 2011). The concentration of IgA, IgG, and IgM isotypes Briefly, the blood samples were mixed with a solution from 19-d-old chickens (16 d after challenge) were mea- of 1% methylcellulose (Sigma-Aldrich, St. Louis, MO) sured in serum samples (8/treatment, experiment 2). dissolved in Roswell Park Memorial Institute (RPMI) Blood samples (2 mL) were obtained by venipuncture 1640 media (Mediatech Inc., Herndon, VA) at a 1:1.5 and allowed to clot at room temperature for 4 h, sam- ratio and centrifuged at 37 × g for 15 min at 4°C. The ples were then centrifuged at 400 × g for 8 min at 4°C supernatant was removed and mixed with an equal vol- and the serum was collected and stored at −80°C until ume of calcium and magnesium free Hank’s balanced assayed. The serum samples were thawed at 4°C to cal- salt solution (Mediatech Inc.). This suspension was culate antibody concentrations using chicken IgA, IgG, layered over a 1.077/1.119 Histopaque (Sigma-Aldrich) and IgM ELISA kits (Bethyl Laboratories, Montgom- gradient and centrifuged at 235 × g for 60 min at 4°C. ery, TX) according to the manufacturer’s instructions. Following centrifugation, the interfaces containing Briefly, flat-bottomed microtiter plates were coated for monocytes or heterophils were collected, washed, and 60 min with capture antibody (goat anti-chicken IgG-Fc resuspended with RPMI. The cells were counted us- or IgM or IgA affinity purified) and coating buffer (0.05 ing a Neubauer hemacytometer and the concentration M carbonate-bicarbonate, pH 9.6). Plates were washed adjusted to 4 × 106 heterophils or monocytes/mL and 3 times with wash solution (50 mM Tris buffered saline, kept on ice until used. Cell viability (>95%) was deter- 0.14 M NaCl, 0.05% Tween 20, pH 8.0), and wells were mined using a trypan blue solution (Sigma-Aldrich) at incubated with blocking (postcoat) solution containing a 1:1 ratio. 50 mM Tris buffered saline, 0.14 M NaCl, 1% BSA, at pH 8.0 during 30 min, then rinsed 3 times with wash Oxidative Burst Assay solution. The calibrator (chicken reference serum) and sample/conjugate diluent (50 mM Tris buffered saline, Oxidative burst activity of heterophils and mono- 0.14 M NaCl, 1% BSA, 0.05% Tween 20, pH 8.0) were cytes was measured using a Wallac fluorescent plate used as standards, whereas serum samples, thawed at reader (Perkin Elmer, Boston, MA), using DCF-DA 4°C overnight, were diluted at 1:1,000 in sample/con- (Molecular Probes Inc., Eugene, OR) as reactive oxy- jugate diluent. Then, they were incubated in wells for gen species indicator, as described previously (Xie et 60 min, and washed 5 times with wash solution. Horse- al., 2002; Stringfellow et al., 2011). Briefly, heterophils radish peroxidase detection antibody (goat anti-chicken and monocytes were preincubated with an agonist (2 IgG-Fc or IgA or IgM) diluted in sample/conjugate di- μg of phorbol-12-myristate-13 acetate (Calbiochem, La luent was added to wells, incubated for 60 min, and Jolla, CA) or an equivalent volume of RPMI during rinsed 5 times with wash solution. Enzyme substrate 30 min at 42°C in a heated orbital shaker plate (Ther- (3,3c,5,5c-tetramethylbenzidine peroxidase substrate mo-Forma, Marietta, OH). Then, 125 μL of DCF-DA and peroxidase solution B) was added and incubated IMMUNE FUNCTION AFTER SALMONELLA CHALLENGE 885 had intermediate MOB values (Figure 1). In experiment 2, the MOB was not different among treatments 7 d postinfection; however, 14 d postinfection birds fed the AVE diet had the highest MOB (P < 0.05) compared with birds in the other treatments (Figure 1). The heterophil oxidative burst (HOB) in experiment 1 was lowest in birds fed the AVE diet, and highest in birds fed the CTL− diet, 5 d postinfection (P < 0.05; Figure 2). At 9 d postinfection, birds fed the control diets (CTL− or CTL+) had significantly higher HOB than birds fed AVE or AVM diet (Figure 2). In experi- Downloaded from ment 2, birds fed the CTL− or the AVE diet had higher (P < 0.05) HOB than birds fed the CTL+ or the AVM diet 7 d postinfection; however, 14 d postinfection birds fed the AVE diet had the highest HOB (P < 0.05). The lymphocyte proliferation (LPA) was signifi- http://ps.oxfordjournals.org/ cantly lower in birds fed the AVE diet compared with the other treatments 5 d postinfection in experiment 1; however, 9 d postinfection birds fed the AVE or AVM Figure 1. In vitro oxidative burst of monocytes (n = 4 replicates per column) isolated from the pooled peripheral blood of broiler chick- diet had significantly higher LPA than birds fed the ens (n = 10 per treatment) at 5 and 9 (experiment 1) or at 7 and 14 CTL− or the CTL+ diets (Figure 3). In experiment (experiment 2) d postinfection with 106 cfu of Salmonella Typhimuri- 2, birds fed the AVE diet had the highest (P < 0.05) um (birds were orally challenged at d 7 in experiment 1 or at d 3 in LPA at both times (7 and 14 d postinfection); in the at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 experiment 2). Chickens were fed an antibiotic-free diet (CTL−), an antibiotic-free diet plus 40 mg of bacitracin/kg of feed (CTL+), a same experiment, birds fed the CTL+ diet had signifi- CTL− plus 0.8% Arg and 40 IU of VE/kg of feed (AVE), or an AVE cantly higher LPA than birds fed the CTL− diet 7 d diet plus 0.2% mannanoligosaccharides (AVM). Monocytes were stim- postinfection, but by d 14 postinfection there were no ulated with phorbol-12-myristate-13-acetate or cultured in an equal volume of Roswell Park Memorial Institute 1640 media as a negative differences in LPA among birds fed the CTL−, CTL+, control. Columns with different letters (a–c) within sampling time are or AVM diet. statistically different (P < 0.05). for 15 min (IgA or IgM) or 30 min (IgG). Finally, 2 M H2SO4 was used to stop the 3,3c,5,5c-tetramethylbenzidine reaction. The absorbance, measured at 450 nm with a microtiter plate reader (Wallac Victor-2 1420 Multilabel Counter), was used to calculate the immunoglobulins (IgG, IgA, or IgM) concentration with a 4 parameter logistic curve-fit.

Statistical Analysis Data were analyzed as a completely randomized design using a one-way ANOVA using the SigmaStat software (Jandel Scientific, 1994). Means were separated using the Student-Newman-Keuls method, and significance was declared at P < 0.05. All data are presented as the least squares means ± SE.

RESULTS Figure 2. In vitro heterophil oxidative burst (n = 4 replicates per column) isolated from the pooled peripheral blood of broiler chickens Immune Function Parameters (n = 10 per treatment) at 5 and 9 (experiment 1) or at 7 and 14 (experiment 2) d postinfection with 106 cfu of Salmonella Typhimurium In experiment 1, the monocyte oxidative burst (birds were orally challenged at d 7 in experiment 1 or at d 3 in experiment 2). Chickens were fed an antibiotic-free diet (CTL−), an (MOB) was higher in birds fed the CTL− or the AVE antibiotic-free diet plus 40 mg of bacitracin/kg of feed (CTL+), a diet than in birds fed the CTL+ or AVM diet 5 d (P CTL− plus 0.8% Arg and 40 IU of VE/kg of feed (AVE), or an AVE < 0.05) postinfection (Figure 1); however, 9 d postin- diet plus 0.2% mannanoligosaccharides (AVM). Heterophils were stim- fection, the MOB was highest (P < 0.05) in birds fed ulated with phorbol-12-myristate-13-acetate or cultured in an equal volume of Roswell Park Memorial Institute 1640 media as a negative the AVM diet, and lowest (P < 0.05) in birds fed the control. Columns with different letters (a–c) within sampling time are CTL+ diet, whereas birds fed the CTL− or AVE diet statistically different (P < 0.05). 886 LIU ET AL.

Table 1. Effects of arginine, vitamin E, and mannanoligosaccharides on serum antibody titers 16 d after an experimental challenge with 106 cfu of Salmonella Typhimurium in broiler chickens1

Treatment2,3

Isotype CTL− CTL+ AVE AVM IgM 121.3 ± 17.1ab 97.6 ± 9.2b 116.2 ± 15.6ab 152.3 ± 11.0a IgG 1,249.3 ± 229.5 936.0 ± 143.3 943.6 ± 81.3 674.3 ± 65.0 IgA 127.1 ± 38.1b 97.3 ± 10.9b 138.3 ± 37.3b 310.3 ± 68.7a a,bValues with different superscripts within the same row are different (P < 0.05). 1Mean of 8 observations ± SEM.

2CTL−, antibiotic-free diet; CTL+, diet with 40 mg of bacitracin/kg of feed; AVE, antibiotic-free diet plus Downloaded from 0.8% l-Arg, and 40 IU of vitamin E/kg of feed; AVM, AVE diet supplemented with 0.2% mannanoligosaccharides (Biomos, Alltech Inc., Nicholasville, KY). 3Birds were orally challenged at d 3 (experiment 2).

Humoral Immunity Salmonella Concentrations in the Ceca http://ps.oxfordjournals.org/ The amount of circulating IgM was significantly In experiment 1, the amount of recovered Salmonella higher in birds fed the AVM diet than in birds fed the (cfu) was significantly lower in birds fed the AVM diet CTL+ diet, but similar to the IgM concentrations of than in birds fed the CTL− diet 3 d postinfection, but birds fed the CTL− or the AVE diet (Table 1). The was not different from the Salmonella concentrations amount of IgG was not affected by dietary treatment. recovered in birds fed the CTL+ or AVE diet. How- Conversely, birds fed the AVM diet had the highest (P ever, 10 and 17 d postinfection, there was no effect of < 0.05) concentrations of circulating IgA, with no dif- dietary treatment on Salmonella recovery in the ceca. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 ferences among the other 3 treatments. In experiment 2, there were no effects of treatment on Salmonella at any measuring time (7, 14, and 21 d postinfection).

DISCUSSION

In previous studies we have documented that the supplementation of Arg and VE has complementary or synergistic effects on the immune response of broilers after vaccination with immune bursal disease virus or after challenges with Eimeria spp. (Abdukalykova et al., 2008; Ruiz-Feria and Abdukalykova, 2009; Perez- Carbajal et al., 2010; Chan-Díaz et al., 2012). We have also documented that birds fed MOS had an improved intestinal integrity, an increase in beneficial bacteria in the ceca, and reduced shedding of E. coli after challenge (Baurhoo et al., 2007a,b). Thus, we hypothesized that the concurrent supplementation of Arg, VE, and MOS, will improve the innate and acquired immune function and will reduce Salmonella colonization in the ceca of broiler chickens after a challenge with Salmonella Ty- phimurium. In experiment 1, when the birds were challenged at 7 d of age, the effects of Arg and VE (AVE) or the combination of AVE and MOS (AVM) on the activ- Figure 3. In vitro proliferation of lymphocytes (n = 4 replicates ity of MOB were not very consistent, but feeding AVE per column) isolated from the pooled peripheral blood of broiler chickens (n = 10 per treatment) at 5 and 9 (experiment 1) or at 7 and 14 improved the MOB 5 d after challenge, and feeding the (experiment 2) d postinfection with 106 cfu of Salmonella Typhimuri- AVM diet improved the MOB 9 d after challenge com- um (birds were orally challenged at d 7 in experiment 1 or at d 3 in pared with the CTL+ treatment. Also, birds fed the experiment 2). Chickens were fed an antibiotic-free diet (CTL−), an CTL− diet had consistently higher (P < 0.05) MOB antibiotic-free diet plus 40 mg of bacitracin/kg of feed (CTL+), a CTL− plus 0.8% Arg and 40 IU of VE/kg of feed (AVE), or an AVE than birds fed the CTL+ diet. Thus, the higher MOB diet plus 0.2% mannanoligosaccharides (AVM). Lymphocytes were in birds fed the CTL− and the AVE diet 5 d after stimulated with concanavalin A or cultured in an equal volume of challenge could be due to a higher immune stimulation Roswell Park Memorial Institute 1640 media as a negative control. Columns with different letters (a–c) within sampling time are statisti- in the absence of antimicrobial agents or prebiotics; in cally different (P < 0.05). other words, the presence of antibiotics or prebiotics IMMUNE FUNCTION AFTER SALMONELLA CHALLENGE 887 may reduce the interaction of microbes with the gastric significantly higher MOB, HOB, and LPR, than birds immune system, resulting in a downregulation of the in the other treatments. It has been reported that neo- MOB capacity. However, birds fed the AVM diet had natal poultry exhibit a transient susceptibility to infec- the highest MOB 9 d after challenge, which could be tious disease during the first week of life and this is at- attributed to the immunomodulatory effects of MOS tributed to a deficient host defense mechanism (Kogut (Ferket, 2004; Janardhana et al., 2009). Because mono- et al., 2012), in particular a functional inefficiency of cytes are important in phagocytosis and activation of heterophils and monocytes (He et al., 2008); thus, it is the acquired immune response through the production of great significance that the supplementation of Arg of cytokines, our results show that Arg and VE supple- and VE increased the oxidative burst of both mono- mentation (AVE) are better than AVM diets, and that cytes and heterophils when the challenge was done at an the presence of a prebiotic may delay the activation of early age. Also, the LPR was consistently improved in Downloaded from MOB capacity. birds fed the AVE diet when the challenge was done at Birds fed the AVE or the AVM diet had consistently an early age. The effects of VE on the immune response lower HOB than birds fed the control diets 9 d after are mediated by direct effects on lymphocyte cytokine challenge in experiment 1. Genovese et al. (2013) re- expression and modulation of immune cell proliferation ported that chickens with less functional heterophils (Li-Weber et al., 2002; Leshchinsky and Klasing, 2003; http://ps.oxfordjournals.org/ are more susceptible to infection, including Salmonella Babu and Raybourne, 2008), whereas the effects of Arg Enteritidis, than those with highly functional hetero- on the immune response to Salmonella have been at- phils. Thus, under these conditions the supplementa- tributed to a higher production of NO for host defense tion of AVE or AVM does not seem to have beneficial (MacFarlane et al., 1999). Also, macrophages from effects in term of heterophil function. Conversely, birds chickens selected for high antibody response produce fed the AVE or the AVM diet had a significantly higher more NO (Guimarães et al., 2011). However, the effects

lymphocyte proliferation (LPR) reaction than birds of Arg alone can be deleterious because an overproduc- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 fed the CTL diets 9 d after challenge. Thus, when birds tion of NO may lead to increased oxidative stress and are challenged with Salmonella Typhimurium at 7 d of immune suppression (Huang et al., 1996; MacFarlane age and are fed diets supplemented with Arg and VE et al., 1999; Xie et al., 2008). Thus, the complementary (AVE), or a combination of Arg, VE, and MOS (AVM) effects of Arg and VE may be explained by the antioxi- they show a reduced (P < 0.05) HOB, and higher LPR dant effects of VE on the overproduction of NO elic- (P < 0.05) than birds fed the control diets, although ited by the immune challenge. In these experiments, we the differences were modest and may not have a major did not test Arg and VE individually because we have biological significance (Figures 2 and 3). Also, it was documented their complementary effects in several im- evident that further supplementing an AVE diet with mune challenge models (Abdukalykova and Ruiz-Feria, MOS did not have additive effects on immune function 2006; Abdukalykova et al., 2008; Ruiz-Feria and Abdu- under the conditions of this experiment. The limited kalykova, 2009; Perez-Carbajal et al., 2010; Chan-Díaz effects of the dietary treatments on immune function et al., 2012). could be explained by the fact that birds were chal- Also in these experiments it was evident that fur- lenged at 7 d of age, when the normal microflora was ther supplementing an AVE diet with MOS did not already established, and under these circumstances the further improve immune response in experiment 1 (ex- Salmonella challenge was not enough to elicit a strong cept MOB 9 d after challenge), and even reversed the immune response. Redmond et al. (2011) found no ef- positive effects of Arg and VE on immune response in fect of dietary β-glucans or ascorbic acid on hetero- experiment 2. The reason for this could be attributed phil function after a Salmonella Typhimurium chal- to the mode of action of MOS, which binds gram-neg- lenge, and they attributed this result to the lack of ative bacteria expressing the type 1 fimbrae, including other stressors or disease, suggesting that under most Salmonella, reducing its adherence to intestinal epithe- challenging environments the effects of prebiotics and lial cells (Spring et al., 2000) and immune activation. antioxidants would be more evident. Recently, Baurhoo et al. (2012) reported that birds fed Thus, to avoid the effects of an established microflo- MOS and challenged with Salmonella LPS showed gene ra, in experiment 2 the birds were challenged at 3 d of expression downregulation of Toll-like receptor-2, gal- age, and the effects of the concurrent supplementation linacin-1 α, and CXC1, an IL-8 receptor that binds of Arg and VE on immune response were more consis- the IL-8 chemoattractant expressed by macrophages, tent, and in agreement with our previous findings (Ab- whereas Wigley et al. (2006) reported that Salmonella- dukalykova et al., 2008; Ruiz-Feria and Abdukalykova, resistant lines of chickens presented a higher expression 2009; Perez-Carbajal et al., 2010; Chan-Díaz et al., of IL-6, IL-8, and IL-18, and Kogut (2002) reported the 2012). Seven days after challenge, MOB was not differ- importance of IL-8 as a major chemotactic factor to in- ent among treatments, whereas at this time birds fed crease heterophil recruitment to the site of Salmonella the AVE diet had significantly higher HOB than birds infection; thus, it appears that downregulation of IL-8 fed the CTL+ or the AVM diet, and higher (P < 0.05) by MOS may explain the effects seen in these experi- LPR than birds fed the other diets. Furthermore, 14 d ments. Our results agree with reports indicating that after challenge birds fed the AVE diet had consistently prebiotics do not have an effect on monocyte function 888 LIU ET AL.

Table 2. Effects of Arg, vitamin E, and mannanoligosaccharides on cecal Salmonella Typhimurium (cfu, log10), and on number of Salmonella-positive samples (out of 10 samples, in parentheses) after an experimental challenge with 106 cfu of Salmonella Typhimurium in broiler chickens1

Treatment2 Days after challenge3 CTL− CTL+ AVE AVM Experiment 1 3 3.0 ± 0.5a (8) 2.1 ± 0.5ab (5) 1.6 ± 0.5ab (3) 0.8 ± 0.4b (1) 10 0.8 ± 0.4 (1) 1.5 ± 0.4 (1) 0.2 ± 0.2 (0) 0.7 ± 0.4 (0) 17 0.5 ± 0.4 (1) 0.7 ± 0.3 (0) 1.0 ± 0.6 (2) 0.6 ± 0.4 (1) Experiment 2 7 4.4 ± 0.6 (10) 4.0 ± 1.1 (10) 4.4 ± 1.1 (10) 4.4 ± 1.14 (9) Downloaded from 14 1.5 ± 0.5 (7) 1.9 ± 0.6 (2) 1.9 ± 0.6 (7) 0.9 ± 0.6 (7) 21 0.3 ± 0.3 (1) 0 ± 0 (0) 0.8 ± 0.4 (0) 0.4 ± 0.4 (0) a,bValues with different superscripts within the same row are different (P < 0.05). 1Mean of 10 observations ± SEM (number of birds tested positive, out of 10, after Rappaport-Vassiliadis broth media enrichment for 24 h). 2CTL−, antibiotic-free diet; CTL+, diet with 40 mg of bacitracin/kg of feed; AVE, antibiotic-free diet plus http://ps.oxfordjournals.org/ 0.8% l-Arg, and 40 IU of vitamin E/kg of feed; AVM, AVE diet supplemented with 0.2% mannanoligosaccharides (Biomos, Alltech Inc., Nicholasville, KY). 3Birds were orally challenged at d 7 (experiment 1) or d 3 (experiment 2).

(Bunout et al., 2002) or leukocyte proliferation (Ja- no significant differences among treatments. Therefore, nardhana et al., 2009). the higher immune response elicited by Arg and VE

In experiment 2, the dietary treatments did not af- in experiment 2 was not correlated with a lower Sal- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 fect antibody titers of IgG, but the antibody titers of monella recovery in the ceca compared with birds fed IgM were significantly higher in birds fed the AVM diet the control diets; this could be attributed to a low dose compared with birds fed the CTL+ diet, and IgA titers of Salmonella in the challenge, or to the lack of other were highest (P < 0.05) in birds fed the AVM diet. stresses in otherwise healthy birds. Also, it has been These results are in contrast with the effects of MOS on suggested that the expression of cytokines and chemo- the innate immune response and LPR discussed above. kines during primary Salmonella infection will result It is well documented that infection with Salmonella in the modulation of CD4 T cell response to improve leads to increased levels of IgG, IgM, and IgA anti- resistance against secondary infection (Withanage et bodies (Withanage et al., 2005). However, the role of al., 2005; Nanton et al., 2012); accordingly, the higher humoral immunity on Salmonella clearance is not well HOB and MOB found in this experiment, along with understood. Beal et al. (2006) demonstrated that al- our previous findings on the effects of Arg and VE on though Salmonella infection in chickens elicited high T cell differentiation (Abdukalykova et al., 2008) may concentrations of antibodies, B cells did not play an be important in the development of acquired immunity essential role in clearance of the primary or secondary after a secondary challenge. Further research is war- infection. More recently, Nanton et al. (2012) suggested ranted on this subject. that B cells play an indirect role in protection against In the same way, the higher concentrations of IgA Salmonella through the development of T cell immunity in birds fed the AVM diet did not correlate with low- after secondary infection. Thus, the higher concentra- er concentrations of Salmonella in birds fed the AVM tions of IgA in birds fed the AVM diet may be benefi- diet. The effects of prebiotics on Salmonella prevention cial in a secondary challenge, or may reduce horizontal have been variable. Agunos et al. (2007) reported re- transmission of Salmonella during growing. ductions in fecal Salmonella 16 and 19 d after a chal- However, in these experiments the dietary treat- lenge with 2 × 107 cfu of Salmonella when birds were ments did not affect Salmonella recovery in the ceca fed β1–4-mannobiose, whereas Revolledo et al. (2009) after challenge. Only in experiment 1, when birds were did not find any effect on ceca Salmonella in birds fed challenged at d 7, birds fed the AVM diet had lower (P β-glucans. In the same way, Burkey et al. (2004) and < 0.05) concentrations of Salmonella than birds fed the Calveyra et al. (2011) reported no effects of MOS on CTL− diet, but not different from the ceca Salmonella Salmonella shedding in pigs experimentally infected levels of birds fed the AVE or the CTL+ diet 3 d after with Salmonella Typhimurium. challenge (Table 2). At 10 and 17 d after challenge, In summary, our results show that supplementing di- the recovery of Salmonella was very low and not dif- ets with Arg and VE improve the innate and acquired ferent among treatments. In experiment 2, when the immune response in chicks challenged with Salmonella, challenge was performed at an earlier age (d 3) Sal- but only when the challenge occurred early in their life monella recovery was high in all treatments 7 d after (3 d old). We also found that further supplementing challenge (4–4.4 logs), was reduced by more than half the diet with MOS did not improve the effects of Arg by 14 d after challenge (0.9–1.9 logs), and was very and VE, except in the production of IgA. Although low by 21 d after challenge (0–0.8 logs), but there were the dietary treatments did not reduce Salmonella con- IMMUNE FUNCTION AFTER SALMONELLA CHALLENGE 889 centrations in the ceca, the better immune response of Erf, G. F., W. G. Bottje, T. K. Bersi, M. D. Headrick, and C. A. birds fed Arg and VE may benefit birds grown under Fritts. 1998. Effects of dietary vitamin E on the immune system in broilers: Altered proportions of CD4 T cells in the thymus and commercial conditions, where they face more stressful spleen. Poult. Sci. 77:529–537. conditions and are exposed to several microbial chal- Fairchild, A. S., J. L. Grimes, F. T. Jones, M. J. Wineland, F. W. lenges. Further studies are needed to evaluate the ef- Edens, and A. E. Sefton. 2001. Effects of hen age, Bio-Mos and fects of Arg and VE on Salmonella clearance after a Flavomycin on poult susceptibility to oral Escherichia coli challenge. Poult. Sci. 80:562–571. secondary Salmonella challenge. FDA. 2013. 2011 summary report on antimicrobials sold or distributed for use in food-producing animals. Accessed June 2013. http://www.fda.gov/AnimalVeterinary/NewsEvents/ REFERENCES CVMUpdates/ucm338178.htm.

Ferket, P. R. 2004. Alternatives to antibiotics in poultry production: Downloaded from Abdukalykova, S., and C. A. Ruiz-Feria. 2006. Arginine and vitamin Responses, practical experience and recommendations. Pages 57– E improve the cellular and humoral immune response of broiler 68 in Nutritional Biotechnology in the Feed and Food Industries. chickens. Int. J. Poult. Sci. 5:121–127. T. P. Lyons and K. A. Jacques, ed. Nottingham University Press, Abdukalykova, S. T., X. Zhao, and C. A. Ruiz-Feria. 2008. Arginine Nottingham, UK. and vitamin E modulate the subpopulations of T lymphocytes in Fernandez, F., M. Hinton, and B. Van Gils. 2002. Dietary mannan- broiler chickens. Poult. Sci. 87:50–55. oligosaccharides and their effect on chicken caecal microflora in Agunos, A., M. Ibuki, F. Yokomizo, and Y. Mine. 2007. Effects of di- relation to Salmonella Enteritidis colonization. Avian Pathol. http://ps.oxfordjournals.org/ etary β1–4 mannobiose in the prevention of Salmonella enteritidis 31:49–58. infection in broilers. Br. Poult. Sci. 48:331–341. Genovese, K. J., H. He, C. L. Swaggerty, and M. H. Kogut. 2013. Babu, U. S., and R. B. Raybourne. 2008. Impact of dietary compo- The avian heterophil. Dev. Comp. Immunol. http://dx.doi. nents on chicken immune system and Salmonella infection. Ex- org/10.1016/j.dci.2013.03.021. pert Rev. Anti Infect. Ther. 6:121–135. Gibson, G. R., and M. B. Roberfroid. 1995. Dietary modulation of Bach Knudsen, K. E. 2001. Development of antibiotic resistance and the human colonic microflora: Introducing the concept of prebiot- options to replace antimicrobials in animal diets. Proc. Nutr. ics. J. Nutr. 125:1401–1412. Soc. 60:291–299. Guimarães, M. C., L. V. Guillermo, M. F. Matta, S. G. Soares,

Baurhoo, B., P. Ferket, C. H. Ashwell, J. de Oliveira, and X. Zhao. and R. A. DaMatta. 2011. Macrophages from chickens select- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 2012. Cell walls of Saccharomyces cerevisiae differentially modu- ed for high antibody response produced more nitric oxide and late innate immunity and glucose metabolism during late sys- have greater phagocytic capacity. Vet. Immunol. Immunopathol. temic inflammation. PLoS ONE 7:e30323. http://dx.doi. 140:317–322. org/101371/journal.pone.0030323. Harrison, E. M., G. K. Paterson, M. T. Holden, J. Larsen, M. Steg- Baurhoo, B., A. Letellier, X. Zhao, and C. A. Ruiz-Feria. 2007b. ger, A. R. Larsen, A. Petersen, R. L. Skov, J. M. Christensen, A. Cecal populations of Lactobacilli and Bifidobacteria and E. coli B. Zeuthen, O. Heltberg, S. R. Harris, R. N. Zadoks, J. Parkhill, populations after in vivo challenge in birds fed diets with puri- S. J. Peacock, and M. A. Holmes. 2013. Whole genome sequenc- fied lignin or mannanoligasaccharides. Poult. Sci. 86:2509–2516. ing identifies zoonotic transmission of MRSA isolates with the Baurhoo, B., L. Phillip, and C. A. Ruiz-Feria. 2007a. Effects of pu- novel mecA homologue mecC. EMBO Mol. Med. 5:509–515. rified lignin and mannanoligosaccharides on intestinal integrity He, H., K. J. Genovese, C. L. Swaggerty, D. J. Nisbet, and M. H. and microbial populations in the ceca and litter of broiler chick- Kogut. 2008. Differential induction of nitric oxide, degranulation, ens. Poult. Sci. 86:1070–1078. and oxidative burst activities in response to microbial agonist Beal, R. K., P. Wigley, C. Powers, P. A. Barrow, and A. L. Smith. stimulations in monocytes and heterophils from young commer- 2006. Cross-reactive cellular and humoral immune responses to cial turkeys. Vet. Immunol. Immunopathol. 123:177–185. Salmonella enterica serovars Typhimurium and Enteritidis are Huang, D., M. G. Schwacha, and T. K. Eisenstein. 1996. Attenu- associated with protection to heterologous re-challenge. Vet. Im- ated Salmonella vaccine-induced suppresion of murine spleen cell munol. Immunopathol. 114:84–93. responses to mitogen is mediated by macrophage nitric oxide: Bunout, D., S. Hirsh, M. P. De la Maza, C. Munoz, F. Haschke, Quantitative aspects. Infect. Immun. 64:3786–3792. P. Steenhout, P. Klassen, G. Barrera, V. Gattas, and M. Pe- Janardhana, V., M. M. Broadway, M. P. Bruce, J. W. Lowenthal, termann. 2002. Effects of prebiotics on the immune response to M. S. Geier, R. J. Hughes, and A. G. D. Bean. 2009. Prebiotics vaccination in the elderly. J. Parent. Enter. Nutr. 26:372–376. modulate immune responses in the gut-associated lymphoid tis- Burkey, T. E., S. S. Dritz, J. C. Nietfeld, B. J. Johnson, and J. E. sue of chickens. J. Nutr. 139:1404–1409. Minton. 2004. Effect of dietary mannanoligosaccharide and so- Jandel Scientific. 1994. SigmaStat Statistical Software User’s Manu- dium chlorate on the growth performance, acute-phase response, al. Jandel Scientific Software, San Rafael, CA. and bacterial shedding of weaned pigs challenged with Salmonella Kidd, M. T. 2004. Nutritional modulation of immune function in enterica serotype Typhimurium. J. Anim. Sci. 82:397–404. broilers. Poult. Sci. 83:650–657. Byrd, J. A., D. E. Corrier, J. R. Deloach, D. J. Nisbet, and L. H. Klasing, K. C. 1998. Nutritional modulation of resistance to infec- Stanker. 1998. Horizontal transmission of Salmonella Typhimuri- tious diseases. Poult. Sci. 77:1119–1125. um in broiler chicks. J. Appl. Poult. Res. 7:75–80. Kogut, M. H. 2002. Dynamics of a protective avian inflammatory Calveyra, J. C., M. G. Nogeira, J. D. Kich, L. L. Biesus, R. Vizzotto, response: The role of an IL-8-like cytokine in the recruitment of L. Berno, A. Coldebella, L. Lopes, N. Mores, G. J. Lima, and M. heterophils to the site of organ invasion by Salmonella enteritidis. Cardoso. 2011. Effect of organic acids and mannanoligosaccha- Comp. Immunol. Microbiol. Infect. Dis. 25:159–172. ride on excretion of Salmonella typhimurium in experimentally Kogut, M. H., H. He, and K. J. Genovese. 2012. Bacterial Toll-like infected growing pigs. Res. Vet. Sci. 93:46–47. receptor agonists induce sequential NF-kB-mediated leukotriene Chan-Díaz, D. J., A. Pro-Martínez, C. A. Ruíz-Feria, J. M. Cuca- B4 and prostaglandin E2 production in chicken heterophils. Vet. García, C. Narciso-Gaytan, M. E. Ramírez-Guzmán, J. Gallegos- Immunol. Immunopathol. 145:159–170. Sánchez, E. Sosa-Montes, and J. C. Rodriguez-Castillo. 2012. Kogut, M. H., H. He, K. J. Genovese, and Y. W. Jiang. 2010. Feed- Effects of arginine and vitamin E supplemented diets on the im- ing the BT cationic peptides to chickens at hatch reduces ce- munological response of broiler chickens. Tropical and Subtropi- cal colonization by Salmonella enterica serovar Enteritidis and cal Agroecosystems 15:367–374. primes innate immune cell functional activity. Foodborne Pat- Denev, S. A., I. Dinev, I. Nikiforov, and V. Koinarski. 2005. Effects hog. Dis. 7:23–30. of mannan oligosaccharides on composition of cecal microflora Kogut, M. H., E. D. McGruder, B. M. Hargis, D. E. Corrier, and and performance of broiler chickens. Pages 351–353 in 15th Eur. J. R. DeLoach. 1995. In vivo activation of heterophil function in Symp. Poult. Nut., Balatonfüred, Hungary. World’s Poult. Sci. chickens following injection with Salmonella Enteritidis-immune Assoc., Budapest, Hungary. lymphokines. J. Leukoc. Biol. 57:56–62. 890 LIU ET AL.

Kubena, L. F., J. A. Byrd, C. R. Young, and D. E. Corroer. 2001. Redmond, S. B., P. Chuammitri, C. B. Andreasen, D. Palic, and S. Effect of tannic acid on cecal volatile fatty acids and suscepti- L. Lamont. 2011. Proportion of circulating chicken heterophils bility to Salmonella typhimurium colonization in broiler chicks. and CXCLi1 expression in response to Salmonella enteriditis are Poult. Sci. 80:1293–1298. affected by genetic line and immune modulating diet. Vet. Im- Kwak, H., R. E. Austic, and R. R. Dietert. 1999. Influence of dietary munol. Immunopathol. 140:323–328. arginine concentration on lymphoid organ growth in chickens. Revolledo, L., C. S. A. Ferreira, and A. J. P. Ferreira. 2009. Pre- Poult. Sci. 78:1536–1541. vention of Salmonella Typhimurium colonization and organ in- Leshchinsky, T. V., and K. C. Klasing. 2003. Profile of chicken cy- vasion by combination treatment in broiler chicks. Poult. Sci. tokines induced by lipopolysaccharides is modulated by dietary 88:734–743. α-tocopheryl acetate. Poult. Sci. 82:1266–1273. Ruiz-Feria, C. A., and S. Abdukalykova. 2009. Arginine and vita- Li-Weber, M., M. Giaisi, M. K. Treiber, and P. H. Krammer. 2002. min E improve antibody titres to infectious bursal disease virus Vitamin E inhibits IL-4 gene expression in peripheral blood T (IBDV) and sheep red blood cells in broiler chickens. Br. Poult.

cells. Eur. J. Immunol. 32:2401–2408. Sci. 50:291–297. Downloaded from Macfarlane, S., G. T. Macfarlane, and J. H. Cummings. 2006. Re- Spring, P., C. Wenk, K. A. Dawson, and K. E. Newman. 2000. The view article: Prebiotics in the gastrointestinal tract. Aliment. effects of dietary mannanoligosaccharides on cecal parameters Pharmacol. Ther. 24:701–714. and the concentrations of enteric bacteria in the ceca of Salmo- MacFarlane, A. S., M. G. Schwacha, and T. K. Eisenstein. 1999. In nella-challenge broiler chicks. Poult. Sci. 79:205–211. vivo blockage of nitric oxide with aminoguanidine inhibits im- Stringfellow, K., D. Caldwell, J. Lee, M. Mohnl, R. Beltran, G. munosuppression induced by an attenuated strain of Salmonella Schatzmayr, S. Fitz-Coy, C. Broussard, and M. Farnell. 2011. typhimurium, potentiates Salmonella infection, and inhibits mac- Evaluation of probiotic administration on the immune response http://ps.oxfordjournals.org/ rophage and polymorphonuclear leukocyte influx into the spleen. of coccidiosis-vaccinated broilers. Poult. Sci. 90:1625–1658. Infect. Immun. 67:891–898. Wigley, P., S. Hulme, L. Rothwell, N. Bumstead, P. Keiser, and Moncada, S., R. M. Palmer, and E. A. Higgs. 1991. Nitric oxide: P. Barrow. 2006. Macrophages isolated from chickens genetically Physiology, pathophysiology, and pharmacology. Pharmacol. resistant or susceptible to systemic salmonellosis show magnitu- Rev. 43:109–142. dinal and temporal differential expression of cytokines and che- Moriguchi, S., H. Miwa, M. Okamura, K. Maekawa, Y. Kishino, mokines following Salmonella enterica challenge. Infect. Immun. and K. Maeda. 1993. Vitamin E is an important factor in T cell 74:1425–1430. differentiation in thymus of F344 rats. J. Nutr. Sci. Vitaminol. Withanage, G. S. K., P. Wigley, P. Kaiser, P. Mastrochi, H. Brooks,

(Tokyo) 39:451–463. C. Powers, R. Beal, P. Barrow, and I. McConnell. 2005. Cytokine at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Nanton, M. R., S. S. Way, M. J. Schlomchik, and S. J. McSorley. and chemokine responses associated with clearance or a primary 2012. Cutting edge: B cells are essential for protective immunity Salmonella enterica serovar Typhimurium infection in the chick- against Salmonella independent of antibody secretion. J. Immu- en and in protective immunity to rechallenge. Infect. Immun. nol. 189:5503–5507. 73:5173–5182. NRC. 1994. Nutrient Requirements of Poultry. Natl. Res. Counc., Xie, H., G. R. Huff, W. E. Huff, J. M. Balog, and N. C. Rath. 2002. Natl. Acad. Sci., Washington, DC. Effects of ovotransferrin on chicken macrophages and heterophil- Pauly, J. L., J. E. Sokal, and T. Han. 1973. Whole blood culture granulocytes. Dev. Comp. Immunol. 26:805–815. techniques for functional studies of lymphocyte reactivity to mi- Xie, Q. W., Y. Kashiwabar, and C. Nathan. 1994. Role of transcrip- togens, antigens, and homologous lymphocytes. J. Lab. Clin. tion factor NF-kB/Rel in induction of nitric oxide synthase. J. Med. 82:500–512. Biol. Chem. 269:4707–4708. Perez-Carbajal, C., D. Caldwell, M. Farnell, K. Stringfellow, G. Xie, X. Q., Y. Shinozawa, J. Sasaki, K. Takuma, S. Akaishi, S. Ya- Casco, S. Pohl, A. Pro-Martinez, and C. A. Ruiz-Feria. 2010. Im- manouchi, T. Endo, R. Nomura, M. Kobayashi, D. Kudo, and N. mune response of broiler chickens fed different levels of arginine Hojo. 2008. The effects of arginine and selective inducible nitric and vitamin E to a coccidiosis vaccine and Eimeria challenge. oxide synthase inhibitor on pathophysiology of sepsis in a CLP Poult. Sci. 89:1870–1877. model. J. Surg. Res. 146:298–303. Research Note

Exposure of juvenile Leghorn chickens to lead acetate enhances antibiotic resistance in enteric bacterial flora

M. Nisanian ,* S. D. Holladay ,* E. Karpuzoglu ,* R. P. Kerr ,* S. M. Williams,† L. Stabler ,† J. Vaun McArthur ,‡ R. Cary Tuckfield,§ and R. M. Gogal Jr. *1

* Department of Veterinary Biosciences and Diagnostic Imaging, and † Poultry Diagnostic and Research Center, College of Veterinary Medicine, University of Georgia, Athens 30602; ‡ Savannah River Ecology Laboratory, Downloaded from University of Georgia, Aiken, SC 29802; and § ECOSTATys LLC, Aiken, SC 29803

ABSTRACT Heavy metals have been implicated for terial cultures on d 0 before Pb exposure, d 7 and 14, their ability to increase antibiotic resistance in bacteria and then birds were euthanized by CO2 gas for nec- collected from polluted waters, independent of antibi- ropsy on d 14, at which time intestinal contents were http://ps.oxfordjournals.org/ otic exposure. Specific-pathogen-free Leghorn chickens also collected for bacterial cultures. Fecal swabs but were therefore given Pb acetate in the drinking water not intestinal samples from Pb-exposed birds contained to expose the enteric bacteria to Pb and to determine if isolates that had significantly elevated antibiotic resis- antibiotic resistance changed in these bacteria. Concen- tance. Some of the isolates contained bacteria that were trations of Pb used were 0.0, 0.01, 0.1, 1.0, or 10.0 mM; resistant to up to 20 antibiotics. These results suggest birds given the highest 2 concentrations showed signs of the need for repeated studies in chickens infected with

moribundity and dehydration and were removed from zoonotic pathogens. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 the study. Vent culture samples were collected for bac- Key words: lead acetate , Leghorn chicken, antimicrobial resistance 2014 Poultry Science 93 :891–897 http://dx.doi.org/ 10.3382/ps.2013-03600

INTRODUCTION Lead is a common heavy metal contaminant released into the environment as a byproduct of many anthropo- Recent reports have implicated heavy metal environ- genic sources. These sources include mining and smelt- mental contaminants as selective agents for indirectly ing of ore, manufacture of Pb-containing products, increasing microbial antibiotic resistance (McArthur combustion of coal and oil, and waste incineration. Due and Tuckfield, 2000; Baker-Austin et al., 2006, 2009; to the environmental persistence of Pb, historic sources Stepanauskas et al., 2006). For instance, isolates from also contribute heavily to present contamination. These coastal waters that contained elevated Pb, Cu, Zn, Sr, included leaded gasoline, Pb-based paints, Pb solder and Cd imparted multiple increased antibiotic resis- in food cans, Pb-arsenate pesticides, and Pb shot and tance in Vibrio vulnificus, a serious human pathogen. sinkers (ATSDR, 2007). Presently, the major sources The antibiotics included doxycycline, tetracycline, ami- of Pb emissions in the air are associated with ore and noglycosides, and cephalosporins, representing those metal processing and leaded aviation fuel for piston en- routinely prescribed for human V. vulnificus infection gines (NAAQS, 2012). Resulting widespread distribu- (Baker-Austin et al., 2009). It was hypothesized that tion of Pb suggests considerable opportunity for bacte- the metal contaminants may increase antibiotic re- rial pathogen exposure. This would include intestinal sistance in microbes, independent of exposure to the bacterial flora of animals exposed to Pb through their antibiotics, by adaptive coselection of linked metal-re- diets. sistance genes and antibiotic-resistance genes (Baker- Lead exposure in poultry, specifically chickens, has Austin et al., 2006). been of interest because of occasional low level but detectable Pb in feeds or water, and the high agricultural value of chickens as a food source for humans. Chick- ens sampled by the USDA Food Safety and Inspection Service showed Pb levels ranging from 25 to 671 ppb © 2014 Poultry Science Association Inc. in kidneys, 25 to 1,293 ppb in liver, and 35 to 71 ppb Received September 4, 2013. Accepted November 30, 2013. in muscle, presumably from trace Pb in feed and wa- 1 Corresponding author: [email protected] ter (Pagan-Rodriguez et al., 2007). Concern has more

891 892 NISANIAN ET AL. recently been expressed regarding Pb accumulation in Experimental Design backyard and free-range chickens in the Berkeley, Cali- fornia, area, where soil Pb levels in the range of 300 to Two chickens were randomly assigned to each of 5 600 ppm are typical. A variety of studies have shown treatments (cages). These treatments corresponded to that laying hens accumulate Pb while foraging such 5 concentrations (0.0, 0.01, 0.1, 1.0, and 10 mM) of Pb soils, and pass this metal into their eggs (Waegeneers et acetate administered ad libitum in the drinking water al., 2009). Lead in the leg muscles of 3 wild bird species for both chickens per cage for the duration of the study. was recently reported in mg/kg of wet weight as pheas- These Pb concentrations were intended to expose en- ants, 0.15 to 1.32; Eurasian coots, 0.02 to 0.63; and teric bacteria to a concentration range that produced mallards, 0.07 to 1.40, with no associated overt toxic- no overt toxicity in the birds. Each treatment regimen ity (Gasparik et al., 2010). The present chickens dosed was replicated 5 times for a total of 50 birds randomly Downloaded from with 0.01 or 0.1 mM Pb acetate had blood Pb levels of assigned among 25 cages. Birds were bled and weighed about 6 and 16 μg/dL at the end of the study, respec- on d 0 before dosing began, and again on d 7 and 14. tively. Muscle Pb levels were not determined; however, Vent culture (cloacal swab) samples were collected on d 16 μg of Pb/dL of blood corresponds to 160 μg of Pb/L 0, 7, and 14 for bacterial cultures, and then birds were of blood or approximately 0.16 mg of Pb/kg of blood, euthanized by cervical dislocation for necropsy on d http://ps.oxfordjournals.org/ a level not dissimilar to that found in muscle tissue 14. Intestinal contents were then collected for bacterial of wild birds. Wild birds may achieve toxic blood Pb cultures. levels as a consequence of mistaking Pb bullets or bul- let fragments for grit or food particles (Bennett et al., Hematological Analyses 2007), as can water-inhabiting species that filter-feed in sediments and inadvertently consume spent Pb shot Blood Collection. Blood parameters are sensitive to

(Mateo et al., 2001). Pb, and were used to monitor toxicity. On d 0, 7, and at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Chickens are a potentially important zoonotic spe- 14, blood was collected aseptically from a jugular vein cies because of close interactions with humans through of each bird, using a 25-ga needle and 1-mL syringe. handling during meat and egg production and through The needle was removed and the blood from the sy- food processing and consumption. For instance, a re- ringe was dispensed into a depressurized, heparinized cent evaluation of zoonotic transfer of Salmonella en- Vacutainer tube and mixed by gentle tapping. Blood teritidis from chicken layers to handlers showed that physiologic parameters including total red blood cell flocks with >60% infection were a high risk for handlers number, packed cell volume, mean corpuscular volume, to contract infection (El-Tras et al., 2010). The pos- blood Pb levels, delta aminolevulinic-acid dehydratase sibility has not been investigated that these bacteria assay (δ-ALAD), and total protein were performed by may show elevated antibiotic resistance if isolated from protocols as described by Gogal et al. (2002). chickens that were exposed to Pb. Should this occur, Whole-Blood Cellularity. A 10-μL aliquot of blood efficacy of antimicrobial therapies may be reduced and was transferred into a tube containing 990 μL of PBS disease threat to humans exacerbated. The present and mixed by gentle vortexing. A 10-μL aliquot of the study was conducted to determine if Pb exposure in blood sample was placed onto a hemacytometer and chickens could enhance microbial antibiotic resistance the number of red blood cells (RBC) was counted per in normal intestinal microbial flora of the chicken. milliliter. Whole-blood cellularity was enumerated and size analyzed with Nexcelom cellometer (Nexcelom Bio- science LLC, Lawrence, MA). MATERIALS AND METHODS Packed Cell Volume and Total Protein. A hematocrit capillary tube was inserted into the tube contain- Birds ing the heparinized blood sample. At 75% volume, the bottom of the capillary was inserted into a clay pad, Fifty 2-wk-old specific-pathogen-free (SPF) Leghorn centrifuged (23°C, 10,000 × g) for 5 min in a Micro chickens, a common breed for egg production, were Hematocrit centrifuge (Unico, Dayton, NJ), and the purchased from Merial Select (Gainesville, GA) and packed cell volume (PCV) determined using a graphic placed on ad libitum broiler grower medicated diet that reader card. Following analysis of the PCV, the capil- contained nondetectable Pb. Birds were housed in an lary was cut above the RBC layer, 30 μL of plasma was open-air flock house in pens sized 1.524 m wide × 3.048 placed onto a refractometer, and the total protein was m long × 2.5908 m high (5 ft × 10 ft × 8.5 ft), kept at determined. 26.7 ± 2°C, and maintained on a 12L:12D cycle. Birds Mean Corpuscular Volume. The mean corpuscular were randomly grouped into 5 treatment groups (n = volume (MCV) was calculated using the following for- 10). All procedures in this study were approved by the mula: MCV = (PCV × 10)/RBC (× 106/μL). University of Georgia’s Institutional Animal Care and Whole-Blood 5-Point Leukocyte Differential. A Use Committee before initiation, and experiments were 3-μL drop of heparinized whole blood was placed on a conducted in compliance with Good Laboratory Prac- microscope slide and a smear was made using standard tice standard operating procedures. techniques. The blood smear was allowed to air dry RESEARCH NOTE 893 and then stained with a modified Wright stain (Sigma and dilution. One hundred microliters of diluted sam- Chemical, St. Louis, MO). A 5-point differential was ple was pipetted onto the agar and a flame-sterilized performed where 100 leukocytes were counted using spreader was used to work all the solution into the agar light microscopy. for each dilution. Spread plates were incubated for 24 h Blood Lead Levels. Heparinized whole blood sam- at 37°C. Plates were examined to determine which dilu- ples (75 μL/bird) were frozen at −80°C and shipped on tion had evenly distributed colonies in suitable num- dry ice for analysis as described by D’Ilio et al. (2010). bers for picking individual colonies. į$/$'Delta aminolevulinic-acid dehydratase ac- A numbered grid was placed under the spread plate tivity is inhibited by Pb exposure in birds and is a and a random number list used to select 10 colonies highly sensitive indicator of Pb exposure (Beyer et al., from locations nearest to the chosen numbers. Selected 1988). Activity of δ-ALAD was determined by using the colonies were circled with pen on the bottom of the Downloaded from European standardized method (Berlin and Schaller, plate. The dilution series of the spread plates used was 1974). Briefly, 50 μL of heparinized whole blood, from recorded with the total number of colonies that grew on each bird, was added to a separate 5.0-mL centrifuge that plate. Two thousand isolates were obtained. tube (Fisher Scientific, Suwanee, GA) containing 1.45 Sterilized 10-mL culture tubes containing 2.5 mL of mL of molecular grade water (Sigma, St. Louis, MO) 50% nutrient broth was labeled with the sample num- http://ps.oxfordjournals.org/ and vortexed for 5 s. To each tube, 1 mL of 10 mM ber, series and individual colonies transferred from the ALA solution (Sigma) was added, vortexed for 5 s, and spread plates into the culture tubes using sterile plastic allowed to incubate for 60 min in the dark at 38°C. One loops. The inoculated culture tubes were then incubat- milliliter of trichloroacetic acid stop solution (Sigma) ed at 37°C at 150 rpm for 24 h. was then added to each tube to stop the reaction and Sterile 2-mL cryovials were labeled with the sample vortexed for 5 s. Tubes were then centrifuged for 10 number, series, and isolate number for each cultured

min at 23°C, 400 × g in a 5810R centrifuge (Eppendorf, colony. Glycerol solution (0.8 mL of sterile 50%) was at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Westbury, NY). One hundred microliters of the result- added to each vial. Each bacterial culture was then ing supernatants was then added in duplicate to clear, vortexed and a 1.0-mL sample pipetted into the vial, round bottom 96-well plates (Corning, Corning, NY) capped, and inverted a few times. Cryovials were placed and read in a Synergy 4 plate reader (Biotek, Winooski, in labeled freezer boxes and stored in an ultra-cold VT) at 555 nm against a reagent blank. Results were (−80°C) freezer until further testing. From the culture expressed as nanomoles of ALA used per minute per collection, 300 were randomly selected from the last milliliter of RBC. sampling date, including 150 cloacal isolates and 150 alimentary canal isolates, and used for antimicrobial Histopathology susceptibility testing. Antimicrobial Susceptibility Testing. Individual For tissue evaluation, the liver, spleen, kidney, and isolates were screened for susceptibility against a panel brain were collected immediately following euthanasia of 23 antibiotics using microdilution onto commercial of each bird and fixed in 10% buffered formalin. Each dehydrated 96-well Sensititer GN2F panels for gram- tissue was trimmed for routine processing, then embed- negative bacteria (Thermo Fisher Scientific, Cleveland, ded in paraffin wax and sectioned at 4 to 5 μm. Slides OH) according to the manufacturer’s instructions us- were then stained with hematoxylin and eosin. The re- ing MacConkey broth. The Sensititer GN2F panels sulting slides were then examined by an avian patholo- include 23 different antibiotics or antibiotic combina- gist (S. B. Williams, University of Georgia, Athens), tions. These antibiotics (concentration ranges in mg/L) who was blinded to the treatment groups, and scored included 9 main structural groups, including aminogly- as showing 0 (nil) to 5 (severe) pathology. cosides: amikacin, 8 to 64; gentamicin 2 to 16; tobra- Bacterial Isolations. Four successive series of sam- mycin 4 to 8; beta-lactams: ampicillin, 4 to 32; aztreo- ple tubes were delivered from the Savannah River Ecol- nam 8 to 32; meropenem 1 to 8; piperacillin 16 to 128; ogy Laboratory every 2 wk to the University of Georgia aminoglycoside/β-lactamase inhibitor: ampicillion/ Veterinary College where test subjects were housed. sulbactam A/S2 4/2 to A/S2 32/16; beta-lactam/β- Fifty capped test tubes, each with 5 mL of 0.85% saline lactamase inhibitor: piperacillin/tazobactam constant and a cotton swab, were prepared by autoclaving at 16/4 to 128/4; ticarcillin/clavulanic acid constant 16/2 120°C for 20 min. Swabs from the cloacae of 50 subject to 64/2; cephalosporins: cefazolin 4 to 32; cefepime 4 to chickens were returned on ice to the Savannah River 32; cefotetan 8 to 32; ceftriaxone, 1 to 64; ceftazidime 1 Ecology Laboratory. to 32; cefuroxime 4 to 32; cefoxitin 4 to 32; cefpodoxime A serial dilution of each sample was prepared as fol- 2 to 16; fluoroquinolone: ciprofloxacin 0.5 to 4; gati- lows: 900 μL of sterile saline was pipetted into each of floxacin 1 to 81; imipenem: imipenem 2 to 16; quino- 3 sterile 10-mL culture tubes from which serial dilu- lone: nitrofurantoin 16 to 128; dihydrofolate reductase tions were from 100 to 10−3. Sterile culture tubes were inhibitor/sulfonamide: trimethoprim/sulfamethoxazole labeled with the sample number and dilution. 0.5/9.5 to 4/76. All concentration levels used in this Levine agar Petri plates had been previously pre- analysis were successive doublings from the range mini- pared and were labeled with the sample number, series, mum to maximum. 894 NISANIAN ET AL. Statistical Analysis required to inhibit isolate growth. A simple linear regression analysis of the MIC index versus the TAR was Ten isolates per chicken were identified from growth performed. The weight of evidence (Tuckfield, 2005) plates, which received bacterial inoculations from each data displays, including box-and-whisker plots of the cloacal swab or the final alimentary canal sample ac- TAR scores and scatterplots of MIC index versus TAR, cording the microbial methods stated above. It was as well as all statistical analyses, were generated using determined from preliminary analyses that the d 7 the JMP Pro 10.0 software (SAS Institute Inc., Cary, isolates did not differ significantly form the d 14 iso- NC). lates regarding their susceptibility to select few antibiotic-laced agar plates (data not shown). Therefore,

RESULTS Downloaded from 300 hundred isolates were randomly selected from deep freeze (−70°C) preservation state subsequent to the experiment, 150 from the d 14 cloacal swabs and 150 Blood Pb Levels from alimentary canal samples. The selection process Providing Pb acetate in the drinking water of chick- ensured that at least 1 of the 2 chickens per cage was ens resulted in dose-related increases in their blood Pb represented in the d 14 cloacal swabs and alimentary levels (Figure 1). These increasing blood Pb levels with http://ps.oxfordjournals.org/ canal sampling schemes. These isolates were subjected increasing Pb in the drinking water verified enteric to antimicrobial susceptibility testing. Each isolate was availability of Pb in the birds. then scored as to the number of antimicrobial resistance traits (TAR) evinced by that isolate, which values ranged from 0 to 23. These scores are a response mea- Signs of Overt Pb Toxicity sure per isolate and were analyzed by a 1-way ANOVA By d 14, the chickens receiving, via the water, the 2 using the corresponding ranks of all isolate responses highest Pb acetate concentrations showed moribundity at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 (Conover, 1999). The main effect constituted the 5 Pb in the form of impaired activity and reduced water in- concentration levels in the experiment. Because it was take, most notable at the highest dose. These 2 higher determined that the 2 highest Pb concentrations (1 and dose groups were therefore not used for additional end- 10 mM) resulted in adipsia, only responses among vi- points. Weight gain data are shown in (Figure 2) for all able enterobacter isolates from chickens in the 3 lowest exposure groups. There were no significant differences Pb treatment levels were considered for final statistical in the 0.01 or 0.1 mM Pb acetate treatment groups in analysis. By the method of linear contrasts (Neter et hematocrit, whole-blood cellularity, leukocyte differen- al., 1996), the control (0.0 or control) average treatment tial, PCV, MCV, or blood proteins (data not shown). level TAR response from the rank-transformed ANOVA was compared with the average treatment level TAR response for each of the 0.01 and 0.1 mM Pb treatment %ORRGį$/$'$FWLYLW\ levels, respectively. In addition, a minimum inhibitory concentration (MIC) index was created and calculated Exposure of birds to Pb acetate significantly de- as the trimmed average Z-score per isolate among the creased δ-ALAD activity in a dose-related manner (Ta- set of antibiotic concentrations per antibiotic used in ble 1). The level of depressed enzyme activity did not the antimicrobial susceptibility testing plates. Z-scores change between d 7 and 14. were created as centered and scaled values of the base 2 logarithms (log2) of the largest antibiotic concentration Histopathology to which an isolate was resistant. The log2 was selected because the antibiotic concentrations were successive There were no noteworthy histopathologic findings in doublings from 0.5 to 128, which transformed these re- the birds exposed to 0.01 or 0.1 mM Pb acetate (data sponses into a range of values from 1 to 8, respectively. not shown). Only those antibiotics to which isolates were resistant were included in the calculation (i.e., a trimmed num- Microbial Analysis ber ≤23). Because the Z-score lacks a unit of measurement, the resulting trimmed average reflects the col- As indicated above, antibiotic resistance patterns lective strength of the concentration among antibiotics analysis was restricted to the control and first 2 Pb

Table 1. Peripheral blood delta aminolevulinic-acid dehydratase assay

Day 0 Day 7 Day 14 Pb acetate (mM) (nmol/min per mL) (nmol/min per mL) (nmol/min per mL) 0 116 ± 5.0 114.9 ± 11.4 128.3 ± 6.3 0.01 125.8 ± 12.7 93.6 ± 13.2 99.3 ± 5.0 0.1 140.5 ± 24.5 58.6 ± 6.5* 64.5 ± 5.0* *P < 0.05. ANOVA, n = 10 birds/treatment. RESEARCH NOTE 895 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 2. Weight gain of chickens exposed to Pb acetate in drinking water. Birds were weighed on d 0, 7, and 14. Measurements of weight gain were determined by the difference in d 7 bird weight from d 0 bird weight (A) and the difference in d 14 bird weight from d 7 bird weight (B) across all treatments. *P < 0.05. ANOVA, n = 10 birds/ treatment. Color version available in the online PDF.

tance traits found in the control isolates was 5, whereas some isolates in the 0.1 mM treatment were resistant to Figure 1. Blood Pb levels in chickens exposed to Pb acetate in 20 of the 23 antibiotics. Linear contrasts from the rank- drinking water. Peripheral blood was collected from a jugular vein of transformed ANOVA showed that isolates exposed to chickens and Pb levels were measured on d 0 (A), d 7 (B), and d 14 (C). *P < 0.05. ANOVA, n = 10 birds/treatment. Color version avail- the 0.01 mM Pb level had a significantly (P = 0.05) able in the online PDF. elevated TAR above the control (0.0 mM) Pb level. In addition, isolates exposed to the 0.1 mM Pb level had a significantly (P < 0.05) elevated TAR above the control. acetate treatments (0, 0.01, 0.1 mM) of the d 14 cloacal Figure 4 shows a linear trend of the TAR as a func- and alimentary canal isolates, due to moribundity and tion of number of the MIC index. Although there is noncontinuous exposure at the higher concentrations. considerable variation in this index for isolates with Restricted analysis showed no treatment effect among few resistance traits, there is a significant (P < 0.05) the alimentary canal isolates. However, as seen in Fig- positive relationship from the inclusion of isolates with ure 3, there was an overall statistically significant (P < multiple antibiotic resistances. These data suggest that 0.05) treatment effect due to Pb concentrations among as the number of individual resistances exhibited by an d 14 cloacal isolates. The maximum number of resis- isolate increases, so too does the average MIC. 896 NISANIAN ET AL. Downloaded from http://ps.oxfordjournals.org/ Figure 3. Number of antibiotic resistance (AR) traits per isolate obtained from d 14 samples exposed to Pb acetate in drinking water. The last 2 concentrations were not included as noted in the text. Color version available in the online PDF. Figure 4. Significant relationship between the number of antibiotic resistance (AR) traits per isolate and index of the minimum inhibitory concentration (MIC) required to suppress growth. Index is a Z-score calculated from the concentrations of all antibiotics. Color version DISCUSSION available in the online PDF.

Calomiris et al. (1984) reported a positive correlation at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 between Cu, Pb, and Zn contaminants, and multiple studies have found zoonotic transmission between live- antibiotic resistance in bacteria in the drinking water stock and humans (Morgan, 2008; Catry et al., 2010). system of an Oregon coastal community. They attrib- Of the livestock species studied, there has been almost uted this finding to a selection process undergone by no focus on chickens. A single study sampling 1,913 bacteria in a drinking water distribution system via specimens from various livestock animals determined exposure to metals from the piping, that was thought that out of the 15 specimens positive for MRSA, 3 of to be plasmid-driven. In support of this idea, penicil- them were chickens, and these isolates’ genomes were linase plasmids were previously implicated in enhanced very similar to those of the human strains, suggesting Pb resistance detected in Staphylococcus aureus strains common ancestry (Lee, 2003). This would implicate the (Novick and Roth, 1968). possibility of MRSA infection from animal to handler. Metal-ion resistance systems have been found on A metal-related increase in antibiotic-resistant plasmids of every eubacterial group tested, from Esch- strains of pathogens such as S. aureus and V. vulnifi- erichia coli to Streptomyces (Silver, 1996). These mi- cus, and the ability of these strains to transfer zoonoti- crobes are able to survive and proliferate, and plasmids cally to humans, may represent an important issue in encoding resistance have generally been found to be the realm of disease management. Fecal swabs from the higher in areas of pollution (Habi and Daba, 2009). present chickens contained isolates that had significant Plasmids that contain metal resistance genes also con- levels of antibiotic resistance. Although the majority tain genes encoding for antimicrobial resistance (Silver, of isolates were susceptible to most of the antibiotics 1996). Thus, it has been suggested that the occurrence screened, some were found in the Pb-treated birds that of antibiotic-resistant and metal-tolerant bacterial were resistant to up to 20 antibiotics. Given that the strains has resulted from selection for this tolerance isolates from the control did not express this same level through environmental exposure to heavy metals (Na- of resistance, it is reasonable to conclude that selection kahara et al., 1977). imposed by the Pb acetate resulted in the observed pat- Karbasizaed et al. (2003) observed that bacterial terns. It is not clear from where the resistance elements strains isolated from nosocomial patients were more originated as these chickens were SPF Leghorns. resistant to metals and antibiotics compared with These results amplify the need for similar studies to strains from healthy humans. Some of the enterobac- be repeated in chickens infected with zoonotic patho- teria isolated from the nosocomial patients harbored gens. Should increased antibiotic resistance occur in a conjugative plasmid (>56.4 kb) encoding resistance these microbial species, this would increase difficulty to antibiotics and heavy metals. One particular noso- managing zoonotic infections derived from animals that comial infectious agent of concern that has increased were also exposed to Pb through feeds or water. Our re- in incidence over recent years is methicillin-resistant sults suggest the potential for indirect selection (Baker- Staphylococcus aureus (MRSA). It has been suggested Austin et al., 2006) occurring in organisms exposed to that MRSA incidence may be increasingly attributed toxic heavy metals producing multiple antibiotic-resis- to sources outside the nosocomial setting, and many tant species. Species that are in close contact with hu- RESEARCH NOTE 897 mans and may in turn pass these traits on to zoonotic ion animals: Epidemiology and control options for human and pathogens and decrease the ability to treat infections, animal health. Epidemiol. Infect. 138:626–644. Conover, W. J. 1999. Practical Nonparametric Statistics. John Wi- such as chickens, may be of greatest concern. ley & Sons, New York, NY. D’Ilio, S., M. Costanza, F. Petrucci, N. Violante, and O. Senofonte. 2010. Method validation for the quantification of As, Cd, Hg and ACKNOWLEDGMENTS Pb in blood by ICP-MS for monitoring purposes. Anal. Methods 2:2049–2054. This research was partially supported by the De- El-Tras, W. F., A. A. Tayel, and A. Samir. 2010. Potential zoonotic partment of Energy under Award Number DE-FC09- pathways of Salmonella Enteritidis in laying farms. Vector Borne 07SR22506 to the University of Georgia Research Zoonotic Dis. 10:739–742. Foundation. We thank Paul Stankus for laboratory as- Gasparik, J., D. Vladarova, M. Capcarova, P. Smehyl, J. Slamacka,

P. Garaj, R. Stawarz, and P. Massanyi. 2010. Concentration of Downloaded from sistance. lead, cadmium, mercury and arsenic in leg skeletal muscles of Disclaimer: This report was prepared as an account three species of wild birds. J. Environ. Sci. Health A 45:818–823. of work sponsored by an agency of the United States Gogal, R. M. J., M. S. Johnson, C. T. Larsen, M. R. Prater, R. B. Duncan, D. L. Ward, and S. D. Holladay. 2002. Influence of Government. Neither the United States Government dietary 2,4,6-trinitrotoluene exposure in the northern bobwhite nor any agency thereof, nor any of their employees, (Colinus virginianus). Environ. Toxicol. Chem. 21:81–86. makes any warranty, express or implied, or assumes any Habi, S., and H. Daba. 2009. Plasmid incidence, antibiotic and met- http://ps.oxfordjournals.org/ legal liability or responsibility for the accuracy, com- al resistance among Enterobacteriaceae isolated from Algerian streams. Pak. J. Biol. Sci. 12:1474–1482. pleteness, or usefulness of any information, apparatus, Karbasizaed, V., N. Badami, and G. Emtiazi. 2003. Antimicrobial, product, or process disclosed, or represents that its use heavy metal resistance and plasmid profile of coliforms isolated would not infringe privately owned rights. Reference from nosocomial infections in a hospital in Isfahan, Iran. Afr. J. herein to any specific commercial product, process, Biotechnol. 2:379–383. Lee, J. H. 2003. Methicillin (Oxacillin)-resistant Staphylococcus au- or service by trade name, trademark, manufacturer, reus strains isolated from major food animals and their potential or otherwise does not necessarily constitute or imply transmission to humans. Appl. Environ. Microbiol. 69:6489– at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 its endorsement, recommendation, or favoring by the 6494. United States Government or any agency thereof. The Mateo, R., A. Geren, C. Jeske, V. Urios, and C. Gerique. 2001. Lead poisoning in the globally threatened marbled teal and white- views and opinions of authors expressed herein do not headed duck in Spain. Environ. Toxicol. Chem. 20:2860–2868. necessarily state or reflect those of the United States McArthur, J. V., and R. C. Tuckfield. 2000. Spatial patterns in Government or any agency thereof. antibiotic resistance among stream bacteria: Effects of industrial pollution. Appl. Environ. Microbiol. 66:3722–3726. Morgan, M. 2008. Methicillin-resistant Staphylococcus aureus and REFERENCES animals: Zoonosis or humanosis? J. Antimicrob. Chemother. 62:1181–1187. Agency for Toxic Substances and Disease Registry (ATSDR). 2007. Nakahara, H., T. Ishikawa, Y. Sarai, and I. Kondo. 1977. Frequency Toxicological Profile for Lead, P.H.S. US Department of Health of heavy-metal resistance in bacteria from inpatients in Japan. and Human Services, Atlanta, GA. Nature 266:165–167. Baker-Austin, C., J. V. McArthur, A. H. Lindell, M. S. Wright, National Ambient Air Quality Standard (NAAQS). 2012. Lead in R. C. Tuckfield, J. Gooch, L. Warner, J. Oliver, and R. Step- Air, US Environmental Protection Agency, Washington, DC. Ac- anauskas. 2009. Multi-site analysis reveals widespread antibiotic cessed Sep. 2013. http://www.epa.gov/airquality/lead/. resistance in the marine pathogen Vibrio vulnificus. Microbial Neter, J., M. H. Kutner, C. J. Nachtsheim, and W. Wasserman. Ecol. 57:151–159. 1996. Applied Linear Statistical Models. 4th ed. McGraw-Hill, Baker-Austin, C., M. S. Write, R. Stepanauskas, and J. V. Mc- Boston, MA. Authur. 2006. Co-selection of antibiotic and metal resistance. Novick, R. P., and C. Roth. 1968. Plasmid-linked resistance to inor- Trends Microbiol. 14:176–182. ganic salts in Staphylococcus aureus. J. Bacteriol. 95:1335–1342. Bennett, J., C. Kaufman, I. Koch, J. Sova, and K. Reimer. 2007. Pagan-Rodriguez, D., M. O’Keefe, C. Deyrup, P. Zervos, H. Walker, Ecological risk assessment of lead contamination at rifle and pis- and A. Thaler. 2007. Cadmium and lead residue control in a haz- tol ranges using techniques to account for site characteristics. ard analysis and critical control point (HACCP) environment. J. Sci. Total Environ. 374:91–101. Agric. Food Chem. 55:1638–1642. Berlin, A., and K. H. Schaller. 1974. European standardized method Silver, S. 1996. Bacterial resistances to toxic metal ions—A review. for the determination of delta-aminolevulinic acid dehydratase Gene 179:9–19. activity in blood. Z. Klin. Chem. Klin. Biochem. 12:389–390. Stepanauskas, R., T. C. Glenn, C. H. Jagoe, R. C. Tuckfield, A. H. Beyer, W. N., J. W. Spann, L. Sileo, and J. C. Franson. 1988. Lead Lindell, C. J. King, and J. V. McArthur. 2006. Co-selection for poisoning in six captive avian species. Arch. Environ. Contam. microbial resistance to metals and antibiotics in freshwater mi- Toxicol. 17:121–130. crocosms. Environ. Microbiol. 8:1510–1514. Calomiris, J. J., J. L. Armstrong, and R. J. Seidler. 1984. Asso- Tuckfield, R. C. 2005. Statistical thinking for students of ecology. ciation of metal tolerance with multiple antibiotic resistance of Chapter 12 in Fundamentals of Ecology. 5th Ed. E. P. Odum and bacteria isolated from drinking water. Appl. Environ. Microbiol. G. Barrett, ed. Thomson Learning Press. 47:1238–1242. Waegeneers, N., H. De Steur, L. De Temmerman, S. Van Steenwin- Catry, B., E. Van Duijkeren, M. C. Pomba, C. Greko, M. A. More- kel, X. Gellynck, and J. Viaene. 2009. Transfer of soil contami- no, S. Pyorala, M. Ruzauskas, P. Sanders, E. J. Threlfall, F. nants to home-produced eggs and preventive measures to reduce Ungemach, K. Torneke, C. Munoz-Madero, and J. Torren-Edo. contamination. Sci. Total Environ. 407:4438–4446. 2010. Reflection paper on MRSA in food-producing and compan- METABOLISM AND NUTRITION

Effects of low ambient temperatures and dietary vitamin C supplement on growth performance, blood parameters, and antioxidant capacity of 21-day-old broilers

X. Yang , Y. H. Luo , Q. F. Zeng, 1 K. Y. Zhang , X. M. Ding , S. P. Bai , and J. P. Wang

Institute of Animal Nutrition, Key Laboratory for Animal Disease Resistance Nutrition of the Ministry Downloaded from of Education, Sichuan Agricultural University, Sichuan, Ya’an 625014, China

ABSTRACT The study was conducted to determine whole experimental period (P < 0.01), whereas it in- the effects of low ambient temperature (LAT) and a creased heart index at 21 d (P < 0.05) and hematocrit vitamin C (VC) dietary supplement on the growth per- and hemoglobin level at 14 d (P < 0.05). Supplement- http://ps.oxfordjournals.org/ formance, blood parameters, and antioxidant capacity ing the diet with VC increased hematocrit, hemoglobin, of 21-d-old broilers. A total of 400 one-day-old male and red blood cell count at 21 d (P < 0.05). At 21 d, Cobb broilers were assigned to 1 of 4 treatments as fol- LAT conditions decreased total antioxidant capacity lows: 1) LAT and a basal diet; 2) LAT and a basal diet in the serum, liver, and lungs (P < 0.05), and it also supplemented with 1,000 mg of VC/kg (LAT + VC); 3) increased the levels of VC in the serum and liver, the normal ambient temperature (NAT) and a basal diet; amount of protein carbonylation in liver and lungs, and 4) NAT and a basal diet supplemented with 1,000 mg the malondialdehyde level in the lungs (P < 0.05). The of VC/kg (NAT + VC). All birds were fed to 21 d of addition of VC tended to increase the total antioxidant at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 age. Broilers in groups 1 and 2 were raised at 24 to capacity level in serum (P < 0.1). Low ambient tem- 26°C during 1 to 7 d, and at 9 to 11°C during 8 to 21 perature resulted in oxidative stress for broilers that d, whereas groups 3 and 4 were raised at 29 to 31°C were fed from 1 to 21 d of age, whereas no significant during 1 to 7 d and at 24 to 26°C during 8 to 21 d. effect was found on the antioxidant activity by dietary The LAT increased the feed conversion ratio during the VC supplementation.

Key words: low ambient temperature, vitamin C, ascites , blood parameter , antioxidant capacity 2014 Poultry Science 93 :898–905 http://dx.doi.org/ 10.3382/ps.2013-03438

INTRODUCTION temperature by 1°C per day from 5 to 22 d resulted in broilers experiencing right ventricular hypertrophy Fast-growing broilers are susceptible to ascites syn- and death from AS. In another study, keeping ambi- AS drome ( ), which causes fluid accumulation in their ent temperatures at 32 and 30°C during 1 and 2 wk, abdominal cavities and is associated with pathological respectively, and then reducing it to 15°C during the damage to the heart, liver, and lungs. This syndrome third week, also led to more AS incidences in broil- leads to increased mortality and results in substantial ers (Daneshyar et al., 2009). Because of the significant economic losses (Maxwell and Robertson, 1997). Hypo- temperature changes in the winter in China, the ex- baric chambers (Owen et al., 1990; Julian and Squires, pensive thermal equipment to mitigate temperature ef- 1994), low ventilation (Enkvetchakul et al., 1993), and fects, and the high incidence of AS affecting broilers in the compound 3,3c,5-triiodothyronine (Decuypere et winter, it is important to study the effects of natural al., 1994) have all been found to induce AS. temperatures on the incidence of AS in winter. LAT Low ambient temperatures ( ), which affect most Previous studies have indicated that a lack of oxy- farming areas in the winter, have also been reported gen in the body of modern broilers, due the imbalance to induce widespread incidences of AS. In a study by between rapid growth and limited cardiopulmonary Julian et al. (1989), holding meat-type chickens at an development, is a major factor in developing AS (Ju- ambient temperature of 30°C and then reducing the lian et al., 1987; Nain et al., 2009). Nain et al. (2008) reported that fast-growing commercial broilers had a high risk of heart failure coupled with increased oxida- © 2014 Poultry Science Association Inc. tive stress resulting from continuous hypoxia; oxida- Received June 22, 2013. Accepted December 16, 2013. tive stress caused damage to tissues and exhausted the 1 Corresponding author: [email protected] broilers’ antioxidant supply. Previously, Enkvetchakul

898 EFFECTS OF LOW TEMPERATURES AND VITAMIN C ON BROILERS 899

Table 1. Temperature scheme1

Indoor temperature Temperature inside the cage

Item 0–7 d 8–21 d 0–7 d 8–21 d LAT (°C) 24–26 9–11 27–29 11–13 NAT (°C) 29–31 24–26 31–33 27–29 1LAT = low ambient temperature, NAT = normal ambient temperature. et al. (1993) showed that levels of ascorbic acid (VC) University in China. The initial BW was similar among and glutathione in the lungs and liver were low in as- all pens. Feed and water were provided ad libitum. The Downloaded from citic broilers housed in a poorly ventilated environment experimental treatments were arranged as a 2 × 2 fac- for 7 wk. Wang et al. (2012) recently observed that torial as follows: 1) LAT and a basal diet; 2) LAT and a LAT conditions can induce body hypoxia and result in basal diet supplemented with 1,000 mg of VC/kg (LAT tissue damage. This study showed an increase in hepat- + VC); 3) Normal ambient temperature (NAT) and ic malondialdehyde (MDA), a biomarker for oxidative a basal diet; 4) NAT and a basal diet supplemented http://ps.oxfordjournals.org/ stress, and a decrease in the activity of total superoxide with 1,000 mg of VC/kg (NAT + VC). Each treatment dismutase (SOD), an antioxidant that protects against group contained 10 replicate pens, with 10 broilers per oxidative stress, in ascitic broilers exposed to LAT con- pen. The temperature scheme is shown in Table 1. The ditions. VC was purchased from Minsheng Biochemical Tech- Incorporating antioxidants into broilers’ diets has nology Co. Ltd. (Hangzhou, Zhejiang province, China), been investigated as an approach to reduce the effects and the content of ascorbic acid was above 98%. The

of oxidative stress resulting from hypoxia. Ladmakhi et basal diet (mash form) was a corn-soybean meal-based at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 al. (1997) showed that incorporating VC (500 mg/kg diet and was formulated to meet or exceed nutrient of diet) into the diet of broilers with AS reduced their requirements of broilers (Table 2). mortality rates. Another study showed that VC (500 mg/kg of diet) played an important role in decreasing Sampling and Measurements the incidence of AS, which was induced in broilers by LAT conditions and supplemented dietary 3,3c,5-triio- Body weight, feed intake, and feed conversion ratio of dothyronine (Xiang et al., 2002). each replicate were determined on a weekly basis. One Considering these previous investigations, the first replicate (10 birds per replicate) was randomly selected aim of our study was to determine whether natural LAT conditions were able to induce AS, and how they influenced the blood parameters and antioxidant capac- Table 2. Composition of the basal diet for birds fed 0 to 21 d ity in broilers. The second aim was to evaluate whether of age dietary supplemental VC could alleviate the adverse Item % effects of LAT by improving the antioxidant capacity of broilers. This study focused on the starter phase Ingredient Corn (%) 59.2 of broilers (1–21 d) to investigate the effects of LAT Soybean meal (%) 29.1 condition and VC supplementation on the growth per- Soybean oil (%) 4.5 formance, blood parameters, and antioxidant capacity Fish meal (%) 3.8 Limestone (%) 1 in 21-d-old broilers. The results of this study will be Calcium hydrophosphate (%) 1.24 important for broiler producers to prevent AS in LAT dl-Methionine (%) 0.13 1 conditions. Mineral premix (%) 0.5 Salt (%) 0.35 Choline chloride (%) 0.15 Vitamin premix2 (%) 0.03 MATERIALS AND METHODS Total (%) 100 Determined composition Birds, Experimental Design, and Diet ME (Mcal/kg) 3.09 CP (%) 19.7 The animal experiment was carried out in accor- Calcium (%) 0.91 Available phosphorus (%) 0.45 dance with the Chinese guidelines for animal welfare Methionine (%) 0.45 and approved by the Animal Health and Care Commit- Lysine (%) 1.07 tee of Sichuan Agricultural University. A total of 400 Methionine + cysteine (%) 0.76 one-day-old male Cobb broilers were purchased from 1Provided per kilogram of diet: iron, 100 mg; copper, 8 mg; manga- a commercial hatchery (Wenjiang Zhengda Co. Ltd., nese, 100 mg; zinc, 75 mg; selenium, 0.15 mg; iodine, 0.45 mg. Chengdu, Sichuan province, China). They were raised 2Provided per kilogram of diet: vitamin A, 8,000 IU; cholecalciferol, 2,000 IU; vitamin E, 5 IU; vitamin K , 1 mg; thiamine, 0.4 mg; ribofla- in cage pens (1.5 m × 1 m × 0.5 m) at the farm of 3 vin, 3.2 mg; pyridoxine, 1.2 mg; vitamin B12, 6 μg; folic acid, 100 μg; the Animal Nutrition Institute of Sichuan Agricultural niacin, 7 mg; calcium pantothenate, 5 mg. 900 YANG ET AL. from each treatment group at 7, 14, and 21 d for blood was used to analyze the effect of time on the RBC collection. Blood (approximately 2 mL) was collected count, HCT, and HGB level. Duncan’s multiple range via the vena brachialis and was placed into tubes con- test was used as a multiple comparison procedure to taining EDTA-Na2 (an anticoagulant) for subsequent compare means when the effect of age was significant. determination of blood parameters, which included Differences in the means were expressed as either sig- the hematocrit (HCT), the hemoglobin (HGB) level, nificant (P < 0.05), or trending toward significance (P and the red blood cell (RBC) count. These blood pa- < 0.1). Statistical analyses were performed using the rameters were tested using an automated hematology SAS system (9.0). Data are presented as the mean with analyzer (Abbott CD 3200) within 2 h after collection. pooled SEM values. Birds were euthanized by cervical dislocation after Downloaded from blood collection. RESULTS Next, cardiac tissue samples from each treatment were obtained. The atria, major vessels, and gross fat Effects of Temperature and VC were removed, and the right ventricle was cut away on the Mortality, Heart Index, from the left ventricle and septum, which were left in-

and Growth Performance of Broilers http://ps.oxfordjournals.org/ tact together. The right ventricle was weighed separately from the left ventricle and septum tissue. Heart Broiler mortality under LAT, both with and without index was calculated as the weight of the right ventricle the VC supplement, was 8% throughout the study pe- divided by the weight of the left ventricle and septum. riod. Broiler mortality under NAT was 0%. Deaths that On d 21, in addition to the blood samples collected occurred between 1 and 14 d were not caused by heart via the vena brachialis, blood samples were also ob- failure or AS, according to the symptoms. After 14 d, tained from a jugular vein before euthanasia. The blood one of the broilers in the LAT group (without the VC samples from the jugular vein were centrifuged (2,000 supplement) died of AS. The morbidities of AS were at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 × g, 10 min, 4°C) to obtain serum. The left liver and low in this study according to the apparent symptoms lung of each of the selected broilers were obtained. Se- and autopsy, which was only caused by LAT conditions rum, liver, and lung samples were then stored at −20°C over the entire study period (21 d; data not shown). until the antioxidant capacity was analyzed. Heart index values and growth performance measure- The antioxidant capacity analyses included measur- ments are shown in Table 3. The LAT environment ing VC levels, total antioxidant capacity (T-AOC), resulted in increased heart index values at 21 d (P < MDA levels, amount of protein carbonylation, SOD ac- 0.05) compared with NAT conditions. There was no tivity, and glutathione peroxidase (GSH-Px) activity. effect of VC supplementation on heart index values (P Serum samples were used directly, whereas the tissue > 0.05). The LAT environment decreased average BW samples were homogenized with saline and centrifuged gain during 15 to 21 d (P < 0.01) and increased feed at 2,000 × g for 10 min at 4°C, and the supernatant was conversion ratio (feed intake/gain, g/g) during the en- collected for the assay. Spectrophotometry was used to tire period (P < 0.05). The addition of VC to the diet measure the reaction products in the VC and T-AOC increased ADFI and feed conversion ratio during 1 to assays. The MDA assay was performed using the TBA 7 d (P < 0.01), but were decreased by the VC supple- colorimetric method (Ohkawa et al., 1979). The activity ment during 8 to 14 d. The interactions for ADFI be- of SOD was measured as the ability of SOD to inhibit tween temperature and VC were significant at 15 to 21 the oxidation of oxymine by the xanthine-xanthine oxi- d (P < 0.05), as well as feed conversion ratio at 8 to 14 dase system (Oyanagui, 1984). The activity of GSH-Px d (P < 0.01). was also determined by spectrophotometry to measure the reaction rate of H2O2 and glutathione. The details Effects of Temperature and VC on the Blood of the measurement of these components followed the kit instructions (Nanjing Jiancheng Bioengineering In- Parameters of Broilers stitute, China). Blood parameter results are shown in Table 4. At In this experiment, broilers that showed abdominal 14 d, the LAT environment increased HCT and HGB fluid accumulation were regarded as having AS. Broil- levels (P < 0.05). Adding the VC supplement increased ers that died during the experimental period underwent the RBC count, HCT, and the HGB level at 21 d in autopsy to examine whether AS was the cause by ob- the LAT groups and NAT groups (P < 0.05). The in- serving the tissue lesions and abdominal fluid accumu- teraction for HGB level between temperature and VC lation. The morbidities and mortalities were recorded. tended to be significant at 7 and 21 d (P < 0.1). The blood parameter results varied with broiler age Statistical Analysis (Table 5). Age of the broiler affected the RBC count, HGB level, and HCT. The RBC count, HGB level, and Two-way ANOVA of GLM was used to test the main HCT increased with age in broilers within the LAT + effects of VC supplement and temperature, and to in- VC group, the NAT group, and the NAT + VC group vestigate their interaction effects. A one-way ANOVA (P < 0.05). EFFECTS OF LOW TEMPERATURES AND VITAMIN C ON BROILERS 901

Table 3. Effects of temperature and vitamin C (VC) on the heart index, average BW gain (ABWG), ADFI, and feed conversion ratio (feed/gain; FCR) of broilers1

P-value Pooled Item LAT LAT + VC NAT NAT + VC SEM Temperature VC Temperature × VC Heart index (g/g) 21 d 0.19 0.20 0.16 0.16 0.01 * 0.863 0.457 ABWG (g) 1–7 d 16.1 15.9 16.6 16.9 0.39 0.057 0.900 0.566 8–14 d 32.6 30.1 36.0 35.3 2.42 0.086 0.505 0.703 15–21 d 37.9 38.9 45.9 46.4 1.68 *** 0.644 0.898

ADFI (g) Downloaded from 1–7 d 27.9 28.6 27.1 28.9 0.42 0.576 ** 0.216 8–14 d 57.0 50.7 54.9 40.1 3.67 0.092 ** 0.259 15–21 d 101 91.5 97.4 102 3.06 0.263 0.452 * FCR (g/g) 1–7 d 1.74 1.80 1.64 1.72 0.03 ** * 0.777 8–14 d 1.77 1.70 1.53 1.14 0.05 *** *** ** 15–21 d 2.71 2.42 2.13 2.23 0.14 ** 0.472 0.155 http://ps.oxfordjournals.org/ 1LAT = low ambient temperature, NAT = normal ambient temperature. *P < 0.05; **P < 0.01; ***P < 0.001.

Effects of Temperature and VC no differences in the VC level, MDA level, T-AOC, or on the Antioxidant Capacity of Broilers amount of protein carbonylation in the liver between groups that were given the VC supplement and those Results of antioxidant capacity analyses of the se- that were not (P > 0.05). at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 rum, liver, and lungs at 21 d are shown in Table 6. In the lungs, LAT environment increased the MDA The LAT environment elevated VC level and reduced level and the amount of protein carbonylation, and de- T-AOC in serum (P < 0.05). The LAT environment creased T-AOC (P < 0.05). There were no differences tended to increase the levels of MDA and the amount of in the VC level in the lungs between the LAT and NAT protein carbonylation in serum (P < 0.1). There were groups (P > 0.05), and no effects of the VC supple- no differences in the activity levels of SOD and GSH- mentation were observed on the VC level, MDA level, Px in serum between the LAT and NAT groups (P > T-AOC, and amount of protein carbonylation in the 0.05). Groups that were given the VC supplement tend- lung (P > 0.05). ed to have increased T-AOC (P < 0.1), whereas the VC supplement decreased the activity of GSH-Px in serum DISCUSSION (P < 0.05). There were no differences in the VC level, MDA level, protein carbonylation level, or the SOD Considering that LAT conditions are a major stressor activity in the serum between groups that were given for broilers in the winter and contribute to the high the VC supplement and those that were not (P > 0.05). occurrence of AS in most areas, it was worthwhile to The LAT environment increased the levels of VC and investigate the response of broilers to natural LAT con- protein carbonylation, and decreased T-AOC in the liv- ditions. The heart index value was indicative of AS er (P < 0.05). The LAT environment tended to increase because right ventricular hypertrophy is an indicator of the level of MDA in the liver (P < 0.1). There were pulmonary hypertension associated with AS (Huchzer-

Table 4. Effects of temperature and vitamin C (VC) on the blood parameters of broilers1

P-value Pooled Item LAT LAT + VC NAT NAT + VC SEM Temperature VC Temperature × VC HGB (g/L) 7 d 77.6 72.2 71.0 75.8 2.42 0.543 0.903 0.051 14 d 94.7 91.0 79.7 81.3 4.44 * 0.828 0.565 21 d 86.7 104 90.8 96.8 2.46 0.613 ** 0.058 HCT (%) 7 d 0.32 0.31 0.30 0.31 0.01 0.146 0.908 0.215 14 d 0.36 0.35 0.31 0.31 0.02 * 0.854 0.585 21 d 0.34 0.40 0.36 0.38 0.01 0.946 * 0.157 RBC (× 1012/L) 7 d 2.28 2.16 2.10 2.24 0.08 0.544 0.886 0.108 14 d 2.64 2.57 2.32 2.40 0.12 0.068 0.967 0.519 21 d 2.57 2.92 2.66 2.79 0.07 0.753 ** 0.127 1LAT = low ambient temperature; NAT = normal ambient temperature; HGB = hemoglobin; HCT = hematocrit; RBC = red blood cell. *P < 0.05. **P < 0.01. 902 YANG ET AL.

Table 5. Effect of broiler age on red blood cell (RBC) count, hemoglobin (HGB) level, and hematocrit (HCT)1

Item LAT LAT + VC NAT NAT + VC RBC (× 1012/L) 7 d 2.28b 2.16c 2.10c 2.24b 14 d 2.64a 2.57b 2.32b 2.40b 21 d 2.57ab 2.92a 2.66a 2.79a Pooled SEM 0.10 0.10 0.04 0.10 P-value * *** *** * HGB (g/L) 7 d 77.6b 72.2c 71.0c 75.8b a b b b 14 d 94.7 91.0 79.7 81.3 Downloaded from 21 d 86.7ab 104a 90.8a 96.8a Pooled SEM 3.19 3.24 1.70 3.70 P-value ** *** *** ** HCT (%) 7 d 0.32b 0.31b 0.30b 0.31b 14 d 0.36a 0.35ab 0.31b 0.31b 21 d 0.34ab 0.40a 0.36a 0.38a http://ps.oxfordjournals.org/ Pooled SEM 0.01 0.02 0.01 0.01 P-value * ** ** ** a–cDifferent lowercase letters in the same column mean significant differences (P < 0.05). 1LAT = low ambient temperature; NAT = normal ambient temperature; VC = vitamin C. *P < 0.05. **P < 0.01. ***P < 0.001.

meyer and De Ruyck, 1986). A previous study showed Hypoxia has been shown to result in significant re- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 that the effect of LAT on the right ventricle was remark- ductions in BW, which are caused by increased energy able, resulting in heavier total ventricular weights and demands (Yahav and McMurtry, 2001). In the current higher values for pulmonary arterial pressure (Wide- study, hypoxia resulting from LAT conditions caused man and Tackett, 2000). The heart index was gener- an increase in the broilers’ feed conversion ratio be- ally increased with age or BW. Because LAT stress cause the birds needed more energy to produce heat resulted in poor growth in the present experiment, the to maintain their temperature and depressed their BW heart index values of both LAT and NAT treatments gain, implying a strong LAT stress on broilers accord- were low at 21 d. However, LAT increased heart index ing to the growth performance. The addition of VC de- values compared with NAT. Therefore, these results creased the feed conversion ratio by reducing the ADFI implied that LAT caused heavier right ventricle, and it during 8 to 14 d. However, there were no differences in increased the possibility of developing AS. growth performance after 14 d as a result of adding VC.

