Prehatch Intestinal Maturation of Turkey Embryos Demonstrated Through Gene Expression Patterns

MOLECULAR, CELLULAR, AND DEVELOPMENTAL BIOLOGY

Prehatch intestinal maturation of turkey embryos demonstrated through gene expression patterns

J. E. de Oliveira ,*1 S. Druyan ,*2 Z. Uni ,† C. M. Ashwell ,* and P. R. Ferket *3

* North Carolina State University, Department of Poultry Science, Raleigh 27695; and † Hebrew University, Faculty of Agriculture, Department of Animal Science, Rehovot, Israel 76100

ABSTRACT Some of the challenges faced by neonatal by gene expression pattern similarity into 4 groups. turkeys include weakness, reduced feed intake, impaired The expression pattern of hormone receptors revealed growth, susceptibility to disease, and mortality. These that intestinal tissues may be less responsive to growth symptoms may be due to depleted energy reserves after hormone, insulin, glucagon, and triiodothyronine dur- hatch and an immature digestive system unable to re- ing the last 48 h before hatch, when developmental plenish energy reserves from consumed feed. To better emphasis switches from cell proliferation to functional understand enteric development in turkeys just before maturation. Based on gene expression patterns, we con- hatch, a new method was used to identify the patterns cluded that at hatch, poults should have the capacity of intestinal gene expression by utilizing a focused mi- to 1) digest disaccharides but not oligopeptides, due croarray. The duodenums of 24 turkey embryos were to increased expression of sucrase-isomaltase but de- sampled on embryonic day (E)20, E24, E26, and hatch creased expression of aminopeptidases and 2) absorb (E28). The RNA populations of 96 chosen genes were monosaccharides and small peptides due to high ex- measured at each time point, from which 81 signifi- pression of sodium-glucose cotransporter-4 and peptide cantly changed (P < 0.01). These genes were clustered transporter-1. Key words: turkey embryo , intestinal development , gene expression , microarray 2009 Poultry Science 88 :2600–2609 doi: 10.3382/ps.2008-00548

INTRODUCTION In order for poults to independently forage and digest food immediately after hatch, they must develop before Modern turkey strains have been selected for fast hatch the necessary complement of digestive enzymes growth, efficient feed conversion, and high meat yield and mucosal maturity with the absorptive capacity to (Havenstein et al., 2007). Moreover, the time to raise use ingested feed. This prenatal development of diges- a turkey to market size decreases each year, and thus tive capacity increases as the avian embryos orally neonatal growth and development become an increasing consume their amniotic fluid before hatch (Romanoff, proportion of the productive life of a turkey (Uni and 1960, 1967; Uni and Ferket, 2004; Moran, 2007). The Ferket, 2004). Some of the challenges neonatal turkeys nutrients within the amnion comprise the first meal must overcome include weakness due to low feed intake of the embryo and along with yolk infusion into the or feed refusal (Christensen et al., 2003). Poor digestion intestine (Esteban et al., 1991a,b; Noy et al., 1996), of feed and malabsorption increases the susceptibility they facilitate enteric development toward hatch (Uni of poults to opportunistic pathogens, resulting in im- et al., 2003), but the details of this development are not paired growth and high “starve out” mortality (Barnes, clearly understood. 1994a,b). During the first few days after hatch, turkey The perinatal development of digestive capacity can be poults often succumb to low energy status (Donaldson, evaluated by studying the activity and gene expression 1995), most likely due to an immature digestive sys- of digestive enzymes and nutrient transporters in the tem (Uni and Ferket, 2004; Christensen et al., 2007). intestine. Pancreatic enzymes carboxypeptidase A and chymotrypsin have been reported in chicken embryos at © 2009 Poultry Science Association Inc. 16 d of incubation or embryonic day (E)16 (Marchaim Received December 15, 2008. and Kulka, 1967). Activities of maltase, aminopepti- Accepted September 3, 2009. 1 Current address: Provimi Research and Innovation Centre, Len- dase (XPNPEP1), sodium-glucose cotransporter-1 neke Marelaan 2, Sint-Stevens-Woluwe, B1932, Belgium. (SGLT1), and ATPase ion channel (P2X1) were found 2 Current address: Institute of Animal Science, Agricultural Re- in chicken embryos on E19, and their mRNA expression search Organization, The Volcani Center, PO Box 6, Bet Dagan 50250, Israel. was detected as early as E15 (Uni et al., 2003). Activity 3 Corresponding author: [email protected] of the brush border enzymes leucine amino peptidase

