Biomedical Sciences Publications Biomedical Sciences

8-2013 Effects of subtherapeutic concentrations of antimicrobials on gene acquisition events in Yersinia, Proteus, Shigella, and Salmonella recipient organisms in isolated ligated intestinal loops of swine Matt .T Brewer Iowa State University, [email protected]

Nalee Xiong Iowa State University, [email protected]

Kristi L. Anderson Iowa State University, [email protected]

SFtoevellow A thi. Cas ralsonnd additional works at: http://lib.dr.iastate.edu/bms_pubs IowaP Satrate of U ntheiversitAynim, stevale [email protected] duCommons, and the Veterinary Pathology and Pathobiology Commons The ompc lete bibliographic information for this item can be found at http://lib.dr.iastate.edu/ bms_pubs/36. For information on how to cite this item, please visit http://lib.dr.iastate.edu/ howtocite.html.

This Article is brought to you for free and open access by the Biomedical Sciences at Iowa State University Digital Repository. It has been accepted for inclusion in Biomedical Sciences Publications by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. Effects of subtherapeutic concentrations of antimicrobials on gene acquisition events in Yersinia, Proteus, Shigella, and Salmonella recipient organisms in isolated ligated intestinal loops of swine

Abstract Objective—To assess antimicrobial resistance and transfer of virulence genes facilitated by subtherapeutic concentrations of antimicrobials in swine intestines.

Animals—20 anesthetized pigs experimentally inoculated with donor and recipient bacteria.

Procedures—4 recipient pathogenic bacteria (Salmonella enterica serotype Typhimurium, Yersinia enterocolitica, Shigella flexneri, or Proteus mirabilis) were incubated with donor bacteria in the presence of subinhibitory concentrations of 1 of 16 antimicrobials in isolated ligated intestinal loops in swine. Donor Escherichia coli contained transferrable antimicrobial resistance or virulence genes. After coincubations, intestinal contents were removed and assessed for pathogens that acquired new antimicrobial resistance or virulence genes following exposure to the subtherapeutic concentrations of antimicrobials.

Results—3 antimicrobials (apramycin, lincomycin, and neomycin) enhanced transfer of an antimicrobial resistance plasmid from commensal E coli organisms to Yersinia and Proteus organisms, whereas 7 antimicrobials (florfenicol, hygromycin, penicillin G, roxarsone, sulfamethazine, tetracycline, and tylosin) exacerbated transfer of an integron (Salmonella genomic island 1) from Salmonella organisms to Yersinia organisms. Sulfamethazine induced the transfer of Salmonella pathogenicity island 1 from pathogenic to nonpathogenic Salmonella organisms. Six antimicrobials (bacitracin, carbadox, erythromycin, sulfathiazole, tiamulin, and virginiamycin) did not mediate any transfer events. Sulfamethazine was the only antimicrobial implicated in 2 types of transfer events.

Conclusions and Clinical Relevance—10 of 16 antimicrobials at subinhibitory or subtherapeutic concentrations augmented specific na timicrobial resistance or transfer of virulence genes into pathogenic bacteria in isolated intestinal loops in swine. Use of subtherapeutic antimicrobials in animal feed may be associated with unwanted collateral effects.

Disciplines Animal Diseases | Veterinary Pathology and Pathobiology

Comments This article is from American Journal of Veterinary Research 74 (2013): 1078–1083, doi:10.2460/ ajvr.74.8.1078. Posted with permission.

This article is available at Iowa State University Digital Repository: http://lib.dr.iastate.edu/bms_pubs/36 This material has been provided by the publisher for your convenience. It may not be further reproduced in any manner, includi ng (but not lim ited to) reprinting, photoco pying, electronic storage or transmission, or uploading onto the Internet. It may not be redistributed, amended, or overp rinted. Reproduction of this material without permissi on of the publisher violates federal law and is punishable under Title 17 of the United States Code (Copyright Act) and various intern ational treaties. Reprints or permission to reprint may be ordered by contacting [email protected] . Effects of subtherapeutic concentrations of antimicrobials on gene acquisition events in Yersinia, Proteus, Shigella, and Salmonella recipient organisms in isolated ligated intestinal loops of swine