Table 6. Effects of temperature and vitamin C (VC) on the VC, total antioxidant capacity (T-AOC), malondialdehyde (MDA), protein carbonylation, superoxide dismutase (SOD), and glutathione peroxidase (GSH-Px) in the serum and on the VC, T-AOC, MDA, and protein carbonylation in the liver and the lungs at 21 d1

P-value

LAT NAT Pooled Temperature Item LAT + VC NAT + VC SEM Temperature VC × VC Serum VC (μg/mL) 16.0 18.2 13.2 14.8 1.42 * 0.207 0.839 T-AOC (U/mL) 3.01 4.62 4.78 4.89 0.46 * 0.082 0.127 SOD (U/mL) 130 141 151 151 17.2 0.372 0.766 0.741 GSH-Px (U/mL) 194 154 193 159 16.4 0.909 * 0.841 MDA (nmol/mL) 15.8 15.6 14.2 10.4 1.74 0.070 0.275 0.317 Protein carbonylation (nmol/mL) 2.03 1.95 1.87 1.20 0.23 0.063 0.124 0.217 Liver VC (μg/mg of protein) 2.09 2.21 1.58 1.80 0.18 * 0.346 0.780 T-AOC (U/mg of protein) 0.82 0.94 0.98 1.07 0.06 * 0.111 0.841 MDA (nmol/mg of protein) 1.04 0.82 0.47 0.73 0.17 0.067 0.926 0.176 Protein carbonylation 4.11 3.98 2.56 3.09 0.45 * 0.666 0.476 (nmol/mg of protein) Lungs VC (μg/mg of protein) 0.06 0.09 0.21 0.08 0.05 0.148 0.339 0.098 T-AOC (U/mg of protein) 0.28 0.40 0.60 0.64 0.10 * 0.434 0.670 MDA (nmol/mg of protein) 3.13 2.32 1.47 1.70 0.37 ** 0.439 0.174 Protein carbonylation (nmol/mg 10.7 11.8 6.82 7.19 1.96 * 0.708 0.847 of protein) 1LAT = low ambient temperature; NAT = normal ambient temperature. *P < 0.05. **P < 0.01. EFFECTS OF LOW TEMPERATURES AND VITAMIN C ON BROILERS 903 A similar result was reported in a previous study, which Hypoxia that is induced by LAT conditions can also revealed that adding VC (500 mg/kg) to the diet had result in oxidation reactions. When oxidation overcomes no significant effect on growth performance (Ladmakhi the body’s ability to protect itself using antioxidants, et al., 1997). oxidative stress occurs (Iqbal et al., 2002). Mujahid Shlosberg et al. (1992) reported that LAT stress and Furuse (2009) have observed some oxidative stress caused elevated HCT, which would lead to AS. Wang in the brains and hearts of broilers exposed to LAT et al. (2012) reported that HCT, HGB levels, and RBC conditions. The MDA compound is an end product of counts of broilers with AS were higher than those of the oxidative degradation of lipids and is an indicator healthy broilers, suggesting that hypoxic conditions of oxidative stress (Del Rio et al., 2005). Heat stress, caused the levels of these substances to increase to ful- which causes hypoxia, has been shown to increase MDA fill tissues’ demand for oxygen. Increases in the HCT levels in liver tissue, both in vitro and in vivo (Feng et Downloaded from and HGB levels have been observed in broilers in LAT al., 2008). One study found that when broilers were conditions, and blood volume values were significant- exposed to LAT environment for 3 wk, MDA level in ly elevated at 10°C than at 20°C (Yahav et al., 1997; the plasma increased (Pan et al., 2005). Along with the Wideman et al., 1998; Wideman and Tackett, 2000; MDA level, the amount of protein carbonylation in the Luger et al., 2001). In the present study, LAT condi- body, which is a form of protein oxidation, is also an http://ps.oxfordjournals.org/ tions markedly increased the HCT and HGB levels at indicator of the body’s degree of oxidation. In agree- 14 d, suggesting that LAT conditions had elevated the ment with Aruoma (1994), it shown that free radicals possibility of birds developing AS. can oxidize both lipids and proteins, worsening disease Even though many studies have investigated the lev- conditions in broilers. els of these blood parameters related to AS, there are In our study, LAT conditions also decreased T-AOC few reports about the effect of VC on blood parameters in the serum, liver, and lungs, but significantly in-

of broilers. In calves experiencing heat stress, one study creased the VC level in serum and liver. It is possible at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 showed that a VC supplementation (20 g/d) decreased that the hypoxic conditions resulted in an inflamma- their RBC counts, HCT, and HGB levels (Kim et al., tory response, as the level of VC is usually higher in 2012). When investigating these blood parameters in regions of inflammation than in areas that are not the broilers, Tras et al. (2000) found that VC (100 mg/ focus of an inflammation response. In tissues and cells, kg of diet) had no significant effect. In our study, RBC many specific enzymes, such as SOD and GSH-Px, can counts, HCT, and HGB levels in broilers that were giv- eliminate ROS that arise from oxidative stress. The ac- en a VC supplement increased at 21 d, especially for tivity of GSH-Px in the mitochondria of the lungs and broilers in LAT conditions. This was a compensation in the liver tissue of broilers with AS have been found mechanism. Because in some condition, the lower HCT to be higher than its activity in normal broilers (Iqbal level indicated less erythrocytes in blood volume, and et al., 2002), indicating that the amount of GSH-Px it induced a greater risk of hypoxia. Broilers with low- in tissues increases to eliminate ROS when necessary. er HCT level had significantly higher mortality other However, when the amount of ROS in mitochondria than mortality caused by AS (Shlosberg and Bellaiche, exceeded the antioxidant abilities provided by SOD 1996). The increase of RBC counts, HGB, and HCT and GSH-Px activities, mitochondrial injury would oc- levels due to VC addition were not anticipated; there- cur (Bottje and Wideman, 1995). However, Wang et fore, the mechanism was not investigated in this study. al. (2012) reported that the total SOD activity in the These results provide further evidence that the increase liver significantly decreased in broilers with AS, which in RBC counts, HCT, and HGB levels induced by LAT indicates that AS may severely damage the mechanism conditions resulted from the increased need for oxygen for SOD production in the liver. In our experiment, the in broilers. activities of SOD and GSH-Px in the serum were simi- Hypoxia has been shown to increase the amount of lar under both LAT and NAT conditions. This result reactive oxygen species (ROS) in cells and mitochon- might be due to the short study period (21 d) and that dria and disturb the oxidation-reduction system in the fact that only LAT was used as an AS-inducing factor. body, as well as stabilize the hypoxia-inducible factor, Vitamin C has been found to be a good scavenger leading to vascular remodeling, which is the foundation of oxygen (Cort, 1974), and it can alleviate oxidative of AS (Chandel et al., 1998, 2000). Free radicals or ROS stress. However, one study showed that VC (100 mg/ can result in lipid and protein peroxidation, and if it is kg of diet) had no significant effect on the level of SOD not reduced by antioxidant protective mechanisms, tis- in the serum (Tras et al., 2000). In our study, the ac- sues would be damaged. Supplementation with VC has tivity of GSH-Px in serum decreased upon addition of been found to alleviate the oxidative damage induced VC to the diet. Our results are in accordance with re- by ROS and therefore allow cells to maintain their sults from a previous study, where dietary vitamin E proper redox state (Blagojeviü, 2007). Broilers living reduced the activity of GSH-Px in ascitic broilers as in conditions of low ventilation have been reported as this vitamin acts also as an antioxidant in vivo (Iqbal having low levels of VC in their liver and lungs, which et al., 2002). The antioxidant effects of VC were only suggests that VC is consumed in the oxidation reac- evident as indicated by the increasing T-AOC levels in tions caused by hypoxia (Enkvetchakul et al., 1993). the current study. Because oxidative stress caused by 904 YANG ET AL. LAT conditions was severe, supplemental VC did not it needs further study when broilers are reared under improve the antioxidant capacity. Vitamin C generally LAT conditions. acted as a co-antioxidant that functioned by cooperating with other antioxidants, for instance, with vitamin ACKNOWLEDGMENTS E. Vitamin C was shown to restore the antioxidant capability of vitamin E by converting the tocopherol We thank Todd. J. Applegate and Chen Xi (Ani- radical back to its reduced state (Witting and Horwitt, mal Science Department, Purdue University, West La- 1964; Carr et al., 2000; Guney et al., 2007). Vitamin fayette, IN) for their role in revising the paper. This C cooperating with vitamin E can become a scavenger research was supported by grants from the National of ROS to protect NO, which was a substance main-

Natural Science Foundation of China (31101733). Downloaded from taining pulmonary vasodilatation (Ruiz-Feria, 2009; Bautista-Ortega and Ruiz-Feria, 2010). In our study, vitamin E was not added to the basal diet, and soybean REFERENCES oil was included, which was apt to be oxidized. Vitamin Aengwanich, W., and S. Simaraks. 2004. Pathology of heart, lung, C with inappropriate supplementation can also become liver and kidney in broilers under chronic heat stress. Pathology http://ps.oxfordjournals.org/ a pro-oxidative agent by reducing Fe3+ to Fe2+, which 26:418. then generated free radicals by interacting with Cu2+ Aruoma, O. I. 1994. Free radicals and antioxidant strategies in (Bottje and Wideman, 1995). This suggests that the sports. J. Nutr. Biochem. 5:370–381. Bautista-Ortega, J., and C. A. Ruiz-Feria. 2010. l-Arginine and proper cooperative substance and the additive amount antioxidant vitamins E and C improve the cardiovascular per- of VC should be continually explored when the broilers formance of broiler chickens grown under chronic hypobaric hy- are reared under LAT conditions. poxia. Poult. Sci. 89:2141–2146. Blagojeviü, D. P. 2007. Antioxidant systems in supporting environ- The change in antioxidant capacity (T-AOC) ob- mental and programmed adaptations to low temperatures. Cryo served in this study indicated that hypoxia, induced Letters 28:137–150. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 by LAT conditions, had a marked effect on the dif- Bottje, W. G., and R. F. Wideman. 1995. Potential role of free radi- ferent tissues, especially the lungs. Edema and hyper- cals in the pathogenesis of pulmonary hypertension syndrome. Poult. Avian Biol. Rev. 6:211–231. emia of the lungs have been observed in broilers expe- Carr, A. C., B. Z. Zhu, and B. Frei. 2000. Potential antiatherogenic riencing chronic heat stress, which causes hypoxia, and mechanisms of ascorbate (vitamin C) and α-tocopherol (vitamin hemorrhaging was observed in their alveoli, indicating E). Circ. Res. 87:349–354. that lung lesions are common consequences of hypoxia Chandel, N. S., E. Maltepe, E. Goldwasser, C. E. Mathieu, M. C. Si- mon, and P. T. Schumacker. 1998. Mitochondrial reactive oxygen (Aengwanich and Simaraks, 2004). When AS is pres- species trigger hypoxia-induced transcription. Proc. Natl. Acad. ent, cell growth and migration of pulmonary arterioles Sci. USA 95:11715–11720. have been found (Nain et al., 2009); pulmonary arteri- Chandel, N. S., D. S. McClintock, C. E. Feliciano, T. M. Wood, oles become more muscular, evident through observa- J. A. Melendez, A. M. Rodriguez, and P. T. Schumacker. 2000. Reactive oxygen species generated at mitochondrial complex III tion of thicker media and cellular proliferation in the in- stabilize hypoxia-inducible factor-1α during hypoxia. J. Biol. tima, which results in pulmonary hypertension (Xiang Chem. 275:25130–25138. et al., 2002). In addition, a study by Iqbal et al. (2001) Cort, W. M. 1974. Antioxidant activity of tocopherols, ascorbyl pal- showed that hydrogen peroxide, another biomarker of mitate, and ascorbic acid and their mode of action. J. Am. Oil Chem. Soc. 51:321–325. oxidative stress, was overproduced in the lung mito- Daneshyar, M., H. Kermanshahi, and A. Golian. 2009. Changes of chondria of broilers experiencing pulmonary hyperten- biochemical parameters and enzyme activities in broiler chickens sion syndrome (Iqbal et al., 2001). Similarly, the pres- with cold-induced ascites. Poult. Sci. 88:106–110. Decuypere, E., C. Vega, T. Bartha, J. Buyse, J. Zoons, and G. A. ent study showed that LAT conditions worsened the A. Albers. 1994. Increased sensitivity to triiodothyronine (T3) oxidative stress in broilers, as evidenced by changes in of broiler lines with a high susceptibility for ascites. Br. Poult. the levels of T-AOC, MDA, and protein carbonylation. Sci. 35:287–297. Although it has been reported that VC (500 mg/kg of Del Rio, D., A. J. Stewart, and N. Pellegrini. 2005. A review of recent studies on malondialdehyde as toxic molecule and bio- diet) can alleviate AS by reducing the muscularization logical marker of oxidative stress. Nutr. Metab. Cardiovasc. Dis. of the pulmonary arterioles (Xiang et al., 2002), it had 15:316–328. little effect on maintaining proper lung function in the Enkvetchakul, B., W. Bottje, N. Anthony, R. Moore, and W. Huff. present study. 1993. Compromised antioxidant status associated with ascites in broilers. Poult. Sci. 72:2272–2280. In conclusion, the natural LAT scheme employed in Feng, J., M. Zhang, S. Zheng, P. Xie, and A. Ma. 2008. Effects of this experiment caused oxidative stress of broilers over high temperature on multiple parameters of broilers in vitro and a 21-d period, as shown by the significant increase in in vivo. Poult. Sci. 87:2133–2139. HCT, poor growth, and low antioxidant capacities. Al- Guney, M., B. Oral, H. Demirin, N. Karahan, T. Mungan, and N. Delibas. 2007. Protective effects of vitamins C and E against though the LAT conditions failed to induce large-scale endometrial damage and oxidative stress in fluoride intoxication. incidence of AS, it can be used as a model for future Clin. Exp. Pharmacol. Physiol. 34:467–474. studies that focusing on the mechanisms by which hy- Huchzermeyer, F. W., and A. M. De Ruyck. 1986. Pulmonary hypoxia leads to oxidative stress. The effect of the di- pertension syndrome associated with ascites in broilers. Vet. Rec. 119:94. etary VC supplementation on the antioxidant capacity Iqbal, M., D. Cawthon, K. Beers, R. F. Wideman, and W. G. Bot- of 21-d-old broilers was not significant, indicating that tje. 2002. Antioxidant enzyme activities and mitochondrial fatty EFFECTS OF LOW TEMPERATURES AND VITAMIN C ON BROILERS 905

acids in pulmonary hypertension syndrome (PHS) in broilers. Oyanagui, Y. 1984. Reevaluation of assay methods and establish- Poult. Sci. 81:252–260. ment of kit for superoxide dismutase activity. Anal. Biochem. Iqbal, M., D. Cawthon, R. F. Wideman, and W. G. Bottje. 2001. 142:290–296. Lung mitochondrial dysfunction in pulmonary hypertension syn- Pan, J. Q., X. Tan, J. C. Li, W. D. Sun, and X. L. Wang. 2005. drome. I. Site-specific defects in the electron transport chain. Effects of early feed restriction and cold temperature on lipid Poult. Sci. 80:485–495. peroxidation, pulmonary vascular remodelling and ascites mor- Julian, R. J., G. W. Friars, H. French, and M. Quinton. 1987. The bidity in broilers under normal and cold temperature. Br. Poult. relationship of right ventricular hypertrophy, right ventricular Sci. 46:374–381. failure, and ascites to weight gain in broiler and roaster chickens. Ruiz-Feria, C. A. 2009. Concurrent supplementation of arginine, vi- Avian Dis. 31:130–135. taminE, and vitamin C improve cardiopulmonary performance in Julian, R. J., I. McMillan, and M. Quinton. 1989. The effect of cold broilers chickens. Poult. Sci. 88:526–535. and dietary energy on right ventricular hypertrophy, right ven- Shlosberg, A., and M. Bellaiche. 1996. Hematocrit values and mor-

tricular failure and ascites in meat-type chickens. Avian Pathol. tality from ascites in cold-stressed broilers from parents selected Downloaded from 18:675–684. by hematocrit. Poult. Sci. 75:1–5. Julian, R. J., and E. J. Squires. 1994. Haematopoietic and right ven- Shlosberg, A., I. Zadikov, U. Bendheim, V. Handji, and E. Berman. tricular response to intermittent hypobaric hypoxia in meat-type 1992. The effects of poor ventilation, low temperatures, type of chickens. Avian Pathol. 23:539–545. feed and sex of bird on the development of ascites in broilers. Kim, J. H., L. L. Mamuad, C. J. Yang, S. H. Kim, J. K. Ha, W. S. Physiopathological factors. Avian Pathol. 21:369–382. Lee, K. K. Cho, and S. S. Lee. 2012. Hemato-biochemical and Tras, B., F. Inal, A. L. Bas, V. Altunok, M. Elmas, and E. Yazar. cortisol profile of Holstein growing-calves supplemented with vi- 2000. Effects of continuous supplementations of ascorbic acid, http://ps.oxfordjournals.org/ tamin C during summer season. J. Anim. Sci. 25:361–368. aspirin, vitamin E and selenium on some haematological param- Ladmakhi, M. H., N. Buys, E. Dewil, G. Rahimi, and E. Decuypere. eters and serum superoxide dismutase level in broiler chickens. 1997. The prophylactic effect of vitamin C supplementation on Br. Poult. Sci. 41:664–666. broiler ascites incidence and plasma thyroid hormone concentra- Wang, Y., Y. Guo, D. Ning, Y. Peng, H. Cai, J. Tan, Y. Yang, and tion. Avian Pathol. 26:33–44. D. Liu. 2012. Changes of hepatic biochemical parameters and Luger, D., D. Shinder, V. Rzepakovsky, M. Rusal, and S. Yahav. proteomics in broilers with cold-induced ascites. J. Anim. Sci. 2001. Association between weight gain, blood parameters, and Biotechnol. 3:41. thyroid hormones and the development of ascites syndrome in Wideman, R. F. Jr., and C. D. Tackett. 2000. Cardio-pulmonary

broiler chickens. Poult. Sci. 80:965–971. function in broilers reared at warm or cool temperatures: Effect at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Maxwell, M. H., and G. W. Robertson. 1997. World broiler ascites of acute inhalation of 100% oxygen. Poult. Sci. 79:257–264. survey 1996. Poult. Int. 36:16–33. Wideman, R. F. Jr., T. Wing, Y. K. Kirby, M. F. Forman, N. Mar- Mujahid, A., and M. Furuse. 2009. Oxidative damage in different son, C. D. Tackett, and C. A. Ruiz-Feria. 1998. Evaluation of tissues of neonatal chicks exposed to low environmental tempera- minimally invasive indices for predicting ascites susceptibility in ture. Comp. Biochem. Physiol. A Mol. Integr. Physiol. 152:604– three successive hatches of broilers exposed to cool temperatures. 608. Poult. Sci. 77:1565–1573. Nain, S., B. Ling, B. Bandy, J. Alcorn, C. Wojnarowicz, B. Laarveld, Witting, L. A., and M. K. Horwitt. 1964. Effect of degree of fatty and A. A. Olkowski. 2008. The role of oxidative stress in the acid unsaturation in tocopherol deficiency-induced creatinuria. development of congestive heart failure in a chicken genotype J. Nutr. 82:19. selected for rapid growth. Avian Pathol. 37:367–373. Xiang, R. P., W. D. Sun, J. Y. Wang, and X. L. Wang. 2002. Effect Nain, S., C. Wojnarowicz, B. Laarveld, and A. A. Olkowski. 2009. of vitamin C on pulmonary hypertension and muscularisation Vascular remodeling and its role in the pathogenesis of ascites of pulmonary arterioles in broilers. Br. Poult. Sci. 43:705–712. in fast growing commercial broilers. Res. Vet. Sci. 86:479–484. Yahav, S., and J. P. McMurtry. 2001. Thermotolerance acquisition Ohkawa, H., N. Ohishi, and K. Yagi. 1979. Assay for lipid peroxides in broiler chickens by temperature conditioning early in life–the in animal tissues by thiobarbituric acid reaction. Anal. Biochem. effect of timing and ambient temperature. Poult. Sci. 80:1662– 95:351–358. 1666. Owen, R. L., R. F. Wideman Jr., A. L. Hattel, and B. S. Cowen. Yahav, S., A. Straschnow, I. Plavnik, and S. Hurwitz. 1997. Blood 1990. Use of a hypobaric chamber as a model system for investi- system response of chickens to changes in environmental tem- gating ascites in broilers. Avian Dis. 34:754–758. perature. Poult. Sci. 76:627–633.

Effect of calcium level and phytase addition on ileal phytate degradation and amino acid digestibility of broilers fed corn-based diets1

A. M. Amerah,* 2 P. W. Plumstead ,* L. P. Barnard ,* and A. Kumar †

* Danisco Animal Nutrition, DuPont Industrial Bioscience, Marlborough, SN8 1XN, UK; and † School of Agriculture and Food Science, University of Queensland, Gatton, Qld-4343, Australia Downloaded from ABSTRACT This study investigated the effect of di- ratio (FCR). In diets with no phytase, Ca:AvP ratio etary Ca to available P (AvP) ratio and phytase sup- had a greater effect on WG, FI, and FCR compared plementation on bone ash, ileal phytate degradation, with those fed diets without phytase. The orthogonal and nutrient digestibility in broilers fed corn-based di- polynomial contrasts showed that the increase in diets. The experimental design was a 4 × 2 factorial ar- etary Ca:AvP ratio significantly decreased WG and http://ps.oxfordjournals.org/ rangement of treatments evaluating 4 Ca:AvP ratios FI in a quadratic manner, whereas FCR increased (P (1.43, 2.14, 2.86, and 3.57) and 2 levels of phytase (0 < 0.05) linearly with higher dietary Ca:AvP ratio. In- and 1,000 phytase units/kg of feed). The 4 Ca:AvP ra- creasing dietary Ca:AvP ratio led to a significant qua- tios were achieved by formulating all diets to a constant dratic decrease in phytate degradation and significant AvP level of 0.28% and varying Ca levels (0.4, 0.6, 0.8, linear decreases in P digestibility and bone ash. Phy- and 1.0%). Each treatment was fed to 6 cages of 8 male tase addition increased (P < 0.05) phytate degradation

Ross 308 broilers from 5 to 21 d. At 21 d, digesta from and improved (P < 0.05) energy, AA, and P digest- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 the terminal ileum was collected and analyzed for ener- ibility at all levels of Ca:AvP with no interaction (P gy, phytate, P, Ca, and amino acids (AA) to determine > 0.05) between the main factors. Digestibility of AA digestibility. Digesta pH was measured in each segment was positively correlated (P < 0.05) with the degree of (crop, gizzard, duodenum, and ileum) of the digestive phytate degradation. Increasing dietary Ca:AvP ratio tract. Data were analyzed by 2-way analysis of covari- significantly increased gizzard pH in a linear manner.

ance. There was a significant interaction between di- In conclusion, phytase (1,000 phytase units/kg of feed) etary Ca:AvP ratio and phytase supplementation for improved phytate, and P and AA digestibility at all weight gain (WG), feed intake (FI), and feed conversion Ca:AvP ratios evaluated in this study. Key words: broiler , available phosphorus, calcium , amino acid, phytate degradation 2014 Poultry Science 93 :906–915 http://dx.doi.org/ 10.3382/ps.2013-03465

INTRODUCTION bial phytase to poultry diets enhanced the utilization of nutrients other than P, including energy and amino Microbial phytases have been widely used in poultry acids (AA; Ravindran et al., 1999; Selle et al., 2000; feeds as a means of improving dietary P availability and Selle and Ravindran, 2007). However, the magnitude reducing P excretion in manure. The effect of microbial of response has not been consistent in all studies, and phytases on phytate degradation and subsequent im- factors contributing to the variability in energy and AA provements in P utilization in poultry diets has been digestibility improvements in response to supplemental well documented (Selle and Ravindran, 2007). Several phytase have not been delineated (Selle and Ravindran, studies have also shown that dietary additions of micro- 2007). Potential factors that have been reported to contribute to the variation in response have included the concentration of the substrate for the enzyme (phytate) in the diet, the level of added phytase, the intrinsic © 2014 Poultry Science Association Inc. properties of the phytase enzyme used, source of phy- Received July 4, 2013. tase, and feed particle size (Ravindran et al., 2006; Am- Accepted November 29, 2013. 1 This is an Open Access article distributed under the terms of erah and Ravindran, 2009; Selle et al., 2009). Another the Creative Commons Attribution-Noncommercial License (http:// factor that may have contributed to the inconsistent re- creativecommons.org/licenses/by-nc/3.0/), which permits noncom- sults in the literature was the formation of Ca-phytate mercial use, distribution, and reproduction in any medium, provided the original work is properly cited. complexes that were not susceptible to degradation by 2 Corresponding author: [email protected] phytase (Angel et al., 2002; Selle et al., 2009).

906 CALCIUM LEVEL AND PHYTASE ADDITION IN BROILER DIETS 907 The ability of phytase enzymes to hydrolyze phytate 3.57), and each basal diet was then divided into 2 equal has been shown to be negatively affected by high levels parts. One part was supplemented with phytase, and of dietary Ca or a high ratio of Ca:AvP (Qian et al., the other part remained unsupplemented. The phytase 1997). As Ca-phytate complexes were formed at pH (Axtra PHY, Danisco Animal Nutrition, Marlborough, ≥5 (Selle et al., 2009), similar to the conditions in the UK) was added to provide 1,000 phytase units (FTU)/ small intestine, rapid phytate degradation in the proxi- kg of feed. Celite, a source of acid-insoluble ash (AIA), mal gut where the digesta was more acidic (gizzard + was added at 0.2% to all experimental diets as an indi- proventriculus) using phytase enzymes has been pro- gestible marker to determine nutrient digestibility. posed to mitigate the negative effect of Ca on phytate degradation and increase nutrient availability for the

Measurements Downloaded from bird. However, phytases have been suggested to vary in their pH optima with efficiency and speed of hydro- At 21 d of age, all birds in each group were eutha- lyzing phytate depending on their source (Tran et al., nized by cervical dislocation. The ileum was then im- 2011). Therefore, the objectives of this study were to mediately excised and divided into 2 parts, the ante- examine the interactions of dietary Ca:AvP ratio and rior and posterior ileum. The ileum was defined as the addition of a bacterial phytase isolated from Buttiaux- portion of the small intestine extending from Meckel’s http://ps.oxfordjournals.org/ ella spp. on ileal phytate degradation, Ca, P, and AA diverticulum to a point 40 mm proximal to the ileo- digestibility in broiler chickens. cecal junction. Contents of the posterior ileum were collected by gently flushing with distilled water into MATERIALS AND METHODS plastic containers. Digesta were pooled within a cage, lyophilized, ground to pass through a 0.5-mm screen Birds and Housing size, and stored at −20°C until analyzed for gross en-

ergy, nitrogen, Ca, P, phytate, and AA. At 21 d of age, at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Experimental procedures were conducted in accor- pH of the contents of the crop, gizzard, duodenum, and dance with the University of Queensland Animal Eth- ileum was also measured using an Alpha pH test meter ics Committee guidelines and model code of practice (Coral Cay Health, Surfers Paradise, Queensland, Aus- for welfare of animals. A total of 384 one-day-old male tralia) and the left tibia from each bird was collected Ross 308 broiler chickens were obtained from a com- for bone ash analysis. mercial hatchery. One-day-old chickens were selected according to mean BW and 8 chickens of 42 ± 3 g were Chemical Analysis allocated to each of 48 digestibility cages housed in an environmentally controlled room. Feed and water for Gross energy was determined from the temperature ad libitum consumption was offered during the trial pe- increase when the sample was ignited in an oxygen-rich riod. Brooding temperature was maintained at 31°C for atmosphere in a bomb calorimeter (Leco AC600, NSW, the first 7 d then gradually reduced to 27°C at the end Australia). Nitrogen was determined by the AOAC of the 3-wk experimental period. The broiler chickens method 990.03 (AOAC, 1990) where the sample was in this trial were on 23 h light with a minimum light ignited in an oxygen atmosphere and the nitrogen de- intensity of 20 lx and 1 h dark per day. The lighting termined in the gas stream by a thermal conductivity program and the environment in the house were con- meter after stripping out the oxygen and replacing it trolled by timers and temperature sensors, respectively. with helium. Calcium and total P in both feed and ileal All chickens were vaccinated against infectious bronchi- samples were determined by ICP-OES method 2011.14 tis at the hatchery. Body weight and feed intake (FI) (AOAC, 1990) analysis following microwave-assisted were recorded by cage at 5 and 21 d of age. Mortality acid digestion. The AA analysis was performed using was recorded daily. Any bird that died was weighed and 6 M HCl liquid hydrolysis with AA quantitation by feed conversion ratio (FCR) values were calculated by AccQTag derivatization (Waters Corporation, Milford, dividing total feed intake by weight gain (WG) of live MA) and reversed-phase ultra-performance liquid chro- plus dead birds. matography analysis (Cohen and DeAntonis, 1994). Left tibia bones were autoclaved for 30 min, and the Diets and Treatments adhering tissue was removed mechanically. The clean bones were dried in an oven at 100°C for 24 h, and dried From 1 to 5 d, all birds in the experiment received bones were ground. According to AOAC method 942.05 a commercial broiler starter diet. The experimental di- (AOAC, 1990), 2 g of ground bone sample was ashed ets were offered from 5 to 21 d of age and were based at 600°C using a thermo gravimetric analyzer. Calcium mainly on corn and soybean meal (Table 1). Diets were and P contents of tibia ash were analyzed by the ICP- formulated to be similar to commercial diets fed to the OES method 2011.14 (AOAC, 1990). Phytate in both Ross 308 strain of broiler, except Ca and P. This study feed and ileal samples was extracted by the method had 8 experimental diets. Four basal diets were formu- described by Latta and Eskin (1980) and determined lated with different Ca:AvP ratios (1.43, 2.14, 2.86, and by ICP-AES. Acid insoluble ash was determined by 908 AMERAH ET AL.

Table 1. Composition and calculated analysis (g/100 g as fed) of the basal diet1

Formulated Ca:available P ratio

Item 1.43 2.14 2.86 3.57 Ingredient Corn 62.36 61.33 60.29 59.25 Soybean meal 48% CP 34.03 34.21 34.40 34.58 Soybean oil 0.68 1.01 1.34 1.66 dl-Methionine 0.31 0.31 0.31 0.31 l-Threonine 0.12 0.12 0.12 0.12 l-Lysine HCl 0.30 0.30 0.29 0.29

Salt 0.32 0.33 0.33 0.33 Downloaded from Limestone 0.30 0.83 1.35 1.87 Dicalcium phosphate 0.98 0.98 0.98 0.98 Vitamin and trace mineral premix1 0.30 0.30 0.30 0.30 Indigestible marker 0.20 0.20 0.20 0.20 Calculated analysis CP 22.00 22.00 22.00 22.00 ME, kcal/kg 2,990 2,990 2,990 2,990 http://ps.oxfordjournals.org/ Total P 0.55 0.55 0.55 0.55 Available P 0.28 0.28 0.28 0.28 Ca 0.40 0.60 0.80 1.00 Digestible lysine 1.24 1.24 1.24 1.24 Digestible methionine + cysteine 0.87 0.87 0.87 0.87 Analyzed value Ca 0.51 0.68 0.91 1.30 Total P 0.51 0.51 0.51 0.51 Phytate P 0.32 0.32 0.32 0.32 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 1Supplied per kilogram of diet: antioxidant, 50 mg; biotin, 0.2 mg; calcium pantothenate, 12 mg; cholecalciferol, 60 μg; cyanocobalamin, 0.017 mg; folic acid, 5.2 mg; menadione, 4 mg; niacin, 50 mg; pyridoxine, 10 mg; trans- retinol, 3.33 mg; riboflavin, 12 mg; thiamine, 3.0 mg; dl-α-tocopheryl acetate, 40 mg; choline chloride, 638 mg; Co, 0.3 mg; Cu, 3 mg; Fe, 25 mg; I, 1 mg; Mn, 125 mg; Mo, 0.5 mg; Se, 200 μg; Zn, 60 mg.

AOAC method 975.12 (AOAC, 1990) by boiling ash in RESULTS 25 mL of HCl for 5 min, collecting insoluble matter on an ashless filter, washing with hot water until washings Diet Analysis are acid-free, igniting until chloride-free, cooling, and weighing the sample. Analyzed dietary Ca, P, and phytate P are shown in Table 1. The analyzed dietary phytate P level was Calculations higher than calculated values, which can be explained by higher phytate P values than expected in the feed The apparent ileal nutrient digestibility/absorption ingredients used. The analyzed phytate P values for coefficients were calculated by the following formula us- corn and soybean meal were 0.27 and 0.61%, respec- ing AIA as the indigestible marker (Ravindran et al., tively. The calculated values for phytate P for corn and 2006): soybean meal were 0.18 and 0.42%, respectively. Phy- tase recovery was below target but within an accept- apparent nutrient digestibility % = {[(NT/AIA)d able range (mean 831 FTU/kg of diet). Analyzed Ca was also slightly higher and analyzed P slightly lower − (NT/AIA)i]/(NT/AIA)d} × 100, than formulated values and would have increased the ratio of Ca:AvP slightly above formulated values. where (NT/AIA)d = ratio of nutrient and AIA diet, and (NT/AIA)i = ratio of nutrient and AIA in ileal digesta. Bird Performance

Data Analysis Effects of dietary Ca:AvP ratio and phytase supplementation on WG (Figure 1), FI, and FCR are sum- Data were analyzed by 2-way analysis of covariance marized in Table 2. There was a significant interaction using the GLM procedure of SAS Institute Inc. (2004, between dietary Ca:AvP ratio and phytase supplemen- Cary, NC) using cage mean as an experimental unit. A tation for WG, FI, and FCR. In diets with no phytase, probability value of P < 0.05 was described to be sta- Ca:AvP ratio had a greater effect on WG, FI, and FCR tistically significant. Orthogonal polynomial contrasts compared with those fed diets without phytase. The or- were used to assess the significance of linear or qua- thogonal polynomial contrasts showed that the increase dratic models to describe the response in the dependent in dietary Ca:AvP ratio significantly decreased WG variable to Ca:AvP level with or without added phy- and FI in a quadratic manner, whereas FCR increased tase using the PROC MIXED procedure of SAS. (P < 0.05) linearly with higher dietary Ca:AvP ratio. CALCIUM LEVEL AND PHYTASE ADDITION IN BROILER DIETS 909 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 1. Regression of weight gain in broilers fed diets containing different Ca:available P (AvP) ratios, in the presence and absence of a supplementary microbial phytase from Buttiauxella. RMSE = root mean square error; FTU = phytase units.

Table 2. Effect of varying dietary Ca:available P (AvP) ratios and microbial phytase on the weight gain (g/bird), feed intake (g/bird), and feed conversion ratio (g/g) in broilers fed a corn/soy-based diet (5–21 d posthatch)1

Feed Phytase Weight Feed conversion Item (FTU/kg) gain intake ratio Ca:AvP ratio 1.43 0 614 819 1.337 2.14 0 589 791 1.345 2.86 0 500 687 1.375 3.57 0 404 596 1.482 1.43 1,000 679 862 1.269 2.14 1,000 688 877 1.274 2.86 1,000 650 854 1.316 3.57 1,000 607 745 1.234 SEM2 9.0 24 0.04 Main effect Ca:AvP ratio 1.43 646 841 1.303 2.14 639 834 1.309 2.86 575 770 1.345 3.57 496 670 1.358 Phytase (FTU/kg) 0 527 723 1.384 1,000 658 834 1.273 P ≤ Ca:AvP <0.0001 <0.0001 0.1164 Phytase <0.0001 <0.0001 0.0003 Ca:AvP × phytase <0.0001 0.0176 0.0406 Ca:AvP (linear)3 <0.0001 <0.0001 <0.0001 Ca:AvP (quadratic)3 0.0077 0.0141 0.1088 1Each value represents the mean of 6 replicates (8 birds per replicate). FTU = phytase units. 2Pooled SEM. 3Orthogonal polynomial contrasts were used to assess the significance of linear or quadratic models to describe the response in the dependent variable to Ca:AvP level with or without added phytase. 910 AMERAH ET AL. Energy, Phytate, P, and Ca Digestibility Table 4. Phytase supplementation improved (P < 0.05) digestibility of all AA. Dietary Ca:AvP ratio influenced The effects of dietary Ca:AvP ratios and microbial (P < 0.05) digestibility of all AA apart from Arg, Gly, phytase on bone ash, energy, phytate, P, and Ca digest- Tyr, Val, and Cys. Increasing dietary Ca:AvP ratio ibility are summarized in Table 3. Phytase supplemen- reduced (P < 0.05) AA digestibility in a linear man- tation increased (P < 0.05) phytate degradation, en- ner. Mean AA digestibility was improved (P < 0.05) by ergy digestibility, and P digestibility. Dietary Ca:AvP phytase supplementation (Figure 2). ratio had no effect (P > 0.05) on energy digestibility but significantly influenced phytate degradation and Regression for AA Digestibility P digestibility. Phytase supplementation and dietary Ca:AvP ratio had no effect (P > 0.05) on Ca digest- The AA digestibility significantly increased in a cur- Downloaded from ibility. There was no significant interaction between di- vilinear manner with increasing phytate degradation (y etary Ca:AvP ratio and phytase supplementation on = −0.0031x2 + 0.4942x + 64.667, R2 = 0.74, P < 0.05), phytate degradation, energy, P, and Ca digestibility. Figure 3. However, there was a significant interaction between dietary Ca:AvP ratio and phytase supplementation for Gastrointestinal pH http://ps.oxfordjournals.org/ bone ash. Increasing the dietary Ca:AvP ratio in the absence of phytase reduced bone ash; conversely, in the The effect of dietary Ca:AvP ratio and microbial presence of phytase, bone ash was increased at higher phytase supplementation on gastrointestinal pH is sum- Ca:AvP ratios. Increasing dietary Ca:AvP ratio led to marized in Table 5. Diets receiving phytase exhibited a significant quadratic decrease in phytate degradation a significant increase in gizzard and ileum pH. The and significant linear decreases in P digestibility and Ca:AvP ratio only influenced (P < 0.05) the pH of bone ash. gizzard digesta. Increasing dietary Ca:AvP ratio signifi- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 cantly increased gizzard pH in a linear manner. There AA Digestibility was no effect (P > 0.05) of dietary Ca:AvP ratio on the pH of the other regions of the gastrointestinal tract: The effects of dietary Ca:AvP ratio and phytase crop, duodenum, and ileum. There was a significant supplementation on individual AA are summarized in interaction between dietary Ca:AvP ratio and phytase

Table 3. Effect of varying dietary Ca:available P (AvP) ratios and microbial phytase on energy digestibility (%), phytate degradation (%), Ca and P digestibility (%), and bone ash (%) of broilers fed a corn/soy-based diet1

Gross energy Phytate Ca P Bone Phytase digestibility degradation digestibility digestibility ash Item (FTU/kg) (%) (%) (%) (%) (%) Ca:AvP ratio 1.43 0 67.4 51.4 53.8 61.5 43.9 2.14 0 64.6 40.4 50.4 52.5 41.1 2.86 0 66.8 43.7 54.9 55.3 39.3 3.57 0 65.8 39.8 61.8 51.2 38.0 1.43 1,000 74.6 88.4 56.1 80.4 45.1 2.14 1,000 74.1 75.2 40.7 72.4 47.6 2.86 1,000 69.7 76.2 49.3 67.6 48.6 3.57 1,000 73.9 75.9 58.6 67.6 46.8 SEM2 1.5 2.6 3.7 2.2 0.7 Main effect Ca:AvP ratio 1.43 71.0 69.9 55.0 71.0 44.5 2.14 69.4 57.8 45.6 62.5 44.4 2.86 68.3 59.9 52.1 61.5 43.9 3.57 69.9 57.9 60.2 59.2 42.4 Phytase (FTU/kg) 0 66.2b 43.8b 55.2 55.1b 40.6 1,000 73.1a 78.9a 51.2 71.9a 47.0 P-value Ca:AvP 0.3795 0.0006 0.0932 <0.0001 0.005 Phytase <0.0001 <0.0001 0.1676 <0.0001 <0.0001 Ca:AvP × phytase 0.6935 0.7821 0.6362 0.2759 <0.0001 Ca:AvP (linear)3 0.3691 0.0003 0.0613 <0.0001 0.0268 Ca:AvP (quadratic)3 0.1514 0.0059 0.0016 0.0568 0.2950 a,bMeans in a column not sharing a common superscript are different (P < 0.05). 1Each value represents the mean of 6 replicates (8 birds per replicate). FTU = phytase units. 2Pooled SEM. 3Orthogonal polynomial contrasts were used to assess the significance of linear or quadratic models to describe the response in the dependent variable to Ca:AvP level with or without added phytase. Table 4. Apparent ileal amino acid (AA) digestibility (%) of broilers fed a corn/soy-based diet with varying dietary Ca:available P (AvP) ratios supplemented with or without microbial phytase1

Phytase Mean CALCIUM LEVELANDPHYTASE ADDITIONINBROILER DIETS Item (FTU/kg) His Ser Arg Gly Asp Glu Thr Ala Pro Lys Tyr Met Val Iso Leu Phe Cys AA Ca:AvP ratio 1.43 0 75.6 72.4 85.6 68.9 79.9 83.8 69.7 73.7 74.9 84.7 77.8 89.5 68.5 75.2 75.6 76.9 57.5 78.0 2.14 0 70.4 65.3 79.3 62.1 73.9 78.6 62.7 67.2 70.4 77.1 73.4 87.4 64.2 69.4 69.9 71.6 51.8 72.3 2.86 0 75.7 70.9 84.9 69.2 76.9 82.2 68.8 72.9 73.7 84.3 77.8 90.4 69.4 75.1 74.5 75.8 54.1 77.0 3.57 0 70.2 65.1 79.3 63.1 71.5 78.4 63.2 65.7 69.8 76.2 73.0 84.8 65.5 68.7 68.9 70.6 53.7 71.8 1.43 1,000 84.6 84.3 90.8 79.6 87.9 89.9 79.2 83.7 83.6 89.6 85.9 93.6 81.4 85.0 85.5 86.3 72.6 86.4 2.14 1,000 83.0 82.0 90.2 77.7 85.9 88.9 77.4 82.0 82.3 89.5 85.1 93.7 77.9 83.0 83.5 84.3 69.3 84.9 2.86 1,000 80.5 79.2 89.7 74.1 83.8 87.0 73.9 78.0 79.3 88.6 82.9 92.5 75.0 80.0 80.3 81.3 65.1 82.4 3.57 1,000 80.3 78.1 89.3 74.6 82.5 87.0 75.4 79.0 79.7 87.9 83.7 91.5 75.2 80.4 80.9 81.6 67.1 82.4 SEM2 1.70 2.08 1.37 2.16 1.52 1.23 2.04 1.94 1.50 1.58 1.55 1.26 2.13 1.86 1.78 1.66 2.33 1.55 Main effect Ca:AvP ratio 1.43 80.1 78.4 88.2 74.3 83.9 86.8 74.4 78.7 79.2 87.2 81.9 91.6 75 80.1 80.6 81.6 65.1 82.2 2.14 76.7 73.7 84.8 69.9 79.9 83.8 70.0 74.6 76.4 83.3 79.2 90.6 71.0 76.2 76.7 78.0 60.6 78.6 2.86 78.1 75.0 87.3 71.6 80.4 84.6 71.4 75.4 76.5 86.4 80.3 91.5 72.2 77.5 77.4 78.5 59.6 79.7 3.57 75.2 71.6 84.3 68.9 77.0 82.7 69.3 72.4 74.7 82.1 78.4 88.1 70.4 74.5 74.9 76.1 60.4 77.1 Phytase (FTU/kg) 0 73.0b 68.4b 82.3b 65.8b 75.6b 80.7b 66.1b 69.9b 72.2b 80.6 75.5b 88.0b 66.9b 72.1b 72.2b 73.7b 54.3b 74.8b 1,000 82.1a 80.9a 90.0a 76.5a 85.0a 88.2a 76.5a 80.7a 81.2a 88.9 84.4a 92.8a 77.4a 82.1a 82.6a 83.4a 68.5a 84.0a P-value Ca:AvP 0.0238 0.0078 0.0656 0.0505 0.0002 0.0072 0.0436 0.0082 0.0083 0.0402 0.0709 0.0291 0.0688 0.0173 0.0079 0.0051 0.0501 0.0085 Phytase <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 Ca:AvP 0.6753 0.7033 0.4103 0.5587 0.6807 0.8126 0.9077 0.9827 0.7708 0.3059 0.8978 0.6813 0.1946 0.8134 0.8808 0.8292 0.4333 0.9883 × phytase Ca:AvP 0.0240 0.0077 0.0676 0.0508 0.0002 0.0070 0.0423 0.0082 0.0082 0.0424 0.0707 0.0268 0.0725 0.0172 0.0077 0.0050 0.0454 0.0090 (linear)3 Ca:AvP 0.8249 0.6873 0.8409 0.6345 0.7724 0.5401 0.4516 0.7228 0.5925 0.8421 0.7810 0.2176 0.4887 0.7539 0.5942 0.6183 0.1093 0.6640 (quadratic)3 a,bMeans in a column not sharing a common superscript are different (P < 0.05). 1Each value represents the mean of 6 replicates (8 birds per replicate). FTU = phytase units. 2Pooled SEM. 3Orthogonal polynomial contrasts were used to assess the significance of linear or quadratic models to describe the response in the dependent variable to Ca:AvP level with or without added phytase.

911

Downloaded from from Downloaded http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 6, April on University Harvard Zoology, Comp Museum the of Library Mayr Ernst at 912 AMERAH ET AL. Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 2. Regression of mean amino acid (AA) digestibility in broilers fed diets containing different Ca:available P (AvP) ratios, in the presence and absence of a supplementary microbial phytase from Buttiauxella. FTU = phytase units, RMSE = root mean square error. supplementation on duodenal pH whereby in the ab- the unsupplemented diets, compared with the phytase- sence of phytase an increase in dietary Ca:AvP ratio supplemented diets, as noted by the significant interac- did not have an effect on duodenal pH; conversely, in tion between dietary Ca:AvP ratio and phytase supple- the presence of phytase an increase in dietary Ca:AvP mentation. This can likely be attributed to diets with resulted in an increase in duodenal pH. no phytase being more P deficient, and a negative effect of Ca on Phytate P utilization. Tamim et al. (2004) DISCUSSION showed that in the absence of dietary Ca, broilers were able to use 69.2% of phytate P by the terminal ileum. In the current study phytase supplementation im- However, this was reduced to 25.4% when Ca levels proved WG at all Ca:AvP ratios. The detrimental effect were increased to 0.5%. Phytase supplementation miti- on WG from increasing Ca:AvP ratios was greater in gated the negative effect of increasing Ca:AvP ratio on

Figure 3. Regression of phytate degradation and mean amino acid digestibility in broilers fed diets containing different Ca:available P (AvP) ratios, in the presence and absence of a supplementary microbial phytase from Buttiauxella. CALCIUM LEVEL AND PHYTASE ADDITION IN BROILER DIETS 913

Table 5. Effect of varying dietary Ca:available P (AvP) ratios and microbial phytase on pH of the gastrointestinal tract in broilers fed a corn/soy-based diet1

Phytase Crop Gizzard Duodenum Ileum Item (FTU/kg) pH pH pH pH Ca:AvP ratio 1.43 0 4.5 2.8 5.9 5.5 2.14 0 4.3 3.2 6.0 5.2 2.86 0 4.8 3.2 5.9 5.3 3.57 0 4.5 3.3 5.7 5.4 1.43 1,000 4.7 2.5 5.7 6.4 2.14 1,000 4.7 2.4 5.7 6.4

2.86 1,000 4.8 2.4 5.6 6.7 Downloaded from 3.57 1,000 4.9 3.2 5.9 6.2 SEM2 0.2 0.2 0.1 0.2 Main effect Ca:AvP ratio 1.43 4.6 2.6 5.8 5.9 2.14 4.5 2.8 5.8 5.8 2.86 4.8 2.8 5.8 6.0 http://ps.oxfordjournals.org/ 3.57 4.7 3.2 5.8 5.8 Phytase (FTU/kg) 0 4.5 2.6b 5.7 5.4b 1,000 4.8 3.2a 5.9 6.4a P-value Ca:AvP 0.0975 0.0003 0.5999 0.4098 Phytase 0.0607 <0.0001 0.0552 <0.0001 Ca:AvP × phytase 0.2203 0.7385 0.0384 0.3655 Ca:AvP (linear)3 0.3364 0.0006 0.8111 0.7955 Ca:AvP (quadratic)3 0.9044 0.3508 0.9723 0.9389 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 a,bMeans in a column not sharing a common superscript are different (P < 0.05). 1Each value represents the mean of 6 replicates (8 birds per replicate). FTU = phytase units. 2Pooled SEM. 3Orthogonal polynomial contrasts were used to assess the significance of linear or quadratic models to describe the response in the dependent variable to Ca:AvP level with or without added phytase.

FI and FCR at all dietary Ca:AvP ratios. These data level of 0.5% to the diet, which resulted in a reduction may be explained by the higher nutrient digestibility of phytate-P disappearance from 69.2 to 25.4%. The observed in this study when the diets were supplement- negative effects of high levels of Ca on phytate degrada- ed with phytase. Improved broiler performance with tion and P digestibility may be explained by the forma- the addition of phytase to low P diets is to be expected tion of Ca-Phytate complexes and the increase in the (Selle and Ravindran, 2007). In contrast, Powell et al. pH of the proximal digestive tract (Selle et al., 2009). (2011) reported improved WG and FI with no effect The reason for the quadratic response to phytate deg- on FCR when the diet was supplemented phytase. In- radation in this study is not clear, but may be due to creasing Ca:AvP ratio reduced FI in the current study, saturation of the phytate binding sites with Ca ions, in which is in agreement with previous reports (Powell et which case increasing Ca concentration further would al., 2011; Delezie et al., 2012). have no additional effect. In the current study, increasing the Ca:AvP ratio re- Phytase supplementation increased phytate degrada- duced phytate degradation quadratically and P digest- tion and improved P digestibility at all Ca:AvP ratios. ibility linearly. One phytate molecule can bind up to In contrast to the findings of the current study, other 5 Ca atoms and the extent of this complex formation studies have shown that phytase effects on P retention has been reported to be driven by gut pH and molar were reduced at higher than optimal Ca:P ratios (Qian ratios of the 2 components (Selle et al., 2009). Previous et al. 1997). These researchers suggested that the extra reports have shown that the degree of phytate degra- Ca may directly suppress phytase activity by compet- dation was highly dependent on the dietary Ca level ing for the active sites of the enzymes. The inconsistent (Tamim et al., 2004; Plumstead et al., 2008; Selle et results of these data compared with that reported pre- al., 2009). Plumstead et al. (2008) reported a linear viously in the literature may be explained, partly, by reduction in ileal phytate P degradation by 71%, when the differences in Ca:AvP ratios tested, phytase level increasing dietary Ca level from 4.7 to 11.6 g/kg in used, and differences in the characteristics of the phy- broiler diets. Similar results were reported by Tamim tase enzymes used. The source of phytase used in these et al. (2004) in both in vitro and in vivo studies. These earlier studies was of fungal origin with a higher pH researchers found that dietary Ca at a level as low as optimum, which may be more prone to Ca inhibition. 0.1% reduced phytate-P hydrolysis at pH 6.5 in vitro. In their review, Selle et al. (2009) hypothesized that This effect has been proven in vivo by adding Ca at a as Ca-phytate complexes were mainly formed in the 914 AMERAH ET AL. small intestine and exogenous phytases of bacterial ori- current study, the addition of a bacterial phytase with gin would be more active in more proximal segments of a low pH optimum (Shukun Yu, DuPont Industrial Bio- the gut where the pH was closer to the optimum pH sciences, Aarhus, Denmark, personal communication) of the phytase. Further, phytate would be less likely would have degraded the phytate in the area of the pro- to bind Ca and the efficacy of these bacterial phytases ventriculous, reducing the production of the HCl from would be influenced to a lesser extent by Ca–phytate the proventriculus, which may have caused the increase complexes. in the digesta pH. Increasing dietary Ca:AvP ratio in- At low Ca:AvP ratios the response to phytase was creased digesta pH in the gizzard. These results were lower compared with higher Ca:AvP ratios. These data in agreement with previous reports showing increased suggest that at a low Ca:AvP ratio, broilers were able digesta pH due to adding Ca as limestone due to lime- to use phytate P better and therefore maintain bone stone’s high acid binding capacity (Selle et al., 2009). Downloaded from ash levels similar to those fed the phytase supplemented Our results supported previous observations that diet. Similar results were observed by Rousseau et al. lowering dietary Ca:AvP positively affected phytate (2012) who showed a tendency for interaction between and P digestibility. However, the comparatively small Ca, nonphytate P (nPP), and phytase supplementa- negative effects on the efficacy of phytase when the tion, such that a lower response to phytase at lower Ca:AvP increased above 2.14 suggest that bacterial http://ps.oxfordjournals.org/ Ca:nPP levels were observed. In contrast, Powell et al. phytases that hydrolyze phytate at low pH in the proxi- (2011) reported that phytase supplementation of a high mal intestinal tract may be less prone to inhibition by Ca diet increased bone breaking strength, bone weight, higher dietary Ca levels. The positive effects of phytate ash weight, and percentage of tibia ash. Interestingly, hydrolysis on AA digestibility support the negative ef- in the present study, there was significantly higher Ca fect that dietary phytate can have on the digestibility digestibility observed at the low Ca:AvP (1.43) com- of nutrients other than P.

pared with the Ca:AvP (2.14). This suggested a higher at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 efficiency of Ca utilization at lower Ca levels, which REFERENCES may be caused by upregulation of Ca transporters at Ca levels below the bird’s requirement (Li et al., 2012). Amerah, A. M., and V. Ravindran. 2009. Influence of maize particle size and microbial phytase supplementation on the performance, Phytase addition improved the digestibility of all nutrient utilisation and digestive tract parameters of broiler measured AA. There was a strong correlation between starters. Anim. Prod. Sci. 49:704–710. AA digestibility and degree of phytate degradation. Angel, R., N. M. Tamim, T. J. Applegate, A. S. Dhandu, and L. E. These positive effects of phytase supplementation on Ellestad. 2002. Phytic acid chemistry: influence on phytin –phosphorus availability and phytase efficacy. J. Appl. Poult. Res. AA digestibility are supported by previous studies 11:471–480. (Rutherfurd et al., 2002; Ravindran et al., 2006). In- AOAC (Association of Official Analytical Chemists). 1990. Official creasing dietary Ca:AvP ratio significantly decreased Methods of Analysis. 15th ed. Association of Official Analytical the digestibility of most AA. The exact mechanism of Chemists, Washington, DC. Cohen, S. A., and K. M. DeAntonis. 1994. Applications of amino the effects of varying Ca:AvP ratio on AA acid digest- acid analysis derivatisation with 6-aminoquinolyl-N-hydroxysuc- ibility remains to be determined, but may be related cinimidyl carbamate: Analysis of feed grains, intravenous solu- to a reduction in phytate hydrolysis and increased for- tions and glycoproteins. J. Chromatog. 661:25–34. mation of Ca-phytate or Ca-protein binary complexes, Delezie, E., L. Maertens, and G. Huyghebaert. 2012. Consequences of phosphorus interactions with calcium, phytase, and chole- increased endogenous losses, an increase in the pH of calciferol on zootechnical performance and mineral retention in the proximal digestive tract, or a combination of these. broiler chickens. Poult. Sci. 91:2523–2531. Calcium may also interact with protein (Selle et al., Latta, M., and M. Eskin. 1980. A simple and rapid colorimet- 2012), which may also explain the lower AA digest- ric method for phytate determination. J. Agric. Food Chem. 28:1313–1315. ibility at higher Ca levels. It should be noted that in Li, J., J. Yuan, Y. Guo, Q. Sun, and X. Hu. 2012. The influence of the current study, birds were killed by cervical disloca- dietary calcium and phosphorus imbalance on intestinal NaPi-IIb tion, which has been suggested to cause mixing of the and calbindin mRNA expression and tibia parameters of broilers. Asian-australas. J. Anim. Sci. 25:552–558. digesta between intestinal segments and thus influence Plumstead, P. W., A. B. Leytem, R. O. Maguire, J. W. Spears, P. ileal digestibility. However, to the authors’ knowledge, Kwanyuen, and J. Brake. 2008. Interaction of calcium and phy- this is, so far, unsupported in the scientific literature. tate in broiler diets. 1. Effects on apparent prececal digestibility Digesta pH was higher in all segments of the diges- and retention of phosphorus. Poult. Sci. 87:449–458. Powell, S., T. D. Bidner, and L. L. Southern. 2011. Phytase supple- tive tract when diets contained phytase. Increased pH mentation improved growth performance and bone characteris- in the gizzard, duodenum, jejunum, and ileum was also tics in broilers fed varying levels of dietary calcium. Poult. Sci. reported by Walk et al. (2012) when diets were supple- 90:604–608. mented to provide 5,000 FTU/kg of feed. Woyengo et Qian, H., E. T. Kornegay, and D. M. Denbow. 1997. Utilization of phytate phosphorus and calcium as influenced by microbial phy- al. (2010) reported that phytate decreased jejunal di- tase, cholecalciferol, and the calcium:total phosphorus ratio in gesta pH in pigs. These authors suggested that phytate broiler diets. Poult. Sci. 76:37–46. reduced pepsin activity in the stomach and resulted in Ravindran, V., S. Cabahug, G. Ravindran, and W. L. Bryden. 1999. higher secretion of HCl and reduced pH of the digesta Influence of microbial phytase on apparent ileal amino acid digestibility in feedstuffs for broilers. Poult. Sci. 78:699–706. in the stomach and the upper part of the small intes- Ravindran, V., P. C. H. Morel, G. G. Partridge, M. Hruby, and J. S. tine. Therefore, it is plausible to suggest that in the Sands. 2006. Influence of an E. coli-derived phytase on nutrient CALCIUM LEVEL AND PHYTASE ADDITION IN BROILER DIETS 915

utilization in broilers fed starter diets containing varying concen- Selle, P. H., V. Ravindran, R. A. Caldwell, and W. L. Bryden. 2000. trations of phytic acid. Poult. Sci. 85:82–89. Phytate and phytase: Consequences for protein utilisation. Nutr. Rousseau, X., M. P. Létourneau-Montminy, N. Même, M. Magnin, Res. Rev. 13:255–278. Y. Nys, and A. Narcy. 2012. Phosphorus utilization in finishing Tamim, N. M., R. Angel, and M. Christman. 2004. Influence of di- broiler chickens: effects of dietary calcium and microbial phytase. etary calcium and phytase on phytate phosphorus hydrolysis in Poult. Sci. 91:2829–2837. broiler chickens. Poult. Sci. 83:1358–1367. Rutherfurd, S. M., T. K. Chung, and P. J. Moughan. 2002. The Tran, T. T., R. Hatti-Kaul, S. Dalsgaard, and S. Yu. 2011. A simple effect of microbial phytase on ileal phosphorus and amino acid and fast kinetic assay for phytases using phytic acid-protein com- digestibility in broiler chickens. Br. Poult. Sci. 44:598–606. plex as substrate. Anal. Biochem. 410:177–184. Selle, P. H., A. J. Cowieson, N. P. Cowieson, and V. Ravindran. Walk, C. L., M. R. Bedford, and A. P. McElroy. 2012. Influence 2012. Protein-phytate interactions in pig and poultry nutrition: of limestone and phytase on broiler performance, gastrointes- A reappraisal. Nutr. Res. Rev. 25:1–17. tinal pH, and apparent ileal nutrient digestibility. Poult. Sci.

Selle, P. H., A. J. Cowieson, and V. Ravindran. 2009. Consequences 91:1371–1378. Downloaded from of calcium interactions with phytate and phytase for poultry and Woyengo, T. A., O. Adeola, and C. M. Nyachoti. 2010. Gastroin- pigs. Livest. Sci. 124:126–141. testinal digesta pH, pepsin activity and soluble mineral concen- Selle, P. H., and V. Ravindran. 2007. Microbial phytase in poultry tration responses to dietary phytic acid and phytase in piglets. nutrition. Anim. Feed Sci. Technol. 135:1–41. Livest. Sci. 134:91–93. http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Lactobacillus sakei modulates mule duck microbiota in ileum and ceca during overfeeding

F. Vasaï ,* K. Brugirard Ricaud,* 1 L. Cauquil ,†‡§ P. Daniel ,# C. Peillod ,Ϧ K. Gontier ,* A. Tizaoui ,¶ O. Bouchez ,** S. Combes ,†‡§ and S. Davail *

* Institut pluridisciplinaire de recherche sur l’environnement et les matériaux–Equipe Environnement et Microbiologie UMR5254, IUT des Pays de l’Adour, Rue du Ruisseau, BP 201, 40004 Mont de Marsan, Downloaded from France; † Institut National de la Recherche Agronomique (INRA), UMR1289 Tissus Animaux Nutrition Digestion Ecosystème et Métabolisme, F-31326 Castanet-Tolosan, France; ‡ Université de Toulouse, Institut National Polytechnique de Toulouse (INPT)–Ecole Nationale Supérieure Agronomique de Toulouse, UMR1289 Tissus Animaux Nutrition Digestion Ecosystème et Métabolisme, F-31326 Castanet-Tolosan, France; § Université de Toulouse INPT Ecole Nationale Vétérinaire de Toulouse, UMR1289 Tissus Animaux Nutrition http://ps.oxfordjournals.org/ Digestion Ecosystème et Métabolisme, F-31076 Toulouse, France; # Laboratoires des Pyrénées et des Landes, 1 rue Marcel David, BP219, 40004 Mont de Marsan, France; Ϧ Institut Technique de l’Aviculture, 28 rue du Rocher, 75008 Paris, France; ¶ Institut Universitaire de Technologie des Pays de l’Adour, Rue du Ruisseau, BP 201, 40004 Mont de Marsan, France; and ** Plateforme Génomique Bâtiment Centre de Ressources de Génotypage & Séquençage-Centre National de Ressources Génomiques Végétales, INRA Auzeville, Chemin de Borderouge-BP 52627, 31326 Castanet-Tolosan Cedex, France at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 ABSTRACT The supplementation with Lactobacillus and diversity decreased in the ileum and increased in sakei as probiotic on the ileal and cecal microbiota of the ceca after overfeeding. Overfeeding increased the mule ducks during overfeeding was investigated using relative abundance of Firmicutes and especially the high-throughput 16S rRNA gene-based pyrosequenc- Lactobacillus group in ileal samples. Nonmetric mul- ing and real-time PCR. The ducks were overfed with tidimensional scaling profiles separated the bacterial or without L. sakei for 12 d with 56% ground corn communities with respect to overfeeding only in cecal and 42% whole corn. Samples were collected before the samples. Richness indicators decreased after L. sakei overfeeding period (at 12 wk), at 13 wk (meal 12 of has been added at mid-overfeeding only in the ileum. overfeeding), and at 14 wk (meal 24), 3 h postfeeding. In the ceca, the decrease of these indexes only occurred Whatever the digestive segment and the level of intake, at the end of overfeeding. The addition of L. sakei trig- Firmicutes, Bacteroidetes, and Proteobacteria were the gers major changes in the ileum, whereas the ceca are dominant phyla in the bacterial community of mule not affected. Lactobacillus sakei decreased the relative ducks (at least 90%). Before overfeeding, ileal samples abundance of Bacteroides at mid-overfeeding and the were dominated by Clostridia, Bacteroidia, and Gam- relative abundance of Enterobacteria at the end of over- maproteobacteria (80% and up), and cecal samples by feeding in the ileum. Bacteroidia and Clostridia (around 85%). The richness Key words: duck microbial diversity, overfeeding , probiotic , 454 pyrosequencing, real-time PCR 2014 Poultry Science 93 :916–925 http://dx.doi.org/ 10.3382/ps.2013-03497

INTRODUCTION al., 1982; Leser and Mølbak, 2009). In chickens, the bacterial activity is more intense in the crop, the small The intestinal microbiota play a major role in ani- intestine (jejunum, duodenum, ileum), and the ceca mal physiology and form a complex ecosystem. For ex- in comparison with proventriculus, gizzard, and pan- ample, intestinal bacteria can affect gut morphology creas (Fuller, 1984). Culture-dependent methods have and nutrition and are known to stimulate the immune shown 2 major bacterial populations: facultative anaer- response and to protect against pathogens (Moreau et obes from the crop to the small intestine and obligate anaerobes in the ceca (Fuller, 1984; Mead, 1989). In accordance with culture-dependent methods, molecu- © 2014 Poultry Science Association Inc. lar methods identified Lactobacillus spp. as the major Received July 17, 2013. Accepted December 16, 2013. bacterial group in the ileum (over 70% were Lactobacil- 1 Corresponding author: [email protected] lus related) and 65% of the sequences in the ceca were

916 PROBIOTICS AND INTESTINAL MICROBIOTA 917 Clostridiaceae related in chickens (Lu et al., 2003). Mo- community in ileum and ceca mule ducks during over- lecular approaches have shown that the composition of feeding. intestinal bacteria is modified by diets (Knarreborg et al., 2002; Hammons et al., 2010). The different phyla MATERIALS AND METHODS identified in bird microbiota are also present in mammals, but molecular inventory in the crop of hoatzin Experimental Design (a tropical bird) showed genera and species specific to birds over 94% of all sequences (Godoy-Vitorino et al., All experimental procedures involving ducks were in 2010). Furthermore, archaebacteria have been identified accordance with the French national guidelines for the in the chicken cecum (Saengkerdsub et al., 2007) and care of animals for research purposes. Male mule ducks the crop of hoatzin (Wright et al., 2009). In goose feces, (MMG × PKL) were raised in a breeding facility be- Downloaded from Bacilli, Clostridia, and Bacteroidetes were predominant longing to the French National Research Institute for (Lu et al., 2009). A study of the 2 parental genetic Agronomy (INRA) at the Domaine d’Artiguères, Ben- types (Pekin and Muscovy) showed that the ileal and quet, France, accreditation number B40–037–1. The cecal microbiota were mainly composed of Firmicutes birds were placed into one pen and reared under usual and Bacteroidetes. Furthermore, overfeeding has an ef- conditions of light [24 h of light for the first 3 d, then http://ps.oxfordjournals.org/ fect on the bacterial composition of the microbiota of natural photoperiod and temperature (26°C for the first these ducks (Vasaï et al., 2013). During this study, the 3 d and then a decrease of 2°C every 3 d to reach 18°C)] nonoverfed animals at the same age have shown no evo- at the Experimental Station for Waterfowl Breeding lution in their microbiota composition between 12 and (INRA, Artiguères, France). From the first day of life 14 wk. Culture-dependent methods have shown that to 4 wk of age, the ducks were fed ad libitum with obligate anaerobes dominate their ceca (Ridsdale and 2 mm pellets of a diet (starter diet) providing 11.93

Corry, 1999). During overfeeding (OF), the birds are MJ/kg of feed and a CP of 17.5%. From 5 to 12 wk of at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 exclusively fed with corn, which is rich in carbohydrates age, birds were fed with a diet providing 11.7 MJ/kg of (especially starch). This diet induces an hepatic steato- feed and a CP of 15% (grower diet). At 12 wk of age, sis resulting in storage of fatty acids in the liver (named all birds were weighed and 228 ducks were selected for foie gras; Davail et al., 2003). We can hypothesize that OF studies based on their live weights being represen- this diet could affect the bacterial community. tative of the live weight variability of the population. Probiotics (live microorganisms) are known as ben- Among the ducks selected, 20 were killed before OF. eficial to their host (Fuller, 1995). A study with chick- The remaining 208 birds were housed in collective cages ens showed that dietary supplementation of probiotics of 4 ducks per box, outdoor access was denied during induced significant changes in the microbial commu- the overfeeding time. Birds were housed in a ventilated nity structure of the gastrointestinal tract (Netherwood room. They were fed hydraulically with an automatic et al., 1999). Different modes of action have been de- feed dispenser (Gaveuse Mg 300 Dussau, Distribution scribed; for example, probiotics can reduce the presence Sas, Pecorade, Landes, France) using a diet consisting of pathogenic bacteria in the digestive tract presumably of 56% cornflour and 42% corn grain (overfeeding diet) by the production of bacteriocins, lactic acid, or both providing 13.9 MJ/kg of feed and a CP of 8.91% with (Czerucka and Rampal, 2002; Servin and Coconnier, or without probiotics (109 L. sakei live cells for one 2003; Servin, 2004). Competition between probiotic meal in the overfeeding food). The composition of the and endogenous bacteria can also occur with respect to different feeds is listed in Table 1. During OF, birds the occupation of ecological niches (La Ragione, 2003). were fed twice a day for 2 wk. Digestive contents from By strengthening the lactic acid bacteria presence, pro- 20 ducks in the groups fed diets with and without probiotics can also help establishing stable environmental biotic were obtained after killing the birds at 13 wk conditions with a low pH limiting the development of (meal 12) and at 14 wk (meal 24). Ducks were killed pathogens (van Winsen et al., 2001). In addition, some by exsanguination after electric stunning 3 h following lactobacilli strains exhibit significant hydrolytic action their last meal to homogenize the filling levels of the and favor digestion and absorption of nutrients, and ducks’ digestive tract. Ducks with similar weights were make certain compounds available to the host. The ad- selected at 5 points. Depending on the time of analysis dition of lactic acid bacteria to feed increased its safety during OF, the experiment had the following 5 experi- and favored endogenous digestive bacteria (van Winsen mental groups, each having n = 20 ducks: baseline be- et al., 2001). In the poultry industry, the interest in fore OF at wk 12, 1 wk during OF without administra- probiotics is increasing because growers in many coun- tion of L. sakei, 1 wk during OF with administration tries can no longer use antibiotics as growth stimulants of L. sakei, 2 wk during OF without administration of (Nava et al., 2005). In the current study, the effect of L. sakei, and 2 wk during OF with administration of L. adding a probiotic (Lactobacillus sakei) during over- sakei. Ileum and ceca were collected immediately after feeding of male mule ducks was investigated using real- slaughter and kept on ice. Digestive contents of the time PCR and high-throughput pyrosequencing. This ileum and ceca were collected by gently applying pres- is the first report of a molecular inventory of bacterial sure along the organ and stored at −20°C. 918 VASAÏ ET AL. Table 1. Ingredients and chemical composition of the experi- Real-Time PCR and Primers mental diets Twenty birds per experimental groups (with or with- Overfeeding Item (%, unless otherwise Starter Grower diet (dry corn) out probiotic according to overfeeding period) were specified) diet diet DM % used for quantitative PCR analysis. All primers sets Ingredient are listed in Table 3. Reactions were run in duplicate Wheat 15 35 — in 48-well plates in a final volume of 15 μL. The PCR Corn 32.6 38.69 98 mix consisted of 7.5 μL of SybrGreen Universal PCR Rapeseed oil free 7 12 — Sunflower cake 8 11.35 — Master Mix (Quanta Bioscience, Gaithersburg, MD), Sodium carbonate 1.17 1.375 — 5.5 μL of 500 nM primers, and 2 μL of template DNA Dicalcium phosphate 0.675 0.575 — of appropriate dilutions or water for the negative con- Downloaded from Lysine 0.336 0.298 — Salt 0.25 0.36 — trol. Real-time PCR were performed in a StepOne in- Methionine 0.235 0.086 — strument (Applied Biosystems, Saint Aubin, France) Choline chloride 0.06 0.04 — with an initial 10-min step at 95°C, 35 cycles for 15 Premix + vitamin 0.3 0.23 2 Nutrient s at 95°C, annealing/extension for 1 min at the tem- ME, MJ/kg 11.93 11.7 13.9 perature specific for each primer set, and 1 final cycle http://ps.oxfordjournals.org/ Humidity 12.46 12.61 — at 95°C for 15 s. Melt curve analyses were done by CP 17.5 15 8.91 Fat 2.66 2.26 3.68 slowly heating the PCR mixtures from 60 to 95°C, and Cellulose 5.46 5.28 2.77 the cycle threshold (Ct) was determined with the Ste- Ashes 5.24 4.99 — pOne Applied Biosystems software. Standard curves Starch 40.21 47.32 75.91 Nutrient level (% raw were generated from 10-fold serial dilutions in water of material) known concentrations of genomic bacterial DNA (Table Lysine 0.92 0.701 — 2) and used to quantify the copy number of the respec- Methionine 0.506 0.36 — at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Methionine + cysteine 0.85 0.705 — tive PCR reaction, with cycle threshold determined by Threonine 0.672 0.548 — the second derivative maximum method (Tichopad et Tryptophan 0.203 0.173 — al., 2003). Results are presented as the number of 16S Linoleic acid 1.43 1.2 — Calcium 0.948 0.95 — rDNA copies and expressed per gram of fresh samples Phosphorus 0.575 0.519 — and log10 transformed for statistical analysis. Sodium 0.139 0.15 — Available phosphorus 0.358 0.299 — pH — — 5.36 High-Throughput Sequencing and Analysis of 16S rRNA Gene Amplicons For pyrosequencing, 3 birds (yielding 3 ileums and 3 DNA Extraction ceca) per experimental groups were used except for ileal sample before overfeeding where only one sample was For DNA extraction from pure bacterial cultures analyzed. Amplicons from the V3 to V4 regions of 16S (Table 2), the QIAamp DNA mini kit was used follow- rRNA genes (corresponding to a 460-bp region of Esch- ing the manufacturer’s recommended procedure (Qia- erichia coli 16S rDNA, GenBank number J01695) were gen, Hilden, Germany). Genomic DNA from 220 mg of produced using bacterial forward 343F (TACGGRAG- ileal and cecal samples were extracted employing the GCAGCAG; Liu et al., 2007) and reverse 784R (TAC- QIA DNA stool kit (Qiagen) according to the manu- CAGGGTATCTAATCCT; Andersson et al., 2008) facturer’s instructions. An additional lysis step using primers. The PCR was performed in a total volume of lysozyme (140 μL of 10 mg/mL in Tris-EDTA buffer; 100 μL containing 1× PCR buffer, 200 μM of dNTP, 1 Sigma, St. Louis, MO) to improve the DNA extraction U of Isis DNA polymerase (MP Biomedicals, Illkirch, of gram-positive bacteria (Johansen et al., 2007). The Graffenstaden, France), 0.5 μM of each primer, and 1 to extracted DNA was eluted in 200 μL of Tris-EDTA 5 ng of DNA template. The amplification program con- buffer and stored at −20°C. sisted of an initial denaturation step at 94°C for 2 min;

Table 2. Strains used for the preparation of standard curves for real-time PCR analysis

Strain Primer targeted Reference Escherichia coli Enterobacteriaceae CIP548T Lactobacillus plantarum All bacteria NCIMB8826 Firmicutes Lactobacillus group Enterococcus faecium Enterococcus CIP5432 Bacteroides dorei Bacteroides NSMZ17855 Bacteroidetes Clostridium spp. Clostridium Laboratory strain Firmicutes PROBIOTICS AND INTESTINAL MICROBIOTA 919

Table 3. Targets and list of primers used for the enumeration of 16S rDNA sequences representing different bacterial groups by real- time PCR

Temperature Target Sequence (5c–3c) (°C) Reference All bacteria TCCTACGGGAGGCAGCAGT 58 Walter et al. (2001) GACTACCAGGGTATCTAATCCTGTT Firmicutes TGAAACTYAAAGGAATTGACG 61.5 Bacchetti De Gregoris et al. (2011) ACCATGCACCACCTGTC Bacteroidetes CRAACAGGATTAGATACCCT 61.5 Bacchetti De Gregoris et al. (2011) GGTAAGGTTCCTCGCGTAT Bacteroides GAGAGGAAGGTCCCCCAC 60 Layton et al. (2006)

CGCTACTTGGCTGGTTCAG Downloaded from Lactobacillus group CACCGCTACACATGGAG 58 Walter et al. (2001) AGCAGTAGGGAATCTTCCA Clostridium TTACTGGGTGTAAAGGG 60 Van Dyke and McCarthy (2002) TAGAGTGCTCTTGCGTA Enterococcus CCCTTATTGTTAGTTGCCATATT 61 Rinttilä et al. (2004) ACTCGTTGTACTTCCCATTGT Enterobacteria CATTGACGTTACCCGCAGAAGAAGC 63 Bartosch et al. (2004) http://ps.oxfordjournals.org/ CTCTACGAGACTCAAGCTTGC

32 cycles of denaturation at 94°C for 30 s, annealing at sequences) or with a homopolymer stretch longer than 60°C for 30 s, and elongation at 72°C for 30 s; and a 8 bases (140 sequences), those that did not match the final extension step at 72°C for 7 min. The PCR prod- proximal PCR primer sequences (with 2 mismatches al- ucts were purified with the QIAquick PCR Purification lowed; 801 sequences), and finally those including both kit (Qiagen) followed by DNA yield quantification and primer sequences but with a length shorter than 350 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 quality estimation using a Nano-Drop instrument. Se- pb (295 sequences). A total of 246,090 sequences were quencing of the 16S rRNA genes was performed with retained corresponding to 8.789 ± 1.510 sequences per a 454 GS FLX instrument (454 Life Sciences, Roche, samples. Branford, CT) and Titanium chemistry according to the manufacturer’s instructions. Amplicons of 450 bp Taxonomic Classification were pooled using equal amount of each PCR product. and Statistical Analysis For the analysis, samples were recognized by couples of multiplex identifier (mid: 11-bp nucleotide sequence), Filtered sequences were analyzed using MOTHUR which were fixed on the universal primers provided by software v. 1.24; (Schloss et al., 2009) Readings were Roche during the amplification. All sequences that did aligned with the SILVA alignment database provided not have a length between 150 and 600 bp, contained by the MOTHUR software (14,956 sequences corre- 10 homopolymers, 1 N (ambiguous base), or had a mis- sponding to unique sequences in the SSU REF data- match in the primer sequence were removed using Py- base v102; Pruesse et al., 2007), and an alignment qual- thon script developed by the bioinformatics platform of ity was calculated using the SILVA secondary structure Toulouse (https://mulcyber.toulouse.inra.fr/). map file (1,072 sequences were removed). After calculat- The remaining sequences were treated with the R ing a pairwise distance between aligned sequences, the software (http://www.r-project.org/) to obtain the sequences were clustered into operational taxonomic composition of the microbiota with respect to treatment units (OTU, cutoff 0.05 using a farthest neighbor clus- conditions. Sequences were analyzed using Mothur soft- tering). Rarefaction curves, abundance-based coverage ware v. 1.24 (Schloss et al., 2009) and were aligned with estimator (ACE; defines the coverage of sampling), the Silva database (14,956 sequences). and Chao1 richness (calculates the estimated true species diversity of a sample) were calculated with the Pyrotag Processing and Analysis Rarefaction and Chao1 Estimator. Shannon α-diversity index (measure biodiversity) was calculated according A total of 292,043 16S rRNA gene sequences (also to Hayek and Buzas (1996), and the results of diversity referred to as 16S pyrotags) were obtained with the 454 indexes were obtained using the Vegan library on the Titanium pyrosequencing run for the 28 samples. The R software (Oksanen et al., 2010). For real-time PCR 16S pyrotags were sorted based on their respective bar- results, the Mann-Whitney test (nonparametric test) codes to form a total of 28 pyrotag library representing in the R software was used. Two independent samples the 28 collected ileal and cecal samples. Sequences were of respective sizes n1 and n2 were available. If n1 and sequentially filtered using a Python script developed by n2 ≥ 8, the U statistic follows a normal distribution. the bioinformatic platform of Toulouse (https://mulcy- The null hypothesis (H0) is the sum of the ranks for the ber.toulouse.inra.fr/), first by removing those sequenc- sample X is analogous to Y. UX, Y = min (UX, UY) es with a sequencing length less than 150 nt (35,880 se- was defined for testing the hypothesis (H0). If the size quences), those with at least one ambiguous base (8,887 n1 and n2 are lower than 20 and higher than 8, for a 920 VASAÏ ET AL. RESULTS Bacterial Community in Mule Ducks Microbial diversity in samples was estimated by calculating the number of OTU. For all samples of the ileum, the number of OTU with a cutoff of 0.05 was 791 ± 63 with coverage by sample of 96.09 ± 0.97%. The average number of sequences was 10,100; 8 different phyla were recorded for 118 different taxa. Fur-

thermore, the Chao1, the ACE, and the Shannon index Downloaded from were, respectively, 1,472 ± 105.3, 1,943 ± 136.3, and 3.98 ± 0.17 (Table 4). The majority of the diversity of the ileum before OF was represented by Firmicutes sequences (70%). The rest was composed of Proteobac-

teria (14.98%) and Bacteroidetes (13.27%) sequences. http://ps.oxfordjournals.org/ Other phyla such as Actinobacteria or Fusobacteria represented less than 2% (Figure 1a). At the class level, the sequencing data indicated that the microbiota were mainly composed of Clostridia (60.50%) and Bacilli (8.63%) from the Firmicutes, Bacteroidia (13.14%) from the Bacteroidetes, and Gammaproteobacteria (10.73%) in the phylum Proteobacteria. Regarding ceca samples, the number of OTU at a at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 cutoff of 0.05 was 2,111 ± 622. The average number of sequences was 14,031 with 9 phyla represented by 138 different taxa. The Chao1, the ACE, and the Shannon index were, respectively, 4,129.49 ± 1,166.61, 6,246.55 ± 1,760.35, and 5.88 ± 0.27 (Table 4). The sequences retrieved from the ceca were mainly composed of Bac- teroidetes sequences (57.03%). Firmicutes sequences amounted to 33.95% of the total, and Proteobacteria se- Figure 1. Percentage contribution of sequences (%) evaluated at quences accounted for 7.26% (Figure 2a). The sequenc- the phylum (a) and class (b) levels to the total number of sequences es from 2 classes were dominant in the ceca, Bacteroidia in ileum during overfeeding (OF) and with or without probiotics (3 samples by experimental groups). Before OF (Bof), mid-OF without (56.34%) from Bacteroidetes, and Clostridia (32.85%) probiotics (Mof WP), mid-OF with probiotics (Mof P), end OF with- from Firmicutes. out probiotics (Eof WP), end OF (Eof P). Effect of Overfeeding on the Microbial Community significance level α (α = 5% = 0,05), Mann-Whitney tables provide a critical value c to decide if UX,Y,obs > Overfeeding triggered changes in bacterial microbi- c, H0 was accepted; if UX,Y,obs ≤ c, H0 was rejected. ota in the ileum because before OF, the microbiota

Table 4. Estimators of diversity during overfeeding (OF) in ileum and ceca of mule ducks

Period1

Item Bof Mof WP Mof P Eof WP Eof P SEM Ileum Number of operational taxonomic unit 791 441 302 320 569 63 Chao1 estimated richness 1,472.1 723.7 530.4 577.3 884.3 105.3 ACE2 1,943.2 956.6 766.4 770.3 1,052.1 136.3 Shannon diversity index 3.98 3.52 2.59 2.94 3.22 0.17 Ceca Number of operational taxonomic unit 2,111 654 1257 780 361 209 Chao1 estimated richness 4,129.5 1,237.2 1,826.8 1,598.5 721.5 403.4 ACE 6,246.6 1,759.5 2,758.8 2,297.8 1,079 609.2 Shannon diversity index 5.88 4.74 5.15 4.63 4.46 0.19 1Bof: before OF; Mof WP: mid-OF without probiotics; Mof P: mid-OF with probiotics; Eof WP: end OF without probiotics; Eof P: end OF. Three samples by experimental groups. 2ACE = abundance-based coverage estimator. PROBIOTICS AND INTESTINAL MICROBIOTA 921 cant increase in Deltaproteobacteria (21.79%), concomitant to a decrease in Bacteroidia (37.01%) and Clostrid- ia (24.85%), but these 3 classes were still dominant at the end of OF, representing 90% of the total bacterial population (Figure 2b). Nonmetric multidimensional scaling (nMDS) profiles (representing the repartition of different families) between different periods of OF in ceca indicated that the bacterial communities in cecal samples were separated with respect to the 3 different OF periods (Figure 3). Analyses of similarity yielded an r of 0.346 and a P-value of 0.09 between before and Downloaded from mid-OF, r = 0.143 and a P = 0.09 between mid and end OF and r = 0.636 and a P = 0 between the first and the last period of OF. These changes were confirmed by real-time PCR as a statistically significant decrease in Bacteroidetes and Enterococcus after 1 wk http://ps.oxfordjournals.org/ of OF. Between mid-OF and the end of OF the total bacterial load, Bacteroides, Firmicutes including Clos- tridium, and Lactobacillus group decreased (Table 5).

Effect of a Probiotic on the Microbial Community at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 In samples from probiotics-fed ducks, richness indicators decreased at mid-OF only in the ileum whereas in the ceca, the decrease of these indexes only occurred at the end of OF. In the ileum at mid-OF, the addition of L. sakei (which is part of the Bacilli class) triggered an increase in the Bacilli class from 15.05% without probiotics to 62.6% with them. This increase coincided with a decrease in Gammaproteobacteria from 51.57 to 14.89%, whereas the Clostridia increased slightly from Figure 2. Percentage contribution of sequences (%) evaluated at the phylum (a) and class (b) levels to the total number of sequences in 16.34 to 20.35%. At the end of OF, when L. sakei was ceca during OF and with or without probiotics (3 samples by experi- added to the meal, we observed a decrease in the per- mental groups). Before overfeeding (OF; Bof), mid-OF without pro- centage of Clostridia (−25.95%), whereas conversely biotics (Mof WP), mid-OF with probiotics (Mof P), end OF without probiotics (Eof WP), end OF (Eof P). there was an increase in Bacteroidia of around 6.5% and of the Fusobacteria, Actinobacteria, Alphaproteo- bacteria, Betaproteobacteria, and Epsilonproteobacte- was mainly composed of Clostridia (60.5%), followed ria. At the end of OF, the Bacilli represented more by Bacteroidia (13.1%), Bacilli (8.6%), and Gamma- than 60% of the total population (Figure 1b). Analyses proteobacteria (10.7%) as indicated by the sequencing by nMDS showed that the bacterial community can data. At mid-OF, the Gammaproteobacteria were most be separated regarding the addition of L. sakei, since represented (51.6%), whereas Clostridia represented an r-value of 0.349 and a P of 0.025 were obtained. only 16.3% of the population and Bacteroidia almost These data indicate separated but strongly overlapping disappeared (0.4%). At the end of the OF phase, Ba- groups (Figure 4). Real-time PCR data confirmed this cilli (58.1%) were dominant but Clostridia still were an effect on bacterial composition: Bacteroides statistical- important component of the bacterial population with ly decreased at mid-OF with or without probiotics but 32.3% (Figure 1b). Significant changes were observed with a strong effect when L. sakei was administrated using real-time PCR analysis of different targets. To- (Table 5). A slight decrease in total bacterial load and ward mid-OF, a statistically significant decrease in the representatives of the Firmicutes including the Lac- Bacteroides and Clostridium was indicated by the py- tobacillus group was observed. Sequences representing rosequencing results. Between mid-OF and the end of the Enterobacteria decreased strongly (Table 5). OF, Bacteroidetes including Bacteroides, Firmicutes Treatment with L. sakei did not trigger changes in including Enterococcus numbers apparently decreased, the number of sequences representing the Bacilli in the whereas Clostridium number increased (Table 5). ceca. At mid-OF, only a small percentage of correspond- The cecal microbiota consisted essentially of Bacte- ing Bacilli sequences (2.36%) were found. In contrast, roidia (56.34%) and Clostridia (32.85%). Some changes Deltaproteobacteria sequences decreased from 21.79 to were detected during OF but these 2 classes stayed 14.40% whereas Clostridia and Bacteroidia sequenc- dominant during OF. At mid-OF, there was a signifi- es increased by 4.20 and 6.83%, respectively. At the 922 VASAÏ ET AL. Downloaded from http://ps.oxfordjournals.org/

Figure 3. Nonmetric multidimensional scaling (nMDS) profile Figure 4. Nonmetric multidimensional scaling (nMDS) profile showing the distribution of different families of the different samples showing the distribution of different families of the different samples studied before, mid, and end of overfeeding in ceca. studied without or with probiotics in 3 periods of overfeeding (OF) in at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 the ileum. Before (1), mid (2), and end (3) of OF in the ileum. end of OF there was a decrease in Clostridia numbers (−5.23%), whereas there was a slight increase in Bac- groups with respect to probiotics-fed and probiotic-free teroidia (+2.95%) and Gammaproteobacteria numbers birds (data not shown). Statistical changes due to the (+3.34%) in samples from probiotics-fed birds (Figure probiotics used were also not observed with the real- 2b). The nMDS analysis did not show a separation of time PCR analysis.