2600 TURKEY EMBRYO INTESTINAL GENE EXPRESSION 2601 and sucrase-isomaltase, as well as jejunum glucose and North Carolina State University Institutional Animal alanine uptake, have been reported in turkey embryos Care and Use Committee. as early as E25 (Foye et al., 2007). Extensive studies on chicken intestinal enzymes and nutrient transporters Statistical Analyses from E18 to 14 d posthatch have been recently pub- lished (Gilbert et al., 2007; Frazier et al., 2008; Mott et In turkey embryos, differences between day of incu- al., 2008). However, there is limited knowledge about bation in YFBW, DYFBW, and duodenum weight data at which phases of embryonic development these genes were tested by 1-way ANOVA according to the follow- start to be expressed, especially in turkeys. A better ing model: Y = μ + Age + e [equation 1], where Y understanding of small intestine maturation and gene = weight; μ = mean; and e = random error, with the expression pattern before hatch may help turkey nutri- fixed effect of age (E20, E22, E24, E26, or E28). All of tionists formulate diets that complement the digestive the statistical analyses were conducted using the JMP capacity at hatch. Genes of interest must include diges- software (SAS Institute, 2005). tive brush border enzymes for carbohydrate, protein and lipid, nutrient transporters, and hormone receptors RNA Isolation, Sample Labeling, involved in intestinal tissue growth, maturation, and and Microarray Procedure function. High-throughput DNA microarrays are able to identify many RNA populations in one single experi- Total RNA was extracted from 100 mg of duodenum ment (Harris, 2000; Spielbauer and Stahl, 2005), quali- tissue using TRI Reagent (Molecular Research Center, fying it as a good tool for the task of identifying the Cincinnati, OH). Equal amounts of RNA from 6 em- pattern of expression of these genes. Even though turkey bryos were pooled and adjusted to 0.5 µg/µL of concen- gene sequences are still not available for much of the tration. Complementary DNA was produced from each genome, it has been demonstrated that gene sequences RNA pool and indirect-labeled either with Cy3 or Cy5 in chickens and turkeys are similar enough that chicken utilizing the ChipShot Indirect Labeling and Clean-up probes can hybridize to turkey target RNA (Reed et System kit (Promega, Madison, WI), according to the al., 2005). More recent data collected for comparative experimental design (Figure 1). Thirty-three picomoles genome organization studies to characterize the physi- of cDNA labeled with each Cy-Dye (Amersham Biosci- cal map for the turkey (Romanov and Dodgson, 2006) ences Corp., Piscataway, NJ) was hybridized in each ar- and the distribution of copy number variants across ray slide for 16 h at 42°C, then was washed and scanned species (chicken:turkey; Griffin et al., 2008) show the to generate digitized image data files. Microarray slides utility of these cross-species hybridizations for studying were produced by the North Carolina State University the turkey genome. The objective of this study was to Domestic Animals Genomic Laboratory. The microar- use microarray technology to map small intestine gene ray platform used was NCSU_Chicken_JS1, registered expression in turkey embryos during the last phase of at the National Center for Biotechnology Informa- incubation from 20 d of incubation until day of hatch. tion Gene Expression Omnibus (GEO) (http://ncbi. nlm.nih.gov/projects/geo) under accession number GPL6041. This is an earlier version of the microarray MATERIALS AND METHODS platform introduced by Druyan et al. (2008), contain- Egg Incubation and Tissue Sampling ing 96 probes. The most important feature of this array is the high number of replication because each gene se- Two hundred Nicholas 300 turkey eggs were ob- quence was spotted 8 times in each slide, and combined tained from a commercial turkey operation (Prestage biological and technical replicates totaled 32 observa- Farms, Clinton, NC) and were incubated under stan- tions per gene. An additional feature of these studies is dard conditions. At E19, eggs were candled, infertile the cross-species hybridization of turkey cDNA with the eggs and dead embryos were removed, and the remain- chicken sequence-based arrays. To validate the utility of ing eggs were divided into 5 groups of 24 eggs with this platform for turkey studies, we performed a bioin- similar weight distribution (74.3 ± 10 g). One group of formatic analysis (BLASTN via http://blast.ncbi.nlm. 24 eggs was sampled at E20, E22, E24, E26, and E28 nih.gov/Blast.cgi) of the chicken 70-mer oligos printed (hatch). The eggs were opened at the blunt end using on the arrays to the available gene sequences for tur- surgical scissors; the embryo was extracted and killed key. In addition, a comparative hybridization of pooled by cervical dislocation. Embryos were weighed with- RNA from liver and intestine (1:1) was made across out yolk sac to calculate yolk-free BW (YFBW). The species with samples matched by collection at hatch- duodenum of each embryo was dissected, weighed, and ing following the labeling, hybridization, detection, and immersed in RNA Later (Applied Biosystems-Ambion, normalization procedures of Druyan et al. (2008). Foster City, CA) according to the protocol of the manu- facturer. Duodenum weight expressed as a percentage Microarray Data Analysis of embryo YFBW (DYFBW) was also calculated. All experimental procedures followed established guidelines The experimental design was a complete interwoven for animal care and handling and were approved by the loop design (Garosi et al., 2005), with 4 biological rep- 2602 de Oliveira et al. Table 1. Turkey embryo weights and tissue weights in the period between embryonic day (E)20 and E28 (hatch)