Matt T. Brewer, DVM; Nalee Xiong, Ph D; Kristi L. Anderson, DVM; Steve A. Carlson, DVM, Ph D

Objective-To assess antimicrobial resistance and transfer of virulence genes faci litated by subthe rapeutic concentrations of antimicrobials in swine intestines. Animals-20 anesthetized pigs experimentally inoculated with donor and recipient bacteria. Procedures-4 recipient pathogenic bacteria (Salmonella enterica serotype Typhim urium, Yersinia enterocolitica, Shigella flexneri, or Proteus mirabilis) were incubated w ith donor bacteria in the presence of subinhibito ry concentrations of 1 of 16 antimicrobials in isolated ligated intestinal loops in swine. Donor Escherichia coli contained transferrable antimicrobial . resistance or virulence genes. After coincubations, intestinal contents were removed and assessed fo r pathogens that acquired new antimicrobial resistance or virulence genes fol­ lowing exposure to the subtherapeutic concentrati ons of antimicro bials. Results-3 antimicrobials (ap ramycin, lincomycin, and neomycin) enhanced transfer of an antimicrobial resistance plasmid from commensal E coli organisms to Yersinia and Prote­ us organisms, whereas 7 antimicrobials (florfenicol, hygromycin, penici llin G, roxa rsone, sulfamethazine, tetracycl ine, and tylosin) exacerbated transfer of an integron (Salmonella genomic island 1) from Salmonella organisms to Yersinia organisms. Sulfamethazine in­ duced the transfer of Salmonella pathogenicity island 1 from pathogenic to nonpathogenic Salmonella organisms. Six a ntimicrobials (bacitracin, carbadox, erythromycin, s ulfathiazole, tiamulin, and virginiamycin) did not mediate any transfer events. Sulfamethazic1e was the only antimicrobial implicated in 2 types of transfer events. Conclusions and Clinical Relevance-10 of 16 antimicrobials at s ubinhibitory or subthera­ peutic concentrations augmented specific antim icrobial resistance or transfer of viru lence genes into pathogenic bacteria in isolated intestinal loops in swine. Use of subtherapeutic antimicrobials in animal feed may be associated with unwanted coll ateral effects. (Am J Vet Res 2013;74 1078-1083)

u bth erapeutic concen trations of antimicrobials have S been used for decades as growth promotants and ABBREVIATIONS prophylactic agents. This practice has been scrutinized ESBL Extended-s pectrum ~- l ac t a m as e SGl1 Salmonella genomic isl a nd 1 as a contributor to th e d issemination of antimicrobial SPl1 Salmonella pathoge ni city isl a nd 1 resistance. Specifi cally, subtherapeu tic concentrations XLD Xylos e lysine d eoxycholate of chlortetracycline have been associated with an in­ crease in resistance to m ultiple antimicrobials for swine intestinal m icrobes. 1 Another study2 revealed that sub­ 3 therapeutic concentrations of antimicrobials can lead to organisms in swine. A recent study indicated that in­ the propagation of an timicrobial-resistant Enterococcus testinal bacteriophages are activated by subtherapeutic concentrations of antim icrobials, and this activation can lead to transfer of antimicrobial resistance genes. Received October 24, 2012. However, these studies have not addressed specific gene Accepted April 1, 2013. From the Department of Biomedical Scienc es, College of Veterinary transfer events that precipitate resistance and virulence. Medicine, Iowa State Universit y, Ames, IA 5001l. Dr. Xiong's pres­ Th us, little is known abou t the effects of these antimi­ ent address is Department of Neuroscience, University of Minnesota crobials on specific gene transfer even ts that promote Medical School, Minneapolis, MN 55414. the dissem ination of antimicrobial resistance and viru­ Supported by the National Pork Board on behalf of the Io wa Pork Pro­ lence genes in en teri c path ogens. ducers Association. Dr. Brewer was supported by the Merck Sum­ mer Scholars Fell owshi p Program. The objective of the study reported here was to The authors thank Diane McDonald and Dale Hinderaker fo r assis- iden tify antimicrobials that at subtherapeutic concen­ tance with care of the animals. trations augment the transfer of certain antimicrobial Address correspondence to Dr. Carlson ([email protected]). resistance and virulence genes. The main goal was to