Table 5. Statistical effect of overfeeding (OF) and addition of probiotics on different phyla, genus, or groups in ileum or ceca of mule ducks1

Period2

Item Bof Mof WP Mof P Eof WP Eof P SE Significance3 P-value Ileum (20 samples by experimental groups) All bacteria 8.39 8.95 9.98 9.32 8.74 0.61 d 0.019 Bacteroidetes 6.25 6.38 6.36 5.37 5.43 0.51 b <0.001 Bacteroides 6.31 5.79 5.60 4.74 4.72 0.69 a 0.011 c 0.026 b <0.001 Firmicutes 9.85 10.00 10.75 10.05 9.25 0.54 b 0.018 d 0.0031 Clostridium 8.74 6.60 6.57 8.6 8.79 1.17 a 0.04 b <0.001 Enterococcus 4.90 5.01 5.33 3.69 4.32 0.65 b 0.0002 Lactobacillus group 8.33 8.66 9.43 9.04 8.74 0.41 d 0.033 Proteobacteria Enterobacteria 6.73 8.00 8.76 8.23 6.82 0.89 d <0.001 Ceca (20 samples by experimental groups) All bacteria 11.27 11.78 11.88 10.10 10.4 0.81 b <0.001 Bacteroidetes 9.98 9.19 8.96 8.46 8.87 0.56 a <0.001 Bacteroides 10.07 10.47 10.30 8.29 8.85 0.97 b <0.001 Firmicutes 10.92 12.23 12.04 11.05 11.07 0.62 b 0.0001 Clostridium 11.92 12.18 12.34 11.01 12.23 0.59 b <0.001 Enterococcus 5.36 3.53 4.27 4.89 4.86 0.71 a <0.001 Lactobacillus group 8.87 8.91 10.05 7.82 7.97 0.90 b <0.001 Proteobacteria Enterobacteria 8.24 9.53 9.03 9.27 9.05 0.48 — — 1 . The median is expressed as log10 copies number of 16S rDNA/g of fresh digestive content 2Bof: before OF; Mof WP: mid-OF without probiotics; Mof P: mid-OF with probiotics; Eof WP: end OF without probiotics; Eof P: end OF. 3Overfeeding effect: a: Bof versus Mof WP, b: Mof WP versus Eof WP. Probiotic effect: c: Mof WP versus Mof P, d: Eof WP versus Eof P. PROBIOTICS AND INTESTINAL MICROBIOTA 923 DISCUSSION Firmicutes/Bacteroides ratios increased in comparison with those of controls (Angelakis and Raoult, 2010), As in many other studies on intestinal microbiota but in the current study no significant effect on growth of birds and mammals, the current study provides performance and weight of the fatty liver was observed evidence for the Firmicutes and Bacteroidetes as the (data not shown). Feeding of L. sakei triggered an in- major phyla, a finding that suggests their important crease in ileal Bacilli numbers and a decrease in se- role in metabolism and host physiology (Ley et al., quences representing the phylum. In poultry species, 2008; Kohl, 2012). The Proteobacteria were the second the use of lactobacilli alone or in mixtures of bacteria most common phylum after Firmicutes in ileal samples reduced the numbers of coliforms including Escherichia based on 454 pyrosequencing, as was also observed in coli, Salmonella enterica, and Clostridium perfringens feces of Canada geese (Lu et al., 2009) and in Muscovy as determined by culture-dependent methods (Fuller, Downloaded from ducks (Cairina moschata; Vasaï et al., 2013). Interest- 1977; Francis et al., 1978; Watkins et al., 1982; Watkins ingly, mule ducks are a hybrid of a male Muscovy and and Kratzer, 1983; Mountzouris et al., 2007). Molecular a female Pekin duck (Anas platyrhynchos). Pekin, Mus- methods confirmed the changes induced by probiotics covy, and mule ducks differ in their ability to produce on bacterial communities (Netherwood et al., 1999). a fatty liver because Mule and Muscovy ducks show a Real-time PCR has been used to confirm the beneficial http://ps.oxfordjournals.org/ lower capacity for fat storage than Pekin ducks. In con- effect of L. sakei by decreasing Enterobacteria, suggest- trast to chickens, where obligate anaerobes dominated ing that L. sakei can protect against potential patho- ceca samples (the Clostridia class from Firmicutes and gens (Chaucheyras-Durand et al., 2006). The number Bacteroidetes) and facultative anaerobes (Bacilli class of sequences representing Bacteroides significantly and especially Lactobacillus) are dominant in the ileum decreased with probiotic supplementation, suggesting (Lu et al., 2008), in mule ducks, obligate anaerobes some level of antagonisms between L. sakei and these

were dominant in both ileal (class Clostridia and Bacte- bacteria. The separation of the sequence data into 2 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 roidia) and cecal samples. Higher numbers of sequences separated groups based on the diets with and with- representing the Clostridia in ducks than in chickens out the probiotic (Figure 4) suggests that the bacte- suggest that the bacterial digestive metabolism in rial communities were more affected by L. sakei than chicken and ducks could be somewhat different. by overfeeding. Because a decrease in the total bacte- Overfeeding increased the relative abundance of Fir- rial load at the end of OF in the ileum was observed, micutes in the ileum and especially a component of it is possible that L. sakei inhibited growth of several the Bacilli class, the Lactobacillus group as previously other bacteria. The cecal samples showed a very differ- described for Pekin and Muscovy ducks (Vasaï et al., ent pattern than the ileal samples in response to pro- 2013). These bacteria are well known as amylolytic bac- biotics. First of all, the probiotics did not change the teria and frequently increase in pigs and rats fed with microbial communities enough to see an effect based on diets rich in starch (Wang et al., 2002; Regmi et al., nMDS analysis and bacterial composition. In chickens, 2011). Interestingly, in cattle, it has been shown that it is well known that Lactobacillus spp. (facultative an- Lactobacillus increase with diets containing more than aerobes) are an important component of the ileal mi- 71% of starch on concentrate with a composition very crobiota, whereas Clostridia and Bacteroides (obligate close to that the feed used in the present study (Brown anaerobes) are predominant in cecal samples (Lu et et al., 2006). According to the PCR results, Bacteroides al., 2003). This observation could explain why feeding and Clostridium numbers decreased, but interestingly, L. sakei did not have a noticeable effect in the ceca. for Clostridium this decrease was reversed at the end of Perhaps, the complex anaerobic environment is very re- OF. The clustering of the results for the cecal samples fractory to changes in the microbial community. Over- shows a good separation into different groups regard- all, the real-time PCR data were in agreement with ing the OF period, before, mid, or end of OF, prob- the sequencing results; however, for samples from the ably due to the negative effect on the total bacterial mid-OF and end OF periods, pyrosequencing showed load especially the Bacteroidetes including Bacteroides, an increase whereas real-time PCR revealed a slight Firmicutes including Clostridium, Lactobacillus group, decrease in Firmicutes numbers. The reason(s) for the and Enterococcus. Bacteroides is well known to ferment discrepancy are not known. To conclude, this work, us- a variety of plant polysaccharides like amylose starch ing both real-time PCR and high-throughput pyrose- (Salyers et al., 1977). Interestingly, when Enteroccoccus quencing based on 16S rRNA genes in ileal and cecal faecium was used as probiotic in piglets, it increased samples, gives a first report on the microbial commu- the nutrient availability and improved the absorption nity in ducks and the effect of administering probiotics of glucose (Lodemann et al., 2006). during overfeeding on the abundance of major groups. Lactobacillus sakei reduced the diversity indexes at The microbial diversity of mule ducks is dominated by mid-OF only in the ileum, whereas in the ceca, the de- Firmicutes, Bacteroidetes, and Proteobacteria, and OF crease of these indexes only occurred at the end of OF. modifies bacterial communities of ceca samples, where- In newborn chicks and ducks inoculated intragastrically as probiotics show an important effect on ileal samples. with Lactobacillus species, growth performance as well The increase in lactobacilli through the OF process 924 VASAÏ ET AL. when adding L. sakei as a probiotic could potentially in diet influences the Lactobacillus strain composition in the crop improve the birds’ health and could be of great interest of broiler chickens. Syst. Appl. Microbiol. 33:275–281. http:// dx.doi.org/10.1016/j.syapm.2010.04.003. to the poultry industry. Hayek, L. C., and M. A. Buzas. 1996. Surveying Natural Popula- tions. Columbia University Press, New York, NY. ACKNOWLEDGMENTS Johansen, C. H., L. Bjerrum, and K. Pedersen. 2007. Impact of sa- linomycin on the intestinal microflora of broiler chickens. Acta Vet. Scand. 49:30. http://dx.doi.org/10.1186/1751-0147-49-30. We thank the Conseil Général des Landes (France) Knarreborg, A., M. A. Simon, R. M. Engberg, B. B. Jensen, and and the Comité Interprofessionnel des Palmipèdes à Foie G. W. Tannock. 2002. Effects of dietary fat source and subther- Gras (CIFOG, France) for financing this work. We also apeutic levels of antibiotic on the bacterial community in the thank the technical staff of INRA Artiguères (France) ileum of broiler chickens at various ages. Appl. Environ. Micro-

biol. 68:5918–5924. http://dx.doi.org/10.1128/aem.68.12.5918- Downloaded from for the ducks. We thank Julien Arroyo for his construc- 5924.2002. tive comments on the manuscript. We are grateful to Kohl, K. D. 2012. Diversity and function of the avian gut microbiota. Jérôme Lluch and Clémence Genthon (Plateforme Gé- J. Comp. Physiol. B 182:591–602. http://dx.doi.org/10.1007/ nomique Bâtiment CRGS-CNRGV, INRA Auzeville, s00360-012-0645-z. La Ragione, R. 2003. Competitive exclusion by Bacillus subtilis Castanet-Tolosan, France) and Béatrice Gabinaud spores of Salmonella enterica serotype Enteritidis and Clostridi- (INRA, UMR1289 Tissus Animaux Nutrition Digestion um perfringens in young chickens. Vet. Microbiol. 94:245–256. http://ps.oxfordjournals.org/ Ecosystème et Métabolisme, Castanet-Tolosan, France) http://dx.doi.org/10.1016/s0378-1135(03)00077-4. for their excellent technical assistance. Layton, A., L. McKay, D. Williams, V. Garrett, R. Gentry, and G. Sayler. 2006. Development of Bacteroides 16S rRNA gene Taq- Man-based real-time PCR assays for estimation of total, human, REFERENCES and bovine fecal pollution in water. Appl. Environ. Microbiol. 72:4214–4224. http://dx.doi.org/10.1128/AEM.01036-05. Andersson, A. F., M. Lindberg, H. Jakobson, F. Bäckhed, P. Nyrén, Leser, T. D., and L. Mølbak. 2009. Better living through microbial and L. Engstrand. 2008. Comparative analysis of human gut action: The benefits of the mammalian gastrointestinal micro-

microbiota by barcoded pyrosequencing. PLoS ONE 3:e2836 biota on the host. Environ. Microbiol. 11:2194–2206. http:// at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 http://dx.doi.org/10.1371/journal.pone.0002836. dx.doi.org/10.1111/j.1462-2920.2009.01941.x. Angelakis, E., and D. Raoult. 2010. The increase of Lactobacil- Ley, R. E., M. Hamady, C. Lozupone, P. J. Turnbaugh, R. R. Ra- lus species in the gut flora of newborn broiler chicks and ducks mey, J. S. Bircher, M. L. Schlegel, T. A. Tucker, M. D. Schren- is associated with weight gain. PLoS ONE 5 http://dx.doi. zel, R. Knight, and J. I. Gordon. 2008. Evolution of mammals org/10.1371/journal.pone.0010463. and their gut microbes. Science 320:1647–1651. http://dx.doi. Bacchetti De Gregoris, T., N. Aldred, A. S. Clare, and J. G. Burgess. org/10.1126/science.1155725. 2011. Improvement of phylum- and class-specific primers for real- Liu, Z., C. Lozupone, M. Hamady, F. D. Bushman, and R. Knight. time PCR quantification of bacterial taxa. J. Microbiol. Meth- 2007. Short pyrosequencing reads suffice for accurate microbial ods 86:351–356. http://dx.doi.org/10.1016/j.mimet.2011.06.010. community analysis. Nucleic Acids Res. 35:e120. http://dx.doi. Bartosch, S., A. Fite, G. T. Macfarlane, and M. E. T. McMurdo. org/10.1093/nar/gkm541. 2004. Characterization of bacterial communities in feces from Lodemann, U., K. Hübener, and H. Martens. 2006. Effects of En- healthy elderly volunteers and hospitalized elderly patients by terococcus faecium NCIMB 10415 as probiotic supplement on in- using real-time PCR and effects of antibiotic treatment on the fe- testinal transport and barrier function of piglets. Arch. Anim. cal microbiota. Appl. Environ. Microbiol. 70:3575–3581. http:// Nutr. 60:35–48. dx.doi.org/10.1128/AEM.70.6.3575-3581.2004. Lu, J., C. Hofacre, F. Smith, and M. D. Lee. 2008. Effects of feed Brown, M. S., C. H. Ponce, and R. Pulikanti. 2006. Adaptation of additives on the development on the ileal bacterial community of beef cattle to high-concentrate diets: Performance and ruminal the broiler chicken. animal 2. 10.1017/s1751731108001894. metabolism. J. Anim. Sci. 84:E25–E33. Lu, J., U. Idris, B. Harmon, C. Hofacre, J. J. Maurer, and M. D. Lee. Chaucheyras-Durand, F., J. Madic, F. Doudin, and C. Martin. 2003. Diversity and succession of the intestinal bacterial com- 2006. Biotic and abiotic factors influencing in vitro growth of munity of the maturing broiler chicken. Appl. Environ. Micro- Escherichia coli O157:H7 in ruminant digestive contents. Appl. biol. 69:6816–6824. http://dx.doi.org/10.1128/aem.69.11.6816- Environ. Microbiol. 72:4136–4142. http://dx.doi.org/10.1128/ 6824.2003. AEM.02600-05. Lu, J., J. W. Santo Domingo, S. Hill, and T. A. Edge. 2009. Mi- Czerucka, D., and P. Rampal. 2002. Experimental effects of Sac- crobial diversity and host-specific sequences of Canada goose charomyces boulardii on diarrheal pathogens. Microbes Infect. feces. Appl. Environ. Microbiol. 75:5919–5926. http://dx.doi. 4:733–739. org/10.1128/AEM.00462-09. Davail, S., N. Rideau, G. Guy, J. M. André, and R. Hoo-Paris. 2003. Mead, G. 1989. Microbes of the avian cecum: Types present and Pancreatic hormonal and metabolic responses in overfed ducks. substrates utilized. J. Exp. Zool. Suppl. 3:48–54. Horm. Metab. Res. 35:439–443. Moreau, M., P. Raibaud, and M. C. Muller. 1982. Relation entre Francis, C., D. M. Janky, A. S. Arafa, and R. H. Harms. 1978. Inter- le développement du système immunitaire intestinal à IgA et relationship of Lactobacillus and zinc bacitracin in the diets of l’établissement de la flore microbienne dans le tube digestif du turkey poults. Poult. Sci. 57:1687–1689. souriceau holoxénique. Ann. Immunol. 133D:29–39. Fuller, R. 1977. The importance of Lactobacilli in maintaining nor- Mountzouris, K. C., P. Tsirtsikos, E. Kalamara, S. Nitsch, G. mal microbial balance in the crop. Br. Poult. Sci. 18:85–94. Schatzmayr, and K. Fegeros. 2007. Evaluation of the efficacy of Fuller, R. 1984. Microbial activity in the alimentary tract of birds. a probiotic containing Lactobacillus, Bifidobacterium, Enterococ- Proc. Nutr. Soc. 43:55–61. cus, and Pediococcus strains in promoting broiler performance Fuller, R. 1995. Probiotics. The Scientific Basis. Chapman & Hall, and modulating cecal microflora composition and metabolic ac- London, UK. tivities. Poult. Sci. 86:309–317. Godoy-Vitorino, F., K. C. Goldfarb, E. L. Brodie, M. A. Garcia- Nava, G. M., L. R. Bielke, T. R. Callaway, and M. P. Castañeda. Amado, F. Michelangeli, and M. G. Domínguez-Bello. 2010. 2005. Probiotic alternatives to reduce gastrointestinal infections: Developmental microbial ecology of the crop of the folivorous The poultry experience. Anim. Health Res. Rev. 6:105–118. hoatzin. ISME J. 4:611–620. http://dx.doi.org/10.1038/is- Netherwood, T., H. J. Gilbert, D. S. Parker, and A. G. O’Donnell. mej.2009.147. 1999. Probiotics shown to change bacterial community structure Hammons, S., P. L. Oh, I. Martínez, K. Clark, V. L. Schlegel, E. in the avian gastrointestinal tract. Appl. Environ. Microbiol. Sitorius, S. E. Scheideler, and J. Walter. 2010. A small variation 65:5134–5138. PROBIOTICS AND INTESTINAL MICROBIOTA 925

Oksanen, J., F. Blanchet, R. Kindt, P. Legendre, R. O’Hara, G. Servin, A. L., and M.-H. Coconnier. 2003. Adhesion of probiotic Simpson, P. Solymos, M. Stevens, and H. Wagner. 2010. Vegan: strains to the intestinal mucosa and interaction with pathogens. Community Ecology Package. R package version 1.17–1, 2010. Best Pract. Res. Clin. Gastroenterol. 17:741–754. http://dx.doi. http://www.R-project. org. org/10.1016/s1521-6918(03)00052-0. Pruesse, E., C. Quast, K. Knittel, B. M. Fuchs, W. Ludwig, J. Pe- Tichopad, A., M. Dilger, G. Schwarz, and M. W. Pfaffl. 2003. Stan- plies, and F. O. Glockner. 2007. SILVA: A comprehensive online dardized determination of real-time PCR efficiency from a single resource for quality checked and aligned ribosomal RNA sequence reaction set-up. Nucleic Acids Res. 31:e122. data compatible with ARB. Nucleic Acids Res. 35:7188–7196. Van Dyke, M. I., and A. J. McCarthy. 2002. Molecular biological http://dx.doi.org/10.1093/nar/gkm864. detection and characterization of Clostridium populations in mu- Regmi, P. R., B. U. Metzler-Zebeli, M. G. Ganzle, T. A. T. G. nicipal landfill sites. Appl. Environ. Microbiol. 68:2049–2053. van Kempen, and R. T. Zijlstra. 2011. Starch with high amylose http://dx.doi.org/10.1128/aem.68.4.2049-2053.2002. content and low in vitro digestibility increases intestinal nutri- van Winsen, R. L., B. A. P. Urlings, L. J. A. Lipman, J. M. A.

ent flow and microbial fermentation and selectively promotes Snijders, D. Keuzenkamp, J. H. M. Verheijden, and F. van Knap- Downloaded from Bifidobacteria in pigs. J. Nutr. 141:1273–1280. http://dx.doi. en. 2001. Effect of fermented feed on the microbial population org/10.3945/jn.111.140509. of the gastrointestinal tracts of pigs. Appl. Environ. Microbi- Ridsdale, J. A., and J. E. L. Corry. 1999. Study of the caecal mi- ol. 67:3071–3076. http://dx.doi.org/10.1128/AEM.67.7.3071- croflora of broiler ducks: A comparison with published data for 3076.2001. chickens. Anaerobe 5:321–323. Vasaï, F., K. Brugirard Ricaud, M. D. Bernadet, L. Cauquil, O. Rinttilä, T., A. Kassinen, E. Malinen, L. Krogius, and A. Palva. 2004. Bouchez, S. Combes, and S. Davail. 2013. Overfeeding and genet- Development of an extensive set of 16S rDNA-targeted primers ics affect the composition of intestinal microbiota in Anas platy- http://ps.oxfordjournals.org/ for quantification of pathogenic and indigenous bacteria in faecal rhynchos (Pekin) and Cairina moschata (Muscovy) ducks. FEMS samples by real-time PCR. J. Appl. Microbiol. 97:1166–1177. Microbiol. Ecol. 12 http://dx.doi.org/10.1111/1574-6941.12217. http://dx.doi.org/10.1111/j.1365-2672.2004.02409.x. Walter, J., C. Hertel, G. W. Tannock, C. M. Lis, K. Munro, and W. Saengkerdsub, S., P. Herrera, C. L. Woodward, R. C. Anderson, D. P. Hammes. 2001. Detection of Lactobacillus, Pediococcus, Leu- J. Nisbet, and S. C. Ricke. 2007. Detection of methane and quan- conostoc, and Weissella species in human feces by using group- tification of methanogenic archaea in faeces from young broiler specific PCR primers and denaturing gradient gel electrophoresis. chickens using real-time PCR. Lett. Appl. Microbiol. 45:629– Appl. Environ. Microbiol. 67:2578–2585. 634. http://dx.doi.org/10.1111/j.1472-765X.2007.02243.x. Wang, X., I. L. Brown, D. Khaled, M. C. Mahoney, A. J. Evans, and

Salyers, A. A., J. R. Vercellotti, S. E. H. West, and T. D. Wilkins. P. L. Conway. 2002. Manipulation of colonic bacteria and vola- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 1977. Fermentation of mucin and plant polysaccharides by strains tile fatty acid production by dietary high amylose maize (amylo- of Bacteroides from the human colon. Appl. Environ. Microbiol. maize) starch granules. J. Appl. Microbiol. 93:390–397. 33:319–322. Watkins, B. A., and F. H. Kratzer. 1983. Effect of oral dosing of Lac- Schloss, P. D., S. L. Westcott, T. Ryabin, J. R. Hall, M. Hartmann, tobacillus strains on gut colonization and liver biotin in broiler E. B. Hollister, R. A. Lesniewski, B. B. Oakley, D. H. Parks, C. J. chicks. Poult. Sci. 62:2088–2094. Robinson, J. W. Sahl, B. Stres, G. G. Thallinger, D. J. Van Horn, Watkins, B. A., B. F. Miller, and D. H. Neil. 1982. In vivo inhibitory and C. F. Weber. 2009. Introducing mothur: Open-source, plat- effects of Lactobacillus acidophilus against pathogenic Escherichia form-independent, community-supported software for describing coli in gnotobiotic chicks. Poult. Sci. 61:1298–1308. and comparing microbial communities. Appl. Environ. Micro- Wright, A.-D. G., K. S. Northwood, and N. E. Obispo. 2009. Ru- biol. 75:7537–7541. http://dx.doi.org/10.1128/AEM.01541-09. men-like methanogens identified from the crop of the folivorous Servin, A. L. 2004. Antagonistic activities of lactobacilli and bifi- South American bird, the hoatzin (Opisthocomus hoazin). ISME dobacteria against microbial pathogens. FEMS Microbiol. Rev. J. 3:1120–1126. http://dx.doi.org/10.1038/ismej.2009.41. 28:405–440. http://dx.doi.org/10.1016/j.femsre.2004.01.003.

Understanding carbohydrate structures fermented or resistant to fermentation in broilers fed rapeseed (Brassica napus) meal to evaluate the effect of acid treatment and enzyme addition

Annemieke M. Pustjens ,* Sonja de Vries ,† Henk A. Schols ,* Harry Gruppen ,* Walter J. J. Gerrits ,† and Mirjam A. Kabel *1 Downloaded from * Laboratory of Food Chemistry, Wageningen University, Bornse Weilanden 9, 6708 WG, Wageningen, the Netherlands; and † Animal Nutrition Group, Wageningen University, De Elst 1, 6708 WD, Wageningen, the Netherlands

ABSTRACT Unprocessed and acid-extruded rapeseed proportion of unextractable carbohydrates increased in http://ps.oxfordjournals.org/ meal (RSM) was fed to broiler chickens, with and with- excreta from broilers fed acid-extruded RSM. Probably, out addition of commercial pectolytic enzymes. Non- acid extrusion resulted in a less accessible NSP matrix, starch polysaccharide (NSP) fermentability and unfer- also decreasing the accessibility for pectolytic enzymes mented NSP structures from RSM were studied in the added in the diet. During alkaline extraction of the ex- excreta in detail. From unprocessed RSM, 24% of the creta, 39 to 52% (wt/wt) of the insoluble carbohydrates nonglucose polysaccharides could be fermented. Acid was released as glucosyl- and uronyl-rich carbohy- treatment did not have a significant effect, but en- drates, probably originally present via ester linkages or zyme addition did improve fermentability to 38%. Most hydrogen bonding within the cellulose-lignin network. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 likely, the significant increase in NSP fermentability These linkages are expected to hinder complete NSP can be ascribed to the addition of pectolytic enzymes, fermentation and indicate that digestibility of RSM which decreased branchiness of the water-soluble arabi- may benefit substantially from an alkaline treatment or nan. Mainly xyloglucan, (glucurono-)xylan, (branched) addition of esterases. arabinan, and cellulose remained in the excreta. The

Key words: Brassica napus , in vivo digestion, broiler , carbohydrate structure , processing 2014 Poultry Science 93 :926–934 http://dx.doi.org/ 10.3382/ps.2013-03519

INTRODUCTION where water solubles can be fermented until secretion (Clench and Mathias, 1995). Also, reverse peristalsis Rapeseed meal (RSM) is a by-product from the pro- in the colon refluxes the soluble fraction of digesta to duction of rapeseed oil and used as animal feed for its the cecum to continue its fermentation (Duke, 1982). high protein content [around 38% wt/wt (Bell, 1993)]. In general, NSP fermentability in broilers is low, which Apart from protein, RSM is rich in nonstarch polysac- leaves part of the energy available in the unused feed. charides [NSP; 16–22% wt/wt (Slominski and Camp- In addition, the presence of NSP in the feed reduces di- bell, 1990)]. The NSP present in RSM are composed gestion of other nutrients due to their physicochemical of 1) pectic polysaccharides (e.g., homogalacturonan, properties; for example, starch and protein digestion rhamnogalacturonan, arabinan, and arabinogalactan), were reduced when isolated soluble NSP from wheat 2) hemicelluloses (e.g., xyloglucan, glucuronoxylan, and and rye were added to poultry diets (Smits and An- galactomannan), and 3) cellulose (Siddiqui and Wood, nison, 1996). 1977; Pustjens et al., 2013). An option to increase the energy value of RSM and Chickens lack the necessary digestive enzymes that to decrease the negative effect of NSP on other nu- can degrade NSP. Still, after the small intestine, the trients is to enhance the degradability of these poly- soluble fraction (and small particles) enters the ceca, saccharides by processing of the RSM before inclusion into the diet. From an in vitro study (Pustjens et al., 2012), it was concluded that acid treatment with addition of commercial pectolytic enzymes significantly © 2014 Poultry Science Association Inc. improves solubilization of NSP from RSM. The RSM Received July 25, 2013. Accepted December 20, 2013. cell wall structure showed to be only accessible for en- 1 Corresponding author: [email protected] zymes added in the feed after technological treatment,

926 UNFERMENTED CARBOHYDRATE STRUCTURES IN BROILERS 927 as was also indicated by the interaction between tech- mL per kilogram of dry feed for each enzyme mixture) nological treatment and enzyme addition found in vitro to the unprocessed RSM+ and acid-extruded RSM+ (Pustjens et al., 2012). As a follow-up experiment, diets diets. containing acid-treated RSM with and without the ad- During d 26 to 29, excreta were collected quantita- dition of commercial pectolytic enzymes were fed to tively per cage to calculate digestibility parameters and broilers. The aim of the current study was to determine perform statistical evaluation of the effect of the acid NSP fermentability and identify which NSP structures treatment, enzyme addition, or both. Also, bird perfor- were recalcitrant to fermentation. This results obtained mance per cage is calculated in this period. Digestibil- will assist in optimization of processing technologies of ity of DM, fat, and protein, and fermentability of NSP RSM to improve its digestibility in broilers. and constituent monosaccharides was calculated using the unprocessed RSM diet as the reference. In 2 cages Downloaded from MATERIALS AND METHODS (unprocessed RSM and acid-extruded RSM+), one or more broilers suffered from Escherichia coli infection, as Chemicals were of analytical grade and purchased confirmed by laboratory analyses of the Dutch animal from Sigma-Aldrich (St. Louis, MO) or VWR Interna- health service (GD, Deventer, the Netherlands). Hence, http://ps.oxfordjournals.org/ tional (Radnor, PA), unless stated differently. these cages were excluded from analyses. To enable in- depth analysis of undigested carbohydrate structures, excreta samples (containing feces plus urine) of 5 or 6 In Vivo Trial: Birds and Diets cages were pooled per diet before extraction and glyco- In a 2 × 2 factorial arrangement, the effect of pro- sidic linkage type analysis. cessing of RSM (unprocessed vs. acid extrusion) and enzyme addition (yes or no) was studied on the fate Extraction of Water- and Alkali-Solubles of the NSP fraction from RSM. The experimental pro- from Excreta at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 cedures were approved by and conformed to the requirements of the Animal Care and Use Committee of Unprocessed and acid-extruded RSM, and corre- Wageningen University, Wageningen, the Netherlands. sponding excreta were freeze-dried and milled using a Birds and Housing. Female one-day-old broiler ball mill (Retsch MM2000, Haan, Germany). The milled chickens (288 birds; Ross 308, Aviagen, Newbridge, samples (2.5 g) were extracted 3 times with 45 mL of UK; Hatchery Lagerweij, Lunteren, the Netherlands) demineralized water during 1 h at 40°C. After each ex- were housed and fed as described by De Vries et al. traction, the solubilized material was separated from (2014). In short, the experiment consisted of 24 cages, the insoluble residue by filtration over Whatman 595.5 with 6 cages per treatment. Each cage housed 12 broil- filters (Whatman, Kent, UK). Water-soluble fractions ers from d 14 to 25 and 11 broilers from d 26 to 29. were combined, freeze-dried, and denoted as water-sol- Experimental Diets. The ingredient composition of uble solids (WSS). Water-unextractable solids (WUS; the experimental diets is presented in Table 1. Rape- ~0.75 g) obtained were further extracted twice using seed meal (Cargill, Antwerp, Belgium, 2011, supplied 45 mL of 6 M NaOH containing 20 mM NaBH4 at 4°C. by Agrifirm, Apeldoorn, the Netherlands) was included After each extraction, material was centrifuged (48,400 in the diet at a level of 35% (wt/wt). Besides RSM, × g, 20 min, 4°C) and the supernatant was further maize was also included in all 4 diets at a level of 30%, separated from the insoluble residue by filtration on which contributes to the NSP content. The first diet G2-glass filters. Alkali-soluble fractions were combined, contained unprocessed RSM. The second diet con- neutralized using 6 M HCl, dialyzed against deminer- tained unprocessed RSM with addition of commercial alized water at 4°C (Visking dialysis tubes, molecular pectolytic enzymes (unprocessed RSM+), the third mass cut-off 12 to 14 kDa, pore diameter approximate- diet contained acid-extruded RSM (acid-extruded ly 25Å, Serva, Heidelberg, Germany), freeze-dried, and RSM), and the fourth diet contained acid-extruded denoted as alkali-soluble solids (ASS). The final resi- RSM with addition of pectolytic enzymes (acid-ex- due was neutralized using 6 M HCl, washed with water, truded RSM+). Prior to extrusion, RSM, water, and freeze-dried, and denoted as residue (RES). maleic acid were mixed using a paddle-mixer (type F60, Halvor Forberg, Bygland, Norway) to final contents of Enzymatic Fingerprinting 19.8% (wt/wt) and 1.4% (wt/wt), respectively. Within 1 h, acid extrusion of this mixture was performed with The WSS, ASS, and RES fractions from excreta (5 a double screw extruder (Baker-Perkins, Peterborough, mg/mL) were incubated with pure, well-characterized UK) at 120°C (product temperature 110.8 ± 1.7°C), enzymes, as described previously (Pustjens et al., 2013). 250 rpm screw speed, feeding rate 36 kg/h, and die size Digests were analyzed by high-pressure size exclusion 6 mm. Commercial enzyme mixtures Pectinex UltraSP chromatography, high-performance anion exchange (Novozymes, Bagsvaerd, Denmark) and Multifect Pec- chromatography, and matrix-assisted laser-induced de- tinase FE (Genencor, Rochester, NY) were added (8.75 sorption/ionization time-of-flight mass spectrometry. 928 PUSTJENS ET AL.

Table 1. Ingredient composition of the diets containing rapeseed meal (RSM; g/kg, as fed), calculated composition (g/kg, as fed) and analyzed content of starch, protein, fat, and NSP (% wt/wt DM) and molar composition of nonstarch polysaccharides (NSP; mol%)

Item Unpr. RSM1 Unpr. RSM+1 Acid RSM1 Acid RSM+1 Ingredient composition (g/kg, as fed) Unprocessed rapeseed meal 350 350 Acid-extruded rapeseed meal 350 350 Maize 300 300 300 300 Maize starch 158.15 158.15 158.15 158.15 Wheat gluten meal 50 50 50 50 Soy oil 45 45 45 45 Fishmeal 40 40 40 40

Soy protein isolate 25 25 25 25 Downloaded from Limestone 7 7 7 7 Mineral and vitamin premix2 5555 Monocalcium phosphate 4.5 4.5 4.5 4.5 Potassium bicarbonate 3 3 3 3 Sodium bicarbonate 1.1 1.1 1.1 1.1 Salt 1 1 1 1 l-Lysine HCl 3 3 3 3 http://ps.oxfordjournals.org/ d-Methionine 1.8 1.8 1.8 1.8 l-Valine 1 1 1 1 l-Threonine 0.8 0.8 0.8 0.8 l-Isoleucine 0.8 0.8 0.8 0.8 l-Tryptophan 0.4 0.4 0.4 0.4 l-Arginine 1.2 1.2 1.2 1.2 Co-EDTA 1 1 1 1 Chromium oxide 0.25 0.25 0.25 0.25 Pectolytic enzymes3 8.75 8.75 Calculated composition (g/kg, as fed) at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 DE (kcal/kg) 2,627 2,627 2,627 2,627 Calcium 7.6 7.6 7.6 7.6 Available phosphorus 3.1 3.1 3.1 3.1 Sodium 16 16 16 16 Digestible lysine 11.6 11.6 11.6 11.6 Digestible lysine + cysteine 9.5 9.5 9.5 9.5 Digestible threonine 7.8 7.8 7.8 7.8 Analyzed composition (% wt/wt DM) Starch 35 34 35 34 Protein 24 24 24 24 Fat 9 9 9 9 Nonstarch polysaccharides (NSP) 17 17 16 11 Molar composition of NSP4 (mol%) Rha 1 1 1 1 Ara 22 21 19 19 Xyl 16 16 14 12 Man 3 3 3 3 Gal 7 6 6 6 Glc 30 29 34 39 UA 21 24 23 20 1Unpr. RSM: unprocessed RSM, Acid RSM: acid-extruded RSM, +: including addition of pectolytic enzymes. 2 The mineral and vitamin premix provided the following per kilogram of feed: vitamin A: 12,000 IU; vitamin D3: 2,400 IU; vitamin E: 50 mg; vitamin K3: 1.5 mg; vitamin B1: 2 mg; vitamin B2: 7.5 mg; vitamin B6: 3.5 mg; vitamin B12: 20 μg; niacin: 35 mg; d-calcium pantothenate: 12 mg; choline chloride: 460 mg; folic acid: 1 mg; biotin: 0.2 mg; Fe: 80 mg (FeSO4·H2O); Cu: 12 mg (CuSO4·H2O); Mn: 85 mg (MnO); Zn: 60 mg (ZnSO4·H2O); Co: 0.4 mg (CoSO4·7H2O); I: 0.8 mg (KI); Se: 0.15 mg (Na2SeO3). 3Enzymes were added as liquid (mL/kg of feed). 4Rha = rhamnose, Ara = arabinose, Xyl = xylose, Man = mannose, Gal = galactose, Glc = glucose, UA = uronyl residues.

Analytical Methods esbaden, Germany). Nitrogen in the excreta was corrected for nitrogen from uric acid. Dry matter content was determined in duplicate by Total starch content was determined enzymatically drying overnight in an oven (WTC Binder, Tüttlingen, using the total starch assay procedure K-TSTA 04/2009 Germany) at 103°C. (Megazyme, Bray, Ireland). Protein content (N × 5.3; Mossé, 1990) was deter- Fat content was determined by extraction with pe- mined in duplicate by the Dumas method (AOAC In- troleum ether using the Soxhlet apparatus after hydro- ternational, 1995) on a Thermo Quest NA 2100 Ni- chloric acid hydrolysis (AOAC 920.39). trogen and Protein Analyzer (Interscience, Troy, NY). Nonstarch polysaccharide extraction from the feed Samples (~10 mg) were weighed into a cup and directly was performed as described elsewhere (Jonathan et al., analyzed. d-Methionine was used for calibration. 2013). In short, the starch present was gelatinized and Uric acid was analyzed enzymatic-colorimetrically enzymatically degraded, after which NSP could be pre- using a test kit (catalog no. 10694, Human GmbH, Wi- cipitated using acidified ethanol. UNFERMENTED CARBOHYDRATE STRUCTURES IN BROILERS 929 Neutral carbohydrate content and composition was can, and arabinan, as described before for B. napus determined in duplicate as per Englyst and Cummings meal (Pustjens et al., 2013). As expected, the content (1984) as described previously (Pustjens et al., 2013). and composition of RSM was not affected by acid ex- Uronic acid content was determined in duplicate ac- trusion (Table 2). Nevertheless, the NSP content as an- cording to the automated colorimetric m-hydroxydiphe- alyzed in the diet containing acid-extruded RSM+ was nyl (Thermo Fisher Scientific) assay including tetrabo- slightly lower (Table 1), which was a result of partly rate (Thibault, 1979) as described previously (Pustjens degradation by acid extrusion and the pectolytic en- et al., 2013). zymes added. The resulting soluble carbohydrates are High-pressure size exclusion chromatography was rinsed away with starch in the analytical procedure for performed as described previously (Pustjens et al., quantification of NSP from the feed and are therefore 2013), with the adaptation that the injection volume not account as NSP (Table 2). The carbohydrate con- Downloaded from was 10 μL and column oven temperature was 55°C. tent in RSM is variable and can differ between cultivars Enzyme digests were analyzed without prior dilution. (Theander et al., 1977) and can be affected by oil ex- High-performance anion exchange chromatography traction processes (Newkirk et al., 2003) and the pres- was performed as described previously (Pustjens et ence of the hull (Bell, 1993). In the diets, in addition to al., 2013). Enzyme digests were 10 times diluted before NSP from RSM, also NSP originating from maize was http://ps.oxfordjournals.org/ analysis. present (Table 1), which is estimated to be 15 to 20% Matrix-assisted laser-induced desorption/ionization (wt/wt) of the total NSP in the diet. Maize contains 6% time-of-flight mass spectrometry was performed using (wt/wt) NSP, besides 62% (wt/wt) starch, 11% (wt/ an Ultraflex workstation (Bruker Daltonics, Bremen, wt) protein, and some lignin. The molar composition of Germany) equipped with a nitrogen laser of 337 nm as maize NSP has been reported as 33 mol% glucosyl, 30 described previously (Pustjens et al., 2013). mol% xylosyl, and 23 mol% arabinosyl residues, being

Glycosidic linkage type analysis was performed as de- present as mixed-linked β-glucans, glucurono-arabinox- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 scribed elsewhere (Pustjens et al., 2013). ylan, and cellulose (Huisman et al., 2000).

Statistical Analysis Growth Performance In a 2 × 2 factorial arrangement, the effects of acid Growth performance of the broilers is presented in extrusion (T) and commercial pectolytic enzyme ad- Table 3. Average daily gain and ADFI were both lower dition (E) were studied. The results from the in vivo in broilers fed acid-extruded RSM (P = 0.001 and P < experiment were statistically analyzed using the GLM 0.001, respectively). Apparently, the acid-treated RSM procedure with a SAS program (version 9.2, SAS Insti- was less palatable compared with untreated RSM, as tute Inc., Cary, NC). The model used to describe the reported for the addition of fumaric and sorbic acid data was (Pirgozliev et al., 2008). The G:F of broilers fed unpro-

Yij = μ + Ti + Ej + Ti × Ej + eij, Table 2. Composition of unprocessed and acid-extruded rape- where Yij is the response variable, μ is the overall mean, seed meal (RSM; g/100 g of DM) Ti is the treatment (i = unprocessed, acid extrusion), Unpr. Acid Ej is the addition of commercial pectolytic enzymes (j Item RSM1 RSM1 = yes or no), Ti × Ej is the interaction between treat- Protein 35 34 ment and enzyme addition, and eij is the error term. Carbohydrate 31 30 Cage was the experimental unit. Residuals were tested Water-soluble saccharides (DP ≤4)2 17 15 for normality. For the excreta from broilers fed unpro- Water-soluble saccharides (DP >4)2 12 12 Water unextractable2 71 73 cessed RSM: n = 5; unprocessed RSM+: n = 6; acid-ex- Others (lignin, ash, fat)3 34 36 truded RSM: n = 6; and acid-extruded RSM+: n = 5. Molar composition of carbohydrates4 Rha 1 1 Ara 19 18 RESULTS AND DISCUSSION Xyl 8 8 Man 3 3 Characteristics of Unprocessed Gal 11 11 Glc 40 40 and Acid-Extruded RSM UA 18 19 1 The RSM (Table 2; unprocessed) contained 35% Unpr. RSM: unprocessed RSM; Acid RSM: acid-extruded RSM. 2Percentage of water-soluble fructose, saccharose, raffinose, and stach- (wt/wt) proteins and 31% (wt/wt) carbohydrates, in- yose [DP (degree of polymerization)DP ≤4], other water-soluble sac- cluding 26% (wt/wt) NSP. Other compounds in RSM charides (DP >4), and water unextractable carbohydrates from total are expected to be lignin, lipids, and ash (Bell, 1993). carbohydrates. 3 The RSM carbohydrates mainly contained glucosyl (40 Calculated as DM minus protein and minus carbohydrates. 4Mol%; molar composition presented as anhydrosugar moieties; Rha mol%), arabinosyl (19 mol%), and uronyl (18 mol%) = rhamnose, Ara = arabinose, Xyl = xylose, Man = mannose, Gal = residues, indicating the presence of cellulose, xyloglu- galactose, Glc = glucose, UA = uronyl residues. 930 PUSTJENS ET AL.

Table 3. Growth performance of broilers fed diets containing unprocessed or acid-extruded rapeseed meal (RSM) with (+) or without (−) addition of pectolytic enzymes

Unpr. RSM1 Acid RSM2 Model P-value

Pooled Processing Enzyme Item − + − + SEM (P) (E) P × E n3 56 65 Initial BW (g) 349.8 349.4 349.7 349.5 — — — — Final mean BW (g) 1,068 1,059 1,008 998 — — — — ADG (g/bird) 64.1 62.4 58.8 56.6 14.3 0.001 0.161 0.849 ADFI (g of DM/bird) 81.3 81.1 75.7 74.4 18.4 <0.001 0.381 0.463

G:F4 (g/g) 0.79 0.77 0.78 0.76 0.18 0.560 0.186 0.893 Downloaded from 1Unpr. RSM: unprocessed RSM. 2Acid RSM: acid-extruded RSM. 3Number of replicate cages of 12 broilers each. 4G:F (g of growth/g of feed). http://ps.oxfordjournals.org/ cessed RSM was 0.79 kg of growth/kg of feed between This can only be explained by the fact that enzyme d 14 and 25 of the trial (Table 3). This was not affected addition can significantly improve NGP fermentabil- by acid extrusion (P = 0.560) or by the addition of ity from unprocessed RSM and not from acid-extrud- pectolytic enzymes (P = 0.186), or by a combination of ed RSM. Opposed to the conclusion from our in vitro acid extrusion and enzyme addition (P = 0.893). study (Pustjens et al., 2012), apparently, acid extrusion made the matrix less accessible for enzymes during in Fermentability of (Enzyme-Added) vivo fermentation rather than more accessible. Even at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Unprocessed and Acid-Extruded RSM though addition of commercial pectolytic enzymes improved NGP fermentability of unprocessed RSM, still The nonglucose polysaccharide (NGP) fermentabil- 62% remained unfermented in this study. These unfer- ity was calculated rather than NSP fermentability, as mented carbohydrate structures were analyzed in more excreta still contained residual starch [analyzed to be 2 detail. Understanding these structures in detail will as- to 3% (wt/wt)]. Twenty-four percent of the NGP in the sist in optimizing pretreatment for improved digestion unprocessed RSM diet was fermented (Table 4). Addi- of RSM. tion of commercial pectolytic enzymes significantly improved NGP fermentability to 38% (P < 0.001; Table Distribution of Unfermented Carbohydrates 4). This was mainly visible in the improved fermentability of arabinosyl residues (P < 0.001) and xylo- The excreta of broilers fed diets containing unpro- syl residues (P = 0.005). Acid extrusion did not affect cessed and acid-extruded RSM contained 40 to 44% NGP fermentability (P = 0.567). (wt/wt) carbohydrates and 30 to 32% (wt/wt) pro- In the diet containing acid-extruded RSM+, NGP teins. Excreta were pooled per diet to enable detailed fermentability improved to 32%. This increase is lower characterization of unfermented carbohydrates after ex- than when enzymes were added to unprocessed RSM. traction with water and 6 M NaOH. Figure 1 shows the In NGP fermentability, an interaction was found be- distribution of carbohydrates fermented or unferment- tween acid extrusion and enzyme addition (P = 0.008). ed. In addition, in Figure 1 the unfermented carbohy-

Table 4. Total tract digestibility (%) of DM, fat, and protein, and fermentability of nonglucose polysaccharides (NGP), nonstarch polysaccharides (NSP), arabinosyl (Ara), xylosyl (Xyl), and uronyl (UA) residues from diets containing unprocessed or acid-extruded rapeseed meal with (+) and without (−) the addition of commercial pectolytic enzymes

Total tract digestibility and fermentability (%)

Item DM Fat Protein NGP NSP Ara Xyl UA Diet Unpr. RSM−1 67 88 72 24 20 22 31 31 Unpr. RSM+ 70 90 74 38 36 41 41 43 Acid RSM−1 67 90 70 29 24 30 37 32 Acid RSM+ 68 89 71 32 26 30 39 34 Pooled SE 0.43 0.40 0.49 0.97 1.51 1.34 1.01 1.34 P-value (acid)2 0.140 0.945 0.008 0.567 0.219 0.719 0.531 0.259 P-value (enzyme)2 0.017 0.374 0.037 <0.001 0.004 <0.001 0.005 0.107 P-value (acid × enzyme)2 0.082 0.087 0.366 0.008 0.020 0.035 0.048 0.193 1Unpr. RSM: unprocessed RSM; Acid RSM: acid-extruded RSM. 2Differences among means with P < 0.05 were considered to represent significant differences. UNFERMENTED CARBOHYDRATE STRUCTURES IN BROILERS 931 Chakraborty, 2006). Thereby, noncovalently bound (e.g., by H-bonding) carbohydrate structures can be released. Third, the small carbohydrates originating from NSP could be bound to the microbial cell surface proteins, which is a known digestive mechanism for Bac- teroidetes (McBride et al., 2009), which make up 23 to 46% of the bacteria in the ceca of chicken (Yeoman et al., 2012). As described for the first explanation, alkali will then release these small carbohydrates from the microbial cell surface. We believe the third theory is less likely because of the relative high amounts of Downloaded from carbohydrates released. Also, first indications (high- pressure size exclusion chromatography and high-performance anion exchange chromatography) about the size of these carbohydrates released (data not shown) show, besides very small carbohydrates, larger material http://ps.oxfordjournals.org/ than those expected to bind to cell surface proteins. Further research will be conducted to analyze the nature of these carbohydrates released. These explana- Figure 1. Distribution of nonglucose polysaccharides from rape- tions indicate that the broilers’ microbiota was able to seed meal being either fermented or unfermented. The unfermented ferment part of the polysaccharides present, but full carbohydrates are further divided as water-soluble solids (WSS), cal- fermentation was hindered by ester bonds or hydrogen culated alkali-soluble solids <14 kDa (ASSd), alkali-soluble solids >14 kDa (ASS), and final residue (RES) of excreta from pigs fed unpro- bridges. Therefore, carbohydrates in the ASSd fraction, at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 cessed rapeseed meal (Unpr. RSM), unprocessed RSM with addition being a substantial part of the unfermented carbohy- of commercial pectolytic enzymes (Unpr. RSM+), acid-extruded RSM drates, could be made available for fermentation by the (Acid RSM), and acid-extruded RSM with the addition of commercial animal by alkaline pretreatment or by esterases added pectolytic enzymes (Acid RSM+). to the RSM. After water and alkaline extraction, still 32 to 37% drates were further divided into a WSS, ASS, and RES. (wt/wt) of the carbohydrates in the excreta of broil- In all WSS, 16 to 18% (wt/wt) of the total amount of ers fed unprocessed RSM, unprocessed RSM+, and carbohydrates in the excreta was extracted (Table 5). acid-extruded RSM+ remained in the RES. This in- Calculated from the NSP in WUS minus the NSP in creased to 45% (wt/wt) in the excreta of broilers fed ASS and RES, 52 to 58% (wt/wt) of NSP present in acid-extruded RSM, suggesting that the NSP matrix WUS was extracted with alkali. However, 39 to 52% became more rigid in these excreta. This explains why (wt/wt) of these NSP was not recovered after dialy- acid extrusion did not result in an improved NGP fer- sis of the ASS (Figures 1 and 2). This suggested that mentability. an important part of the water unextractable carbohydrates was released by alkali and had a rather low Characterization of Unfermented Water- molecular mass (<14 kDa, cut-off value of the dialysis Soluble Carbohydrate Structures membrane). In Figure 1, these alkali-extracted carbohydrates are referred to as ASSd <14 kDa (alkali-soluble The water-soluble fraction of excreta consisted of 34 solids dialyzed). Acid extrusion or addition (or both) of to 47 mol% glucosyl, 15 to 16 mol% arabinosyl, 13 to commercial pectolytic enzymes tended to decrease the 19 mol% uronyl, and 12 to 13 mol% galactosyl residues relative amount of the released fraction (ASSd), where- (Table 5). In the WSS of excreta from broilers fed un- as the amounts recovered for RES remained the same. processed RSM+ and broilers fed acid-extruded RSM+, A similar alkaline extraction of NSP from unpro- the molar proportion of glucosyl residues decreased and cessed RSM, as added in the diet, did not result in those of xylosyl and uronyl residues increased, compared such a release of small carbohydrates (no further data with excreta from broilers fed unprocessed RSM. The shown). For this phenomenon observed, 3 possible ex- significant improvement in fermentability of arabinosyl planations were considered. First, small carbohydrates residues, even though the molar proportion of arabi- (<14 kDa), remaining after partly fermentation of the nosyl residues in the WSS-fraction was not affected by original RSM polysaccharide structures, were present acid extrusion or addition of pectolytic enzymes. Part in the excreta linked via ester linkages. Ester linkages of the small saccharides, such as fructose, raffinose, and to proteins and carbohydrates in RSM are reported to stachyose [together 4–8% (wt/wt) of the carbohydrates be present through sinapic, ferulic, and Coumaric acid in the WSS fraction], present in the RSM in the diets (Kozlowska et al., 1983). Alkaline extraction will break (Table 2) remained in the excreta (data not shown). As such bonds, thereby solubilizing these partly fermented expected, saccharose was completely digested (Klein et carbohydrates. Second, 6 M NaOH makes cellulose fi- al., 2001). Apart from the small saccharides mentioned, brils swell, as published for bamboo fibers (Das and the constituent monosaccharide composition of larger 932 PUSTJENS ET AL.

Table 5. Yield and carbohydrate composition of water-soluble solids (WSS), water-unextractable solids (WUS), alkali-soluble solids >14 kDa (ASS), alkali-soluble solids <14 kDa (ASSd), and residue (RES) fractions of excreta after in vivo digestion in broilers fed diets containing unprocessed and acid-extruded rapeseed meal (RSM) with (+) or without (−) addition of commercial pectolytic enzymes

Carbohydrate

Yield Content Yield Item (%)1 (wt/wt %) (%) Rha2 Ara2 Xyl2 Man2 Gal2 Glc2 UA2 Excreta diet 1 (Unpr. RSM−)3 100 41 100 1 16 7 1 4 44 27 WSS 38 17 16 1 15 7 5 12 47 13 WUS 46 66 74 1 21 15 2 6 31 24 Recovery 14 84 90 Downloaded from ASS 4 78 5 tr.5 20 32 5 10 24 9 RES 46 48 43 1 18 7 1 4 40 29 Recovery 26 50 48 Excreta diet 2 (Unpr. RSM+)7 100 44 100 1 11 7 1 4 53 23 WSS 40 20 18 1 16 12 5 13 34 19 WUS 54 59 73 1 14 18 2 6 38 21 http://ps.oxfordjournals.org/ Recovery 1 94 91 ASS 6 70 7 tr. 17 35 7 9 24 8 RES 64 59 48 1 11 7 1 4 53 23 Recovery 2 70 55 Excreta diet 3 (Acid RSM−)3 100 41 100 1 17 8 1 4 45 24 WSS 39 17 17 1 15 7 5 13 46 13 WUS 57 63 88 1 17 15 2 6 32 27 Recovery 1 96 105 ASS 5 89 7 tr. 20 32 5 10 24 9 RES 57 72 51 1 17 8 1 4 45 24 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Recovery 2 62 58 Excreta diet 4 (Acid RSM+)7 100 40 100 1 13 7 1 4 53 21 WSS 38 19 18 1 16 12 5 13 34 19 WUS 56 63 89 1 15 15 3 6 39 21 Recovery 1 94 107 ASS 15 79 19 tr. 16 34 5 10 26 9 RES 77 59 42 1 13 7 1 4 53 21 Recovery 2 92 61 1Dry matter. 2Mol%; molar composition presented as anhydrosugar moieties; Rha = rhamnose; Ara = arabinose; Xyl = xylose; Man = mannose; Gal = galactose; Glc = glucose; UA = uronyl residues. 3Unpr. RSM: unprocessed RSM; acid RSM: acid-extruded RSM. 4Recovery = recovery after water extraction. 5tr.: trace amounts. 6Recovery 2 = recovery after alkaline extraction. 7Numbers in italics represent recoveries; numbers in bold represent the carbohydrate composition of the excreta before extraction. carbohydrates was calculated to be 34 to 48 mol% glu- mainly 31 mol% glucosyl, 24 mol% uronyl, 21 mol% cosyl, 15 to 22 mol% uronyl, 13 to 14 mol% arabinosyl, arabinosyl, and 15 mol% xylosyl residues (Table 5). and 13 to 14 mol% galactosyl residues. Due to the higher fermentability of arabinosyl residues, The carbohydrate structures in WSS-fractions from in the WUS of excreta from broilers fed unprocessed excreta were further analyzed using glycosidic linkage RSM+ and broilers fed acid-extruded RSM+, arabino- type analysis (Supplemental Table 1; http://dx.doi. syl residues decreased and glucosyl residues increased org/10.3382/ps.2013-03519) and enzymatic fingerprint- compared with excreta from broilers fed unprocessed ing (data not shown). The main water-soluble poly- RSM. saccharides present in the excreta of broilers fed un- The analyzed alkali-soluble fractions (ASS > 14 kDa) processed RSM were branched arabinan (both at the in all the excreta samples were rather similar and con- O2- and O3-position), xyloglucan (both XXGG- and tained mainly 32 to 35 mol% xylosyl, 24 to 26 mol% XXXG-type), linear xylan and galactomannan. The ad- glucosyl, and 16 to 20 mol% arabinosyl residues (Table dition of pectolytic enzymes decreased the branchiness 5). With linkage-type analysis and enzymatic finger- of the arabinan, thereby helping in fermenting arabi- printing, these alkali-soluble polysaccharides were ana- nans in broilers. lyzed to be arabinan, homogalacturonan, rhamnogalacturonan, branched glucurono-arabinoxylan, xyloglucan, Characterization of Unfermented Water and galactomannan. The arabinoxylan is probably Unextractable Carbohydrate Structures originating from maize present in the diets (Table 1), which is known to contain highly branched arabinox- Water-unextractable solids from the excreta of broil- ylan structures (Huisman et al., 2000). The other poly- ers fed unprocessed and acid-extruded RSM contained saccharides originate from RSM. UNFERMENTED CARBOHYDRATE STRUCTURES IN BROILERS 933 As discussed above, 39 to 52% of the unextractable carbohydrates from excreta was calculated to be released by 6 M NaOH, but not recovered in the ASS or RES fractions. In the ASSd, the main constituent monosaccharides were calculated to be 25 to 37 mol% glucosyl and 24 to 36 mol% uronyl, next to 11 to 19 mol% xylosyl, 11 to 17 mol% arabinosyl, 6 to 8 mol% galactosyl, and 3 mol% mannosyl residues. The calculated release of glucosyl- and uronyl-residues in ASSd was relatively high compared with the other constituent monosaccharides (as can be seen in Figure 2). The re- Downloaded from leased glucosyl residues probably originate from (xylo) glucans previously bound to cellulose by H-bonding, as previously suggested (Pauly et al., 1999). The released uronyl residues may originate from pectins that were ester linked in the cell wall material, which has been http://ps.oxfordjournals.org/ hypothesized to occur via the acid group of galacturon- ic acid (Brown and Fry, 1993). The carbohydrate structures found in the ASS fractions of excreta from broilers fed all 4 diets were similar. Apparently, the effect of acid extrusion and enzyme addition only had an effect on the proportion of alkali-

soluble carbohydrate, whereas there was no selective at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 effect of acid extrusion and addition of pectolytic enzymes on the structure of alkali-soluble carbohydrates analyzed. The unextractable residual fraction (RES) from excreta from broilers fed unprocessed RSM contained mainly 43 mol% glucosyl, 27 mol% uronyl, and 16 mol% arabinosyl residues (Table 5). This molar composition indicates the presence of cellulose and tightly bound pectins (Oechslin et al., 2003; Zykwinska et al., 2008). In conclusion, NGP fermentability of unprocessed RSM significantly improved from 24 to 38% after enzyme addition, mostly caused by an increase in fermentability of arabinosyl and xylosyl residues. Acid extrusion was not able to improve NSP digestibility significantly. Acid extrusion resulted in a more rigid NSP matrix in vivo, which also decreased the accessibility for pectolytic enzymes added in the diet. In addition, our study indicated that during alkaline extraction of the WUS from excreta, half of the carbohydrates from excreta were not recovered in ASS and RES. These unrecovered carbohydrates were expected to be linked via H-bonding or ester linkages within the cellulose-lignin network of RSM. Therefore, alkaline pretreatment or addition of esterases may substantially improve NSP fermentability from RSM.

ACKNOWLEDGMENTS The authors thank MS students Hsuan Chen, Dorien Figure 2. Distribution of arabinosyl (A), xylosyl (B), and uro- te Peele, Maria Shipandeni, and personnel of experi- nyl (C) residues being fermented or unfermented. The unfermented mental farm De Haar for contributing to the research. residues are further divided as water-soluble solids (WSS), calculated alkali-soluble solids <14 kDa (ASSd), alkali-soluble solids >14 kDa This project is jointly financed by the European (ASS), and final residue (RES) of excreta from pigs fed unprocessed Union, European Regional Development Fund and The rapeseed meal (Unpr. RSM), unprocessed RSM with addition of com- Ministry of Economic Affairs, Agriculture and Inno- mercial pectolytic enzymes (Unpr. RSM+), acid-extruded RSM (Acid RSM), and acid-extruded RSM with addition of commercial pectolytic vation, Peaks in the Delta, the Municipality of Gron- enzymes (Acid RSM+). ingen, and the Province of Groningen as well as the 934 PUSTJENS ET AL. Dutch Carbohydrate Competence Center (CCC WP7), Novel features of the polysaccharide-digesting gliding bacterium supported by Agrifirm Group (Apeldoorn, the Neth- Flavobacterium johnsoniae as revealed by genome sequence analysis. Appl. Environ. Microbiol. 75:6864–6875. erlands), Duynie Holding (Alphen aan den Rijn, the Mossé, J. 1990. Nitrogen to protein conversion factor for ten cereals Netherlands), Nutreco Nederland B.V. (Boxmeer, the and six legumes or oilseeds. A reappraisal of its definition and Netherlands), and Wageningen University (Wagenin- determination. Variation according to species and to seed protein gen, the Netherlands). content. J. Agric. Food Chem. 38:18–24. Newkirk, R. W., H. L. Classen, and M. J. Edney. 2003. Effects of prepress-solvent extraction on the nutritional value of canola REFERENCES meal for broiler chickens. Anim. Feed Sci. Technol. 104:111–119. Oechslin, R., M. V. Lutz, and R. Amadò. 2003. Pectic substances AOAC International. 1995. Official Methods of Analysis. AOAC isolated from apple cellulosic residue: Structural characterisa-

Int., Arlington, VA. tion of a new type of rhamnogalacturonan I. Carbohydr. Polym. Downloaded from Bell, J. M. 1993. Factors affecting the nutritional value of canola 51:301–310. meal: A review. Can. J. Anim. Sci. 73:689–697. Pauly, M., L. N. Andersen, S. Kauppinen, L. V. Kofod, W. S. York, Brown, J. A., and S. C. Fry. 1993. The preparation and susceptibil- P. Albersheim, and A. Darvill. 1999. A xyloglucan-specific endo- ity to hydrolysis of novel O-galacturonoyl derivatives of carbohy- β-1,4-glucanase from Aspergillus aculeatus: Expression cloning in drates. Carbohydr. Res. 240:95–106. yeast, purification and characterization of the recombinant enzyme. Glycobiology 9:93–100.

Clench, M. H., and J. R. Mathias. 1995. The avian cecum: A review. http://ps.oxfordjournals.org/ Wilson Bull. 107:93–121. Pirgozliev, V., T. C. Murphy, B. Owens, J. George, and M. E. E. Das, M., and D. Chakraborty. 2006. Influence of alkali treatment on McCann. 2008. Fumaric and sorbic acid as additives in broiler the fine structure and morphology of bamboo fibers. J. Appl. feed. Res. Vet. Sci. 84:387–394. Polym. Sci. 102:5050–5056. Pustjens, A. M., S. De Vries, W. J. J. Gerrits, M. A. Kabel, H. A. De Vries, S., A. M. Pustjens, M. A. Kabel, R. P. Kwakkel, and W. Schols, and H. Gruppen. 2012. Residual carbohydrates from in J. J. Gerrits. 2014. Effects of processing technologies and pecto- vitro digested processed rapeseed (Brassica napus) meal. J. Ag- lytic enzymes on degradability of nonstarch polysaccharides from ric. Food Chem. 60:8257–8263. rapeseed meal in broilers. Poult. Sci. 93:589–598. Pustjens, A. M., H. A. Schols, M. A. Kabel, and H. Gruppen. 2013. Duke, G. E. 1982. Gastrointestinal motility and its regulation. Characterisation of cell wall polysaccharides from rapeseed Poult. Sci. 61:1245–1256. (Brassica napus) meal. Carbohydr. Polym. 98:1650–1656. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Englyst, H. N., and J. H. Cummings. 1984. Simplified method for Siddiqui, I. R., and P. J. Wood. 1977. Carbohydrates of rapeseed: A the measurement of total non-starch polysaccharides by gas- review. J. Sci. Food Agric. 28:530–538. liquid chromatography of constituent sugars as alditol acetates. Slominski, B. A., and L. D. Campbell. 1990. Non-starch polysac- Analyst (Lond.) 109:937–942. charides of canola meal: Quantification, digestibility in poultry Huisman, M. M. H., H. A. Schols, and A. G. J. Voragen. 2000. and potential benefit of dietary enzyme supplementation. J. Sci. Glucuronoarabinoxylans from maize kernel cell walls are more Food Agric. 53:175–184. complex than those from sorghum kernel cell walls. Carbohydr. Smits, C. H. M., and G. Annison. 1996. Non-starch plant polysac- Polym. 43:269–279. charides in broiler nutrition—Towards a physiologically valid ap- Jonathan, M. C., D. Haenen, C. Souza da Silva, G. Bosch, H. A. proach to their determination. World’s Poult. Sci. J. 52:217–221. Schols, and H. Gruppen. 2013. Influence of a diet rich in resistant Theander, O., P. Aman, G. E. Miksche, and S. Yasuda. 1977. Carbo- starch on the degradation of non-starch polysaccharides in the hydrates, polyphenols, and lignin in seed hulls of different colors large intestine of pigs. Carbohydr. Polym. 93:232–239. from turnip rapeseed. J. Agric. Food Chem. 25:270–273. Klein, M., M. Neubert, L. Hoffmann, W. Jentsch, M. Beyer, H. Thibault, J.-F. 1979. Automisation du dosage des substances pec- Scholze, and S. Kuhla. 2001. Energy metabolism of cocks and tiques par la methode au meta-hydroxydihpenyl. Lebensm. Wiss. broiler chickens fed on diets with different carbohydrate sources. Technol. 21:247–251. Arch. Tierernahr. 55:207–220. Yeoman, C. J., N. Chia, P. Jeraldo, M. Sipos, N. D. Goldenfeld, and Kozlowska, H., D. A. Rotkiewicz, R. Zadernowski, and F. W. Sos- B. A. White. 2012. The microbiome of the chicken gastrointesti- ulski. 1983. Phenolic acids in rapeseed and mustard. J. Am. Oil nal tract. Anim. Health Res. Rev. 13:89–99. Chem. Soc. 60:1119–1123. Zykwinska, A., J. F. Thibault, and M. C. Ralet. 2008. Competitive McBride, M. J., G. Xie, E. C. Martens, A. Lapidus, B. Henrissat, R. binding of pectin and xyloglucan with primary cell wall cellulose. G. Rhodes, E. Goltsman, W. Wang, J. Xu, D. W. Hunnicutt, A. Carbohydr. Polym. 74:957–961. M. Staroscik, T. R. Hoover, Y.-Q. Cheng, and J. L. Stein. 2009.

Dietary electrolyte balance influences ileal endogenous amino acid losses in broiler chickens

S. A. Adedokun1 and T. J. Applegate

Department of Animal Sciences, Purdue University, West Lafayette, IN 47907-2054

ABSTRACT The objective of the current study was to etary dextrose level (corn starch-to-dextrose ratio of Downloaded from determine the effect of nitrogen-free diets formulated 0.31) resulted in higher (P < 0.05) ileal DM and energy with 2 ratios of corn starch and dextrose and 2 levels digestibility. A high level of DEB resulted in higher (P of dietary electrolyte balance (DEB; calculated as Na < 0.05) ileal endogenous nitrogen loss. Arginine, Ile, + K – Cl−, in mEq/kg of diet) on ileal endogenous Leu, Phe, Val, Ala, Glu, Gly, Pro, and Tyr secretion amino acid (EAA) losses in 48-d-old broiler chickens. into the gut increased (P ≤ 0.05) with an increasing http://ps.oxfordjournals.org/ On d 43, 240 broiler chickens were allotted to 4 di- level of DEB. Endogenous His, Lys, Thr, Asp, Cys, and etary treatments with 6 replicate cages per diet in a Ser showed a tendency for increased (P ≤ 0.1) losses completely randomized design. Each experimental diet with a high level of DEB. Mean ileal EAA losses for the was fed for 5 d (d 43–48). All diets were free of nitro- indispensable and dispensable amino acids for the low- gen with 2 ratios of corn starch-to-dextrose (0.31 and DEB diet were 81 and 82%, respectively, that of the 1.04) and 2 levels of DEB (108 and 219 mEq/kg of birds fed the high-DEB diet. The 4 amino acids with

diet). Final BW and BW loss between birds fed the 2 the lowest endogenous flow were Met, His, Tyr, and at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 corn starch-to-dextrose ratios and between birds fed 2 Cys, whereas the highest endogenous flow was found in levels of DEB were not different (P > 0.05). Birds fed Glu, Asp, Val, and Leu. Data from the current study the diet containing a high level of DEB consumed less showed that EAA losses in the ileum of broiler chickens (P < 0.05) feed than birds on the low-DEB diet. Birds are significantly affected by DEB, but not by the ratio on the high-corn starch-to-dextrose (1.04) diet showed of corn starch to dextrose. P a tendency for higher ( = 0.08) feed intake. High di- Key words: broiler , corn starch , dextrose , endogenous amino acid, dietary electrolyte balance 2014 Poultry Science 93 :935–942 http://dx.doi.org/ 10.3382/ps.2013-03661

INTRODUCTION lian et al., 2008), proportion of the different sources of energy (corn starch vs. dextrose) in a nitrogen-free diet The importance of accurately determining ileal en- (NFD; Kong and Adeola, 2013), fiber levels in healthy EAA dogenous amino acid ( ) losses in poultry to stan- and stressed (challenged with mild coccidial vaccine) AA dardize apparent ileal amino acid ( ) digestibility broilers (Adedokun et al., 2011), as well as phytase values has been highlighted (Angkanaporn et al., 1994; and phytic acid (Cowieson et al., 2004; Cowieson and Lemme et al., 2004; Adedokun et al., 2007c, 2011; Klu- Ravindran, 2007; Onyango et al., 2009) on ileal EAA th and Rodehutscord, 2009; Kong and Adeola, 2013). losses have been studied. In a review on EAA losses To accurately determine EAA flow at the terminal il- in poultry, Adedokun et al. (2011) discussed some of eum of poultry, examining the effects of different fac- these factors, including factors that could influence en- tors that may contribute to or affect ileal EAA losses in dogenous mucin turnover rate, gut health, and the role poultry is important. Several of these factors have been of dietary electrolyte balance (DEB). However, the examined. The effects of age and strain of birds (Ra- DEB of the different formulation of nitrogen-free diet vindran and Hendriks, 2004; Adedokun et al., 2007b,c, (NFD) that have been employed varies considerably, 2009), dietary protein and AA levels, methods of es- as reported by Adedokun et al. (2011). Dietary electro- timating EAA losses (Ravindran and Hendriks, 2004; lyte balance has been reported to influence acid base Ravindran et al., 2004; Adedokun et al., 2007a,c; Go- balance (Patience, 1990), broiler performance, and the absorption of monosaccharides and AA (Johnson and Karunajeewa, 1985; Ravindran et al., 2008). Different © 2014 Poultry Science Association Inc. types of NFD, especially as it relates to the DEB, in Received September 30, 2013. + + − Accepted December 16, 2013. milliequivalent values of Na + K − Cl per kilogram 1 Corresponding author: [email protected] of diet, have been used to estimate ileal EAA losses.

935 936 ADEDOKUN AND APPLEGATE The DEB of the NFD used in the studies reported by complete randomized design in such a way that the av- Adedokun et al. (2007b,c) was 219 mEq/kg of diet, erage weight per cage and treatment were similar across whereas the NFD used by Golian et al. (2008) was 250 treatments using the Experimental Animal Allotment mEq/kg of diet. Program of Kim and Lindemann (2007). Cage served Kong and Adeola (2013) showed that the response of as the experimental unit. Each experimental diet was broiler chickens fed an NFD with energy supplied solely fed for 5 d (d 43–48) and all birds had unrestricted ac- by corn starch or dextrose were different. They report- cess to feed and water. Birds were housed and raised, ed higher ileal EAA losses in birds fed an NFD whose including bird care and experimental procedures, in energy source was solely from dextrose compared with line with a protocol approved by the Purdue University birds fed an NFD with corn starch as the only source of Animal Care and Use Committee. energy. The DM digestibility values increased with in- At the end of the 5-d feeding period, birds were Downloaded from creasing dietary dextrose levels, with a concomitant de- weighed individually and euthanized by cage by CO2 crease in the quantity of ileal digesta collected from the asphyxiation, then ileal contents (entire ileum) were ileum (Kong and Adeola, 2013). Despite the plethora flushed with distilled water into clean containers and of available information on ileal EAA losses in broilers, stored at −20°C until they were freeze-dried. The ileum the need to further investigate the effect of different was considered to be the portion of the small intestine http://ps.oxfordjournals.org/ components of NFD on EAA losses in broilers exists. It from Meckel’s diverticulum to about 5 mm proximal to was hypothesized that high DEB and low dietary corn the ileo-cecal-colonic junction. Dried ileal digesta and starch-to-dextrose ratio would result in an increase in diets were analyzed for DM, complete AA profile, gross basal EAA flow in 48-d-old broiler chickens. The aim of energy, and chromium. Additionally, diets were ana- the current study was to investigate the importance of lyzed for Na, K, Cl−, Ca, and P. 2 levels of DEB and 2 ratios of corn starch to dextrose on ileal EAA losses in mature broilers fed an NFD. Chemical Analyses at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 MATERIALS AND METHODS Dry matter content was determined on ground diets and ileal digesta by drying the samples at 105°C for Diet Formulation 16 h (method 934.01; AOAC International, 2006). The gross energy contents of diets were determined by adia- A total of 4 dietary treatments (A, B, C, and D) were batic bomb calorimetry (model 1261; Parr Instrument used in the current study. Diets A and B were similar Co., Moline, IL), with benzoic acid as a standard. Ileal in corn starch-to-dextrose level (0.31), but differed in digesta and dietary AA contents, Na, K, Cl−, Ca, and DEB (108 and 219 mEq/kg of diet, respectively). Like- P in diets were determined at the University of Missouri wise, diets C and D had a similar corn-to dextrose ratio Experiment Station Chemical Laboratory (Columbia). (1.04), but differed in DEB (108 and 219 mEq/kg of Preparation and analysis of AA [method 982.30 E (a, b, diet, respectively). To obtain a desirable DEB, sodium c); AOAC International, 2006] in diets and ileal digesta chloride (salt) was added to diets B and D at the rate were as described by Adedokun et al. (2012). of 2.0 g/kg of diet. Purified cellulose (Solka-Floc) and Chromium concentrations in diets and ileal digesta soy oil were added to each diet at 50 g/kg of diet (Table were determined by spectrophotometry (AOAC Inter- 1). Chromium III oxide was added to each diet at 5 g/ national, 2000; method 946.06) and absorbance was kg of diet as an indigestible marker for digestibility read using a Dynex plate reader (Dynex Technologies calculation. Birds had free access to feed and water Inc., Chantilly, VA). Calcium and P concentrations in throughout the duration of the study. The analyzed the diets were also determined. Phosphorus concentra- sodium (Na), potassium (K), chloride (Cl−), Ca, P, N, tion was determined by spectrophotometry (AOAC In- and gross energy contents of the experimental diets are ternational, 2000; method 946.06) and Ca concentra- reported in Table 1, and the analyzed dietary AA con- tions were determined from the same digested samples tents are reported in Table 2. using the flame atomic absorption spectroscopy method (Varian Spectr.AA 220FS, Varian Australia Pty Ltd., Birds, Housing, Sampling, and Ileal Mulgrave, Australia). Sodium, K, and Cl− concentra- Digesta Processing tions in the diets were also determined (AOAC Interna- tional, 2000; method 976.25). Nitrogen was determined Two hundred sixty 1-d-old Ross 708 broiler chicks by the combustion method (model FP2000, Leco Corp., were raised on deep litter until d 43. They were fed in- St. Joseph, MI; AOAC International, 2000; method dustry representative broiler corn-soybean meal starter 990.03), with EDTA serving as the internal standard. (d 0–14), grower (d 14–28), and finisher (d 28–42) diets. All diets were adequate in all nutrients and energy Calculations and Statistical Analysis for the appropriate phases of growth (NRC, 1994). On d 43, all birds were tagged, weighed individually, and Dietary electrolyte balance, in milliequivalents, was sorted, with 240 birds randomized into 4 dietary treat- calculated as the mEq of Na + mEq of K – mEq of Cl−. ments with 10 birds/cage and 6 cages/treatment in a The milliequivalent value of each of Na, K, and Cl− ELECTROLYTE BALANCE IN BROILERS 937

Table 1. Diet composition of the experimental diets

Diet

Item A B C D Ratio (corn starch:dextrose) 0.31 0.31 1.04 1.04 DEB1 (calculated) 108 219 108 219 Ingredient, g/kg Corn starch2 180.5 175.8 409.0 398.3 Dextrose3 640.0 623.2 411.5 400.7 Solka-Floc4 50 50 50 50 Soy oil 50 50 50 50 5 Vitamin-mineral premix 5.0 5.0 5.0 5.0 Downloaded from Monocalcium phosphate 19.0 19.0 19.0 19.0 NaHCO3 7.5 20.0 7.5 20.0 KCl 2.9 12.0 2.9 12.0 MgO 2.0 2.0 2.0 2.0 Choline chloride 2.5 3.0 2.5 3.0 Limestone 13.0 13.0 13.0 13.0 NaCl 0.0 2.0 0.0 2.0 http://ps.oxfordjournals.org/ K2CO3 2.6 0.0 2.6 0.0 Chromic oxide premix6 25 25 25 25 Total 1,000 1,000 1,000 1,000 Analyzed value, % (unless otherwise noted) Sodium 0.20 1.35 0.24 0.65 Potassium 0.23 0.57 0.20 0.63 Chloride 0.21 1.63 0.21 0.64 Calcium 0.75 0.85 0.81 0.85 Phosphorus 0.39 0.41 0.43 0.42 CP 0.93 0.99 0.87 0.81 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Gross energy, kcal/kg 3,575 3,447 3,636 3,550 DEB (based on analyzed values)7 87 273 96 263 1DEB = dietary electrolyte balance per milliequivalent value for dietary electrolyte balance. Calculated from dietary sodium, potassium, and chloride concentration (DEB, mEq/kg of diet = Na + K + Cl−, mEq/kg of diet); calculated values. 2Clabber Girl, Terre Haute, IN. 3Dextrose monohydrate, ADM Processing, Decatur, IL. 4Solka-Floc 40, International Fiber Corporation, Urbana, IL. 5Provided per kilogram of diet: iron, 71.6 mg; copper, 11.0 mg; manganese, 178.7 mg; zinc, 178.7 mg; iodine, 3.0 mg; selenium, 0.4 mg; vitamin A (retinyl acetate), 18,904.3 IU; vitamin D3 (cholecalciferol), 9,480.0 IU; vitamin E (dl-α-tocopheryl acetate), 63.0 IU; vitamin K activity, 6.4 mg; thiamine, 3.2 mg; riboflavin, 9.4 mg; pantothenic acid, 34.7 mg; niacin, 126.0 mg; pyridoxine, 4.7 mg; folic acid, 1.6 mg; biotin, 0.5 mg; vitamin B12, 35.4 μg; choline, 956.9 mg. 6Prepared by mixing 1 g of chromic oxide with 4 g of corn starch. 7Dietary electrolyte balance, in milliequivalents, was calculated as the mEq of Na + mEq of K – mEq of Cl−. The mEq of each of Na, K, and Cl− was determined using Y% × 10,000/molecular weight of Y, where Y is the percentage of either Na, K, or Cl− in the diets. was determined using Y% × 10,000/molecular weight concentration of DM, N, energy, or AA in ileal digesta of Y, where Y is the percentage of Na, K, or Cl− in the in milligrams per kilogram of DM. diets. Ileal EAA losses from birds on different dietary Data were analyzed using the GLM procedure of SAS treatments were calculated as milligrams of AA loss (SAS Institute Inc., 2006) appropriate for a completely per kilograms of DM intake (DMI) using the formula randomized design for a 2 × 2 factorial arrangement previously reported by Adedokun et al. (2007c), of treatments. The main effects of the ratio of energy sources (corn starch-to-dextrose ratio: 0.31 vs. 1.04) Ileal EAA losses (mg/kg of DMI) = (Cri/Cro) × (No). and DEB (108 vs. 219 mEq/kg of diet) were tested, as well as their interaction (ratio of energy sources × Apparent ileal digestibility was calculated by DEB). The interaction term was removed from the statistical analysis and the data were reanalyzed whenever the probability values for the interaction term were not Apparent ileal digestibility (%) = [1 – (Cri/Cro) significant. In this case, only the main effects (ratio × (No/Ni)] × 100, and DEB) are reported, otherwise simple effects are reported. Where necessary, means separation was by where Cri represents the concentration of chromium in Tukey’s test. Level of significance was set at 5%. the diet in grams per kilogram of DM; Cro represents the concentration of chromium in the ileal digesta in RESULTS grams per kilogram of DM output; Ni represents the concentration of DM, N, energy, or AA in the diet in Analyzed mineral composition of the experimental milligrams per kilogram of DM; and No represents the diets is close to what was formulated except for diet 938 ADEDOKUN AND APPLEGATE

Table 2. Analyzed amino acid composition of the experimental diets (on an as-is basis)

Diet

Item A B C D Ratio (corn starch:dextrose) 0.31 0.31 1.04 1.04 DEB1 (calculated values) 108 219 108 219 Indispensable amino acid, % Arg 0.01 0.00 0.01 0.01 His 0.00 0.00 0.00 0.00 Ile 0.01 0.01 0.01 0.01 Leu 0.01 0.01 0.01 0.01

Lys 0.01 0.01 0.01 0.01 Downloaded from Met 0.00 0.00 0.00 0.00 Phe 0.01 0.01 0.01 0.01 Thr 0.00 0.00 0.00 0.00 Val 0.01 0.01 0.01 0.01 Dispensable amino acid, % Ala 0.01 0.01 0.01 0.01 Asp 0.02 0.01 0.01 0.01 http://ps.oxfordjournals.org/ Cys 0.00 0.00 0.00 0.00 Glu 0.02 0.01 0.02 0.02 Gly 0.01 0.01 0.01 0.01 Pro 0.01 0.01 0.01 0.01 Ser 0.01 0.00 0.01 0.00 Tyr 0.01 0.01 0.01 0.01 Total amino acid, % 0.23 0.19 0.23 0.21 1DEB = dietary electrolyte balance per milliequivalent value for dietary electrolyte balance. Calculated from − dietary sodium, potassium, and chloride concentration (DEB, mEq/kg of diet = Na + K + Cl , mEq/kg of diet); at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 calculated values.

B, where the Na and Cl− values were higher than cal- the highest endogenous flow was found in Glu, Asp, culated (Table 1). Based on the analyzed values, the Val, and Leu. The ratio of each of the AA to that of DEB for diets A, B, C, and D were 87, 273, 96, and Lys is reported in Table 5. Methionine had the lowest 263, respectively (Table 1). Total AA contents of the ratio to Lys (a range of 0.31 to 0.33), whereas Glu had experimental diets ranged between 0.19 and 0.23% (Ta- the highest ratio (a range of 2.94 to 3.21). ble 2). Final BW and BW loss between birds fed the 2 corn starch-to-dextrose ratios and between birds fed 2 levels of DEB were not different (P > 0.05). Birds fed DISCUSSION the diet containing a high level of DEB consumed less (P < 0.05) feed than birds on the low-DEB diet. Birds Ileal EAA losses in poultry have been reported by on the high-corn starch-to-dextrose (1.04) diet showed several researchers for different ages of birds and using a tendency for higher (P = 0.08) feed intake (Table different methods. However, differences in methodol- 3). Consumption by birds on diets A and C were not ogy still exist within a particular method of estimation. different for the amount of K and Cl−, but lower (P < The NFD method is one of the methods that has been 0.05) than what was consumed by birds on diets B and extensively used (Siriwan et al., 1993; Ravindran et al., D. Birds on diet B consumed more (P < 0.05) Na com- 2004; Adedokun et al., 2007c; Golian et al., 2008; Kong pared with birds fed diets A, C, and D (Table 3). Birds and Adeola, 2013). However, NFD with different pro- on diet C consumed the least (P < 0.05) amount of Na portions or sources of energy and different levels of di- (Table 3). The high level of DEB resulted in higher (P etary fibers have been used in evaluating EAA losses in < 0.05) ileal endogenous nitrogen loss (Table 4). The poultry. The variations in the composition of the NFD high dietary dextrose level (corn starch-to-dextrose ra- fed to birds may explain some of the variations in EAA tio of 0.31) resulted in higher (P < 0.05) ileal DM and losses values obtained for the NFD (Adedokun et al., energy digestibility (Table 4). 2011). Recently, Kong and Adeola (2013) reported that Endogenous Arg, Ile, Leu, Phe, Val, Ala, Glu, Gly, the ratio of corn starch to dextrose in NFD significantly Pro, and Tyr secretion into the gut increased (P ≤ influenced ileal EAA losses in broilers, with birds fed 0.05) with an increasing level of DEB (Table 4). En- dextrose as the only source of energy in the diet hav- dogenous His, Lys, Thr, Asp, Cys, and Ser showed a ing significantly higher ileal EAA losses compared with tendency for increased (P ≤ 0.1) losses with a high birds fed corn starch as the only source of energy. level of DEB (Table 4). Mean ileal EAA losses for the In addition to understanding the effect of varying indispensable and dispensable AA for the low-DEB diet ratios of sources of energy in NFD on ileal EAA losses, were 81 and 82%, respectively, of that from the birds the influence of other factors, such as dietary fiber and fed the high-DEB diet. The 4 AA with the lowest en- type and DEB, is important (Angkanaporn et al., 1994; dogenous flow were Met, His, Tyr, and Cys, whereas Adedokun et al., 2011). The physiological importance ELECTROLYTE BALANCE IN BROILERS 939

, − − of dietary Na, K, and Cl intake in acid-base homeostasis in poultry has been studied and a DEB value of about 250 mEq/kg of diet has been reported to be adequate for broilers (Mongin, 1981). However, Ravindran dextrose et al. (2008) reported that broiler performance (BW DEB × starch: DEB × starch: gain), AMEn, and ileal N digestibility values decreased at a DEB value of 375 mEq/kg of diet. According to Ravindran et al. (2008), feed intake was P -value not affected by increasing DEB value and apparent il- Starch: Starch: dextrose eal AA digestibility was not affected with dietary DEB levels between 150 and 225 mEq of diet; however, ap- Downloaded from parent ileal AA digestibility decreased at DEB of 300

DEB and 375 mEq/kg of diet. As those authors did not determine ileal EAA losses in their study, 2 possibilities exist for that observation. First, an actual decrease in AA digestibility as a result of a decrease in AA diges- http://ps.oxfordjournals.org/ SEM tion or absorption; second, this decrease may be a result of an increase in EAA secretion into the gut as a result of an imbalance in acid-base homeostasis. Based 1.04 on the results from the current study, the decrease in apparent ileal AA digestibility could be attributed to the latter effect. 2.951.533.50 1.77 1.67 1.70 0.04 0.02 <0.0001 0.04 <0.0001 <0.0001 <0.0001 0.0001 <0.0001 <0.0001 <0.0001 <0.0001 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Starch:dextrose

0.31 To increase the reliability of ileal EAA values generated from poultry, it is important to evaluate how each of these dietary factors influences ileal EAA losses; once these have been established, it would be necessary to derive a consensus protocol for the NFD that would be 219 widely acceptable. Based on the need to further evalu- 2 ate the effect of different components of the NFD on

DEB ileal EAA losses, we conducted this study to evaluate

0.880.870.86 3.83 2.33 4.34 the effect of feeding NFD with 2 levels of corn starch- 108 to-dextrose ratios (0.31 and 1.04) and 2 DEB levels (108 and 219 mEq/kg of diet) on ileal EAA losses in 48-d-old broiler chickens.

b a b Due to the short duration of the current study, no 2.55 2.50 2.53 D difference in BW loss were observed between birds fed the different DEB levels or birds fed diets containing different corn starch-to-dextrose ratios; however, feed

c c c intake in birds fed diets with high DEB dropped by

0.98 0.83 0.87 5%. The drop in feed intake maybe related to the rela- C tively high levels of dietary Na, K, and Cl−. Although the final DEB values obtained from the analyzed val- Diet

a b a ues were different from the calculated values (108 and 219 vs. 91 and 268 mEq/kg of diet), they were close 5.12 2.15 6.15 B enough to give us the separation between the low and high levels as hypothesized. Dry matter digestibility and energy digestibility values were similar with birds d c c on high-dextrose diets having higher digestibility values 0.31 0.31 1.04 1.04 A

108 219 108 219 (79%) compared with birds on higher corn-starch diets (75%). Kong and Adeola (2013) reported an increase in ileal DM digestibility from 78 (corn starch only) to 88% (dextrose only). The increase in DM digestibility with an increasing level of dextrose is an indication of the ability of broilers to digest and absorb dextrose bet-

1 ter than corn starch. The similarity in DM and energy digestibility in the current study could be attributed to the fact that the bulk of each diet (at least 80%) was 3 Means within a row with different superscripts are significantly different ( P ≤ 0.05). different superscripts are significantly with different Means within a row 2

Each treatment consisted of 6 replicate cages 10 birds/replicate. treatment Each (DEB, mEq/kg of diet = Na + K Cl for dietary electrolyte balance. Calculated from sodium, concentration potassium, value and chloride Dietary electrolyte balance per milliequivalent Ratio of corn starch:dextrose. corn starch and dextrose. a–d 1 2 3 Sodium 0.78 Final BW Final gain BW intake Feed g/bird5-d mineral intake, 2,723 −149 399 2,756 −124 378 2,748 −133 412 2,763 −122 396 2,736 −141 2,760 −123 406 2,739 −137 387 2,756 −128 49.1 389 10.7 0.73 404 0.26 0.82 5.9 0.55 0.04 0.90 0.65 0.08 0.79 DEB g/birdPerformance, Ratio Performance and mineral intake of 48-d-old broilers fed nitrogen-free diets with different levels of dietary electrolyte balance (DEB) and different proportion of corn starch proportion of corn starch (DEB) and different of dietary electrolyte balance levels of 48-d-old broilers fed nitrogen-free diets with different and mineral intake 3. Performance Table and dextrose Item mEq/kg of diet); calculated values. Potassium Chloride Calcium Phosphorus 0.90 0.85 1.55 3.02 1.53 3.22 1.77 3.33 1.65 3.35 1.66 3.18 1.59 3.28 1.54 3.12 1.71 3.34 0.02 0.05 0.080 0.12 0.0002 0.003 0.18 0.18 940 ADEDOKUN AND APPLEGATE

Table 4. Ileal DM and energy digestibility and ileal endogenous nitrogen and amino acid losses in 48-d-old broilers fed nitrogen-free diets with different levels of dietary electrolyte balance and different proportion of corn starch and dextrose1

DEB2 Ratio3 P-value

Item 108 219 0.31 1.04 SEM DEB Ratio Ileal digestibility, % DM 76.9 76.6 78.6 74.9 0.89 0.85 0.01 Energy 76.9 76.7 78.7 74.9 0.89 0.85 0.01 Nitrogen 2,603 3,222 2,861 2,964 172.7 0.02 0.68 Indispensable amino acid, mg/ kg of DM intake

Arg 568 714 649 633 48.1 0.04 0.82 Downloaded from His 251 297 262 286 19.4 0.11 0.40 Ile 590 716 651 655 43.8 0.05 0.96 Leu 927 1,163 1,041 1,048 76.1 0.04 0.95 Lys 482 625 554 553 59.6 0.10 0.99 Met 158 193 176 175 16.6 0.16 0.97 Phe 536 663 603 596 41.8 0.04 0.90 Thr 777 931 837 871 54.4 0.06 0.67 http://ps.oxfordjournals.org/ Val 969 1,200 1,063 1,105 75.1 0.04 0.70 Mean 584 722 649 658 Dispensable amino acid, mg/ kg of DM intake Ala 736 923 800 859 61.6 0.04 0.50 Asp 1,151 1,367 1,244 1,274 84.0 0.08 0.80 Cys 471 569 515 525 36.1 0.07 0.84 Glu 1,507 1,870 1,652 1,725 123.9 0.05 0.68 Gly 823 1,016 889 950 63.6 0.04 0.51 Pro 784 966 865 885 62.6 0.05 0.83 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Ser 784 940 868 856 65.5 0.11 0.90 Tyr 397 481 438 440 28.6 0.05 0.96 Mean 832 1,016 909 939 Total amino acid, mg/kg of 13,220 16,114 14,448 14,886 936.0 0.04 0.74 DM intake 1Each treatment consisted of 6 replicate cages of 10 birds/replicate. 2DEB = dietary electrolyte balance per milliequivalent value for dietary electrolyte balance (DEB, mEq/kg of diet = Na + K + Cl−, mEq/kg of diet); calculated values. 3Ratio of corn starch:dextrose.