Embryonic age (d) YFBW1 (g) Duodenum (g) DYFBW2 (%) E20 22.38e 0.045d 0.202c E22 29.23d 0.062c 0.213c E24 37.89c 0.061c 0.161d E26 41.40b 0.102b 0.246b E28 48.36a 0.223a 0.463a P-value <0.0001 <0.0001 <0.0001

a–eMeans followed by different letters within the same column are statistically different. 1YFBW = embryo weight expressed as yolk-free BW. 2DYFBW = duodenum weight expressed as percentage of embryo yolk-free BW. licates for each time point (2 labeled with each Cy-Dye; time. The data discussed in this publication have been Figure 1). Raw data were transformed to a log2 base and deposited in the National Center for Biotechnology In- analyzed using JMP Genomics (SAS Institute, 2007), formation GEO (Edgar et al., 2002) and are accessible using the normalization method previously proposed through GEO Series accession number GSE9474. by Wolfinger et al. (2001) and modified by Druyan et al. (2008). This method is best to analyze data with Pathway Map complex experimental designs and a minimum of 3 replicates per spot (Kerr et al., 2000). The residuals from To summarize data concerning all main gene expres- this model were analyzed by mixed ANOVA according sion discussed in this paper, a pathway map was cre- to the following gene-specific model: Y = µ + E + Dye ated using the MapEditor tool of GeneGo (GeneGo + Hyb + Slide + e [equation 2], where Y = hybridiza- Inc., St Joseph, MI). GeneGo is a Web-based systems tion intensity; µ = mean intensity; and e = random biology application. error, with day of incubation (E) and Cy-Dye (Dye) as fixed effects and hybridization batch (Hyb) and Slide RESULTS AND DISCUSSION as random effects. Mean intensities were compared using false discovery rate at P < 0.01. Genes were clus- Embryonic Tissue Weights tered by expression pattern similarity using hierarchical clustering, with distances between clusters defined The measured YFBW of the embryos, duodenum by the Ward method (SAS Institute, 2007), with time weight, and calculated DYFBW during the studied pe- as a nominal parameter (distances = 0). Standardized riod (E20 to E28) are presented in Table 1. Embry- least squares means of each gene were used to produce onic YFBW increased daily from E20 until hatch (E28; parallel plots presenting a pattern of expression over Table 1). The observed linear pattern of increasing embryonic weight is supported by previous studies done with chickens and turkeys (Hamburger and Hamilton, 1951; Mun and Kosin, 1960; Romanoff, 1960; Tona et al., 2005). Duodenum weight also increased during the same period, except between E22 and E24, whereas DYFBW showed smaller changes between E20 and E26 followed by aggressive growth between E26 and E28 when duodenum weight and DYFBW doubled in the 48 h before hatch (Table 1). This change indicates that a great portion of embryonic investment in intestinal tissue is concentrated during the last 48 h before hatching, probably in preparation for posthatch digestion of nutrients.

Cross-Species Characterization Figure 1. Array hybridization scheme for 2-color labeling. Each of the Array Platform sampling time point is shown as embryonic day (E)20, E22, E24, E26, and E28 (hatch). Each array is represented by an individual arrow To assess the degree of similarity of the sequences (10 total). The sample groups located at each end of the respective used to construct the microarrays and the sequence of arrow were hybridized on the same array. Samples at the tail of the ar- rows were hybridized with Cy3 (black) and samples at the arrowhead the turkey genome, a comparison was made (BLAST). were labeled with Cy5 (white). Samples were combined according to a The range of nucleotide conservation between the 2 complete interwoven loop design, resulting in 4 measurements for each species ranged from 91 to 100% identity, with the aver- time point. Each of these measures corresponded to a pooled sample from 6 individual embryos and was labeled with either Cy3 or Cy5 as age identity being 96%. These comparisons were made shown. where sequences were available for the turkey (78 of the TURKEY EMBRYO INTESTINAL GENE EXPRESSION 2603

Figure 2. Cross-species analysis of the microarray platform. The level of nucleotide identity of the chicken oligo sequences (70-mers) with available turkey sequence (GenBank) is shown in panel A. Panel B shows the correlation of the normalized hybridization intensities for each gene (oligo) across species. Gene expression as measured by the microarray was highly correlated across species (R2 = 0.9875). Color version available in the online PDF.