1078 AJVR , Vol 74, No. 8, August 2013 assess 3 types of transfer events (plasmid transfer, inte­ were excised. Total volume of intestinal content in each gron transfer, and horizontal transfer of a pathogenicity isolated loop was 1.1 to 1.2 ml; approximately 10% of island) mediated by subtherapeutic concentrations of the total content of each loop was removed for plating antimicrobials in vivo. Specifically, the study was con­ on media selective for each of the 4 pathogens (XLD ducted to determine the relative rates of antimicrobial­ agarg for Salmonella and Shigella organisms, Yersinia mediated transfer events involving 3 representative anti­ selective agar base11 for Yersinia organisms, and phenyl­ microbial resistance plasmids, 1 model integron, and 1 alanine agar& for Proteus organisms); agar contained 1 pathogenicity island in the presence of antimicrobials of 3 antimicrobials (ceftiofur,; 32 µg/ml for the ESBL that have been approved for use (currently or in the plasmid; amikacin,< 64 µg/ml; or enronoxacin,i 8 µg/ past) at subtherapeutic concentrations as feed additives ml) at their respective breakpoint concentrations.5 The for swine. content of each loop was cultered in triplicate (3 agar plates/loop). Control experiments revealed that the do­ Materials and Met hods nor E coli did not grow on the selective media, except for the XLD agar; however, the E coli colonies that grew Animals- Twenty juvenile swine were used to as­ on XLD agar were biochemically distinct from the Sal­ sess in vivo transfer events. Pigs were of mixed breeds monella and Shigella colonies. and both sexes; pigs weighed between 5 and 10 kg. Ani­ Plates were incubated at 3 7°C for 16 hours. Bacte­ mal experiments were approved by the Iowa State Uni­ ria were then enumerated, and the number of translo­ versity Institutional Animal Care and Use Committee. cants/109 recipients was calculated by multiplying the Pigs were anesthetized with pentobarbitala ( 40 mg/ number of recovered colonies by 10 to account for the kg, intraperitoneal). Isolated ligated loops of intestine fact that only 10% of the intestinal contents were plat­ (9 loops/pig; each ligated loop was 10 cm in length) 4 ed. The log 10 of the product was derived and used for were created with 2-0 silk sutures.b At the end of the final data analysis and statistical evaluation. experiments, anesthetized pigs were euthanized by Transfer of the ESBL plasmid was assessed with a intracardiac administration of an overdose of pentobar­ PCR assay with primers (5'-ATGATGAAAAAATCGT­ bital (100 mg/kg). TATGCT-3' and 5'-TTATTGCAGCTTTTCAAGAAT-3') Plasmid transfer- To determine plasmid trans­ specific to the blacMr-z gene present on the ESBL plas­ fer from donor bacteria to recipient pathogenic bacte­ mid.8 Each transfer event was determined for 100 rep­ ria, each loop was injected with approximately 109 to resentative colonies. 