In the current study, no effect of the ratio of corn The order of EAA losses for all treatments in the starch and dextrose on EAA losses was observed. This current study is similar irrespective of the DEB values. mirrors what was reported by Kong and Adeola (2013) In general, the first 7 AA with the largest flow were in 26-d-old broilers, where dietary dextrose effect on Glu > Asp > Val > Leu > Gly > Pro > Ser. This or- EAA losses was significant only when it completely re- der, especially the first 2 AA (Glu and Asp), is similar placed corn starch in the NFD. Those authors observed to what has been reported for 21-d-old broilers fed an no effect of dextrose at ratios of 0:1, 1:2, or 2:1 dex- NFD with mEq of 219 by Adedokun et al. (2007c) and trose-to-corn starch. At low DEB (108) and irrespective 26-d-old broilers fed an NFD containing 283 and 566 of the ratio of corn starch to dextrose (0.31 or 1.04), g of corn starch and dextrose/kg of diet with an mEq the total EAA loss values from the current study are of 115, respectively (Kong and Adeola, 2013). Further- similar to what Kong and Adeola (2013) reported in more, data from 37-d-old broiler chickens fed an NFD 26-d-old broiler chickens fed diets containing a mixture showed similar trends (Ravindran et al., 2004). How- of corn starch and dextrose (12,527 vs. 13,220 mg/kg ever, in the data reported by Adedokun et al. (2007c) of DMI). and Ravindran et al. (2004), Thr was the third most Unlike the effect of the ratio of sources of energy, abundant AA, whereas Cys was third most abundant high DEB values increased the amount of AA secretion for Kong and Adeola (2013). Glutamine, Asp, Ser, Leu, from endogenous origin significantly; although intake and Pro were among the first 7 most abundant AA of dropped by about 5%, total EAA losses increased by endogenous origin in the current study and the studies about 13% in birds on a high DEB diet. It is interest- of Ravindran et al. (2004), Adedokun et al. (2007c), ing to note that despite the fact that a high DEB level and Kong and Adeola (2013). It is interesting to note resulted in increased secretion of EAA, it did not influ- that Thr was the third and fifth most abundant EAA ence the DM digestibility (DMD), but the higher pro- in 21- and 26-d-old birds (Adedokun et al., 2007c; Kong portion of dextrose in the NFD resulted in an increase and Adeola, 2013), but at d 37 and 48 they were the in DMD (about 4 percentage units). This observation third and eighth most abundant EAA, respectively. is similar to what Kong and Adeola (2013) reported for The observed difference in the order of EAA may be 26-d-old broiler fed different types of NFD. due to the fact that different dietary components (en- ELECTROLYTE BALANCE IN BROILERS 941

Table 5. Ratio of individual amino acids to lysine of ileal endogenous amino acid losses in broilers fed nitrogen-free diets with different levels of dietary electrolyte balance and different proportion of corn starch and dextrose1

Current study diet (d 48) Ravindran Adedokun Kong and et al., 20042 et al., 2007a3 Adeola, 20134 Item A B C D (d 37) (d 21) (d 26) Ratio5 0.31 0.31 1.04 1.04 82% dextrose 0.26 0.50 DEB6 108 219 108 219 250 219 115 Indispensable amino acid Arg 1.14 1.19 1.21 1.09 1.34 0.93 0.91 His 0.51 0.45 0.53 0.51 0.76 0.40 0.38 Ile 1.20 1.16 1.24 1.13 1.37 0.90 0.83

Leu 1.88 1.88 1.96 1.84 2.10 1.39 1.36 Downloaded from Lys 1.00 1.00 1.00 1.00 1.00 1.00 1.00 Met 0.33 0.31 0.33 0.31 0.48 0.28 0.27 Phe 1.10 1.08 1.13 1.04 1.37 0.85 0.79 Thr 1.59 1.45 1.63 1.52 2.45 1.51 1.13 Val 1.96 1.89 2.06 1.95 2.00 1.18 1.00 Dispensable amino acid Ala 1.48 1.41 1.57 1.54 1.40 0.98 0.85 http://ps.oxfordjournals.org/ Asp 2.38 2.15 2.40 2.22 2.90 1.88 1.58 Cys 0.96 0.91 1.00 0.91 1.08 0.75 1.35 Glu 3.04 2.94 3.21 3.04 3.46 2.32 2.03 Gly 1.64 1.58 1.77 1.67 2.43 1.13 0.97 Pro 1.56 1.56 1.69 1.52 NA7 1.33 0.97 Ser 1.57 1.56 1.68 1.44 2.03 1.44 1.05 Tyr 0.82 0.77 0.83 0.77 1.21 0.69 0.61 1Values are based on the ratio of each amino acid and the corresponding value for endogenous Lys losses within each study.

2Data used were from 21-d-old broiler chickens fed the N-free diet. Corn starch was not used in the N-free diet. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 3Data used were from 21-d-old broiler chickens fed the N-free diet. 4Data used were for birds fed diet containing 283 and 566 g of corn starch and dextrose/kg of diet, respectively (diet 3). 5Ratio of corn starch:dextrose. 6DEB = dietary electrolyte balance per milliequivalent value for dietary electrolyte balance (DEB, mEq/kg of diet = Na + K + Cl−, mEq/kg of diet); calculated values. 7NA = not available. ergy source, fiber source and level, and DEB) may have REFERENCES influenced sources of EAA, such as digestive secretions, Adedokun, S. A., O. Adeola, C. M. Parsons, M. S. Lilburn, and T. sloughed epithelial cells, and intestinal mucous, differ- J. Applegate. 2011. Review: Factors affecting endogenous amino ently. The relative ratio of individual AA to that of Lys acid flow in chickens and the need for consistency in methodol- have been postulated by Kong and Adeola (2013) to be ogy. Poult. Sci. 90:1737–1748. stable across studies conducted using similar methodol- Adedokun, S. A., K. M. Ajuwon, L. F. Romero, and O. Adeola. 2012. Ileal endogenous amino acid loss: Response of broiler chick- ogy. Similar relationships between individual AA and ens to fiber and mild coccidial vaccine challenge. Poult. Sci. Lys are observed in this study irrespective of the DEB 91:899–907. or ratio of corn starch to dextrose. Adedokun, S. A., M. S. Lilburn, C. M. Parsons, O. Adeola, and T. J. Values obtained from this relationship (AA-to-Lys Applegate. 2007a. Effect of age and method on ileal endogenous amino acid flow in turkey poults. Poult. Sci. 86:1948–1954. ratio) in the current study and that from Ravindran Adedokun, S. A., C. M. Parsons, M. S. Lilburn, O. Adeola, and T. et al. (2004) are similar compared with values from J. Applegate. 2007b. Comparison of ileal endogenous amino acid 21-d-old broilers fed an NFD (Adedokun et al., 2007c) flows in broiler chicks and turkey poults. Poult. Sci. 86:1682– and 26-d-old broilers (Kong and Adeola, 2013). The 1689. Adedokun, S. A., C. M. Parsons, M. S. Lilburn, O. Adeola, and T. J. reason for the similarity in AA-to-Lys ratio between Applegate. 2007c. Endogenous amino acid flow in broiler chicks the current study and that of Ravindran et al. (2004) is affected by the age of birds and method of estimation. Poult. may be related to the fact that birds in these studies Sci. 86:2590–2597. were older (48 and 37 d, respectively) compared with Adedokun, S. A., P. Utterback, C. M. Parsons, O. Adeola, M. S. Lilburn, and T. J. Applegate. 2009. Comparison of endogenous 21- (Adedokun et al., 2007c) and 26-d-old (Kong and amino acid flow in broilers, laying hens, and cecectomized roost- Adeola, 2013) birds. ers. Br. Poult. Sci. 50:359–365. Data from the current study showed that changes in Angkanaporn, K., M. Choct, W. L. Bryden, E. F. Annison, and G. Annison. 1994. Effect of wheat pentosans on endogenous amino DEB could influence EAA losses in the ileum of broil- acid losses in chickens. J. Sci. Food Agric. 66:399–404. ers, which underscores the importance of a consensus AOAC International. 2000. Official Methods of Analysis. 17th ed. NFD in estimating basal EAA losses in poultry. Fur- AOAC Int., Arlington, VA. thermore, apparent ileal AA digestibility coefficients AOAC International. 2006. Official Methods of Analysis. 18th ed. AOAC Int., Gaithersburg, MD. of feeds or feed ingredients may be influenced by high Cowieson, A. J., T. Acamovic, and M. R. Bedford. 2004. The effects DEB as a result of its effects on ileal EAA losses into of phytase and phytic acid on the loss of endogenous amino acids the gut. and minerals from broiler chickens. Br. Poult. Sci. 45:101–108. 942 ADEDOKUN AND APPLEGATE

Cowieson, A. J., and V. Ravindran. 2007. Effect of phytic acid and NRC. 1994. Nutrient Requirements of Poultry. 9th rev. ed. Natl. microbial phytase on the flow and amino acid composition of en- Acad. Press, Washington, DC. dogenous protein at the terminal ileum of growing broiler chick- Onyango, E. M., E. K. Asem, and O. Adeola. 2009. Phytic acid ens. Br. J. Nutr. 98:745–752. increases mucin and endogenous amino acid losses from the gas- Golian, A., W. Guenter, D. Hoehler, H. Jahanian, and C. M. trointestinal tract of chickens. Br. J. Nutr. 101:836–842. Nyachoti. 2008. Comparison of various methods for endogenous Patience, J. F. 1990. A review of the role of acid-base balance in ileal amino acid flow determination in broiler chickens. Poult. amino acid nutrition. J. Anim. Sci. 68:398–408. Sci. 87:706–712. Ravindran, V., and W. H. Hendriks. 2004. Endogenous amino acid Johnson, R. J., and H. Karunajeewa. 1985. The effects of dietary flows at the terminal ileum of broilers, layers, and adult roosters. minerals and electrolytes on the growth and physiology of the Anim. Sci. 79:265–271. young chick. J. Nutr. 115:1680–1690. Ravindran, V., L. I. Hew, G. Ravindran, and W. L. Bryden. 2004. Kim, B. G., and M. D. Lindemann. 2007. A new spreadsheet method Endogenous amino acid flow in the avian ileum: Quantification

for the experimental animal allotment. J. Anim. Sci. 85(Suppl. using three techniques. Br. J. Nutr. 92:217–223. Downloaded from 2):218. Ravindran, V., A. J. Cowieson, and P. H. Selle. 2008. Influence Kluth, H., and M. Rodehutscord. 2009. Effect of inclusion of cellu- of dietary electrolyte balance and microbial phytase on growth lose in the diet on the inevitable endogenous amino acid losses in performance, nutrient utilization, and excreta quality of broiler the ileum of broiler chickens. Poult. Sci. 88:1199–1205. chickens. Poult. Sci. 87:677–688. Kong, C., and O. Adeola. 2013. Ileal endogenous amino acid flow SAS Institute. 2006. SAS/STAT User’s guide. Release 9.1. SAS Inst. response to nitrogen-free diets with differing ratios of corn starch Inc., Cary, NC. to dextrose in broiler chickens. Poult. Sci. 92:1276–1282. Siriwan, P., W. L. Bryden, Y. Mollah, and E. F. Annison. 1993. http://ps.oxfordjournals.org/ Lemme, A., V. Ravindran, and W. L. Bryden. 2004. Ileal digestibil- Measurement of endogenous amino acid losses in poultry. Br. ity of amino acid in feed ingredients for broilers. World’s Poult. Poult. Sci. 34:939–949. Sci. J. 60:423–437. Mongin, P. 1981. Recent advances in dietary cation-anion balance: Applications in poultry. Proc. Nutr. Soc. 40:285–294. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 MOLECULAR, CELLULAR, AND DEVELOPMENTAL BIOLOGY

Bone characteristics and femoral strength in commercial toms: The effect of protein and energy restriction

R. C. Van Wyhe ,* P. Regmi ,* B. J. Powell ,† R. C. Haut ,† M. W. Orth ,* and D. M. Karcher *1

* Department of Animal Sciences, and † Orthopaedic Biomechanics Laboratories, Michigan State University, East Lansing 48824 Downloaded from ABSTRACT Selection for rapid growth in turkeys has mur length was longer in the 60% NRC birds at 5 and resulted in skeletal problems such as femoral fractures. 10 kg of BW compared with control (P < 0.05); this Slowing growth rate has improved bone structure, but significance was lost by the time birds reached 16 kg the effect on mechanical properties of the bone is un- of BW. At 5 and 10 kg of BW, ash content was higher clear. The current study’s hypothesis was that slowing in the control birds than in the 60% NRC birds (P < http://ps.oxfordjournals.org/ the growth of turkeys by reducing energy and CP in 0.05). At 16 kg of BW, the 60% NRC birds had the the diet would result in increased femur integrity. Com- highest femur ash (P < 0.05). The mechanical test- mercial turkeys were fed 1 of 3 diets: control with 100% ing parameters were failure torque, shear strength, and of NRC energy and CP levels, as well as a diet feed- shear modulus of the bones. The 60% diet produced ing 80 or 60% of NRC energy and CP levels. All other the highest failure torque (P < 0.05), at 16 kg of BW nutrients met or exceeded NRC requirements. Control and onward. The shear strength was greater (P = 0.01)

birds were grown to 20 wk of age, whereas the 80 and once the birds reached 5 kg of BW for the 60% diet at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 60% NRC birds were sampled when BW matched that than other diets. In conclusion, reducing the energy of control birds at wk 4, 8, 12, 16, and 20. Both femurs and protein in the diet to 60% of NRC recommenda- were extracted, with one being measured and ashed tions, thus slowing growth, improved bone strength, as and the other twisted to failure to evaluate mechanical measured by failure torque, and bone quality, as mea- properties. Total bone length, diameter, cortical thick- sured by shear strength, without altering bone length

ness, and cortical density were measured. The total fe- or ash content by the time birds reached market weight. Key words: femur , turkey , break , ash , mechanics 2014 Poultry Science 93 :943–952 http://dx.doi.org/ 10.3382/ps.2013-03604

INTRODUCTION pigs (Roberson et al., 2005; Shaw et al., 2006; Her- trampf et al., 2009). Thus, at a time when maximal The growth rate of the commercial turkey has ap- support is needed, the bones are still trying to remodel proximately doubled over the last 45 yr (Havenstein to accommodate the weight, leaving it vulnerable to et al., 2007). In toms, one problem likely related to injury. Others have suggested similar issues with the increased size is the occurrence of femoral fractures. increasing BW of skeletally immature broilers (Skinner These fractures are a major concern to turkey produc- and Waldroup, 1995). ers, as the fractured femur may lacerate the femoral ar- Growth rate of fast-growing meat birds has been re- tery and result in death. Larger body mass in toms de- ported to be mainly responsible for the increased po- mands a concomitant increase in the size and strength rosity and reduced stiffness of bones (Williams et al., of the skeletal structure for support. The bones respond 2004). Thus, one way to reduce skeletal problems might to increased BW by continually increasing their rates of be to restrict the growth of turkey toms and allow more remodeling (Turner and Robling, 2004; Warden, 2006). time for the bones to mature to accommodate the BW. The rates of systemic bone resorption in turkeys ap- Feed-restricted 33- to 35-wk-old breeder tom femurs proaching market weight, as measured by serum pyr- have been shown to have increased cross-sectional di- idinoline (PYD) concentrations, are much higher than ameter, decreased cortical thickness, and decreased what are seen in other species with accelerated bone cross-sectional cortical area (Crespo et al., 2000). The turnover, such as ovariectomized rats and market age smaller cortical thickness and cross-sectional area is likely a result of a smaller BW. However, when the © 2014 Poultry Science Association Inc. bones were mechanically tested and data was adjusted Received September 5, 2013. Accepted December 16, 2013. for BW, the measurements translated into stiffer and 1 Corresponding author: [email protected] stronger femurs in the feed-restricted turkeys (Crespo

943 944 VAN WYHE ET AL. et al., 2000). It is unclear at what point these chang- ment. Toms were raised to market weight (20 kg). Feed es might start to take place and if a difference would consumption and weight gain were recorded at each be noted in commercial turkeys. Turner and Lilburn phase (every 4 wk) and average feed conversion ratio (1992) restricted growth in toms by feeding a control was calculated. Mortality was monitored continuously and a protein-deficient diet for the first 8 wk. The tib- throughout the trial. Tom turkeys (80 and 60% NRC ia and femur length was reduced in protein-deficient diets) were weight matched at the end of each phase (4, toms, but the effect was not observed after 12 to 14 8, 12, 16, and 20 wk; Table 3). At each sampling point, wk. The width was, however, permanently reduced in 2 average-weight toms per pen were selected for a total both tibia and femur compared with the control group.. of 8 toms per diet and strain combination. The sampled Whereas the bones were smaller, no difference in bone birds had both femurs collected for analysis. The final ash was observed between the 2 groups. It is unclear collection of femur samples (n = 30 per treatment com- Downloaded from whether changes in size altered bone quality as the bination) occurred at slaughter, when toms achieved mechanical properties were not evaluated. In turkeys, the 20-kg market weight. most longitudinal bone growth has occurred by 10 wk of age (Lilburn et al., 2006), whereas mineralization

PYD http://ps.oxfordjournals.org/ has occurred at a fairly constant rate throughout the bird’s life, reaching a plateau as early as 5 wk (Apple- Brachial vein blood samples were collected and se- gate et al., 2006). With mineralization rates plateau- rum was separated and frozen at −20°C before analysis. ing early in life, slowing the growth of the birds early Serum PYD concentration was quantified using com- may allow the skeletal system to better support BW mercially available ELISA tests (Quidel Corporation, and prevent femoral fractures that are normally seen as San Diego, CA). Sample filtrate was diluted 8-fold in birds approach market weight. Therefore, the objective assay buffer for the analysis. of the present study was to evaluate the structural and at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 functional quality of appendicular bones in normal and feed-restricted turkeys by determining their mechani- Peripheral Quantitative CT Imaging cal properties, the alteration of strength and stiffness and Bone Ash of the bone, and correlating rate of growth in birds to skeletal turnover. Quantitative computed tomography scans of the left femurs were performed using a GE BrightSpeed scan- ner (General Electric Healthcare, Princeton, NJ). To- MATERIALS AND METHODS tal bone length was divided by 4 to obtain proximal The experimental protocol was approved by the (one-fourth), medial, and distal (three-fourths) regions. Michigan State University Institutional Animal Care Mimics software (Materialise, Plymouth, MI) was used and Use Committee. to analyze the resulting 1-mm cross-sections for diameter, cortical bone thickness, and cortical density in each of the 3 regions. Following tomographic scanning, Birds, Housing, and Diets the left femur was prepared for ashing. The femur was The experiment consisted of 2 turkey strains (A and ether extracted, dried, and ashed according to methods B) and 3 different diets with 4 replicate pens per diet described by Zhong et al. (2012). and strain combination for a total of 600 toms. Three hundred one-day-old Hybrid toms and 300 one-day-old Mechanical Testing Nicholas toms were placed at day of hatch in an environmentally controlled house. All toms were started on The right femurs were allowed to thaw at 4°C for 10 a control starter for the first 7 d. On d 7, toms were h before preparation for mechanical tests. The bones weighed and reassigned to pens to ensure consistent were scraped with a scalpel at the proximal and distal average pen weights. Twenty-five poults were placed in ends to prevent slippage when potted in an air-harden- each 8- × 10-ft pen with 12 pens of each strain. The ing epoxy (Fiber Strand, Martin Senour Corp., Cleve- industry control diet was formulated to meet or exceed land, OH). Each femur was then individually secured the NRC (1994) recommendations. Experimental diets in a custom-built alignment fixture with a tri-axis bone were formulated to limit amino acids and energy while clamp. The center of each bone was aligned vertically maintaining other nutrients at or above NRC (1994) and horizontally with respect to the potting cups with requirements. The 80% diet was formulated so that the use of a laser-centering diode. The potting cups protein and energy met 80% of the recommend NRC were separated by 1 of 3 spacers having lengths of 40.6, (1994) value, whereas the 60% diet met 60% of the 60.9, and 81.7 mm (L1, L2, and L3, respectively). The protein and energy requirement (Table 1). The 80 and spacers accommodated the increase in femur length 60% NRC diets contained extra fiber as a filler to ob- with age. These distances were set based on an aver- tain low protein and energy values. Diets were analyzed age length-to-diameter ratio of approximately 4 for the to confirm nutritive values (Table 2). Feed and water femurs in each age group. The 4-wk femurs were tested were provided ad libitum during the entire experi- with L1, the 8-wk with L2, and the 12-, 16-, and 20- Table 1. Formulated turkey diets containing 100, 80, or 60% of NRC (1994) energy and protein requirements

Control1 80%1 60%1 Ingredient (%) Starter GP2-1 GP-2 GP-3 Fin3 Starter GP-1 GP-2 GP-3 Fin Starter GP-1 GP-2 GP-3 Fin Maize 43.83 42.80 55.61 56.31 69.90 17.89 32.11 35.00 45.00 40.00 8.35 20.92 23.00 30.00 30.00 Soybean meal 48 48.00 47.37 34.13 31.74 19.96 37.00 31.37 29.80 20.00 13.00 26.03 22.24 19.00 11.61 8.00 Wheat middlings — — — — — 39.26 25.00 25.00 25.00 30.00 40.00 25.00 30.00 30.00 30.00 Soy oil 2.19 4.54 5.00 2.51 1.20 0.54 1.00 1.80 0.91 0.72 0.60 1.00 1.13 0.44 0.97 White grease — — — 5.00 5.00 — — — 1.00 5.00 — — — 1.00 2.00 dl-Met 0.32 0.22 0.19 0.06 0.07 0.11 0.13 0.01 0.02 — 0.07 0.09 0.02 0.02 — Lys HCl 0.27 0.15 0.34 0.01 0.14 — 0.18 — 0.02 0.02 — 0.14 0.05 0.06 0.03 Thr 0.05 0.01 0.05 0.05 0.06 — 0.01 — 0.01 — — 0.01 — 0.03 — Limestone 1.42 1.47 1.39 1.28 1.10 1.52 1.54 1.44 1.36 1.15 1.50 1.52 1.44 1.35 1.13 Dicalcium P 3.08 2.58 2.43 2.20 1.77 2.87 2.53 2.28 2.10 1.63 3.02 2.67 2.39 2.19 1.73 Salt 0.51 0.46 0.46 0.43 0.41 0.48 0.44 0.44 0.41 0.39 0.48 0.45 0.44 0.42 0.40

Copper sulfate 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 FEED RESTRICTIONONTURKEYFEMURS Coban 90 0.04 — — — — 0.04 — — — — 0.04 ———— Celite — — — — — — 5.28 3.83 3.78 7.69 19.60 25.55 22.13 22.48 25.34 Vitamin-mineral premix 0.25 0.35 0.35 0.35 0.35 0.25 0.35 0.35 0.35 0.35 0.25 0.35 0.35 0.35 0.35 Calculated ME4 2,850 3,000 3,150 3,270 3,325 2,280 2,399 2,520 2,616 2,659 1,710 1,800 1,890 1,961 1,994 1Control = birds fed a diet containing 100% of NRC (1994) protein and energy requirements; 80% = birds fed a diet containing 80% of NRC (1994) protein and energy requirements; 60% = birds fed a diet containing 60% of NRC (1994) protein and energy requirements. 2GP = grower phase of the diet. 3Fin = finisher phase of the diet. 4kcal/kg.

Table 2. Turkey diets analyzed nutrient values containing 100, 80, or 60% of NRC (1994) energy and protein requirements

Control1 80%1 60%1 Analyzed value Starter GP2-1 GP-2 GP-3 Fin3 Starter GP-1 GP-2 GP-3 Fin Starter GP-1 GP-2 GP-3 Fin DM (%) 91.59 90.44 91.05 91.81 91.16 91.35 91.84 91.42 90.64 91.87 92.44 93.18 93.34 na4 91.46 Gross energy5 3,890 4,133 4,060 4,417 4,260 3,911 3,709 3,831 3,781 3,687 3,262 2,978 3,059 na 3,710 CP (%) 25.18 26.56 21.55 20.77 14.15 23.03 19.88 19.50 16.44 11.64 15.66 14.38 14.87 na 9.33 Ca (%) 24.20 17.29 16.37 14.00 13.12 23.49 19.64 16.27 15.92 12.62 20.28 21.96 15.06 na 11.47 Total P (%) 8.85 7.22 6.47 5.76 5.15 8.24 8.52 7.83 7.08 5.03 8.34 7.81 9.25 na 4.45 Met (%) 0.66 0.56 0.49 0.37 0.30 0.43 0.40 0.31 0.29 0.20 0.30 0.30 0.24 na 0.18 Cys (%) 0.36 0.38 0.32 0.31 0.23 0.35 0.32 0.31 0.28 0.22 0.28 0.26 0.25 na 0.18 Lys (%) 1.73 1.70 1.62 1.23 0.95 1.32 1.28 1.14 0.93 0.69 0.98 1.06 0.95 na 0.62 Ile (%) 1.09 1.15 0.93 0.94 0.67 0.98 0.84 0.84 0.68 0.52 0.77 0.74 0.68 na 0.44 Arg (%) 1.68 1.74 1.43 1.41 0.98 1.53 1.39 1.35 1.14 0.81 1.13 1.11 1.12 na 0.65 Thr (%) 0.98 0.97 0.85 0.83 0.61 0.84 0.78 0.75 0.65 0.48 0.60 0.60 0.59 na 0.40 Trp (%) 0.32 0.34 0.26 0.24 0.22 0.28 0.24 0.25 0.21 0.16 0.22 0.20 0.21 na 0.13 Phe (%) 1.21 1.26 1.05 1.05 0.74 1.10 0.96 0.96 0.80 0.62 0.83 0.79 0.75 na 0.50 1Control = birds fed a diet containing 100% of NRC (1994) protein and energy requirements; 80% = birds fed a diet containing 80% of NRC (1994) protein and energy requirements; 60% = birds fed a diet containing 60% of NRC (1994) protein and energy requirements. 2GP = grower phase of the diet. 3Fin = finisher phase of the diet. 4na = missing data. 945

5kcal/kg.

Table 3. The sampling age for turkeys of equal BW fed a control or an energy- and protein-restricted diet

Age (d) Treatment1 (kg of BW) 1 5 10 16 20 Control 28 56 84 112 140 80% 28 56 88 117 140 60% 35 78 111 155 178 1Control = birds fed a diet containing 100% of NRC (1994) protein and energy requirements; 80% = birds fed a diet containing 80% of NRC (1994) protein and energy requirements; 60% = birds fed a diet containing 60% of

NRC (1994) protein and energy requirements. Downloaded from wk with L3. Once the bone was in place, 4 set screws where a = the major inner radius; and b = the mi- were threaded into each potting cup to accommodate nor inner radius (as adapted from Young and Budynas, fixation of the epoxy to the test fixture. Phosphate- 2002). buffered saline was routinely sprayed onto the exposed Bone quality before failure was quantified by com- http://ps.oxfordjournals.org/ bone to prevent drying. After both ends of the speci- puting its shear modulus using the expression men were potted, maximum and minimum diameter measurements were taken at proximal, distal, and cen- TL× G = f , [3] tral locations over the test length. Proximal and distal Kavg ×ϕ measurements were taken at the interface of the bone with the epoxy. The bone was then wrapped in PBS- where L = the test length; φ = the failure angle of soaked gauze and placed in a 4°C refrigerator for 24 h at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 before testing. rotation; and Kavg = an average value of the structural A servo-hydraulic testing machine (Model 1331, In- torsion constant K, stron Corp., Canton, OH) was adapted with a custom 33 4 πγAB ×−()1 axial-to-torsion fixture to produce rotation control in = . K 22 [4] the experiments. A torque load cell (Model 5330–1200, AB+ Interface Inc., Scottsdale, AZ) was attached to one rotation cup and used to capture force-time data. Angu- K was calculated using the average of the major and lar rotation data were recorded with a rotary encoder minor outer radii at the proximal, distal, and central fastened to a drive gear (Model BHW 16.05A72000- locations. An average K was used to reduce error due BP-A, Baumer Electric Corp., Southington, CT). Ex- to potential lengthwise variations along the femur (Lev- perimental data was sampled at 1,000 Hz. enston et al., 1994). Each specimen was secured into the fixture with screws into preexisting threads in the epoxy. The PBS gauzes were removed and the specimens were twisted to Statistical Analysis failure at a speed of 180°/sec. After fracture, the bone Performance parameters were analyzed using the was removed from the fixture and 3 cortical thickness PROC MIXED analysis of SAS (SAS Institute Inc., measurements were taken around the circumference of Cary, NC) with the LSMeans procedure. All compari- the fracture site with a digital caliper (IP66, Mitutoyo sons between treatments were done on a BW basis. All America). bone measurements, such as length, section thickness, The structural strength of the bones was based on PYD, and ash data, were analyzed using the repeated failure torque as measured by the peak torque gener- measures statement. For repeated measurements, the ated in each experiment. The cross-sectional geometry model included the fixed effect of diet and pen, random of the femur was assumed to be a uniform thickness, effects of week, the interaction between diet and week, hollow ellipse. To measure the structural quality of the and the residual error. Differences between means were bone as a material, its shear strength was calculated tested using the pdiff option of the lsmeans statement using with significance accepted at P < 0.05. If a main effect of strain was observed, the 2 strains were analyzed 2×T τ= f , independently. 24 [1] πγAB ×−()1 RESULTS where Tf = the failure torque; A = the major outer radius; B = the minor outer radius; and Performance The feed conversion ratio was higher for turkeys on a b 2γ =+, [2] the 60 and 80% NRC diets compared with control birds, A B FEED RESTRICTION ON TURKEY FEMURS 947 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 1. The effect of dietary energy and protein restriction on feed intake and feed conversion of commercial toms. Means within same BW lacking a common superscript (a–c) differ significantly (P < 0.05). Control = birds fed a diet containing 100% of NRC (1994) protein and energy requirements; 80% = birds fed a diet containing 80% of NRC (1994) protein and energy requirements; 60% = birds fed a diet containing 60% of NRC (1994) protein and energy requirements. FCR = feed conversion ratio. with the largest cumulative feed conversion ratio (4.31) Average cortical thickness was largest in the 60% for the 60% group (P < 0.05; Figure 1). Feed intake NRC group at 16 kg of BW, but smallest at 20 kg of was higher (P < 0.05) for birds on the 60% NRC diet BW (P < 0.05; Table 4). Strain A cortex was thickest compared with the 80% NRC diet and control, except at mid femur shaft for control birds at 16 kg of BW, for the period from 16 to 20 kg of BW. The 60% group whereas strain B was thickest in the 60% NRC group took an additional 5 wk compared with the other treat- at 10, 16, and 20 kg of BW (P < 0.05; Table 5). The ments to reach the final BW of 20 kg (Table 1).

Bone Physical Parameters and Quality Though the current study looked at factors that could contribute to bone weakness, no spiral fractures were recorded. The femur length increased with age and BW for each of the treatment groups. Birds fed the 60% diet had consistently longer femurs once the birds reached 5 kg of BW compared with control birds (P < 0.05; Figure 2). However, diameter along the proximal and distal femur shaft was widest in control birds at 20 kg and the 60% diet resulted in the narrowest distal diameter at the end of the trial (Table 4). A difference was noted between strains for the femoral diameter, cortical thickness, and cortical bone density only at mid diaphyseal shaft of the femur (Table 5). Strain A birds fed 80 and 60% NRC diets had a smaller medial Figure 2. The effect of dietary energy and protein restriction on diameter compared with the control group at 5 and femur length in commercial turkeys. Control = birds fed a diet con- again at 16 kg of BW. For strain B, control birds had taining 100% of NRC (1994) protein and energy requirements; 80% widest medial diameter at 10 and 20 kg of BW (P < = birds fed a diet containing 80% of NRC (1994) protein and energy requirements; 60% = birds fed a diet containing 60% of NRC (1994) 0.05), whereas the 80% diet resulted in a similar diam- protein and energy requirements; significant differences (*) were de- eter to the control and 60% group at 16 kg of BW. fined as P < 0.05. 948 VAN WYHE ET AL.

Table 4. Effect of dietary energy and protein restriction at a common BW on turkey proximal and distal femur parameters

BW (kg)

Item1 1 5 10 16 20 Proximal diameter (mm) Control 12.88 ± 0.39 20.60 ± 0.39 26.81 ± 0.39 30.51 ± 0.39 32.16 ± 0.26a 80% 12.60 ± 0.39 20.10 ± 0.39 27.00 ± 0.39 29.75 ± 0.39 31.69 ± 0.26b 60% 12.29 ± 0.39 21.05 ± 0.39 27.71 ± 0.39 29.56 ± 0.39 30.64 ± 0.26b Proximal cortical thickness (mm) Control 1.52 ± 0.07 1.71 ± 0.07 1.81 ± 0.07 1.73 ± 0.07b 1.88 ± 0.04a b ab 80% 1.45 ± 0.07 1.65 ± 0.07 1.71 ± 0.07 1.76 ± 0.07 1.83 ± 0.04 Downloaded from 60% 1.60 ± 0.07 1.70 ± 0.07 1.73 ± 0.07 2.07 ± 0.07a 1.72 ± 0.04b Proximal density (g/g) Control 764.57 ± 20.74b 1,175.96 ± 20.74 985.80 ± 20.74 979.74 ± 20.74b 997.91 ± 18.18b 80% 842.06 ± 20.74a 1,175.17 ± 20.74 983.98 ± 20.74 959.51 ± 20.74b 985.92 ± 12.78b 60% 872.83 ± 20.74a 1,182.78 ± 20.74 967.68 ± 20.74 1,102.78 ± 20.74a 1,032.44 ± 13.07a Distal diameter (mm) Control 11.67 ± 0.34 19.64 ± 0.34a 25.48 ± 0.34 28.26 ± 0.34a 29.55 ± 0.20a http://ps.oxfordjournals.org/ 80% 11.46 ± 0.34 18.66 ± 0.34b 24.94 ± 0.34 27.56 ± 0.34a 28.84 ± 0.20b 60% 11.79 ± 0.34 19.41 ± 0.34ab 25.54 ± 0.34 26.74 ± 0.34b 27.13 ± 0.29c Distal cortical thickness (mm) Control 1.52 ± 0.09 2.02 ± 0.09a 1.91 ± 0.09 1.97 ± 0.09ab 1.74 ± 0.06b 80% 1.52 ± 0.09 1.93 ± 0.09ab 1.73 ± 0.09 1.89 ± 0.09b 1.89 ± 0.06ab 60% 1.63 ± 0.09 1.73 ± 0.09b 1.81 ± 0.09 2.19 ± 0.09a 1.97 ± 0.08a Distal density (g/g) Control 865.47 ± 25.09b 1,218.70 ± 25.09 1,098.19 ± 25.09ab 1,129.83 ± 25.09ab 1,157.37 ± 14.46b 80% 988.66 ± 25.09a 1,219.60 ± 25.09 1,115.65 ± 25.09a 1,072.02 ± 25.09b 1,132.39 ± 14.83b at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 60% 924.54 ± 25.09ab 1,218.72 ± 25.09 1,036.49 ± 25.09b 1,186.90 ± 25.09a 1,234.80 ± 22.36a a,bMeans within column for each treatment with no common superscript differ significantly (P < 0.05). 1Control = birds fed a diet containing 100% of NRC (1994) protein and energy requirements; 80% = birds fed a diet containing 80% of NRC (1994) protein and energy requirements; 60% = birds fed a diet containing 60% of NRC (1994) protein and energy requirements. distal cortical bone was thinnest in toms fed the 60% creasing until the end of the trial in strain A (Figure 3). diet at 5 kg of BW but thickest at 20 kg of BW (P As the birds reached 20 kg of BW, PYD was lowest in < 0.05; Table 4). The opposite trend was observed in the 60% group regardless of strain. Strain B PYD levels the control birds during the same time period. Bone peaked at 10 kg of BW in the 60% group, whereas birds density analysis found that the birds fed the 60% diet fed the 80% diet and control toms peaked at 16 kg of had denser proximal and distal cortex after the birds BW (Figure 3). Strain B toms fed at the 60% diet had were 16 kg (P < 0.05; Table 4). Table 5 indicates strain lower PYD levels at 16 and 20 kg of BW compared with A birds’ medial femoral cortex was denser in the 80% control toms (P < 0.05; Figure 3). group at 10 kg of BW but the control and 60% group had a more dense cortex at 16 kg of BW (P < 0.05). Bone Strength However, no difference in cortical density was observed between the treatment groups by the end of the trial. Failure torque was not different between treatment Strain A ash content was highest in the control group groups until the birds reached 16 kg of BW (Table 7). at 10 and 20 kg of BW, but lowest at 16 kg of BW (P < At that point, the 60% group had the highest failure 0.05; Table 6). Strain B controls had more ash at 5 and torque and the 80% group had the lowest (P < 0.05). 10 kg of BW, whereas the 60% group had the lowest Shear strength of all treatment groups increased from 5 (P < 0.05). However, at 16 kg of BW the percent ash to 20 kg of BW, with birds fed the 60% diet having the flipped, with the control birds being lower than those highest shear strength among the groups (P < 0.05). fed the 60% diet (P < 0.05). Table 6 also reports tibia Shear modulus was greater in the 60% group compared ash, but differences observed in strain B at 16 kg and with controls from 5 to 20 kg of BW, which was higher strain A at 20 kg did not reflect differences observed in than birds fed the 80% diet at 10 and 16 kg of BW the femur ash. (P < 0.05).

PYD DISCUSSION Serum PYD was measured as a marker of systemic Previous research examined bone strength in breeder bone resorption. Pyridinoline concentration increased toms or young feed-restricted birds. The breeder toms with BW and age and peaked at 16 kg of BW for con- in previous work were handled twice a week for semen trol birds, whereas it peaked at 10 kg of BW for the collection. This could result in fractures or stresses to birds fed the 60 and 80% diets and then started de- the legs of the turkey that are not seen in commercial Table 5. Effect of dietary energy and protein restriction on turkey on physical attributes of the medial portion of the femur of 2 different commercial strains

BW (kg)

1 5 10 16 20

Item1 A2 B A B A B A B A B

Diameter (cm) Control 7.25 ± 0.34 7.01 ± 0.32 13.84 ± 0.34a 13.56 ± 0.32 17.44 ± 0.34 18.74 ± 0.32a 19.94 ± 0.34a 19.89 ± 0.32a 20.37 ± 0.20a 20.90 ± 0.21a 80% 6.79 ± 0.34 6.83 ± 0.32 12.12 ± 0.34b 12.91 ± 0.32 17.26 ± 0.34 17.24 ± 0.32b 18.62 ± 0.34b 19.38 ± 0.32ab 19.43 ± 0.20b 19.99 ± 0.20b 60% 6.98 ± 0.34 7.36 ± 0.32 12.84 ± 0.34b 13.24 ± 0.32 16.47 ± 0.34 17.41 ± 0.32b 18.16 ± 0.34b 18.71 ± 0.32b 19.03 ± 0.20b 19.30 ± 0.02c Thickness (cm) Control 1.69 ± 0.13 1.63 ± 0.13 2.21 ± 0.13 2.33 ± 0.13 1.99 ± 0.13 2.20 ± 0.13ab 2.58 ± 0.13a 1.98 ± 0.13b 2.46 ± 0.08 2.28 ± 0.13b 80% 1.39 ± 0.13 1.54 ± 0.13 1.94 ± 0.13 2.18 ± 0.13 1.98 ± 0.13 2.02 ± 0.13b 2.04 ± 0.13b 2.21 ± 0.13ab 2.47 ± 0.08 2.09 ± 0.13b 60% 1.52 ± 0.13 1.54 ± 0.13 2.06 ± 0.13 2.11 ± 0.13 1.99 ± 0.13 2.49 ± 0.13a 2.32 ± 0.13ab 2.55 ± 0.13a 2.48 ± 0.11 2.65 ± 0.13a Density (g/g) FEED RESTRICTIONONTURKEYFEMURS Control 1,401.0 ± 46.8 1,293.0 ± 46.8b 1,778.8 ± 46.8 1,773.9 ± 47.8 1,700.3 ± 46.8ab 1,679.4 ± 47.8 1,638.4 ± 46.8a 1,686.7 ± 47.8 1,802.6 ± 26.9 1,714.0 ± 29.1 80% 1,403.2 ± 46.8 1,453.2 ± 46.8a 1,777.2 ± 46.8 1,756.1 ± 47.8 1,751.3 ± 46.8a 1,699.3 ± 47.8 1,571.3 ± 46.8b 1,625.6 ± 47.8 1,761.2 ± 26.9 1,698.7 ± 40.4 60% 1,777.3 ± 46.8 1,313.8 ± 46.8b 1,849.1 ± 46.8 1,660.4 ± 47.8 1,576.9 ± 46.8b 1,599.5 ± 47.8 1,774.4 ± 46.8a 1,632.9 ± 47.8 1,857.2 ± 41.5 1,775.7 ± 27.7 a–cMeans within column for each treatment with no common superscript differ significantly (P < 0.05). 1Control = birds fed a diet containing 100% of NRC (1994) protein and energy requirements; 80% = birds fed a diet containing 80% of NRC (1994) protein and energy requirements; 60% = birds fed a diet containing 60% of NRC (1994) protein and energy requirements. 2A = commercial genetic strain A; B = commercial genetic strain B.

Table 6. The effect of dietary energy and protein restriction on femoral and tibial ash in turkey toms

BW (kg)

1 5 10 16 20

Item1 A2 B A B A B A B A B Femoral ash (%) Control 45.92 ± 0.84 45.40 ± 0.78 50.01 ± 0.84 47.16 ± 0.78a 49.97 ± 0.84a 48.87 ± 0.78a 48.90 ± 0.84b 44.67 ± 0.78b 51.27 ± 0.53a 49.12 ± 0.78 80% 45.06 ± 0.84 46.13 ± 0.78 47.89 ± 0.84 44.78 ± 0.78b 47.63 ± 0.84b 48.29 ± 0.78a 48.97 ± 0.84b 46.44 ± 0.78b 51.51 ± 0.53a 49.89 ± 0.78 60% 44.19 ± 0.84 45.07 ± 0.78 48.55 ± 0.84 43.68 ± 0.78b 45.56 ± 0.84b 44.12 ± 0.78b 52.61 ± 0.84a 50.07 ± 0.78a 48.99 ± 0.76b 49.76 ± 0.78 Tibial ash (%) Control 44.40 ± 0.66 45.82 ± 0.77 50.24 ± 0.66a 48.93 ± 0.77 54.00 ± 0.66 53.84 ± 0.77 55.09 ± 0.66 53.55 ± 0.77a 54.00 ± 0.59b 52.45 ± 0.77 80% 43.88 ± 0.66 45.53 ± 0.77 47.92 ± 0.66b 49.19 ± 0.77 53.57 ± 0.66 53.06 ± 0.77 55.76 ± 0.66 50.56 ± 0.77b 55.91 ± 0.59a 52.92 ± 0.77 60% 42.85 ± 0.66 46.27 ± 0.77 50.49 ± 0.66a 50.35 ± 0.77 53.23 ± 0.66 52.43 ± 0.77 54.59 ± 0.66 53.21 ± 0.77a 55.61 ± 0.59a 53.79 ± 0.77 a,bMeans within column for each treatment with no common superscript differ significantly (P < 0.05). 1Control = birds fed a diet containing 100% of NRC (1994) protein and energy requirements; 80% = birds fed a diet containing 80% of NRC (1994) protein and energy requirements; 60% = birds fed a diet containing 60% of NRC (1994) protein and energy requirements. 2A = commercial genetic strain A; B = commercial genetic strain B.

949

Downloaded from from Downloaded http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 6, April on University Harvard Zoology, Comp Museum the of Library Mayr Ernst at 950 VAN WYHE ET AL. (1995), and Bruno et al. (2007). Between the groups, nutrient-restricted birds had longer bones at the 10-kg mark and onward compared with control, which the age of the birds fed the 60% diet may account for. The difference in bone length may also be attributed to different early BW gains among the experimental groups with protein- and energy-restricted birds having lighter weights at the same point in life. Early mechanical loading in broiler chicks has been associated with accelerated cartilage resorption and early mineralization mediated by enhanced osteopontin and matrix metal- Downloaded from loproteinases expression, resulting in narrower growth plates and shorter bones (Reich et al., 2005). Bruno et al. (2000) restricted growth at an earlier stage in broiler chickens and documented significant decreases in femur, tibia, and humerus lengths in the restricted http://ps.oxfordjournals.org/ group compared with the ad libitum group. Their results are contrary to the results of the current study, but the bone lengths in that study were averaged over the trial period and that does not take into account the compensatory growth that may have taken place later in the bird’s life. However, control birds had wider bone

diameters, which might be an adaptive response to the at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 compression forces occurring during a rapid weight gain (Akhter et al., 1998). Femurs of heavier commercial toms had wider cross-sectional areas than the slow- growing strain of the same age, but differences observed were due to genotype, not diet (Zhong et al., 2012). When comparisons were made between the heavier and lighter birds of the same age and genetic strain, lighter birds had wider diameters of the proximal femur when data accounted for BW, while still maintaining the narrower cross-sectional area (Crespo et al., 2000). Cross- sectional areas of the femur were not measured in our study. The cortical dimensions were altered by the nutrient restriction. Cortical thickness in protein- and energy- Figure 3. The effect of dietary energy and protein restriction on restricted birds, particularly the 60% group, caught up the pyridinoline (PYD) in 2 commercial turkey strains. (A) The effect or even exceeded the control as the birds reached 20 of dietary energy and protein restriction on the PYD of strain A. (B) The effect of dietary energy and protein restriction on the PYB of kg of BW in all 3 regions of the bone, except for the strain B. Control = birds fed a diet containing 100% of NRC (1994) proximal femur, indicating time is essential for bone to protein and energy requirements; 80% = birds fed a diet containing mature. A bone with thinner cortices is more prone to 80% of NRC (1994) protein and energy requirements; 60% = birds fed a diet containing 60% of NRC (1994) protein and energy requirements; fracture (Crabtree et al., 2001), but the strength also significant differences (*) were defined as P < 0.05. depends upon the density of the bone. The wider cortical bone at the medial region, even comparing at common ages for 60% and control-fed birds, suggests that toms. Our work examines the difference in nutrient re- it might be the site where the femur experiences the striction in commercial toms until market weight. The major effect of mechanical loading and the subsequent performance result from the current study corroborates bone formation taking place, making the bone stronger the findings of Turner and Lilburn (1992), in which at this point. However, Cheng and Ward (2007) suggest they restricted protein in the toms’ diet for the first 8 no common skeletal orientation exists. Birds restricted wk and documented that the BW was lower in protein- in diet by 60% of nutritional requirements had denser deficient toms at 18 wk of age compared with normal femora in the proximal and distal regions of the bone toms. This suggests compensatory gain was insufficient compared with the other 2 groups after they reached 10 to return protein-deficient toms to normal weight gain kg of BW. Toms fed the 60% diet had similar femoral in the last 10 wk of growth. The length of the femur density at 16 kg of BW compared with femoral density increased within birds of the same group with age and of 20-kg control toms. The results of cortical thickness BW in the current study, which is in accordance with and density indicate that maturation of cortical bone studies by Bond et al. (1991), Skinner and Waldroup is more age-dependent rather than diet-dependent. A FEED RESTRICTION ON TURKEY FEMURS 951

Table 7. The effect of dietary energy and protein restriction on femoral mechanical properties

BW (kg)

Item1 1 5 10 16 20 Failure torque (N·m) Control 1.81 ± 0.94 10.67 ± 0.97 25.04 ± 0.97 38.43 ± 0.94b 48.44 ± 0.87b 80% 1.95 ± 1.00 9.80 ± 0.94 23.52 ± 0.97 33.44 ± 0.97c 41.98 ± 0.87c 60% 2.03 ± 1.00 11.72 ± 0.94 25.59 ± 0.94 40.73 ± 0.94a 52.06 ± 0.85a Shear strength (MPa) Control 42.76 ± 2.70b 30.10 ± 2.76b 36.38 ± 2.76b 58.31 ± 2.70b 70.54 ± 2.53b 80% 53.74 ± 2.84a 33.92 ± 2.76b 40.57 ± 2.70b 55.26 ± 2.84b 72.30 ± 2.56b b a a a a 60% 44.56 ± 2.84 37.65 ± 2.70 53.62 ± 2.70 70.99 ± 2.76 79.62 ± 2.50 Downloaded from Shear modulus (GPa) Control 0.86 ± 0.16b 0.97 ± 0.16b 1.57 ± 0.16b 2.07 ± 0.16b 2.30 ± 0.16b 80% 1.16 ± 0.16a 1.15 ± 0.16ab 1.64 ± 0.16b 2.02 ± 0.16b 2.36 ± 0.16ab 60% 1.00 ± 0.16ab 1.31 ± 0.16a 2.10 ± 0.16a 2.27 ± 0.16a 2.51 ± 0.16a a–cMeans within same BW lacking a common superscript differ significantly (P < 0.05).

1Control = birds fed a diet containing 100% of NRC (1994) protein and energy requirements; 80% = birds fed a diet containing 80% of NRC (1994) http://ps.oxfordjournals.org/ protein and energy requirements; 60% = birds fed a diet containing 60% of NRC (1994) protein and energy requirements. comparison between the birds fed the 60% NRC diet of turkey bones may help determine the mechanism be- at 10 kg (111 d of age) to control-fed birds at 16 kg sides the changes in bone strength. (112 d of age) shows that the proximal cortical thick- The serum PYD data showed a trend for a peak at ness and density, as well as the distal density, are not around 10 kg of BW in the 60% nutrient-restricted

different at the same age, despite different BW (10 and groups of both strains and then started decreasing. In at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 16 kg) for those birds. Turner and Lilburn (1992) found the control group, PYD peaked at 16 kg of BW and only age-dependent rather than diet-dependent differences started decreasing afterward. In birds from a previous in bone ash in turkeys fed diets with restricted protein study, serum PYD increased with age, peaking at wk 17 and Lys. Genetically slower growing toms had higher before decreasing (Roberson et al., 2005). Our results bone mineral content and better mechanical proper- suggest that bone is experiencing a high turnover rate ties when their weight matched with commercial fast- in the control and 80% group compared with the 60% growing toms, suggesting matrix maturation and opti- group when market BW was achieved. This is probably mum mineralization takes time in meat turkeys (Zhong due to the reduced nutrients in the diet rather than et al., 2012). A correlation of 93% between ash and the differences in age or feed intake. Toms in the 60% whole BMD was reported for chickens (Onyango et al., group consumed more feed than control birds, but not 2003), but unlike the consistent improvement in density more than the 80% group. The rate of systemic bone of proximal and distal femur in birds fed the 60% NRC resorption observed in the current study suggests that diet, ash data were more variable at the end of the even though fast-growing toms modulated the same ge- trial. A difference in age-related response of bone ash in ometry by increasing the diameter of femur, they could the epiphyseal and diaphyseal region has been observed not attain optimal material strength, as mineralization before (Turner and Lilburn, 1992). The variability ob- and matrix maturation is time-dependent (Zhong et al., served in the present study could be age dependent 2012). rather than diet dependent. Also, ash measurement in The results herein demonstrate that a manipulation the current study did not take into account the differ- of the growth rate via the diet during the production ence in the epiphyseal and diaphyseal section of femur period affects numerous bone parameters. Bone min- and could be the reason for variable whole bone ash at eral density and shear strength improved at the end the end of the trial. of the trial when protein and energy was restricted Shear strength and failure torque were higher for by 60%, but some variations occurred throughout the birds fed the 60% diet as they attained 10 kg of BW. study. Bone strength appears to correlate with bone This might be the result of an overall better densi- density and also with stability of organic matrix, as ty profile associated with this group. Increased bone shown by PYD. Though it is not economically advis- mineral density is important for imparting stiffness to able to increase feed conversion ratio and increase feed bones and will result in strong bones (Frost and Ro- intake, a complete study with different restriction levels land, 1991; Orban et al., 1993) together with appropri- and their effect on bone quality could further elucidate ate collagen fiber orientation (Rath et al., 2000; Davi- the actual effects of nutrient restriction on bone quality son et al., 2006). The decreased load on the bone early in turkeys. in life may allow for the birds to reorient their bone structure to correspond to changing loads. In the tra- ACKNOWLEDGMENTS becular bone of sheep, the bone adjusted and realigned itself in response to peak joint loading (Barak et al., The authors thank the Animal Agriculture Initiative 2011). Further studies into loading and mineral density at Michigan State University for providing funding for 952 VAN WYHE ET AL. the project. We also thank Cara Robison for her as- Hertrampf, T., B. Schleipen, C. Offermanns, M. Velders, U. Lauden- sistance in sample collection and analysis, Angelo Na- bach, and P. Diel. 2009. Comparison of the bone protective effects of an isoflavone-rich diet with dietary and subcutaneous ad- politano, farm manager, and the rest of the farm crew ministrations of genistein in ovariectomized rats. Toxicol. Lett. at Michigan State University Poultry Research and 184:198–203. Teaching Farm for their assistance in this study. Levenston, M. E., G. S. Beaupre, and M. C. H. Van Der Meulen. 1994. Improved method for analysis of whole bone torsion tests. J. Bone Miner. Res. 9:1459–1465. REFERENCES Lilburn, M. S., A. Mitchell, and J. Anderson. 2006. The relationship between growth of commercial toms and linear skeletal develop- Akhter, M. P., D. M. Cullen, E. A. Pedersen, D. B. Kimmel, and R. ment. Poult. Sci. 85(Suppl.1):31. R. Reeker. 1998. Bone response to in vivo mechanical loading in NRC. 1994. Nutrient Requirements for Poultry. 9th rev. ed. Natl.

two breeds of mice. Calcif. Tissue Int. 63:442–449. Acad. Press, Washington, DC. Downloaded from Applegate, T. J., P. Jaynes, J. J. Dibner, M. Quiroz, and J. D. Rich- Onyango, E. M., P. Y. Hester, R. L. Stroshine, and O. Adeola. 2003. ards. 2006. Growth and mineralization of the femur of the male Bone densitometry as an indicator of percentage tibia ash in turkey. Poult. Sci. 85(Suppl. 1):30. broiler chicks fed varying dietary calcium and phosphorus levels. Barak, M. M., D. E. Lieberman, and J. J. Hublin. 2011. A Wolff Poult. Sci. 82:1787–1791. in sheep’s clothing: Trabecular bone adaptation in response to Orban, J. L., D. A. Roland, and M. M. Bryant. 1993. Factors influ- changes in joint loading orientation. Bone 49:1141–1151. encing bone mineral content, density, breaking strength and ash Bond, P. L., T. W. Sullivan, J. H. Douglas, and L. B. Robeson. 1991. as a response to criteria for assessing bone quality in chickens. http://ps.oxfordjournals.org/ Influence of age, sex, and method of rearing on tibia length and Poult. Sci. 72:437–446. mineral deposition in broilers. Poult. Sci. 70:1936–1942. Reich, A., N. Jaffe, A. Tong, I. Lavelin, O. Genina, M. Pines, D. Bruno, L. D. G., R. L. Furlan, and E. B. Malheiros. 2000. Influence Sklan, A. Nussinovitch, and E. Monsonego-Ornan. 2005. Weight of early quantitative food restriction on long bone growth at dif- loading young chicks inhibits bone elongation and promotes ferent environmental temperatures in broiler chickens. Br. Poult. growth plate ossification and vascularization. J. Appl. Physiol. Sci. 41:389–394. 98:2381–2389. Bruno, L. D. G., B. C. Luquetti, R. L. Furlan, and M. Macari. 2007. Roberson, K. D., M. W. Klunzinger, R. A. Charbeneau, and M. Influence of early qualitative feed restriction and environmental W. Orth. 2005. Evaluation of phytase concentration needed

temperature on long bone development of broiler chickens. J. for growing-finishing commercial turkey toms. Br. Poult. Sci. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Therm. Biol. 32:349–354. 46:470–477. Cheng, T., and T. Ward. 2007. Impact of Availa/M on growth, feed Shaw, D. T., D. W. Rozeboom, G. M. Hill, M. W. Orth, D. S. Rosen- conversion, femur spiral fracture related mortality, bone morpho- stein, and J. E. Link. 2006. Impact of supplement withdrawal and metrics and carcass yield in tom turkeys. Poult. Sci. 86(Suppl. wheat middling inclusion on bone metabolism, bone strength, 1):781. and the incidence of bone fractures occurring at slaughter in pigs. Crabtree, N., N. Loveridge, M. Parker, N. Rushton, J. Power, and K. J. Anim. Sci. 84:1138–1146. L. Bell. 2001. Intracapsular hip fracture and the region-specific Skinner, J. T., and P. W. Waldroup. 1995. Allometric bone develop- loss of cortical bone: Analysis by peripheral quantitative com- ment in floor reared broilers. J. Appl. Poult. Res. 4:265–270. puted tomography. J. Bone Miner. Res. 16:1318–1328. Turner, C. H., and A. G. Robling. 2004. Mechanical loading and Crespo, R., S. M. Stover, K. T. Taylor, R. P. Chin, and H. L. bone formation. Bonekey. Osteovision 1:15–23. Shivaprasad. 2000. Morphometric and mechanical properties of Turner, K. A., and M. S. Lilburn. 1992. The effect of early protein femora in young adult male turkeys with and without femoral restriction (zero to eight weeks) on skeletal development in tur- fractures. Poult. Sci. 79:602–608. key toms from two to eighteen weeks. Poult. Sci. 71:1680–1686. Davison, K. S., K. Siminoski, J. D. Adachi, D. A. Hanley, D. Warden, S. J. 2006. Breaking the rule for bone adaptation to me- Goltzman, A. B. Hodsman, and R. Josse. 2006. Bone strength: chanical loading. J. Appl. Physiol. 100:1441–1442. The whole is greater than the sum of its parts. Semin. Arthritis Williams, B., D. Waddington, D. H. Murray, and C. Farquharson. Rheum. 36:22–31. 2004. Bone strength during growth: Influence of growth rate on Frost, T. J., and D. A. Roland. 1991. Current methods used in de- cortical porosity and mineralization. Calcif. Tissue Int. 74:236– termination and evaluation of tibia strength: A correlation study 245. involving birds fed various levels of cholecalciferol. Poult. Sci. Young, W. C., and R. G. Budynas. 2002. Roark’s Formulas for 70:1640–1643. Stress and Strain. 7th ed. McGraw-Hill, New York, NY. Havenstein, G. B., P. R. Ferket, J. L. Grimes, M. A. Qureshi, and K. Zhong, Z., M. Muckley, S. Agcaoglu, M. E. Grisham, H. Zhao, M. E. Nestor. 2007. Comparison of the performance of 1966- versus Orth, M. S. Lilburn, O. Akkus, and D. M. Karcher. 2012. The 2003-type turkeys when fed representative 1966 and 2003 turkey morphological, material-level, and ash properties of turkey fe- diets: Growth rate, livability, and feed conversion. Poult. Sci. murs from 3 different genetic strains during production. Poult. 86:232–240. Sci. 91:2736–2746.

Spatiotemporal expression patterns of doublesex and mab-3 related transcription factor 1 in the chicken developing gonads and Müllerian ducts

T. Omotehara ,* C. A. Smith ,† Y. Mantani ,* Y. Kobayashi ,* A. Tatsumi ,* D. Nagahara ,* R. Hashimoto ,* T. Hirano ,* Y. Umemura ,* T. Yokoyama ,* H. Kitagawa ,* and N. Hoshi *1

* Department of Animal Science, Graduate School of Agricultural Science, Kobe University, Kobe, Hyogo 657-8501, Japan; and † Murdoch Childrens Research Institute, Downloaded from Royal Children’s Hospital, Melbourne, Victoria, 3052, Australia

ABSTRACT Sex of birds is genetically determined by toli cells forming testicular cords expressed DMRT1 in the inheritance of sex chromosomes (ZZ for male and the testes after sex determination. Germ cells expressed ZW for female), and the Z-linked gene named doublesex DMRT1 equally in males and females after sex determi- http://ps.oxfordjournals.org/ and mab-3 related transcription factor 1 (DMRT1) is a nation. The expression was continuous in males, but in candidate sex-determining gene in avian species. How- females it gradually disappeared from the germ cells in ever, the mechanisms underlying sex determination in the central part of the cortex of the left ovary toward birds are not yet understood, and the expression pat- both edges. The DMRT1 was also detected in the tubal terns of the DMRT1 protein in urogenital tissues have ridge, which is a precursor of the Müllerian duct, and not been identified. In the current study, we used im- at the mesenchyme and outermost coelomic epithelium

munohistochemistry to investigate the detailed expres- of the Müllerian duct in both sexes. Strong expression at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 sion patterns of the DMRT1 protein in the urogenital was observed in the males, but it was restricted to coe- systems (including Müllerian ducts) in male and female lomic epithelium after the regression of the duct start- chicken embryos throughout embryonic development. ed. Thus, we observed the detailed spatiotemporal ex- Gonadal somatic cells in the male indifferent gonads pression patterns of DMRT1 in the developing chicken showed stronger expressions of DMRT1 compared with urogenital systems throughout embryonic development, those in the female indifferent gonads well before the suggesting its various roles in the development of uro- presumptive period of the sex determination, and Ser- genital tissues in the chicken embryo. Key words: doublesex and mab-3 related transcription factor 1, embryonic development, sex determination and differentiation, urogenital tissue, immunohistochemistry 2014 Poultry Science 93 :953–958 http://dx.doi.org/ 10.3382/ps.2013-03672

INTRODUCTION (in which many germ cells are contained), but the cortex in the right gonad fails to proliferate and regresses Gonadal development is a conserved event among the during hatching (Romanoff, 1960). vertebrates. Gonads arise on the ventromedial region Although the sex in birds is determined genetically of both the right and left mesonephroi as indifferent (ZZ for male, ZW for female), the molecular mechanism gonads called genital ridges at first, and then they dif- that determines whether a chicken embryo develops as ferentiate into testes or ovaries via a sex-determining a male or female is poorly understood. The most like- process that varies among vertebrate species. The geni- ly possible candidate for the chicken sex-determining tal ridges in the chicken embryo are formed by embry- gene is a Z-linked gene, doublesex and mab-3 related onic day (E) 3.5, which is equivalent to the Hamburger transcription factor 1 (DMRT1), a DM domain tran- and Hamilton stage (HH) 21 or 22 (Hamburger and scription factor that has a conserved role for testis dif- Hamilton, 1951). After the indifferent period, sexual ferentiation among vertebrates (Matson and Zarkower, differences of the gonads are observed at around E6.5 2012). Knockdown of DMRT1 induces ovarian forma- (HH30). In the male gonad, the testicular cord com- tion in the genetically male chick (Smith et al., 2009). posed by Sertoli cells surrounding germ cells is formed. A cell-autonomous sex identity was identified in each The cortex is thickened in the left gonad of the female male and female cell of the chicken (Zhao et al., 2010), and chicken homolog of hemogen, which is known as a © 2014 Poultry Science Association Inc. hematopoietic tissue-specific gene in mice and humans Received October 7, 2013. Accepted December 16, 2013. (Yang et al., 2001), is specifically involved in testis 1 Corresponding author: [email protected] development (Nakata et al., 2013), suggesting that a

953 954 OMOTEHARA ET AL. Downloaded from http://ps.oxfordjournals.org/

Figure 1. Doublesex and mab-3 related transcription factor 1 (DMRT1) localization in the left developing gonad at embryonic day (E) 4.5 (A, E), E6.5 (B, F), E8–9 (C, G), and E15 (D, H). The DMRT1 protein expression is observed in the nucleus of the gonadal somatic cells (arrows), which are (pre-) Sertoli cells in the male at E6.5 onward (B–D), and the germ cells (arrow heads) of both sexes. The DMRT1 protein is first expressed at E4.5 in both male and female gonadal somatic cells (A, E) and its expression in the male (A–D) is stronger than that in the female (E–H) at each embryonic day. Immunoexpression of DMRT1 in the germ cells is first detectable at E6.5 in both sexes (B, F). The expression level in the male germ cells is equivalent to that in the Sertoli cells (arrows) at E8–9 (C) and is higher at E15 (D). The DMRT1 protein expressions of the female germ cells in both the cortex and medulla of the left ovary are increased at E8–9 (G) from E6.5 (F). At E15, very weak expressions at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 are observed in some germ cells in the cortex (c), although the germ cells in the medulla (m) still express DMRT1 (H). Dashed lines show borders between the cortex and medulla. Scale bar = 25 μm. Color version available in the online PDF. unique sex-determining system exists in avian species germ cells (Matson et al., 2010; Krentz et al., 2011). among vertebrates. Although DMRT1 mRNA expression in the urogenital Differences also exist between the male and female tissues of the chicken embryo was examined by real in accessory sex structures, such as the Müllerian duct. time reverse transcription PCR and (whole-mount) in The precursor of the duct, which is called the tubal situ hybridization (Raymond et al., 1999; Smith and ridge, first appears as a groovelike invagination at the Sinclair, 2001; Yamamoto et al., 2003), it is not yet anterior end of the mesonephros at E4.0 (HH24). The known which cells express DMRT1, and the expres- more caudal region of the groove is closed with the fu- sion patterns of both mRNA and protein through em- sion of the edges to form the tubelike structure at E5.0 bryonic development have not been identified. In the (HH27; Romanoff, 1960). The Müllerian duct is distin- current study, to investigate the elaborate expression guishable into 3 cellular components: the inner epithe- patterns and the fundamental characteristics of chick- lial cells (Müllerian duct epithelium), the mesenchymal en DMRT1, we conducted immunohistochemistry on cells surrounding the duct [Müllerian duct mesenchyme the DMRT1 protein at various embryonic stages and (MDM)], and the outermost coelomic epithelial cells we describe the spatiotemporal expression patterns of [Müllerian coelomic epithelium (MCE)]. All of these DMRT1 in the chicken urogenital system, including the components are derived from the coelomic epithelial Müllerian ducts. cells on the mesonephros (Guioli et al., 2007). The chicken Müllerian duct develops as an oviduct MATERIALS AND METHODS only in the left side of the female, and the right side in the female and both in the male are regressed dur- Embryos ing embryonic development. The regression starts at around E8.0 (HH34) in the male and E12 (HH38) in Fertilized eggs of White Leghorn chickens were pur- the female (Romanoff, 1960; Teng, 1987), and anti- chased from Nihon Layer Ltd. (Gifu, Japan) and incu- Müllerian hormone (AMH) is involved in the degen- bated at 38°C and 70% RH. The embryos were collected eration in both the male and female chick (Teng, 1987; every half day from E3.5 (HH21 or 22) to E7.5 (HH32) Weniger, 1991). The DMRT1 mRNA is also expressed and every day from E8.0 (HH34) to E17 (HH43). in the Müllerian duct in the chicken (Smith and Sin- clair, 2001), but the detailed expression pattern and the Genetic Sexing function are not characterized yet. The DMRT1 protein in the mouse is well charac- A small piece of limb tissue was boiled in distilled terized regarding its roles in male gonadal somatic water for 10 min and homogenized to obtain genomic cells (Matson et al., 2011) and both male and female DNA. Both the W chromosome-linked (female specific) EXPRESSION PATTERNS OF A TESTIS-DETERMINING GENE 955 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 2. Doublesex and mab-3 related transcription factor 1 (DMRT1) localization in the left ovary at embryonic day (E) 12 (A), E13 (B), E15 (C), and E17 (D). The DMRT1 protein is expressed in all germ cells in the cortex (c) of the left ovary at E12 (A) and is not detected in some germ cells present at the central region of the cortex at E13 (B). The DMRT1 protein is negative in the majority of the germ cells in the cortex at E15 (C) and in almost all of them at E17 (D). Germ cells scattered in the medulla (m) express DMRT1 at each embryonic day (A–D). Each double-headed arrow shows the range in which DMRT1 was not expressed in the germ cells in the cortex. Note that the arrows gradually lengthen as the ovary develops. Scale bar = 200 μm. Color version available in the online PDF.

XhoI repeat sequence and 18S ribosomal RNA gene Japan) at room temperature for 1 h, and then reacted serving as an internal control were amplified with spe- with anti-chicken DMRT1 rabbit polyclonal antibody cific primers (Clinton et al., 2001). The PCR conditions (1:24,000; Smith et al., 2003) in a humidified chamber were 30 cycles of 95°C for 30 s, 56°C for 30 s, and 72°C at 4°C for 18 h. for 1 min. The amplified products were electrophoresed The sections were washed with PBS and incubated in 2% agarose gel, dyed by ethidium bromide, and visu- with a Dako EnVision+ system-HRP labeled polymer alized with a UV transilluminator. (Dako, Glostrup, Denmark) at room temperature for 1 h. Immunoreactivities of the primary antibody were Tissue Preparation detected with 3,3-diaminobenzidine tetrahydrochloride solution (Dako), and Mayer’s hematoxylin was used as Urogenital tissues containing gonads were fixed in counterstaining. Negative controls, in which the prima- 4% paraformaldehyde in 0.1 M phosphate buffer (pH ry antibodies were replaced with nonimmunized serum, 7.4) for 24 h at 4°C. The specimens were dehydrated did not show nonspecific staining. with an ethanol series followed by xylene and then embedded in paraffin. Then, 4-μm-thick sections were cut RESULTS AND DISCUSSION by microtome and placed on slide glasses (Platinum; Matsunami Glass Ind. Ltd., Kishiwada, Japan). The expression patterns of chicken DMRT1 in the urogenital tissues during embryonic development were Immunohistochemistry examined. The DMRT1 protein was first detected in the indifferent gonads of both sexes at E4.5 (HH25), After deparaffinization and hydration, the sections before the sex determination, and the expression in the were heated at 121°C for 20 min by immersion in 10 male was stronger than that in the female (Figure 1A mM sodium citrate buffer (pH 6.0) for antigen retriev- and E). The expression levels became higher, especially al, and then immersed in absolute methanol and 0.5% in male gonadal somatic cells at E6.5 (HH30) onward H2O2 for 30 min at room temperature to quench the (Figure 1B–D). The somatic cells, (pre-) Sertoli cells, endogenous peroxidase activity. The sections, which formed sex cords containing germ cells. The expres- were rinsed in 10 mM PBS (pH 7.4), were incubated sion of DMRT1 in the somatic cells of the right gonad with Blocking One Histo (Nacalai Tesque Inc., Kyoto, showed the same patterns as in the left gonad in the 956 OMOTEHARA ET AL. Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 3. Doublesex and mab-3 related transcription factor 1 (DMRT1) localization in the developing Müllerian ducts at embryonic day (E) 4.5 (A–D), E7.5 (E–H), E8.0 (I–L), and E13 (M–P). The DMRT1 protein is first detectable at E4.5 in both the male and female tubal ridge (t; A–D) and is expressed at Müllerian duct mesenchyme (MDM; arrow) and Müllerian coelomic epithelium (MCE; arrow head) at E7.5 in both sexes, although the expression in the male is higher than that in the female (E–H). Regression of the Müllerian ducts occurs at both sides in the males, and most expressions of DMRT1 are restricted to the MCE (I, J), whereas the ducts develop continuously and no changes in the DMRT1 expression pattern are observed in the females at E8.0 (K, L). At E13, the ducts are merely connective tissues on both sides of the males and on only the right side of the females, whereas the Müllerian duct is developed well on the left side of the females (M–P) and DMRT1 is also expressed at the MCE in the males (M, N) and is weakly expressed at the MDM and MCE in the females (O, P). w = Wolffian duct. Scale bar = 25 μm. Color version available in the online PDF. male and as in the medullary region of the left gonad in of the male gonads was equivalent to that in the Ser- the female (data not shown). toli cells at E8–9 (HH35), and then its expression was These expression patterns of DMRT1 in the gonadal stronger in the germ cells than in the Sertoli cells (Fig- somatic cells support the Z dosage hypothesis (Chue ure 1C and D). and Smith, 2011), which proposes that high levels of The DMRT1 protein in the mouse spermatogonia DMRT1 expression in male gonads determine testis directly reduces responsiveness to a meiosis inducer, differentiation to promote male-specific genes contain- retinoic acid. It also suppresses STRA8 expression, ing SOX9 and to suppress the genes involving ovar- which is required for the onset of meiosis, and it in- ian development, such as FOXL2 and RSPO1. Indeed, duces SOHLH1 transcription (Matson et al., 2010). As DMRT1 in the postnatal male mouse induces the male a result, the onset of meiosis is repressed and spermato- gene expression and represses female genes to maintain gonial differentiation is induced in the male fetal germ the testis (Matson et al., 2011). cells. Because chicken DMRT1 was expressed after the Germ cells—which are distinguishable by the round sex determination, which is also the timing of the deter- and larger nucleus and larger cytoplasm compared mination of germ cell fate, whether the germ cell differ- with gonadal somatic cells—began to express DMRT1 entiates into a spermatogonium or an oocyte, chicken in both sexes at E6.5 (HH30), just after the sex de- DMRT1 may play a role in promoting male germ cells termination, and its expression levels were forced with to proliferate and differentiate into spermatogonia. the progression of the gonadal differentiation (Figure Most female germ cells of the left ovary were present 1B and F). The intensity of DMRT1 expression in the in the cortex, and the rest were scattered in the me- germ cells surrounded by Sertoli cells in the medulla dulla, and those in the right ovary that failed to form EXPRESSION PATTERNS OF A TESTIS-DETERMINING GENE 957 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 4. A scheme of the spatiotemporal and sexual dimorphic expression patterns of doublesex and mab-3 related transcription factor 1 (DMRT1) in the urogenital system of the chicken embryo. The line width represents the relative expression level compared with that in the opposite sex. A shaded line indicates the presumptive period of sex determination. the cortex were also spread in the medulla as is in the The DMRT1 protein was also expressed in both the left ovary. The DMRT1 protein was expressed in these right and left Müllerian ducts and was localized in germ cells of both the cortex and medulla until E12 the MDM and MCE. Its expression was detectable at (HH38; Figure 1F and G), and then the expressions E4.5 (HH25) in both the male and female tubal ridges disappeared in order from the central region of the cor- (Figure 3A–D). The expression levels in the male were tex in the left ovary (Figures 1H and 2). In contrast, stronger than those in the female (as in the gonad- DMRT1 was still expressed at the germ cells in the al somatic cells), and those expression patterns were medulla of both the right and left ovaries. observed until E7.5 (HH32; Figure 3E–H). At E8.0 Female germ cells in the left ovary enter meiosis his- (HH34), regression of the Müllerian ducts had occurred tologically at E15.5 in the cortex, especially at its medi- in the male and almost all DMRT1 localization was al part, and express STRA8 at E12.5 onward (Smith et restricted to the MCE, whereas both the right and left al., 2008). The DMRT1 protein directly induces STRA8 Müllerian ducts was developed equally and the DMRT1 expression in the mouse fetal ovarian germ cells and expression patterns were not changed in the female disappears from the nuclei of the germ cells expressing ducts (Figure 3I–L). STRA8 (Krentz et al., 2011). Thus, the concordance of Previous studies confirmed the necessity of both the disappearance of DMRT1 (Figure 2) and the initia- AMHR2 and BMPR1A for the mediation of AMH action of STRA8 transcription in the chicken left ovary tivity to degenerate the Müllerian ducts and their mor- suggests that chicken DMRT1 promotes meiosis via the phological localization in MDM (Mishina et al., 1996; upregulation of STRA8, as in the female germ cells of Jamin et al., 2002; Arango et al., 2008). The strong the fetal mouse. expression of DMRT1 in the male MDM just before 958 OMOTEHARA ET AL. the regression of the male Müllerian ducts indicates the gression during male sexual development. Nat. Genet. 32:408– possibility that DMRT1 is involved in the degeneration 410. Krentz, A. D., M. W. Murphy, A. L. Sarver, M. D. Griswold, V. J. of the ducts to regulate the AMH signaling (e.g., the Bardwell, and D. Zarkower. 2011. DMRT1 promotes oogenesis by activity of AMHR2 or BMPR1A). transcriptional activation of Stra8 in the mammalian fetal ovary. In the present study, both the right and left Mül- Dev. Biol. 356:63–70. lerian ducts in the male and only the right Müllerian Matson, C. K., M. W. Murphy, M. D. Griswold, S. Yoshida, V. J. Bardwell, and D. Zarkower. 2010. The mammalian Doublesex duct in the female were retrogressed, becoming rem- homolog DMRT1 is a transcriptional gatekeeper that controls nants of connective tissues at E13 (HH39), and DMRT1 the mitosis versus meiosis decision in male germ cells. Dev. Cell was weakly expressed in the MCE of the males and 19:612–624. the MDM of the females (Figure 3M–P). Our finding Matson, C. K., M. W. Murphy, A. L. Sarver, M. D. Griswold, V.

J. Bardwell, and D. Zarkower. 2011. DMRT1 prevents female Downloaded from that the Müllerian ducts of the males were degener- reprogramming in the postnatal mammalian testis. Nature ated at around E8.0 (HH34) and those of the females 476:101–104. became residual around E13 (HH39) suggests that a Matson, C. K., and D. Zarkower. 2012. Sex and the singular DM domain: Insights into sexual regulation, evolution and plasticity. sex-dependent mechanism underlying the degeneration Nat. Rev. Genet. 13:163–174. of Müllerian ducts exists. Mishina, Y., R. Rey, M. J. Finegold, M. M. Matzuk, N. Josso, R. L. The current study revealed the detailed spatiotem- Cate, and R. R. Behringer. 1996. Genetic analysis of the Mülleri- http://ps.oxfordjournals.org/ poral expression profiles of chicken DMRT1 in the de- an-inhibiting substance signal transduction pathway in mammalian sexual differentiation. Genes Dev. 10:2577–2587. veloping gonads and Müllerian ducts (Figure 4). The Nakata, T., M. Ishiguro, N. Aduma, H. Izumi, and A. Kuroiwa. DMRT1 protein expression in the male gonadal somat- 2013. Chicken hemogen homolog is involved in the chicken-spe- ic cells was intense compared with the female gonadal cific sex-determining mechanism. Proc. Natl. Acad. Sci. USA somatic cells throughout embryonic development. Its 110:3417–3422. Raymond, C. S., J. R. Kettlewell, B. Hirsch, V. J. Bardwell, and strong expression was continuous in the male germ cells D. Zarkower. 1999. Expression of Dmrt1 in the genital ridge of after the sex determination of the gonads, whereas its mouse and chicken embryos suggests a role in vertebrate sexual at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 localization to the cortical germ cells of the left ovary development. Dev. Biol. 215:208–220. disappeared during embryonic development, and the Romanoff, A. L. 1960. The Avian Embryo. The Macmillan Com- pany, New York, NY. expression level of the male Müllerian ducts was high Smith, C. A., M. Katz, and A. H. Sinclair. 2003. DMRT1 is upregu- just before the ducts were degenerated. lated in the gonads during female-to-male sex reversal in ZW chicken embryos. Biol. Reprod. 68:560–570. Smith, C. A., K. N. Roeszler, J. Bowles, P. Koopman, and A. H. ACKNOWLEDGMENTS Sinclair. 2008. Onset of meiosis in the chicken embryo; Evidence of a role for retinoic acid. BMC Dev. Biol. 8:85. This work was supported in part by a Grant-In-Aid Smith, C. A., K. N. Roeszler, T. Ohnesorg, D. M. Cummins, P. for Scientific Research C (24590401) from the Ministry G. Farlie, T. J. Doran, and A. H. Sinclair. 2009. The avian Z- linked gene DMRT1 is required for male sex determination in the of Education, Culture, Sports, Science, and Technology chicken. Nature 461:267–271. of Japan (Tokyo; to N. Hoshi). Smith, C. A., and A. H. Sinclair. 2001. Sex determination in the chicken embryo. J. Exp. Zool. 290:691–699. Teng, C. S. 1987. Quantification of Müllerian inhibiting substance REFERENCES in developing chick gonads by a competitive enzyme-linked im- munosorbent assay. Dev. Biol. 123:255–263. Arango, N. A., A. Kobayashi, Y. Wang, S. P. Jamin, H.-H. Lee, G. Weniger, J. P. 1991. Embryonic sex hormones in birds. Int. J. Dev. D. Orvis, and R. R. Behringer. 2008. A mesenchymal perspective Biol. 35:1–7. of Müllerian duct differentiation and regression in Amhr2-lacZ Yamamoto, I., A. Tsukada, and N. Saito. 2003. Profiles of mRNA mice. Mol. Reprod. Dev. 75:1154–1162. expression of genes related to sex differentiation of the gonads in Chue, J., and C. A. Smith. 2011. Sex determination and sexual dif- the chicken embryo. Poult. Sci. 82:1462–1467. ferentiation in the avian model. FEBS J. 278:1027–1034. Yang, L. V., R. H. Nicholson, J. Kaplan, A. Galy, and L. Li. 2001. Clinton, M., L. Haines, B. Belloir, and D. McBride. 2001. Sexing Hemogen is a novel nuclear factor specifically expressed in mouse chick embryos: A rapid and simple protocol. Br. Poult. Sci. hematopoietic development and its human homologue EDAG 42:134–138. maps to chromosome 9q22, a region containing breakpoints of Guioli, S., R. Sekido, and R. Lovell-Badge. 2007. The origin of the hematological neoplasms. Mech. Dev. 104:105–111. Mullerian duct in chick and mouse. Dev. Biol. 302:389–398. Zhao, D., D. McBride, S. Nandi, H. A. McQueen, M. J. McGrew, Hamburger, V., and H. L. Hamilton. 1951. A series of normal stages P. M. Hocking, P. D. Lewis, H. M. Sang, and M. Clinton. 2010. in the development of the chick embryo. J. Morphol. 88:49–92. Somatic sex identity is cell autonomous in the chicken. Nature Jamin, S. P., N. A. Arango, Y. Mishina, M. C. Hanks, and R. R. 464:237–242. Behringer. 2002. Requirement of Bmpr1a for Müllerian duct re- PHYSIOLOGY, ENDOCRINOLOGY, AND REPRODUCTION

Effects of inorganic and organic manganese supplementation on gonadotropin-releasing hormone-I and follicle-stimulating hormone expression and reproductive performance of broiler breeder hens

Jingjing Xie, 1,2 Chuanhuan Tian ,1 Yongwen Zhu , Liyang Zhang , Lin Lu , and Xugang Luo

Institute of Animal Sciences, Chinese Academy of Agricultural Sciences, Beijing, P. R. China 100193 Downloaded from

ABSTRACT Manganese is an essential microelement. 240 mg of Mn/kg (P < 0.05). Furthermore, inorganic Manganese deficiency affects reproduction performance Mn supplementation doubled GnRH-I expression com- and reproductive hormones in layers. However, little pared with supplementation with the organic form (P is known about its effects and the possible mechanism < 0.05), although serum Mn was comparable between http://ps.oxfordjournals.org/ in regulating reproduction in broiler breeder hens. In these 2 supplements. No obvious difference was shown the current study, broiler breeder hens at peak produc- in gene expression of luteinizing hormone, prolactin, tion were fed with diets supplemented with 0, 120, or GnRH-I receptor, inducible NO synthase, and dopa- 240 mg of Mn/kg as MnSO4 or Mn proteinate for 13 mine receptor D1 in the pituitary as well as tyrosine wk. Manganese supplementation did not alter egg lay- hydroxylase and inducible NO synthase in the hypo- ing rate, egg weight, fertility, hatchability, or hatchling thalamus. This suggests that dietary Mn supplementa- weight over a 13-wk trial period. However, 240 mg of tion could improve eggshell quality in the long term.

Mn/kg supplementation significantly increased serum The central mechanism of nontoxic high doses of Mn at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Mn (P < 0.05). Manganese supplements increased the suggested the transcriptional activation of GnRH-I and eggshell breaking strength (P < 0.05) without affect- follicle-stimulating hormone genes. The central effect of ing the eggshell thickness. There was no difference in inorganic Mn activating GnRH-I genes compared with serum cholesterol and estradiol. Expression of follicle- the reduced effect by organic Mn could possibly be due stimulating hormone) and gonadotropin-releasing hor- to a decreased capacity of the latter passing through mone-I (GnRH-I) genes was significantly elevated by the blood-brain barrier. Key words: manganese , gene expression, gonadotropin-releasing hormone-I, follicle-stimulating hormone, broiler breeder 2014 Poultry Science 93 :959–969 http://dx.doi.org/ 10.3382/ps.2013-03598

INTRODUCTION prevent bone diseases and ensure the normal development of fast-growing broiler chickens. It has been shown The micronutrient Mn participates in many physi- by those studies that the organic forms of Mn supple- ological processes in animals and humans. Manganese ment had higher absorptivity and bioavailability than can be a cofactor of biological enzymes in the process of the inorganic form (Li et al., 2004; Ji et al., 2006; Bai et growth, carbohydrate and lipid metabolism, and clot- al., 2012). The bioavailability of organic Mn in broiler ting (Cao, 1987; Greger, 1999). Insufficient dietary Mn chicks fed with corn-soybean-based diets is closely as- leads to the malfunction of ovulation, the regression of sociated with the chelation strength between Mn and testis, and the early death of offspring (Cooper et al., the organic complex (Li et al., 2004, 2011). To date, few 1963; Offiong and Abed, 1980). In chickens, Mn needs studies have been conducted on the dietary effect of dif- to be supplemented in the diet to meet the nutrition re- ferent forms of Mn supplementation in broiler breeder quirements because of low Mn content in the diet ingre- hens with a special focus on reproduction. dients and the low absorption in the gut (Cao, 1987). Dietary Mn may function to regulate reproductive The requirement of Mn has been extensively studied to performance in several ways. As the cofactor of meval- onate kinase and farnesyl pyrophosphate synthase, Mn is speculated to regulate the synthesis of sexual hormone © 2014 Poultry Science Association Inc. precursor cholesterol and consequently to affect the re- Received September 3, 2013. production of animals. Furthermore, previous studies in Accepted December 16, 2013. 1 These authors contributed equally. layer hens have demonstrated the adverse effect of Mn 2 Corresponding author: [email protected] deficiency on egg production while decreasing circulat-

959 960 XIE ET AL. ing progesterone, estradiol, luteinizing hormone (LH), tions before the onset of the dietary treatments. The and follicle-stimulating hormone (FSH; Cao and Chen, photoperiodic lighting was programmed with light on 1987; Feng and Feng, 1998). Those results suggest that at 0800 h and off at 1600 h for the first 3 wk and gradu- dietary Mn insufficiency can affect the function of the ally extended to 2300 h to promote sexual maturity. At hypothalamic-pituitary-gonadal axis. the age of 30 wk, all birds were fed with a basal diet Manganese is known to be capable to pass through without Mn supplementation for 2 wk to deplete the the blood-brain barrier and deposit in the central ner- body’s deposits of Mn. Then the birds were subjected vous system, such as the cerebellum, hypothalamus, to the 13-wk dietary treatments. Eggs were collected on and pituitary (Rehnberg et al., 1981). Although there is a daily basis and laying rate was recorded every week. no evidence showing direct effects of Mn2+ on LH and On the last day of wk 4 and 13 of the trial, serial blood FSH secretion, a few studies have demonstrated that samples were withdrawn every 4 h from 1000 h to 0600 Downloaded from Mn2+ can act as the antagonist of Ca2+ to regulate h of the following day and serum was separated and the secretion of prolactin (PRL; Merritt and Brown, stored at −20°C. All birds were killed by an overdose 1984). In the brain, Mn2+ also can act on the soluble of anesthesia at the end of the trial, and heads were guanylyl cyclase (GC) to activate the protein kinase removed to collect samples of pituitary and hypothala- G pathway to control the release of gonadotropin-re- mus. Tissue samples were snap-frozen in liquid nitrogen http://ps.oxfordjournals.org/ leasing hormone-I (GnRH-I; Lee et al., 2007). The and then transferred to −80°C for storage. central toxic effect of exposures to high amount of Mn Artificial insemination was carried out twice every relates to the regenerative degradation of dopaminergic week after the first egg. Each hen was inseminated by neurons and the resulting decrease of dopamine synthe- 50 μL of semen collected from the same group of roost- sis and secretion (Guilarte, 2011; Rivera-Mancia et al., ers. Reproductive traits, such as laying rate, hatchabil- 2011). As known, dopamine is a potent regulator for ity, egg weight, eggshell quality, were determined at wk

pituitary PRL (Fitzgerald and Dinan, 2008), thereby 4 and wk 8 to 9 of the trial. Eggs were collected for 7 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 affecting FSH and LH secretion (Gregory et al., 2004; or 14 d, stored at 13°C, and incubated at 37.8°C and Henderson et al., 2008). It remains unknown whether 40 to 60% humidity for 18 d before transferred to the the function of dopaminergic neurons is also modified hatchery. Eggs were candled at embryonic d 6 and 18 by a nontoxic dose of Mn administered orally. to determine fertility. Each hatchling was weighed and Broilers are genetically selected for fast growth, and hatchling weight was recorded. broiler breeders have lower reproduction and more abnormalities in reproductive behaviors compared with Experimental Diets layers. To maintain and promote the production of broiler breeder hens, the effects of dietary nutrition Birds from each treatment were fed accordingly with need to be better understood. The current study is basal diets supplemented with 0 mg of Mn/kg (con- aimed at determining the effect of dietary supplemen- trol), 120 mg of Mn/kg as both MnSO4 (120 mg/kg of tation of Mn in both organic and inorganic forms on iMn) and Mn proteinate (120 mg/kg of oMn), and 240 the reproductive performance of broiler breeder hens. mg of Mn/kg as both MnSO4 (240 mg of iMn/kg) and Attempts were also made to reveal the possible central Mn proteinate (240 mg of oMn/kg), respectively. The mechanism via which Mn may function to affect repro- Mn proteinate was provided by a commercial company ductive performance. (Hebei Amino Acid Company, Hebei, China) and had a content of 10.22% Mn and a medium chelation strength MATERIALS AND METHODS of 61.91 (Qf). The ingredient and nutritional components of the basal diet are listed in Table 1. Contents of The procedure for current animal experiment was lysine and methionine were kept the same over different approved by the Animal Welfare Committee in the In- experimental diets by adding lysine and methionine to stitute of Animal Sciences, Chinese Academy of Agri- the basal diet and diets supplemented with MnSO4. All cultural Sciences. nutrients except Mn in each diet met the requirement of broiler breeder hens according to the Arbor Acres Bird Management and Sample Collection manufacturer’s instructions. A total of 120 female Arbor Acres broiler breeders at Eggshell Quality Assessment 16 wk of age were purchased from a commercial farm (Huadu Broiler Company, Beijing, China), randomly A total of 6 to 8 eggs from each replicate were used allocated to 5 treatments with 6 replicates per treat- to assess the quality of eggshells. The eggshell strength ment (n = 6), and kept in 3-story battery cages. The measurement was determined by Egg Shell Force Gauge temperature and humidity were maintained at 21°C (F0241, Robotmation Co. Ltd. , Tokyo, Japan), and and 40%, respectively. All birds were feed restricted the thickness of the eggshell was measured by TI-PVX with the same amount of commercial diet to maintain Egg Shell thickness Gauge of Orka Food Technology the growth curve according to manufacturer’s instruc- Corp. (Herzliya, Israel). EFFECTS OF MANGANESE ON REPRODUCTION 961 Table 1. Composition of ingredients and nutrient in corn-soy- China). Serum was first extracted by ether as described bean basal diet fed to broiler breeder hens from 30 to 44 wk of by (Xie et al., 2010) and the content of estradiol (E2) age was determined using the RIA of the Beijing North In- Item Value stitute of Biological Technology (Beijing, China). Ingredient (%) Corn 65.15 Isolation of RNA, Reverse Transcription, Soybean 23.44 Soybean oil 1.65 and Q-PCR CaHPO4 1.45 CaCO3 7.3 Total RNA was isolated and purified using the col- Salt 0.3

l-Lysine HCl 0.06 umn-based method (74104, RNeasy Mini Kit from Downloaded from Mineral and vitamin premix1 0.4 Qiagen, Hilden, Germany) and the manufacturer’s in- Cornstarch + Mn2 0.25 structions were followed. Because of the large volume of Nutrient ME (kcal/kg) 2,800 hypothalamus, the whole tissue sample was first homog- CP (%) 15.15 enized with 5 volumes of TRIzol (15596018, Life Tech- Calcium (%) 3.31

nologies, Foster City, CA). For each sample, 0.5 mL of http://ps.oxfordjournals.org/ Nonphytate phosphorus (%) 0.33 Lysine (%) 0.8 homogenate was then used for the following procedure Methionine (%) 0.32 of RNA extraction. The concentration of each isolated Manganese3 (mg/kg) 13.7 RNA sample was determined by NanoDrop Spectro- 1 Provided per kilogram of diet: vitamin A, 15,000 IU; vitamin D3, photomer (ND-1000, Gene Company Ltd., Wilmington, 4,500 IU; vitamin E, 36 IU; vitamin K3, 3.9 mg; vitamin B1, 4.5 mg; DE) and the integrity of RNA was checked by dena- vitamin B2, 10.5 mg; vitamin B6, 4.5 mg; vitamin B12, 0.024 mg; pantothenic acid calcium, 18 mg; niacin, 39 mg; folic acid, 1.5 mg; biotin, 0.18 tured RNA electrophoresis. mg; choline, 1,000 mg; Cu(CuSO4·5H2O), 8 mg; Fe(FeSO4·7H2O), 50 mg; A total of 1 μg of RNA was used to obtain cDNA by Zn(ZnSO4·7H2O), 100 mg; I(CaI2), 2 mg; Se(CaSe), 0.3 mg. reverse transcription using QuantiTech Reverse Tran- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 2 Manganese supplement added in place of equivalent weights of corn scription Kit (205311, Qiagen). To eliminate the con- starch. 3Analyzed composition. tamination of genomic DNA, RNA samples were first incubated with gDNA Wipeout Buffer at 42°C for 2 min. A mixture of reverse transcriptase, RT buffer, and Measurement of Serum Microelements Oligo dT primer mix was then added to each sample, followed by incubation at 42°C for 30 min. Reactions A volume of 1 mL of serum samples was digested in 3 were inactivated by a 3-min incubation at 95°C and mL of grade-pure nitric acid using a microwave-assisted cDNA was stored at −20°C until use. digestion method. After the digestion, samples were left Expression of genes coding for GnRH-I, tyrosine to cool and then diluted with ultrapure water to the hydroxylase (TH), inducible nitric oxide synthase final volume of 25 mL. Blank digestion was carried out (iNOS) in the hypothalamus and genes coding for with the same procedure by using 1 mL of ultrapure FSH, LH, PRL, GnRH-I receptor (GnRH-R), dopa- water instead of serum sample. Manganese content in mine receptor D1 (DRD1), iNOS in the pituitary were each digested sample was then analyzed by graphite quantified by real-time PCR using SYBR green dye. furnace atomic absorption spectrometry (Hitachi Po- Primers (Invitrogen, Beijing, China) used in the PCR larized Zeeman atomic absorption spectrophotometer reactions were designed with the Primer-BLAST in Na- Z-2000 series, Tokyo, Japan) with a procedure as fol- tional Center for Biotechnology Information services. lows: dried at 140°C for 40 s, ashed at 1,000°C for 20 To determine the amplification efficiency for each pair s, atomized at 2,300°C for 4 s, and cleaned at 2,500°C of primers, 5-fold serial dilutions of a cDNA template for 4 s. Ammonium dihydrogen phosphate (2%) was were analyzed by real-time PCR. The efficiency was used as the chemical modifier and coinjected with each calculated as the slope of the linear regression of plot- digested sample placed into the furnace. The certified ting Ct verse log2-transformed template dilution. The MnSO4 stock standard solution (obtained from Nation- target genes, primers, and amplification efficiency of al Research Center for Certified Reference Materials, each gene are given in Table 2. All samples were ar- Beijing, China) was serially diluted with 2% nitric acid ranged in each plate to ensure the same reaction con- to a final concentration of 0, 2, 4, or 10 μg of Mn/L to dition. The PCR reactions (20 μL) contained 1 μL of generate the standard curve, and the concentration of diluted cDNA, 250 nM each primer, and PCR buffer Mn in each digested sample was calculated accordingly. master mix from the Power SYBR Green Master Mix (4367659, Life Technologies) and were carried out in Determination of Serum Cholesterol the ABI 7500 Real-Time PCR Detection System (Ap- and Estradiol plied Biosystems, Foster City, CA). Each PCR reaction was conducted in triplicate, and the Ct value used in Total cholesterol in the serum was measured with subsequent calculations was the mean of the triplicate a commercial kit based upon the enzymatic method reactions. The protocol of PCR was as follows: dena- (Nanjing Jiancheng Bioengineering Institute, Nanjing, turation at 95°C for 2 min followed by 40 cycles of 95°C 962 XIE ET AL.