92 genes included in the microarray). The distribution hormone, and thyroid hormones). A general discussion of homology of the chicken oligo sequences with avail- on general gene expression will be presented first. able turkey sequences is shown in Figure 2A. Overall Gene Expression. Generally, gene expres- Upon comparing the hybridization intensities of sion relates to the amount of mRNA present in a tissue genes expressed in hatching chicken liver-duodenum to at a certain time. It can be illustrated by changes in ex- that of hatching turkey liver-duodenum to the array, a pression of a housekeeping gene, which should be ubiq- highly significant correlation was observed across spe- uitously expressed at a common level across tissues and cies. In Figure 2B, the normalized hybridization inten- developmental times. Glyceraldehyde-3-phosphate de- sities for each gene (oligo) on the array were compared hydrogenase (GAPDH) is commonly used as a house- across species. The correlation of observed expression keeping gene for gene expression studies. The expres- was very high (R2 = 0.9875), but the slope of the re- sion of this gene was significantly affected by embryonic gression line was not unity (1). There was a reduction age as presented in Figure 5. In the same period, the in hybridization signal from the turkey samples as com- total genomic DNA detected on the microarray showed pared with the chicken (a slope of 0.93 when compar- an inverse pattern in comparison to GAPDH (Figure ing turkey:chicken), which suggests that only a small 5). This inverse response may indicate that there is a percentage of the expression in the turkey was being developmental emphasis toward cell proliferation (pe- missed by the array. Based on these data, we concluded riod between E22 and E26), represented by increased that the chicken-based oligo array used in these studies genomic DNA, and reduction in cell proliferation and was suitable for use in the turkey, with the understand- shift to cell hypertrophy (period between E26 and E28), ing that the absolute expression levels may be slightly represented by increased expression of GAPDH. Evi- underestimated but the relative expression levels would dently, the great increase in duodenum weight observed provide reliable and accurate data as described previ- between E26 and E28 (Table 1) may be mostly due to ously for the chicken (Druyan et al., 2008). increases in cell size and maturation rather than cell number. A similar increase in GAPDH gene expression Intestinal Gene Expression between the time of pipping and hatch has also been reported in chickens (Gilbert et al., 2007). The results of cluster analysis are presented in Fig- Brush Border Enzymes. Expression of the sucrase- ure 3. There were a total of 81 genes affected by age isomaltase (SI) gene (SUIS in Figure 6) follows the throughout the study. To facilitate understanding, the pattern described in cluster C (Figures 3 and 4), which number of clusters was set to 4 and discriminated by shows increasing expression after E24, with peak ex- capital letters A, B, C, and D based on similarity of pression between E25 and E26. This observation agrees expression patterns. The plots with the pattern of gene with those reported by Foye et al. (2007), who found SI expression representing each cluster are presented in activity in turkey embryos at E25. Because very small Figure 4. Many genes involved in intermediary me- amounts of carbohydrates are expected to be present tabolism, especially energy metabolism, had expres- in the intestine during the last period of incubation, sion changing with embryonic age. For the purpose of this increase in SI gene expression before hatch may this paper, we will focus on discussing the expression be considered part of the normal maturation process patterns presented as parallel plots of genes involved of small intestinal epithelium (Figures 3 and 4). Both in nutrient digestion (carbohydrate, protein, and fat), peptidase genes, XPNPEP1 and aminopeptidase Ey nutrient transporters (glucose, amino acids, and fatty (ANPEP) follow a pattern of decreasing expression to- acids), and hormone receptors linked to tissue matu- ward hatch as described in cluster B (Figures 3 and ration and nutrient uptake (insulin, glucagon, growth 4). Downregulation of these genes can be explained 2604 de Oliveira et al. Table 2. List of genes with respective gene symbols (abbreviations)

Gene symbol Gene name ACAA Acetyl-coenzyme A acyltransferase ACACA Acetyl-coenzyme A carboxylase ACADS Butyryl-coenzyme A dehydrogenase AK1 Adenylate kinase ALDH2 Aldehyde dehydrogenase ALDOB Fructose-biphosphate aldolase B ALDOC Fructose-biphosphate aldolase C AMYP α-Amylase ANPEP Aminopeptidase Ey atCAB2 atCAB2-arapdopsis BPEF1 Protein phosphatase 1 CAMK2G CaM kinase II gamma chBACT chBACT chEF2 chEF2-elongation factor 2 CKM Creatine kinase CPT2 Carnitine O-palmitoyltransferase CRAT Carnitine acetylase CS Adenosine triphosphate citrate synthase DIO1 Thyroxine deiodinase I ECHDC1 3-Hydroxyacyl-coenzyme A dehydrogenase FBP1 Fructose 1,6 biphosphatase (F1,6 BPase) G3P2 Glycerol-3-phosphate dehydrogenase G6PT Glucose-6-phosphatase (G-6-Pase) GAB3 Aconitate hydratase GALT Galactose-1-phosphate uridyltransferase GAPDH Glyceraldehyde 3-phosphate dehydrogenase GBE1 Glycogen branching enzyme GCGR Glucagon receptor GCK Glucokinase GHR Growth hormone receptor GHRH Growth hormone-releasing hormone GHRHR Growth hormone-releasing hormone precursor GLUD1 Glutamate dehydrogenase GLUL Glutamine synthetase GPI Glucose-6-phosphate isomerase GSK3B Glycogen synthase kinase-3 β GYS1 Glycogen synthase 1 (muscle) HAGH Hydroxyacylglutathione hydrolase HK2 Hexokinase II HMDH 3-Hydroxy-3-methylglutaryl-coenzyme A reductase HXK1 Hexokinase 1 IDH2 Isocitrate dehydrogenase (NADP+) IDH3A Isocitrate dehydrogenase (NAD+) IGF2R Insulin-like growth factor-2 receptor INSR Insulin receptor K6PL 6-Phosphofructokinase (PFK-1) LDHA l-Lactate dehydrogenase LDHD d-Lactate dehydrogenase LPL Lipoprotein lipase LYRM5 Growth hormone-inducible soluble protein MDH1 Malate dehydrogenase ME1 Malic enzyme ME2 Malate dehydrogenase MGAM Maltase-glucoamylase NUDT7 Acetyl coenzyme A hydrolase OGDH Oxoglutarate dehydrogenase P2X1 Adenosine triphosphate-gated ion channel receptor PC Pyruvate carboxylase PDHB Pyruvate dehydrogenase PEPT1 Peptide transporter PFKL Phosphofructokinase (PFK-1) PFKM 6-Phosphofructokinase (PFK-1) PGD Phosphogluconate dehydrogenase PGM1 Phosphoglucomutase PHKB Phosphorylase kinase B α regulatory chain PKM2 Pyruvate kinase PYGL Glycogen phosphorylase PYGM Glycogen phosphorylase (muscle) SDHB Succinate dehydrogenase SI/SUIS Sucrase-isomaltase intestinal SLC5A9 Solute carrier family 5 (sodium-glucose cotransporter), member 9 SUCLG1 Succinate coenzyme A ligase TALDO1 Transaldolase