11 9 10 CFUs of donor bacteria and 10 CFUs of recipient SGil transfer-The possibility that subinhibito­ bacteria in 1 ml of saline (O. 9% NaCl) solution that ry concentrations of antimicrobials can modulate the contained 1 of 16 feed additive antimicrobials< (apra­ transfer of SGll from Salmonella organisms to Yersinia mycin, bacitracin, carbadox, erythromycin, norfeni­ recipients, Shigella recipients, or Proteus recipients was col, hygromycin, lincomycin, neomycin, penicillin G, evaluated. The SG ll is a multi resistance integron that roxarsone, sulfamethazine, sulfathiazole, tetracycline, encodes resistance to 5 antimicrobials in Salmonella or­ tiamulin, tylosin, and virginiamycin); antimicrobial­ ganismsn lntegrons are mobile genomic elements pu­ 12 13 15 free saline solution was used as a negative control treat­ tatively transferred by bacterioµhages, and studies - ment. Concentration of each antimicrobial was 1 µg/ have indicated that Yersinia spp, Shigella spp, and Pro­ ml. This concentration was chosen because it was less teus spp are capable of receiving integrons. than the established breakpoints for all of the antimi­ To evaluate SGil transfer, approximately 109 CFUs crobial-bacteria combinations,5 and preliminary experi­ of SGil-free recipient (Yersinia enterocolitica, Shigella ments conducted by our laboratory group on the mini­ flexneri, or Proteus mirabilis [all of which were sus­ mum inhibitory concentration of these antimicrobials ceptible to most antimicrobials, except tetracycline revealed that this was the lowest common concentration and streptomycin, as determined with microdilution that permitted growth of all donor and recipient bacteria broth assays performed in accordance with standards for each of the 16 antimicrobials (data not shown). established by the Clinical and Laboratory Standards 5 9 Plasmid transfer from a donor commensal intesti­ Institute ]) were each coinoculated with 10 CFUs of nal Escherichia coli 6 (antimicrobial susceptible and lack­ SGil-bearing donor S enterica serotype Typhimurium ing virulence genes) to each of 4 recipient pathogenic phagetype DT104 strain LNWI4 into ligated intestinal bacteria (Sa lmonella enterica serotype Typhimurium loops. Coincubations included 1 of the 16 aforemen­ strain Sll344,7 Yersinia enterocolitica,d Shigellaflexneri,' tioned antimicrobials (concentration, 1 µglml) and the or Proteus mirabilisr) was measured. All of the recipi­ negative control treatment (saline solution without an ent bacteria were susceptible to most antimicrobials antimicrobial)_ (except tetracycline and streptomycin), as determined Bacteria were allowed to incubate in the isolated li­ by use of microdilution broth assays performed in ac­ gated intestinal loops for 1 hour, after which loops were cordance with standards established by the Clinical and excised. Total volume of the intestinal content in each Laboratory Standards Institute.5 The E coli donor strain isolated loop was 1.1to1.2 ml; approximately 10% of was experimentally transformed with 1 of 3 conjuga­ the total content of each loop was removed for plating tive plasmids encoding an ESBL, 8 amikacin resistance on media selective for each of the 3 pathogens (XLD (aacC4), 9 or fluoroquinolone resistance (via the qnr agar for Salmonella and Shigella organisms, Yersinia se­ gene). 10 lective agar base for Yersinia organisms, and phenylala­ Bacteria were allowed to incubate in the isolated nine agar for Proteus organisms); agar contained 1 of ligated intestinal loops for 1 hour, after which loops 2 antimicrobials relevant for SGil (ampicillin,< 32 µg/