Table 2. Information of target genes and primers

Product Temperature Amplification Gene1 ID Primer set2 (5c–3c) size (bp) (°C) efficiency (%) β-Actin NM205518.1 F: 5c-ACCTGAGCGCAAGTACTCTGTCT-3c 95 60 106 R: 5c-CATCGTACTCCTGCTTGCTGAT-3c GAPDH NM204305.1 F: 5c-CTTTGGCATTGTGGAGGGTC-3c 128 60 96 R: 5c-ACGCTGGGATGATGTTCTGG-3c FSH-β NM204257.1 F: 5c-AGCAGTGGAAAGAGAAGAATGTGA-3c 151 60 99 R: 5c-TGTTTCATACACAACCTCCTTGAAG-3c LH-β HQ872606 F: 5c-GCTGATGACCCTTTTGGGGA-3c 147 60 90 R: 5c-ACCGCCACCGTTACGTTTAT-3c

PRL AB011438 F: 5c-GCTGTTAATGGCTGCCACAC-3c 120 60 94 Downloaded from R: 5c-ATCATTCCAGGAACGCAGCA-3c DRD1 NM001144848 F: 5c-TGGAACGACACCACTATGGA-3c 153 60 96 R: 5c-GAGGTGGCGAAACCTAATGA-3c iNOS NM204961.1 F: 5c-CCAGCTGATTGGGTGTGGAT-3c 145 60 105 R: 5c-CCTACGGGTCTCATCATGCC-3c GnRH-R NM204653 F: 5c-TCCGAATCATTTGGGAGATCAG-3c 123 60 97 R: 5c-TGGCAACAATCACAATGGTCA-3c http://ps.oxfordjournals.org/ GnRH-I NM001080877.1 F: 5c-ACACTGGTCTTATGGCCTGCA-3c 116 60 103 R: 5c-ATTCAGCCTTCTGCCCTTCTC-3c TH AJ251387.1 F: 5c-CCAGAGGAGGAAATCAATAGCA-3c 142 60 101 R: 5c-GGCATGGGTTGGATACAGG-3c 1GAPDH, glyceraldehyde-3-phosphate dehydrogenase; FSH, follicle-stimulating hormone; LH, luteinizing hormone; PRL, prolactin; DRD1, dopamine receptor D1; iNOS, inducible nitric oxide synthase; GnRH-R, gonadotropin-releasing hormone-I receptor; TH, tyrosine hydroxylase. 2F, forward; R, reverse. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 for 60 s, 60°C for 30 s, and 72°C for 30 s. Dissociation ments within birds; εijk is the random error and the curve analyses were run to ensure a single product of variance between measures within birds. each reaction. For each reaction plate of the same tar- Specific orthogonal contrasts were carried out to get gene, the cDNA pool of all samples was used as the stress the effects of Mn supplementation (control vs. reference control sample. The geometric mean of inter- 120 mg of iMn/kg, 120 mg of oMn/kg, 240 mg of iMn/ nal references, β-actin and glyceraldehyde 3-phosphate kg, 240 mg of oMn/kg), levels of dietary Mn (120 mg dehydrogenase, was used to normalize the expression of of iMn/kg, 120 mg of oMn/kg vs. 240 mg of iMn/kg, the targeted genes (Vandesompele et al., 2002). The 2− 240 mg of oMn/kg), and forms of Mn supplementation ΔΔCt was used to calculate mRNA level of each target (120 mg of iMn/kg, 240 mg of iMn/kg and 120 mg of gene, where the average mean of Ct from the group fed oMn/kg, 240 mg of oMn/kg). The significance of cor- with the basal diet was used as the calibrator. relations between expression of mRNA and levels of serum hormone was determined using Pearson correla- Statistical Analyses tion analyses. All data were presented as mean ± SE with significance considered at the level of P < 0.05. All data were analyzed using JMP 10.2 software (SAS Institute Inc., Cary, NC). Data were first subjected to homogeneity of variance analyses. One-way ANOVA RESULTS was used to test the main effect on relative gene expression with log-transformed data. Data collected at wk Egg Production, Egg Quality, 4 and 13 were considered as repeated measurements. and Hatchability Thus the main effect of dietary treatment on laying rate, fertility, hatchability, egg weight, eggshell force, At the age of 35 wk, the average laying rate of broil- eggshell thickness, hatchling weight, serum cholesterol, er breeder hens reached 78.9 ± 1.5% and dropped re- and E2 were determined using the restricted maximum markably to 66.4 ± 2.0% at the end of the trial (F1,25 likelihood method for ANOVA with the following sta- = 71.733, P < 0.01), whereas average egg weights in- tistical model: creased from 54.0 ± 0.5 g to 66.4 ± 0.4 g (F1,25 = 810.047, P < 0.01). There were no significant differ- Yijk = μ + τi + δij + tk + (τ*t)ik + εijk, ences in average laying rate (Figure 1A, F4,36 = 0.521, P = 0.721) and average egg weight (Figure 1B, F4,37 (i = 1, 2, 3; j = 1, …, 30; k = 1, 2), where Yijk is the = 1.563, P = 0.204) among dietary treatment groups. observation ijk; μ is the overall mean; τi is the effect of Manganese supplementation had a significant effect dietary Mn treatment; tk is the effect of 2 observation on eggshell strength (Figure 1C, F4,44 = 3.556, P = times (wk 4 and 13); (τ*t)ik is the interaction between 0.0027); however, eggshell thickness of sampled eggs treatment and observation time; δij is the random error did not differ among treatments (Figure 1D, F4,38 = and equal to the covariance between repeated measure- 0.913, P = 0.466). EFFECTS OF MANGANESE ON REPRODUCTION 963 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 1. Laying performance and egg quality of broiler breeder hens after dietary supplementation of Mn as MnSO4 (iMn) or Mn proteinate (oMn) for 4 and 13 wk. A. Laying rate on a weekly basis; B. average weight of eggs collected at wk 4 and 13 of the trial; C. eggshell strength, determined by 6 to 8 eggs per replicate; D. eggshell thickness, determined by 6 to 8 eggs from each replicate. All data were presented as mean ± SE. *An asterisk indicates significant difference between the Mn-supplemented group and the control (CON) basal diet, which was determined by post hoc contrast analyses. Significant effects were considered when P < 0.05.

The fertility of eggs after artificial insemination re- dietary supplementation of Mn had a tendency to el- mained at around 98%. There was no indication that evate the serum Mn compared with the control basal the fertility of eggs was affected by dietary treatments diet (P = 0.057). Birds fed with 240 mg of Mn/kg had (Table 3, F4,49 = 0.504, P = 0.733). Mean hatchability significantly higher serum Mn than those fed with 0 or of fertilized eggs varied among treatment groups from 120 mg of Mn/kg (Figure 2B, P < 0.05). There was no 91.7 ± 1.4% to 88.5 ± 1.6%, but these differences were difference in serum Mn between inorganic and organic not statistically significant (Table 3, F4,48 = 0.411, P = Mn supplementation groups (Figure 2C, P = 0.137). 0.800). Hatchling weight correlated with egg weight (P Dietary treatments with different Mn supplements < 0.001, r2 = 0.40), but did change after Mn supple- did not change serum cholesterol content (Figure 3, mentation (Table 3, F4,36 = 0.5821, P = 0.678). F4,48 = 0.204, P = 0.935). Serum E2 displayed diurnal changes with the bottom level after 1800 h and the peak level at 1000 h (Figure 4B), and the circulating Serum Mn, Cholesterol, and E2 level dropped remarkably as the birds aged (F1,288 = There was a significant dietary treatment effect on 70.398, P < 0.001). The average content of E2 of all 6 the serum Mn content (Figure 2A, F4,28 = 3.081, P = bleeds did not differ among treatment groups (Figure 0.035). It was revealed by contrast analyses that the 4A, F4,199 = 0.926, P = 0.450). 964 XIE ET AL.

Table 3. Fertility, hatchability, and hatchling weight of eggs collected from wk 4 and wk 8 to 9 of the trial1

Wk 4 Wk 8 to 9

Fertility Hatchability Hatchling weight Fertility Hatchability Hatchling weight Item (%) (%) (g) (%) (%) (g) Control 98.3 ± 1.1 94.3 ± 2.0 43.0 ± 1.0 98.6 ± 0.7 87.1 ± 4.2 46.5 ± 0.7 120 mg of iMn/kg 98.0 ± 1.2 92.0 ± 3.4 44.4 ± 1.5 97.7 ± 1.0 92.5 ± 3.9 46.3 ± 1.3 120 mg of oMn/kg 98.4 ± 1.0 90.4 ± 2.6 42.3 ± 0.8 97.0 ± 0.8 86.6 ± 3.5 46.0 ± 1.7 240 mg of iMn/kg 97.4 ± 2.6 93.6 ± 2.2 43.3 ± 0.8 99.3 ± 0.5 90.0 ± 4.2 45.3 ± 0.9 240 mg of oMn/kg 95.7 ± 3.5 88.8 ± 5.3 44.1 ± 0.7 97.7 ± 0.8 87.6 ± 2.8 47.4 ± 1.1 1iMn, Mn supplement as inorganic form, MnSO ; oMn, Mn supplement as Mn proteinate.

4 Downloaded from

Relative Abundance of Gene mRNA sults suggest that there might be a central mechanism in the Pituitary and Hypothalamus for dietary Mn in affecting the reproduction of broiler breeder hens, and it may take a longer time to show In the pituitary, only in the FSH mRNA abundance these effects in reproductive behavior. http://ps.oxfordjournals.org/ was the significant main effect of dietary treatment ob- In the current study, the reproductive performance served (Figure 5A, F4,24 = 3.2651, P = 0.033). Contrast and behavior of broiler breeders, such as egg produc- analyses showed that birds fed with diets supplemented tion and hatchability, were not altered by dietary Mn with 240 mg of Mn/kg had higher FSH gene expression deficiency, although previous studies confirmed that than birds fed with basal diets and experimental diets low dietary Mn negatively affected egg productions in with 120 mg of Mn/kg supplement (insert of Figure birds. Attia et al. (2010) found that supplementation of

5A, P < 0.001). No difference was observed between 10 to 40 mg of Mn/kg to the corn-soybean meal signifi- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 inorganic and organic Mn supplements (P = 0.791). cantly increased the laying rate and egg weight in In- Other genes in the pituitary, including PRL (Figure cha hens (Siena X Plymouth Rock) after 16 wk. Layer 5B, F4,24 = 1.0529, P = 0.404), LH (Figure 5C, F4,24 hens fed with a Mn-deficient diet from 1 d old had re- = 1.275, P = 0.313), DRD1 (Figure 5D, F4,24 = 0.485, duced laying rate, hatchability, eggshell thickness, and P = 0.746), GnRH-R (Figure 5E, F4,27 = 0.567, P = force, and increased broken egg rate and soft egg rate 0.689), and iNOS (Figure 5F, F4,26 = 1.081, P = 0.390) at peak egg production (Zhang et al., 1992). A study did not express differently among treatment groups. using Lohmann layers at the peak production did not Correlation analyses showed that mRNA abundance of find an effect of Mn supplementation on laying rate and PRL positively correlated with GnRH-R (P = 0.022, r2 egg production for 14-wk dietary treatments (Chen and = 0.227), whereas FSH negatively correlated to DRD1 Wu, 2002), whereas another trial using the same layers gene expression (P = 0.012, r2 = 0.235). showed a decrease in egg production after exposure to In the hypothalamus, relative expression of 3 genes Mn-unsupplemented diets for 32 wk (Feng and Feng, coding for GnRH-I, TH, and iNOS was quantified. 1998). The inconsistency in the behavioral effects of There was a significant effect of dietary treatment on Mn among those studies may be due to the differences GnRH-I mRNA abundance (Figure 6A, F4,25 = 4.684, in chicken breeds, the duration of dietary treatments, P = 0.007). Manganese supplements significantly in- as well as Mn contents in feed ingredients. In present creased GnRH-I gene expression in the brain. Contrast study, there was no difference in the serum Mn content analyses also showed that 240 mg of Mn/kg signifi- between birds receiving diets supplemented with 0 and cantly induced GnRH-I mRNA expression in the brain 120 mg of Mn/kg. The eggshell strength, however, was (Figure 6B, P < 0.05). Different forms of Mn exhib- lowered by the Mn-deficient diet treatment. Previous ited very different effects on GnRH-I mRNA expres- studies found that the ultrastructure of eggshell could sion with a 2-fold GnRH-I mRNA increase with the be altered by a Mn-deficient diet, where large irregular inorganic Mn supplement over the organic form (P = mammillary knobs were observed (Leach and Gross, 0.016). Abundance of TH mRNA (Figure 6C, F4,28 = 1983; Luo et al., 2003). Hexosamine and hexuronic acid 0.696, P = 0.602) or iNOS (Figure 6D, F4,28 = 0.587, P contents in the eggshell were also decreased by low Mn = 0.675) was not affected by dietary treatments. intake (Leach and Gross, 1983). It is likely that the homeostasis of serum Mn was maintained by redirecting DISCUSSION Mn deposition in the body, such as eggs. Manganese deficiency was found to lead to changes Over the 13 wk of trial, there were no significant in circulating ovarian steroids in layer hens (Cao and differences observed in egg production, hatchability, Chen, 1987; Feng and Feng, 1998). Steroidogenesis be- hatchling weight, or serum cholesterol despite the in- gins with cholesterol. Because Mn is one of the cofac- creased serum Mn after dietary supplementation of tors of enzymes in the biosynthesis of cholesterol (Kli- Mn. Dietary Mn supplementation, however, did affect mis-Tavantzis et al., 1983b), it was speculated to also gene expression of GnRH-I in the brain and FSH in regulate the steroid synthesis. Although an increased the pituitary and as well as eggshell quality. These re- serum Mn was found in the birds supplemented with EFFECTS OF MANGANESE ON REPRODUCTION 965 Downloaded from http://ps.oxfordjournals.org/

Figure 3. Serum cholesterol in broiler breeder hens fed with diets supplemented with Mn as MnSO4 (iMn) or Mn proteinate (oMn) for 4 and 13 wk. All data were presented as mean ± SE. CON = control. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 2. Serum Mn of broiler breeder hens after dietary supplementation of Mn as MnSO4 (iMn) or Mn proteinate (oMn) for 13 wk. A. Average content of Mn in serum from birds fed with diets supplemented with 0, 120, or 240 mg of Mn/kg as MnSO4 or Mn proteinate; B. Average serum Mn of birds fed with diets supplemented with 0, 100, or 200 mg of Mn/kg; C. Serum Mn of birds fed with diets with inorganic or organic Mn supplements. All data were presented as mean ± SE. Different letters (A,B) indicate significant differences among each level of Mn supplementation, which was determined by contrast analyses. Significant effects were considered when P < 0.05. CON = control.

Figure 4. Effects of dietary Mn supplements as MnSO4 (iMn) or Mn proteinate (oMn) on serum estradiol (E2) in broiler breeder hens. 240 mg of Mn/kg, there were no differences in serum A. Average serum E2 in birds fed with 0, 120, and 240 mg of Mn/ cholesterol and E2 among all treatments of present kg. B. Diurnal pattern of serum E2 in broiler breeder hens. All data study. This finding was consistent with previous stud- were presented as mean ± SE. Different letters indicate significant differences among time points (uppercase letters for wk 4, lowercase ies, which showed that Mn did not affect the hepatic for wk 13). Significant effects were considered when P < 0.05. CON cholesterol and fatty acid synthesis and circulating cho- = control. 966 XIE ET AL. Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 5. Gene expression of follicle-stimulating hormone (FSH; A), prolactin (PRL, B), luteinizing hormone (LH; C), dopamine receptor D1 (DRD1, D), GnRH-R (GnRH receptor, E), and iNOS (inducible nitric oxide synthase, F) in the pituitary of broiler breeder hens after dietary Mn treatments for 13 wk. The gene expression was quantified by real-time PCR using geometric mean of β-actin and glyceraldehyde 3-phosphate dehydrogenase as the internal standard. Insert in the FSH (A) showed the effect of different levels of Mn supplementation on the abundance of FSH mRNA. All data are presented as mean ± SE. Different uppercase letters (A,B) indicate significant differences among each level of Mn supplementation. Significant effects were considered when P < 0.05. iMn = dietary supplementation of Mn as MnSO4; oMn = dietary supplementation with Mn proteinate; CON = control. lesterol in rats (Klimis-Tavantzis et al., 1983b), hens The comparable levels of LH among each treatment, (Klimis-Tavantzis et al., 1983a), and roosters (Cao et to a certain extent, supported the unaltered serum E2. al., 1990) in vivo. In avian species, the steroidogenesis We found that dietary supplementation of 240 mg of in follicles is mainly regulated by LH, whereas FSH Mn/kg significantly increased the serum Mn content stimulates the maturation of granulosa cells and at- as well as pituitary FSH mRNA abundance, but not tained ability of cells to respond to LH, FSH, and other LH mRNA. The unchanged LH mRNA expression af- signals (Scanes, 1999). Due to a lack of antibodies, cir- ter 13-wk dietary treatments suggested that it might culating LH was not directly determined in the current take a longer time for dietary Mn to have a signifi- study. Analysis of LH mRNA expression did not show cant impact on ovulation and egg production in broiler Mn to influence LH after a 13-wk dietary treatment. breeders. The upregulated FSH mRNA expression in EFFECTS OF MANGANESE ON REPRODUCTION 967 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 6. Gene expression of gonadotropin-releasing hormone-I (GnRH-I; A), tyrosine hydroxylase (TH, D), and inducible nitric oxide (NO) synthase (E) in the hypothalamus of broiler breeder hens fed with 0, 120, or 240 mg of Mn/kg supplements as MnSO4 (iMn) or Mn proteinate (oMn) for 13 wk. The gene expression was quantified by real-time PCR using geometric mean of β-actin and glyceraldehyde 3-phosphate dehydrogenase as the internal standard. B. The effects of 0, 120, and 240 mg of Mn/kg supplementation on the abundance of GnRH-I mRNA. C. Abundance of GnRH-I mRNA in birds fed with diets with inorganic or organic Mn supplements. All data are presented as mean ± SE. Different uppercase letters (A,B) indicate significant differences among each level of Mn supplementation. *An asterisk indicates significant difference between inorganic and organic supplements. Significant effects were considered when P < 0.05. CON = control. the high-Mn dietary treatment could be induced by at the pituitary level. Third, the regulatory effect of several mechanisms. First, Mn could directly target Mn could depend on the GnRH-I, the primary regula- the FSH-producing cells. Although there was no direct tor of the hypothalamic-pituitary-gonadal axis. Man- evidence, it has been shown that Mn can regulate the ganese is capable of passing through the blood-brain release of PRL (Merritt and Brown, 1984) and also can barrier, which makes it possible for Mn to regulate the activate MnSOD expression at the transcriptional level synthesis and secretion of GnRH-I in the brain. Previ- (Li et al., 2011). Second, Mn could regulate FSH syn- ous work in both male and female rats showed that thesis and secretion by affecting dopamine system. It injecting MnCl2 into the brain could induced LH, FSH, is known that excess Mn exposure causes degenerative and E2 secretion (Pine et al., 2005; Lee et al., 2006; changes in dopaminergic neurons (Guilarte, 2011). In Prestifilippo et al., 2008). The central regulatory effect the current study, the nontoxic high doses of Mn had of Mn depends on the release of GnRH-I (Pine et al., few effect on mRNA expression of tyrosine hydroxylase 2005). In addition to FSH, GnRH-I gene expression in (rate-limited enzyme in dopamine synthesis and com- the brain also exhibited a dose-independent increment monly used as the marker for dopaminergic neurons) from 0 to 240 mg of Mn/kg in our study. The tran- in the hypothalamus of broiler breeder hens. In the pi- scriptionally increased GnRH-I expression suggested tuitary, FSH was shown to be negatively correlated to that the central mechanism was involved in the regu- DRD1 gene expression, indicating that 240 mg of Mn/ lation of Mn on reproductive performance of broiler kg interacted with the dopamine system most likely breeder hens. 968 XIE ET AL. Lee et al. (2007) found that Mn regulated GnRH-I re- 2012cj-1), and Key International Cooperation Program lease depending on the GC/protein kinase G pathway. of the National Natural Science Foundation of China to It functions at the soluble GC level but not at the pro- X. Luo (No. 31110103916). duction of nitric oxide (NO; Lee et al., 2007). Although we did not directly analyze the NO in the hypothalamus because of technical issues, the analysis of NO syn- REFERENCES thase mRNA expression revealed no difference in iNOS Attia, Y. A., E. M. Qota, F. Bovera, A. E. Tag El-Din, and S. mRNA expression among dietary treatments with dif- A. Mansour. 2010. Effect of amount and source of manganese ferent levels of Mn. It suggested that the NO pathway and/or phytase supplementation on productive and reproductive may not be activated by Mn supplementation. Dietary performance and some physiological traits of dual purpose cross- Mn supplementation increased the expression of both bred hens in the tropics. Br. Poult. Sci. 51:235–245. http:// Downloaded from dx.doi.org/10.1080/00071661003786111. GnRH-I and FSH genes; however, no correlations were Bai, S. P., L. Lu, R. L. Wang, L. Xi, L. Y. Zhang, and X. G. Luo. found between the amounts of RNA encoding GnRH-I, 2012. Manganese source affects manganese transport and gene GnRH-R, and FSH. In the pituitary, GnRH-R locates expression of divalent metal transporter 1 in the small intestine not only at gonadotropin-expression cells, but also at of broilers. Br. J. Nutr. 108:267–276. http://dx.doi.org/10.1017/

S0007114511005629. http://ps.oxfordjournals.org/ somatotropin-, somatolactin-, and PRL-expressing cells Cao, S. F. 1987. Manganese and growth and reproduction of animals (Stefano et al., 1999). This makes the regulatory effect and poultry. Journal of Shanghai Agricultural College 5:79–85. of GnRH-I in the pituitary more complicated. Cao, S. F., and L. J. Chen. 1987. Effects of manganese of the con- Comparisons between inorganic and organic Mn sup- centrations of plasma, LH, estrogen and progesterone in white ear hens. Journal of Shanghai Agricultural College 5:109–116. plementation showed that the supplementation of Mn Cao, S. F., C. L. Ye, X. Y. Chen, G. G. Chen, and Y. Y. Gu. 1990. as MnSO4 had a more profound stimulatory effect on The effect of manganese on the level of cholesterol and the activ- gene expression of GnRH-I than supplemental Mn pro- ity of adenosine triphosphatase in sud-cocks. Journal of Shang- hai Agricultural College 8:291–296. teinate. Extensive studies in broiler chickens found that Chen, H. Q., and J. Q. Wu. 2002. Effect of manganese levels and at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 the absorption of organic forms of Mn was higher than sources on performance of laying hens. Journal of Anhui Agri- that of inorganic forms in the intestine (Ji et al., 2006) cultural University 29:44–48. and the organic form of Mn with medium chelation Cooper, D. M., L. G. Chubb, and J. G. Rowell. 1963. The effect of manganese on reproduction in the domestic fowl. Br. Poult. Sci. strength (Qf of 10–100) had the highest bioavailability 4:83–90. http://dx.doi.org/10.1080/00071666308415484. reflecting by higher expression of heart MnSOD mRNA Feng, J., and Z. G. Feng. 1998. Effect of Mn-deficiency on repro- and Mn intestinal transporter DMT1 mRNA (Li et al., ductive performance in egg-laying chickens. Acta Veterinaria et 2004). In contrast to the different bioavailability be- Zootechnica Sinica 29:499–505. Fitzgerald, P., and T. G. Dinan. 2008. Prolactin and dopamine: tween inorganic and organic Mn, serum Mn contents What is the connection? A review article. J. Psychopharmacol. were found to be comparable both in previous studies 22:12–19. http://dx.doi.org/10.1177/0269216307087148. and the current one. Because Mn needs to pass through Greger, J. L. 1999. Nutrition versus toxicology of manganese in the blood-brain barrier to target brain neurons, the or- humans: Evaluation of potential biomarkers. Neurotoxicology 20:205–212. ganic source of Mn may not favor the permeability of Gregory, S. J., J. Townsend, A. S. McNeilly, and D. J. Tortonese. Mn ions because of its complex structure. Therefore, re- 2004. Effects of prolactin on the luteinizing hormone response garding central stimulation of GnRH-I mRNA by Mn, to gonadotropin-releasing hormone in primary pituitary cell cul- the inorganic form performed better than the organic tures during the ovine annual reproductive cycle. Biol. Reprod. 70:1299–1305. http://dx.doi.org/10.1095/biolreprod.103.022806. form with moderate chelation strength, although it had Guilarte, T. R. 2011. Manganese and Parkinson’s disease: A criti- lower bioavailability. cal review and new findings. Cien. Saude Colet. 16:4549–4566. In conclusion, dietary supplementation of Mn could http://dx.doi.org/10.1289/ehp.0901748. Henderson, H. L., J. Townsend, and D. J. Tortonese. 2008. Direct improve the eggshell quality in the long-term. The cen- effects of prolactin and dopamine on the gonadotroph response to tral mechanism of nontoxic dose of Mn might employ GnRH. J. Endocrinol. 197:343–350. http://dx.doi.org/10.1677/ the transcriptional activation of hypothalamic GnRH-I JOE-07-0536. gene and pituitary FSH gene, which may not be medi- Ji, F., X. G. Luo, L. Lu, B. Liu, and S. X. Yu. 2006. Effect of manganese source on manganese absorption by the intestine of broilers. ated by dopamine or NO pathway. The inorganic form Poult. Sci. 85:1947–1952. of Mn supplement showed a higher stimulatory effect Klimis-Tavantzis, D. J., P. M. Kris-Etherton, and R. M. Leach Jr.. of the GnRH-I mRNA than the Mn proteinate despite 1983a. The effect of dietary manganese deficiency on cholesterol similar serum Mn contents. and lipid metabolism in the estrogen-treated chicken and the laying hen. J. Nutr. 113:320–327. Klimis-Tavantzis, D. J., R. M. Leach Jr., and P. M. Kris-Etherton. ACKNOWLEDGMENTS 1983b. The effect of dietary manganese deficiency on cholesterol and lipid metabolism in the Wistar rat and in the genetically The authors thank Li Tang and Linyan Zhang for hypercholesterolemic RICO rat. J. Nutr. 113:328–336. Leach, R. M. Jr., and J. R. Gross. 1983. The effect of manganese the help in managing hens and sample collection. The deficiency upon the ultrastructure of the eggshell. Poult. Sci. authors also appreciate the professional assistance from 62:499–504. Xueyin Qu in quantifying serum Mn. This study was Lee, B., J. K. Hiney, M. D. Pine, V. K. Srivastava, and W. L. Dees. funded by the National Natural Science Foundation of 2007. Manganese stimulates luteinizing hormone releasing hormone secretion in prepubertal female rats: Hypothalamic site and China to J. Xie (Grant no. 31201818), National Non- mechanism of action. J. Physiol. 578:765–772. http://dx.doi. profit Institute Research Grant to J. Xie (Grant no. org/10.1113/jphysiol.2006.123083. EFFECTS OF MANGANESE ON REPRODUCTION 969

Lee, B., M. Pine, L. Johnson, V. Rettori, J. K. Hiney, and W. L. of specific neurotransmitter systems. Toxicol. Sci. 105:295–302. Dees. 2006. Manganese acts centrally to activate reproductive http://dx.doi.org/10.1093/toxsci/kfn135. hormone secretion and pubertal development in male rats. Re- Rehnberg, G. L., J. F. Hein, S. D. Carter, R. S. Linko, and J. W. Las- prod. Toxicol. 22:580–585. http://dx.doi.org/10.1016/j.repro- key. 1981. Chronic ingestion of Mn3O4 by young rats: tissue accu- tox.2006.03.011. mulation, distribution, and depletion. J. Toxicol. Environ. Health Li, S., L. Lu, S. Hao, Y. Wang, L. Zhang, S. Liu, B. Liu, K. Li, and 7:263–272. http://dx.doi.org/10.1080/15287398109529977. X. Luo. 2011. Dietary manganese modulates expression of the Rivera-Mancia, S., C. Rios, and S. Montes. 2011. Manganese ac- manganese-containing superoxide dismutase gene in chickens. J. cumulation in the CNS and associated pathologies. Biometals Nutr. 141:189–194. http://dx.doi.org/10.3945/jn.110.126680. 24:811–825. http://dx.doi.org/10.1007/s10534-011-9454-1. Li, S., X. Luo, B. Liu, T. D. Crenshaw, X. Kuang, G. Shao, and S. Scanes, C. G. 1999. Introduction to endocrinology: Pituitary gland. Yu. 2004. Use of chemical characteristics to predict the relative Pages 437–443 in Sturkie’s Avian Physiology. 5th ed. G. C. Whit- bioavailability of supplemental organic manganese sources for tow, ed. Academic Press, San Diego, CA, and London, UK.

broilers. J. Anim. Sci. 82:2352–2363. Stefano, A. V., P. G. Vissio, D. A. Paz, G. M. Somoza, M. C. Downloaded from Luo, X. G., X. Q. Guo, B. Liu, X. C. Zhang, G. Z. Shao, S. F. Li, Maggese, and G. E. Barrantes. 1999. Colocalization of GnRH and S. X. Yu. 2003. Study on an optimal manganese level in a binding sites with gonadotropin-, somatotropin-, somatolactin-, practical diet for brown-egg layers. Chinese Journal of Animal and prolactin-expressing pituitary cells of the pejerrey, Odontes- and Veterinary Sciences 34:105–111. thes bonariensis, in vitro. Gen. Comp. Endocrinol. 116:133–139. Merritt, J. E., and B. L. Brown. 1984. Dual effects of manganese on http://dx.doi.org/10.1006/gcen.1999.7354. prolactin secretion. Cell Calcium 5:159–165. Vandesompele, J., K. De Preter, F. Pattyn, B. Poppe, N. Van Roy, A. Offiong, S. A., and S. M. Abed. 1980. Fertility, hatchability and De Paepe, and F. Speleman. 2002. Accurate normalization of real- http://ps.oxfordjournals.org/ malformations in guinea fowl embryos as affected by di- time quantitative RT-PCR data by geometric averaging of mul- etary manganese. Br. Poult. Sci. 21:371–375. http://dx.doi. tiple internal control genes. Genome Biol. 3:RESEARCH0034. org/10.1080/00071668008416683. http://dx.doi.org/10.1186/gb-2002-3-7-research0034. Pine, M., B. Lee, R. Dearth, J. K. Hiney, and W. L. Dees. 2005. Xie, J., W. J. Kuenzel, N. B. Anthony, and A. Jurkevich. 2010. Manganese acts centrally to stimulate luteinizing hormone secre- Subpallial and hypothalamic areas activated following sexual and tion: A potential influence on female pubertal development. Tox- agonistic encounters in male chickens. Physiol. Behav. 101:344– icol. Sci. 85:880–885. http://dx.doi.org/10.1093/toxsci/kfi134. 359. http://dx.doi.org/10.1016/j.physbeh.2010.06.004. Prestifilippo, J. P., J. Fermandez-Solari, A. De Laurentiis, C. E. Zhang, J. Y., S. Q. Wang, Z. G. Zhao, W. Y. Wang, and W. G. Qiu.

Mohn, C. de la Cal, R. Reynoso, W. L. Dees, and V. Rettori. 1992. A study of the effects of dietary manganese upon tissue at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 2008. Acute effect of manganese on hypothalamic luteinizing hor- manganese contents and performances in chicken. Acta Zoonu- mone releasing hormone secretion in adult male rats: Involvement trimenta Sinica 4:32–40.

Effects of genetic strain and light intensity on blood physiological variables of broilers grown to heavy weights 1

H. A. Olanrewaju ,2 J. L. Purswell , S. D. Collier , and S. L. Branton

USDA, Agricultural Research Service, Poultry Research Unit, PO Box 5367, Mississippi State 39762-5367

ABSTRACT The effects of genetic strain, light intensi- intensity was observed on all examined variables. In Downloaded from ty, and their interaction were examined on blood physi- addition, no significant main effects were observed for ological variables of broilers maintained in environmen- strain, light intensity, or their interaction on saturated + tally controlled rooms in each of 5 trials. The study O2, hematocrit, hemoglobin, K , mean corpuscular he- consisted of a 2 × 5 factorial arranged in a randomized moglobin concentration, glucose, osmolality, and anion complete block design with 10 treatments of 2 strains gap. Physiologically, although Ross × Ross 308 had a http://ps.oxfordjournals.org/ (Ross × Ross 308 and 708) and exposure to 5 light in- significantly increased BW in comparison with Ross × tensities (25, 10, 5, 2.5, and 0.2 lx) with chicks equally Ross 708, Ross × Ross 708 showed a significant (P ≤ and randomly distributed at 1 d of age. The 5 light 0.05) increase in partial pressure of O2 and triiodothy- intensities were randomly assigned from 22 to 56 d of ronine along with lower blood partial pressure of CO2 age. Feed and water were provided ad libitum. Venous and thyroxine. The results of this study supplement blood samples were collected on d 21 (base line), 28, current knowledge of physiological response to differing

42, and 56 of age. Main effects indicate that Ross × lighting levels and may be used to establish the normal at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Ross 308 significantly (P ≤ 0.05) increased BW, partial blood values for commercial broilers grown to heavy 2+ + − pressure of CO2, Ca , Na , Cl , and thyroxine, along weights. Plasma corticosterone was not affected by with significantly (P ≤ 0.05) reduced pH level, partial either treatments or strain, indicating that the treat- − pressure of O2, HCO3 , and triiodothyronine in comments did not induce physiological stress in broilers. parison with Ross × Ross 708. No main effect of light

Key words: genetic strain, light intensity, acid-base balance, broiler , well-being 2014 Poultry Science 93 :970–978 http://dx.doi.org/ 10.3382/ps.2013-03613

INTRODUCTION nological, biochemical, and anatomical changes in the broiler (Havenstein et al., 2003a; Schmidt et al., 2009) The poultry industry has made rapid progress in im- and the consequences are not clear. This intense genetic proving the efficiency of broiler growth and production. selection for rapid growth may compromise animal wel- Advances include improved genetics and nutrition along fare as related to systemic physiological, immunologi- with changes in environmental management, resulting cal, and biochemical changes, among others. Some of in more rapid broiler growth. Rate of development dif- these changes may be accountable for the metabolic or fers among genetic strains, and lighting needs may be skeletal disorders reported for high-yield modern broil- genetic strain-specific to optimize growth performance. er strains (Julian, 1998). In addition, Julian (2005) re- A new strain of broiler, Ross × Ross 708, was devel- ported rapid growth rate of modern broiler chickens is oped primarily for breast meat production and has a associated with a series of physiological disorders lead- slower initial growth rate than the Ross × Ross 308 ing to higher mortality during grow-out (Julian, 2005). strain that is used for multipurpose broiler production. Manipulation of lighting programs is a strategy used This improvement in growth performance and carcass to reduce the incidence of metabolic and skeletal dis- characteristics have come with physiological, immu- orders in broiler chickens. High light intensity will increase activity, whereas lower intensities are effective © 2014 Poultry Science Association Inc. in improving production parameters (Classen and Rid- Received September 9, 2013. dell, 1989; Gordon, 1994). Exposure of broilers to sub- Accepted December 16, 2013. optimal environmental factors, including temperature, 1 Mention of trade names or commercial products in this publication light, diets, and gases, among others, during the course is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the USDA. of poultry production has an effect on blood physiologi- 2 Corresponding author: [email protected] cal variables such as blood acid-base balance, electro-

970 EFFECTS OF GENETIC STRAIN AND LIGHT INTENSITY 971 lytes, and metabolites (Olanrewaju et al., 2007, 2008, ly distributed to 10 environmentally controlled rooms 2009). (30 male and 30 female chicks per room) with 5 rooms Hematological analyses, in combination with other per strain. Each room was randomly assigned 1 of 5 biochemical methods, have been used as a physiologi- light intensities for each strain. Each environmentally cal method to assess the health status of animals (Kral controlled room had a floor area of 6 m2 (2.3 × 2.6 m; and Suchy, 2000). The changes in key selected param- width by depth) with a room volume of 15.3 m3 (2.5 m eters are routinely used to determine various influences high). At 12 d of age, birds received a Gumboro vacci- of environmental, nutritional, and pathological factors nation via water administration. Each room contained (Vecerek et al. 2002). Furthermore, these factors, in- 7.62 cm (depth) of fresh pine shavings, tube feeders, cluding physiological response (Alodan and Mashaly, and a 7-nipple watering system. The chicks remained 1999), environmental condition (Vecerek et al. 2002), in their respective rooms from 1 d old throughout the Downloaded from and age (Forlan et al. 1999), may affect the blood experimental period (1 to 56 d of age). Birds were pro- chemistry profile of healthy birds. In addition, these vided a 4-phase feeding program (starter = 1 to 14 d; factors, such as age, nutrition, degree of physical activ- grower = 15 to 28 d; finisher = 29 to 42 d; withdrawal ity, sex, and environmental conditions, also affect blood = 43 to 56 d). They were fed standard broiler diets chemistry in animals (Swenson, 1970). Changes in acid- formulated to meet or exceed NRC (1994) nutrient rec- http://ps.oxfordjournals.org/ base balance may signal early symptoms of diseases ommendations. Starter feed was provided as crumbles and influence the early manifestation of clinical signs and subsequent feeds were provided as whole pellets. and therapeutic effectiveness in both domestic animals Feed and water were offered ad libitum. Temperature and human beings (Brobst, 1975; Gunnerson, 2005). and RH on d 1 were maintained at 32 ± 1.1°C and Stress responses are also integrally involved with acid- 50 ± 5%, respectively, and remained constant across base balance in several species (Sandercock et al., 2001; all treatments. Temperature was decreased by 2°C per

Olanrewaju et al., 2006, 2007). week until it reached 15.6°C at 49 d of age and then at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 It is generally accepted that genetic differences be- maintained through 56 d of age. tween strains exist, but limited data are available in the literature on blood physiological responses to com- Experimental Treatments mercial lighting programs. Hence, evaluating blood physiological responses of modern broiler strains to For all treatments, the photoperiod consisted of (1) commercial lighting programs is warranted. Improved continuous lighting (24L:0D) with 20 lx of intensity understanding of the blood physiological response of from placement to 7 d of age; (2) 10 lx from 8 to 21 different strains to different light will enhance manage- d, 20L:4D from 22 through 53 d; and (3) continuous ment strategies to improve production efficiency and lighting (24L:0D) from 54 to 56 d. At d 22, chickens breast meat yield while ensuring bird health. Most of were subjected to 1 of the 5 varying levels of light in- the previous studies have not evaluated gradient lev- tensity (25, 10, 5, 2.5, and 0.2 lx) typical of those found els of light intensity used in commercial practice with in poultry commercial production. Each room was modern early- and late-developing broilers. To address equipped with incandescent bulbs, which peak in the this knowledge gap, the current study evaluated the red portion of the visible spectrum (750 nm), controlled effects of strain, light intensity, and their interaction by a dimmer and digital timer for each room typical of on blood gas, blood acid-base balance, electrolytes, that used in commercial housing. Light intensity set- and metabolites of early- and late-developing modern tings were verified from the center and 4 corners of each broiler strains. room at bird level (30 cm) using a photometric sen- sor with National Institute of Standards and Technol- MATERIALS AND METHODS ogy-traceable calibration (403125, Extech Instruments, Waltham, MA) for each intensity adjustment. The light Bird Husbandry fittings and tubes were cleaned weekly to minimize dust build-up, which would otherwise reduce the intensity. All bird care and use was approved by the USDA- Agricultural Research Service Animal Care and Use Blood Collections and Chemical Analyses Committee at the Mississippi State location. In each of 5 trials with trial as replicates, 600 one-day-old Ross On d 21 (the day before initiation of the treatments), × Ross 308 (Strain A) and Ross × Ross 708 (Strain 28, 42, and 56, blood samples were collected between B) chicks were purchased from the same commercial 0800 and 0900 h on sampling day from a brachial vein hatchery to remove any variation from the onset of the of 6 (3 male and 3 female chicks per room) randomly experiment (Aviagen, Huntsville, AL). Chicks were vac- selected chickens from each room. The birds were then cinated for Marek’s, Newcastle, and infectious bronchi- returned to the appropriate rooms using a standard tis diseases at the hatchery. Ross × Ross 308 and 708 handling procedure (Olanrewaju et al., 2008, 2010, were selected for differences in early and late weight 2013). In addition, unnecessary discomfort to the birds gain and performance, respectively. Upon arrival, the was avoided by using proper housing and handling tech- chicks were feather sexed, group-weighed, and random- niques, as described by the NRC (1996). Blood samples 972 OLANREWAJU ET AL. (3 mL) were collected directly into heparinized (50 IU/ otherwise stated. Analyses of variance combined across mL) monovette syringes. All bleedings were completed days were performed to obtain treatment comparisons within 45 s after birds were caught. Blood samples were averaged across days and to test for treatment interac- drawn directly from the syringes into a blood gas elec- tions with equal variances between days. The repeated trolyte analyzer (ABL-80 Flex, Radiometer America, measures were modeled using a compound symmetry Westlake, OH) for immediate analysis of partial pres- error structure. sure of CO2 (pCO2), partial pressure of O2 (pO2), pH, hematocrit (Hct), hemoglobin (Hb), and electrolytes + + 2+ − − RESULTS (Na , K , Ca , HCO3 , and Cl ). In addition, immediate glucose, osmolality (mOsm), and anion gap Table 1 shows the main effects of strain and light in- − Downloaded from were analyzed. The pH, pCO2, pO2, and HCO3 val- tensity on major selected blood physiological variables. ues were corrected to reflect a body temperature of In comparison with Ross × Ross 708 (Strain B), Ross 41.5°C (Burnett and Noonan, 1974). In addition, pH × Ross 308 (Strain A) had significantly higher BW (P + 2+ + was converted to H concentration to determine signifi- < 0.000), pCO2 (P < 0.023), Ca (P < 0.000), Na − cance levels between treatments. The mean corpuscular (P < 0.017), Cl (P < 0.000), and T4 (P < 0.000), hemoglobin concentration (McHc) in percentage was along with significantly (P < 0.000) reduced pH level, http://ps.oxfordjournals.org/ − calculated using the standard formula [(Hb × 100)/ pO2, HCO3 , and T3. No main effect of light intensity Hct]. The needle mounted on each monovette syringe was observed on all examined variables. In addition, no was then removed, a cap was placed over the needle significant main effects of strain, light intensity, or their + port, and the syringes containing the blood samples interaction on saturated O2 (sO2), Hct, Hb, K , McHc, were plunged into ice. After all birds were bled, the iced mOsm, anion gap, and glucose were observed. samples were transferred to the laboratory and centri- The influence of strain, light intensity, and their in- − fuged at 4,000 × g for 20 min at 4°C. Two milliliters teraction on whole blood pH and HCO3 is presented at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 of each of the plasma samples from the syringes were in Table 2. Strain A, in comparison to strain B, had restored in 2.5-mL graduated tubes at −20°C for later duced blood pH on d 21 (P < 0.003), 28 (P < 0.000), 42 chemical analyses. (P < 0.035), and 56 (P < 0.000). No main effect of light Plasma samples were removed from the freezer, intensity or strain × light intensity interaction was ob- thawed, and analyzed for corticosterone using a uni- served on pH on any of the sampling days. In addition, − versal microplate spectrophotometer (Bio-Tek Instru- strain A had a significant reduction of HCO3 on d 21 ments Inc., Winooski, VT) with ELISA reagent assay (P < 0.005), 28 (P < 0.003), and 42 (P < 0.006), but no test kits (EIA-CS Kit, Enzo Life Sciences, Farmingdale, strain effect was observed on d 56 in comparison with NY) according to the manufacturer’s instructions and strain B. Also, no main effect was noted of light inten- − previously used with broilers (Olanrewaju et al., 2008, sity or strain × light intensity interaction on HCO3 on 2010). Levels of plasma triiodothyronine (T3) and thy- any of the sampling days. roxine (T4) concentrations were measured using a uni- As shown in Table 3, strain A had significantly reversal microplate spectrophotometer (Bio-Tek Instru- duced pO2 on d 21, 42, and 56 (P < 0.0001, P < 0.007, ments Inc.) with ELISA reagent assay test kits from P < 0.000, respectively) in comparison with strain B. ALPCO Diagnostics (Salem, NH) according to the In addition, no effect of light intensity or strain × light manufacturer’s instructions and previously used with intensity was noted on pO2 on any of the sampling broilers (Olanrewaju et al., 2012). days. Strain A, in comparison to strain B, had a significantly increased blood concentration of Na+ on d 21(P Statistical Analysis < 0.015), 28 (P < 0.049), and 42 (P < 0.000), but no effect of strain on d 56. Only a significant (P < 0.043) The experimental design was a randomized complete main effect of light intensity on Na+ was observed on block design. Treatment structure was a 2 × 5 factorial d 42. No effect of strain × light intensity on Na+ was arrangement with the main factors being strain source found on any of the other sampling days. [Ross × Ross 308 (early developing) and Ross × Ross Table 4 shows the effects of strain and light intensity 708 (late developing)] and light intensity (25, 10, 5, 2.5, on blood concentrations of Ca2+ and Cl−. Strain A, 0.2 lx) with trial as replicates. Room was considered the in comparison to strain B, had increased blood Ca2+ experimental unit and treatments were replicated over on d 21(P < 0.001), 28 (P < 0.001), 42 (P < 0.001), time. Rooms used were switched within trials to remove and 56 (P < 0.021). Furthermore, strain A had a sig- room effects so that treatments were not confounded. nificant (P < 0.001) increase in blood concentration Five trials were conducted with trial as replicates. The of Cl− from d 21 through 56 of age. No main effect of main effects of strain and light intensity and the inter- light intensity or strain × light intensity on both blood action of these 2 factors on physiological variables were Ca2+ and Cl− was noted on any of the sampling days. tested by using the MIXED procedure (SAS Institute Table 5 shows the influence of strain and light intensity Inc, 2008). Means comparisons on d 21, 28, 42, and 56 on blood concentrations of T3 and T4. Strain A had were assessed by least significant differences, and state- significantly (P < 0.000) reduced T3 from d 21 through ments of significance were based on P ≤ 0.05, unless 56 in comparison with strain B. However, strain A had Table 1. Main effects of strain (S) and light intensity (LI) on selected blood physiological variables of heavy broiler chicken EFFECTS OFGENETICSTRAINANDLIGHTINTENSITY

Strain2 LI3 S:LI

Variable1 A B P-value 1 2 3 4 5 P-value SEM P-value BW (kg) 4.323a 3.952b 0.000 4.147 4.183 4.141 4.135 4.084 0.523 0.039 0.619 pH 7.352b 7.384a 0.000 7.361 7.370 7.371 7.365 7.373 0.224 0.003 0.123 a b pCO2 (mmHg) 48.994 47.841 0.023 49.018 48.245 47.637 48.926 48.261 0.391 0.351 0.331 − b a HCO3 (mmHg) 26.350 27.522 0.000 26.781 27.069 26.820 26.974 27.035 0.744 0.116 0.729 b a pO2 (mmHg) 36.494 40.040 0.000 37.158 37.930 38.104 39.075 39.070 0.321 0.473 0.575 sO2 (%) 67.008 69.5 0.074 68.697 67.970 66.489 69.134 68.974 0.737 1.542 0.696 Hct (%) 24.895 24.439 0.199 25.046 24.739 24.440 24.594 24.515 0.829 0.393 0.931 Hb (g/dL) 7.990 8.246 0.551 8.039 8.942 7.844 7.899 7.866 0.437 0.477 0.424 Na+ (mEq/L) 149.53a 146.00b 0.017 148.56 148.67 150.69 148.95 149.45 0.615 1.064 0.709 K+ (mEq/L) 4.80 4.77 0.801 4.947 4.654 4.568 4.908 4.849 0.059 0.101 0.207 Ca2+ (mEq/L) 3.176a 3.106b 0.000 3.140 3.130 3.136 3.146 3.151 0.338 0.005 0.724 Cl− (mEq/L) 107.58a 105.18b 0.000 106.02 106.08 106.39 106.49 106.92 0.831 0.377 0.980 McHc (%) 32.090 33.744 0.331 32.092 36.244 32.066 32.107 32.076 0.430 1.890 0.404 GLU (mg/dL) 220.78b 227.98a 0.000 222.47 223.72 222.31 228.38 225.02 0.185 1.246 0.171 mOsm (mmol/kg) 311.3 312.5 0.647 313.68 309.80 310.08 314.47 311.41 0.707 2.867 0.771 Angap (mmol/L) 20.324 20.290 0.842 20.4 20.247 20.216 20.457 20.176 0.735 0.186 0.222 b a T3 (ng/mL) 1.667 2.161 0.000 1.922 1.919 1.827 1.923 1.979 0.702 0.047 0.549 a b T4 (μg/dL) 3.836 3.363 0.000 3.581 3.595 3.621 3.589 3.612 0.996 0.058 0.888 CS (pg/L) 1,868.78 2,009.36 0.298 1,783.70 1,982.24 2,098.78 2,022.47 1,808.17 0.505 150.04 0.932 a,bMeans within a row and treatment that lack common superscripts differ significantly (P ≤ 0.05). 1 pCO2 = partial pressure of CO2; pO2 = partial pressure of O2; sO2 = saturated O2; Hct = hematocrit; Hb = hemoglobin; McHc = mean corpuscular hemoglobin concentration; GLU = glucose; mOsm = osmolality; Angap = anion gap; T3 = triiodothyronine; T4 = thyroxine; CS = corticosterone. 2Strain A = (Ross × Ross 308); Strain B = (Ross × Ross 708). 3Light intensity: 1 = 25 lx, 2 = 10 lx, 3 = 5 lx, 4 = 2.5 lx, 5 = 0.2 lx.

973

− 1 Table 2. Influence of strain and light intensity on blood pH and HCO3 of heavy broiler chickens

− pH HCO3 (mmHg)

Item 21 d 28 d 42 d 56 d 21 d 28 d 42 d 56 d Strain A (Ross × Ross 308) 7.360b 7.345b 7.365b 7.337b 26.343b 26.285b 26.161b 26.608 B (Ross × Ross 708) 7.388a 7.388a 7.377a 7.382a 27.230a 27.902a 27.545a 27.410 Light intensity treatment (lx) 25 7.371 7.358 7.363 7.353 26.832 27.089 26.361 26.842 10 7.374 7.373 7.377 7.357 26.809 27.301 27.075 27.090 5 7.375 7.375 7.375 7.359 26.542 26.961 26.956 26.821

2.5 7.366 7.362 7.372 7.358 26.555 27.000 26.994 27.348 Downloaded from 0.2 7.386 7.365 7.369 7.371 27.195 27.117 26.882 26.945 Pooled SEM 0.005 0.005 0.003 0.004 0.174 0.298 0.283 0.490 Strain-light intensity treatment Strain A + 25 7.351 7.333 7.351 7.323 26.094 26.186 25.337 26.383 Strain A + 10 7.353 7.351 7.374 7.326 26.424 26.531 26.611 27.039 Strain A + 5 7.360 7.352 7.375 7.335 25.932 26.034 26.541 26.407 Strain A + 2.5 7.348 7.338 7.37 7.331 26.083 26.113 26.397 26.892 http://ps.oxfordjournals.org/ Strain A + 0.2 7.390 7.350 7.357 7.369 27.182 26.561 25.922 26.320 Strain B + 25 7.390 7.382 7.375 7.383 27.569 27.992 27.385 27.301 Strain B + 10 7.394 7.396 7.379 7.388 27.194 28.071 27.538 27.140 Strain B + 5 7.391 7.398 7.375 7.384 27.151 27.889 27.371 27.235 Strain B + 2.5 7.384 7.385 7.374 7.385 27.027 27.888 27.591 27.804 Strain B + 0.2 7.383 7.380 7.382 7.372 27.208 27.673 27.841 27.570 Pooled SEM 0.012 0.013 0.008 0.010 0.388 0.667 0.634 0.693 P-value Strain 0.003 0.000 0.035 0.000 0.005 0.003 0.006 0.097 Light intensity 0.533 0.557 0.478 0.471 0.481 0.987 0.803 0.935 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Strain × light intensity 0.279 0.922 0.352 0.065 0.449 0.978 0.806 0.940 a,bMeans within a column and effect that lack common superscripts differ significantly (P ≤ 0.05). 1Values are least squares of 5 replicate rooms with 60 birds per room. a significant increase in blood concentration of T4 on comparison with strain B. In addition, no effect of light d 21(P < 0.000), 28 (P < 0.000), and 42 (P < 0.000), intensity or strain × light intensity was noted on either but no effect of strain on T4 was observed at d 56 in T3 or T4 on any of the sampling days. No effects of

+ 1 Table 3. Influence of strain and light intensity on blood partial pressure of O2 (pO2) and Na of heavy broiler chickens

+ pO2 (mmHg) Na (mEq/L)

Item 21 d 28 d 42 d 56 d 21 d 28 d 42 d 56 d Strain A (Ross × Ross 308) 35.991b 40.997 36.978b 32.012b 147.09a 147.73a 150.17a 153.12 B (Ross × Ross 708) 40.009a 41.131 39.315a 39.707a 146.34b 147.18b 148.33b 154.16 Light intensity treatment (lx) 25 37.181 39.532 37.688 34.232 146.56 147.41 148.43b 151.84 10 38.276 40.080 37.333 36.029 146.64 147.35 148.95b 151.72 5 35.910 41.723 38.745 36.040 146.69 147.69 149.45ab 158.95 2.5 40.256 41.795 38.431 35.819 146.60 147.31 148.98b 152.9 0.2 38.377 42.189 38.535 37.177 147.08 147.51 150.43a 152.77 Pooled SEM 0.632 0.680 0.488 0.629 0.181 0.175 0.248 3.268 Strain-light intensity treatment Strain A + 25 37.718 38.108 36.110 28.887 147.31 147.41 149.99 152.82 Strain A + 10 37.085 40.350 35.800 32.147 147.13 147.58 149.35 152.49 Strain A + 5 31.853 41.234 38.401 32.691 147.05 148.34 149.98 152.97 Strain A + 2.5 39.679 42.656 37.286 32.615 146.74 147.35 149.78 154.43 Strain A + 0.2 36.619 42.635 37.293 33.721 147.22 147.97 151.72 152.87 Strain B + 25 39.644 40.955 39.266 39.577 145.81 147.40 146.88 150.85 Strain B + 10 39.466 39.810 38.867 39.911 146.16 147.12 148.55 150.95 Strain B + 5 39.967 42.211 39.089 39.389 146.33 147.04 148.91 164.94 Strain B + 2.5 40.833 40.933 39.577 39.023 146.47 147.28 148.19 151.37 Strain B + 0.2 40.137 41.744 39.777 40.633 146.93 147.04 149.14 152.68 Pooled SEM 1.414 0.962 1.092 1.406 0.405 0.392 0.555 4.622 P-value Strain 0.001 0.830 0.007 0.000 0.015 0.049 0.000 0.729 Light intensity 0.102 0.072 0.663 0.398 0.715 0.875 0.043 0.523 Strain × light intensity 0.206 0.213 0.795 0.566 0.556 0.451 0.253 0.500 a,bMeans within a column and effect that lack common superscripts differ significantly (P ≤ 0.05). 1Values are least squares of 5 replicate rooms with 60 birds per room. EFFECTS OF GENETIC STRAIN AND LIGHT INTENSITY 975

Table 4. Influence of strain and light intensity on blood Ca2+ and Cl− of heavy broiler chickens1

Ca2+ (mEq/L) Cl− (mEq/L)

Item 21 d 28 d 42 d 56 d 21 d 28 d 42 d 56 d Strain A (Ross × Ross 308) 3.159a 3.189a 3.178a 3.179a 104.77a 106.16a 109.02a 110.36a B (Ross × Ross 708) 3.070b 3.113b 3.112b 3.128b 103.07b 104.19b 105.07b 108.39b Intensity treatment (lx) 25 3.125 3.150 3.143 3.144 103.89 104.99 106.25 108.93 10 3.114 3.139 3.128 3.139 103.71 105.06 106.61 108.96 5 3.103 3.144 3.142 3.155 103.98 105.54 107.16 108.90

2.5 3.118 3.151 3.143 3.175 103.93 105.04 107.10 109.87 Downloaded from 0.2 3.113 3.170 3.167 3.155 104.09 105.26 108.10 110.23 Pooled SEM 0.013 0.019 0.011 0.013 0.258 0.287 0.357 0.481 Strain-light intensity treatment Strain A + 25 3.177 3.183 3.171 3.178 105.20 105.63 108.68 109.89 Strain A + 10 3.166 3.175 3.146 3.158 104.34 105.82 108.05 109.55 Strain A + 5 3.135 3.183 3.164 3.180 104.77 107.00 108.70 110.09 Strain A + 2.5 3.168 3.191 3.185 3.217 104.52 105.88 109.23 111.47 http://ps.oxfordjournals.org/ Strain A + 0.2 3.150 3.210 3.222 3.160 105.02 106.50 110.42 110.83 Strain B + 25 3.072 3.116 3.115 3.120 102.58 104.35 103.82 107.98 Strain B + 10 3.062 3.103 3.111 3.120 103.07 104.30 105.17 108.37 Strain B + 5 3.070 3.105 3.119 3.129 103.19 104.08 105.62 107.70 Strain B + 2.5 3.068 3.110 3.100 3.132 103.34 104.20 104.97 108.27 Strain B + 0.2 3.077 3.130 3.112 3.150 103.16 104.02 105.78 109.64 Pooled SEM 0.019 0.015 0.016 0.029 0.576 0.642 0.505 0.681 P-value Strain 0.001 0.001 0.001 0.021 0.001 0.001 0.000 0.001 Light intensity 0.822 0.355 0.244 0.781 0.973 0.903 0.039 0.226 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Strain × light intensity 0.732 0.987 0.167 0.738 0.735 0.682 0.247 0.556 a,bMeans within a column and effect that lack common superscripts differ significantly (P ≤ 0.05). 1Values are least squares of 5 replicate rooms with 60 birds per room. strain, light intensity, or strain × light intensity was den changes can result in cellular damage via protein + observed on sO2, Hct, Hb, K , McHc, mOsm, anion ionization (Eckert, 1988). However, any speculation re- gap, or glucose on any of the sampling days (data not garding the role of blood pH in the regulation of res- shown). Furthermore, blood glucose and plasma corti- piration must be tempered by a consideration of other costerone concentrations were not significantly affected factors that are also influential in the chemical control by treatments on any of the sampling days in the pres- of respiration (Gesell, 1925). In addition, the carbonic ent study (data not shown). acid-bicarbonate system is the most important buffer for maintaining blood acid-base balance. In the present study, Ross × Ross 308 had significant DISCUSSION + 2+ − (P ≤ 0.05) increases in BW, pCO2, Na , Ca , Cl , Previous studies demonstrated partial effects of light and T4 along with significantly (P ≤ 0.05) reduced pH, − intensity on blood physiological parameters (Olanre- pO2, HCO3 , and T3 in comparison with Ross × Ross waju et al., 2012) and in combination with ammonia 708 under the same light intensity throughout the ex- (Olanrewaju et al., 2008), temperature (Olanrewaju et perimental period (d 0–56 of age). However, all these al., 2010), and photoperiod (Olanrewaju et al., 2013). acid-base changes are still within the normal acid-base Similar to our previous studies, the present study indi- homeostasis and physiological ranges. The association cates that light intensity with or without strain had a of CO2 retention, which we observed in strain A in minor effect on most of the examined blood physiologi- comparison with strain B, may be due to the rate of cal variables of broilers grown to heavy weights (Olan- CO2 elimination as related to BW. For instance, when rewaju et al., 2008, 2012, 2013). the respiration rate decreased in strain A in the cur- The acid-base status of poultry is challenged daily rent study, the levels of pCO2 in the blood increased by environmental factors, such as light, temperature, and more H+ ions accumulated, causing the pH of humidity, nutrition, water, and air quality, as well as by blood to decrease, thereby pushing the equilibrium re- other factors that influence respiratory and metabolic action to the right and creating more hydrogen ions, activities. The principal organ systems used in acid- thereby increasing hydrogen ion concentration, which base homeostasis in birds are the lungs and kidneys, lowered blood pH, and thus the blood acid-base status supported by the gastrointestinal tract (Long, 1982). in this study reflected acute respiratory acidosis (sig- The cardiovascular system also participates in thermo- nificantly increased level of pCO2, decreased levels of − + regulatory processes through modulation of heat dis- pH and HCO3 ) caused by hydrogen ion (H , acid) sipation and by oxygen transport. The pH of the blood accumulation (Julian and Mirsalimi, 1992; Wideman et is maintained within a very narrow range because sud- al., 2000). Sometimes, acute respiratory acidosis results 976 OLANREWAJU ET AL.

1 Table 5. Influence of strain and light intensity on blood plasma triiodothyronine (T3) and thyroxine (T4) of heavy broiler chickens

T3 (ng/mL) T4 (μg/dL)

Item 21 d 28 d 42 d 56 d 21 d 28 d 42 d 56 d Strain A (Ross × Ross 308) 2.114b 1.691b 1.596b 1.269b 3.781a 4.104a 3.858a 3.601 B (Ross × Ross 708) 2.376a 2.279a 2.060a 1.929a 3.112b 3.169b 3.406b 3.765 Intensity treatment (lx) 25 2.293 1.944 1.867 1.584 3.483 3.564 3.626 3.671 10 2.251 2.117 1.787 1.522 3.503 3.611 3.546 3.72 5 2.068 1.957 1.788 1.496 3.371 3.668 3.808 3.638

2.5 2.312 1.925 1.754 1.702 3.322 3.590 3.692 3.751 Downloaded from 0.2 2.300 1.981 1.944 1.690 3.555 3.749 3.508 3.635 Pooled SEM 0.067 0.0327 0.029 0.031 0.088 0.116 0.060 0.141 Strain-light intensity treatment Strain A + 25 2.128 1.568 1.600 1.222 3.814 3.971 3.794 3.413 Strain A + 10 2.146 1.967 1.564 1.240 3.848 4.086 3.701 3.646 Strain A + 5 1.964 1.836 1.633 1.222 3.684 4.162 4.201 3.511 Strain A + 2.5 2.165 1.480 1.499 1.358 3.697 4.076 3.946 3.720 http://ps.oxfordjournals.org/ Strain A + 0.2 2.165 1.602 1.685 1.304 3.861 4.223 3.649 3.712 Strain B + 25 2.458 2.320 2.135 1.945 3.152 3.156 3.417 3.928 Strain B + 10 2.356 2.266 2.010 1.803 3.157 3.136 3.391 3.794 Strain B + 5 2.173 2.078 1.944 1.770 3.059 3.174 3.415 3.765 Strain B + 2.5 2.459 2.369 2.008 2.047 2.946 3.103 3.437 3.781 Strain B + 0.2 2.434 2.360 2.203 2.076 3.248 3.274 3.366 3.559 Pooled SEM 0.149 0.073 0.064 0.069 0.197 0.163 0.133 0.199 P-value Strain 0.019 0.000 0.000 0.000 0.000 0.000 0.000 0.220 Light intensity 0.484 0.143 0.077 0.041 0.747 0.780 0.241 0.967 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Strain × light intensity 0.991 0.213 0.432 0.437 0.996 0.984 0.374 0.573 a,bMeans within a column and effect that lack common superscripts differ significantly (P ≤ 0.05). 1Values are least squares of 5 replicate rooms with 60 birds per room. from hypoventilation that may be due to loss of respi- relative to BW, the heart essentially pushes the blood ratory drive, body mass, chest wall capacity, or rapid- through the lungs with more pressure to increase the shallow breathing, but this can be quickly corrected amount of oxygen available to the bird’s metabolism. through a compensatory mechanism. This is based on However, because the lung volume and cardiovascular the fact that 3 systems (buffer, respiratory, and renal volume within the lung tissue is fixed in birds, unlike systems) function interdependently to regulate and in mammals, eventually a point is reached whereby the maintain acid-base balance. The buffer system can be lungs may no longer accommodate more blood being activated within seconds and the respiratory system supplied by the heart and this may have negative ef- can activate changes in pH within minutes, whereas fects on the body (poor oxygenation). Disturbances in the third line of defense is the renal system, which can venous blood acid-base status (pCO2 and pH) are fre- take days to compensate fully for the increase in CO2. quently observed in older birds, which may be due to These compensatory changes result in increased oxy- relative differences in their body sizes, which is normal. gen transport to tissues, increased chemoreceptor pO2 This may also be due to age-dependent differences in + with a concomitant decrease in CO2 and H , increased ventilation rate or may reflect the consequences of an vascular resistance, increased lactate production, and increased metabolic demand in the larger birds (Korte decreased work capacity. et al., 1999). The measurement of blood oxygen saturation in Plasma T3 level of strain A in the present study was broilers is very important because it has been reported significantly lower than that of strain B, whereas T4 that any condition that increases the oxygen demand concentration was opposite. It has been documented at the tissue level or decreases the oxygen-carrying ca- that T3 is the main physiological thyroid hormone reg- pacity of the circulatory system can result in pulmo- ulating oxygen consumption and daily active activities, nary hypertension and right heart failure with the clini- particularly in young chickens (Bobek et al., 1977), and cal signs of ascites (Julian, 1993). Reduced pO2 and is metabolically more active than T4 (Klandorf et al., slightly reduced sO2, as we observed in strain A, may 1981). It has been reported that the T3 hormone is be due to inadequate blood oxygenation and hypox- closely associated with feeding and is also a key factor emia, which increases the risk of hypoxia (Prakash and influencing conversion of T4 to T3 (McNabb, 2000), Madanmohan, 2006). Modern broiler chickens are able and that a higher T3 level is associated with increased to consume large quantities of feed and grow rapidly protein deposition. due to genetic selection, resulting in high demand for Body fluid electrolyte concentrations, such as Na+, oxygen. This in turn, may cause an increase in blood K+, and Cl−, and acid-base balance are interconnected pressure. When oxygen intake is low (low pO2, sO2) and are also associated with the condition producing EFFECTS OF GENETIC STRAIN AND LIGHT INTENSITY 977 acidosis or alkalosis in mammals, which may be true in Forlan, R. L., M. Macari, R. D. Matherious, V. M. Moraes, and E. birds (Terzano et al., 2012). In the present study, we B. Malherious. 1999. Effect of age and strain on hematology and + 2+ blood biochemical parameters in broiler chickens. Rev. Bras. observed significantly higher levels of Na , Ca , and Dec. Avian 1:77–84. − Cl in strain A in comparison to strain B, but no strain Gesell, R. 1925. The chemical regulation of respiration. Physiol. effects on K+ level. Although, the essential electrolytes Rev. 5:551–595. for the maintenance of the acid-base balance are Na+, Gordon, S. H. 1994. Effects of day-length and increasing day length + − programs on broiler welfare and performance. World’s Poult. Sci. K and Cl , acute respiratory acidosis does not have J. 50:269–282. a great effect on serum electrolyte levels. Small effects Gunnerson, K. J. 2005. Clinical review: The meaning of acid-base occur in calcium and potassium levels, if any. Acidosis abnormalities in the intensive care unit – Epidemiology. Crit. decreases binding of calcium to albumin and tends to Care 9:508–516.

Havenstein, G. B., P. R. Ferket, and M. A. Qureshi. 2003a. Growth, Downloaded from increase serum ionized calcium levels. In addition, aci- livability, and feed conversion of 1957 versus 2001 broilers when dosis causes an extracellular shift of potassium. How- fed representative 1957 and 2001 broiler diets. Poult. Sci. ever, respiratory acidosis rarely causes clinically signifi- 82:1500–1508. Julian, R. J. 1998. Rapid growth problems: Ascites and skeletal de- cant hyperkalemia. Replacing potassium that shifts out formities in broilers. Poult. Sci. 77:1773–1780. of cells during the acidosis state and correction of this Julian, R. J. 1993. Ascites in poultry. Avian Pathol. 22:419–454. may cause a relative serum hypokalemia as potassium Julian, R. J. 2005. Production and growth related disorders and oth- http://ps.oxfordjournals.org/ shifts back into cells. er metabolic diseases of poultry: A review. Vet. J. 169:350–369. Julian, R. J., and S. M. Mirsalimi. 1992. Blood oxygen concentration In conclusion, a specific main effect of genetic strain of fast-growing and slow-growing broiler chickens, and chickens in comparison to light intensity on most examined with ascites from right ventricular failure. Avian Dis. 36:730– blood physiological variables was observed. It follows 732. that the blood oxygen-carrying capacity of strain A, Klandorf, H., P. J. Sharp, and W. S. Newcomer. 1981. The influence of feeding patterns on daily variation in the concentrations of assessed by Hb, Hct, pO2, and SO2 values, was lower plasma thyroid hormones in the hen. IRCS Med. Sci. 9:82–87. in comparison to strain B. However, all these acid-base Korte, S. M., A. Sgoifo, W. Ruesink, C. Kwakernaak, S. van Voorst, at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 changes are still within the normal acid-base homeo- C. W. Scheele, and H. J. Blokhuis. 1999. High carbon dioxide stasis and physiological ranges. This data may be used tension (pCO2) and the incidence of cardiac arrhythmias in rapidly growing broiler chickens. Vet. Rec. 145:40–43. to establish the normal blood values for commercial Kral, I., and P. Suchy. 2000. Hematological studies in adolescent broilers grown to heavy weights. The results of this breeding cocks. Acta. Vet. Bras. 69:189–194. study supplement current knowledge of the hematology Long, S. 1982. Acid-base balance and urinary acidification in birds. Comp. Biochem. Physiol. A 71:519–526. and plasma biochemistry of early- and late-developing McNabb, F. M. A. 2000. Thyroids. Pages 461–471 in Sturkie’s Avian broiler strains grown to heavy weights under environ- Physiology. 5th ed. G. C. Whittow, ed. Academic Press, San Jose, mentally controlled conditions. In addition, the light CA. intensities used in this study apparently did not affect NRC. 1994. Nutrient Requirements of Poultry. 9th ed. Natl. Acad. Press, Washington, DC. plasma corticosterone, suggesting that these levels of NRC. 1996. Guide for the Care and Use of Laboratory Animals. light intensities did not pose as stressors to modern Natl. Acad. Press, Washington, DC. heavy broiler chickens. Olanrewaju, H. A., J. L. Purswell, S. D. Collier, and S. L. Branton. 2009. Age-related effects of varying ammonia concentrations on hematophysiological variables in broiler chickens. Int. J. Poult. ACKNOWLEDGMENTS Sci. 8:138–144. Olanrewaju, H. A., J. L. Purswell, S. D. Collier, and S. L. Bran- The authors thank Larry N. Halford and M. Robin- ton. 2010. Effect of ambient temperature and light intensity on physiological reactions of heavy broiler chickens. Poult. Sci. son, both of the USDA, Agricultural Research Service, 89:2668–2677. Poultry Research Unit, for their contributions to this Olanrewaju, H. A., J. L. Purswell, S. D. Collier, and S. L. Branton. study. 2012. Effect of varying light intensity on blood physiological reactions of broiler chickens grown to heavy weights. Int. J. Poult. Sci. 11:81–87. REFERENCES Olanrewaju, H. A., J. L. Purswell, S. D. Collier, and S. L. Branton. 2013. Interactive effects of photoperiod and light intensity on Alodan, M. A., and M. M. Mashaly. 1999. Effect of induced molting blood physiological and biochemical reactions of broilers grown in laying hens on production and immune parameters. Poult. to heavy weights. Poult. Sci. 92:1029–1039. Sci. 78:171–177. Olanrewaju, H. A., J. P. Thaxton, W. A. Dozier III, and S. L. Bobek, S., M. Jastrzebski, and M. Pietras. 1977. Age-related changes Branton. 2007. Electrolyte diets, stress, and acid-base balance in in oxygen consumption and plasma thyroid hormone concentra- chickens. Poult. Sci. 86:1363–1371. tion in the young chicken. Gen. Comp. Endocrinol. 31:169–174. Olanrewaju, H. A., J. P. Thaxton, W. A. Dozier III, J. L. Purswell, Brobst, D. 1975. Evaluation of clinical disorders of acid-base bal- S. D. Collier, and S. L. Branton. 2008. Interactive effects of am- ance. J. Am. Vet. Med. Assoc. 166:359–364. monia and light intensity on hematochemical variables in broiler Burnett, R. W., and D. C. Noonan. 1974. Calculations and correc- chickens. Poult. Sci. 87:1407–1414. tion factors used in determination of blood pH and blood gases. Olanrewaju, H. A., S. Wongpichet, J. P. Thaxton, W. A. Dozier III, Clin. Chem. 20:1499–1506. and S. L. Branton. 2006. Stress and acid-base balance in chick- Classen, H. L., and C. Riddell. 1989. Photoperiodic effects on per- ens. Poult. Sci. 85:1266–1274. formance and leg abnormalities in broiler chickens. Poult. Sci. Prakash, E. S., and T. Madanmohan. 2006. What does one mean 68:873–879. by “arterial blood oxygenation?” Adv. Physiol. Educ. 30:46–47. Eckert, R. 1988. Physical and chemical concepts. Pages 8–34 in Ani- Sandercock, D. A., R. R. Hunter, G. R. Nute, M. A. Mitchell, and mal Physiology. R. Eckert and D. Randall, ed. Freeman, New P. M. Hocking. 2001. Acute heat stress-induced alterations in York, NY. acidbase status and skeletal muscle membrane integrity in broiler 978 OLANREWAJU ET AL.

chickens at two ages: Implications for meat quality. Poult. Sci. lactate production in hypercapnic respiratory failure: The role of 80:418–425. noninvasive ventilation. PLoS ONE 7:e35245–e35253. SAS Institute Inc. 2008. SAS Proprietary Software Release 9.2. SAS Vecerek, V., E. Strakova, P. Suchy, and E. Volslacrova. 2002. Influ- Institute Inc., Cary, NC. ence of high environmental temperature on production and hae- Schmidt, C. J., M. E. Persia, E. Feierstein, B. Kingham, and W. W. matological and biochemical indexes in broiler chickens. Czech Saylor. 2009. Comparison of a modern broiler line and a heritage J. Anim. Sci. 47:176–182. line unselected since the 1950s. Poult. Sci. 88:2610–2619. Wideman, R. F., M. R. Fedde, C. D. Tackett, and G. E. Weigel. Swenson, M. J. 1970. Physiological properties, cellular and chemical 2000. Cardio-pulmonary function in preascitic (hypoxemic) or constituents of blood. Pages 75–83 in Dukes’ Physiology of Do- normal broilers inhaling ambient air or 100% oxygen. Poult. Sci. mestic Animals. 8th ed. Comstock Publishing Associates, Cornell 79:415–425. University Press, Ithaca, NY. Terzano, C., F. Di Stefano, V. Conti, M. Di Nicola, and G. Paone.

2012. Mixed acid-base disorders, hydroelectrolyte imbalance and Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Supplemental L-arginine and vitamins E and C preserve xanthine oxidase activity in the lung of broiler chickens grown under hypobaric hypoxia

J. Bautista-Ortega ,* A. Cortes-Cuevas ,† E. A. Ellis ,‡ and C. A. Ruiz-Feria *1

* Department of Poultry Science, Texas A&M University, College Station 77843-2472; † Facultad de Medicina Veterinaria y Zootecnia, UNAM, Mexico D.F., 04510; and ‡ Microscopy and Imaging Center, BSBW 119/MS 2257, Texas A&M University, College Station 77843 Downloaded from

ABSTRACT The effects of l-Arg, vitamin C (VC), and munocytochemistry colloidal gold particles (NTY). The vitamin E (VE) on xanthine- (XO) and NAD(P)H-oxi- XO and NOX localized in cell membranes and within dase (NOX) activities, and nitric oxide (NO) availabil- vesicles of pulmonary vessel endothelial cells. The XO ity of hypoxic broilers were evaluated. Chickens were activity was higher in CTL-NOR birds (586 ± 43 re- http://ps.oxfordjournals.org/ kept in wire cages with free access to feed and water. flectance units) than in both AEC-HYP (456 ± 39) and One-day-old chicks were assigned to 1 of 3 diets: con- HA-HYP birds (394 ± 31), whereas CTL-HYP birds trol (CTL; ME 3,200 kcal/kg, CP 23%), high Arg (HA; had the lowest XO activity (313 ± 27). The NO deple- CTL + Arg 0.8%), or high Arg plus VE and VC (AEC; tion was not affected by dietary or hypoxia conditions HA + 200 IU of VE/kg of feed + 500 mg of VC/L in clinically healthy birds; nevertheless, hypoxic birds of water), and grown under hypobaric hypoxia (HYP) that developed pulmonary hypertension had higher

from d 7 to 30. A fourth group of birds was fed the CTL NTY levels (less NO, 145 ± 19) than hypoxic but clini- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 diet and grown under normoxia (CTL-NOR). At d 30, cally healthy birds (56 ± 11). Thus, the concurrent sup- chickens were euthanized, their lungs fixed in vivo, ex- plementation of Arg, VE, and VC restored XO activity cised, and processed for cyto- and histochemistry. The without affecting NOX activity or NO availability. The enzymes XO and NOX were localized and activities as- dual role of XO, which produces superoxide and uric sessed histochemically and in lung homogenates. The acid, may have buffered the effects of superoxide in

NO depletion was assessed through nitrotyrosine im- broiler chickens grown under hypobaric hypoxia. Key words: arginine , antioxidant vitamin, pulmonary hypertension, xanthine oxidase, NAD(P)H oxidase 2014 Poultry Science 93 :979–988 http://dx.doi.org/ 10.3382/ps.2013-03698

INTRODUCTION with pathological conditions of the vasculature, including PH (Selemidis et al., 2008; Lassègue and Griendling, The onset of pulmonary hypertension syndrome 2010). The role of NOX on PHS in broiler chickens has (PHS) in broiler chickens has been associated with an not been studied. On the other hand, increased levels of increased production of reactive oxygen species (ROS) uric acid (UA), attributed to an increased XO activity, and endothelial cell damage (Enkvetchakul et al., 1993; have been reported in humans with PH (Warwick et al., Wedgwood and Black, 2003; Pan et al., 2007; Nain et 2008). Poss et al. (1996) provided evidence that cell- al., 2008a,b). Sources of ROS related to pulmonary bound XO can impair vascular cell function by inhibit- hypertension (PH) include uncoupled endothelial ni- ing nitric oxide (NO)-dependent signal transduction. tric oxide synthase (eNOS; Sud et al., 2007), oxidases Although XO has been reported to be undetectable in such as xanthine oxidase (XO) and NAD(P)H oxidase the lung of healthy broilers (Carro et al., 2009), it has (NOX; Zalba et al., 2000; Jankov et al., 2008; War- been shown to be attached to red blood cell membranes wick et al., 2008; Guzik and Griendling, 2009), and (Al-Khalidi and Chaglassian, 1965), and is therefore uncoupled mitochondria (Maxwell et al., 1996; Iqbal et in close contact with the pulmonary vascular endothe- al., 2001a,b). lium. The enzyme NOX has been described as the only l-Arginine is an essential amino acid for birds and known enzyme whose sole purpose is to produce ROS is the substrate of endothelial NO synthase (eNOS), in various species and tissues, and is often associated the enzyme that synthesizes NO, a potent vasodilator (Dudzinski and Michel, 2007). Endothelial NO synthase © 2014 Poultry Science Association Inc. is constitutively expressed and basal levels of NO cause Received October 18, 2013. Accepted December 3, 2013. the pulmonary vascular smooth muscle to maintain vas- 1 Corresponding author: [email protected] cular tone (Govers and Rabelink, 2001). Supplemental

979 980 BAUTISTA-ORTEGA ET AL. Arg has been shown to improve cardiovascular perfor- tive stress and NO depletion using nitrotyrosine as a mance in broiler chickens exposed to low temperatures marker, in lung tissue of broiler chickens grown under (Lorenzoni and Ruiz-Feria, 2006; Ruiz-Feria, 2009) and hypobaric hypoxia or normoxia, and fed supplemental has been reported to reduce PHS incidence (Wideman, Arg, VE, and VC. et al., 1995; Tan et al., 2007b), although the results have not been consistent. Vitamin E (VE) is a cell MATERIALS AND METHODS membrane antioxidant that protects cell integrity by reducing polyunsaturated fatty acid oxidation, whereas Experimental Design vitamin C (VC) is hydrophilic and functions better in aqueous environment (i.e., cytosolic and extracellular One-day-old chicks (Cobb 500, n = 200) were wing- compartments). Furthermore, VC restores the antioxi- banded and allocated to 1 of 3 dietary treatments: Downloaded from dant capability of oxidized VE (Guney et al., 2007). a basal control diet (CTL, containing 3,200 kcal of Previously we found that high levels of VE (200 IU/ ME/kg of feed, 23% CP, 1.55% (wt/wt) Arg, and 40 kg of feed) and Arg (1%, wt/wt) improved cardiopul- IU of VE/kg of feed); a high-Arg diet (HA, the basal monary response after an acute challenge with epineph- diet plus 0.8% (wt/wt) supplemental l-arginine HCl rine compared with birds fed high levels of Arg alone (SAFC Supply Solutions, St. Louis MO), or a high Arg http://ps.oxfordjournals.org/ in birds grown at low temperature (Ruiz-Feria, 2009). and vitamin diet (AEC; the HA diet plus 200 IU of In the same work, we found that birds fed a combina- α-tocopherol/kg of feed (Producers Cooperative Asso- tion of Arg, VE, and VC had a better cardiopulmonary ciation, Bryan, TX) and 500 mg of ascorbic acid/L of response than birds fed either a combination of Arg drinking water (Sigma-Aldrich Corp., St. Louis, MO). and VE or a combination of Arg and VC. However, The diets were isocaloric and isonitrogenous and formu- the amplification of PHS through cold stress is dif- lated to meet or exceed all of the NRC (1994) require-

ferent from that caused by hypobaric hypoxia; in the ments. Supplemented VC in the drinking water was at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 first case, pulmonary arterial pressure (PAP) is pri- prepared daily. The chicks were brooded convention- marily increased by an enhanced cardiac output in really with temperature starting at 32°C and decreasing sponse to an increase in metabolic demand (Wideman, 2°C each week until wk 3, under a constant lighting 2001), whereas in the second case, the increase in PAP program. The chicks were housed in 6 wire cages (56 is caused primarily by pulmonary vascular constric- cm long × 33 cm wide × 28 cm high) under normoxic tion increasing pulmonary vascular resistance (Julian, conditions for 7 d and then the cages were placed in 2007). We recently reported that feeding supplemental 2 plastic cylindrical hypobaric chambers (185 cm long Arg also improved the pulmonary vascular performance and 58 cm in diameter) in which the chickens were ex- of broiler chickens grown under hypoxic conditions, and posed, until they were 30 d old, to a simulated altitude its effects were further improved by the addition of the of 3,000 m above sea level (oxygen partial pressure of antioxidant VE and VC (Bautista-Ortega and Ruiz- 14.5 KPa) to amplify the incidence of PHS. The desired Feria, 2010; Bautista-Ortega et al., 2013). simulated altitude was achieved by pulling a vacuum One of the mechanisms proposed through which ox- through the chamber with a Roots Universal blower idative stress is implicated in the pathophysiological (Dresser Industries, Connersville, IN). The blower was progression of PHS is by reducing availability of the reversed to pull air through the air plenum. On a daily NO through its reaction with superoxide to produce basis, the chicks were kept inside the hypobaric cham- peroxinitrite, reducing smooth muscle vasodilation and bers continuously except for 40 min to clean and refeed contributing to a high PAP (Martinez-Lemus et al., the chicks, 15 min to decompress the chambers, and 1999; Tan et al., 2007a). Previously we have shown that 15 min to reach the desired vacuum level. The dietary supplemental Arg, VE, and VC improve cardiovascu- groups exposed to hypobaric hypoxia were identified lar performance in chickens exposed to low tempera- as CTL-HYP, HA-HYP, and AEC-HYP. Additionally, ture or hypobaric hypoxia (Ruiz-Feria, 2009; Bautista- one-day-old chicks were raised for 10 d under normoxic Ortega and Ruiz-Feria, 2010; Bautista-Ortega et al., conditions (CTL-NOR), in an adjacent facility to the 2013). We hypothesize that NOX and XO are impor- hypobaric challenge unit, and they were brooded and tant sources of superoxide in hypoxic broiler chickens, raised following the same procedures. The normoxic and concurrent supplementation of Arg, VE, and VC birds provided lung tissue samples as technical control will modulate their activities, preserving NO bioavail- for cyto- and histo-chemical studies. ability. However, to our knowledge, the demonstration of NOX and XO in the vicinity of the pulmonary artery Lung Tissue Collection and Sample endothelium of hypoxic broiler chickens is still pending. Processing for Electron Microscopy Thus, the objectives of the present investigation were: Examination (a) to ultracytochemically localize NOX and XO, (b) to semiquantitatively determine NOX and XO activ- A surgical plane of anesthesia was induced in the ity through the quantitation of H2O2 deposition using 30-d-old chickens by intramuscular injections of allo- scanning laser reflectance confocal microscopy, and (c) barbital (5,5-diallyl-barbituric acid; 25 mg/kg of BW). to semiquantitatively determine the degree of oxida- The birds were fastened in dorsal recumbency on a sur- XANTHINE OXIDASE ACTIVITY IN HYPOXIC CHICKENS 981 gical board, the heart was exposed through an opening ed with the substrate the enzyme produces superoxide in the abdominal cavity, and the lungs were perfused (Ellis and Grant, 2002). In turn, superoxide is dis- in vivo as follows. An 18-ga hypodermic needle was mutased by superoxide dismutase to H2O2. Hydrogen inserted in the left ventricle and perfusion was initiat- peroxide then reacts with cerium chloride, in the pres- ed by injecting 10 mL of heparinized saline to dissolve ence of aminotriazole that inhibits endogenous catalase blood clots followed by 10 mL of perfusion buffer that and peroxidases, to produce a water-insoluble, electron- contained 1% glutaraldehyde and 4% paraformaldehyde dense precipitate, cerium perhydroxide. in 0.1 M HEPES buffer (pH 7.4). The lungs were subsequently removed and placed in 10 volumes of perfusion Determination of XO and NOX Activity buffer in an ice bath until further processing (Wideman

in Intact Lung Parenchyma Downloaded from et al., 2011). The lungs were subsequently fixed for 1 h at 4°C in cold 5.0% (vol/vol) acrolein in 0.1 M HEPES Cryostat Sectioning of Lung Samples. Five-milli- buffer (pH 7.4). Specimens were washed overnight at meter-thick cross sections were cut from each lung and 4°C in cold 0.1 M HEPES buffer (pH 7.4) plus 5% (wt/ then equilibrated in 30% sucrose in 0.1 mM HEPES vol) sucrose and 1% (vol/vol) dimethylsulfoxide. Bio- (pH 7.4) overnight at 4°C. Frozen lung tissue samples logical tissues fixed using the aforementioned procedure were then cut to a thickness of 20 μm and mounted as http://ps.oxfordjournals.org/ can be safely further processed for light or electron mi- suggested by Robinson and Batten (1990). In short, the croscopy studies (Ellis and Grant, 2002). coverslips were rinsed twice in the fixation buffer and then mounted on glass slides in glycerol/0.1 M HEPES Cytochemical Localization of XO and NOX buffer (1:1 ratio, pH 7.4), and then mounted on microscope slides for immediate observation. The cytochemical localization of XO and NOX was Histochemical Localization of XO and NOX. His-

conducted as described by Ellis and Grant (2002). For tochemical localization of XO and NOX in the lung at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 the localization of XO, the lung samples that were pre- parenchyma was done as described above for the cyto- viously processed for electron microscopy were brought chemical localization of both enzymes with the differ- to room temperature in the final 2 buffer washes which ence that the preincubation and incubation media were contained 0.1 M glycine and then preincubated 2 times scaled up to accommodate several mounted specimens for a 6-min cycle at 37°C (temperature range 37 to to be reacted successfully. In this case, the amount of 42°C) using a microwave set to 250 W and in the follow- cerium perhydroxide that precipitates in the lung pa- ing medium: 10 mM cerium chloride, 10 mM 3-amino- renchyma correlates with the activity of XO or NOX. 1,2,4-triazole, 0.1 M HEPES buffer (pH 8), 7% sucrose, The oxidase activity is then measured by the average and 0.0002% Triton X-100. They were then incubated reflectance intensity of a laser beam cast on the mount- in the following complete reaction medium: 10 mM ce- ed tissue sample using a confocal microscope (Robinson rium chloride, 10 mM 3-amino-1,2,4-triazole, 10 mM and Batten, 1990). For XO, the technical control con- hypoxanthine, 0.1 M HEPES buffer (pH 8), 7% sucrose, sisted of 2 mM allopurinol added to the preincubation and 0.0002% Triton X-100. The reaction during the in- and incubation media. Treatment with 2 mM allopu- cubation was performed 4 times for 6-min cycles (2 min rinol inhibited 82% of the XO activity. In the same on, 2 min off, 2 min on) using a microwave as described samples, incubation with higher allopurinol concentra- above. The reaction was terminated by washing the tions (10 mM allopurinol) did not further inhibited XO sample twice in cold 0.1 M HEPES buffer (pH 7.4) plus activity. 5% sucrose. In the case of NOX localization, the follow- We decided to conduct in situ determination of oxi- ing preincubation medium was used: 2.0 mM cerium dase activity in fixed intact lung parenchyma because chloride, 10 mM 3-amino-1,2,4-triazole, 0.1 M HEPES these procedures preserve active sites of enzymes while buffer (pH 7.4), 7% sucrose, and 0.0002% Triton X-100. maintaining tissue/cell structure to provide a picture The samples were then incubated in the following com- as close as possible to what happens in the live bird. plete reaction medium: 2.0 mM cerium chloride, 10 mM Also, in the present investigation we used the cerium- 3-amino-1,2,4-triazole, 0.8 mM NADH, 0.1 M HEPES based method to assess oxidase activity in intact lung buffer (pH 7.4), 7% sucrose, and 0.0002% Triton X-100. tissue of broiler chickens; this method has been well Specificity of the reactions was demonstrated by (a) established to demonstrate oxidase activity in cells omitting the substrate (hypoxanthine or NADH), and (Robinson and Batten, 1990; Ellis and Grant, 2002). (b) using an inhibitor of XO (1.0 mM allopurinol) or an In the cerium-based method, oxidases present in the inhibitor of NOX (1.0 mM diphenyleneiodonium chlo- intact lung parenchyma sample (i.e., enzymes preserve ride, DPI). Specimens were then postfixed overnight in its active sites when tissues are properly fixed) were re- 1% (wt/vol) osmium tetroxide, dehydrated, infiltrated, acted with its substrate (i.e., hypoxanthine for XO and and embedded. Grids were examined and photographed NADH for NOX) to produce superoxide, which in turn at 100 kV (1200EX transmission electron microscope, was dismutased by tissue superoxide dismutase to pro- JEOL USA Inc., Peabody, MA). duce H2O2. The addition of aminotriazole to the reac- With this method, the intact active site of the oxi- tion medium inactivated catalase and peroxidases pres- dases (XO and NOX) is preserved and when it is react- ent in the tissue, thus sparing the H2O2 that had been 982 BAUTISTA-ORTEGA ET AL. produced; H2O2 was then reacted with cerium chloride Xanthine oxidase activity was determined in lung to produce cerium perhydroxide. The specificity of this homogenates using an assay kit based on a multistep staining (i.e., cerium perhydroxide precipitates) was enzymatic reaction where the end product was resoru- demonstrated by the fact that perhydroxide deposits fin, a highly fluorescent compound, which was subse- were only seen in cells when substrate was present; no quently analyzed using an excitation wavelength of 520 staining was detected in the absence of substrate (pic- nm and an emission wavelength of 595 nm. The XO ture not shown). activity measure by the kit was in μU/mL homogenate Determination of NOX and XO Activity. The speci- (Cayman Chemical Company), from which specific XO mens were examined with an Olympus FV1000 confo- activity was calculated based on protein determination cal microscope (Olympus America Inc., Center Valley, in the lung homogenates and presented as μU/g of pro- PA). The reflective mode was used with the laser (λ tein. Downloaded from = 488). Routine observations were made by simultaneously generating a phase-contrast image on a split Determination of Nitrotyrosine Level screen. The digitalized images were captured by com- puter and analyzed using the software (Image J). Be- in Lung Tissue http://ps.oxfordjournals.org/ cause the majority of reflectance units were deemed to Sections from lungs in which XO had been localized have come from the oxidase activity (XO or NOX), the were picked up on nickel grids and processed in the direct readings of reflectance units were presented. microwave set at 250 W and 37 to 42°C. Grids were oxidized with 2% (wt/vol) periodic acid in a 6-min mi- Determination of XO and NOX Activity crowave cycle (2 min on, 2 min off, 2 min on) at 250 in Lung Homogenates W, 37 to 42°C followed by four × 1-min washes in PBS (pH 7.2) in a microwave cycle (1 min on). Grids were Frozen lung samples (0.5 g) were homogenized in 0.5 floated on PBS blocker [PBS plus 4% (wt/vol) cold wa- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 mL of ice-cold 50 mM Tris-HCl buffer (pH 7.4) contain- ter fish gelatin (Sigma Chemical Co., St. Louis, MO)] ing 10 μg/mL of antipain, 1 mM EDTA, and 1 mM of for a single 6-min microwave cycle. The grids were then dithiothreitol, centrifuged at 10,000 × g for 10 min at floated with polyclonal rabbit anti-nitrotyrosine, IgG 4°C and then frozen at −80°C until assayed. Homoge- fraction (Millipore Corporation, Billerica, MA) diluted nate protein content was determined using an assay kit 1:50 with PBS blocker and reacted for a single 6-min based on the Bradford method, which takes advantage microwave cycle. After 2 washes with PBS blocker (2 of the color change of Coomassie dye when it binds to × 1 min in the microwave), grids were washed 2× with proteins in acidic medium. When the dye binds, there Tris-HCI saline buffer (pH 7.6; TBS) plus 4% (wt/vol) is an immediate shift of the absorption maximum from cold water fish gelatin with 1 min/wash in the micro- 465 to 595 nm with a simultaneous change in color from wave. Grids were then reacted for a single 6-min micro- brown to blue. The assay was conducted according to wave cycle with donkey anti-rabbit IgG labeled with 12 the manufacturer’s instructions (Cayman Chemical nm colloidal gold (Jackson Immunoresearch Labs Inc., Company, Ann Arbor, MI). West Grove, PA) diluted 1:30 with TBS blocker. Grids The NAD(P)H oxidase activity in lung tissue ho- were then washed 3 × 1 min/microwave cycle with TBS mogenates was assessed through the oxidation of added blocker, followed by 3 × 1 min/microwave cycle with NADH oxidase-generated superoxide (Vincent et al., deionized water. Control grids were incubated with 2007). On a white 96-well plate, a 50-μL sample was only the gold-labeled secondary antibody. Data on NO mixed with 80 μL of Tris-HCl with either 20 μL of ad- depletion were presented as the number of immunogold ditional Tris-HCl or 20 μL of 1 mM DPI (final concen- particles/100 μm2. tration 100 μM) to inhibit NOX oxidase. The reaction was started by adding 50 μL of 600 μM NADH solution to a final concentration of 150 μM NADH and the ki- Determination of Physiological netics of the reaction curve was sampled every 1 min and Morphological Parameters for 10 min. The decrease in absorbance in the presence of DPI was subtracted from the decrease in absorbance In the birds that were humanely killed at the sam- in the absence of DPI to give a measure of oxidation of pling times, the heart was dissected to determine right NADH by NOX (NOX activity measured as a change in ventricle weight/total ventricle weight ratio (RV/TV; absorbance units per minute). Subsequently, the volu- Burton et al., 1968). The hematocrit (%, Hc) was re- metric NOX activity was calculated (U/L) as follows, corded at the sampling times. Clinical PHS was con- using an extinction coefficient (ε) of 6.22 L/mM per cm firmed by a RV:TV ratio higher than 0.28 and by the (Fasman, 1989): presence of ascitic fluid in the abdominal cavity.

volumetric NOX activity (U/L) = absorbance Statistical Analysis at 340 nm × 1,000 × dilution × volume/ε All data were analyzed as a completely random- × homogenate volume. ized design using a 1-way ANOVA (Jandel Scientific, XANTHINE OXIDASE ACTIVITY IN HYPOXIC CHICKENS 983 1994), and when statistical differences were detected activity compared with birds in the CTL-NOR group. the means were separated using the Student-Newman- In general, hypoxic chickens, regardless of dietary treat- Keuls method. A probability level of P ≤ 0.05 was ment, showed a lower (P < 0.05) XO activity than the considered statistically significant. All results were pre- CTL-NOR birds. To our knowledge, no work has been sented as means ± SE. published using an in situ approach to study the effects of diet on oxidase activity, namely XO and NOX, in relation to PHS in broiler chickens. RESULTS AND DISCUSSION Previously, we showed that supplemental Arg and antioxidant vitamins E and C improved cardiovascu- Cytochemical Localization of Xanthine-(XO) lar performance in vivo in broiler chickens exposed to and NAD(P)H-Oxidase (NOX) chronic hypobaric hypoxia or cold temperatures (Ruiz- Downloaded from Feria, 2009; Bautista-Ortega and Ruiz-Feria, 2010), Both enzymes, NOX (Figure 1A and B) and XO and improved pulmonary arterial reactivity in vitro (Figure 1C), localized in the vicinity of the pulmonary (Bautista-Ortega et al., 2013), suggesting that anti- vascular endothelial cell of all chickens. In the CTL- oxidant vitamins reduced the effects of ROS, whereas NOR chickens (picture not included) and CTL-HYP supplemental Arg ensured enough substrate for NO http://ps.oxfordjournals.org/ chickens, NOX localized within vesicles, outer mito- production, resulting in a healthier endothelium and chondrial membrane, as well as in the cell membrane higher availability of NO. However, we did not have of endothelial cells (Figure 1A). Hypoxic chickens from information regarding the distribution or the activity all dietary treatments also showed that NOX localized of XO and NOX in hypoxic chicken lung, including in endothelial cell junctions (Figure 1B). The XO local- the pulmonary artery endothelium. The present study ized in similar cellular domains as NOX, as well as in shows that there is XO activity in the lung parenchyma the cell membrane of red blood cells (Figure 1B and C).