Continued TURKEY EMBRYO INTESTINAL GENE EXPRESSION 2605 Table 2 (Continued). List of genes with respective gene symbols (abbreviations)

Gene symbol Gene name TKT Transketolase TR-α Tyroid hormone receptor α TR-β1 Tyroid hormone receptor β-1 Turkey DNA Turkey genomic DNA TYR Tyrosinase UGDH Uridine diphosphate-glucose 6-dehydrogenase UGP2 Uridine triphosphate-glucose 1-phosphate uridyltransferase XPNPEP1 Aminopeptidase P1 by the decrease in protein substrates in the intestine after amniotic fluid consumption is completed around E26 (De Oliveira et al., 2008). Uni et al. (2003) found XPNPEP1 enzyme activity in chickens at E19 (which according to De Oliveira et al. (2008) corresponds to turkey embryo E24) during the peak of amniotic fluid consumption by the embryo. This response may also indicate that protein digestive enzymes are dependent upon the presence of substrate to stimulate gene expression, whereas carbohydrate digestive enzymes are expressed as a natural part of the enterocyte maturation process. The lipoprotein lipase (LPL) enzyme gene expression also followed the pattern in cluster B (Fig- ures 3 and 4), also decreasing expression toward hatch. This may be a consequence of reduced lipid metabolism during the hatching period because general metabolism switches to gluconeogenesis (Donaldson, 1995; Moran 2007; De Oliveira et al., 2008). The importance of gluconeogenesis during pipping and hatching, even in the small intestine, is corroborated by increased expression of l-lactate dehydrogenase (LDHD) and fructose 1,6 biphosphatase (FBP1) (gluconeogenic enzymes) in cluster C (Figures 3 and 4), and the downregulation of glycolytic enzymes hexokinase 1 (HXK1), 6-phosphofructokinase (PFKM), and pyruvate kinase (PKM2). The expression pattern of the α-amylase (AMYP) gene matches cluster A (Figures 3 and 4). The α-amylase (AMYP) gene has been previously reported as only expressed in the pancreas; therefore, conclusions will not be drawn for this gene because its presence may represent some intestinal expression of this gene, which needs to be confirmed. Another possible explanation would be that duodenum samples contained traces of pancreas that could not be removed during dissection because both tissues are closely associated. Brush Border Nutrient Transporters. The gene expression patterns of 2 key nutrient transporters, short peptide transporter 1 (PEPT1) and sodium-glucose cotransporter 4 (SLC5A9), were both included in cluster D (Figure 3) and they were upregulated at hatch (E28; Figure 4). The expression patterns of PEPT1 in turkey embryos during the period between E23 and hatch have been previously reported by Van et al. (2005), where they showed an almost identical pattern Figure 3. Hierarchical clustering of genes significantly affected by to the one found in this present work (Figure 4, cluster age in turkey embryo duodenum in the period between embryonic day D). This agreement in PEPT1 gene expression response (E)20 and E28 (hatch). Genes with similar expression patterns were grouped into 4 clusters identified as clusters A, B, C, and D. The list demonstrates that the focused-array technique used in of gene symbols and their correspondent names is provided in Table 2. our study is as precise as the Northern blot technique Color version available in the online PDF. 2606 de Oliveira et al.