AJVR, Vol 74, No. 8, August 2013 1079 mL; chloramphenicol,' 32 µglmL) at their respective loops (153 total loops) and 3 agar plates/loop (459 to­ breakpoint concentrations.5 The content of each loop tal loops). Frequency data for all transfer events were was cultured in triplicate (3 agar plates/loop). analyzed en masse to allow for interantimicrobial and Plates were incubated at 3 7°C for 16 hours. Bacte­ intraevent comparisons and intra-antimicrobial and ria then were enumerated, and the number of translo­ interevent comparisons (ie, comparisons were made cants/109 recipients was calculated by multiplying the among the antimicrobials within the 3 transfer events number of recovered colonies by 10 to account for the [plasmid, SGil, and SPil] and among the 3 transfer fact that only 10% of the intestinal contents were plat­ events within a specific antimicrobial). The Scheffe F ed. The log of the product was derived and used for test was chosen as the ad hoc test because our research 10 final data analysis and statistical evaluation. group has empirically found it to be the most conserva­ A PCR assay specific to the jloR-tetR sequence in tive for detecting differences, in contrast to the Bonfer­ SGil 16 was used to assess the presence of the SGil inte­ roni test. Values of P < 0.05 were considered significant. gron in 100 representative recipient bacteria that grew on selective media containing ampicillin and chloram­ Results phenicol. Additionally, transferrants were assessed for Plasmid transfer-In vivo transfer events of 3 the absence of a Salmonella virulence gene segment clinically relevant antimicrobial resistance plasmids (si pBIC) .16 Control experiments revealed that the donor were assessed by use Salmonella organisms were not able to grow on the me­ of ligated intestinal loops of swine. Loops were coinoculated with donor commensal dia selective for Yersinia spp and Proteus spp, whereas E bearing 1 of 3 antimicrobial resistance plasmids, Salmonella organisms were distinguishable from Shi­ coli 1 of 4 recipient pathogenic Enterobacteriaceae, and an gel la organisms on XLD agar. antimicrobial (1 µglmL) approved as a feed additive in SPil transfer-Transfer of SP il from pathogenic swine. Commensal E coli. and recipient Enterobacte­ to nonpathogenic Salmonella organisms was evaluated. riaceae were chosen because these microbes are highly The SPil is a major determinant of virulence in Sal­ representative of enteric bacteria that transfer genetic monella organisms, and avirulent Salmonella organisms information. Three antimicrobials mediated a signifi­ lack SPil. 17 The genomic structure of SPil suggests that cant increase in the frequency of a specific plasmid this island is transferrable.18 In vivo coincubations with transfer event in vivo (Figure 1). An increased frequen­ the 16 antimicrobials and the negative control treat­ cy of transfer of the ESBL plasmid from E coli to Yer­ ment (saline solution without an antimicrobial) were sinia recipients and from E coli to Proteus recipients was perfo rmed as described previously for the plasmid and evident in the presence of apramycin, lincomycin, or integron transfer experiments. The SPil-bearing S en­ neomycin (Table 1). None of the antimicrobials caused terica serotype Typhimurium (antimicrobial-suscepti­ significant changes in the frequencies of ESBL plasmid 7 ble strain SL1344 ) was incubated with 1 of 4 SPil-free transfer from E coli to Salmonella recipients and from strains (S enteri ca serotypes Litchfield, 17 Senftenberg, 17 E coli to Shigella recipients. Additionally, none of the Seminole, or Betiocky), all 4 of which were transformed antimicrobials caused significant changes in transfer of with a nonco njugative plasmid encoding green fluores­ amikacin resistance plasmids or fluoroquinolone resis­ cent protein. 19 For this experiment, 1011 CFUs of donor tance plasmids. bacteria and 1011 CFUs of recipient bacteria were used for the incubations. Bacteria were allowed to incubate in the ligated loops for 1 hour. Total volume of each loop was 1.1 to 1.2 mL; approximately 10% of the total content of each loop was rem oved and used in a large-volume tis­ sue culture invasion assay20 in which recovered bacteria were plated on XLD agar that contained 50 µ g of zeo­ cink/mL (zeocin is the selective marker for the fluores­ 19 cence plasmid ). The content of each loop was cultured in triplicate (3 agar plates/loop). Fluorescent colonies that were invasive (ie, recovered from inside tissue cul­ ture cells) were individually expanded in fresh nutrient b roth and then subjected to a second invasion assay. Yersinia onterocolilica Proteus mirobilis Amount of invasion was compared with that of strain Recipient bacteria SL1344. 7 Serotype analysis was conducted at another 1 Fi gu re 1-Plasmid transfer from commensal Escherichia coli to laboratory, and a PCR assay that detected the sipB­ Yersinia organisms and Proteus organisms after co in cubation sipC sequence in SPil 16 was performed on each clone with a subtherapeutic concentration (1 µg/mU of lincomycin for which invasion was indistinguishable from that of (white bars), apramycin (gray bars). or neomycin (striped ba rs) or with sa li ne (0.9% NaCl) solution that did not contain an antimicro­ strain SL1344 (approx 1% invasion). bial (negative control treatment [b lack bars]) for 1 hour in isolated ligated intestinal loops in swine. Results represe nt the mean ± Statistical analysis-Statistical differences were SEM for 3 experiments that were each performed in triplicate. assessed with an ANOVA.m There were 51 combina­ Notice that there was enhancement of transfer of the ESB L plas­ tions of antimicrobials and transfer events ( [ 16 anti­ mid8 in the presence of subtherapeutic concentrations of anti­ microbials in Yersinia recipients and Proteus recipients. *Within a microbials plus 1 antimicrobial-free treatment] times bacterium, value differs significantly (P < 0.05) from the value for 3 transfer events), each assessed in 3 separate ligated the negative control treatment