(composed of different anatomic elements including at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Although XO and NOX have been associated with small vessels, small arteries, and parabronchial walls), vascular diseases, including PH in humans (Warwick which may be related to local stimuli (i.e., hypoxia) in et al., 2008; Guzik and Griendling, 2009) and in the parenchymal cell populations as well as from circulat- rodent model (Zalba et al., 2000; Jankov et al., 2008), ing XO attached to the cell membrane. in chickens the role of these enzymes has been less stud- However, in the present work we found that exposure ied, especially in the pulmonary artery endothelial cell to hypobaric hypoxia led to a reduction in XO activ- vicinity, the tissue where the primary lesion in chicken ity, which is in contrast with the response of mamma- PHS occurs. To our knowledge, this is the first time lian endothelial cell cultures exposed to either severe that NOX and XO are demonstrated ultrastructural- or moderate hypoxia (Poss et al., 1996; Kelley et al., ly in the intact lung of hypoxic and normoxic broiler 2006). Also, chickens respond differently from humans chickens. These results suggest that both enzymes may to dietary allopurinol, an inhibitor of XO; allopurinol- be important sources of ROS in the vicinity of chicken treated humans show improvements in PH symptoms pulmonary artery endothelial cell. (Pacher et al., 2006), whereas allopurinol-fed chickens Our results on NOX localization agree with previ- showed signs of toxicity manifested as significant BW ous reports that showed NOX activity confined to cell loss and increased oxidative stress (Carro et al., 2010). membrane rafts or within endosomal/phagosomal com- This could be due to the fact that XO is necessary for partments (Terada, 2006; van der Vliet, 2008). In ad- the production of UA, which acts as a natural anti- dition, our findings that XO localized within vesicles oxidant; for instance, it has been reported that a 33% on endothelial cells agree with the results by Poss et decrease in UA production in birds leads to an increase al. (1996), who reported that intracellular XO was con- in oxidative stress (Stinefelt et al., 2005). In the present tained in organelles, possibly endosomes. These authors investigation hypoxic chickens had a 60% lower average measured XO activity in whole blood and detected light intensity (235.7 ± 32 reflectance units) than the negligible XO activity in the serum and white blood normoxic controls (586 ± 43 reflectance units). A reduc- cells, suggesting that the XO activity was present in tion in XO activity may be related to a lower produc- red blood cells. tion of UA and a concomitant decrease in antioxidant capacity. Our results showed that, compared with the NOX and XO Activity in Intact Lung nonsupplemented hypoxic chickens (CTL-HYP; RLI = Parenchyma and in Lung Homogenates 305), supplemented chickens (AEC-HYP; RLI = 454) had higher (P < 0.05) XO activity, comparable with Chickens in the CTL-HYP group had lower (P < those found in young chickens grown under normoxia 0.05) XO activity than birds in the AEC-HYP and (RLI = 540; Figure 2), suggesting that the nutritional CTL-NOR groups, whereas the XO activity in CTL- supplementation slowed the negative effects of hypo- HYP birds did not differ from that in chickens from the baric hypoxia on XO activity, which may be associated HA-HYP group (Figure 2). Although chickens in the with an increase in UA. Uric acid may function as an AEC-HYP group had a higher (P < 0.05) XO activity antioxidant in the plasma; however, it may switch to than birds in the CTL-HYP group, still had a lower XO prooxidant (urate radical) when other antioxidants be- 984 BAUTISTA-ORTEGA ET AL. Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 1. Localization of NAD(P)H-oxidase (NOX) and xanthine oxidase (XO) in broiler chicken lung histological sections (transmission electron microscope). For NOX, sections were reacted with 2.0 mM cerium chloride, 10 mM 3-amino-1,2,4-triazole and 0.8 mM NADH. For XO, sections were reacted with 10.0 mM cerium chloride, 10 mM 3-amino-1,2,4-triazole, and 1 mM NADH. A. A 28-d-old hypoxic chicken. Deposition of H2O2 (arrowheads) is evident in cell membrane of parabronchial epithelium (PBE) end of endothelial cells (EC). B. A 28-d-old hypoxic chicken. Deposition of H2O2 (arrowheads) is evident in endothelial cell junction, cell membrane of red blood cells (RBC). C. A 28-d-old hypoxic chicken. Deposition of H2O2 is evident (arrowheads) in cell membrane of granular cells (GC). AC = air capillary. XANTHINE OXIDASE ACTIVITY IN HYPOXIC CHICKENS 985 es in NOX activity among dietary treatments, whereas they differ in XO activity. In this regard, it is possible that during chronic hypobaric hypoxia activation of NOX and XO do not correlate as in in vitro conditions. The NOX activity measured in the lung homogenates showed no effect due to dietary treatment or hypoxia (Table 1), agreeing with the results obtained using the in situ approach. The sources of H2O2 in hypertensive individuals include uncoupled eNOS (Sud et al., 2007), oxidases such as XO and NOX (Zalba et al., 2000; Jankov et al., 2008; Downloaded from Warwick et al., 2008; Guzik and Griendling, 2009), and uncoupled mitochondria (Maxwell et al., 1996; Iqbal et al., 2001a,b). In the present investigation almost all the H2O2 measured was produced by oxidases, namely http://ps.oxfordjournals.org/ Figure 2. Xanthine- (XO) and NADH-oxidase (NOX) activities XO and NOX. To reduce the possibility of confounding measured through the average reflected light intensity (RLI, confocal sources of H2O2, including uncoupled mitochondria, the microscopy) of perhydroxide deposits in lung tissue of hypoxic (H, 28 following considerations were taken into account: first, d old, n = 5–6 per dietary treatment) or normoxic (N, 10 d old, n = 3) broiler chickens. Oxidases in cryostat sections were reacted with 2.0 to XO and NOX activity was assessed in clinically healthy 10.0 mM cerium chloride, 10 mM 3-amino-1,2,4-triazole, and 0.8 mM broiler chickens. It has been shown that basal produc- NADH (NOX), or 10 mM hypoxanthine (XO). Hypoxic birds were fed tion of H2O2 in the mitochondria of clinically healthy a control diet (CTL, commercial corn-soybean meal-based diet), a high Arg diet (HA, control diet plus 0.8% Arg), or a high Arg and vitamin chicken is lower than in the mitochondria of birds with diet (AEC, HA plus vitamin E at 200 IU/kg of feed and vitamin C PHS (Iqbal et al., 2001a,b). Second, to ensure that the at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 at 500 mg/L of drinking water); normoxic birds were fed the control sources of H2O2 came from XO and NOX, we used diet. Means with a different letter (a–c) for the same enzyme differ specific inhibitors. For example, regarding XO activity, significantly (P < 0.05). the reported activity was that inhibited by allopurinol, and the NOX activity reported was the one that was come limited, but the prooxidant capability of UA is able to be inhibited by diphenyleneiodonium chloride. reversed by increasing levels of VC (Abuja, 1999). In Overall, if there was any H2O2 coming from uncoupled addition, VC has been shown to restore the antioxidant mitochondria or other sources, such contributions were capability of VE (Guney et al., 2007). Thus, with high equally present in all treatment groups. levels of VE and VC in the diet, UA may be an effec- One of the limitations of the present study was that tive antioxidant. Furthermore, the results on in situ XO the assessment of the activity of the different NOX iso- activity were similar to those obtained on XO activity forms expressed in parenchymal cell populations was in lung homogenates (Table 1). not performed. However, it has been documented that The NOX activity was affected by neither diet nor the analysis of total lung or airway mRNA revealed hypoxia in clinically healthy birds (Figure 2). It has the presence of substantial amount of NOX isoform been reported that in endothelial cells in vitro NOX 2 (NOX2), dual oxidase (DUOX) 1, and DUOX2, activation is necessary for increased XO activity and and low amounts of NOX1 and NOX4 (Lassègue and superoxide production in a NOX-dependent manner Griendling, 2010). In addition, in the lung, expression (McNally et al., 2003). Our results do not agree with of NOX primarily originates from recruited lung mac- this observation because we did not find any differenc- rophages (NOX2), although other NOX/DUOX iso-

Table 1. Effect of diet and hypobaric hypoxia on parameters related to pulmonary hypertension (right ventricle to total ventricle weight ratio, RV/TV; hematocrit, HT, %), NADH activity (NOX, U/L), and xanthine oxidase activity (XO, μU/g of protein) in lung homogenates of broiler chickens

Treatment

Parameter CTL1-HYP2 HA-HYP AEC-HYP CTL-NOR

RV/TV 0.26 ± 0.03a 0.26 ± 0.01a 0.27 ± 0.01a 0.18 ± 0.01b HT 39 ± 4.6a 41 ± 1.9a 39 ± 3.3a 27 ± 0.9b NOX 1.59 ± 0.4 1.36 ± 0.4 2.0 ± 0.4 1.49 ± 0.4 XO 0.43 ± 0.6b 0.66 ± 0.6b 1.21 ± 0.6b 3.87 ± 0.6a a,bMeans within a row, and within the same parameter, lacking a common superscript differ (P ≤ 0.05). 1CTL, control basal diet containing 3,200 kcal of ME/kg of feed, 23% CP, 1.55% (wt/wt) arginine, and 40 IU of VE/kg of feed; HA, the CTL diet plus 0.8% (wt/wt) supplemental l-arginine HCl; AEC, the Arg diet plus 200 IU α-tocopherol/kg of feed and 500 mg of ascorbic acid/L of drinking water. 2HYP, hypoxic birds raised at a simulated 10,000 ft (3,000 m above sea level); NOR, normoxic birds raised at sea level in a pen next to the hypobaric chamber. 986 BAUTISTA-ORTEGA ET AL. Downloaded from http://ps.oxfordjournals.org/ Figure 3. Nitrotyrosine concentration (particles), an indicator of Figure 4. Nitrotyrosine concentration (particles), an indicator of oxidative stress and nitric oxide depletion, in the lung parenchyma of nitric oxide depletion, in the lung parenchyma of clinically healthy clinically healthy broiler chickens grown under hypoxia (H) or nor- chickens (n = 9) and chickens with clinical signs of pulmonary hyper- moxia (N). Lung tissue samples, processed for electron microscopy, tension (IPAH; n = 3). Lung tissue samples, processed for electron were picked up in nickel grids and were reacted with the following microscopy, were picked up in nickel grids and were reacted with the reagents in the order shown: 2% periodic acid, polyclonal rabbit anti- following reagents in the order shown: 2% periodic acid, polyclonal nitrotyrosine antibody, blocked with 4% fish gelatin, and donkey anti- rabbit anti-nitrotyrosine antibody, blocked with 4% fish gelatin, don- rabbit antibody labeled with 12 nm colloidal gold (visualized as par- key anti-rabbit antibody labeled with 12 nm colloidal gold (visualized ticles). Control grids were incubated with only gold-labeled secondary as particles). Control grids were incubated with only gold-labeled sec- antibody. Hypoxic birds were fed a control diet (CTL-H, commercial ondary antibody. Means with different letters (a,b) differ significantly at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 corn-soybean meal based diet; n = 3), a high Arg diet (HA-H, control (P < 0.05). diet plus 0.8% Arg; n = 4), or a high Arg and vitamin diet (AEC-H, Arg plus vitamin E at 200 IU/kg of feed and vitamin C at 500 mg/L of drinking water; n = 2); normoxic birds were fed the control diet Once formed, NO can react with hemoglobin (form- (CTL-N; n = 3). ing iron-nitrosyl-hemoglobin or nitrate), superoxide (forming peroxynitrite), soluble guanylyl cyclase, and cytochrome C oxidase (Chen et al., 2008). We were in- enzymes are largely expressed in nonphagocytic cells terested in the reaction of NO with superoxide because within the lung such as airway epithelium, pulmonary protein tyrosine nitration is an established biomarker of endothelial cells, fibroblasts, and smooth muscle cells oxidative stress and the molecular signal of peroxyni- (Lassègue and Griendling, 2010). Thus, it is likely that trite formation during development, oxidative stress, in the present study the majority of NOX activity de- and aging (Yeo et al., 2008). In the present study, the tected in lung tissue came from recruited macrophages reaction between NO and superoxide that took place (NOX2). in the live bird led to the formation of peroxynitrite; in turn, peroxynitrite reacted with tyrosine residues of Nitrotyrosine (NO Availability) proteins (nitrotyrosine), which were then labeled with in Lung Tissue immunogold particles. The reaction between superoxide and NO disturbs the steady state concentration of NO; There was no difference in nitrotyrosine levels among superoxide generated either intracellularly or extracel- birds grown under hypobaric hypoxia, regardless of di- lularly during NO generation results in a concomitant etary treatment, or compared with birds grown under increase in oxidative intermediates with a decrease in normoxia (Figure 3). However, there was a tendency steady-state NO concentration and a proportional re- (P = 0.10) for the CTL-HYP birds to have a higher duction in the levels of soluble guanylyl cyclase, the tyrosine level than the rest of the hypoxic groups of mediator of NO in smooth muscle relaxation (Thom- birds. In the same line, the CTL-NOR birds had the as et al., 2006). Our results suggest that in chickens same nitrotyrosine level as the hypoxic supplemented that developed PHS there was a higher number of NO birds. The lack of differences among treatments may molecules reacting with superoxide molecules and were have been the result of the limited number of samples subsequently lost to peroxynitrite at the expense of NO processed for cytochemistry. However, we found that bioavailability, reducing the ability of the pulmonary birds with clinical signs of PHS (i.e., ascitic fluid in arteries to relax, further amplifying the incidence of the abdominal cavity and a high RV:TV ratio) showed pulmonary hypertension. Furthermore, peroxynitrites higher (P < 0.001) levels of peroxynitrite formation in are reactive species of nitrogen and oxygen that modify the vascular endothelium and its vicinity than in clini- biomolecules, including DNA, lipids, and proteins, lead- cally healthy hypoxic chickens (Figure 4), indicating a ing to cell and tissue damage (Yeo et al., 2008). The higher oxidative stress in chickens with PHS, and lower present investigation is the first to propose the use of levels of NO bioavailability. nitrotyrosine level in lung tissue of chickens as a marker XANTHINE OXIDASE ACTIVITY IN HYPOXIC CHICKENS 987 for NO bioavailability, apart from its well-established challenge with epinephrine in broiler chickens grown use as a marker for oxidative stress. under cold temperatures or hypobaric hypoxia and fed Arg, VE, and VC (Ruiz-Feria, 2009; Bautista-Ortega PHS-Related Parameters and Ruiz-Feria, 2010; Bautista et al., 2013).

Hematocrit (%) is an indicator of hypoxemia (low ACKNOWLEDGMENTS oxygen tension in the blood) because when chickens are exposed to chronic hypoxia there is an increase in This research was funded by Hatch project H70280. the production of red blood cells. As expected, hypoxic J. Bautista-Ortega is a Tom Slick Senior Graduate broiler, regardless of diet, had higher Hc than chickens Fellow at Texas A&M University and was financially grown under normoxia (Table 1). Although PHS broil- supported by Colegio de Postgraduados, Mexico. We Downloaded from ers presented almost 4 percentage units higher Hc than appreciate the technical assistance of Stan Vitha at the healthy-looking hypoxic birds, this difference did the Microscopy and Imaging Centre. Results on NOX not reach statistical significance (43 ± 5.0% and 40 ± activity and nitrotyrosine (oxidative stress) have been 1.6%, for the PHS and the healthy-looking group of previously presented at the Microscopy and Microanal- broilers, respectively). ysis Society meeting, 2010. http://ps.oxfordjournals.org/ The RV/TV ratio is used as an indicator of pulmonary hypertension, which directly causes right ventricu- REFERENCES lar hypertrophy in broiler chickens. Clinically healthy broiler chickens with normal PAP have RV/TV ratios Abuja, P. M. 1999. Prooxidant effects of uric acid and other aque- ranging from 0.15 to 0.27, whereas pulmonary hyperten- ous antioxidants are prevented by ascorbic acid. Atherosclerosis 144:46–47. sion causes RV/TV ratios to rise above 0.28 (Wideman, Al-Khalidi, U. A. S., and T. H. Chaglassian. 1965. The species distri-

2001). Thus, although the hypoxic chickens appeared bution of xanthine oxidase. Biochem. J. 97:318–320. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 clinically healthy, they all were facing pulmonary hy- Bautista-Ortega, J., and C. A. Ruiz-Feria. 2010. L-arginine and pertension as demonstrated by the RV/TV ratios close antioxidant vitamins E and C improve the cardiovascular performance of broiler chickens grown under chronic hypobaric hy- to the 0.28 threshold (0.26 ± 0.031, 0.26 ± 0.006, and poxia. Poult. Sci. 89:2141–2146. 0.27 ± 0.009 for the CTL-HYP, HA-HYP, and AEC- Bautista-Ortega, J., J. N. Stallone, and C. A. Ruiz-Feria. 2013. Ef- HYP, respectively; Table 1). However, hypoxic broilers fects of arginine and antioxidant vitamins on pulmonary artery were significantly more hypertensive (P < 0.05) than reactivity to phenylephrine in the broiler chicken. Poult. Sci. 92:1062–1072. CTL-NOR broilers (Table 1). In addition, broilers that Burton, R. R., E. Besch, and A. Smith. 1968. Effect of chronic hy- developed PHS were significantly (P < 0.05) more hypoxia on the pulmonary arterial blood pressure of the chicken. pertensive than hypoxic healthy-looking broiler chick- Am. J. Physiol. 214:1438–1442. ens (0.42 ± 0.060 and 0.26 ± 0.090 for the PHS and the Carro, M. D., E. Falkenstein, K. P. Blemings, and H. Klandorf. 2009. Determination of xanthine oxidoreductase activity in broil- healthy-looking group of broiler chickens, respectively). ers: Effect of pH and temperature of the assay and distribution These results agree with those reported elsewhere using in tissues. Poult. Sci. 88:2406–2414. a larger population of broilers exposed to chronic hypo- Carro, M. D., E. Falkenstein, W. J. Radke, and H. Klandorf. 2010. Effects of allopurinol on uric acid concentrations, xanthine oxido- baric hypoxia (Bautista-Ortega and Ruiz-Feria, 2010). reductase activity and oxidative stress in broiler chickens. Comp. In conclusion, the present investigation provides evi- Biochem. Physiol. C Toxicol. Pharmacol. 151:6. dence that XO and NOX are important sources of su- Chen, K., R. N. Pittman, and A. S. Popel. 2008. Nitric oxide in the peroxide in the vicinity of the pulmonary artery endo- vasculature: Where does it come from and where does it go? A quantitative perspective. Antioxid. Redox Signaling 10:1185– thelium. In addition, concurrent supplementation with 1198. Arg and antioxidant VE and VC did not have an effect Cheng, J., J. S. Ou, H. Singh, J. R. Falck, D. Narsimhaswamy, K. A. on NOX activity, and on oxidative stress and NO bio- Pritchard, and M. L. Schwartzman. 2008. 20-hydroxyeicosatet- availability in the vicinity of pulmonary artery endothe- raenoic acid causes endothelial dysfunction via eNOS uncoupling. Am. J. Physiol. Heart Circ. Physiol. 294:H1018–H1026. lium, as measured by the nitrotyrosine levels, but there Dudzinski, D. M., and T. Michel. 2007. Life history of eNOS: Part- was a tendency (P = 0.10) for the CTL-HYP birds to ners and pathways. Cardiovasc. Res. 75:247–260. have a higher tyrosine level (lower NO bioavailability) Ellis, E. A., and M. B. Grant. 2002. Cytochemical localization of than the hypoxic birds fed Arg or Arg, VE, and VC. H2O2 in biological tissues. Pages 3–12 in Methods in Molecular Biology. D. Armstrong, ed. Humana Press Inc., Totowa, NJ. On the other hand, we found that chronic hypobaric Enkvetchakul, B., W. Bottje, N. Anthony, and R. Moore. 1993. hypoxia reduced the activity of XO, but the concur- Compromised antioxidant status associated with ascites in broil- rent supplementation of Arg, VE, and VC restored the ers. Poult. Sci. 72:2272–2280. XO activity. The mechanism(s) through which Arg and Fasman, G. D. 1989. Practical Handbook of Biochemistry and Mo- lecular Biology. CRC Press, Boca Raton, FL. antioxidant vitamins modulate XO activity in broiler Govers, R., and T. J. Rabelink. 2001. Cellular regulation of en- chickens under hypobaric hypoxia and undergoing pul- dothelial nitric oxide synthase. Am. J. Physiol. Renal Physiol. monary hypertension remains to be determined. Thus, 280:F193–F206. Guney, M., B. Oral, H. Demirin, N. Karahan, T. Mungan, and N. the dual role of XO, which produces superoxide but Delibas. 2007. Protective effects of vitamins C and E against also UA, a potent antioxidant, may explain the im- endometrial damage and oxidative stress in fluoride intoxication. proved pulmonary arterial relaxation in vivo after a Clin. Exp. Pharmacol. Physiol. 34:467–474. 988 BAUTISTA-ORTEGA ET AL.

Guzik, T. J., and K. K. Griendling. 2009. NADPH oxidases: Molecu- xanthine oxidase activity by hypoxia. Am. J. Physiol. 270:L941– lar understanding finally reaching the clinical level? Antioxid. L946. Redox Signal. 11:2365–2370. Robinson, J. M., and B. E. Batten. 1990. Localization of cerium- Iqbal, M., D. Cawthon, R. F. Wideman, and W. G. Bottje. 2001a. based reaction products by scanning laser reflectance confocal Lung mitochondrial dysfunction in pulmonary hypertension syn- microscopy. J. Histochem. Cytochem. 38:315–318. drome. I. Site-specific defects in the electron transport chain. Ruiz-Feria, C. A. 2009. Concurrent supplementation of arginine, vi- Poult. Sci. 80:485–495. tamin E, and vitamin C improve cardiopulmonary performance Iqbal, M., D. Cawthon, R. F. Wideman, and W. G. Bottje. 2001b. in broilers chickens. Poult. Sci. 88:526–535. Lung mitochondrial dysfunction in pulmonary hypertension syn- Selemidis, S., C. G. Sobey, K. Wingler, H. H. H. V. Schmidt, and G. drome. II. Oxidative stress and inability to improve function R. Drummond. 2008. NADPH oxidases in the vasculature: Mo- with repeated additions of adenosine diphosphate. Poult. Sci. lecular features, roles in disease and pharmacological inhibition. 80:656–665. Pharmacol. Ther. 120:254–291.

Jandel Scientific. 1994. SigmaStat Statistical Software User’s Manu- Stinefelt, B., S. S. Leonard, K. P. Blemings, X. Shi, and H. Klandorf. Downloaded from al. Jandel Scientific Software, San Rafael, CA. 2005. Free radical scavenging, DNA protection, and inhibition of Jankov, R. P., C. Kantores, J. Pan, and J. Belik. 2008. Contribution lipid peroxidation mediated by uric acid. Ann. Clin. Lab. Sci. of xanthine oxidase-derived superoxide to chronic hypoxic pulmo- 35:37–45. nary hypertension in neonatal rats. Am. J. Physiol. Lung Cell. Sud, N., S. Sharma, D. A. Wiseman, C. Harmon, S. Kumar, R. C. Mol. Physiol. 294:L233–L245. Venema, J. R. Fineman, and S. M. Black. 2007. Nitric oxide Julian, R. J. 2007. The response of the heart and pulmonary arter- and superoxide generation from endothelial NOS: Modulation by ies to hypoxia, pressure, and volume. A short review. Poult. Sci. HSP90. Am. J. Physiol. Lung Cell. Mol. Physiol. 293:L1444– http://ps.oxfordjournals.org/ 86:1006–1011. L1453. Kelley, E. E., T. Hock, N. K. H. Khoo, G. R. Richardson, K. K. Tan, X., H. Song-Hua, and W. Xiao-Long. 2007a. Possible role of ni- Johnson, P. C. Powell, G. I. Giles, A. Agarwal, J. R. Lancaster, tric oxide in the pathogenesis of pulmonary hypertension in broil- and M. M. Tarpey. 2006. Moderate hypoxia induces xanthine ers: A synopsis. Avian Pathology. Taylor & Francis Ltd. 261–267. oxidoreductase activity in arterial endothelial cells. Free Radic. Tan, X., W.-D. Sun, J.-C. Li, J.-Q. Pan, Y.-J. Liu, J.-Y. Wang, and Biol. Med. 40:952–959. X.-L. Wang. 2007b. l-Arginine prevents reduced expression of Lassègue, B., and K. K. Griendling. 2010. NADPH oxidases: Func- endothelial nitric oxide synthase (NOS) in pulmonary arterioles tions and pathologies in the Vasculature. Arterioscler. Thromb. of broilers exposed to cool temperatures. Vet. J. 173:151–157.

Vasc. Biol. 30:653–661. Terada, L. S. 2006. Specificity in reactive oxidant signaling: Think at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Lorenzoni, A. G., and C. A. Ruiz-Feria. 2006. Effects of vitamin E globally, act locally. J. Cell Biol. 174:615–623. and L-arginine on cardiopulmonary function and ascites param- Thomas, D. D., L. A. Ridnour, M. G. Espey, S. Donzelli, S. Ambs, eters in broiler chickens reared under subnormal temperatures. S. P. Hussain, C. C. Harris, W. DeGraff, D. D. Roberts, J. B. Poult. Sci. 85:2241–2250. Mitchell, and D. A. Wink. 2006. Superoxide fluxes limit nitric Martinez-Lemus, L. A., R. K. Hester, E. J. Becker, J. S. Jeffrey, oxide-induced signaling. J. Biol. Chem. 281:25984–25993. and T. W. Odom. 1999. Pulmonary artery endothelium-depen- van der Vliet, A. 2008. NADPH oxidases in lung biology and pathol- dent vasodilation is impaired in a chicken model of pulmonary ogy: Host defense enzymes, and more. Free Radic. Biol. Med. hypertension . Am. J. Physiol. Regul. Integr. Comp. Physiol. 44:938–955. 277:R190–R197. Vincent, A. M., L. Perrone, K. A. Sullivan, C. Backus, A. M. Sastry, Maxwell, M.H., G.W. Robertson, and C. Farquharson. 1996. Evi- C. Lastoskie, and E. L. Feldman. 2007. Receptor for advanced dence of ultracytochemical mitochondria-derived hydrogen per- glycation end products activation injures primary sensory neu- oxide activity in myocardial cells from broiler chickens with an rons via oxidative stress. Endocrinology 148:548–558. ascites syndrome. Res. Vet. Sci. 61:7–12. Warwick, G., P. S. Thomas, and D. H. Yates. 2008. Biomarkers in McNally, J. S., M. E. Davis, D. P. Giddens, A. Saha, J. Hwang, S. pulmonary hypertension. Eur. Respir. J. 32:503–512. Dikalov, H. Jo, and D. G. Harrison. 2003. Role of xanthine oxido- Wedgwood, S., and S. M. Black. 2003. Role of reactive oxygen spe- reductase and NAD(P)H oxidase in endothelial superoxide pro- cies in vascular remodeling associated with pulmonary hyperten- duction in response to oscillatory shear stress. Am. J. Physiol. sion. Antioxid. Redox Signal. 5:759–769. Heart Circ. Physiol. 285:H2290–H2297. Wideman, R. F. 2001. Pathophysiology of heart/lung disorders: Nain, S., B. Ling, B. Bandy, J. Alcorn, C. Wojnarowicz, B. Laarveld, Pulmonary hypertension syndrome in broiler chickens. World’s and A. A. Olkowski. 2008a. The role of oxidative stress in the Poult. Sci. J. 57:289–307. development of congestive heart failure in a chicken genotype Wideman, R. F., K. R. Hamal, M. T. Bayona, A. G. Lorenzoni, D. selected for rapid growth. Avian Pathol. 37:367–373. Cross, F. Khajali, D. D. Rhoads, G. F. Erf, and N. B. Anthony. Nain, S., C. Wojnarowicz, B. Laarveld, and A. A. Olkowski. 2008b. 2011. Plexiform lesions in the lungs of domestic fowl selected for Effects of dietary vitamin E and C supplementation on heart susceptibility to pulmonary arterial hypertension: Incidence and failure in fast growing commercial broiler chickens. Br. Poult. histology. Anat. Rec. (Hoboken) 294:739–755. Sci. 49:697–704. Wideman, R. F., Y. K. Kirby, M. Ismail, W. G. Bottje, R. W. NRC. 1994. Nutrient Requirements of Poultry. 9th rev. ed. ed. Natl. Moore, and R. C. Vardeman. 1995. Supplemental L-arginine at- Acad. Press, Washington, DC. tenuates pulmonary-hypertension syndrome (ascites) in broilers. Pacher, P., A. Nivorozhkin, and C. Szabo. 2006. Therapeutic effects Poult. Sci. 74:323–330. of xanthine oxidase inhibitors: Renaissance half a century after Yeo, W. S., S. J. Lee, J. R. Lee, and K. P. Kim. 2008. Nitrosative the discovery of allopurinol. Pharmacol. Rev. 58:87–114. protein tyrosine modifications: Biochemistry and functional sig- Pan, J. Q., J. C. Li, X. Tan, W. D. Sun, J. Y. Wang, and X. L. nificance. BMB Rep. 41:194–203. Wang. 2007. The injury effect of oxygen free radicals in vitro on Zalba, G., F. J. Beaumont, G. S. Jose, A. Fortuno, M. A. Fortuno, cultured pulmonary artery endothelial cells from broilers. Res. J. C. Etayo, and J. Diez. 2000. Vascular NADH/NADPH oxidase Vet. Sci. 82:382–387. is involved in enhanced superoxide production in spontaneously Poss, W. B., T. P. Huecksteadt, P. C. Panus, B. A. Freeman, and hypertensive rats. Hypertension 35:1055–1061. J. R. Hoidal. 1996. Regulation of xanthine dehydrogenase and PROCESSING, PRODUCTS, AND FOOD SAFETY

Preparation and characterization of gelatin from collagen biomass obtained through a pH-shifting process of mechanically separated turkey meat

L. Du ,* L. Keplová ,† Z. Khiari ,* and M. Betti *1

* Department of Agricultural, Food and Nutritional Science, University of Alberta, Edmonton, T6G 2P5, AB, Canada; and † Department of Dairy, Fat and Cosmetic Science, Institute of Chemical Technology, Prague, 166 28, Czech Republic Downloaded from

ABSTRACT Gelatins were extracted from mechani- recovered from the collagen biomass had higher mo- cally separated turkey meat following 2 different ap- lecular weight components and significantly greater (P proaches. The first method was based on a 2-stage < 0.05) bloom value (353.2 g) compared with thermally batch extraction at 50 and 60°C, respectively, where- extracted gelatins. However, gelatin extracted at 60°C http://ps.oxfordjournals.org/ as the second method consisted of recovering gelatin possessed higher (P < 0.05) foaming properties, as well from a collagen biomass obtained during a pH-shifting as better emulsifying activity, than gelatin extracted process. The yield of gelatin produced by the latter from the 50°C treatment and the collagen biomass. The method was twice that obtained by the former method present study revealed that high-quality gelatins can (13.51 and 6.36%, respectively). The chemical composi- be prepared from mechanically separated turkey meat tion, as well as the rheological and the functional prop- through precipitation and thermal solubilization of col-

erties, of all extracted gelatins were evaluated. Gelatin lagen biomass obtained during a pH-shifting process. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Key words: mechanically separated turkey meat, pH-shifting process, collagen biomass, gelatin , rheological and functional property 2014 Poultry Science 93 :989–1000 http://dx.doi.org/ 10.3382/ps.2013-03609

INTRODUCTION Several studies were undertaken to investigate the recovery of valuable proteins from MSPM (Hrynets et al., According to the Food and Agriculture Organiza- 2010; Hrynets et al., 2011). Functional proteins from tion of the United Nations, the meat sector is grow- MSPM have been successfully isolated by either the ing rapidly as a result of the rising demand for poul- surimi or the pH-shifting processes. Due to high water try meat and its derivatives (FAO, 2012). The latest consumption and the elevated level of organic matter statistic estimated the global poultry meat production in the waste water associated with the surimi process, to be around 102 million t (FAOSTAT, 2013). Along the pH-shifting process became a popular method for with this large production, significant quantities of by- protein isolation from MSPM. In this technology, sarco- products and wastes are generated. In an attempt to plasmic and myofibrillar proteins are first solubilized at recover the meat adhering to the frames and bones and acidic or alkaline conditions then precipitated at their to add value to these by-products, a mechanical debon- isoelectric point. Protein isolates are the main prod- ing process was developed. This process is capable of ucts of the pH-shifting process. However, significant recovering up to 70% of the meat, usually referred to as amounts of fat and insoluble biomass (mainly consist- mechanically separated poultry meat (MSPM; Fron- ing of collagen) are also produced and discarded. The ing and McKee, 2001). The MSPM is considered to be collagen biomass, which is obtained as a by-product a low-cost meat product and has been widely incorpo- during the pH-shifting operation, may represent a valu- rated into a variety of meat product formulations, such able source for gelatin preparation. as sausages, nuggets, and patties (Froning and McKee, Gelatin is a multifunctional ingredient that has long 2001). However, the relatively high-fat bone and con- been used in the food industry as a gelling, thicken- nective tissue content in MSPM limits its industrial ing, and film-forming agent, as well as an emulsifier usage (Liang and Hultin, 2003). and stabilizer (Schrieber and Gareis, 2007). The source of collagen and the manufacturing process significantly affect the physicochemical and functional properties © 2014 Poultry Science Association Inc. of gelatin (Montero and Gómez-Guillén, 2000). Tra- Received September 6, 2013. Accepted December 19, 2013. ditional gelatin production involves the pretreatment 1 Corresponding author: [email protected] of raw material followed by extraction and purification

989 990 DU ET AL. steps (Karim and Bhat, 2009). Acid or alkali is usually used in the pretreatment steps to remove impurities and cleave the collagen crosslinks, then gelatin is a produced by a partial thermal denaturation of collagen. The majority of commercial gelatins come from porcine and bovine sources. However, due to some cultural, religious, or safety reasons (Gómez-Guillén et al.., 2007), mammalian gelatins may not meet the growing demand for halal and kosher products (Arnesen and Gildberg, 2007). Poultry by-product contains a high proportion of collagen, making it a potential alterna- Downloaded from tive source for gelatin extraction (Nakano et al., 2012). Recent studies showed that gelatin can be recovered from chicken skins (Sarbon et al., 2013), chicken and turkey heads (Du et al., 2013), as well as mechanically deboned chicken and turkey residues (Fonkwe and http://ps.oxfordjournals.org/ Singh, 1997; Rafieian et al., 2013). The preparation of gelatin from MSPM has not yet Figure 1. Preparation of collagen biomass by pH-shifting technol- been investigated. Therefore, the main objective of ogy. MTSM = mechanically separated turkey meat. this research was to explore the possibility of gelatin preparation from mechanically separated turkey meat (MSTM) using 2 different extraction methods: the Gelatin Preparation conventional manufacturing method and a newly de- Gelatin from MSTM was obtained from 2 different at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 veloped process. In the conventional method, MSTM extraction methods. was chemically pretreated to remove noncollagenous Conventional Thermal Process. This process was proteins, fat, and impurities, followed by an acid-aided based on chemical elimination of noncollagenous pro- swelling step, and then gelatin was obtained in a series teins, fat, and impurities, followed by thermal solubi- of batch extractions. In the newly developed approach, lization of collagen (Du et al., 2013). One kilogram of gelatin was produced after the recovery of a collagen MSTM was washed with distilled water at a ratio of 1:5 biomass from MSTM, through a pH-shifting process, (wt/vol) for 15 min followed by centrifugation at 8,000 followed by an acid solubilization and a thermal de- × g for 10 min at 4°C using an Avanti J-E High-Perfor- naturation. The advantage of the latter method would mance Centrifuge (Beckman Coulter Inc., Palto Alto, be the recovery of other important functional proteins CA). This step resulted in removing soluble impuri- (for example actin and myosin in the form of protein ties, and to dispose of noncollagenous proteins and fat isolate), along with collagen, making the process more the sample was successively treated with NaCl, NaOH, profitable. Physicochemical and functional properties and butanol. Briefly, the water washed sample was first of the extracted gelatins were also studied. mixed with 0.5 M NaCl solution at a ratio of 1:5 (wt/ vol) under stirring at 4°C with regular changes every 15 MATERIALS AND METHODS min for a total period of 45 min. The residue obtained after centrifugation at 20,000 × g for 10 min at 4°C was Materials treated with 0.25 M NaOH at a ratio of 1:10 (wt/vol) for 6 h with regular changes every 2 h. After centrifu- For the conventional extraction process, 3 separate gation, the residue was treated with 2.5% butanol at a batches of frozen MSTM were purchased from Lily- ratio of 1:10 (wt/vol) for 24 h with 3 changes. dale (Edmonton, AB, Canada) and thawed overnight The clean, defatted collagen material was recovered at 4°C before usage. Collagen biomass, a by-product after centrifugation and was subjected to a swelling obtained during the preparation of poultry protein iso- step using 0.05 M acetic acid at a ratio of 1:10 (wt/ late through the pH-shifting technology (Figure 1) and vol) for 4 h under stirring. The acid-treated sample was used in the newly developed process, was collected from washed twice with distilled water for 10 min each time. 3 separate batches of 20 kg of MSTM. The extraction Gelatin from MSTM was produced in a series of and separation of poultry protein isolate and collagen batch extractions at 50 and 60°C (Du et al., 2013). biomass were performed in the pilot plant at Food This was carried out by mixing the swelled sample with Processing Development Centre (Leduc, AB, Canada) distilled water at a ratio of 1:10 (wt/vol) in a jack- according to the acid-aided pH-shifting process as de- eted glass beaker (Chemglass Life Sciences, Vineland, scribed by Hrynets et al. (2011). Collagen biomass was NJ) connected to a Haake S7 heated bath circulator stored at −20°C in polyethylene bags until use. All the (Thermo Fisher Scientific, Newington, NH). In the first chemicals and reagents used in this study were of ana- stage, the mixture was kept under continuous stirring lytical grade. for 18 h at 50°C. Soluble gelatin was collected by filtra- GELATIN FROM COLLAGEN BIOMASS OF TURKEY 991 tion through a 1-mm pore size sieve (Fisherbrand US Gelatin extraction yield (%) = [Dry weight Standard Brass Test, Thermo Fisher Scientific Inc.). of extracted gelatin (g)/Dry weight of collagen Gelatin in the insoluble material was then extracted at 60°C for 6 h in the second stage. in the raw material (g)] × 100. Extracted gelatins were filtered using Whatman No.4 filter papers and deionised using Rexyn I-300 (H-OH) Color Measurement. The lightness (L*), redness or beads (Acros, Geel, Belgium) until the conductivity greenness (a*), and yellowness or blueness (b*) values value was lower than 50 μS/cm (Oakton Acorn CON6, of 6.67% (wt/vol) gelatin gels were obtained using a Vernon Hills, IL). Gelatin solution was then evaporated Minolta CR-400 colorimeter (Konica Minolta Sensing at 50°C by a Heidolph Rotavapor (Hei-VAP Collegiate Americas Inc., Ramsey, NJ). WB/G3) fitted with a Heidolph HB digital heating bath Electrophoretic Analysis. Sodium dodecyl sulfate- Downloaded from and Brinkmann auto-purge vacuum system (Heidolph poly acrylamide gel electrophoresis of gelatins was per- Brinkmann LLC., Schwabach, Germany), and freeze- formed according to the method of Khiari et al. (2013). dried (Labconco Corporation, Kansas City, MO). The gelatin sample was first dissolved in distilled water Newly Developed Process. This process was based at 55°C to a final concentration of 5 mg/mL and then on recovering collagen biomass from MSTM using the mixed with Laemmli sample buffer [65.8 mM Tris-HCl, http://ps.oxfordjournals.org/ pH-shifting technology as developed by Hrynets et al. pH 6.8, 2.1% (wt/vol) SDS, 26.3% (wt/vol) glycerol, (2011; Figure 1), followed by solubilization of collagen 0.01% (wt/vol) bromophenol blue] containing 5% (vol/ and thermal transformation. Collagen biomass was vol) β-mercaptoethanol at a ratio of 1:1 (vol/vol). The mixed with distilled water at a ratio of 1:5 (wt/vol) and mixture was heated to 90°C for 5 min to denature the stirred for 15 min followed by centrifugation at 8,000 proteins. Ten microliters of each sample and molecular × g for 10 min at 4°C. This step was repeated 3 more weight markers (Bio-Rad Laboratories Inc., Hercules,

times to remove fat. Collagen was solubilized in 0.5 M CA) were loaded on a precast 4 to 20% ready gel (Bio- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 acetic acid at a ratio of 1:100 (wt/vol) containing pep- Rad Laboratories Inc.). The analysis was performed sin (MP Biomedicals, LLC, Solon, OH) at a ratio 1:20 in a PowerPack Basic electrophoresis apparatus (Bio- (wt/wt) for 18 h at 4°C. Soluble collagen was recovered Rad Laboratories Inc.) at 120 V. The protein bands of by centrifugation at 8,000 × g for 10 min at 4°C and analyzed samples were visualized after staining in Coo- precipitated at pH 8 using 1M NaOH. The precipitated massie Brilliant Blue R250 and destaining in a mixture collagen was collected by centrifugation and washed 3 of distilled water, methanol, and acetic acid (70/20/10; times with water (1:5; wt/vol) to remove the salt (so- vol/vol/vol). dium acetate). Purified collagen was finally obtained Amino Acid Composition. Gelatin samples were after centrifugation at 20,000 × g for 10 min at 4°C. hydrolyzed under a vacuum at 160°C for 1 h in 6 M Extracted collagen was homogenized with distilled HCl containing 0.1% phenol and Norleucine was added water at a ratio of 1:3 (wt/vol) and gelatin was pro- as an internal standard. The amino acids were labeled duced after thermally treating the sample at 60°C for 6 with AccQ-Tag Reagent Kit (Waters, Milford, MA), h. The soluble gelatin was recovered by centrifugation and then determined by HPLC (Agilent 1200 Series) at 25°C and then freeze-dried. In total, 3 gelatin groups using AccQ-Tag C18 column (3.9 × 150 mm; Waters) were obtained: gelatin extracted in the first stage from with an absorbance detection at 254 nm. The amino MSTM (MSTMS1), gelatin extracted in the second acid composition was analyzed at the Institute for Bio- stage from MSTM (MSTMS2), and gelatin extracted molecular Design (University of Alberta, Edmonton, from collagen biomass (CB) Canada) and results were presented as residues per 100 residues. Characterization of Extracted Gelatins Rheological and Textural Properties Proximate Analysis. Moisture, ash and fat contents of Extracted Gelatins of the starting materials and extracted gelatins were determined following the method of AOAC (2000). Pro- Determination of Dynamic Viscoelastic Behav- tein content was obtained by a TruSpec CN carbon/ ior. The dynamic viscoelastic behavior study of gelatin nitrogen determinator (Leco Corp., St. Joseph, MI), samples was investigated by a Physica MCR 301 rhe- and the nitrogen conversion factors used in this study ometer (Anton Paar GmbH, Ashland, VA) according to were 6.25 for raw materials and 5.4 for extracted gela- the method of Du et al. (2013). Gelatin was dissolved in tins (Eastoe and Eastoe, 1952). Hydroxyproline (Hyp) distilled water to a concentration of 6.67% (wt/vol) at content was measured following the method of Edwards 55°C. The analysis was performed under a temperature and O’Brien (1980), using 7.14 as a factor for collagen sweep in the range of 45 to 5°C with a temperature content calculation. change rate of 0.5°C/min using a 2.5-cm parallel plate, Extraction Yield. The gelatin extraction yield was a gap of 0.5 mm, a strain of 5%, and a frequency of 10 calculated on a dry basis according to Du et al. (2013) s−1. During the whole test, an Anton Paar H-PTD200 as hood was used to prevent water evaporation from the 992 DU ET AL. sample. The elasticity (Gc), viscosity (Gcc), and phase volume at 0, 30, and 60 min was recorded. The FE and angle were measured and presented as a function of FS were calculated as temperature. The gelling and melting temperatures were obtained as the crossover point of the storage (G’) FE (%) = (VT/V0) × 100 and and the loss (G’’) moduli during the cooling and heating ramps, respectively. FS (%) = (Vt/V0) × 100, Bloom Strength. The bloom strength of the extracted gelatins was determined according to the British where VT = total volume at 0 min after homogeniza- Standard 757:1975 (BSI, 1975), with a slight modification; V0 = 10 mL (initial volume); and Vt = total vol- tion. Gelatin was dissolved in distilled water at 55°C ume at 30 or 60 min after homogenization. Downloaded from to a final concentration of 6.67% (wt/vol) and trans- Determination of Emulsifying Properties. Emulsi- ferred into a bloom jar (59-mm internal diameter and fying properties of the extracted gelatins were deter- 85-mm height, Schott-Duran, Elmsford, NY). Gelatin mined by measuring the emulsion particles size. Gela- solution was kept under 7°C in a cold room for 17 h and tin-oil emulsion was prepared according to the method the bloom strength was measured by a TA-XT2 Tex- of Pearce and Kinsella (1978) with some modifications. http://ps.oxfordjournals.org/ ture Analyzer (Stable Micro System Ltd., Surrey, UK) Different gelatin concentrations (0.5, 1, 2, and 3%; wt/ equipped with a 1.27-cm diameter flat-faced cylindrical vol) were first prepared with distilled water at 55°C. Teflon plunger and using a load cell of 5 kN. A constant Eight milliliters of each gelatin solution was then mixed speed of 0.5 mm/s was set during the measurement and with 2 mL of corn oil (Mazola, ACH Food Companies the maximum force (g) was recorded at a penetration Inc., Cordova, TN). After homogenization at 20,000 depth of 4 mm. rpm for 1 min, the emulsion was diluted 100-fold with Determination of Gel Microstructure. The micro- 0.1% (wt/vol) SDS solution and the particle size of the structure of gelatin gels was observed by cryo-scanning oil-gelatin emulsion system was analyzed using a laser at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 electron microscopy following the method of Du et al. light-scattering analyzer (Mastersizer 2000S, Malvern (2013). Gelatin gels were prepared at a concentration of Instruments, Malvern, UK). The volume mean diame- 6.67% (wt/vol) then cut into small cubes with a thick- ter (D [4, 3]) value, which stands for the mean droplet ness of 2 to 3 mm. The gel cubes were first frozen at size, was calculated as −207°C by Slush Nitrogen and then fractured, warmed −5 43 at −40°C, and set in a 10 Torr vacuum for 30 min D [43 , ]= ∑∑nd / nd , to sublime the water. Finally the fractured sample was ii ii sputtered with gold coating and visualized with a scanning electron microscope (JEOL JSM-6301, Tokyo, Ja- where ni = the number of particles with diameter di. pan) at an acceleration voltage of 5 kV. The emulsion activity was determined immediately after preparation, whereas the emulsion stability was Functional Properties of Extracted Gelatins estimated by the change of D [4, 3] value (ΔD [4, 3]) after storage for 1 wk at 4°C. Determination of Foaming Properties. Foaming properties, including foam expansion (FE) and foam Statistical Analysis stability (FS), were measured following the method of Shahidi et al. (1995) with some modifications. Gelatin The entire experiment was replicated 3 times and sample was dissolved in 10 mL of distilled water at each assay was replicated at least twice. Data was pre- 55°C to different final concentrations (0.5, 1, 2, and 3%; sented as mean ± SD. One- and two-way ANOVA were wt/vol). The gelatin solution was then transferred into used for mean comparison at the 0.05 significance level a 25-mL glass cylinder and homogenized at room tem- by performing a studentized range test (Tukey HSD perature by a Power Gen 1000 S1 homogenizer (Fisher Test). All statistical analyses were done using SPSS Scientific, Schwerte, Germany) at 13,000 rpm for 1 min. statistical software (version 20.0, SPSS Inc., Chicago, After whipping, the sample was retained for 1 h and the IL).

Table 1. Proximate analysis of mechanically separated turkey meat (MSTM) and collagen biomass1

Moisture Ash Protein Fat Hydroxyproline Collagen Material (%) (%) (%) (%) (dry-weight basis, %) (%) MSTM 59.29 ± 2.47b 3.69 ± 0.76a 16.08 ± 0.79a 21.38 ± 0.77a 1.54 ± 0.06b 11.00 ± 0.44b Collagen biomass 90.11 ± 0.89a 0.20 ± 0.02b 3.35 ± 0.01b 6.06 ± 0.22b 5.73 ± 0.53a 40.91 ± 3.79a P-value <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 a,bDifferent letters in the same column indicate significant (P < 0.05) differences between the means. 1Values given as mean ± SD. GELATIN FROM COLLAGEN BIOMASS OF TURKEY 993

Table 2. Proximate analysis and recovery yields of gelatins extracted from mechanically separated turkey meat (MSTM) and collagen biomass1

Moisture Ash Protein Fat Yield Gelatin (%) (%) (%) (%) (%) MSTMS12 11.21 ± 0.32a 0.07 ± 0.01b 85.22 ± 0.25b 1.84 ± 0.02a 4.13 ± 0.48b MSTMS23 10.14 ± 0.25b 0.07 ± 0.01b 88.45 ± 0.76a 1.52 ± 0.04b 2.23 ± 0.03b CB4 11.57 ± 0.35a 1.87 ± 0.03a 86.07 ± 0.13b 0.44 ± 0.08c 13.51 ± 1.51a P-value 0.003 <0.001 <0.001 <0.001 <0.001 a–cDifferent letters in the same column indicate significant (P < 0.05) differences between the means. 1Values given as mean ± SD.

2MSTMS1 = gelatin extracted in the first stage from MSTM. Downloaded from 3MSTMS2 = gelatin extracted in the second stage from MSTM. 4CB = gelatin extracted from collagen biomass.

RESULTS AND DISCUSSION et al., 2013). However, the application of severe temperatures and long processing times result in lowering http://ps.oxfordjournals.org/ Characterization of Raw Materials both the visual and the functional properties of gelatin (Schrieber and Gareis, 2007). As collagen in the The chemical compositions of MSTM and collagen newly developed process has been already solubilized biomass are shown in Table 1. Moisture and fat were and precipitated, it was only heated to 60°C for 6 h to the major components of the 2 raw materials, followed efficiently maximize its conversion into gelatin, without by protein and ash. The MSTM had significantly high- negatively affecting its properties.

er protein (16.08%), fat (21.38%), and ash (3.69%) Gelatin extraction yields (on a dry basis) from both at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 contents than collagen biomass (3.35, 6.06, and 0.20%, MSTM and collagen biomass were expressed as the respectively). However, significantly higher moisture weight of extracted gelatin (g) in respect to the total content (90.11%) was observed for the collagen biomass of collagen in the raw materials (g; Table 2). The mass compared with the MSTM (59.29%). The higher amount of gelatin extracted from the collagen biomass moisture in collagen biomass may have been due to the (13.51%) was significantly higher than that obtained greater water absorption during the pH-shifting pro- from MSTM (6.36%, in stages 1 and 2). This might cess. not only be due to the higher Hyp content in collagen On a dry weight basis, collagen biomass had sig- biomass (Table 1), but also due to the usage of pep- nificantly higher Hyp content (5.73%) compared with sin during the collagen solubilisation process. Pepsin MSTM (1.54%). The collagen content in the collagen is known to cleave crosslinks at the telopeptide region biomass (41%) was ~4 times higher than that from in collagen fibers, resulting in a higher extraction yield MSTM (11%), which could be due to the accumulation (Nalinanon et al., 2008). of collagen during the pH-shifting process. However, it is also important to mention that the extraction yield of the newly developed process would Characterization of Extracted Gelatins be 5.41 ± 0.60% if calculated based on the collagen content found in the initial MSTM. This yield was not Gelatin Extraction Yield. The temperature and the significantly different from the total extraction yield process duration play a major role in converting col- (MSTMS1 + MSTMS2) obtained through the tradi- lagen to gelatin. It was reported that increasing the tional process. Despite of the same yield obtained from temperature improves the extraction yield (Rafieian these 2 different methods, the newly developed method

Table 3. Color and bloom strength of gelatins extracted from mechanically separated turkey meat (MSTM) and collagen biomass1

Bloom Gelatin L* a* b* strength (g) MSTMS12 80.91 ± 1.87a −2.05 ± 0.08b 4.27 ± 0.45b 338.4 ± 11.4b MSTMS23 79.98 ± 0.80a −2.16 ± 0.06c 7.22 ± 0.24a 309.2 ± 8.2c CB4 36.42 ± 0.71b −1.79 ± 0.10a 0.96 ± 0.20c 353.2 ± 5.2a P-value <0.001 <0.001 <0.001 <0.001 a–cDifferent letters in the same column indicate significant (P < 0.05) differences between the means. 1Values given as mean ± SD. 2MSTMS1 = gelatin extracted in the first stage from MSTM. 3MSTMS2 = gelatin extracted in the second stage from MSTM. 4CB = gelatin extracted from collagen biomass. 994 DU ET AL. was more efficient than the conventional method because of the reduced processing time (67 and 28 h, respectively). Chemical Composition. The proximate analysis of the extracted gelatins is presented in Table 2. In the current study, all extracted gelatins were high in protein (85.22 –88.45%) and low in ash (0.07–1.87%) and fat (0.44–1.84%). The moisture content of all gelatins varied from 10.14 to 11.57%. This moisture range is considered to be mi- crobiologically safe, as higher water content may lead to Downloaded from microbial growth (Schrieber and Gareis, 2007). Accord- ing to GME (2000), gelatin should be fat-free and the recommended level of ash should be less than 2%. The higher (P < 0.05) ash content in CB (1.87%) compared with MSTMS1 and MSTMS2 may be due to the usage http://ps.oxfordjournals.org/ of NaOH during the precipitation of collagen, which resulted in salt formation. It seems that the newly developed process is more efficient than the traditional method in removing fat, as proven by the lower amount Figure 2. The SDS-PAGE patterns of gelatins extracted from me- of fat in CB (0.44%). With regard to the protein con- chanically separated turkey meat (MSTM) and collagen biomass. Lane tent, MSTMS2 contained significantly more protein 1 = gelatin extracted from collagen biomass (CB); lane 2 = gelatin

(88.45%) than MSTMS1 (85.22%) and CB (86.07%). at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 extracted in the first stage from MSTM (MSTMS1); lane 3 = gelatin extracted in the second stage from MSTM (MSTMS2); lane 4 = mo- The higher protein content of MSTMS2 compared with lecular weight markers. MSTMS1 was in agreement with the result reported by Rammaya et al. (2012), who observed an increase in protein content at higher extraction temperature for gelatin extracted from mechanically deboned chicken meat residue.

Table 4. Amino acid composition (residue/100 residue) of gelatins extracted from mechanically separated turkey meat (MSTM) and collagen biomass1

Amino acid MSTMS12 MSTMS23 CB4 P-value Asx5 4.3 ± 0.7 3.9 ± 0.4 4.7 ± 0.9 0.401 Glx6 8.0 ± 0.4 8.0 ± 0.5 7.4 ± 0.5 0.271 His 0.0 ± 0.0 0.0 ± 0.0 0.3 ± 0.5 0.422 Arg 5.6 ± 0.2 5.7 ± 0.3 5.5 ± 0.4 0.754 Thr 1.6 ± 0.2 1.7 ± 0.2 1.6 ± 0.4 0.961 Ala 12.7 ± 0.2 12.8 ± 0.7 12.1 ± 0.5 0.258 Ser 1.7 ± 0.6 1.8 ± 0.6 2.2 ± 0.7 0.645 Cys ND7 ND ND — Tyr 1.0 ± 0.6 0.4 ± 0.1 0.8 ± 0.2 0.212 Gly 31.8 ± 0.6 31.5 ± 0.5 31.5 ± 1.0 0.868 Pro 12.6 ± 0.2 12.3 ± 0.4 12.5 ± 0.9 0.802 Hydroxyproline (Hyp) 9.5 ± 0.4 9.8 ± 0.7 9.6 ± 1.4 0.893 Val 2.0 ± 0.2 2.5 ± 0.6 2.3 ± 0.3 0.270 Met 0.8 ± 0.1 0.8 ± 0.1 0.7 ± 0.1 0.171 Lys 2.7 ± 0.0 2.9 ± 0.1 3.0 ± 0.2 0.072 Ile 1.3 ± 0.1 1.4 ± 0.1 1.3 ± 0.1 0.422 Leu 2.8 ± 0.1 3.0 ± 0.1 2.8 ± 0.2 0.190 Phe 1.6 ± 0.1 1.6 ± 0.1 1.5 ± 0.1 0.317 Trp ND ND ND — Total 100 100 100 — Imino acids (Pro + Hyp) 22.1 ± 0.3 22.1 ± 0.6 22.1 ± 0.8 1.0 Hydrophobic amino acid 21.1 ± 0.2b 22.1 ± 0.4a 20.7 ± 0.3b 0.003 a,bDifferent letters in the same row indicate significant (P < 0.05) differences between the means. 1Values given as mean ± SD. 2MSTMS1 = gelatin extracted in the first stage from MSTM. 3MSTMS2 = gelatin extracted in the second stage from MSTM. 4CB = gelatin extracted from collagen biomass. 5Asx = the sum of aspartic acid and asparagine. 6Glx = the sum of glutamic acid and glutamine. 7ND = not detected. GELATIN FROM COLLAGEN BIOMASS OF TURKEY 995 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 3. Evolution of viscoelastic properties of mechanically separated turkey meat (MSTM) and collagen biomass (CB) gelatins upon cooling (45–5°C) and heating (5–45°C). MSTMS1 = gelatin extracted in the first stage from MSTM; MSTMS2 = gelatin extracted in the second stage from MSTM. (A) Evolution of elastic properties (Gc) of gelatins upon cooling (45–5°C). (B) Evolution of elastic properties (Gc) of gelatins upon heating (5–45°C). (C) Evolution of viscous properties (Gcc) of gelatins upon cooling (45–5°C). (D) Evolution of viscous properties (Gcc) of gelatins upon heating (5–45°C). (E) Change of phase angle (°) of gelatins upon cooling (45–5°C). (F) Change of phase angle (°) of gelatins upon heating (5–45°C).

Color of Gelatin. The colors of the extracted gelatin MSTMS2. This result was in agreement with that re- gels were evaluated by the L*, a*, and b* parameters ported by Nagarajan et al. (2012), who observed that (Table 3). The CB showed lower (P < 0.05) L* and b* the increase in color parameters is correlated with the but higher a* compared with MSTMS1 and MSTMS2. increase in temperature, as higher temperatures would The low L* value (36.42) and the high a* value (−1.79) raise the number of free amino groups capable of induc- for CB, which reflected a darker gelatin color, could be ing the nonenzymatic browning reaction. due to the result of a Maillard reaction between the Protein Patterns of Gelatins. The molecular weight amino acids in gelatin and residual sugars. In fact, in distribution of gelatins produced from the 2 processes the newly developed method, high pH and tempera- was estimated by SDS-PAGE and is shown in Figure 2. ture were used to precipitate collagen and convert it The CB comprised 3 major bands: 2 α-chains (α1 upper to gelatin, respectively. These conditions (elevated pH and α2 lower) and one β-chain (Figure 2, Lane 1). The values and temperature) are known to accelerate the α and β chains were also observed in MSTMS1, along Maillard reaction (Rich and Foegeding, 2000), resulting with some low-molecular weight bands (Figure 2, Lane in a lower L* value. 2). However, MSTMS2 had different protein patterns Gelatin gels, obtained from MSTM, had similar than both MSTMS1 and CB. All 3 bands were totally lightness values (80.91 and 79.98 for MSTMS1 and hydrolyzed into lower molecular weight components MSTMS2, respectively). However, significantly lower (Figure 2, Lane 3), probably due to the effect of higher a* (−2.16) and higher b* (7.22) were observed with temperature and the long extraction time. 996 DU ET AL. amino acid content. All gelatins, regardless of the extraction process, had very low amounts of His and Tyr, whereas Trp and Cys were not detected. The absence of Trp and Cys may suggest that all extracted gelatins were from type I collagen. The imino acids (Hyp and Pro) are believed to play a major role in stabilization of the collagen triple helix structure (Fernandez-Diaz et al., 2001). In the current study, no significant differences among the imino acid contents were observed for all the extracted gelatins. However, MSTMS2 had significantly higher amounts of Downloaded from hydrophobic amino acids (Ala, Val, Met, Ile, Leu, Phe) compared with both MSTMS1 and CB. According to Khiari et al. (2013), differences among the hydrophobic amino acid contents may affect the functional properties of gelatins. http://ps.oxfordjournals.org/

Rheological and Textural Properties of Extracted Gelatins Dynamic Viscoelastic Behavior of Gelatins. The dynamic viscoelastic behavior of gelatins was investi-

gated under a temperature sweep test in the range of at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 5 to 45°C (Figure 3). During the cooling step, both the elastic (Gc) and the viscous (Gcc) moduli of all gelatins (Figure 3A and C) increased rapidly from 25 to 5°C. The sharp decrease of the phase angle at 25°C (Figure 3E) indicated the transition of gelatins from solution to gel. During the heating step, the elastic (Gc) and viscous (Gcc) moduli gradually decreased from around 8°C and a marked decrease of both elasticity and viscosity was observed between 15 and 33°C, which represented the gelatin-melting zone. The gelling temperatures of these gelatins varied between 27.0 to 27.9°C, whereas the melting temperatures ranged from 31.4 to 34.5°C. The CB showed better viscoelastic properties in terms of higher gelling and melting temperatures, as well as elastic and viscous moduli, compared with gela- Figure 4. Scanning electron microscopy (×500) of gelatin gels from tin extracted from MSTM. This might be associated mechanically separated turkey meat (MSTM) and collagen biomass. with the molecular weight distribution of gelatins. Gel- (A) Gelatin extracted in the first stage from MSTM; (B) gelatin extracted in the second stage from MSTM; (C) gelatin extracted from atin obtained from the new process had a large amount collagen biomass. of higher molecular weight chains (α and β chains), whereas those obtained from the traditional process were more prone to thermal degradation. The presence The presence of high-molecular weight protein bands of high-molecular weight components in CB might have in CB is an indication that the newly developed pro- resulted in partial renaturation of the collagen triple cess resulted in less degradation of collagen molecules helices and therefore enhanced its viscoelastic proper- compared with the traditional method. Although CB ties (Gómez-Guillen et al., 2002). and MSTMS2 were extracted at the same temperature Bloom Strength of Gelatins. Bloom strength is the (i.e., 60°C), the greater protein degradation observed in most important quality associated with the price and MSTMS2 may be associated with the longer treatment industrial application of gelatins. Significant differences time rather than the temperature. were observed among the gel strengths of the differ- Amino Acid Composition of Gelatins. The amino ent extracted gelatins (Table 3). The CB exhibited the acid composition of gelatins, presented as number of highest (P < 0.05) bloom strength (353.2 g) followed amino acid residues per 100 residues, is shown in Table by MSTMS1 (338.4 g) and MSTMS2 (309.2 g). The 4. Glycine, Pro, Ala, and Hyp were the most abundant molecular weight distribution and the amino acid com- amino acids in all gelatins. The 3 amino acids, Gly, Pro, position of gelatins are the main factors affecting the and Hyp, accounted for more than 53% of the total bloom strength. High-molecular weight gelatins can ef- GELATIN FROM COLLAGEN BIOMASS OF TURKEY 997

Table 5. Foaming properties of gelatins extracted from mechanically separated turkey meat (MSTM) and collagen biomass1

FS3 (%)

Item FE2 (%) 30 min 60 min Gelatin MSTMS14 126.2 ± 12.4b 122.4 ± 12.3b 116.6 ± 11.0c MSTMS25 134.2 ± 9.8a 130.5 ± 10.8a 129.2 ± 10.3a CB6 126.2 ± 5.2b 122.0 ± 5.8b 118.0 ± 7.3b Concentration (%) 0.5 116.6 ± 5.2d 112.4 ± 4.7d 108.1 ± 6.3d

1 125.2 ± 4.4c 121.2 ± 3.6c 118.3 ± 5.0c Downloaded from 2 134.6 ± 5.2b 130.3 ± 7.7b 127.3 ± 9.0b 3 139.2 ± 5.6a 136.0 ± 5.4a 131.4 ± 5.0a Interaction (gelatin × concentration) MSTMS1 0.5 109.8 ± 0.3h 106.3 ± 1.2g 100.8 ± 1.4g 1 120.3 ± 0.6g 119.3 ± 0.6d 115.1 ± 0.4e 2 134.5 ± 0.9c 124.7 ± 0.6c 120.5 ± 0.9d 3 140.1 ± 0.1b 139.2 ± 1.4a 130.0 ± 0.0b http://ps.oxfordjournals.org/ MSTMS2 0.5 120.8 ± 1.4g 115.8 ± 1.4ef 115.1 ± 0.2e 1 130.5 ± 0.5de 125.8 ± 1.4c 125.0 ± 0.0c 2 140.5 ± 0.9b 140.5 ± 0.9a 139.2 ± 1.4a 3 145.1 ± 0.4a 140.0 ± 0.0a 137.7 ± 0.3a CB 0.5 119.0 ± 1.0g 115.0 ± 0.5f 108.5 ± 1.3f 1 124.7 ± 0.6f 118.5 ± 0.9de 114.8 ± 0.3e 2 128.7 ± 1.2e 125.7 ± 0.6c 122.3 ± 0.6d 3 132.3 ± 0.6cd 128.8 ± 0.3b 126.5 ± 0.5c P-value Gelatin <0.001 <0.001 <0.001 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Concentration <0.001 <0.001 <0.001 Gelatin × concentration <0.001 <0.001 <0.001 a–hDifferent letters in the same column indicate significant (P < 0.05) differences between the means. 1Values given as mean ± SD. 2FE = foaming expansion. 3FS = foaming stability. 4MSTMS1 = gelatin extracted in the first stage from MSTM. 5MSTMS2 = gelatin extracted in the second stage from MSTM. 6CB = gelatin extracted from collagen biomass. fectively form strong gels compared with low-molecular pores (Figure 4A and B, respectively). However, gela- weight ones (Nagarajan et al., 2012). As all gelatins tin extracted from collagen biomass had highly porous had similar imino acid contents (Table 4), the differenc- microstructure (Figure 4C). Compared with MSTMS1 es among the gel strengths may be due to the molecular and MSTMS2, CB gels contained greater number of weight distribution (Figure 2). As discussed previously, small-size pores, which is an indication of a rigid gel CB contained large amounts of high-molecular weight (Wangtueai and Noomhorm, 2009). Both the bloom components compared with MSTMS1 and MSTMS2, strength analysis and the scanning electron microscopy which might have imparted greater bloom strength. were in accordance with this finding. Conversely, the partially thermally degraded MSTMS2 had significant amounts of low-molecular weight frac- Functional Properties of Gelatin tions compared with MSTMS1 and, as a consequence, exhibited lower gel strength. Foaming Properties of Gelatin. The FE and FS of The bloom strength of all gelatin gels, regardless of extracted gelatin are shown in Table 5. The interac- the extraction process, varied from 309.2 to 353.2 g, tive effect of gelatin groups and different levels of gela- which corresponded to high-bloom gelatin range and tin concentrations on the foaming properties of gela- was comparable to the strength of porcine and bovine tins was significant. It was observed that, regardless of gelatins (Schrieber and Gareis, 2007). the extraction method, both FE and FS significantly Gel Microstructure. The internal microstructure of increased with the increment of gelatin concentration the cryo-fractured gelatin gels was determined using a (from 0.5 to 3%). Conversely, the concentration effect scanning electron microscope. The gelatin gel micro- was less pronounced on the gelatins obtained from CB. graphs, obtained at 500× magnifications, are shown In general, higher protein concentration leads to higher in Figure 4. Gelatins extracted from MSTM through amounts of proteins migrating to the surface of air bub- the conventional process (MSTMS1 and MSTMS2) had bles. As a consequence, a rapid foam formation with a similar internal microscopic structures characterized by denser stable film at the air-water interface is observed well-organized protein networks with evenly distributed (Zayas, 1997). 998 DU ET AL.

Table 6. Emulsifying properties of gelatins extracted from mechanically separated turkey meat (MSTM) and collagen biomass1

Item D [4, 3]2 (μm) ΔD [4, 3]3 (μm) Gelatin MSTMS14 8.54 ± 0.39b 2.95 ± 3.32a MSTMS25 7.39 ± 0.50c 2.48 ± 2.36b CB6 9.64 ± 0.57a 2.14 ± 1.88b Concentration (%) 0.5 7.88 ± 0.94d 6.39 ± 1.69a 1 8.37 ± 0.97c 2.40 ± 0.46b 2 8.72 ± 1.01b 1.00 ± 0.36c

3 9.11 ± 1.01a 0.30 ± 0.18d Downloaded from Interaction (gelatin × concentration) MSTMS1 0.5 8.05 ± 0.01 8.27 ± 0.27a 1 8.39 ± 0.12 2.49 ± 0.26de 2 8.71 ± 0.21 0.76 ± 0.07fgh 3 9.00 ± 0.19 0.28 ± 0.12gh MSTMS2 0.5 6.73 ± 0.17 6.19 ± 1.25b 1 7.25 ± 0.22 1.86 ± 0.18def http://ps.oxfordjournals.org/ 2 7.57 ± 0.02 1.44 ± 0.28efg 3 8.02 ± 0.03 0.45 ± 0.24gh CB 0.5 8.87 ± 0.15 4.72 ± 0.41c 1 9.46 ± 0.13 2.85 ± 0.09d 2 9.89 ± 0.02 0.81 ± 0.14fgh 3 10.32 ± 0.03 0.17 ± 0.00h P-value Gelatin <0.001 <0.001 Concentration <0.001 <0.001 Gelatin × concentration 0.102 <0.001 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 a–hDifferent letters in the same column within the same effect indicate significant (P < 0.05) differences between the means. 1Values given as mean ± SD. 2D [4, 3] = mean droplet size of emulsion immediately after homogenization. 3ΔD [4, 3] = the change of mean droplet size of emulsion after 1-wk storage. 4MSTMS1 = gelatin extracted in the first stage from MSTM. 5MSTMS2 = gelatin extracted in the second stage from MSTM. 6CB = gelatin extracted from collagen biomass.

At the same gelatin concentration, MSTMS2 showed activity of gelatins was observed. With respect to the greater (P < 0.05) FE compared with MSTMS1 and concentration effect, a significant improvement in the CB. This may due to the presence of low-molecular emulsion activity, as indicated by the reduction of D [4, weight chains (Figure 2, Lane 3) and larger numbers of 3] value, was noticed with the decrease in the gelatin hydrophobic groups (Table 4), which might have result- concentration (from 3 to 0.5%). This might be due to ed in an efficient protein adsorption at the air-water the fact that the diffusion and adsorption of protein interface (Khiari et al., 2013). The FS can be influenced at the oil-water interface is favorable at lower protein by various factors, such as the surface tension, viscos- concentrations (Cheftel et al., 1985). ity, and electrical repulsion (Ahmad and Benjakul, Regarding the gelatin group’s effect, the average 2011). The MSTMS2 exhibited the highest (P < 0.05) droplet size of emulsions prepared with MSTMS1 and FS (especially after 60 min) at all gelatin concentra- MSTMS2 were smaller (P < 0.05) than that prepared tions, which was probably due to its higher (P < 0.05) with CB. This was probably due to differences in mo- amount of hydrophobic amino acid residues (Ala, Val, lecular weight distributions (Figure 2) and hydrophobic Ile, Leu, Met, Phe) compared with the other 2 gelatins amino acid contents (Table 4). The smaller size pep- (Table 4). The positive effect of proteins hydrophobic- tide chains present in gelatins extracted from MSTMS2 ity on the foam properties has been previously demon- may have enhanced the emulsion capacity by facilitat- strated by Townsend and Nakai (1983). ing the diffusion at the surface of oil droplets (Surh et Emulsifying Properties of Gelatin. The emulsion al., 2006). In addition, the significantly larger amounts activity and stability of gelatin samples from different of hydrophobic amino acid residues in MSTMS2 may processes were evaluated by the mean droplet size (D have improved the emulsion activity by lowering the [4, 3]) of emulsions (Table 6). A smaller emulsion drop- interfacial tension through binding to the oil fraction at let size and smaller change in the droplet size (ΔD [4, the oil-water interface (Kato and Nakai, 1980) 3]) during storage indicate better emulsion activity and Emulsion Stability. The stability of emulsions was stability, respectively. evaluated by the ΔD [4, 3] after 1 wk of storage (Table Emulsion Activity. No significant interaction of gela- 6). The interaction effect of gelatin groups and concen- tin groups and different concentrations on the emulsion trations on the emulsion stability of gelatins was signifi- GELATIN FROM COLLAGEN BIOMASS OF TURKEY 999 cant. Contrary to emulsion activity, increasing gelatin Cheftel, J. C., J. L. Cuq, and D. Lorient. 1985. Amino acids, pep- concentration resulted in better emulsion stability for tides, and proteins. Pages 245–369 in Food Chemistry. 2nd ed. O. R. Fennema, ed. Marcel Dekker Inc., New York, NY. these 3 gelatin groups. Higher protein concentrations Du, L., Z. Khiari, Z. Pietrasik, and M. Betti. 2013. Physico-chemical lead to higher viscosities of dispersions and subsequent- and functional properties of gelatins extracted from turkey and ly reduced the coalescence rate and improve the emul- chicken heads. Poult. Sci. 92:2463–2474. sion stability (Sajjadi, 2007). Eastoe, J. E., and B. Eastoe. 1952. A method for the determination of total nitrogen in proteins. Pages 1–17 in The British Gelatin Gelatin groups strongly affected (P < 0.05) the emul- and Glue Research Association Research Report, Series B5. Na- sion stability. No significant differences were observed ture Publishing Group, London, UK. among the emulsion stabilities of all gelatins at pro- Edwards, C. A., and W. D. J. R. O’Brien. 1980. Modified assay for tein concentrations of 1, 2, or 3% (wt/vol). However, determination of hydroxyproline in a tissue hydrolyzate. Clin.

Chim. Acta 104:161–167. Downloaded from at 0.5% gelatin concentration, CB showed the smallest FAO. 2012. World Food and Agriculture. Statistical Yearbook. Food ΔD [4, 3] value (4.72) compared with MSTMS1 and and Agriculture Organization of the United Nations, Rome, Italy. MSTMS2 (8.27 and 6.19, respectively), indicating bet- FAOSTAT. 2013. Food and Agriculture Organization of the United Nations, Rome, Italy. Accessed Mar. 12, 2013. http://faostat.fao. ter emulsion stability. org/site/569/DesktopDefault.aspx?PageID=569#. Fernandez-Diaz, M. D., P. Montero, and M. C. Gomez-Guillen. 2001. Gel properties of collagens from skins of cod (Gadus morhua) http://ps.oxfordjournals.org/ Conclusions and hake (Merluccius merluccius) and their modification by the coenhancers magnesium sulphate, glycerol and transglutaminase. Gelatins were extracted from MSTM following 2 dif- Food Chem. 74:161–167. ferent methods. One was a traditional 2-stage batch Fonkwe, L. G., and R. K. Singh. 1997. Production and characterization of gelatinous protein extracts from turkey deboned residue. thermal extraction and the other was a new process Process Biochem. 32:309–318. based on recovering gelatin from a collagen biomass Froning, G. W., and S. R. McKee. 2001. Mechanical separation of obtained through the pH-shifting technology. Both pro- poultry meat and its use in products. Pages 243–244 in Poultry cesses significantly affected the properties and quality Meat Processing. A. R. Sams, ed. CRC Press LLC, Boca Raton, at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 FL. of the final products. Gelatin prepared by the newly GME. 2000. Standard Methods for the Testing of Edible Gelatin. developed process had better rheological and textural Version 1. Gelatine Manufacturers of Europe, Brussels, Belgium. characteristics but darker color compared with those Gómez-Guillén, M. C., M. Ihl, V. Bifani, A. Silva, and P. Montero. 2007. Edible films made from tuna-fish gelatin with antioxidant produced through the traditional thermal way. Results extracts of two different murta ecotypes leaves (Ugni molinae from the current study showed that high-quality gela- Turcz). Food Hydrocoll. 21:1133–1143. tin can be extracted from MSTM using an alternative Gómez-Guillen, M. C., J. Turnay, M. D. Fernndez-Daz, N. Ulmo, process to the traditional thermal extraction method. M. A. Lizarbe, and P. Montero. 2002. Structural and physical properties of gelatin extracted from different marine species: A Despite the darker color of gelatin, the newly devel- comparative study. Food Hydrocoll. 16:25–34. oped process seems to be industrially more advanta- Hrynets, Y., D. A. Omana, Y. Xu, and M. Betti. 2010. Effect of geous compared with the traditional process in terms acid- and alkaline-aided extractions on functional and rheologi- of recovery of valuable meat proteins (protein isolate cal properties of proteins recovered from mechanically separated turkey meat (MSTM). J. Food Sci. 75:E477–486. and collagen biomass), as well as time saving and in the Hrynets, Y., D. A. Omana, Y. Xu, and M. Betti. 2011. Compara- reduction of water and chemical reagent consumption. tive study on the effect of acid- and alkaline-aided extractions on Mechanically Separated Turkey Meat (MSTM): Chemical and functional properties of recovered proteins. Process Biochem. 46:335–343. ACKNOWLEDGMENTS Karim, A. A., and R. Bhat. 2009. Fish gelatin: Properties, challenges, and prospects as an alternative to mammalian gelatins. The authors thank the Alberta Livestock and Meat Food Hydrocoll. 23:563–576. Agency (Edmonton, AB, Canada) for providing fund- Kato, A., and S. Nakai. 1980. Hydrophobicity determined by a fluo- ing. The authors also extend appreciation to Takuo Na- rescent probe method and its correlation with surface properties of proteins. Biochim. Biophys. Acta Protein Struct. 624:13–20. kano (University of Alberta, Edmonton, AB, Canada) Khiari, Z., D. Rico, A. B. Martin-Diana, and C. Barry-Ryan. 2013. for his assistance in laboratory session. Comparison between gelatines extracted from mackerel and blue whiting bones after different pre-treatments. Food Chem. 139:347–354. Liang, Y., and H. O. Hultin. 2003. Functional protein isolates from REFERENCES mechanically deboned turkey by alkaline solubilization with isoelectric precipitation. J. Muscle Foods 14:195–205. Ahmad, M., and S. Benjakul. 2011. Characteristics of gelatin from Montero, P., and M. C. Gómez-Guillén. 2000. Extracting conditions the skin of unicorn leatherjacket (Aluterus monoceros) as influ- for megrim (Lepidorhombus boscii) skin collagen affect functional enced by acid pretreatment and extraction time. Food Hydro- properties of the resulting gelatine. J. Food Sci. 65:434–438. coll. 25:381–388. Nagarajan, M., S. Benjakul, T. Prodpra, P. Songtipy, and H. AOAC. 2000. Official Methods of Analysis of AOAC International. Kishimura. 2012. Characteristics and functional properties of W. Horowitz, ed. Association of Official Analytical Chemists In- gelatin from splendid squid (Loligo formosana) skin as affected ternational, Gaithersburg, MD. by extraction temperatures. Food Hydrocoll. 29:389–397. Arnesen, J. A., and A. Gildberg. 2007. Extraction and characteriza- Nakano, T., Z. Pietrasik, L. Ozimek, and M. Betti. 2012. Extraction, tion of gelatine from Atlantic salmon (Salmo salar) skin. Biore- isolation and analysis of chondroitin sulfate from broiler chicken sour. Technol. 98:53–57. biomass. Process Biochem. 47:1909–1918. BSI. 1975. Methods for Sampling and Testing Gelatin (Physical and Nalinanon, S., S. Benjakul, W. Visessanguan, and H. Kishimura. Chemical Methods). British Standards Institution, London, UK. 2008. Improvement of gelatin extraction from bigeye snapper skin 1000 DU ET AL.

using pepsin-aided process in combination with protease inhibi- Schrieber, R., and H. Gareis. 2007. Gelatine Handbook: Theory and tor. Food Hydrocoll. 22:615–622. Industrial Practice. Wiley-VCH GmbH & Co., Weinhem, Ger- Pearce, K. N., and J. E. Kinsella. 1978. Emulsifying properties of many. proteins: Evaluation of a turbidimetric technique. J. Agric. Food Shahidi, F., H. Xiao-Qing, and J. Synowiecki. 1995. Production and Chem. 26:716–723. characteristics of protein hydrolysates from capelin (Mallotus vil- Rafieian, F., J. Keramat, and M. Kadivar. 2013. Optimization of gel- losus). Food Chem. 53:285–293. atin extraction from chicken deboner residue using RSM method. Surh, J., E. A. Decker, and D. J. McClements. 2006. Properties and J. Food Sci. Technol. 50:374–380. stability of oil-in water emulsions stabilized by fish gelatin. Food Rammaya, K., V. Q. Ying, and A. S. Babji. 2012. Physicochemical Hydrocoll. 20:596–606. analysis of gelatin extracted from mechanically deboned chick- Townsend, A. A., and S. Nakai. 1983. Relationships between hydro- en meat (MDCM) residue. Int. J. Food Nutr. Public Health phobicity and foaming characteristics of food proteins. J. Food 5:147–168. Sci. 48:588–594.