Figure 4. Gene expression patterns identified in turkey embryo duodenum in the period between embryonic day (E)20 and E28 (hatch). Clus- ters A, B, C, and D correspond to the same clusters found in Figure 3. The list of genes represented by each pattern is found in the respective cluster in Figure 3. used by those authors, but with the advantage of high- ing this period is P2X1, which is found in the decreas- er throughput. According to Gilbert et al. (2007), the ing pattern of cluster B (Figure 3). The expression of Na+-dependent glucose transporter (SGLT) present on the P2X1 gene decreased toward hatch (Figure 4); thus, chromosome 15 in chicken (GenBank accession number active transport (at expense of adenosine triphosphate) XM_415247) was the most similar to human SGLT1, of ions (Ca2+, K+, and Na+) may be reduced during the whereas the one on chromosome 8 (GenBank accession number XM_422459) was similar to SGLT5. These authors found expression of both SGLT in chicken small intestine. Currently, the XM_422459 record has been removed from GenBank due to annotation updates. The SGLT isoform present on chromosome 8 (Gen- Bank accession number XM_422460) as used in this study (SLC5A9) is similar to human SGLT4, capable of transporting d-glucose, d-manose, and d-fructose and it is expressed in small intestine, kidney, and liver (http://www.uniprot.org/uniprot/Q2M3M2). Based on the expression patterns of transporters, PEPT1 and SLC5A9, small peptides and glucose could potentially be transported from intestinal lumen to the gut mucosa around the time of hatching, independent of enzymatic digestion. Similar results were reported for many other nutrient transporters in chickens around time of hatch (Gilbert et al., 2007; Frazier et al., 2008; Mott et al., 2008). Based on these results, we recom- mend that poults should be fed highly digestible nutri- Figure 5. Expression patterns of glyceraldehyde-3-phosphate dehy- ents at hatch because even though digestive capacity drogenase (GAPDH) and measured genomic DNA in turkey duodenum in the period between embryonic day (E)20 and E28 (hatch). Opposing may be limited, the nutrient transport potential is pres- patterns starting on E22 until E28 illustrate change in duodenum tis- ent. Another membrane transporter gene affected dur- sue from increasing cell number to hypertrophy. TURKEY EMBRYO INTESTINAL GENE EXPRESSION 2607