1080 AJVR, Vol 74, No. 8, August 2013 Table 1-Gene transfer events for donor bacteria to recipient bacteria after coincubation with a subtherapeutic concentration (1 µg/ml) of 16 antimicrobials or saline (0 .9 % NaCl) solution that did not contain an antimicrobial (negative control treatment) for 1 hour in isolated ligated intestinal loops in swine.

Ratio of total transfer Antimicrobial ESBL plasmid transfer SGl1 transfer SPl1 transfer frequencies Saline so lution NA NA NA 1.0

Apramycin Yersinia and Proteus recipients* 3,471 Bacitracin 2.3 Carbadox 0.6 Eryth romycin 0.7 Florfenicol Yersinia recipients* 9.8 Hygromycin Yersinia recipients* 5.1 Lincomycin Yersinia and Proteus recipients* 4,596 Neomycin Yersinia and Proteus recipients* 4,641

Penicillin G Yersinia recipients* 4.4 Roxarsone Yersinia recipients* 4.7 Sulfamethazine Yersinia recipients* 3 recipient nonpathogenic 13.6 serovars of Salmonella enterica*t Sulfathiazole 1.1 Tetracycline Yersinia recipients* 4.1 Tiamulin 1.2 Tylosin Yersinia recipients* 10.3 Virginiamycin 0.9

Mean transfer frequency in the absence of an antimicrobial was approximately 1.4 X 10-a transconjugates/recipient for the ESBL plasmid, 2.6 X 10-• translocants/recipient for SGll, and 3 X 10-11 invasive clones/noninvasive rec ipient for SPl1; these frequencies were calculated by dividing the total number of all recovered recipients by t he total number of donor bacteria added to all of the coincubations in the absence of an antimicrobial. For each antimicrobial, the ratio of total transfer frequencies= (total number of all bacteria that received the new genetic element when coincubated with an antimicrobial/total number of recipients coincubated with an antimicrobial)/(total number of all bacteria that received the new genetic element when coi ncubated without an antimicrobial/total number of recipients coincubated without an antimicrobia l). *Transfer frequency was significantly (P < 0.05) greater than the transfer frequency for the antimicrobial-free (negative control) treatment. t Salmonella enterica serotypes Betiocky, Li tc hfield, and Seminole. NA= Not applicable. - =Transfer frequency was not significantly greater than the transfer frequency for the antimicrobial-free (negative control) treatment.

The PCR assay of 100 representative 2 ceftiofur-resistant colonies from each of T* Jij * the aforementioned transfer events re­ .,c:: * vealed that 96% to 97% of the putative ·15. * -~ , . * transconjugates contained blacM1 2 Colo­ T nies that lacked this sequence were nu­ ~ merically discounted as transconjugates in the final calculations (Figure 1). SGil transfer-In vivo transfer events 0 of an integron were assessed with ligated o intestinal loops coinoculated with donor ! Salmonella organisms bearing SGll,11 1 of 4 recipient pathogenic Enterobacteriaceae, and an antimicrobial (1 µglmL) approved as a feed additive in swine. Seven antimicro­ 0 -, bials mediated a significant increase in the frequency of in vivo transfer of a specific in­ tegron (Figure 2). An increased frequency of transfer of the SGll integron from Salmonel­ la organisms to Yersinia organisms was evi­ Figure 2-Transfer of SGl1 from Salmonella enterica serotype Typhimurium phage­ dent in the presence of florfenicol, hygromy­ type DT104 to Yersinia enterocolitica after coincubation w ith a subtherapeutic con­ cin, penicillin G, roxarsone, sulfamethazine, centration (1 µg/ml) of 16 antimicrobials or saline solution that did not contain an antimicrobial (negative control treatment) for 1 hour in isolated ligated intestinal tetracycline, and tylosin (Table 1). None of loops in swine. Results represent the mean ± SEM for 3 experiments that were the antimicrobials caused significant chang­ each performed in triplicate. Notice that transfer of the integron was enhanced in es in the frequencies of SGil transfer events the presence of 7 antimicrobials. *Value differs significantly (P < 0.05) from the from Salmonella organisms to either of the value for coincubation with the negative control treatment. SMZ = Sulfamethazine. other 2 pathogens examined. The PCR assay of 100 representative colonies from each of the 16 afore­ indicative of SGll. 16 Colonies that lacked this sequence mentioned transfer events revealed that 98% to 99% of the were numerically discounted as translocants in the final putative SGil recipients contained the floR-tetR sequence calculations.