Rich, L. M., and E. A. Foegeding. 2000. Effects of sugars on whey Wangtueai, S., and A. Noomhorm. 2009. Processing optimization Downloaded from protein isolate gelation. J. Agric. Food Chem. 48:5046–5052. and characterization of gelatin from lizardfish (Saurida spp.) Sajjadi, S. 2007. Formation of fine emulsions by emulsification at scales. LWT Food Sci. Technol. 42:825–834. high viscosity or low interfacial tension; A comparative study. Zayas, J. F. 1997. Solubility of proteins. Pages 6–22 in Functionality Colloids Surf. A Physicochem. Eng. Asp. 299:73–78. of Proteins in Food. Springer-Verlag, Berlin, Germany. Sarbon, N. M., F. Badii, and N. K. Howell. 2013. Preparation and characterisation of chicken skin gelatin as an alternative to mammalian gelatin. Food Hydrocoll. 30:143–151. http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Sequential separation of lysozyme, ovomucin, ovotransferrin, and ovalbumin from egg white

E. D. N. S. Abeyrathne,*† H. Y. Lee ,* and D. U. Ahn *‡§1

* WCU Biomodulation Major, Department of Agricultural Biotechnology, College of Agriculture and Life Science, Seoul National University, Seoul 151-742, Korea; † Department of Animal Science, Uva Wellasa University, Badulla, Sri Lanka 90000; ‡Department of Animal Science, Iowa State University, Ames 50010; Downloaded from and § Department of Animal Science and Technology, Sunchon National University, Sunchon 540-742, South Korea

ABSTRACT Ovalbumin, ovotransferrin, ovomucin, and used for ovalbumin separation and the precipitant was lysozyme are a few of the egg white proteins that can dissolved in water, and reprecipitated using 2.0% am- be used as functional components. The objective of this monium sulfate (wt/vol) and 1.5% citric acid (wt/vol) http://ps.oxfordjournals.org/ study was to develop a simple, sequential separation combination. The precipitant was used as ovotransfer- method for multiple proteins from egg white. Separated rin fraction, and the supernatant (S2) was pooled with proteins are targeted for human use, and thus any toxic the first supernatant (S1), desalted using ultrafiltra- compounds were excluded. The methods for individual tion, and then heat-treated to remove impurities. The components and the sequential separation were prac- yield of ovomucin and ovalbumen was >98% and that ticed in laboratory scale first, and then tested for scale- of ovotransferrin and lysozyme was >82% for both

up. Lysozyme was separated first using FPC3500 cat- laboratory and scale-up preparations. The SDS-PAGE at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 ion exchange resin and then ovomucin using isoelectric and western blotting of the separated proteins, except precipitation. Ovalbumin and ovotransferrin were sepa- for ovomucin, showed >90% purity. The ELISA results rated from the lysozyme- and ovomucin-free egg white indicated that the activities of separated ovalbumin, by precipitating ovotransferrin first using 5.0% (wt/ ovotransferrin, and lysozyme were >96%. The proto- vol) (NH4)2SO4 and 2.5% (wt/vol) citric acid combi- col separated 4 major proteins in sequence, and the

nation. After centrifugation, the supernatant (S1) was method was simple and easily scaled up. Key words: egg white , sequential separation, protein , yield , purity 2014 Poultry Science 93 :1001–1009 http://dx.doi.org/ 10.3382/ps.2013-03403

INTRODUCTION in immunological and nutritional studies (Datta et al., 2009). Albumin was first separated using saturated am- Ovalbumin (54%), ovotransferrin (12%), ovomucin monium sulfate and acetic acid (Hopkins, 1900; Chick (3.5%), and lysozyme (3.4%) are among the major and Martin, 1913), but no clear record about the purity proteins in egg white (Stadelman and Cotterill, 2001). and yield is available. Recently, ovalbumin was separat- These proteins are known to have unique functions and ed using 2-stage polyethersulfone flat disk membranes can be used in food processing and as pharmaceuticals (Datta et al., 2009), electrophoretic method (Desert et or antimicrobial agents after separation. Ovalbumin is al., 2001), foam fractionation (Ward et al., 2007), or the major egg white protein with a molecular weight liquid chromatographic method (Awade and Efstahiou, of 45 kDa and is one of the first proteins isolated from 1999). However, all these methods are difficult to scale egg white. It is known as a phosphoglycoprotein com- up for industrial applications due to their complicated posed of 385 amino acids (Huopalahti et al., 2007). procedures, material costs, or sample handling capac- Half of the amino acids in ovalbumin are hydrophobic ity. and one-third is charged (Nisbet et al., 1981). Oval- Ovotransferrin is another major protein found in egg bumin is widely used as a standard protein in protein white. It is a monomeric glycoprotein with 686 amino assays (Huntington and Stein, 2001) and is important acids and 76 kDa molecular weight with 15 disulfide bonds (Williams, 1968). Ovotransferrin is known to bind and transport irons in animal body. Ovotransfer- © 2014 Poultry Science Association Inc. rin is present in apo- (iron free) and holo- (iron bound) Received June 11, 2013. Accepted December 20, 2013. forms, and the chemical and physical properties of 1 Corresponding author: [email protected] these 2 forms differ significantly (Wu and Acer-Lopez,

1001 1002 ABEYRATHNE ET AL. 2012). The holo-form of ovotransferrin is more resistant is composed of 2 subunits called α1 and α2 (Hiidenhovi to chemical and physical conditions such as pH and et al., 1999) and can be solubilized using dissociating heat than the apo-form (Ko and Ahn, 2008). Ovotrans- agents such as urea, β-mercaptoethanol, SDS, guani- ferrin is well known to have a strong iron binding capa- dine hydrochloride, and dithiothreitol (Hiidenhovi et bility, and thus, can be used as an antimicrobial agent, al., 1999; Huopalahti et al., 2007). Ovomucin was sepa- antioxidant, or an iron-supplementing agent (Fraenkel- rated mainly by isoelectric precipitation. MacDonnell Conrat and Feeney, 1950; Valenti et al., 1982; Ibrahim et al. (1951) separated ovomucin by bringing down the et al., 2000; Ko et al., 2009; Wu and Acer-Lopez, 2012). pH of egg white to 4.5, the pI of ovomucin and centrifu- Over the past several years, different techniques have gation. Others used various approaches, which include been developed to separate ovotransferrin from chick- isoelectric precipitation (Donovan et al. 1970; Rabouille en egg white, but most of the methods developed are et al., 1990; Dubiard et al., 2005; Omana et al., 2010b), Downloaded from laboratory scale. Anion exchange chromatography was dual-column gel filtration (Hiidenhovi et al., 1999), gel developed by Omana et al. (2010a) and Wu and Acer- electrophoresis (Desert et al., 2001), β-mercaptoethanol Lopez (2012) with 80% yield. It was also separated us- (Hiidenhovi et al., 2002), calcium chloride (Omana and ing CM-Toyopearl 650M cation exchange (Tankrathok Wu, 2009b), and NaCl + pH adjustment (Wang and et al., 2009) and SDS-PAGE and immunoelectropho- Wu, 2012). Among the methods, however, isoelectric http://ps.oxfordjournals.org/ resis (Desert et al., 2001). However, the methods have precipitation was the best way of separating ovomucin. difficulties in handling and the resins are not appro- All the methods discussed above were separating sin- priate for large-scale separation of ovotransferrin. Ovo- gle protein from egg white and rest of the proteins were transferrin was also separated using ammonium sulfate either denatured during the separation processes or dis- precipitation (Fraenkel-Conrat and Feeney, 1950), but carded. Separation of more than one protein has been the amount of ammonium sulfate used was too high done by a few research groups, but none of them were

for a large-scale method. Recently, ovotransferrin was in large scale. Tankrathok et al. (2009) separated oval- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 produced in large scale by ethanol precipitation (Ko bumin, lysozyme, ovotransferrin, and ovomucoid using and Ahn, 2008). However, using ethanol has some limi- Q-Sepharose fast-flow anion exchange chromatography tations because it denatures the rest of the proteins, and then with CM-Toyopearl 650M cation exchange which makes it difficult to use the remaining proteins. chromatography, but the yields were 54, 55, and 21%, Another important protein found in egg white is ly- respectively. Also, lysozyme, ovotransferrin, and oval- sozyme. The molecular weight of lysozyme is 14.4 kDa bumin were separated in sequence by Vachier et al. and is consisted of a single polypeptide chain with 129 (1995), but the yield of lysozyme was as low as 60%. amino acids (Radziejewska et al., 2008). It is a strongly Both of these methods have low yields and cannot be basic protein with isoelectric point (pI) of 10.7 and scaled up for commercial production. The objective of has 4 disulfide bridges leading to high thermal stabil- this study was to develop a simple, economical, sequen- ity (Huopalahti et al., 2007). Lysozyme was first sepa- tial, and scalable method to separate lysozyme, ovomu- rated using high levels of ammonium sulfate, but the cin, ovotransferrin, and ovalbumin from egg white. characteristics of lysozyme were changed by pH and salts (Alderton et al., 1945). The most frequent sepa- MATERIALS AND METHODS ration technique used for lysozyme is based on cation exchange chromatography mainly due to its high pI Materials value. Carboxymethyl cellulose (CMC) was used in Chicken eggs (large size, less than 3 d old) were pur- the past to trap lysozyme (Strang, 1984). However, due chased from a local market and used within a day. For to its fine granule sizes, handling is difficult and flow small scale 280 g and for large scale 3,625 g of egg white rate is very slow for column chromatography. Magnetic was used for the separation of the proteins. Amber- cation exchange isolation (Safarik et al., 2007), use of lite FPC 3500 (styrene-divinylbenzene, total exchange reductants such as β-mercaptoethanol (Chang et al., capacity >2.6 mEq/g, H+-form), ammonium sulfate, 2000), and ultrafiltration (Wan et al., 2006) were some and citric acid were purchased from Fisher Scientific other methods used to separate the protein. Howev- (Waltham, MA). Standard ovalbumin, lysozyme, and er, all these methods have problems in scaling up and ovotransferrin were purchased from Sigma-Aldrich β-mercaptoethanol cannot be used in the food industry. (St. Louis, MO), and rabbit polyclonal to lysozyme, This is one of the first egg white proteins that was iso- lysozyme marker HRP, anti-ovotransferrin, and anti- lated and used in industry. It has the ability to control ovalbumin antibodies were purchased from Abcam a limited spectrum of bacteria and fungi that cause (Cambridge, MA). Rest of the standards and rabbit spoilage of food (Durance, 1994). Other than that, it is anti-mouse IgG (H+L) conjugated AP12 antibody from widely used in kimchi pickles, sushi, Chinese noodles, Bio-Rad (Richmond, CA). and cheese and wine production (Mine et al., 2004). Ovomucin is a glycoprotein consisting of 2 subunits (α- and β-subunits) with a molecular weight of 5.5- 8.3 Separation of Lysozyme × 103 kDa (Omana et al., 2010b), and is responsible for Separation of lysozyme was done with cation ex- the gel-like properties of thick egg white. α-Ovomucin change chromatography. Amberlite FPC 3500 resin was SEPARATION OF EGG WHITE PROTEINS 1003 used as the cation exchange resin. Egg white was dilut- and concentrated using an ultrafiltration unit equipped ed with an equal volume of distilled water, Amberlite with a hollow fiber cartridge (30 kDa cut-off size, Quix- FPC 3500 resin was added (0.5 g/10 mL of diluted egg stand for laboratory scale; Flexstand for scale-up study, white), and then the mixture was stirred 12 h in a cold GE Healthcare Bio-Sciences Corp.) at 22°C, and heated room (4°C) using an overhead stirrer set at the low- at 70°C for 15 min to precipitate impurities. After re- est speed (RW20 digital, IKA Works Inc., Wilmington, moving the precipitant by centrifugation at 3,000 × NC). The solution was filtered through a paper towel g for 30 min at 4°C, the final supernatant containing (for laboratory scale) or centrifugation for 3,400 × g for ovalbumin was lyophilized as above (Labconco Corp.). 20 min at 4°C (for scale-up study), and the resin was collected and washed several times with distilled water,

Yield and Purity Calculation Downloaded from and then once with 0.1 M glycine-NaOH buffer, pH 9.3. The lysozyme trapped to the resin was eluted with The yields of lysozyme, ovomucin, ovotransferrin, 0.1 M glycine-NaOH buffer, pH 9.3, containing 0.5 M and ovalbumin were calculated using their theoretical NaCl. The eluent was desalted using an ultrafiltration values in egg white. Yield was calculated with the ratio unit equipped with a hollow fiber cartridge (10 kDa cut- between lyophilized protein and the theoretical value off sizes, Quixstand for laboratory scale and Flexstand could be presented. To check the separation efficiency, http://ps.oxfordjournals.org/ for scale-up study, GE Healthcare Bio-Sciences Corp., SDS-PAGE (Ko and Ahn, 2008) was conducted under Piscataway, NJ) and lyophilized (Labconco Corp., Kan- reduced conditions using Mini-Protein II cell (Bio- sas City, MO). The freeze dryer conditions were as fol- Rad). Ten percent SDS gel and Coomassie brilliant lows: collector temperature −51°C, vacuum set at 0.002 blue R-250 (Sigma-Aldrich) staining were used. The mbar, and drying chamber temperature set at −15°C. purity of protein was calculated by converting the density of protein bands in the gel picture using the ImageJ

Separation of Ovomucin software (NIH, Bethesda, MD) as the percent of the at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 total gel density. The actual protein content (10.95%) The lysozyme-free egg white (from the first step) was obtained from the egg white sample was used to calcu- adjusted to pH 4.75 using 3 N HCl to precipitate ovo- late the yields. mucin and centrifuged at 3,400 × g for 30 min at 4°C. Western blot was also used to confirm lysozyme, ovo- The precipitated ovomucin was collected, homogenized, transferrin, and ovalbumin. Western blot was carried and washed with 10 volumes of distilled water to re- out using the method of Xie et al. (2002) with some move impurities, and the washing was repeated 3 times. modifications. After running the SDS-PAGE, proteins The pH of the final homogenate with 10 volumes of dis- were transferred onto a nitrocellulose membrane (Bio- tilled water was adjusted to pH 12.0 using 3 N NaOH Rad) at 90 V for 2 h under controlled temperature. The to dissolve the protein. transferred membrane was blocked with 5% skim milk powder dissolved in PBS with Tween 20 (PBST). Separation of Ovotransferrin To identify lysozyme, the membrane was treated with and Ovalbumin rabbit polyclonal antibody to lysozyme (Abcam, after 1:10,000 dilution) and kept overnight at 4°C with shak- Ovotransferrin and ovalbumin was separated by pre- ing. The membrane was washed 3 times with PBST so- cipitating ovotransferrin using ammonium sulfate and lution at 15-min intervals, exposed to Amersham ECL citric acid combinations. The ovomucin- and lysozyme- Prime (GE Healthcare, after 1:1 dilution) for 5 min, free egg white solution (from the second step) was added and then analyzed using a Chemidoc (Bio-Rad). with 5.0% (wt/vol) ammonium sulfate and 2.5% (wt/ For ovalbumin, anti-ovalbumin antibody (Abcam) vol) citric acid combination (final pH 2.90), and held at was used after diluting 1:15,000 with distilled water 4°C for 12 h to precipitate ovotransferrin. After centrif- and incubating overnight at 4°C. For ovotransferrin, ugation at 3,000 × g for 30 min at 4°C, the precipitant ovotransferrin chicken antibody was used after dilut- and supernatant (S1) were collected. The supernatant ing 1:15,000. Rabbit anti-mouse IgG (H+L) conjugated was used for ovalbumin and precipitant for ovotransfer- AP12 (Abcam) was used after diluting 1:20,000 as the rin separation. The precipitant was dissolved with 4 secondary antibody, incubated 1 h at room tempera- volumes of distilled water and then reprecipitated using ture, and analyzed. a 2.0% ammonium sulfate and 1.5% citric acid combination (final pH 3.35). After centrifugation (3,400 × g ELISA Assay for the Activity for 30 min at 4°C), the precipitant was collected and of Purified Proteins dissolved in distilled water, desalted using an ultrafiltration unit equipped with a hollow fiber cartridge (30 The activity of the proteins was checked with the kDa cut-off size, Quickstand for laboratory scale; Flex- ELISA method of Vidal et al. (2005) with some modi- stand for scale-up study, GE Healthcare Bio-Sciences fications. Because the primary antibody for ovomucin Corp.) at 22°C, and then lyophilized (Labconco Corp.). was not available, ELISA was done only for ovalbu- The supernatant from the second precipitation (S2) min, ovotransferrin, and lysozyme. Standard curves was pooled with the first supernatant (S1), desalted were prepared using the standards purchased (lyso- 1004 ABEYRATHNE ET AL. zyme from Sigma-Aldrich, ovotransferrin and ovalbu- batch system, the resin was directly added to the di- min from Abcam). Each of the standards and separated luted egg white solution. Batch method was easy, fast, proteins were dissolved in a carbonate buffer (pH 9.6) and compatible for large-scale production of lysozyme and diluted to the final concentration of 10 μg/mL. For from egg white. The yield of lysozyme from egg white each protein, 100 μL of the diluted standard or sample using Amberlite FPC 3500 was around 90% and the were coated onto a microplate and incubated overnight purity over 95% purity (Table 1). No pretreatment, ex- at 4°C. The microplate wells were washed with PBS cept for 1:1 dilution of egg white solution, was used Tween 20 (200 μL) twice and blocked with 5% skim because the pH of fresh egg white is around 9.0 to 9.3, milk (200 μL) for 2 h at room temperature. The wells which was similar to the optimal pH conditions for were washed again with PBS Tween 20 twice, incubat- trapping lysozyme from egg white using the Amberlite ed with 100 μL of primary antibody (1:10,000, Abcam) FPC 3500 resin. Compared with the ammonium sulfate Downloaded from dissolved in 5% skim milk, incubated for 2 h at room separation (Alderton et al., 1945), CMC chromatogra- temperature in the dark, and then washed with PBST phy (Strang, 1984), β-mercaptoethanol with thermal 4 times. After removing incubation solution, 100 μL of treatment (Chang et al., 2000), ultrafiltration (Wan et substrate (ABTS substrate) was added and the samples al., 2006), and magnetic cation exchange chromatog- were kept at room temperature for 30 min in the dark. raphy (Safarik et al., 2007), the use of Amberlite FPC http://ps.oxfordjournals.org/ The reaction was stopped by adding 0.1 M citric acid 3500 resin was much easier, simpler, and more efficient and the absorbance was measured at 416 nm using a than egg white. Amberlite FPC 3500 performed better microplate reader (xMark, Bio-Rad). A secondary an- than CMC: with CMC, only part of the lysozyme in the tibody (rabbit anti-mouse IgG (H+L) conjugated AP12 egg white was trapped and a significant amount of ly- (Abcam) was used for ovotransferrin and ovalbumin sozyme was still remaining in egg white solution (data after diluting 1:10,000 ratio. not shown). Another associated issue of CMC resin was

that the fine particles of CMC floated on the surface of at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Statistical Analysis the liquid and resulted in loss of the resin during the equilibration and washing steps for scale-up study. The The separation protocol was replicated 3 times, and Amberlite FPC 3500 resin had larger granules (0.3–1.18 data were analyzed using Microsoft Excel 2010 (Micro- mm vs. 25–60 μm in diameter), which made it easier soft Corp., Redmond, WA). Differences in mean values to handle, had higher ion exchange capability (>2.6 were compared by 1-way ANOVA using Minitab 16.0 vs. 1.0 mEq/mL), and lower price than CM cellulose (State College, PA). resin. CM-Toyopearl 650M cation exchange resin also had small bead size and had much lower ion exchange RESULTS AND DISCUSSION capacity but much higher price than Amberlite FPC 3500 resin (0.24 vs. 2.6 mEq/mL). Thus, Amberlite Lysozyme Separation FPC 3500 is superior cation exchange resin for lysozyme separation from egg white to other resins. Carboxymethyl cellulose was used first by Strang Because no chemical was added to egg white, the (1984) to separate lysozyme from egg white. Howev- proteins in the lysozyme-free egg white solution had no er, because of its fine granule sizes and low efficiency, physical or chemical changes. Thus, separation of lyso- separation of lysozyme from egg white using CMC was zyme using FPC 3500 resins was selected as the first not easy especially in large scale. Therefore, Amber- step for the sequential separation of multiple egg white lite FPC 3500, a cation exchange chromatography resin proteins. After removing lysozyme, the resin could be with different granule size and ion exchange capacity, easily regenerated by washing the resin 2 to 3 times was used to separate lysozyme from egg white. When with 5 volumes of 3% HCl, rinsing 2 to 3 times with 10 Amberlite FPC3500 resin was used, either batch or col- volumes of distilled water, and then equilibrating with umn method could be used to separate lysozyme from 0.1 M glycine-NaOH buffer, pH 9.0 to 9.3, before using egg, but batch method was used in this study. For the it again.

Table 1. Yield and purity of ovalbumin, lysozyme, ovotransferrin, and ovomucin using the sequential separation method1

Laboratory scale Large scale

Weight Yield Purity Weight Yield Purity Activity Sample (g) (%) (%) (g) (%) (%) (10 μg/mL) Ovalbumin 16.240 ± 0.75 98.54 ± 4.47 94.2 213.23 ± 2.67 99.99 ± 1.24 88.2 9.85 ± 0.11 Lysozyme 0.96 ± 0.02 89.72 ± 1.93 96.5 12.27 ± 0.50 88.72 ± 3.61 96.2 9.86 ± 0.55 Ovotransferrin 3.07 ± 0.17 83.39 ± 5.33 96.2 39.36 ± 1.29 83.01 ± 2.69 94.3 9.77 ± 0.23 Ovomucin 1.08 ± 0.05 100.93 ± 2.25 82.2 13.95 ± 0.14 100.87 ± 2.69 82.2 ND2 1The calculated amount of egg white proteins in egg white solution (total egg white proteins in egg white was 10.95%). Laboratory scale (280 g of egg white): ovalbumin, 16.48 g; ovotransferrin: 3.66 g; lysozyme, 1.07 g; ovomucin, 1.07 g. Large scale (3,625 g of egg white): ovalbumin, 213.38 g; ovotransferrin, 47.42 g; lysozyme, 13.83 g; ovomucin, 13.83 g). n = 3. 2Not determined. SEPARATION OF EGG WHITE PROTEINS 1005 Separation and Solublization of Ovomucin In our sequential separation protocol, we have used isoelectric precipitation of ovomucin by adjusting the Ovomucin has a tendency to bind with other pro- pH of egg white to 4.75. Brining the pH of egg white teins, especially lysozyme. Therefore, removal of lyso- down to pH 4.5 to 5.0 neutralized most of the charges zyme from egg white solution at the first step can help of ovomucin and facilitated the separation of ovomucin increasing the purity of ovomucin. Centrifugation with from the rest of the proteins. The precipitated ovomu- different buffers and pH conditions was practiced by cin was easily separated by centrifugation at 3,400 × g many researchers to separate ovomucin from egg white. for 30 min at 4°C and did not require high-speed cen- The main reason for using pH adjustment and centrifu- trifugation. Although the pI of ovalbumin (pH 4.5) is gation to separate ovomucin is that it can be easily similar to that of the ovomucin, only a minimal amount precipitated by adjusting pH close to its pI ranges (pH of ovalbumin was precipitated along with ovomucin. On Downloaded from 4.5–5.0). Robinson and Monsey (1975) used isoelectric the other hand, ovomucin is easily separated due to its precipitation of ovomucin with Tris-HCl buffer and cen- large molecular weight once it reaches its pI value. Most trifugation at 35,000 × g and others (Omana and Wu, of the coprecipitated ovalbumin and other proteins 2009a) used a combination of low concentration NaCl trapped in ovomucin coagulates were washed off by ho- (100 mM) and high-speed centrifugation (11,200 × g). mogenizing and centrifuging the ovomucin precipitant. http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 1. The SDS-PAGE of egg white proteins collected over the sequential separation steps. Lane 1 = marker, lane 2 = egg white, lane 3 = egg white after removing lysozyme, lane 4 = lysozyme separated, lane 5 = egg white after removing lysozyme and ovomucin, lane 6 = ovomucin dissolve at pH 12.0, lane 7 = supernatant after ammonium sulfate and citric acid precipitation, lane 8 = separated crude ovotransferrin, lane 9 = supernatant after removing ovotransferrin, lane 10 = purified ovotransferrin, lane 11 = purified ovalbumin after heat treatment. Color version available in the online PDF. 1006 ABEYRATHNE ET AL. Thus, the separation of ovomucin from egg white can be The supernatants collected from the second ovo- done in large scale as well as in laboratory scale because transferrin precipitation (S2) was pooled with the first only pH adjustment and a low-speed centrifugation is supernatant (S1) from the 5.0% (wt/vol) ammonium required. Ovomucin is not soluble in water, but it could sulfate and 2.5% (wt/vol) citric acid precipitation and be solubilized by adjusting the pH to alkaline conditions used for ovalbumin separation. The pooled supernatant (pH 12.0, Figure 1a). Burgess and Deutscher (2009) re- was mainly composed of ovalbumin, but some impuri- ported that alkaline conditions increased the solubility ties remained in the solution. To remove the impurities, of ovomucin by interacting with disulfide bonds. Even the pooled supernatant was subjected to heat treatment though ovomucin bands were seen in all SDS-PAGE after desalting because ovalbumin has higher thermal gel tested (pH 7–12), the ovomucin was not completely resistance than ovotransferrin, the major impurities re- soluble until the pH reached to 12.0 (Figure 1a and 1b). maining in the supernatant (Stadelman and Cotterill, Downloaded from Because there is no commercially available ovomucin 2001). The heat treatment results indicated that heat- standard and anti-ovomucin antibody to identify ovo- ing the pooled, desalted supernatant at 70°C for 15 min mucin using western blotting, only SDS-PAGE results removed most of the impurities (Figure 2, lane 11), in- were shown. Previous studies reported 6 to 8 bands on a dicating that 2-step, low concentrations of ammonium SDS-PAGE gel when ovomucin was dissolved in alkaline sulfate and citric acid combinations efficiently separat- http://ps.oxfordjournals.org/ conditions (Huopalahti et al., 2007). Figure 1b showed ed ovalbumin and ovotransferrin from the lysozyme- 6 bands in SDS-PAGE as reported by Huopalahti et al. free and ovomucin-free egg white solution. This method (2007). The molecular weight of α-ovomucin varies from used a much smaller amount of ammonium sulfate com- 150 to 220 kDa, whereas β-ovomucin varies from 400 pared with the Hopkins (1900) method. Western blot to 523 kDa (Omana et al., 2010b). The multiple bands results in Figure 3 confirmed the proteins separated as shown in SDS-PAGE gel should be the products of the ovalbumin, lysozyme, and ovotransferrin.

2 subunits during the sample preparation, which uses at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 β-mercaptoethanol for the SDS-PAGE. Previous studies Yield and Purity of Lysozyme, Ovomucin, showed that ovomucin can also be dissolved using urea Ovotransferrin, and Ovalbumin and SDS (Omana et al., 2010b). After separating ovomucin from egg white, the remaining supernatant was In both laboratory scale and scale-up preparations, used to separate other proteins in the subsequent steps. the yield of ovalbumin 98 to 99%, lysozyme was 91 to 92%, ovomucin >100%, and ovotransferrin 82 to Separation of Ovotransferrin 83% (Table 1), which was much higher than that of and Ovalbumin the Tankrathok et al. (2009; lysozyme 55%, ovotransferrin 21%, and ovalbumin 54%). However, no separa- Hopkins (1900) used high levels of ammonium sulfate tion method can produce 100% yield even though there under acidic conditions to precipitate ovalbumin and are some impurities in each of the separated proteins. separate it from egg white. Our protocol also uses am- The exceptionally high yield for ovalbumin and >100% monium sulfate in acidic conditions, but our approach yield for ovomucin could be attributed to the following: is different from that of Hopkins. Instead of using high our proximate analysis of egg white (moisture 86.53%, levels of ammonium sulfate (saturated) in acidic condi- protein 10.95%, carbohydrate 1.77%, and ash 0.75%) tions to precipitate ovalbumin, we have used low levels indicated that the carbohydrate content was about 1% of ammonium sulfate and citric acid combination to higher than the reported value (Stadelman and Cotter- precipitate ovotransferrin and leave the rest of the egg ill, 2001). All egg white proteins are glycoproteins, but white proteins including ovalbumin soluble. The addi- the content of carbohydrate in ovomucin was higher tion of citric acid helped precipitation of ovotransferrin than other egg white proteins. Also, the sum of individ- by ammonium sulfate probably because citric acid low- ual protein content in egg white used for calculation in ered the pH of egg white solution toward the pI values this study was 95.8% (Stadelman and Cotterill, 2001), (6.0) of ovotransferrin. Our previous studies showed indicating that protein content, especially that of oval- that a 5.0% (wt/vol) ammonium sulfate and 2.5% (wt/ bumin and ovomucin could have been underestimated. vol) citric acid combination showed an excellent capa- The purity of lysozyme and ovotransferrin from the bility in precipitating ovotransferrin from egg white so- laboratory-scale preparation analyzed using the ImageJ lution (Abeyrathne et al., 2013). However, along with software was >96% and that of ovalbumin was over the ovotransferrin, a significant amount of ovalbumin 94%. The purity of lysozyme in large-scale preparation (about 35% of total precipitated proteins) was also pre- remained at 96%, but that of the ovotransferrin was cipitated. To increase the purity of ovotransferrin and 94%. The purity of ovalbumin with scale-up prepara- to recover some of the ovalbumin coprecipitated with tion was 88%, which were slightly lower than that with ovotransferrin, the precipitant mainly containing ovo- laboratory-scale preparation. The purity of ovomucin transferrin was dissolved with water and reprecipitated for both small- and large-scale preparations was 85%. using a lower level of ammonium sulfate (2.0% wt/vol) Although a sequential separation protocol for ovalbu- and citric acid (1.5% wt/vol) combination. min, ovomucin, ovotransferrin, and lysozyme from egg SEPARATION OF EGG WHITE PROTEINS 1007 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 2. a) Effect of pH on the solubility of ovomucin. Tube 1 = pH 7.0, tube 2 = pH 8.0, tube 3 = pH 9.0, tube 4 = pH 10.0, tube 5 = pH 11.0, and tube 6 = pH 12.0. b) Effect of pH on the SDS-PAGE of ovomucin. Lane 1 = marker, lane 2 = pH 7.0, lane 3 = pH 8.0, lane 4 = pH 9.0, lane 5 = pH 10.0, lane 6 = pH 11.0, and lane 7 = pH 12.0. Color version available in the online PDF.

Figure 3. Western blot pictures of ovotransferrin, ovalbumin, and lysozyme. M = marker, S = purified protein, Std = standard protein. 1008 ABEYRATHNE ET AL. Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 4. Schematic diagram for the separation of lysozyme, ovotransferrin, ovomucin, and ovalbumin from egg white. DW = distilled water. Color version available in the online PDF. white using counter-current chromatography has been combination of cation exchange chromatography, iso- reported (Shibusawa et al., 1998, 2001), scaling up was electric precipitation, ammonium sulfate and citric acid not practical. There had been several separation meth- precipitation, and heat treatment. This protocol sepa- ods developed for individual as well as multiple pro- rated the 4 major egg white proteins with >98% yield teins from egg white (Vachier et al., 1995; Tankrathok and >85% purity of ovalbumin; >88% yield and >95% et al., 2009; Geng et al., 2010), but this protocol is the purity of lysozyme; >80% yield and >90% purity of first one that showed the separation of the 4 major ovotransferrin, and >99% yield and >80% purity of egg white proteins in sequence is possible in both labo- ovomucin. Also, ovotransferrin, ovalbumin, and lyso- ratory and large-scale processes. Figure 4 is the final zyme showed over 95% activity. protocol for the sequential separation of lysozyme, ovomucin, ovotransferrin, and ovalbumin from egg white. ACKNOWLEDGMENTS Lysozyme and ovomucin could be separated within 2 d and ovotransferrin and ovalbumin within the next 2 d. This study was supported jointly by the Coopera- Therefore, all 4 proteins could be separated within 4 d tive Research Program for Agriculture Science & Tech- even in large-scale preparation. nology Development (Project No. PJ008460), Rural In conclusion, lysozyme, ovomucin, ovalbumin, and Development Administration, Republic of Korea, and ovotransferrin can be separated in sequence using a WCU (World Class University) program (R31-10056) SEPARATION OF EGG WHITE PROTEINS 1009 through the National Research Foundation of Korea MacDonell, L. R., H. Lineweaver, and R. E. Feeney. 1951. Chemistry funded by the Ministry of Education, Science and Tech- or shell egg deterioration; effect of reducing agents. Poult. Sci. 130:856–863. nology, Korea. Mine, Y., F. Ma, and S. Lauriau. 2004. Antimicrobial peptides released by enzymatic hydrolysis of hen egg white lysozyme. J. Agric. Food Chem. 52:1088–1094. REFERENCES Nisbet, A. D., R. H. Saundry, A. J. G. Moir, L. A. Fothergill, and J. E. Fothergill. 1981. The complete amino-acid sequence of hen ovalbumin. Eur. J. Biochem. 115:335–345. Abeyrathne, E. D. N. S., H. Y. Lee, J. S. Ham, and D. U. Ahn. 2013. Omana, D. A., J. Wang, and J. Wu. 2010a. Co-extraction of egg Separation of ovotransferrin from chicken egg white without us- white proteins using ion-exchange chromatography from ovo- ing organic solvents. Poult. Sci. 92:1091–1097. mucin-removed egg white. J. Chromatogr. B Analyt. Technol. Alderton, G., W. H. Ward, and H. L. Fevold. 1945. Isolation of lyso- Biomed. Life Sci. 878:1771–1776. Downloaded from zyme from egg white. J. Biol. Chem. 157:43–58. Omana, D. A., J. Wang, and J. Wu. 2010b. Ovomucin—A glyco- Awade, A. C., and T. Efstahiou. 1999. Comparison of three liq- protein with promising potential. Trends Food Sci. Technol. uid chromatography methods for egg-white protein analysis. J. 21:455–463. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 723:69–74. Omana, D. A., and J. Wu. 2009a. A new method of separating Burgess, R. R., and M. P. Deutscher. 2009. Methods in Enzymol- ovomucin from egg white. J. Agric. Food Chem. 57:3596–3603. ogy, Guide to Protein Purification. 2nd ed. Academic Press, Omana, D. A., and J. Wu. 2009b. Effect of different concentra- Waltham, MA. tions of calcium chloride and potassium chloride on egg white http://ps.oxfordjournals.org/ Chang, H. M., C. C. Yang, and Y. C. Chang. 2000. Rapid separation proteins during isoelectric precipitation of ovomucin. Poult. Sci. of lysozyme from chicken egg white by reductants and thermal 88:2224–2234. treatment. J. Agric. Food Chem. 48:161–164. Rabouille, C., M. A. Aon, G. Muller, J. Cartaud, and D. Thomas. Chick, H., and C. J. Martin. 1913. The precipitation of egg albumin 1990. The supramolecular organization of ovomucin—Biophysi- by ammonium sulfate. A contribution to the theory of the “Salt- cal and morphological studies. Biochem. J. 26:697–706. ing Out” of proteins. Biochem. J. 7:380–398. Radziejewska, R. C., G. LeĞnierowski, and J. Kijowski. 2008. Prop- Datta, D., S. Bhattacharjee, A. Nath, R. Das, C. Bhattacharjee, erties and application of egg white lysozyme and its modified and S. Datta. 2009. Separation of ovalbumin from chicken egg preparations—A review. Polish J. Food Nutr. Sci. 58:5–10.

white using two-stage ultrafiltration technique. Separ. Purif. Robinson, D. S., and J. B. Monsey. 1975. The composition and pro- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Tech. 66:353–361. posed subunit structure of egg-white β-ovomucin. Biochem. J. Desert, C., C. G. Dubiard, F. Nau, G. Jan, F. Val, and J. Mallard. 147:55–62. 2001. Comparison of different electrophoretic separations of hen Safarik, I., Z. Sabatkova, T. Oldrich, and M. Safarikov. 2007. Mag- egg white proteins. J. Agric. Food Chem. 49:4553–4561. netic cation exchange isolation of lysozyme from native hen egg Donovan, J. W., J. G. Davis, and L. M. White. 1970. Chemical and white. Food Technol. Biotechnol. 45:355–359. physical characterization of ovomucin, a sulfated glycoprotein a Shibusawa, Y., S. Kihira, and Y. Ito. 1998. One-step purification complex from chicken eggs. Biochim. Biophys. Acta 207:190– proteins from egg white using counter-current chromatography. 201. J. Chromatogr. B Biomed. Sci. Appl. 709:301–305. Dubiard, C. G., M. Pasco, A. Hietanen, A. Q. Bosque, F. Nau, Shibusawa, Y., S. Lino, H. Shindo, and Y. Ito. 2001. Separation of and T. Croguennec. 2005. Hen egg white fractionation by ion- chicken egg white proteins by high-speed countercurrent chroma- exchange chromatography. J. Chromatogr. A 1090:58–67. tography. J. Liquid Chromatogr. Relat. Technol. 24:2007–2016. Durance, T. D. 1994. Separation, purification and thermal stability Stadelman, W. J., and O. J. Cotterill. 2001. Egg Science and Tech- of lysozyme and avidin from chicken egg white. Pages 77–93 in nology. 4th ed. AVI Publ. Co., Westport, CT. Egg Uses and Processing Technologies, New Developments. CAB Strang, R. H. 1984. Purification of egg white lysozyme by ion ex- Int., Wallingford, UK. change chromatography. Biochem. Educ. 12:57–59. Fraenkel-Conrat, H., and R. Feeney. 1950. The metal-binding activ- Tankrathok, A., S. Daduang, R. Patramanon, T. Arakai, and S. ity of conalbumin. Arch. Biochem. 29:101–113. Thammasirirak. 2009. Purification process for the preparation Geng, F., Q. Huang, X. Wu, G. Ren, Y. Shan, G. Jin, and M. Ma. and characterization of hen egg white ovalbumin, lysozyme, ovo- 2010. Co-purification of chicken egg white proteins using polyeth- transferrin and ovomucoid. Prep. Biochem. Biotechnol. 39:380– ylene glycol precipitation and anion-exchange chromatography. 399. Separ. Purif. Tech. 96:75–80. Vachier, M. C., M. Piot, and A. C. Awade. 1995. Isolation of hen egg Hiidenhovi, J., H. S. Aro, and V. Kankare. 1999. Separation of ovo- white lysozyme, ovotransferrin and ovalbumin using a quaternary mucin subunits by gel filtration: Enhanced resolution of subunits ammonium bound to a highly cross-linked agarose matrix. J. by using a dual-column system. J. Agric. Food Chem. 47:1004– Chromatogr. A 664:201–210. 1008. Valenti, P., G. Antonini, M. R. R. Fanelli, N. Orsi, and E. Antonini. Hiidenhovi, J., J. Makinen, R. Hupalathti, and E. L. Ryhanen. 2002. 1982. Antibacterial activity of matrix-bound ovotransferrin. An- Comparison of different egg albumin fractions as source of ovo- timicrob. Agents Chemother. 21:840–841. mucin. J. Agric. Food Chem. 50:2840–2845. Vidal, M. L., J. Gautron, and Y. Nys. 2005. Development of an Hopkins, F. G. 1900. On the separation of a pure albumin from egg ELISA for quantifying lysozyme in hen egg white. J. Agric. Food white. J. Physiol. 25:306–330. Chem. 53:2379–2385. Huntington, J. A., and P. E. Stein. 2001. Structure and properties of Wan, Y., J. Lu, and Z. Cui. 2006. Separation of lysozyme from ovalbumin. J. Chromatogr. B Biomed. Sci. Appl. 756:189–198. chicken egg white using ultrafiltration. Separ. Purif. Tech. Huopalahti, R., R. L. Fandino, M. Anton, and R. Schade, ed. 2007. 48:133–142. Egg White Compounds. Subpart 1b. Pages 33–66 in Bioactive Wang, J., and J. Wu. 2012. Effect of operating conditions on the Egg Compounds. Springer, New York, NY. extraction of ovomucin. Process Biochem. 47:94–98. Ibrahim, H. R., Y. Sugmito, and T. Akoi. 2000. Ovotransferrin anti- Ward, T. M., R. A. Edwards, and R. D. Tanner. 2007. Separating microbial peptide (OTAP-92) kills bacteria through a membrane a mixture of egg yolk and egg white using foam fractionation. damage mechanism. Biochim. Biophys. Acta 1523:196–205. Appl. Biochem. Biotechnol. 137–140:927–934. Ko, K. Y., and D. U. Ahn. 2008. An economic and simple purifica- Williams, J. 1968. A comparison of glycopeptides from the ovotrans- tion procedure for the large scale production of ovotransferrin ferrin and serum transferrin of the hen. Biochem. J. 108:57–67. from egg white. Poult. Sci. 87:1441–1450. Wu, J., and A. Acer-Lopez. 2012. Ovotransferrin: Structure, bioac- Ko, K. Y., A. F. Mendonca, and D. U. Ahn. 2009. Effect of ethyl- tivities and preparation. Food Res. Int. 46:480–487. enediaminetetraacetate and lysozyme on the antimicrobial activ- Xie, H., G. R. Huff, W. E. Huff, J. M. Balog, P. Holt, and N. C. ity of ovotransferrin against Listeria monocytogenes. Poult. Sci. Rath. 2002. Identification of ovotransferrin as an acute phase 87:1649–1658. protein in chickens. Poult. Sci. 81:112–120.

Separation of ovotransferrin and ovomucoid from chicken egg white

E. D. N. S. Abeyrathne,*† H. Y. Lee ,* and D. U. Ahn *‡§1

* Department of Agricultural Biotechnology, Seoul National University, Seoul 151-921, Korea; † Department of Animal Science, Uva Wellassa University, Badulla, Sri Lanka 90000; ‡ Department of Animal Science, Iowa State University, Ames 50010; and § Department of Animal Science

and Technology, Sunchon National University, Sunchon 540-742, South Korea Downloaded from

ABSTRACT Ovotransferrin and ovomucoid were sepa- the acidic salt precipitation method, the ethanol in the rated using 2 methods after extracting the ovotransfer- supernatant was removed first. The ethanol-free solu- rin- and ovomucoid-containing fraction from egg white. tion was then concentrated and treated with a 2.5% 3+ Diluted egg white (2×) was added to Fe and treated ammonium sulfate and 2.5% citric acid combination. http://ps.oxfordjournals.org/ with 43% ethanol (final concentration). After centrifu- After centrifugation, the precipitant was used as the gation, the supernatant was collected and treated with ovotransferrin and the supernatant as the ovomucoid either a high-level ethanol (61% final concentration) fraction. The ovomucoid fraction from both of the pro- or an acidic salt combination (2.5% ammonium sulfate tocols was further purified by heating at 65°C for 20 and 2.5% citric acid) to separate ovotransferrin and min and the impurities were removed by centrifugation. ovomucoid. For the high-level of ethanol method, ovo- The yields of ovomucoid and ovotransferrin were >96 transferrin was precipitated using 61% ethanol. After and >92%, respectively. The purity of ovomucoid was centrifugation, the precipitant was dissolved in 9 vol. >89% and that of the ovotransferrin was >88%. The at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 of distilled water and the residual ethanol in the solu- ELISA results confirmed that the activity of the sepa- tion was removed using ultrafiltration. The superna- rated ovotransferrin was >95%. Both of the protocols tant, mainly containing ovomucoid, was diluted with separated ovotransferrin and ovomucoid effectively and 4 vol. of water, had ethanol removed, and was then the methods were simple, fast, and easy to scale up. concentrated and used as the ovomucoid fraction. For

Key words: ovotransferrin , ovomucoid , ethanol precipitation, acidic salt, yield , purity 2014 Poultry Science 93 :1010–1017 http://dx.doi.org/ 10.3382/ps.2013-03649

INTRODUCTION SDS-PAGE and immunoelectrophoresis, Tankrathok et al. (2009) used Carboxymethyl-Toyoperal 650M cation Ovotransferrin (12%) and ovomucoid (11%) are 2 exchange, and Omana et al. (2010) and Wu and Acer- major functional proteins found in the egg white (St- Lopez (2012) used anion exchange chromatography. adelman and Cotterill, 2001; Yousif and Kan, 2002). However, these methods were on a laboratory scale and Ovotransferrin can be present in 2 forms, apo- (iron were not practical for the scale-up processes. Ko and free) and holo-form (iron bound), which significantly Ahn (2008) separated ovotransferrin using ethanol pre- differ in their chemical and physical properties. Holo- cipitation, and the ethanol method used in that study form is more resistant to chemical and physical condi- was applicable for the large-scale production and hu- tions than apo-form (Ko and Ahn, 2008). Ovotransfer- man consumption. Recently, Abeyrathne et al. (2014) rin is known to have a strong iron-binding capability, developed an acidic salt method using an ammonium and thus can be used as an antioxidant, antimicrobial, sulfate and citric acid combination for ovotransferrin or iron-supplementing agent (Valenti et al., 1982; Ibra- separation from egg white. However, both methods him et al., 2000; Ko et al., 2008). Different approaches were targeted only for the separation of ovotransferrin have been used to separate ovotransferrin from the from egg white. egg white: Fraenkel-Conrat and Feeney (1950) used a Ovomucoid is the most highly glycosylated protein high-level ammonium sulfate, Desert et al. (2001) used among the egg white proteins (Kovacs-Nolan et al., 2000). The molecular weight of ovomucoid is 20.1 kDa, but it is visualized in SDS-PAGE gel at a range between © 2014 Poultry Science Association Inc. 30 and 40 kDa (Kovacs-Nolan et al., 2000), however, Received September 24, 2013. Accepted December 19, 2013. the reason for such an observation is still not clear. 1 Corresponding author: [email protected] Ovomucoid is well known as a trypsin inhibitor and is

1010 SEPARATION OF EGG WHITE PROTEINS 1011 considered the main food allergen present in the egg Abcam (Cambridge, UK). Rabbit antimouse IgG (H + white. Ovomucoid binds with trypsin at a 1:1 ratio, and L) conjugated AP12 antibody was purchased from Bio- its three-dimensional structure is secured with the 3 Rad (Richmond, CA). disulfide bonds in it (Oliveria et al., 2009). Ovomucoid was first separated using trichloroacetic acid and ac- Separation of Ovotransferrin etone by Lineweaver and Murray (1947), and then tri- and Ovomucoid chloroacetic acid and ethanol combination by Fredericq and Deutsch (1949). Ovomucoid was also separated us- The initial preparation was similar to that of Ko and ing 50% ethanol (Forsythe and Foster, 1950), but the Ahn (2008) for ovotransferrin separation. Egg white purity and the yield of the ovomucoid separated was was manually separated from yolk, diluted with 1 vol. Downloaded from not reported. Davies et al. (1971) separated ovomucoid of distilled water, added to Fe3+ (0.8 mL/L of 500 mM using 3-step chromatography, but they were not easy to FeCl3) to saturate ovotransferrin with iron, and then separate in large quantities. Tanabe et al. (2000) used homogenized using a hand mixer (KitchenAid, St. Jo- 25% ethanol (final concentration) to separate ovomu- seph, MI) for 2 min at high speed. The proteins in the coid from egg white. The recovery of ovomucoid from diluted egg white, except for ovotransferrin and ovo- the 25% ethanol method was around 70%, but the pu- mucoid, were precipitated by slowly adding 100% etha- http://ps.oxfordjournals.org/ rity of the protein was not reported. Yousif and Kan nol to the final concentration of 43%, and centrifuged (2002) separated ovomucoid from egg white using SDS- at 3,400 × g for 20 min at 4°C. To recover remaining PAGE with a linear gradient (4–20%), but this method ovotransferrin and ovomucoid in the precipitant, the cannot be scaled up and the protein was denatured dur- precipitant was reextracted with 2 vol. of 43% ethanol ing the separation due to the 2-mercaptoethanol used and centrifuged again. The 2 supernatants containing in the protocol. Most of the methods discussed were for ovotransferrin and ovomucoid were pooled and used as the separation of a single protein from egg white and the starting material for the separation of ovotransfer- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 were completed on a laboratory scale. rin and ovomucoid using two approaches. Vachier et al. (1995) first tried to separate lysozyme, The first method was treating the supernatant with ovotransferrin, and ovalbumin in sequence using anion a 2.5% (wt/vol) ammonium sulfate and 2.5% (wt/vol) exchange chromatography with Q-Sepharose fast flow citric acid combination (pH 3.0). Ammonium sulfate column. As of that trial, a few other separation pro- and citric acid combinations tested were from 2.5 to tocols for more than one protein from egg white have 10% (wt/vol) at intervals of 2.5% and the final con- been developed. Tankrathok et al. (2009) separated ov- centrations selected were 2.5% (wt/vol) each. Ethanol albumin, lysozyme, ovotransferrin, and ovomucoid in in the supernatant was removed using ultrafiltration sequence using Q-Sepharose fast-flow anion exchange (hollow fiber, 10,000 cut-off size; Quickstand for labora- chromatography in the first step and then with Car- tory scale and Flexstand for scale-up preparation, GE boxymethyl-Toyopearl 650M cation exchange chroma- Healthcare Bio-Sciences Corp., Piscataway, NJ), and tography in the second step. The yields of ovotransfer- then 2.5% (wt/vol) ammonium sulfate and 2.5% (wt/ rin and ovomucoid were 21 and 21%, respectively. Geng vol) citric acid combination was slowly added to the et al. (2012) also used a fast-flow anion exchange chro- ethanol-free supernatant to precipitate ovotransferrin. matography to separate lysozyme, ovotransferrin, oval- After holding overnight in a 4°C cold room, the sample bumin, and ovoflavoprotein from egg white. However, was centrifuged at 3,400 × g for 20 min at 4°C. To col- none of the sequential separation methods were used lect ovotransferrin, the precipitant was dissolved with 4 for the large-scale separation of the egg white proteins. vol. of distilled water with adjusting the pH to 9.0 using Therefore, the objective of the current study was to NaOH, desalted, and concentrated using ultrafiltration develop simple, economical methods, with the ability to (hollow fiber, 30,000 cut-off size; Quickstand for labora- scale up, to separate ovotransferrin and ovomucoid us- tory scale and Flexstand for scale-up preparation, GE ing the supernatant from the 43% ethanol precipitation Healthcare), and then lyophilized (Labconco Corp., of egg white proteins developed by Ko and Ahn (2008). Kansas City, MO). The supernatant mainly consisted of ovomucoid, but contained some ovotransferrin. After MATERIALS AND METHODS desalting using ultrafiltration (hollow fiber, 10,000 cut- off size; Quickstand for laboratory scale and Flexstand Materials for scale-up preparation, GE Healthcare), the supernatant was adjusted to pH 4.0 using citric acid, heat- Medium-size brown chicken eggs (<5 d after pro- treated at 65°C for 20 min, centrifuged at 3,400 × g for duction) were purchased from a local market and used 20 min at 4°C to remove the denatured ovotransferrin, within the next day. Ammonium sulfate, citric acid, and then lyophilized. The 2 protocols were tested for and ethanol were purchased from Fisher Scientific both laboratory- and large-scale separation of ovotrans- (Waltham, MA). Standard ovomucoid and ovotransfer- ferrin and ovomucoid. The laboratory scale refers to rin were purchased from Sigma-Aldrich (St. Louis, MO) separation of the proteins with about 300 to 400 mL of and antiovotransferrin antibodies were purchased from egg white and large scale used about 3 L of egg white. 1012 ABEYRATHNE ET AL. The second approach used the same method as Ko ple were coated onto a microplate and incubated over- and Ahn (2008) to precipitate ovotransferrin from the night at 4°C. The microplate wells were washed twice 43% ethanol extract. The ethanol concentration of the with 200 μL of PBST and blocked with 5% skim milk extract was increased to 61% (final) by slowly adding (200 μL) for 2 h at room temperature. The wells were 100% ethanol. The solution was kept at 4°C overnight washed again with PBST twice, incubated with 100 and then centrifuged at 3,400 × g for 20 min at 4°C. μL of primary antibody (1:10,000, Abcam) dissolved in The precipitant was collected, dissolved with 10 vol. of 5% skim milk, incubated for 2 h at room temperature distilled water, and the residual ethanol was removed in the dark, and then washed with PBST 2 times. A using an ultrafiltration unit (hollow fiber, 30,000 cut- secondary antibody [rabbit antimouse IgG (H+L) con- off sizes). The supernatant was subjected to ultrafiltra- jugated AP12; Abcam] was diluted to a 1:10,000 ratio tion (hollow fiber, 10,000 cut-off size), heat-treated to and incubated for 2 h at room temperature in the dark Downloaded from remove residual ovotransferrin, and then lyophilized as and washed with PBST 4 times. Then, 100 μL of sub- mentioned above. strate [2,2’-azino-bis(3-ethylbenzothiazoline-6-sulphon- ic acid)] was added and kept at room temperature for Yield and Purity Analyses 30 min in the dark. The reaction was stopped by adding 0.1 M citric acid and the absorbance was measured http://ps.oxfordjournals.org/ The yields of ovotransferrin and ovomucoid were cal- at 416 nm using a microplate reader (Bio-Rad). culated by comparing the amount of separated proteins after freeze-drying with the calculated values (Stadel- RESULTS AND DISCUSSION man and Cotterill, 2001) of each protein in egg white. To check the separation efficiency and yields, SDS- Separation of Ovotransferrin PAGE was conducted under reduced conditions using and Ovomucoid

Mini-Protein II cell (Bio-Rad). Ten percent polyacryl- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 amide gel and Coomassie Brilliant Blue R-250 (Sigma) The holo-form of ovotransferrin is more stable to staining was used. To check the purity of ovotransferrin chemical and pH changes than the iron-free (apo) form. and ovomucoid, gel pictures were taken after distaining. Ko and Ahn (2008) separated ovotransferrin from egg The purity of proteins was calculated by converting the white using 43% ethanol after converting apo-ovotrans- density of protein bands in the gel picture using the ferrin to holo-ovotransferrin. The same strategy was ImageJ software (NIH, Bethesda, MD). used to separate ovotransferrin from egg white. With Western blot was used to confirm ovotransferrin (Xie 43% ethanol (final concentration), all the egg-white et al., 2002). After running the SDS-PAGE, proteins proteins, except for ovotransferrin and ovomucoid, were were transferred onto a nitrocellulose membrane (Bio- denatured and (Figure 1, lane 3) precipitated. After Rad) at 90V for 2 h under a controlled temperature. centrifugation, ovotransferrin and ovomucoid in the su- The transferred membrane was blocked with 5% skim pernatant fraction were separated using 2 strategies: milk powder solution dissolved in phosphate-buffered one with a combination of ammonium sulfate and citric saline with tween-20 (PBST). To identify standard acid and the other with a high-level of ethanol (61% ovotransferrin, the membrane was treated with antiovo- final concentration). Each of the methods has its ad- transferrin antibodies to ovotransferrin (after 1: 15,000 vantages and disadvantages in separating ovotransfer- dilution as a primary antibody; Abcam) and kept over- rin and ovomucoid from the supernatant. Separation of night at 4°C with shaking. The membrane was washed ovotransferrin using the ammonium sulfate and citric 3 times with the PBST solution at 15-min intervals. acid combination requires an additional step to remove Rabbit antimouse IgG (H+L) conjugated AP12 (Ab- ethanol from the supernatant before the acidic-salt pre- cam) was used after diluting 1:20,000 as the second- cipitation, but the purity of the ovotransferrin separat- ary antibody, incubated for 1 h at room temperature ed was better than that with the high-level of ethanol and washed with PBST 3 times with 15-min intervals, precipitation (Table 1). Precipitation of ovotransferrin exposed to Amersham ECL Prime (after 1:1 dilution; with 61% ethanol denatured some of the ovotransferrin GE Healthcare) for 5 min, and then analyzed using a in the supernatant, but the protocol was simpler than Chemidoc (Bio-Rad). the acidic-salt precipitation method. Both protocols took 2 d to separate the 2 proteins and both were ap- ELISA Assay for the Activity plicable for scale-up production. of Purified Proteins The purities of ovotransferrin and ovomucoid were not changed by increasing ammonium sulfate concen- Only the activity of purified ovotransferrin was tration from 2.5 to 7.5% (Figure 2), and the 2.5% (wt/ checked using the ELISA method of Vidal et al. (2005) vol) ammonium sulfate concentration was selected as a because no antibody for ovomucoid was available. For salt concentration of the acidic salt method. Figure 2 the ELISA assay, ovotransferrin standard or separated indicated that in the combination of 2.5% (wt/vol) am- ovotransferrin was dissolved in a carbonate buffer (pH monium sulfate and 2.5% (wt/vol) citric acid separated 9.6) and diluted to the final concentration of 10 ng/mL. ovotransferrin from the supernatant, some ovotransfer- One hundred microliters of diluted standard and sam- rin still remained in the supernatant (ovomucoid) frac- SEPARATION OF EGG WHITE PROTEINS 1013

Table 1. Yield and purity of ovotransferrin and ovomucoid using 2 sequential separation protocols1

Small scale Large scale

Item Weight (g) Yield (%) Purity (%) Weight (g) Yield (%) Purity (%) Sample 4.70 ± 0.14 98.12 ± 2.81 98.10 38.66 ± 0.46 98.07 ± 1.16 97.70 Ovomucoid2 4.41 ± 0.05 100.46 ± 1.04 92.30 36.29 ± 0.40 100.44 ± 2.18 90.16 Ovotransferrin3 4.56 ± 0.13 95.20 ± 2.62 90.25 36.64 ± 0.54 92.94 ± 1.34 88.00 Ovomucoid3 4.42 ± 0.04 100.61 ± 0.92 91.40 34.99 ± 0.40 96.84 ± 1.11 89.29 1Calculated amount of egg white proteins in egg white solution (total egg white proteins in egg white was 10.95%). Small scale (365 g of egg white) = ovotransferrin, 4.79 g; ovomucoid, 4.39 g. Large scale (3,000 g of egg white) = ovotransferrin 39.42 g; ovomucoid, 36.13 g. n = 3. 2

Ovotransferrin and ovomucoid produced using ethanol precipitation protocol. Downloaded from 3Ovotransferrin and ovomucoid produced using acidic ammonium sulfate protocol. tion. The supernatant with the high-level of ethanol Moon et al., 2013). Ovomucoid was reported as one of precipitation of ovotransferrin also had some ovotrans- the main egg allergens present in the egg white (Stadel- ferrin and other impurities. Figures 2 and 3 indicated man and Cotterill, 2001), but separated ovomucoid can http://ps.oxfordjournals.org/ that crude ovomucoid solution after centrifugation had be used to produce bioactive peptides that have metal- some ovotransferrin and other impurities (Figure 2, binding, antioxidant, antihypertensive, and antimicro- lane 3 and Figure 3, lane 2). To remove the residual bial activities (unpublished data). proteins, crude ovomucoid solution was heat-treated. Stadelman and Cotterill (2001) reported that ovomu- Purity, Yield, and Activity coid is a more heat-stable protein than ovotransferrin, of the Separated Proteins

and heating crude ovomucoid solution at 65°C did not at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 damage ovomucoid. In Figure 3, different heating times The yield of ovotransferrin in the small scale was 98% (15, 20, 25, and 30 min) at 65°C were tested and 20 min for ethanol separation and 95% for the acid-salt precip- was selected as a heating condition for both methods to itation method, whereas that of the large scale was 98 improve the purity of ovomucoid. and 92%, respectively. The yield of ovomucoid in small The separated ovotransferrin can be used to produce scale was over 100% for both methods, whereas that functional peptides with growth-inhibiting effects on of the large scale was over 100% for the ethanol pre- human cancer cell lines (Ibrahim and Kiyono, 2009; cipitation protocol and >96% for the acidic ammonium

Figure 1. The SDS-PAGE of egg white proteins collected over the sequential separation steps. Lane 1 = marker; lane 2 = diluted egg white; lane 3 = supernatant after 43% ethanol addition; lane 4 = supernatant after 61% ethanol before heat treatment; lane 5 = ovomucoid from ethanol treatment; lane 6 = ovotransferrin from ethanol treatment; lane 7 = supernatant after 2.5% ammonium sulfate and citric acid and before heat treatment; lane 8 = ovomucoid after heat treatment; lane 9 = ovotransferrin after 2.5% ammonium sulfate and citric acid. Color version available in the online PDF. 1014 ABEYRATHNE ET AL. Downloaded from http://ps.oxfordjournals.org/

Figure 2. The SDS-PAGE picture of the lower levels of ammonium sulfate (AS) and citric acid (CA) treatments for purifying ovotransferrin at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 from ovomucoid. Lane 1 = marker; lane 2 = crude ovomucoid and ovotransferrin; lane 3 = supernatant of AS 2.5 and CA 2.5; lane 4 = supernatant of AS 5.0 and CA 2.5; lane 5 = supernatant of AS 7.5 and CA 2.5; lane 6 = supernatant of AS 10.0 and CA 2.5; lane 7 = precipitant of AS 2.5 and CA 2.5; lane 8 = precipitant of AS 5.0 and CA 2.5; lane 9 = precipitant of AS 7.5 and CA 2.5; lane 10 = precipitant of AS 10.0 and CA 2.5. Color version available in the online PDF.

Figure 3. The SDS-PAGE picture of the heating treatment to remove residual proteins at 65°C. Lane 1 = marker; lane 2 = crude ovomucoid; lane 3 = supernatant after heating for 15 min; lane 4 = supernatant after heating for 20 min; lane 5 = supernatant after heating for 25 min; lane 6 = supernatant after heating for 30 min. Color version available in the online PDF. SEPARATION OF EGG WHITE PROTEINS 1015 sulfate method (Table 1). The purity of ovomucoid was and 21 vs. 97% for ovomucoid and ovotransferrin, re- 90 to 92% for the ethanol precipitation method and spectively). Recently, Geng et al. (2012) developed a 89 to 91% for the acidic ammonium sulfate method, separation method for lysozyme, ovotransferrin, oval- whereas that of ovotransferrin was 97 to 98% for the bumin, and ovoflavoprotein with a yield of over 90%, ethanol precipitation protocol and 88 to 90% for the but this method was not easy for scale-up production acidic ammonium sulfate method (Table 1). Further and is not practical due to the sophisticated separation analyses of ovomucoid and ovotransferrin using their procedure used. enzyme hydrolysates indicated that the major impuri- The 2 protocols (ethanol and acidic ammonium sul- ties of ovomucoid were riboflavin-binding proteins, ser- fate) were tested for both in laboratory- and large-scale ine protease inhibitor, and ovotransferrin, whereas that protocols. Both sizes showed good separation of the of the ovotransferrin was ovalbumin (data not shown). 2 proteins. The differences in the yield between labo- Downloaded from The yield was calculated according to the calculated ratory- and large-scale processes were mainly due to values using the literature values for ovotransferrin and loss of the proteins during ultrafiltration, which proves ovomucoid (Stadelman and Cotterill, 2001) and the that the methods can be used for large preparation actual protein content measured from egg white used of ovotransferrin and ovomucoid. Figure 1 shows the in the current study. The yield and purity results of proteins during the separation steps of the 2 methods. http://ps.oxfordjournals.org/ our study suggested that the amount of ovomucoid in According to Figure 1, both proteins showed similar egg white could be higher than the reported value. We separation patterns. Figure 4 confirms the separated have seen similar results with ovomucin in our previous ovotransferrin with Western blotting and ovomucoid study (Abeyrathne et al., 2014). with SDS-PAGE with relevant standards. Figures 5A This work is different from that of Ko and Ahn (2008) and B are the flow charts of the 2 protocols for sepa- and Abeyrathne et al. (2013) who used ethanol extrac- rating ovotransferrin and ovomucoid. The activity of

tion and ammonium sulfate and citric acid combination, the ovotransferrin separated using both of the protocols at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 respectively, to separate ovotransferrin from egg white. was over 95%. The activity of ovotransferrin separated The previous 2 methods separated only ovotransferrin, using the ethanol protocol was 95.8%, and the activ- whereas current protocols separated both ovotransfer- ity of that separated using the ammonium sulfate and rin and ovomucoid in a single sequence. Compared with citric acid combination was 97.8%, indicating that the the sequential separation of Tankrathok et al. (2009), ovotransferrin separated using both of the protocols both of the current protocols produced much higher maintained their activity. Even though this value was yields of ovomucoid and ovotransferrin (21 vs. 100% slightly lower than that of the previous studies (Abey-

Figure 4. Western blot pictures of ovotransferrin and SDS-PAGE picture of ovomucoid. Mr = marker; OTA = ovotransferrin separated with ammonium sulfate and citric acid; OTE = ovotransferrin separated with ethanol; St = standard protein; OME = ovomucoid separated using the ethanol precipitation method; OMA = ovomucoid from ammonium sulfate and citric acid precipitation method. Color version available in the online PDF. 1016 ABEYRATHNE ET AL. Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 5. (A) Schematic diagram for the separation of ovotransferrin and ovomucoid from egg white using ethanol. (B) Schematic diagram for the separation of ovotransferrin and ovomucoid from egg white using ethanol followed with ammonium sulfate and citric acid. Color version available in the online PDF. SEPARATION OF EGG WHITE PROTEINS 1017 rathne et al., 2013), the activity reported herein is still Desert, C., C. G. Dubiard, F. Nau, G. Jan, F. Val, and J. Mallard. very high. The ELISA of ovomucoid was not working 2001. Comparison of different electrophoretic separations of hen egg white proteins. J. Agric. Food Chem. 49:4553–4561. using currently available ovomucoid antibodies. Forsythe, R. H., and J. F. Foster. 1950. Egg white proteins II. An ethanol fractionation scheme. J. Biochem. 184:385–392. Fraenkel-Conrat, H., and R. Feeney. 1950. The metal-binding activ- Comparison of the 2 Protocols ity of conalbumin. Arch. Biochem. 29:101–113. for Separating Ovotransferrin Fredericq, E., and H. F. Deutsch. 1949. Studies on ovomucoid. J. Biol. Chem. 181:499–510. and Ovomucoid Geng, F., Q. Huang, X. Wu, G. Ren, Y. Shan, G. Jin, and M. Ma. 2012. Co-purification of chicken egg white proteins using glycol Separating ovotransferrin and ovomucoid using etha- purification and anion exchange chromatography. Separ. Purif.

nol was simple and fast, but the amount of ethanol Tech. 96:75–80. Downloaded from needed is high. The ammonium sulfate and citric acid Ibrahim, H. R., and T. Kiyono. 2009. Novel anticancer activity of combination protocol uses less ethanol and low levels the autocleaved ovotransferrin against human colon and breast cancer cells. J. Agric. Food Chem. 57:11383–11390. of ammonium sulfate and citric acid to precipitate ovo- Ibrahim, H. R., Y. Sugmito, and T. Akoi. 2000. Ovotransferrin anti- transferrin, but needs an extra step to remove etha- microbial peptide (OTAP-92) kills bacteria through a membrane damage mechanism. Biochim. Biophys. Acta 1523:196–205.

nol before the precipitation step. Heating is required http://ps.oxfordjournals.org/ Ko, K. Y., and D. U. Ahn. 2008. An economic and simple purifica- for both protocols to increase the purity of ovomucoid. tion procedure for the large scale production of ovotransferrin Both protocols produced ovotransferrin and ovomucoid from egg white. Poult. Sci. 87:1441–1450. with similar yield and purity, but the ovotransferrin Ko, K. Y., A. F. Mendonca, and D. U. Ahn. 2008. Effect of ethyl- produced is holo-form and should be converted to apo- enediaminetetraacetate and lysozyme on the antimicrobial activity of ovotransferrin against Listeria monocytogenes. Poult. Sci. form if necessary. 87:1649–1658. Kovacs-Nolan, J. K., J. W. Zhang, S. Hayakawa, and Y. Mine. 2000. Conclusions Immunochemical and structural analysis of pepsin-digested egg white ovomucoid. J. Agric. Food Chem. 48:6261–6266. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Lineweaver, H., and C. W. Murray. 1947. Identification of trypsin Ovomucoid and ovotransferrin can be easily separated inhibitor of egg white ovomucoid. J. Biol. Chem. 171:565–581. in sequence from the supernatant of ethanol extraction Moon, S. H., J. H. Lee, K. H. Chang, J. Y. Paik, D. U. Ahn, and H. using a high concentration of ethanol or ammonium D. Paik. 2013. Screening for cytotoxic activity of ovotransferrin sulfate and citric acid combination. The yields of the and its enzyme hydrolysates. Poult. Sci. 92:424–434. Oliveria, F. C., J. S. R. Coimbra, L. H. M. Silva, E. E. G. Rojas, separated ovotransferrin and ovomucoid were >92 and and M. C. H. Silva. 2009. Ovomucoid partitioning in aqueous >95%, respectively, the purity of the separated proteins two-phase systems. Biochem. Eng. J. 47:55–60. ranged from 88 to 98%, and the activity of the purified Omana, D. A., J. Wang, and J. Wu. 2010. Co-extraction of egg white ovotransferrin using both of the protocols remained at proteins using ion exchange chromatography from ovomucin-removed egg white. J. Chromatogr. B Analyt. Technol. Biomed. >95%. The protocols are simple, easy, and scalable to Life Sci. 878:1771–1776. separate the 2 proteins in sequence. Stadelman, W. J., and O. J. Cotterill. 2001. Egg Science and Tech- nology. 4th ed. AVI Publishing Co., Westport, CT. Tanabe, S., S. Tesaki, and M. Watanabe. 2000. Producing a low ACKNOWLEDGMENTS ovomucoid egg white preparation by precipitation with aqueous ethanol. Biosci. Biotechnol. Biochem. 64:2005–2007. This study was supported jointly by the Coopera- Tankrathok, A., S. Daduang, R. Patramanon, T. Arakai, and S. tive Research Program for Agriculture Science & Tech- Thammasirirak. 2009. Purification process for the preparation and characterization of hen egg white ovalbumin, lysozyme, ovo- nology Development (Project No. PJ008460), Rural transferrin and ovomucoid. Prep. Biochem. Biotechnol. 39:380– Development Administration, Republic of Korea, and 399. World Class University program (Republic of Korea, Vachier, M. C., M. Piot, and A. C. Awade. 1995. Isolation of hen egg R31-10056) through the National Research Foundation white lysozyme, ovotransferrin and ovalbumin using a quaternary ammonium bound to a highly cross-linked agarose matrix. J. of Korea funded by the Ministry of Education, Science Chromatogr. A 664:201–210. and Technology. Valenti, P., G. Antonini, M. R. R. Fanelli, N. Orsi, and E. Antonini. 1982. Antibacterial activity of matrix-bound ovotransferrin. An- timicrob. Agents Chemother. 21:840–841. REFERENCES Vidal, M. L., J. Gautron, and Y. Nys. 2005. Development of an ELISA for quantifying lysozyme in hen egg white. J. Agric. Food Abeyrathne, E. D. N. S., H. Y. Lee, and D. U. Ahn. 2014. Sequential Chem. 53:2379–2385. separation of lysozyme, ovomucin, ovotransferrin, and ovalbumin Wu, J., and A. Acer-Lopez. 2012. Ovotransferrin: Structure, bioac- from egg white. Poult. Sci. 93:1001–1009. 10.3382/ps.2013-03403. tivities and preparation. Food Res. Int. 46:480–487. Abeyrathne, E. D. N. S., H. Y. Lee, J. S. Ham, and D. U. Ahn. 2013. Xie, H., G. R. Huff, W. E. Huff, J. M. Balog, P. Holt, and N. C. Separation of ovotransferrin from chicken egg white without us- Rath. 2002. Identification of ovotransferrin as an acute phase ing organic solvents. Poult. Sci. 92:1091–1097. protein in chickens. Poult. Sci. 81:112–120. Davies, J. G., C. J. Mapes, and J. W. Donovan. 1971. Batch purifi- Yousif, A. N., and J. W. Kan. 2002. Visualization of chicken ovomu- cation of ovomucoid and characterization of the purified product. coid in polyacrylamide gels. Anal. Biochem. 311:93–97. Biochemistry 10:39–42.

Bioactive amines and microbiological quality in pasteurized and refrigerated liquid whole egg 1

I. O. P. Rêgo ,* L. D. M. Menezes,* T. C. Figueiredo ,† D. D. Oliveira ,† J. S. R. Rocha ,† L. J. C. Lara ,† A. L. Lima ,† M. R. Souza ,† and S. V. Cançado† 2

* Instituto Mineiro de Agropecuária, BR 040-KM 527, Contagem, 32145-900, Brazil; and † Departamento de Tecnologia e Inspeção de Produtos de Origem Animal, Escola de Veterinária da Downloaded from Universidade Federal de Minas Gerais, Avenida Antônio Carlos, 6627, Belo Horizonte, 30123-970, Brazil

ABSTRACT The objective of this study was to evaluate < 0.05) in the numbers of microorganisms. Levels of the correlation between the levels of bioactive amines the amines putrescine, cadaverine, and tyramine were and the microbiological quality of liquid pasteurized detected only in fertile liquid pasteurized egg, and the http://ps.oxfordjournals.org/ egg stored under refrigeration. Pasteurized whole egg storage period contributed to the increase (P < 0.05) liquid was obtained from 2 types of different raw mate- in the levels of these amines. There was a high correla- rials, fresh eggs, and commercial fertile eggs. They were tion between total coliform most probable number and stored under refrigeration over a period of 21 d. The cadaverine levels, and a moderate correlation between treatments were arranged in a completely randomized the numbers of aerobic mesophilic microorganisms and split plot, with the plots being the 2 types of liquid tyramine levels. It was concluded that the most con-

pasteurized egg, and the subplots being the 4 storage taminated liquid pasteurized eggs were the fertile liquid at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 periods (1, 7, 14, and 21 d). The storage period did not pasteurized eggs and this caused the highest levels of contribute to the increase (P > 0.05) of contamination bioactive amines in them compared with all the eggs by mesophilic aerobic microorganisms and total coli- that had been subjected to pasteurization and refriger- forms in commercial liquid pasteurized egg. However, ated storage. for fertile eggs, the storage period led to an increase (P Key words: liquid pasteurized egg, quality , storage , bioactive amine, microbiology 2014 Poultry Science 93 :1018–1022 http://dx.doi.org/ 10.3382/ps.2013-03694

INTRODUCTION a potentially nutritive food and prevent economic loses to the industry. Eggs are one of the most complete foods for human To deliver a nutrient-rich, highly digestible, quality consumption because they are rich in vitamins, min- food product to the consumer, some care is required erals, fatty acids, and proteins that provide several to prevent eggs from becoming an excellent medium essential amino acids of excellent biological value. In for microbial growth. Liquid pasteurized egg may be the food industry, dried pasteurized egg are used pref- contaminated in various ways, including microbial erentially rather than fresh eggs (shell eggs), because contamination of raw materials, failure of the bino- in addition to retaining flavor, color, nutritional value, mial time-temperature for pasteurization, or postpro- and functional properties, they offer advantages such cessing contamination. Proper pasteurization hygiene as better uniformity, less storage space, and ease of and equipment are factors that determine the final portion measurement (Aragon-Alegro et al., 2005). In residual microbiota. These factors, together with the general, liquid pasteurized egg products are processed cooling process and storage conditions, are often what using nonfertile eggs. However, the use of fertile eggs, determine the product’s shelf life (Hamid-Samimi and still fresh and before being placed into the hatching ma- Swartzel, 1985; Ball et al., 1987; Hara-Kudo and Taka- chine, would be an alternative to avoid the discharge of tori, 2009). Bioactive amines are organic bases of low molecu- © 2014 Poultry Science Association Inc. lar weight that can be classified as biogenic amines or Received October 15, 2013. polyamines. The former are growth factors, whereas the Accepted December 16, 2013. latter are formed by decarboxylation of amino acids by 1 Support: Colegiado de Pós-Graduação em Ciência Animal and Pró-Reitoria de Pesquisa da Universidade Federal de Minas Gerais. microbial enzymes. Studies indicate that many amine- 2 Corresponding author: [email protected] producing bacteria are especially representatives of the

1018 BIOACTIVE AMINES AND MICROBIOLOGICAL QUALITY 1019 Enterobacteriaceae family, especially the genera Esch- et al. (2001). For the presumptive test, 1 mL of liquid erichia, Enterobacter, Salmonella, Shigella, and Proteus pasteurized egg (dilutions of 100, 10−1, and 10−2) was alongside species of the genera Achromobacter, Lacto- incubated (48 ± 2 h at 35°C) in lauryl tryptose broth bacillus, Leuconostoc, Pseudomonas, Pediococcus, Strep- (Difco, Sparks, MD). For the confirmatory test for co- tococcus, Propionibacterium, and Clostridium (Russell liforms and thermotolerant coliforms, a loop of suspen- and Snyder, 1968; Halász et al., 1994; Lima and Glória, sion from each lauryl tryptose test tube producing gas 1999). was transferred to test tubes containing brilliant green, Because the deterioration of food is associated with lactose bile broth 2%, and EC broth (Difco), which increased levels of some of these amines, their use as an were then incubated at 48 ± 2 h at 35°C and 48 ± 2 h index of chemical quality has been proposed for some at 45.5°C, respectively. For the total aerobic mesophilic foods such as fish (Veciana-Nogués et al., 1997) and count (cfu/mL), 1 mL of liquid pasteurized egg (dilu- Downloaded from eggs (Saito et al., 1992; Bardócz et al., 1995; Okamoto tions of 10−2, 10−3, and 10−4) was placed in a sterile et al., 1997; Oliveira et al., 2009). Knowledge of the dish of Petrifilms Aerobic Count Plate (3M, St. Paul, levels of bioactive amines in liquid pasteurized egg is MN) and incubated 48 ± 2 h at 35°C (AOAC, 1990). important because amines are thermally stable, making Determination of Bioactive Amines. Amines were their quantification an indication of the quality of the extracted from 3-g samples with 20 mL of 50 g/L tri- http://ps.oxfordjournals.org/ final product. chloroacetic acid (TCA) in 3 successive extractions (7, The objective of this study was to evaluate the cor- 7, and 6 mL). After agitation for 10 min in a vortex relation between the levels of bioactive amines and the mixer, the slurries were centrifuged at 10,000 × g at microbiological quality of liquid pasteurized whole egg 4°C for 21 min and the supernatants were collected, obtained from 2 different types of raw materials, com- combined, and then filtered through 13-mm-diameter mercial eggs and fertile eggs, which were subjected to and 45 mM HAWP pore membranes (Millipore, Bed-

pasteurization and refrigerated storage. ford, MA). The amines were separated by ion-pair at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 HPLC using a μBondapak C18 column, 300 × 3.9 mm, MATERIALS AND METHODS 10 mm (Waters, Milford, MA). They were quantified by fluorescence at 340 and 445 nm of excitation and Samples and Experimental Conditions emission, respectively, after postcolumn derivation with o-phthalaldehyde (Moreira et al., 2008). The quantifi- Twelve batches of liquid whole egg subjected to pas- cation of amines was achieved by interpolation in an teurization and refrigerated storage were analyzed. external standard curve. Commercial eggs (fresh eggs disqualified due to weight The amine standards (spermine tetrahydrochloride, or being cracked or broken) and fertile eggs (disquali- spermidine trihydrochloride, putrescine dihydrochlo- fied for hatching due to low weight or surplus produc- ride, cadaverine dihydrochloride, histamine dihydro- tion) were used as raw materials for pasteurization. chloride, serotonine hydrochloride, β-phenylethylamine After the pasteurization of 6 batches of commercial hydrochloride, agmatine sulfate, tyramine, and trypt- eggs, a 1-L sample of each batch was collected. Six 1-L amine) were purchased from Sigma Chemical Co. (St. samples were also collected from the 6 batches of fer- Louis, MO). The reagents were of analytical grade, and tile liquid pasteurized egg. Hence, each sample of liquid HPLC solvents were of chromatographic grade. pasteurized egg represented one batch or one repetition and each batch of both eggs had 54.546 units. All eggs Statistical Analyses were broken in an automatic machine Diamond 120s model (Moba, Barneveld, the Netherlands) and soon The treatments were arranged in a completely ran- after were subjected to high-temperature, short-time domized split plot, with the plots being the 2 types of pasteurization following the binomial time-temperature liquid pasteurized egg, and subplots being the 4 storage 60°C for 3.5 min (USDA, 2008) using an on-plate heat periods (1, 7, 14, and 21 d), making a total of 8 treat- exchanger (Inoxil, Guarulhos, Brazil). Hydrogen perox- ments with 6 replicates (n = 48). The microbiological ide was not added to the eggs. variables and levels of bioactive amines were subjected All samples of liquid pasteurized egg were divided to nonparametric statistical analysis. The Kruskal-Wal- into 4 equal portions and placed into 4 sterile contain- lis test was applied to the plot, and the Friedman test ers of 250 mL each. The 4 containers of each batch of was applied to the subplot. The Pearson correlation at samples were dated and stored under refrigeration at 5% probability was also performed to correlate the mi- 3.5°C for up to 21 d. Refrigerated liquid pasteurized egg crobiological results at the level of bioactive amines. was analyzed at 1, 7, 14, and 21 d. RESULTS AND DISCUSSION Methods of Analysis Microbiological Analyses Microbiological Analyses. The most probable numbers (MPN/mL) of total and thermotolerant co- Commercial liquid pasteurized egg showed no signifi- liforms were determined in accordance with Hitchins cant differences (P < 0.05) for counts of mesophilic aer- 1020 RÊGO ET AL.

Table 1. Mean counts of mesophilic aerobic microorganisms (log cfu/mL) in fertile and commercial liquid pasteurized egg stored under refrigeration for up to 21 d

Days of storage

Type of egg 1 7 14 21 Commercial 3.4 ± 0.28a,x 3.36 ± 0.27a,x 3.41 ± 0.41a,x 5.65 ± 1.79a,x Fertile 2.52 ± 0.42b,x 5.94 ± 0.35a,y 6.34 ± 0.37a,y >6.40 ± 0.51a,y a,bMeans followed by different letters in the same row are different (P < 0.05; Kruskal-Wallis test). x,yMeans followed by different letters in the same column are different (P < 0.05; Friedman test). Downloaded from obic microorganisms (Table 1) with increasing storage liquid pasteurized egg are biogenic amines whose for- period. However, for samples of fertile liquid pasteur- mation is due to the decarboxylation of amino acids by ized egg, numbers of mesophilic aerobic microorgan- microbial enzymes (Gloria, 2005). isms were higher with increasing storage period. Their With the lengthening of the storage period, there was counts reached above 6.40 log cfu/mL on d 21. an increase (P < 0.05) in the levels of putrescine. At 21 http://ps.oxfordjournals.org/ Because the fertile egg remains in contact with the d, this value reached 15.3 mg/kg of sample. Saito et al. nest and the birds longer, it is therefore exposed to (1992) observed putrescine level equal to 15.2 mg/kg in more contamination and can carry a larger number of deteriorated eggs. However, Bardócz et al. (1995) found microorganisms on the inside. However, the pasteuri- 0.35 mg/kg of putrescine in boiled eggs, and Oliveira zation temperature should be sufficient to inactivate et al. (2009) observed only 0.17 mg/kg of putrescine these microorganisms. Therefore, the primary hypoth- in eggs after 40 d of refrigerated storage. Figueiredo et esis regarding fertile liquid pasteurized egg samples al. (2013) did not observe the presence of this amine in at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 suggested for the present study is that raw material eggs stored for up to 28 d either at room temperature before pasteurization was excessively contaminated, or refrigerated. considering that the other factors such as control of the Cadaverine (2.9 mg/kg) and tyramine (9.1 mg/kg) binomial time-temperature, packaging, and transport were detected in samples of fertile liquid pasteurized were controlled. egg at 21 d of storage. Cadaverine was found in all 6 Table 2 shows the microbiological results for total samples of fertile eggs and tyramine in 4 samples. Saito coliforms. During the entire commercial liquid pasteur- et al. (1992) observed higher levels of cadaverine (55.2 ized egg evaluation period, only one sample (4.1%) on mg/kg) in deteriorated eggs. Nishimura et al. (2006) d 7 was contaminated by total coliforms. However, all found 26.6 mg/kg of cadaverine in egg yolk. In contrast, fertile liquid pasteurized egg samples were positive for Ramos et al. (2009) observed decreased levels of cadav- total coliforms on d 21. In this experiment, MPN of erine in egg yolk (10.80 mg/kg to 7.39 mg/kg) after 26 total coliforms ranged from <0.3 to 110. d of refrigerated storage. All commercial liquid pasteurized egg samples were The literature contains no reports on amine research negative for thermotolerant coliforms. However, on the in liquid pasteurized egg or on the detection of tyra- first day of storage, thermotolerant coliforms were de- mine either in shell eggs or in boiled eggs. According tected in 1 (4.1%) fertile liquid pasteurized egg sample. to Halász et al. (1994), tyramine is responsible for food poisoning, and its detection is therefore important in Bioactive Amines food safety programs. In the present study, low tyramine content was detected (9.1 mg/kg) in liquid pas- Of the 10 amines evaluated, none was detected in teurized egg. commercial liquid pasteurized egg. However, in fertile Although the levels of toxic bioactive amines have liquid pasteurized egg, low levels of the amines putres- not been clearly established yet, ten Brink et al. (1990) cine, cadaverine, and tyramine (Table 3) were detected consider levels from 100 to 800 mg/kg of tyramine in on d 14 and 21 of storage. The 3 amines found in fertile foods to be potentially hazardous to human health. The

Table 2. Means of total coliforms (most probable number/mL) in fertile and commercial liquid pasteurized egg stored under refrigeration for up to 21 d

Days of storage

Type of egg 1 7 14 21

Commercial 0a,x 0.06a,x 0a,x 0a,y Fertile 0.12b,x 0b,x 0b,x 80a,x a,bMeans followed by different letters in the same row are different (P < 0.05; Kruskal-Wallis test). x,yMeans followed by different letters in the same column are different (P < 0.05; Friedman test). BIOACTIVE AMINES AND MICROBIOLOGICAL QUALITY 1021

Table 3. Means of amine levels (mg/100 g) in whole fertile liquid pasteurized egg stored under refrigeration for 21 d

Days of storage Amine content (mg/kg) 1 7 14 21 Total Putrescine NDc NDc 1.6 ± 1.4b 15.3 ± 2.1a 16.9 Cadaverine NDb NDb NDb 2.9 ± 2.5a 2.9 Tyramine NDb NDb NDb 9.1 ± 12.1a 9.1 Total ND ND 1.6 27.3 28.9 a–cMeans followed by distinct letters differ by the Friedman test (P < 0.05). ND = not detected (amine content

<0.05 mg/kg). Downloaded from pharmacological effects of tyramine in the human body analyses, and the Laboratory of Food Biochemistry at are peripheral vasoconstriction, increased cardiac out- the Faculdade de Farmácia of the UFMG. put, salivation and lacrimation, increased respiratory rate, increased blood sugar levels, release of noradrena- http://ps.oxfordjournals.org/ line by the sympathetic nervous system, and migraine. REFERENCES Cadaverine poisoning causes hypotension, bradycardia, AOAC (Association of Official Analytical Chemists). 1990. Aero- exaggerated contraction of the jaw muscles, paralysis bic count plate, dry rehydratable film PetrifilmTM aerobic plate of the extremities, and potentiation of the toxicity of method. J. AOAC Int. 83:635. other amines (Shalaby, 1996). Aragon-Alegro, L. C., K. L. O. Souza, P. S. C. Sobrinho, M. Land- graf, and M. T. Destro. 2005. Avaliação da qualidade micro- The increased levels of the aforementioned 3 amines biológica de ovo integral pasteurizado produzido com e sem a with increasing storage period (P < 0.05), particularly etapa de lavagem no processamento [Influence of washing in the at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 in fertile eggs, confirm that bioactive amines are re- microbial quality of pasteurized egg]. Ciênc. Tecnol. Aliment. lated to egg deterioration, which was bacterially me- 25:618–622. Ball, J. H. R., M. P. Hamid-samimi, M. Foegeding, and K. R. diated. Emphasis should be placed on the possibility Swartzel. 1987. Functionality and microbial stability of ultrapas- of using the analysis of amine levels as a food quality teurized, aseptically packaged refrigerated whole egg. J. Food index. Sci. 52:1212–1218. Using the Pearson correlation to 5% probability, a Bardócz, S., T. J. Duguid, D. S. Brown, G. Grant, A. Pusztai, A. White, and A. Ralph. 1995. The importance of dietary poly- high positive correlation was found (0.84) between the amines in cell regeneration and growth. Br. J. Nutr. 73:819–828. MPN of total coliforms in fertile liquid pasteurized egg Figueiredo, T. C., R. P. Viegas, L. J. C. Lara, N. C. Baiao, M. and cadaverine levels, and a moderate positive correla- R. Souza, L. G. D. Heneine, and S. V. Cancado. 2013. Bioac- tion (0.72) was found between the counts of mesophilic tive amines and internal quality of commercial eggs. Poult. Sci. 92:1376–1384. aerobic microorganisms and putrescine levels. These Gloria, M. B. A. 2005. Amines. Pages 1–38 in Handbook of Food Sci- results are similar to those of Lima and Glória (1999) ence. H. Hui and L. L. Nollet, ed. Marcel Dekker, New York, NY. and Jay et al. (2005) who reported that cadaverine is Halász, A., A. Baráth, L. Simon-Sarkadi, and W. Holzapfel. 1994. Biogenic Amines and their production by microorganisms in produced mainly by members of the Enterobacteriaceae food. Trends Food Sci. Technol. 5:42–49. family. Hamid-Samimi, M. H., and K. R. Swartzel. 1985. Pasteurization The quality of refrigerated liquid pasteurized egg is design criteria for production of extended shelf-life refrigerated influenced by the type of raw material (commercial or liquid whole egg. J. Food Proc. Preserv. 8:219–224. Hara-Kudo, Y., and K. Takatori. 2009. Microbial quality of liquid fertile eggs) used to make it. Fertile eggs produce a egg and Salmonella infection status in Japan. Shokuhin Eisei- lower quality end product than commercial eggs. The gaku Zasshi 50:34–40. presence of biogenic amines in liquid pasteurized egg Hitchins, A. D., P. Feng, W. D. Watkins, S. R. Rippey, and L. A. prepared from fertile eggs reflects the microbiological Chandler. 2001. Escherichia coli and the Coliform bacteria. In Bacteriological Analytical Manual. Accessed Feb. 2010. http:// contamination detected in this product. www.fda.gov. Jay, J. M., M. J. Loessner, and D. A. Golden. 2005. Modern Food Microbiology. 7th ed. Springer, New York, NY. ACKNOWLEDGMENTS Lima, A. S., and M. B. A. Glória. 1999. Aminas bioativas em alimen- tos (Bioactive amines in foods). Bol. SBCTA 33:70–79. The authors acknowledge the assistance of the Col- Moreira, A. P. S., A. Giombeli, R. A. Labanca, D. L. Nelson, and legiate of Postgraduate Studies on Animal Science of M. B. A. Gloria. 2008. Effect of bioactive amines, microbial flora, the Escola de Veterinária da Universidade Federal de physico-chemical characteristics, and tenderness of broiler breast Minas Gerais (UFMG) and Pró-Reitoria de Pesquisa meat. Poult. Sci. 87:1868–1873. Nishimura, K., R. Shiina, K. Kashiwagi, and K. Igarashiy. 2006. at the UFMG for providing funds for researching and Decrease in polyamines with aging and their ingestion from food publication and the Conselho Nacional para Desen- and drink. J. Biochem. 139:81–90. volvimento Científico e Tecnológico for providing schol- Okamoto, A., E. Sugi, Y. Koizumi, F. Yanagida, and S. Udaka. 1997. Polyamine content of ordinary foodstuffs and various fer- arship. The authors also thank the Laboratory of Mi- mented foods. Biosci. Biotechnol. Biochem. 61:1582–1584. crobiological Safety of Food at the Instituto Mineiro Oliveira, G. E., T. C. Figueiredo, M. R. Souza, A. L. Oliveira, S. V. de Agropecuária for assisting with the microbiological Cançado, and M. B. A. Gloria. 2009. Bioactive amines and qual- 1022 RÊGO ET AL.

ity of egg from Dekalb hens under different storage conditions. ten Brink, B. T., C. Damink, H. M. Joosten, and J. H. Veld. 1990. Poult. Sci. 88:2428–2434. Occurrence and formation of biologically active amines in foods. Ramos, B., O. Pinho, and I. M. P. L. V. O. Ferreira. 2009. Changes Int. J. Food Microbiol. 11:73–84. of yolk biogenic amine concentrations during storage of shell hen USDA. 2008. Code of Federal Regulations. Title 9, Chapter 3, Vol. eggs. Food Chem. 116:340–344. 2, Sec. 590.570, Inspection of Eggs and Egg Products. Accessed Russell, D., and S. H. Snyder. 1968. Amine synthesis in develop- Feb. 2013. http://www.gpo.gov/fdsys/pkg/CFR-2008-title9- ing rapidly growing tissues: Ornithine decarboxylase activity in vol2/pdf/CFR-2008-title9-vol2-chapIII-subchapI.pdf. regenerating rat liver, chick embryo, and various tumors. Proc. Veciana-Nogués, M. C., A. Mariné-Font, and M. C. Vidal-Carou. Natl. Acad. Sci. USA 60:1420–1427. 1997. Biogenic amines as hygienic quality indicators of tuna. Saito, K., M. Horie, N. Nose, K. Nakagomi, and H. Nakazawa. 1992. Relationship with microbial counts, ATP-related compounds, Determination of polyamines in foods by liquid chromatography volatile amines, and organoleptic changes. J. Agric. Food Chem. with fluorescence on-column derivatization. Anal. Sci. 8:675– 45:2036–2041.