Figure 6. Schematic map of key genes involved in enterocyte function. Genes representing nutrient digestion, absorption, and hormonal control in the intestinal epithelium are connected according to MetaCore software algorithms (GeneGo Inc., St. Joseph, MI). Thermometer level repre- sents the expression level of the correspondent gene, red meaning upregulation and blue meaning downregulation, whereas the numbers 1, 2, 3, 4, and 5 correspond to turkey embryonic day (E)20, E22, E24, E26, and E28 (hatch), respectively. A complete list of gene symbols is provided in Table 2, and a reference list of object symbols is provided in Appendix Figure 1a. *In this figure, the SLC5A1 symbol was used to represent the SLC5A9 gene. Color version available in the online PDF. period between completion of amniotic fluid consump- 1995; Canfield and Kornfeld, 1989), the observed ex- tion and hatch (E26 to E28). pression pattern of IGF2R found in this present study Hormone Receptors. The genes for growth hormone matches the concentration pattern of IGF-II in embry- receptor (GHR), insulin-like growth factor-2 receptor onic fluids reported by Karcher et al. (2008; Figure 4, (IGF2R), insulin receptor (INSR), glucagon receptor cluster B). Although TR-α gene expression decreased, (GCGR), and thyroid receptor α (TR-α) were all pres- the other receptor variant, thyroid receptor β-1 (TR- ent in cluster B (Figure 3), indicating downregulation β1), peaked in expression at hatch (Figures 3 and 4, toward hatch (Figure 4). This observation indicates an cluster D). It is generally accepted that only the α vari- overall reduction in tissue response to growth hormone, ant of thyroid receptors plays a major role in intestinal insulin, glucagon, and triiodothyronine. Insulin-like epithelial cell proliferation during development (Plat- growth factor (IGF) is very important for avian devel- eroti et al., 2001, 2006); therefore, TR-β1 may be in- opment (Karcher et al., 2008). Both IGF-I and IGF-II volved in other processes such as gut maturation. These bind to IGF-I receptor (IGFIR) to play their role in findings give further support to our earlier observation signal transduction (Zhou et al., 1995), and IGF-II also about the inverse relationship between GAPDH gene binds to INSR (Korner et al., 1995). Karcher et al. expression and genomic DNA (Figure 5), indicating less (2008) observed that IGF-II concentrations in the am- cell proliferation and more tissue maturation toward niotic and allantoic fluids of turkey embryos declined hatch. dramatically throughout development (Karcher et al., Successful utilization of a focused microarray tech- 2005). Even though it has been reported that IGF-II nique made it possible to determine the expression does not bind to IGF2R in avian species (Zhou et al., pattern of a large number of genes in turkey embryo 2608 de Oliveira et al. duodenum. The summary expression of the key genes Foye, O. T., P. R. Ferket, and Z. Uni. 2007. The effects of in ovo observed in this present study can be seen in Figure feeding arginine, β-hydroxy-β-methyl-butyrate, and protein on jejunal digestive and absorptive activity in embryonic and neona- 6. The top of the map suggests that at hatch (ther- tal turkey poults. Poult. Sci. 86:2343–2349. mometer 5) the membrane can absorb monosaccharides Frazier, S., K. Ajiboye, A. Olds, T. Wyatt, E. S. Luetkemeier, and (SLC5A9) and peptides (PEPT1), whereas oligopeptide E. A. Wong. 2008. Functional characterization of the chicken digestion (aminopeptidase P1) and adenosine triphos- peptide transporter 1 (pept1, slc15a1) gene. Anim. Biotechnol. 19:201–210. phate-dependent ion transport (P2X1) may be limited Garosi, P., D. De Filippo, M. van Erk, P. Rocca-Serra, S. A. San- (Figure 6). The expression of hormone receptors (TR- sone, and R. Elliott. 2005. Defining best practice for microarray α, GCGR, GHR, INSR) in the duodenum mucosa and analyses in nutrigenomic studies. Br. J. Nutr. 93:425–432. Gilbert, E. R., H. Li, D. A. Emmerson, K. E. Webb Jr., and E. A. lipid mobilization (LPL) also are reduced at hatch (Fig- Wong. 2007. Developmental regulation of nutrient transporter ure 6, thermometer 5). Evidently, enterocyte cell divi- and enzyme mRNA abundance in the small intestine of broilers. sion in the small intestine of the turkey embryo slows Poult. Sci. 86:1739–1753. down as it approaches hatch, and enterocyte matura- Griffin, D. K., L. B. Robertson, H. G. Tempest, A. Vignal, V. Fillon, R. P. Crooijmans, M. A. Groenen, S. Deryusheva, E. Gaginskaya, tion and function peaks to accommodate the assimila- W. Carré, D. Waddington, R. Talbot, M. Völker, J. S. Masa- tion of glucose, simple sugars (isomaltose and sucrose), banda, and D. W. Burt. 2008. Whole genome comparative stud- and small peptides, whereas enzymes associated with ies between chicken and turkey and their implications for avian energy metabolism still favor gluconeogenesis. Further genome evolution. BMC Genomics 9:168. Hamburger, V., and H. L. Hamilton. 1951. A series of normal stages studies must be conducted to verify if in ovo feeding of in the development of the chick embryo. J. Morphol. 88:49–92. prestarter or starter diets formulated based on these Harris, E. D. 2000. Differential PCR and DNA microarrays: The observations could change or influence turkey neonatal modern era of nutritional investigations. Nutrition 16:714–715. performance. Havenstein, G. B., P. R. Ferket, J. L. Grimes, M. A. Qureshi, and K. E. Nestor. 2007. Comparison of the performance of 1966- versus 2003-type turkeys when fed representative 1966 and 2003 turkey diets: Growth rate, livability, and feed conversion. Poult. Sci. ACKNOWLEDGMENTS 86:232–240. Karcher, D. M., J. N. Fleming-Waddell, and T. J. Applegate. 2008. Developmental changes in insulin-like growth factor (IGF)-I and This research was supported by Embrex Inc. (Dur- -II mRNA abundance in extra-embryonic membranes and small ham, NC) through the USDA Small Business Innova- intestine of avian embryos. Growth Horm. IGF Res. 19:31–42. tion Research (SBIR) Phase II grant 2005-33610-16491, Karcher, D. M., J. P. McMurtry, and T. J. Applegate. 2005. De- September 15, 2005 to September 14, 2007. velopmental changes in amniotic and allatoic fluid insulin-like growth factor (IGF)-I and II concentrations of avian embryos. Comp. Biochem. Physiol. A Mol. Integr. Physiol. 142:404–409. Kerr, M. K., M. Martin, and G. A. Churchill. 2000. Analysis of variance for gene expression microarray data. J. Comput. Biol. REFERENCES 7:819–837. Korner, C., B. Nurnberg, M. Uhde, and T. Braulke. 1995. Mannose Barnes, H. J. 1994a. Clues to possible control of poult growth de- 6-phosphate/insulin-like growth factor II receptor fails to interact pression. Turkey World 70:24–26. with G-proteins. J. Biol. Chem. 270:287–295. Barnes, H. J. 1994b. Poult growth depression costs industry big Marchaim, U., and R. G. Kulka. 1967. The non-parallel increase of bucks. Turkey World 70:12–14. amylase, chymotrypsinogen and procarboxypeptidase in the de- Canfield, W. M., and S. Kornfeld. 1989. The chicken liver cation- veloping chick pancreas. Biochim. Biophys. Acta 146:553–559. independent mannose 6-phosphate receptor lacks the high affin- Moran, E. T. Jr. 2007. Nutrition of the developing embryo and ity binding site for insulin-like growth factor II. J. Biol. Chem. hatchling. Poult. Sci. 86:1043–1049. 264:7100–7103. Mott, C. R., P. B. Siegel, K. E. Webb Jr., and E. A. Wong. 2008. Christensen, V. L., D. T. Ort, and J. L. Grimes. 2003. Physiologi- Gene expression of nutrient transporters in the small intestine of cal factors associated with weak neonatal poults (Meleagris gal- chickens from lines divergently selected for high or low juvenile lopavo). Int. J. Poult. Sci. 2:7–14. body weight. Poult. Sci. 87:2215–2224. Christensen, V. L., D. T. Ort, K. E. Nestor, S. G. Velleman, and Mun, A. M., and I. L. Kosin. 1960. Development of the broad breast- G. B. Havenstein. 2007. Genetic control of neonatal growth and ed bronze turkey embryo. Biol. Bull. 119:90–97. intestinal maturation in turkeys. Poult. Sci. 86:476–487. Noy, Y., Z. Uni, and D. Sklan. 1996. Routes of yolk utilisation in the De Oliveira, J. E., Z. Uni, and P. R. Ferket. 2008. Important meta- newly-hatched chick. Br. Poult. Sci. 37:987–995. bolic pathways in poultry embryos prior to hatch. World’s Poult. Plateroti, M., K. Gauthier, C. Domon-Dell, J.-N. Freund, J. Sa- Sci. J. 64:488–499. marut, and O. Chassande. 2001. Functional interference between Donaldson, W. E. 1995. Carbohydrate, hatchery stressors affect thyroid hormone receptor α (TRα) and natural truncated TRΔα poult survival. Feedstuffs 67:16–17. isoforms in the control of intestine development. Mol. Cell. Biol. Druyan, S., J. E. de Oliveira, and C. M. Ashwell. 2008. Focused mi- 21:4761–4772. croarrays as a method to evaluate subtle changes in gene expres- Plateroti, M., E. Kress, J. I. Mori, and J. Samarut. 2006. Thy- sion. Poult. Sci. 87:2418–2429. roid hormone receptor α1 directly controls transcription of the Edgar, R., M. Domrachev, and A. E. Lash. 2002. Gene Expression β-catenin gene in intestinal epithelial cells. Mol. Cell. Biol. Omnibus: NCBI gene expression and hybridization array data 26:3204–3214. repository. Nucleic Acids Res. 30:207–210. Reed, K. M., T. P. Knutson, S. B. Krueth, L. R. Sullivan, and L. D. Esteban, S., M. Moreno, I. Mestre, J. M. Planas, and J. A. Tur. Chaves. 2005. In silco mapping of ESTs from turkey (Meleagris 1991a. Regional development of α-methyl-d-glucoside transport gallopavo). Anim. Biotechnol. 16:81–102. in the small intestine of chick embryos and newly-hatched chicks. Romanoff, A. L. 1960. The Avian Embryo: Structural and Func- Arch. Int. Physiol. Biochim. Biophys. 99:425–428. tional Development. Macmillian, New York, NY. Esteban, S., M. Moreno, J. M. Rayo, and J. A. Tur. 1991b. Gastro- Romanoff, A. L. 1967. Biochemistry of the Avian Embryo: A Quan- intestinal emptying in the final days of incubation of the chick titative Analysis of Prenatal Development. Interscience Publish- embryo. Br. Poult. Sci. 32:279–284. ers, New York, NY. TURKEY EMBRYO INTESTINAL GENE EXPRESSION 2609

Romanov, M. N., and J. B. Dodgson. 2006. Cross-species overgo Uni, Z., E. Tako, O. Gal-Garber, and D. Sklan. 2003. Morphological, hybridization and comparative physical mapping within avian molecular, and functional changes in the chicken small intestine genomes. Anim. Genet. 37:397–399. of the late-term embryo. Poult. Sci. 82:1747–1754. SAS Institute. 2005. JMP Version 7.0 User’s Guide. SAS Institute Van, L., Y.-X. Pan, J. R. Bloomquist, K. E. Webb Jr., and E. A. Inc., Cary, NC. Wong. 2005. Development regulation of a turkey intestinal pep- SAS Institute. 2007. JMP Genomics Version 3.0 User’s Guide. SAS tide transporter (PepT1). Poult. Sci. 84:75–82. Institute Inc., Cary, NC. Wolfinger, R. D., G. Gibson, E. D. Wolfinger, L. Bennett, H. Hama- Spielbauer, B., and F. Stahl. 2005. Impact of microarray technology deh, P. Bushel, C. Afshari, and R. S. Paules. 2001. Assessing gene in nutrition and food research. Mol. Nutr. Food Res. 49:908– significance from cDNA microarray expression data via mixed 917. models. J. Comput. Biol. 8:625–637. Tona, K., O. Onagbesan, V. Bruggeman, K. Mertens, and E. Decuy- Zhou, M., Z. Ma, and W. S. Sly. 1995. Cloning and expression of pere. 2005. Effects of turning during incubation on embryo the cDNA of chicken cation-independent mannose-6-phosphate growth, utilization of albumen, and stress regulation. Poult. Sci. receptor. Proc. Natl. Acad. Sci. USA 92:9762–9766. 84:315–320. Uni, Z., and P. R. Ferket. 2004. Methods for early nutrition and their potential. World’s Poult. Sci. J. 60:101–111.

APPENDIX

Figure 1a. Reference list with detailed description of symbols used in pathway map in Figure 6. Color version available in the online PDF.