680. Downloaded from Shalaby, A. R. 1996. Significance of biogenic amines to food safety and human health. Food Res. Int. 29:675–690. http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Prevalence and antimicrobial resistance of Listeria, Salmonella, and Yersinia species isolates in ducks and geese

Hossein Jamali ,* Behrad Radmehr ,† and Salmah Ismail *1

* Biohealth Science Program, Institute of Biological Science, Faculty of Science, University of Malaya, 50603, Kuala Lumpur, Malaysia; and †Department of Food Hygiene and Quality Control, Faculty of Veterinary Medicine, Islamic Azad University, Karaj Branch, Karaj, Iran Downloaded from

ABSTRACT The aims of this study were to determine among Listeria, Salmonella, and Yersinia spp., respec- the prevalence and antimicrobial resistance of Liste- tively. Moreover, resistance to tetracycline was com- ria, Salmonella, and Yersinia spp. isolated from duck mon in Listeria (48.3%) and Salmonella spp. (63.6%), and goose intestinal contents. A total of 471 samples, whereas 51.3% of the Yersinia spp. isolates were resis- http://ps.oxfordjournals.org/ including 291 duck and 180 goose intestinal contents, tant to cephalothin. Therefore, continued surveillance were purchased from wet markets between November of the prevalence of the pathogens and also of emerging 2008 and July 2010. Listeria, Salmonella, and Yersinia antibiotic resistance is needed to render possible the spp. were isolated from 58 (12.3%), 107 (22.7%), and recognition of foods that may represent risks and also 80 (17%) of the samples, respectively. It was conclud- ensure the effective treatment of listeriosis, salmonel- ed that Listeria ivanovii, Salmonella Thompson, and losis, and yersiniosis.

Yersinia enterocolitica were the predominant serovars at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Key words: Listeria spp. , Salmonella spp. , Yersinia spp. , duck , goose 2014 Poultry Science 93 :1023–1030 http://dx.doi.org/ 10.3382/ps.2013-03699

INTRODUCTION ics (Castanon, 2007; Mathew et al., 2007). The growing problem of antimicrobial resistance among poultry- The public health concern of microbiological food associated pathogens may represent a risk of human safety is increasing worldwide. Listeria, Salmonella, infections with these organisms. and Yersinia spp. that cause foodborne diseases are To the best of our knowledge, there is no published the common significant pathogens in terms of food information regarding antibiotic susceptibility of Liste- safety and present a public health risk for consumers, ria, Salmonella, and Yersinia spp. from ducks and geese especially in developing countries (Soltan-Dallal et al., in Iran. The aim of this study was to investigate the 2010; Adzitey et al., 2012; Jamali et al., 2013). Some prevalence and antimicrobial resistance profiles of these studies have reported that animal origin foods are con- foodborne pathogens isolated from duck and goose in- sidered as vehicles related with infections caused by testinal contents in Varamin, Tehran province, Iran. Listeria, Salmonella, and Yersinia spp. (Cretikos et al., 2008; Jamali et al., 2013). Human antimicrobial treatment for listeriosis, salmo- MATERIALS AND METHODS nellosis, and yersiniosis in most cases is not indicated. Although, antibiotic therapy is necessary in some criti- Sample Collection cal cases such as infections of immune-compromised A total of 471 samples, including intestinal contents persons and extraintestinal disease, where antimicro- of ducks (n = 291) and geese (n = 180), were randomly bial resistance of the pathogens could complicate the collected from different wet markets located in Vara- treatment (Engberg et al., 2004). The use of antimicro- min, Tehran province, Iran, between November 2008 bials in animal foods for the growth promotion, control, and July 2010. The samples were transferred into ster- or treatment of poultry diseases has been increased. ile plastic bags and transported in an ice box to the However, the overuse of these antibiotics may increase laboratory within 3 h and analyzed immediately. the rates of antimicrobial resistance to several antibiot- Isolation and Identification of Isolates © 2014 Poultry Science Association Inc. Received October 18, 2013. Accepted December 16, 2013. Listeria spp. In this study, the isolation and detec- 1 Corresponding author: [email protected] tion of Listeria spp. was done by the USDA method

1023 1024 JAMALI ET AL. (McClain and Lee, 1988). Briefly, 25 g of each sample (30 μg), gentamicin (10 μg), chloramphenicol (30 μg), was added to 225 mL of Listeria enrichment broth (Ox- trimethoprim (15 μg), ampicillin (30 μg), amoxicillin oid, Basingstoke, UK) as first enrichment broth, mixed (30 μg; for all isolates); vancomycin (30 μg), rifampi- for 2 min, and incubated at 37°C for 24 h. Then, 1 mL cin (5 μg), penicillin G (10 unit), kanamycin (30 μg), of Listeria enrichment broth was added to 9 mL of erythromycin (15 μg), clindamycin (2 μg), amoxicillin- Fraser broth (Oxoid) as a second enrichment culture clavulanic acid (20/10 μg; for Listeria spp.), streptomy- and incubated for 24 h at 37°C. Then, enriched Fraser cin (30 μg), nalidixic acid (30 μg), ciprofloxacin (5 μg), broth-culture was streaked onto Palcam agar (Oxoid) cephalothin (30 μg; for Yersinia spp.), and streptomy- and Oxford agar (Oxoid) and incubated at 37°C for 24 cin (30 μg), nalidixic acid (30 μg), ciprofloxacin (5 μg), to 48 h. Presumptive Listeria colonies were confirmed and colisitin (10 μg; for Salmonella spp.). by using the biochemical tests (Aygun and Pehlivanlar, Downloaded from 2006) and API Listeria (bioMérieux). Statistical Analysis Salmonella spp. The standard conventional culture method was used in the isolation of Salmonella spp. in The relationship between the contaminated samples this study as described by Varnam and Evans (1991). and the different kinds of samples was analyzed us- Briefly, 25 g of each sample was added to 225 mL of ing chi-squared analysis. All statistical and chi-squared http://ps.oxfordjournals.org/ buffered peptone water (Merck, Darmstadt, Germany), analyses were performed using SPSS 18.0 (SPSS Inc., mixed for 2 min and incubated for 24 h at 37°C. Then, Chicago, IL). A P-value < 0.05 was used for statistical 0.1 mL of the broth was added to 10 mL of selenite significance. cystine broth (Merck) and was incubated at 42°C overnight. Finally, enriched selenite cystine broth-culture RESULTS was streaked onto Salmonella Shigella agar (Merck) and

brilliant green agar (Merck). All plates were then incu- The obtained results showed that 58 (12.3%), 107 at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 bated at 37°C for overnight. Presumptive Salmonella (22.7%), and 80 (17%) of duck and goose intestinal colonies were confirmed by using API 20E (bioMérieux contents samples were contaminated with Listeria, Sal- 20100, Marcy l’Etoile, France). Salmonella polyvalent monella, and Yersinia spp., respectively. O and H antisera (Mast Diagnostics, Merseyside, UK) were used for agglutination tests. Listeria spp. Yersinia spp. Twenty-five gram of each sample was added to 225 mL of PBS. The samples were mixed Prevalence of Listeria spp. in Ducks and Geese. for 2 min and incubated for 3 wk at 4°C. Then, the In this study, 58 (12.3%) samples were contaminated cold-enriched samples were subjected to alkali treat- with Listeria spp., and among them, 41 (14.1%) and 17 ment by adding 0.5 mL of 0.5% KOH into 4.5 mL of (9.4%) duck and goose samples were contaminated with cold-enriched samples. The cold-enriched and alkali- Listeria spp., respectively (Table 1). There was a sig- treated cultures were streaked onto cefsulodin-Irgasan- nificant difference between contaminated duck samples novobiocin agar and incubated for 18 to 24 h at 25°C. and goose samples (P < 0.05). The highest prevalence Presumptive Yersinia colonies were confirmed by using of Listeria spp. was Listeria ivanovii (43.1%) followed the API 20E (bioMérieux 20100). by Listeria monocytogenes (32.8%), Listeria innocua (8.6%), and Listeria seeligeri (15.5%). Antimicrobial Susceptibility Test Antimicrobial Resistance of Listeria spp. Iso- lates. The resistance profiles of Listeria spp. to 10 an- All Listeria, Salmonella, and Yersinia spp. were sub- timicrobial agents tested in this study are shown in jected to antimicrobial susceptibility tests. The Kirby- Table 2. Thirty-two (55.2%) and 8 (13.8%) out of 58 Bauer disc diffusion method using Mueller Hinton agar isolates of Listeria spp. were resistant to 1 and 2 an- (Oxoid) was applied to antimicrobial susceptibility test tibiotic agents, respectively. However, 3 Listeria spp. according to the Clinical and Laboratory Standards isolates (5.2%) showed multidrug resistance (MDR; Institute (CLSI, 2006). However, Muller-Hinton agar resistance to >2 antimicrobial). Resistance to tetra- supplemented with 5% defibrinated sheep blood was cycline (48.3%) was the most common finding in this used for Listeria spp. The following panel of antimicro- study, followed by resistance to penicillin G (24.1%), bial agents and concentrations was applied: tetracycline erythromycin (12.1%), clindamycin (10.3%), and amox-

Table 1. Prevalence (% in parentheses) of Listeria spp. in duck and goose samples

Listeria Listeria Listeria Listeria Listeria Samples Total spp. monocytogenes ivanovii innocua seeligeri Duck 291 41 (14.1) 14 (4.8) 16 (5.5) 4 (1.4) 7 (2.4) Goose 180 17 (9.4) 5 (2.8) 9 (5) 1 (0.6) 2 (1.1) Total 471 58 (12.3) 19 (4) 25 (5.3) 5 (1.1) 9 (1.9) Table 2. Number and percentages of antimicrobial resistance of Listeria spp. isolated from duck and goose samples1

Samples Pen G AMC CL E GN TET VA R CD K R to 1 R to 2 R to >2 Item (No.) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) a.m.2 (%) a.m.3 (%) a.m.4 (%) Listeria spp. Duck (41) 11 (26.8) 4 (9.8) 0 4 (9.8) 0 20 (48.8) 0 0 5 (12.2) 0 23 (56.1) 5 (12.2) 3 (7.3) Goose (17) 3 (17.6) 0 0 3 (17.6) 0 8 (47.1) 0 0 1 (5.9) 0 9 (52.9) 3 (17.6) 0 Total (58) 14 (24.1) 4 (6.9) 0 7 (12.1) 0 28 (48.3) 0 0 6 (10.3) 0 32 (552) 8 (13.8) 3 (5.2) Listeria monocytogenes Duck (14) 4 (28.6) 1 (7.1) 0 1 (7.1) 0 8 (57.1) 0 0 2 (14.3) 0 7 (50) 3 (21.4) 1 (7.1) Goose (5) 2 (40) 0 0 1 (20) 0 3 (60) 0 0 0 0 4 (80) 1 (20) 0

Total (19) 6 (31.6) 1 (5.3) 0 2 (10.5) 0 11 (57.9) 0 0 2 (10.5) 0 11 (57.9) 4 (21.1) 1 (5.3) LISTERIA, SALMONELLA, ANDYERSINIAINDUCKSGEESE Listeria seeligeri Duck (7) 1 (14.3) 0 0 0 0 2 (28.6) 0 0 0 0 3 (42.9) 0 0 Goose (2) 0 0 0 0 0 0 0 0 0 0 0 0 0 Total (9) 1 (11.1) 0 0 0 0 2 (22.2) 0 0 0 0 3 (33.3) 0 0 Listeria innocua Duck (4) 0 0 0 0 0 1 (25) 0 0 0 0 1 (25) 0 0 Goose (1) 0 0 0 0 0 0 0 0 0 0 0 0 0 Total (5) 0 0 0 0 0 1 (20) 0 0 0 0 1 (20) 0 0 Listeria ivanovii Duck (16) 6 (37.5) 3 (18.8) 0 3 (18.8) 0 9 (56.3) 0 0 3 (18.8) 0 12 (75) 2 (12.5) 2 (12.5) Goose (9) 1 (11.1) 0 0 2 (22.2) 0 5 (55.6) 0 0 1 (11.1) 0 5 (55.6) 2 (22.2) 0 Total (25) 7 (28) 3 (12) 0 5 (20) 0 14 (56) 0 0 4 (16) 0 17 (68) 4 (16) 2 (8) 1Pen G: penicillin G; AMC: amoxicillin-clavulanic acid; CL: chloramphenicol; E: erythromycin; GN: gentamicin; TET: tetracycline; VA: vancomycin; R: rifampicin; CD: clindamycin; K: kanamycin. 2R to 1 a.m.: resistance to 1 antimicrobial. 3R to 2 a.m.: resistance to 2 antimicrobials. 4R to >2 a.m.: resistance to >2 antimicrobials.

Table 3. Number and percentages of isolates of serotypes of Salmonella from ducks and geese

No. of Salmonella Salmonella Salmonella Salmonella Salmonella Salmonella Source samples Thompson Paratyphi C Enteritidis Hadar Virginia Typhimurium Total Duck 291 56 (59.6%) 6 (6.4%) 8 (8.5%) 7 (7.4%) 5 (5.3%) 2 (6.5%) 84 (28.9%) Goose 180 14 (60.9%) 3 (13%) 2 (8.7%) 1 (4.3%) 3 (13%) 0 23 (12.8%) Total 471 70 (65.4%) 9 (8.4%) 10 (9.3%) 8 (7.5%) 8 (7.5%) 2 (1.9%) 107 (22.7%)

1025

Table 4. Number and percentages of antimicrobial resistance of serotypes of Salmonella isolated from ducks and geese1

Samples AMO AMP NAL TMP CL TET GN CIP STR COL R to 1 R to 2 R to >2 Item (No.) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) a.m.2 (%) a.m.3 (%) a.m.4 (%)

Salmonella spp. Duck (84) 6 (7.1) 3 (3.6) 33 (39.3) 24 (28.6) 5 (6) 54 (64.3) 0 0 22 (26.2) 0 18 (21.4) 14 (16.7) 26 (31) Goose (23) 1 (4.3) 1 (4.3) 10 (43.5) 6 (26.1) 0 14 (60.9) 0 0 10 (43.5) 0 7 (30.4) 1 (4.3) 9 (39.1) Total (107) 7 (6.5) 4 (3.7) 43 (40.2) 30 (28) 5 (4.7) 68 (63.6) 0 0 32 (29.9) 0 25 (23.4) 15 (14) 35 (32.7) Salmonella Duck (56) 4 (7.1) 2 (3.4) 21 (37.5) 18 (32.1) 4 (7.1) 39 (69.6) 0 0 19 (34) 0 10 (17.9) 8 (14.3) 21 (37.5) Thompson Goose (14) 0 1 (7.1) 8 (57.1) 6 (42.9) 0 11 (78.6) 0 0 9 (64.3) 0 3 (21.4) 1 (7.1) 8 (57.1) Total (70) 4 (5.7) 3 (4.3) 29 (41.4) 24 (34.3) 4 (5.7) 50 (71.4) 0 0 28 (40) 0 13 (18.6) 9 (12.9) 29 (41.4) Salmonella Duck (6) 0 1 (16.7) 2 (33.3) 2 (33.3) 0 4 (66.7) 0 0 3 (50) 0 1 (16.7) 1 (16.7) 2 (33.3) JAMALI ETAL. Paratyphi C Goose (3) 0 0 1 (33.3) 0 0 1 (33.3) 0 0 1 (33.3) 0 3 (100) 0 0 Total (9) 0 1 (11.1) 3 (33.3) 2 (22.2) 0 5 (55.6) 0 0 4 (44.4) 0 4 (44.4) 1 (11.1) 2 (22.2) Salmonella Duck (8) 1 (12.5) 0 1 (12.5) 0 1 (12.5) 3 (37.5) 0 0 0 0 2 (25) 0 1 (12.5) Enteritidis Goose (2) 0 0 0 0 0 0 0 0 0 0 0 0 0 Total (10) 1 (10) 0 1 (10) 0 1 (10) 3 (30) 0 0 0 0 2 (20) 0 1 (10) Salmonella Duck (7) 0 0 4 (57.1) 2 (28.6) 0 3 (42.9) 0 0 0 0 3 (42.9) 3 (42.9) 0 Hadar Goose (1) 0 0 0 0 0 1 (100) 0 0 0 0 1 (100) 0 0 Total (8) 0 0 4 (50) 2 (25) 0 4 (50) 0 0 0 0 4 (50) 3 (37.5) 0 Salmonella Duck (5) 1 (20) 0 3 (60) 2 (40) 0 4 (80) 0 0 0 0 1 (20) 1 (20) 2 (40) Virginia Goose (3) 1 (33.3) 0 1 (33.3) 0 0 1 (33.3) 0 0 0 0 0 0 1 (33.3) Total (8) 2 (25) 0 4 (50) 2 (25) 0 5 (62.5) 0 0 0 0 1 (12.5) 1 (12.5) 3 (37.5) Salmonella Duck (2) 0 0 2 (100) 0 0 1 (50) 0 0 0 0 1 (50) 1 (50) 0 Typhimurium Goose (0) 0 0 0 0 0 0 0 0 0 0 0 0 0 1AMO: amoxicillin, AMP: ampicillin, NAL: nalidixic acid, TMP: trimethoprim, CL: chloramphenicol, TET: tetracycline, GN: gentamicin, CIP: ciprofloxacin, STR: streptomycin, COL: colisitin. 2R to 1 a.m.: resistance to 1 antimicrobial. 3R to 2 a.m.: resistance to 2 antimicrobials.

4R to >2 a.m.: resistance to >2 antimicrobials.

Downloaded from from Downloaded http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 6, April on University Harvard Zoology, Comp Museum the of Library Mayr Ernst at LISTERIA, SALMONELLA, AND YERSINIA IN DUCKS AND GEESE 1027 icillin-clavulanic acid (6.9%). All the isolates of Listeria intermedia isolates, were MDR. Moreover, 21 (36.3%) spp. were susceptible to half of the antibiotic examined of the Yersinia spp. isolates were sensitive to all tested including chloramphenicol, gentamicin, vancomycin, ri- antibiotics. fampicin, and kanamycin. DISCUSSION Salmonella spp. Our results on the prevalence of Listeria, Salmonella, Distribution of Salmonella Serovars in Ducks and and Yersinia spp. demonstrated a high level of con- Geese. Out of 107 Salmonella spp., 84 (78.5%) and tamination among purchased duck and goose samples 23 (21.5%) isolates were detected from duck and goose in Varamin, Tehran province, Iran. Our findings are in samples, respectively (Table 3). There was a significant agreement with earlier findings, whereby Chipilev et al. Downloaded from difference between contaminated duck and goose sam- (2010) and Adzitey et al. (2013) isolated L. monocyto- ples (P < 0.05). The most common serovar was Salmo- genes and other species of Listeria from duck products nella Thompson (65.4%). The remaining isolates were and intestinal contents, respectively. However, no Lis- Salmonella Enteritidis (9.3%), Salmonella Paratyphi C teria spp. was recovered from healthy indigenous ducks (8.4%), Salmonella Hadar/Salmonella Virginia (7.5%), (Njagi et al., 2004). It is not clear whether ducks and http://ps.oxfordjournals.org/ and Salmonella Typhimurium (1.9%). geese are Listeria primary reservoirs or whether they Antimicrobial Resistance of Salmonella Se- were contaminated by environments. Several resources rovars. The resistance profiles for the 107 Salmonella such as feces, soil, vegetation, and sewage have been re- spp. are as follows: tetracycline, 63.6%; nalidixic acid, ported as primary reservoir for Listeria species. Listeria 40.2%; streptomycin, 29.9%; trimethoprim, 28%; amox- spp. can survive and be transmitted to farm equipment icillin, 6.5%; chloramphenicol, 4.7%; and ampicillin, and future flocks. Isolation of L. monocytogenes in the

3.7%. However, all the Salmonella isolates were sensi- current study suggests that ducks and geese could be a at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 tive to colisitin, ciprofloxacin, and gentamicin (Table potential risk of foodborne listeriosis in customers. 4). Thirty-five (32.7%) isolates of Salmonella spp. were The prevalence of Salmonella spp. in duck samples is MDR. Multidrug resistance was more frequently ob- in agreement with earlier findings in Iran (Dilmaghani served among Salmonella Thompson as well as duck et al., 2011), the United Kingdom (Little et al., 2008), isolates. In addition, 42 (39.3%) Salmonella spp. iso- and Malaysia (Adzitey et al., 2012). Moreover, our relates were susceptible to all of the examined antibiotics. sults showed the high level of Salmonella contamination among ducks in Iran compared with the previous reports Yersinia spp. in other countries such as the United States (McCrea et al., 2006), China (Pan et al., 2010), Taiwan (Yu et Prevalence of Yersinia spp. in Ducks and Geese. al., 2008), and Egypt (Osman et al., 2010). Salmonella Of 80 Yersinia spp. isolates, 58 (72.5%) and 22 (27.5%) prevalence in ducks from Brazil was greater than our isolates were detected from duck and goose samples, re- study (Hofer et al., 1997; Ribeiro et al., 2004). The high spectively (Table 5). The isolates of Yersinia spp. were prevalence of Salmonella spp. in ducks could depend on identified as Y. enterocolitica (9.6%), Y. frederiksenii environmental exposure and contact with other animals (4.5%), and Y. intermedia (4%). There was a signifi- in the rearing area. However, a few reports regarding cant difference between contaminated duck and goose Salmonella prevalence in goose have been published. samples (P < 0.05). The prevalence of Salmonella spp. (12.8%) in our study Antimicrobial Resistance of Yersinia spp. Iso- was less than previously reported (32.6%) in Iran (Dil- lates. The resistance profiles for the 80 Yersinia spp. maghani et al., 2011). Trawinska et al. (2008) and Pan are as follows: cephalothin, 51.3%; ampicillin, 21.3%; et al. (2010) have isolated Salmonella spp. in goose trimethoprim, 17.5%; amoxicillin, 11.3%; ciprofloxacin, samples in Poland and China, respectively. 10%; nalidixic acid, 7.5%, streptomycin, 6.3%, and tet- Salmonella Thompson was the predominant serovar racycline, 2.9% each (see Table 6). All the isolates were in this study as well as previously described in Iran sensitive to gentamicin and chloramphenicol. Thirteen (Soltan-Dallal et al., 2010). It is also considered as (16.3%) isolates of Yersinia spp., including 8 Yersinia one of the most common serovars that causes infec- enterocolitica, 3 Yersinia frederiksenii, and 2 Yersinia tion and outbreaks. Two clinically significant Salmonel-

Table 5. Number and percentages (in parentheses) of isolates of serotypes of Yersinia from duck and goose samples

No. of Yersinia Yersinia Yersinia Source samples enterocolitica fredreksenni intermediate Total Duck 291 32 (11) 15 (5.2) 11 (3.8) 58 (19.9) Goose 180 13 (7.2) 6 (3.3) 3 (1.7) 22 (12.2) Total 471 45 (9.6) 21 (4.5) 14 (3) 80 (17) 1028 JAMALI ET AL. la serovars, Salmonella Typhimurium and Salmonella Enteritidis, have been frequently reported as predomi- (%) 4 nant serovars in many countries (Bennasar et al., 2000; R to >2 a.m. Little et al., 2008; Trawinska et al., 2008; Pan et al., 2010; Osman et al., 2010; Adzitey et al., 2012). How- ever, these serovars were not found as the predominant (%) 3 serovars in our study among duck and goose samples. R to 2

a.m. The prevalence of Yersinia spp. in foods and animals and its significance is still unknown in Iran (Hanifian and Khani, 2012). In this study, 9.6% of duck and goose (%) Downloaded from 2 samples were contaminated by Yersinia spp. To the R to 1 1 (33.3) 1 (33.3) 0

a.m. best of our knowledge, there is no published information regarding to prevalence of Yersinia spp. in ducks and geese in Iran. However, a few studies have been

(%) reported the prevalence of Yersinia spp. in chicken and CEF 6 (46.2) 1 (7.7) 3 (23.1) 2 (15.4) http://ps.oxfordjournals.org/ 2 (33.3) 4 (66.7) 2 (33.3) 0 1 (9.1) 4 (36.4) 1 (9.1) 0 beef with the rate ranged of 13 to 16% (Soltan-Dallal et al., 2010; Yazdi et al., 2011). Yersinia enterocolitica

was the predominant species in this study and is in 1 (%) STR agreement with earlier findings (Jiang and Kang, 2000; Capita et al., 2002; Soltan-Dallal et al., 2010; Yazdi et al., 2011; Hanifian and Khani, 2012). Antimicrobial resistance in pathogens and therapeu- (%) CIP

tic interference is always a significant issue in public at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 health. The investigation of susceptibility pattern and 0 0 1 (7.1) 1 (7.1) 5 (35.7) 2 (14.3) 0 0 0 0 0 0 1 (33.3) 0 antibiotic resistance is important to treatment (Soltan- GN (%) Dallal et al., 2010). Antimicrobial agents that are used in growth promotion, treatment, or prophylaxis in poultry and broiler rearing have the potential to pro- (%) TET mote antimicrobial resistance among pathogens present. Therefore, antibiotic resistance strains could be transferred to humans via contaminated poultry prod- CL (%) ucts (Mayrhofer et al., 2004; Ezekiel et al., 2011). Although only 3 isolates of Listeria spp. showed MDR in this study, more than half (55.2%) of all Listeria spp. (%)

TMP isolates were resistant to at least one antibiotic agent. 1 (6.7) 0 3 (20) 0 2 (13.3) 0 11 (73.3) 7 (46.7) 1 (6.7) 3 (20) 2 (33.3) 0 3 (50) 0 0 0 3 (14.3) 0 6 (28.6) 0 2 (9.5) 0 13 (61.9) 11 (52.4) 3 (14.3) 3 (14.3) 2 (18.2) 0 0 3 (21.4) 0 0 1 (33.3) 0 0 The MDR isolates were divided into L. ivanovii (66.7%) and L. monocytogenes (33.3%), and the susceptibility of L. monocytogenes isolates were observed in gentamicin, (%) NAL chloramphenicol, vancomycin, kanamycin, and rifampicin, which are in agreement with earlier findings (Hof and Emmerling, 1984; Srinivasan et al., 2005; Arslan and Özdemir, 2008; Conter et al., 2009). Furthermore, (%) AMP 1 (16.7) 0 1 (33.3) 0 in an earlier study, susceptibility of food L. monocytogenes isolates to vancomycin and gentamicin also was reported (Harakeh et al., 2009).

(%) In the current study, all Salmonella spp. isolates were AMO sensitive to ciprofloxacin, which is in agreement with previous studies in Iran and other countries (Mayrhofer et al., 2004; Bhatia et al., 2007; Soltan-Dallal et al., 2010). Resistance to tetracycline and cephalothin was greater in Salmonella and Yersinia spp., respectively. A Duck (58)Duck 7 (12.1) 13 (22.4) 5 (8.6) 9 (15.5) 0 15 (25.9) 0 7 (12.1) 4 (6.9) 33 (56.9) 20 (34.5) 7 (12.1) 10 (17.2) Samples (No.) Goose (22) 2 (9.1) 4 (18.2) 1 (4.5) 5 (22.7) 0 8 (36.4) 0 1 (4.5) 1 (4.5) 8 (36.4) 6 (27.3) 6 (27.3) 2 (9.1) Total (80)Total 9 (11.3) 17 (21.3) 6 (7.5) 14 (17.5) 0 23 (28.8) 0 8 (10) 5 (6.3) 41 (51.3) 26 (32.5) 13 (16.3) 12 (15) Duck (32)Duck 5 (15.6) 7 (21.9) 5 (15.6) 6 (18.8) 0 (37.5) 12 0 5 (15.6) 4 (12.5) 21 (65.6) 9 (28.1) 5 (15.6) 7 (21.9) Goose (13) 2 (15.4) 2 (15.4) 1 (7.7) 2 (15.4) 0 5 (15.6) 0 1 (7.7) 0 Total (45)Total 7 (15.6) 9 (20) 6 (13.3) 8 (17.8) 0 (37.8) 17 0 6 (13.3) 4 (8.9) 27 (60) 10 (22.2) 8 (17.8) 9 (20) Duck (15)Duck 1 (6.7) 4 (26.7) 0 Goose (6) 0 Total (21)Total 1 (4.8) 5 (23.8) 0 Duck (11)Duck 1 (9.1) 2 (18.2) 0 Goose (3) 0 Total (14)Total 1 (7.1) 3 (21.4) 0 high prevalence of resistance to tetracycline in Salmo- nella was found by Yildirim et al. (2011) and Thai et al. (2012), whereas in an earlier study a low number of the pathogen showed resistance to tetracycline (Cook et al., 2011). Furthermore, a high prevalence of resistance to cephalothin in Yersinia spp. was reported by Bhaduri

AMO: amoxicillin, AMP: ampicillin, NAL: nalidixic acid, TMP: trimethoprim, CL: chloramphenicol, TET: tetracycline, GN: gentamicin, CIP: ciprofloxacin, STR: streptomycin, CEF: cephalothin. STR: streptomycin, CIP: ciprofloxacin, GN: gentamicin, TET: tetracycline, AMP: ampicillin, NAL: nalidixic acid, TMP: trimethoprim, CL: chloramphenicol, AMO: amoxicillin, R to 1 a.m.: resistance antimicrobial. R to 2 a.m.: resistance antimicrobials. R to >2 a.m.: resistance antimicrobials. and Wesley (2012) and Bolton et al. (2013). Ciprofloxa- 1 2 3 4 isolated from duck and goose samples isolated from duck of Yersinia resistance of serotypes of antimicrobial and percentages 6. Number Table Item Yersinia spp. Yersinia Y. enterocolitica Y. fredreksenni Y. intermediate cin and tetracycline are used as disease treatment and LISTERIA, SALMONELLA, AND YERSINIA IN DUCKS AND GEESE 1029 prevention, and a growth promoter in poultry rearing (M45-A). Clinical and Laboratory Standards Institute, Wayne, in Iran (Soltan-Dallal et al., 2010). Therefore, antimi- PA. Conter, M., D. Paludi, E. Zanardi, S. Ghidini, A. Vergara, and crobial substances that are used for animal treatment A. Ianieri. 2009. Characterization of antimicrobial resistance could be reflected by antimicrobial resistance profile in of foodborne Listeria monocytogenes. Int. J. Food Microbiol. foodborne pathogens. 128:497–500. In summary, the presence of Listeria, Salmonella, Cook, A., R. J. Reid-Smith, R. J. Irwin, S. A. McEwen, V. Young, and C. Ribble. 2011. Antimicrobial resistance in Campylobacter, and Yersinia spp. in raw duck and goose intestinal Salmonella, and Escherichia coli isolated from retail grain- contents showed that consumption of these kinds of fed veal meat from southern Ontario, Canada. J. Food Prot. meats could be a potential risk of foodborne listeriosis, 74:1245–1251. salmonellosis, and yersiniosis in consumers. Thus, to Cretikos, M., B. Telfer, and J. McAnulty. 2008. Enteric disease out-

break reporting, New South Wales, Australia, 2000 to 2005. N. Downloaded from decrease the prevalence of developed pathogen resis- S. W. Public Health Bull. 19:3–7. tance, resistance monitoring and educational programs Dilmaghani, M., M. Ahmadi, Z. T. Salahi, and A. Talebi. 2011. De- are required. Moreover, continual surveillance on the tection of Salmonella enterica serovar Typhimurium from avians using multiplex-PCR. Vet. Res. Forum 2:157–165. prevalence of the pathogens as well as on emerging an- Engberg, J., J. Neimann, E. M. Nielsen, F. M. Aarestrup, and V. timicrobial resistance is needed to recognize foods that Fussing. 2004. Quinolone-resistant Campylobacter infections in may represent risks and ensure the effective treatment Denmark: Risk factors and clinical consequences. Emerg. Infect. http://ps.oxfordjournals.org/ of foodborne infections. Dis. 10:1056–1063. Ezekiel, C. N., A. O. Olarinmoye, J. M. A. Oyinloye, O. B. Olaoye, and A. O. Edun. 2011. Distribution, antibiogram and multidrug ACKNOWLEDGMENTS resistance in Enterobacteriaceae from commercial poultry feeds in Nigeria. Afr. J. Microbiol. Res. 5:294–301. This study was supported by OCAR chancellery of Hanifian, S., and S. Khani. 2012. Prevalence of virulent Yersinia enterocolitica in bulk raw milk and retail cheese in northern-west University of Malaya with grant number A-21010-DA674 of Iran. Int. J. Food Microbiol. 155:89–92. and A-21010-DA677. Harakeh, S., I. Saleh, O. Zouhairi, E. Baydoun, E. Barbour, and N. at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Alwan. 2009. Antimicrobial resistance of Listeria monocytogenes isolated from dairy-based food products. Sci. Total Environ. REFERENCES 407:4022–4027. Hof, H., and P. Emmerling. 1984. Murine model for therapy of lis- Adzitey, F., G. R. Rahmat Ali, N. Huda, T. Cogan, and J. Corry. teriosis in the compromised host. Chemotherapy 30:125–130. 2013. Prevalence, antibiotic resistance and genetic diversity of Hofer, E., S. J. Silva Filho, and E. M. F. Reis. 1997. Prevalência de Listeria monocytogenes isolated from ducks, their rearing and sorovares de Salmonella isolados de aves no Brasil. Pesqui. Vet. processing environments in Penang, Malaysia. Food Contr. Bras. 17:55–62. (In Portuguese.) 32:607–614. Jamali, H., B. Radmehr, and K. L. Thong. 2013. Prevalence, char- Adzitey, F., G. Rusul, and N. Huda. 2012. Prevalence and antibi- acterisation, and antimicrobial resistance of Listeria species and otic resistance of Salmonella serovars in ducks, duck rearing and Listeria monocytogenes isolates from raw milk in farm bulk tanks. processing environments in Penang, Malaysia. Food Res. Int. Food Contr. 34:121–125. 45:947–952. Jiang, G. C., and D. H. Kang. 2000. Enrichment procedures and Arslan, S., and F. Özdemir. 2008. Prevalence and antimicrobial re- plating media for isolation of Yersinia enterocolitica. J. Food sistance of Listeria spp. in homemade white cheese. Food Contr. Prot. 63:1483–1486. 19:360–363. Little, C. L., J. F. Richardson, R. J. Owen, E. de Pinna, and E. Aygun, O., and S. Pehlivanlar. 2006. Listeria spp. in the raw milk J. Threlfall. 2008. Prevalence, characterisation and antimicrobial and dairy products in Antakya, Turkey. Food Contr. 17:676–679. resistance of Campylobacter and Salmonella in raw poultry meat Bennasar, A., G. de Luna, B. Cabrer, and J. Lalucat. 2000. Rapid in the UK, 2003–2005. Int. J. Environ. Health Res. 18:403–414. identification of Salmonella Typhimurium, S. Enteritidis, and S. Mathew, A. G., R. Cissell, and S. Liamthong. 2007. Antibiotic resis- Virchow isolates by polymerase chain reaction based fingerprint- tance in bacteria associated with food animals: A United States ing methods. Int. Microbiol. 3:31–38. perspective of livestock production. Foodborne Pathog. Dis. Bhaduri, S., and I. V. Wesley. 2012. Prevalence, serotype, virulence 4:115–133. characteristics, clonality, and antibiotic susceptibility of patho- Mayrhofer, S., P. Paulsen, F. J. M. Smulders, and H. Friederike. genic Yersinia enterocolitica from swine feces. Adv. Exp. Med. 2004. Antimicrobial resistance profile of five major food-borne Biol. 954:111–116. pathogens isolated from beef, pork and poultry. Int. J. Food Bhatia, L. C. J., C. A. Mathur, and C. M. Arora. 2007. Reemergence Microbiol. 97:23–29. of chloramphenicol sensitivity in enteric fever. Med. J. Armed McClain, D., and W. H. Lee. 1988. Development of USDA-FSIS Forces India 63:212–214. method for isolation of Listeria monocytogenes from raw meat Bolton, D. J., C. Ivory, and D. McDowell. 2013. A small study of and poultry. J. Assoc. Off. Anal. Chem. 71:660–664. Yersinia enterocolitica in pigs from birth to carcass and charac- McCrea, B. A., K. H. Tonooka, C. VanWorth, C. L. Boggs, E. R. terisation of porcine and human strains. Food Contr. 33:521– Atwill, and J. S. Schrader. 2006. Prevalence of Campylobacter 524. and Salmonella species on farm, after transport, and at process- Capita, R., C. A. Calleja, M. Prieto, M. C. Garcia-Fernandez, and ing in specialty market poultry. Poult. Sci. 85:136–143. B. Moreno. 2002. Incidence and pathogenicity of Yersinia spp. Njagi, L. W., P. G. Mbuthia, L. C. Bebora, P. N. Nyaga, U. Minga, isolates from poultry in Spain. Food Microbiol. 19:295–301. and J. E. Olsen. 2004. Carrier status for Listeria monocytogenes Castanon, J. I. 2007. History of the use of antibiotic as growth pro- and other Listeria species in free range farm and market healthy moters in European poultry feeds. Poult. Sci. 86:2466–2471. indigenous chickens and ducks. East Afr. Med. J. 81:529–533. Chipilev, N., H. Daskalov, and T. Stoyanchev. 2010. Distribution Osman, K. M., M. M. Y. Ashgan, M. M. Aly, and M. I. Radwan. of Listeria spp. in duck breast and liver during gavaging, plant 2010. Salmonella spp. infection in imported 1-day-old chicks, processing and vacuum-packing. Bulg. J. Vet. Med. 13:87–91. ducklings, and turkey poults: A public health risk. Foodborne Clinical and Laboratory Standards Institute (CLSI). 2006. Meth- Pathog. Dis. 7:383–390. ods for antimicrobial dilution and disk susceptibility testing of Pan, Z. M., S. Z. Geng, Y. Q. Zhou, Z. Y. Liu, Q. Fang, B. B. Liu, infrequently isolated or fastidious bacteria; Approved Guideline and X. A. Jiao. 2010. Prevalence and antimicrobial resistance of 1030 JAMALI ET AL.

Salmonella sp. isolated from domestic animals in Eastern China. Trawinska, B., L. Saba, L. Wdowiak, O. Ondrasovicova, and B. J. Anim. Vet. Adv. 9:2290–2294. Nowakowicz-Debek. 2008. Evaluation of Salmonella rod incidence Ribeiro, S. A. M., M. C. M. Galleti, M. A. Orsi, A. R. Ferrati, and in poultry in the Lublin Province over the years 2001–2005. Ann. A. O. Mendonc.¸Doretto Junior, J. L., S. C. A. Camillo, and Agric. Environ. Med. 15:131–134. D. Reischak. 2004. Incidence of Salmonella in imported day-old Varnam, A. H., and M. G. Evans. 1991. Salmonella. Pages 71–76 in ducklings. Brazil, 1998–2003. Braz. J. Poult. Sci. 8:39–43. Foodborne Pathogens. An Illustrated Text. A. H. Varnam, and Soltan-Dallal, M. M., P. M. Doyle, M. Rezadehbashi, H. Dabiri, M. M. G. Evans, ed. Wolfe Publishing Ltd., London, UK. Sanaei, S. Modarresi, R. Bakhtiari, K. Sharifiy, M. Taremi, M. R. Yazdi, S. M. K., M. M. Soltan-Dallal, M. R. Zali, S. Avadisians, Zali, and M. K. Sharifi-Yazdi. 2010. Prevalence and antimicrobial and R. Bakhtiari. 2011. Incidence and antibiotic susceptibilities resistance profiles of Salmonella serotypes, Campylobacter and of Yersinia enterocolitica and other Yersinia species recovered Yersinia spp. isolated from retail chicken and beef, Tehran, Iran. from meat and chicken in Tehran, Iran. Afr. J. Microbiol. Res. Food Contr. 21:388–392. 5:2649–2653.

Srinivasan, V., H. M. Nam, L. T. Nguyen, B. Tamilselvam, S. E. Yildirim, Y., Z. Gonulalan, S. Pamuk, and N. Ertas. 2011. Incidence Downloaded from Murinda, and S. P. Oliver. 2005. Prevalence of antimicrobial and antibiotic resistance of Salmonella spp. on raw chicken car- resistance genes in Listeria monocytogenes isolated from dairy casses. Food Res. Int. 44:725–728. farms. Foodborne Pathog. Dis. 2:201–211. Yu, C. Y., C. Chu, S. J. Chou, M. R. Chao, C. M. Yeh, D. Y. Lo, Y. Thai, T. H., T. Hirai, N. T. Lan, A. Shimada, P. T. Ngoc, and R. C. Su, Y. M. Horng, B. C. Weng, J. G. Tsay, and K. C. Huang. Yamaguchi. 2012. Antimicrobial resistance of Salmonella serovars 2008. Comparison of the association of age with the infection of isolated from beef at retail markets in the North Vietnam. J. Salmonella and Salmonella enterica serovar Typhimurium in Pe- Vet. Med. Sci. 74:1163–1169. kin ducks and Roman geese. Poult. Sci. 87:1544–1549. http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 PRODUCTION, MODELING, AND EDUCATION

Alternative growth functions for predicting body, carcass, and breast weight in ducks: Lomolino equation and extreme value function

A. Faridi ,*1 D. Murawska ,† A. Golian ,* M. Mottaghitalab ,‡ A. Gitoee ,§ S. Lopez ,# and J. France Ϧ

* Center of Excellence in the Animal Sciences Department, Ferdowsi University of Mashhad, Mashhad, Iran, 91775-1163; † Department of Commodity Science and Animal Improvement, Faculty of Animal Bioengineering, University of Warmia and Mazury in Olsztyn, Oczapowskiego 5, 10-719 Olsztyn, Poland; Downloaded from ‡ Department of Animal Science, Faculty of Agricultural Science, University of Guilan, PO Box 41635-1314, Rasht, Iran; § Department of Animal Science, College of Agriculture, University of Kurdistan, PO Box 416, Sanandaj, Iran; # Departamento de Produccion Animal, Universidad de Leon, E-24007 Leon, Spain; and Ϧ Centre for Nutrition Modelling, Department of Animal and Poultry Science, University of Guelph, Guelph, Ontario, N1G 2W1, Canada http://ps.oxfordjournals.org/

ABSTRACT In this study, 2 alternative growth func- were obtained with the exponential. Analysis of residu- tions, the Lomolino and the extreme value function als indicated that, for the different traits investigated, (EVF), are introduced and their ability to predict the least biased and the most accurate equations were body, carcass, and breast weight in ducks evaluated. the Gompertz, EVF, Richards, and generalized Michae- A comparative study was carried out of these equa- lis-Menten, whereas the exponential was the most bi-

tions with standard growth functions: Gompertz, ex- ased and least accurate. Based on the Durbin-Watson at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 ponential, Richards, and generalized Michaelis-Menten. statistic, all models generally behaved well and only Goodness of fit of the functions was evaluated using R2, the exponential showed evidence of autocorrelation for mean square error, Akaike information criterion, and all 3 traits investigated. Results showed that with all Bayesian information criterion, whereas bias factor, ac- functions, estimated final weights of males were higher curacy factor, Durbin-Watson statistic, and number of than females for the body, carcass, and breast weight

runs of sign were the criteria used for analysis of re- profiles. The alternative functions introduced here have siduals. Results showed that predictive performance of desirable advantages including flexibility and a low all functions was acceptable, though the Richards and number of parameters. However, because this is prob- exponential equations failed to converge in a few cases ably the first study to apply these functions to predict for both male and female ducks. Based on goodness-of- growth patterns in poultry or other animals, further fit statistics, the Richards, Gompertz, and EVF were analysis of these new models is suggested. the best equations whereas the worst fits to the data Key words: duck , growth function, Lomolino equation, extreme value function 2014 Poultry Science 93 :1031–1042 http://dx.doi.org/ 10.3382/ps.2013-03375

INTRODUCTION BW changes over time, allowing information from recorded measurements to be combined into a few summa- Although most poultry meat and eggs come from ry parameters with meaningful interpretation. Different chickens, significant amounts of meat are produced growth functions have been used to describe growth from ducks in certain parts of the world. Meat and eggs curves in ducks (Knizetova et al., 1991; Maruyama of waterfowl are foods of high nutritional quality and et al., 1999). Historically, the Gompertz equation has are marketed at relatively low prices. Duck and goose been the function of choice for describing growth in production accounts for about 7.5% of total world poul- broilers. To describe duck and goose growth, the Rich- try meat production (Stipkovits and Szathmary, 2012). ards equation offers a flexible model (Mignon-Grasteau Age has a significant effect on body growth rate and and Beaumont, 2000). The generalized Michaelis-Men- carcass tissue composition. Growth functions describe ten (GMM) equation was recently proposed by Lopez et al. (2000) to describe animal growth. The extreme value function (EVF; Williams, 1995) and Lomolino © 2014 Poultry Science Association Inc. equation (Lomolino, 2000) were originally developed to Received June 5, 2013. Accepted November 7, 2013. study the species-area relationship, one of the key tools 1 Corresponding author: [email protected] to evaluate species diversity in conservation biology

1031 1032 FARIDI ET AL. =−−exp[αβ exp( )]; and landscape ecology (Rosenzweig, 1995). Up to the W WWfft present, these functions have not been used to describe growth patterns of avian species. Richards (Thornley and France, 2007): The aims of the present study, therefore, were 1) to introduce and determine the suitability of the EVF and WW = 0 f ; the Lomolino equation for predicting body, carcass, W 1 and breast weight in male and female ducks, and 2) to ααα− [(WWW+− )]e kt α compare the performance of these models with stan- 00f dard growth functions such as the Gompertz, Richards, GMM, and exponential equations. GMM (Lopez et al., 2000): Downloaded from =+()/();ααk kkk + MATERIALS AND METHODS W WWtt0 f

Birds Lomolino (Lomolino, 2000):

Experimental subjects consisted of 200 Pekin ducks http://ps.oxfordjournals.org/ =+/[1; (log(kt / ) )] (sex ratio 1:1) raised to 10 wk of age and fed standard W Wbf diets ad libitum: 1 to 14 d, starter (20.0% CP, 2,700 kcal of ME/kg of feed); 15 to 28 d, grower (18.3% CP, EVF (Williams, 1995): 2,749 kcal of ME/kg of feed); and from d 29 until the =−−(1 exp[ exp(α + )]). end of the experiment, grower (17.0% CP, 2,798 kcal of W Wtf k ME/kg of feed). Throughout the experiment, all birds

were weighed individually at 7-d intervals. Starting at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 In these equations, W (g) represents weight (body, car- from the first week of rearing, 10 males and 10 females cass, or breast) at age t (wk); W and W are the were selected randomly every 7 d for carcass analysis. 0 f zero- and infinite-time values of W, respectively; β is Ducks were fasted for approximately 12 h and slaugh- the intrinsic growth rate; b is a direct measure of the tered by cervical dislocation. Details are reported by slope of the curve through the inflection point; and α Murawska (2012). and k are constants. Data Curve Fitting and Model Evaluation The data used in this study comprised 20 profiles Models were fitted to the growth curves by nonlinear (10 for each sex). Each data profile consisted of body, regression using the NLIN procedure of the SAS pack- carcass, and breast weights of male and female ducks age (SAS Institute Inc., 2008). Statistical criteria used over 10 wk. The range of the data used to investigate to evaluate goodness of fit for all models were mean the growth functions is summarized in Table 1. square error, R2, Bayesian information criterion (BIC), whereas the Durbin-Watson (DW) statistic, number Growth Functions of runs of similar sign, bias factor, and accuracy factor (Ross, 1996) were used for analysis of residuals. In this study 6 different growth functions were fitted Differences between actual and predicted values were to the duck data: considered as residuals. exponential (Thornley and France, 2007):

=−()exp() − −β ; RESULTS AND DISCUSSION W WWWff0 t Several growth functions have been used for ana- Gompertz (Ahmadi and Mottaghitalab, 2007): lyzing the growth profiles of different animals. These

Table 1. Range of data used to investigate the growth functions in male and female ducks

Sex

Male Female

BW Carcass Breast BW Carcass Breast Entity (g) weight (g) weight (g) (g) weight (g) weight (g) Minimum 178.5 86.9 13.1 100.2 41.4 6.6 Maximum 4,010 2,755 925 3,910 2,661 866 Mean 2,403 1,574 436 2,266 1,486 425 SD 1,261.3 882.2 296.8 1,132.8 794.8 276.2 MODELING GROWTH IN DUCKS 1033 models can be regarded as an integration of knowledge behavior. Because of space limitations, not all 60 pro- of the effects of genetic potential, nutrient intake, and files (20 individuals and 3 traits: body, carcass, and environmental conditions on animal growth (Schinckel breast weight) fitted here could be shown. Therefore, a and de Lange, 1996). The Gompertz, Richards, and random selection of male and female body and carcass exponential equations are considered classical growth weight profiles are shown in Figures 1, 2, 3, and 4 by functions and have been used extensively to describe way of example. growth profiles of different animals (Darmani Kuhi et Goodness-of-fit results for the growth functions fitted al., 2003). The GMM equation was introduced by Lopez to the body, carcass, and breast data for both sexes are et al. (2000) after reparameterization of the Michaelis- shown in Tables 3, 4, and 5, respectively. Based on the Menten. The Lomolino equation and EVF were devel- statistical indices, all models were able to predict the oped to investigate the relationship between area and growth patterns satisfactorily. Considering the R2 in- Downloaded from species, often referred to as the closest thing in ecology dex, the highest and lowest values for all 3 investigated to a rule (Scheiner, 2004). However, these latter 2 mod- traits for either sex were obtained with the Richards els have not been used to analyze growth patterns in and exponential equations, respectively. In agreement animal species. To our knowledge this is the first study with our results, Knizetova et al. (1991) claimed that to apply the EVF and the Lomolino to growth patterns growth of ducks will be best represented by the Rich- http://ps.oxfordjournals.org/ describing animal data. ards, because their high maturing rate is reflected in a To investigate convergence, different initial values of more sigmoidal curve than provided for by the Gom- the parameters were used. As the initial values moved pertz. Performance of the newly introduced functions closer to the final solution, convergence was met in a (EVF and Lomolino) was close to that of the Richards, fewer number of iterations. In analyzing the curves, it Gompertz, and GMM. The lowest mean square error was generally found that all functions were able to fit values for all investigated traits were obtained with the

all growth patterns, though the Richards was unable to Richards, followed by EVF on breast meat (both sexes) at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 fit one curve for each of male carcass, male breast, and and female carcass data. For male carcass data, the female breast weight. The exponential failed to con- GMM ranked second. It is worth noting that EVF like verge in 3 cases for male breast weight. In agreement the Gompertz has a fixed inflection point, which for the with these results, previous studies have shown that EVF occurs at 63.2% of its asymptote (Tjørve, 2009). the Richards can be difficult to fit (e.g., Mignon-Gras- However, despite possessing fixed inflection points, teau and Beaumont, 2000; Porter et al., 2010). Limita- both the EVF and Gompertz performed well in predict- tions of different functions in fitting duck growth data ing the growth behavior of body, carcass, and breast in have been reported previously. Maruyama et al. (1999) both the male and female duck lines. found that the Gompertz failed to converge when fit- Overall BIC values calculated for the fitted equations ting BW data from 4 duck lines. The newly introduced for the 3 traits are shown in Tables 3, 4, and 5. Lower models, however, fitted the investigated data without values of this index indicate the preferred model (i.e., difficulty. This is beneficial. With models that fail con- the one with the fewest parameters that still provides vergence, the corresponding data lines will be omitted an adequate fit to the data). Bayesian information cri- from further analysis, and therefore valuable data will terion, therefore, is a useful index for comparing models be lost. with different numbers of parameters (Köhn and Shar- Average estimated parameters for body, carcass, and ifi, 2007). Analyzing the results for BW showed that breast weight curves of the duck lines are given in Table the Gompertz and Richards provided the lowest BIC 2. An advantage of the newly introduced models is the values, followed by the Lomolino, whereas the highest small number of parameters (n = 3), which is within the values were obtained with the exponential (Table 3). range for the classical models compared here. The fewer A similar trend was seen with carcass weight for which the parameters, the easier the function is to fit and the lowest BIC values were obtained with the Gom- interpret (Faridi et al., 2011). A function with mean- pertz and Richards and the highest with the exponen- ingful parameters has much to recommend it compared tial (Table 4). The lowest BIC values for breast weight with an equation that merely provides a close fit to the were obtained with the Richards and EVF followed by study data. The asymptotic weight Wf is the common the Lomolino, Gompertz, and GMM. The exponential biologically interpretable parameter across models. Re- gave the highest BIC for this particular trait (Table 5). sults showed the models gave higher estimates of Wf in Analysis of residuals was conducted using the sta- males compared with females for all traits (body, car- tistics: bias factor, accuracy factor, DW, and number cass, and breast weight). As Table 2 shows, the highest of runs of similar sign. Results for body, carcass, and and the lowest values of Wf for the 3 traits for either breast are summarized in Tables 6, 7, and 8, respec- sex were achieved with the exponential and EVF, re- tively. Bias factor is a measure of the relative aver- spectively. Based on the results obtained, the estimated age deviation of predicted from observed values. An value of the asymptote for BW was higher than that accuracy factor of 1, like bias factor, represents perfect for carcass weight, and carcass was higher than that agreement between observed and predicted values. The for breast. This suggests the functions (including the larger the deviation from 1 the value, the less accurate newly introduced ones) could effectively track growth is the average prediction (Ross, 1996). 1034 Table 2. Range of estimated parameters (SD) for the functions investigated

Trait

BW Carcass weight Breast weight

Model1,2 Male Female Male Female Male Female EVF Wf 3,589 to 3,895 3,180 to 3,501 2,456 to 2,619 2,103 to 2,386 772 to 914 658 to 808 (88.43) (105.21) (51.37) (88.58) (47.31) (46.95) α 0.45 to 0.57 0.53 to 0.66 0.4848 to 0.573 0.555 to 0.681 0.44 to 0.60 0.47 to 0.65 (0.034) (0.041) (0.029) (0.039) (0.059) (0.049) k −2.63 to −2.27 −2.76 to −2.49 −2.83 to −2.52 −2.94 to −2.74 −3.56 to −2.92 −3.53 to −2.98 (0.104) (0.093) (0.101) (0.076) (0.233) (0.178) Gompertz Wf 3,828 to 4,380 3,287 to 3,693 2,648 to 2,969 2,168 to 2,535 859 to 1,222 714 to 984 (163.21) (122.22) (99.22) (108.69) (132.72) (86.52) α 3.86 to 5.07 4.46 to 6.86 4.38 to 5.64 5.04 to 7.89 4.66 to 7.32 5.33 to 9.12 (0.398) (0.685) (0.382) (0.811) (0.875) (1.103) β 0.0001 to 0.00013 0.00014 to 0.002 0.00014 to 0.0002 0.0002 to 0.0003 0.00023 to 0.0005 0.00035 to 0.00072 (0) (0.00002) (0) (0) (0.00001) (0.00013) Lomolino Wf 4,120 to 5,248 3,390 to 3,987 2,841 to 3,403 2,226 to 2,717 958 to 1,666 753 to 1,182

(357.61) (186.12) (196.31) (148.38) (280.19) (136.49) FARIDI ETAL. b 6.57 to 12.74 9.6 to 24.16 8.2 to 14.9 11.68 to 30.1 7.34 to 23.71 10.25 to 36.75 (2.065) (4.276) (2.138) (5.371) (5.468) (7.721) k 4.01 to 5.08 3.45 to 4.15 4.33 to 5.16 3.57 to 4.51 5.51 to 8.97 4.67 to 6.79 (0.367) (0.214) (0.289) (0.259) (1.421) (0.705) GMM Wf 3,977 to 5,222 3,339 to 3,960 2,745 to 3,256 2,200 to 2,683 880 to 1,602 738 to 1,202 (438.8) (188.5) (195.2) (142.0) (267.4) (147.3) α 4.09 to 5.12 3.57 to 4.17 4.4 to 5.1 3.68 to 4.52 5.43 to 8.61 4.71 to 6.87 (0.350) (0.173) (0.243) (0.229) (1.261) (0.676) k 1.82 to 2.96 2.6 to 3.6 2.2 to 3.1 2.86 to 3.77 2.13 to 4.04 2.27 to 4.27 (0.385) (0.38) (0.296) (0.352) (0.709) (0.614) W0 27 to 201 41.1 to 284.1 35.6 to 158.1 48.7 to 172.3 6.86 to 59.51 4.86 to 47 (88.58) (64.5) (34.2) (34.44) (18.56) (14.82) Richards Wf 3,751 to 4,615 3,245 to 3,688 2,523 to 2,879 2,135 to 2,476 778 to 1,383 681 to 841 (313.11) (139.42) (131.70) (104.56) (218.69) (48.84) α 0.28 to 0.63 0.516 to 1.01 0.37 to 0.92 0.60 to 1.11 0.21 to 2.15 0.63 to 1.32 (0.134) (0.169) (0.161) (0.177) (0.652) (0.233) k 0.048 to 0.444 0.024 to 1.43 0.14 to 1.47 0.203 to 1.79 0.26 to 27.21 0.63 to 2.61 (0.370) (0.441) (0.478) (0.468) (9.154) (0.658) W0 0 to 115 24 to 168 19 to 125 21 to 112 2.24 to 50.10 10.8 to 30.5 (46.11) (42.91) (33.44) (25.45) (15.60) (6.28) Exponential Wf 4,651 to 6,370 3,708 to 4,395 3,411 to 4,944 2,466 to 3,193 2,200 to 4,253 1,066 to 6,849 (521.1) (220.6) (476.5) (240.1) (924.1) (1,889.4) β 0.128 to 0.196 0.19 to 0.27 0.103 to 0.16 0.15 to 0.27 0.03 to 0.06 0.015 to 0.13 (0.022) (0.025) (0.024) (0.032) (0.012) (0.038) W0 −933 to −694 −1,242 to −787 −697 to −515 −868 to −522 −162 to −144 −189 to −128 (106.53) (127.51) (51.59) (90.91) (6.18) (23.47) 1EVF = extreme value function; GMM = generalized Michaelis-Menten.

Wf = final weight; W0 = initial weight; β = intrinsic growth rate; b = direct measure of the slope of the curve through the inflection point; α, k are constants.

Downloaded from from Downloaded http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 6, April on University Harvard Zoology, Comp Museum the of Library Mayr Ernst at MODELING GROWTH IN DUCKS 1035 Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 Figure 1. Observed (dots) and predicted (solid lines) values for a randomly selected individual male BW profile. EVF = extreme value function; GMM = generalized Michaelis-Menten.

The average bias and accuracy factors across the tors across the curves give an indication of the ability of curves have to be interpreted with caution because large models to fit curves with outliers (Lopez et al., 2004). underestimates may balance with large overestimates The residuals were analyzed further using the DW giving an average factor approaching unity. Therefore, statistic to test whether a model has been successful the median and the range of values should be examined in describing the underlying trend. Statistically nonsig- (Tables 6, 7, and 8). Maximum and minimum bias fac- nificant DW values are around 2, obtained when serial

Figure 2. Observed (dots) and predicted (solid lines) values for a randomly selected individual female BW profile. EVF = extreme value function; GMM = generalized Michaelis-Menten. 1036 FARIDI ET AL. Downloaded from http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014

Figure 3. Observed (dots) and predicted (solid lines) values for a randomly selected individual male carcass weight profile. EVF = extreme value function; GMM = generalized Michaelis-Menten. correlation is small and the residuals are distributed al., 2000). As Tables 6, 7, and 8 show, all models gen- randomly around the zero line. However, when the DW erally behaved well and only the exponential showed is significant, serial correlation is significant because of evidence of autocorrelation in a few instances over all 3 the presence of cycles in the plot of residuals (Lopez et traits investigated here.

Figure 4. Observed (dots) and predicted (solid lines) values for a randomly selected individual female carcass weight profile. EVF = extreme value function; GMM = generalized Michaelis-Menten. Table 3. Goodness of fit of the functions used to predict BW in male and female ducks

Model2

EVF Gompertz Lomolino GMM EXP Richards

Entity1 Male Female Male Female Male Female Male Female Male Female Male Female

R2 Minimum 0.978 0.971 0.981 0.974 0.979 0.972 0.979 0.972 0.962 0.961 0.982 0.974 Maximum 0.991 0.993 0.996 0.993 0.995 0.992 0.995 0.993 0.992 0.983 0.998 0.995 Mean 0.984 0.986 0.991 0.988 0.990 0.987 0.99 0.987 0.981 0.971 0.992 0.989 MSE Minimum 19,082 11,919 8,205 10,011 11,447 12,445 9,860 11,313 15,907 25,158 3,855 8,778 Maximum 42,693 57,763 35,827 47,939 40,707 51,584 38,863 51,502 59,964 69,853 34,702 47,942 Mean 31,672 24,944 18,091 19,293 20,348 22,817 19,534 21,068 37,441 46,989 16,157 17,651 BIC

Minimum 105.5 100.8 97.3 99.2 100.4 101.1 101.2 102.6 103.6 108.2 91.8 100.0 MODELING GROWTH INDUCKS Maximum 113.5 116.6 111.8 114.7 113.1 115.4 114.9 117.7 116.9 118.4 113.8 116.9 Mean 110.0 107.1 103.8 104.3 104.9 106.2 106.8 107.6 111.4 114.0 104.1 105.6 1MSE = mean square error; BIC = Bayesian information criterion. 2EVF = extreme value function; GMM = generalized Michaelis-Menten; EXP = exponential.

Table 4. Goodness of fit of the functions used to predict carcass weight in male and female ducks

Model2

EVF Gompertz Lomolino GMM EXP Richards

Entity1 Male Female Male Female Male Female Male Female Male Female Male Female R2 Minimum 0.975 0.972 0.968 0.971 0.965 0.970 0.966 0.969 0.949 0.943 0.971 0.980 Maximum 0.994 0.997 0.998 0.998 0.998 0.997 0.998 0.997 0.987 0.981 0.998 0.997 Mean 0.986 0.987 0.985 0.986 0.985 0.983 0.984 0.985 0.971 0.965 0.988 0.988 MSE Minimum 7,200 2,801 1,666 1,923 2,403 2,386 1,596 2,212 12,325 14,889 1,597 1,779 Maximum 26,893 25,668 32,295 26,565 35,353 28,689 34,540 28,545 52,084 41,597 26,856 25,824 Mean 14,312 10,712 13,943 11,025 15,587 14,924 13,631 11,875 28,167 28,042 10,752 9,510 BIC Minimum 95.7 86.3 81.1 82.6 84.7 84.7 83.0 86.3 101.1 103.0 83.1 84.0 Maximum 108.9 108.4 110.7 108.8 111.6 109.5 113.7 111.8 115.6 113.3 111.2 110.9 Mean 101.7 98.1 99.0 97.7 100.6 100.8 100.1 100.8 108.4 108.7 99.2 98.2 1MSE = mean square error; BIC = Bayesian information criterion.

2EVF = extreme value function; GMM = generalized Michaelis-Menten; EXP = exponential. 1037

Table 5. Goodness of fit of the functions used to predict breast weight in male and female ducks

Model2

EVF Gompertz Lomolino GMM EXP Richards

Entity1 Male Female Male Female Male Female Male Female Male Female Male Female FARIDI ETAL. R2 Minimum 0.973 0.975 0.959 0.969 0.957 0.968 0.961 0.968 0.936 0.945 0.976 0.977 Maximum 0.996 0.998 0.998 0.996 0.999 0.996 0.999 0.997 0.976 0.991 0.998 0.997 Mean 0.987 0.988 0.982 0.985 0.982 0.985 0.982 0.985 0.961 0.964 0.990 0.990 MSE Minimum 526 261 189 435 146 354 161 352 2,853 970 161 254 Maximum 3,012 2,614 4,532 3,260 4,669 3,378 4,309 3,349 6,963 5,369 2,606 1,787 Mean 1,423 1,304 1,993 1,406 2,031 1,448 1,900 1,358 4,263 3,396 1,041 925 BIC Minimum 89.5 62.6 84.3 67.7 80.1 65.5 82.9 67.8 111.4 75.7 83.6 64.6 Maximum 112.0 85.6 116.1 87.8 116.3 88.1 115.5 90.4 120.3 92.8 111.5 84.1 Mean 100.4 76.9 102.2 77.5 101.9 77.6 102.1 79.2 115.0 87.3 99.1 75.9 1MSE = mean square error; BIC = Bayesian information criterion.

2EVF = extreme value function; GMM = generalized Michaelis-Menten; EXP = exponential.

Downloaded from from Downloaded http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 6, April on University Harvard Zoology, Comp Museum the of Library Mayr Ernst at Table 6. Analysis of residuals for the functions used to predict BW in male and female ducks

Model2

EVF Gompertz Lomolino GMM EXP Richards

Entity1 Male Female Male Female Male Female Male Female Male Female Male Female

BF MODELING GROWTH INDUCKS Mean 1.078 1.075 1.009 0.997 0.969 0.956 1.009 1.011 0.892 0.897 1.013 1.021 Minimum 1.069 1.011 0.983 0.966 0.935 0.912 1.001 1 0.764 0.717 0.999 1 Maximum 1.112 1.156 1.037 1.049 1.007 1.007 1.014 1.032 1.033 1.311 1.026 1.053 Range 0.040 0.145 0.054 0.084 0.073 0.095 0.013 0.028 0.268 0.591 0.027 0.048 Median 1.080 1.074 1.003 0.994 0.964 0.95 1.009 1.009 0.919 0.874 1.014 1.014 AF Mean 1.321 1.308 1.078 1.105 1.122 1.215 1.068 1.082 1.937 2.472 1.082 1.097 Minimum 1.267 1.245 1.043 1.042 1.043 1.08 1.032 1.043 1.201 1.187 1.032 1.038 Maximum 1.390 1.621 1.147 1.248 1.225 1.427 1.101 1.171 3.361 5.422 1.127 1.21 Range 0.123 0.376 0.104 0.206 0.182 0.347 0.069 0.127 2.160 4.235 0.096 0.172 Median 1.310 1.267 1.059 1.076 1.124 1.196 1.066 1.067 1.585 2.282 1.070 1.073 DW Minimum 1.46 1.64 1.66 1.15 1.45 1.09 1.63 1.24 1.19 0.89 1.79 1.51 Maximum 2.37 2.13 3.13 3.16 3.4 2.87 3.4 3.03 2.8 2.03 3.25 3.25 Mean 1.69 1.83 2.40 2.18 2.34 1.91 2.41 2.1 1.81 1.49 2.55 2.38 No. of SDW 0 0 0 2 0 2 0 1 2 2 0 0 NRS3 4 22.2 11.1 — 22.2 22.2 44.5 — — 55.6 55.6 — — 5, 6 66.7 77.8 33.3 44.5 55.6 44.5 55.6 66.7 44.4 44.4 22.2 66.7 7, 8 22.2 11.1 66.7 33.3 11.1 11.1 44.4 33.3 — — 66.7 33.3 9 — — — — 11.1 — — — — — 11.1 — 1BF = bias factor; AF = accuracy factor; DW = Durbin-Watson; SDW = curves with significant Durbin-Watson (P < 0.05); NRS = number of runs of sign. 2EVF = extreme value function; GMM = generalized Michaelis-Menten; EXP = exponential. 3As a percentage of all curves.

1039

Table 7. Analysis of residuals for the functions used to predict carcass weight in male and female ducks

Model2

EVF Gompertz Lomolino GMM EXP Richards

Entity1 Male Female Male Female Male Female Male Female Male Female Male Female

BF Mean 1.088 1.092 1 0.986 0.957 0.938 1.018 1.022 0.809 0.943 1.037 1.042 Minimum 0.998 1.015 0.975 0.931 0.920 0.871 0.991 0.999 0.510 0.653 1.004 1.008 Maximum 1.116 1.196 1.038 1.059 0.998 1.008 1.029 1.065 1.047 1.124 1.059 1.117 Range 0.118 0.181 0.064 0.128 0.078 0.137 0.038 0.066 0.536 0.471 0.055 0.108 Median 1.099 1.092 0.996 0.985 0.955 0.937 1.019 1.018 0.830 0.969 1.035 1.037 FARIDI ETAL. AF Mean 1.378 1.369 1.089 1.145 1.164 1.277 1.131 1.137 1.728 1.407 1.218 1.183 Minimum 1.306 1.270 1.039 1.045 1.048 1.112 1.078 1.054 1.167 1.229 1.072 1.040 Maximum 1.451 1.788 1.159 1.322 1.261 1.477 1.198 1.287 2.757 2.132 1.599 1.469 Range 0.145 0.518 0.119 0.277 0.213 0.365 0.119 0.232 1.589 0.903 0.528 0.429 Median 1.366 1.334 1.081 1.107 1.183 1.245 1.119 1.112 1.705 1.307 1.201 1.146 DW Minimum 1.22 1.09 1.63 0.99 1.48 0.99 1.63 1.21 1.08 0.84 1.99 1.30 Maximum 2.70 2.33 3.10 2.84 2.61 2.35 2.83 2.73 2.16 1.52 3.23 3.05 Mean 1.87 1.73 2.31 1.97 2.02 1.83 2.21 2.00 1.53 1.29 2.52 2.26 No. of SDW 0 2 0 1 0 1 0 1 3 4 0 1 NRS3 4 22.2 11.1 11.1 33.3 55.6 77.8 — — 77.8 88.9 — — 5, 6 66.6 77.8 77.8 66.7 44.6 22.2 77.8 88.9 22.2 11.1 37.5 88.9 7, 8 11.2 11.1 11.1 — — — 22.3 11.1 — — 62.5 11.1 9 ———————————— 1BF = bias factor; AF = accuracy factor; DW = Durbin-Watson; SDW = curves with significant Durbin-Watson (P < 0.05); NRS = number of runs of sign. 2EVF = extreme value function; GMM = generalized Michaelis-Menten; EXP = exponential.

3As a percentage of all curves.

Downloaded from from Downloaded http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 6, April on University Harvard Zoology, Comp Museum the of Library Mayr Ernst at Table 8. Analysis of residuals for the functions used to predict breast weight in male and female ducks

Model2

EVF Gompertz Lomolino GMM EXP Richards

Entity1 Male Female Male Female Male Female Male Female Male Female Male Female

BF MODELING GROWTH INDUCKS Mean 1.104 1.139 0.970 0.951 0.899 0.891 1.049 1.049 1.075 1.152 1.078 1.070 Minimum 1.006 1.035 0.906 0.798 0.815 0.752 0.978 0.960 0.978 1.089 1.000 1.023 Maximum 1.161 1.229 1.058 1.101 0.998 1.026 1.096 1.130 1.145 1.215 1.153 1.130 Range 0.155 0.194 0.152 0.303 0.183 0.274 0.118 0.170 0.167 0.126 0.130 0.108 Median 1.130 1.136 0.957 0.931 0.881 0.870 1.055 1.053 1.082 1.162 1.074 1.066 AF Mean 1.441 1.517 1.190 1.293 1.358 1.446 1.344 1.319 1.311 1.393 1.370 1.301 Minimum 1.289 1.319 1.125 1.119 1.072 1.082 1.037 1.182 1.185 1.186 1.035 1.231 Maximum 1.561 1.961 1.279 1.668 1.618 1.891 1.524 1.582 1.508 1.641 1.603 1.487 Range 0.272 0.641 0.154 0.549 0.545 0.808 0.486 0.399 0.323 0.455 0.567 0.256 Median 1.469 1.449 1.192 1.259 1.374 1.444 1.403 1.259 1.305 1.395 1.351 1.273 DW Minimum 1.07 1.25 1.64 1.51 1.61 1.54 1.76 1.65 1.10 1.08 1.74 1.87 Maximum 3.28 3.17 3.41 3.02 3.33 2.93 3.53 3.00 1.99 2.02 3.74 3.24 Mean 2.41 2.31 2.40 2.29 2.39 1.27 2.53 2.42 1.43 1.43 2.70 2.72 No. of SDW 1 1 0 0 0 0 0 0 4 4 0 0 NRS3 4 11.1 11.1 55.6 33.3 66.7 33.3 — — 83.4 88.9 — — 5, 6 44.4 55.6 22.2 55.6 11.1 44.5 66.7 44.4 16.7 — 25.0 42.9 7, 8 44.5 33.3 22.2 11.1 22.2 22.2 11.1 55.6 — 11.1 37.5 42.9 9 — — — — — — 22.2 — — — 37.5 14.3 1BF = bias factor; AF = accuracy factor; DW = Durbin-Watson; SDW = curves with significant Durbin-Watson (P < 0.05); NRS = number of runs of sign. 2EVF = extreme value function; GMM = generalized Michaelis-Menten; EXP = exponential. 3As percentage of all curves.

1041

Downloaded from from Downloaded http://ps.oxfordjournals.org/ at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 6, April on University Harvard Zoology, Comp Museum the of Library Mayr Ernst at 1042 FARIDI ET AL. The runs test is a simple and robust method for Faridi, A., M. Mottaghitalab, F. Rezaee, and J. France. 2011. Na- determining whether data differ systematically from rushin-Takma models as flexible alternatives for describing economic traits in broiler breeder flocks. Poult. Sci. 90:507–515. a theoretical curve. A run is a series of consecutive Knizetová, H., J. Hyanek, B. Knize, and H. Prochazkova. 1991. points with a residual of the same sign (positive or Analysis of growth curves of fowl. II. Ducks. Br. Poult. Sci. negative). A high number of runs of sign is indicative 32:1039–1053. of a more random distribution of residuals (Motulsky Köhn, F., and A. R. Sharifi. 2007. Modeling the growth of the Goet- tingen minipig. J. Anim. Sci. 85:84–92. and Ransnas, 1987). A small number of runs of similar Lomolino, M. V. 2000. Ecology’s most general, yet protean pattern: sign is obtained when the residuals are not randomly The species–area relationship. J. Biogeogr. 27:17–26. distributed, so residuals of the same sign tend to cluster López, S., J. France, W. J. J. Gerrits, M. S. Dhanoa, D. J. Humphries, together on some parts of the curve. Results showed and J. Dijkstra. 2000. A generalized Michaelis–Menten equation

for the analysis of growth. J. Anim. Sci. 78:1816–1828. Downloaded from that for body, carcass, and breast weight, the exponen- López, S., M. Prieto, J. Dijkstra, M. S. Dhanoa, and J. France. tial was the model with highest percentage of minimum 2004. Statistical evaluation of mathematical models for microbial number of runs of sign (n = 4). The Lomolino, GMM, growth. Int. J. Food Microbiol. 96:289–300. Maruyama, K., M. K. Akbar, and C. M. Turk. 1999. Growth pattern and Richards showed the maximum number of runs of and carcase development in male ducks selected for growth rate. sign (n = 9), though none of models gave 9 runs of sign Br. Poult. Sci. 40:233–239. for carcass weight (Tables 6, 7, and 8). Mignon-Grasteau, S., and C. Beaumont. 2000. Growth curves in http://ps.oxfordjournals.org/ In conclusion, the newly introduced growth functions birds. INRA Prod. Anim. 13:337–348. Motulsky, H. J., and L. A. Ransnas. 1987. Fitting curves to data showed reasonable performance compared with other using nonlinear regression: A practical and nonmathematical re- more familiar ones in predicting body, carcass, and view. FASEB J. 1:365–374. breast weight of ducks. These functions have desirable Murawska, D. 2012. The effect of age on the growth rate of tissues advantages including flexibility and a low number of and organs and the percentage content of edible and nonedible carcass components in Pekin ducks. Poult. Sci. 91:2030–2038. parameters. They were readily fitted to the investigat- Porter, T., E. Kebreab, H. Darmani Kuhi, S. Lopez, A. B. Strathe, ed data, whereas Richards and exponential equations and J. France. 2010. Flexible alternatives to the Gompertz equa- at Ernst Mayr Library of the Museum Comp Zoology, Harvard University on April 6, 2014 failed to reach convergence in some cases. This is a tion for describing growth with age in turkey hens. Poult. Sci. noteworthy advantage of the newly introduced mod- 89:371–378. Rosenzweig, M. L. 1995. Species Diversity in Space and Time. Cam- els. However, this study is probably the first to use bridge University Press, Cambridge, UK. these equations to predict animal growth patterns, and Ross, T. 1996. Indices for performance evaluation of predictive mod- therefore further studies involving these functions are els in food microbiology. J. Appl. Bacteriol. 81:501–508. SAS Institute Inc. 2008. SAS/STAT 9.2 Users Guide. SAS Institute desirable. Inc., Cary, NC. Scheiner, S. M. 2004. A mélange of curves—Further dialogue about ACKNOWLEDGMENTS species–area curves. Glob. Ecol. Biogeogr. 13:479–484. Schinckel, A. P., and C. F. M. de Lange. 1996. Characterization of growth parameters needed as inputs for pig growth models. J. The Canada Research Chairs Program is thanked for Anim. Sci. 74:2021–2036. partial funding. Stipkovits, L., and S. Szathmary. 2012. Mycoplasma infection of ducks and geese. Poult. Sci. 91:2812–2819. Thornley, J. H. M., and J. France. 2007. Mathematical Models in REFERENCES Agriculture. Rev. 2nd ed. CAB International, Wallingford, UK. Tjørve, E. 2009. Shapes and functions of species–area curves (II): Ahmadi, H., and M. Mottaghitalab. 2007. Hyperbolastic models as A review of new models and parameterizations. J. Biogeogr. a new powerful tool to describe broiler growth kinetics. Poult. 36:1435–1445. Sci. 86:2461–2465. Williams, M. R. 1995. An extreme-value function model of the spe- Darmani Kuhi, H., E. Kebreab, S. Lopez, and J. France. 2003. An cies incidence and species–area relations. Ecology 76:2607–2616. evaluation of different growth functions for describing the profile of live weight with time (age) in meat and egg strains of chicken. Poult. Sci. 82:1536–1543. ASSOCIATION NOTES

Nunc Dimittis

Ari van Tienhoven Patterns of Mammalian Reproduction (a revision of S. A. Asdell’s Patterns of Mammalian Reproduction) and Reproductive Physiology of Vertebrates. In 1960 he was Ari van Tienhoven, 91, of Ithaca, NY, retired emeri- elected a Fellow of the American Association for the tus professor of animal physiology, Cornell Univer- Advancement of Science and in 1987 a Fellow of the sity, died January 31, 2014. Professor van Tienhoven Poultry Science Association. was born April 22, 1922, Although most notably recognized for his research in The Hague, the Neth- and publications, van Tienhoven derived his greatest erlands. He came to the pleasure from teaching. His students and colleagues United States in 1947 and considered him to be an outstanding instructor and obtained his US citizen- mentor, honoring him with the College of Agriculture ship in 1959. van Tien- and Life Sciences, Cornell University Teacher of Merit hoven received his mas- Award, and the Eagerton Career Teaching Award in ter of science in 1951 and 1987. He continued to teach well after his retirement, doctor of philosophy in developing a new and unique course for Cornell Univer- 1953 from the University sity titled AIDS and Society. of Illinois in reproduc- van Tienhoven was a longtime patron and benefactor tive physiology. He was of Mann Library, Cornell University. Mann Library was an assistant professor at a very special place to him. He attributed much of his Mississippi State Univer- success as a scholar to the tremendous resources the sity from 1953 to 1955. In library provided, including the library staff. In 1987 he 1955 he joined the faculty established the Ans van Tienhoven Award in honor of of Cornell University where he remained as professor of his wife. The award is presented annually to a Mann poultry science until his retirement in 1987. librarian to assist in travel and professional develop- van Tienhoven was a renowned scholar in avian phys- ment. iology with particular emphasis on reproductive endo- Ari van Tienhoven married Ans van Haselen in 1950. crinology of domestic fowl. His research and writings They were married for fifty-five years until her death contributed significantly to the understanding of the in 2005. They are survived by three children, Richard relationships between endocrine secretions and ovula- A. van Tienhoven of Warwick, Rhode Island, Arianne tion in chickens, the avian nervous system, thermoregu- J. Tepper of Pen Yan, New York, and Andrew W. van lation of birds, and animal welfare. He authored some Tienhoven of Ithaca, New York, as well as six grand- 60 refereed journal publications as well as two books, children.

1043 Positions Available

Gordon Cain Endowed Chair in Animal Nutrition three of these facilities are located on the LSU main campus in Associate or full professor, Gordon A. Cain Endowed Chair Baton Rouge. In addition, the university also operates two com- in Animal Nutrition. Work location: Louisiana State University mercial broiler houses located in north Louisiana which can be (LSU) School of Animal Sciences, College of Agriculture, Ba- utilized for poultry research and production demonstration proj- ton Rouge. Position description: This position is a 12-month ects. Qualification requirements: PhD in animal nutrition, poul- appointment with joint responsibilities in research (70%) and try science, or related field with an established record of refereed teaching (30%). The selected candidate will also have the op- publications and extramural funding. Salary and benefits: Salary portunity for supplemental salary through extramural grant will be commensurate with qualifications and experience. The funding. Research: The successful candidate will be expected to LSU AgCenter has an attractive benefits package with a wide develop an independent research program in nonruminant nutri- variety of benefit options. Benefits offered include retirement, tion with primary emphasis in the area of poultry nutrition along multiple medical insurance options, supplemental insurances with the opportunity to use swine for specific research projects. (dental, life, long-term disability, accident, vision, long-term The applicant will be expected to conduct research appropriate care, etc.), university holidays (14 per year, typically includes for publication in refereed journals; recruit and advise graduate a week off at Christmas), generous annual (vacation) and sick students; attract extramural funding; participate in collaborative leave benefits, Employee Assistance Program, and possible edu- research within the LSU system and with other outside enti- cational leave and tuition exemption for coursework at campuses ties. Teaching: Candidates must be able to teach undergraduate/ of the LSU system. Specific benefits depend on job category, graduate courses in one or more of the following areas: animal percent effort, and length of employment. Application deadline: nutrition, poultry biology, and poultry management. Teaching June 1, 2014, or until a suitable candidate is identified. Applica- expectations also include shared responsibility for other depart- tion procedure: Apply on line at www.lsusystemcareers.lsu.edu. mental courses and advising both graduate and undergraduate Include a letter of application, curriculum vitae, transcripts, and students. Candidate will also be expected to develop a strong three letters of recommendation. Questions about the on-line graduate mentoring program, especially at the PhD level. Facili- system should be directed to the HRM Office at 225/578-2258. ties: Excellent farm and laboratory research facilities are avail- Questions about the position and search process may be directed able. The farm facilities include capabilities for broilers, laying to Dr. Gary Hay, Director, School of Animal Sciences, J. B. hens, and broiler breeders, with an onsite feed mill. The labo- Francioni Hall, LSU College of Agriculture, Baton Rouge, LA; ratory facilities include five Petersime Brooder batteries with phone: 225-578-3241; fax 225-578-3279; e-mail: ghay@agcenter. onsite diet mixing capability. Laboratory facilities also include lsu.edu. Web site: School of Animal Sciences, www.lsuagcenter. a new state-of-the-art animal nutrition research laboratory. All com/en/our_offices/departments/Animal_Sciences.

1044 ABBREVIATIONS

Revised March 2011 7KHIROORZLQJDEEUHYLDWLRQVPD\EHXVHGZLWKRXWGH¿QLWLRQLQPoultry Science. Plural abbreviations do not require “s”. &KHPLFDOV\PEROVDQGWKUHHOHWWHUDEEUHYLDWLRQVIRUDPLQRDFLGVGRQRWQHHGGH¿QLWLRQ8QLWVRIPHDVXUHH[FHSWWKRVH VKRZQEHORZVKRXOGEHDEEUHYLDWHGDVOLVWHGLQWKHCRC Handbook for Chemistry and Physics &5&3UHVV&RU- SRUDWH%OYG%RFD5DWRQ)/ DQGGRQRWQHHGWREHGH¿QHG A adenine 0+& PDMRUKLVWRFRPSDWLELOLW\FRPSOH[ ADG average daily gain mRNA messenger ribonucleic acid $'), DYHUDJHGDLO\IHHGLQWDNH min minute $0( DSSDUHQWPHWDEROL]DEOHHQHUJ\ mo month AMEn QLWURJHQFRUUHFWHGDSSDUHQWPHWDEROL]DEOHHQHUJ\ 06 PHDQVTXDUH ANOVA analysis of variance n number of observations %FHOO EXUVDOGHULYHGEXUVDOHTXLYDOHQWGHULYHGFHOO N normal ES EDVHSDLUV NAD nicotinamide adenine dinucleotide %6$ ERYLQHVHUXPDOEXPLQ NADH reduced nicotinamide adenine dinucleotide %: ERG\ZHLJKW NRC National Research Council C cytosine 16 QRWVLJQL¿FDQW F'1$ FRPSOHPHQWDU\'1$ 3$*( SRO\DFU\ODPLGHJHOHOHFWURSKRUHVLV cfu colony-forming units 3%6 SKRVSKDWHEXIIHUHGVDOLQH &, FRQ¿GHQFHLQWHUYDO 3&5 SRO\PHUDVHFKDLQUHDFWLRQ &3 FUXGHSURWHLQ 47/ TXDQWLWDWLYHWUDLWORFL FSP FRXQWVSHUPLQXWH SIX SODTXHIRUPLQJXQLWV &9 FRHI¿FLHQWRIYDULDWLRQ U FRUUHODWLRQFRHI¿FLHQW d day r FRHI¿FLHQWRIGHWHUPLQDWLRQVLPSOH df degrees of freedom R FRHI¿FLHQWRIGHWHUPLQDWLRQPXOWLSOH DM dry matter 5)/3 UHVWULFWLRQIUDJPHQWOHQJWKSRO\PRUSKLVP '1$ GHR[\ULERQXFOHLFDFLG RH relative humidity EDTA ethylenediaminetetraacetate 5,$ UDGLRLPPXQRDVVD\ (/,6$ HQ]\PHOLQNHGLPPXQRVRUEHQWDQWLERG\DVVD\ RNA ribonucleic acid (67 H[SUHVVHGVHTXHQFHWDJ USP UHYROXWLRQVSHUPLQXWH g gram s second g gravity 6' VWDQGDUGGHYLDWLRQ G guanine 6'6 VRGLXPGRGHF\OVXOIDWH GAT glutamic acid-alanine-tyrosine 6( VWDQGDUGHUURU *) JDLQWRIHHGUDWLR 6(0 VWDQGDUGHUURURIWKHPHDQ */0 JHQHUDOOLQHDUPRGHO 613 VLQJOHQXFOHRWLGHSRO\PRUSKLVP h hour 65%& VKHHSUHGEORRGFHOOV +(3(6 NK\GUR[\HWK\OSLSHUD]LQHNc-ethane-sulfonic acid T thymine +3/& KLJKSHUIRUPDQFH KLJKSUHVVXUH OLTXLGFKURPDWRJUDSK\ 7%$ WKLREDUELWXULFDFLG ,&8 LQWHUQDWLRQDOFKLFNXQLWV T cell thymic-derived cell ,J LPPXQRJOREXOLQ 70( WUXHPHWDEROL]DEOHHQHUJ\ ,/ LQWHUOHXNLQ TMEn QLWURJHQFRUUHFWHGWUXHPHWDEROL]DEOHHQHUJ\ ,8 LQWHUQDWLRQDOXQLWV 7ULV WULV K\GUR[\PHWK\O DPLQRPHWKDQH NE NLOREDVHSDLUV 76$$ WRWDOVXOIXUDPLQRDFLGV N'D NLORGDOWRQ 8 XULGLQH / OLWHU 86'$ 8QLWHG6WDWHV'HSDUWPHQWRI$JULFXOWXUH /' KRXUVOLJKWKRXUVGDUNQHVVLQDSKRWRSHULRG 89 XOWUDYLROHW m meter vol/vol volume to volume μ micro vs. versus M molar wt/vol weight to volume 0$6 PDUNHUDVVLVWHGVHOHFWLRQ wt/wt weight to weight 0( PHWDEROL]DEOHHQHUJ\ ZN ZHHN MEn QLWURJHQFRUUHFWHGPHWDEROL]DEOHHQHUJ\ yr year

$OVRFDSLWDOL]HGZLWKDQ\FRPELQDWLRQHJP/ THE POULTRY SCIENCE ASSOCIATION

PATRONS Ajinomoto Heartland LLC Archer Daniels Midland Co. Cobb-Vantress Inc. Mosaic Feed Ingredients Novus International Inc. Tyson Foods Inc. Zinpro Corporation

SUSTAINING MEMBERS Adisseo USA Inc. Foster Farms Akey Inc. Huvepharma Inc. Aviagen Inc. Hy-Line International Cargill Animal Nutrition Land O’Lakes / Purina Feeds LLC Diamond V Maple Leaf Farms Inc.

2013 PSA Annual Meeting Sponsors Platinum Level Gold Level Silver Level Bronze Level DSM Nutritional Products AB Vista Feed Ingredients Ajinomoto Heartland LLC Akey Inc. Jones-Hamilton Co. Adisseo USA Inc. Alltech Inc. Chr. Hansen Inc. Novus International Inc. DuPont – Danisco Animal Archer Daniels Midland Co. Fieldale Farms Corporation Nutrition Aviagen Turkeys Inc. Kerry Ingredients & Flavors Evonik Degussa Corporation Biomin USA Inc. Midwest Poultry Federation Kemin Animal Nutrition & BioResource International Inc. Zoetis Health Cargill Animal Nutrition United Soybean Board Cobb-Vantress Inc. Diamond V Elanco Animal Health Enzyvia LLC Foster Farms Hendrix Genetics Huvepharma Inc. Jefo Lallemand Animal Nutrition Maple Leaf Farms Inc. Mosaic Feed Ingredients Prince Agri Products Inc. QualiTech Inc. Quality Technology International Inc. Southern Poultry Science Society Tyson Foods Inc. Vetagro S.p.A.. Zinpro Corporation