AJVR, Vol 74, No. 8, August 2013 1081 22 to mimic field conditions without directly harming the donor or recipient bacteria. Although this concentra­ tion may be lower than concentrations in the intesti­ 120 nal tract of swine fed feed that contains antimicrobials, 110 17 this concentration was considered relevant for swine in 100 12 90 which the antimicrobial is removed from the diet prior '#so, to slaughter. ~ 70 1 The present study revealed that there were transfer - ~ GO i events in the absence of an antimicrobial and that cer­ ~ 50 j tain antimicrobials mediated gene transfer events at a £ 40 higher frequency than did other antimicrobials (Table 30 20 1). Sulfamethazine mediated 2 separate transfer events 10 (5Gll transfer and 5Pll transfer) , whereas a related 0 -<--8-e-l~io-ck-y~~~-S-e-m~in-o-le~·~~ drug, sulfathiazole, did not mediate any transfer events. Litchfield Bacitracin, carbadox, erythromycin, tiamulin, and vir­ Recipient Salmonella entedca serovar giniamycin were also not implicated in transfer events. Figure 3-Transfer of SPI 1 from pathogenic Salmonella organ­ Transfer events were confirmed with a PCR assay, al­ isms to nonpathogenic Salmonella organisms after coincubation though a few (1 % to 4%) transconjugates and translo­ with a subtherapeutic concentration (1 µg/ml) of sulfamethazine cants did not harbor the transferrable element, which (white bars) or saline solution that did not contain an antimicro­ bial (negative control treatment [black bars]) for 1 hour in isolated suggested that efflux systems may have been activated ligated intestinal loops in swine. Fluorescent bacteria were sub­ in these few clones. jected to 2 invasion assays, and colonies that were invasive (ie, Three classes of antimicrobials were implicated in recovered from inside tissue culture cells) were quantitated by enumerating the bacteria recovered from tissue culture cell ly­ transfer of the ESBL plasmid. Lincomycin is a lincos­ sates at the end of the second assay. Invasion represents the amide, apramycin is an aminocyclitol, and neomycin is percentage as compared with that of the donor Salmonella or­ an aminoglycoside (although the latter 2 are sometimes ganisms. Results represent the mean ± SEM for 3 experiments grouped in the same class). Lincomycin is an inhibitor that were each performed in triplicate. Notice that transfer of SPl1 was only detected in the presence of sulfamethazine and of the 50S ribosome in bacteria, yet 3 other 50S inhibi­ that 1 clone of S enterica serotype Litchfield was invasive in the tors (erythromycin, tiamulin, and virginiamycin) did absence of an antimicrobia l. The number above each column indi­ not mediate transfer events. Apramycin and neomycin ca tes the cumulative number of serovar-s pecific clones that were invasive across 3 separate experiments. are inhibitors of the 305 ribosome, but another 30S in­ hibitor, tetracycline, did not exacerbate transfer of the ESBL plasmid. It is possible that apramycin, lincomy­ SPil transfer-In vivo transfer events of a patho­ cin, and neomycin can selectively alter protein synthe­ genicity island were assessed with ligated intestinal sis that impacts sex pheromones and conjugation. loops coinoculated with donor Salmonella organisms, 1 Seven antimicrobials (florfenicol, hygromycin, of 4 recipient nonpathogenic Salmonella organisms that penicillin G, roxarsone, sulfamethazine, tetracycline, lacked SPil, and an antimicrobial (1 µglmL) approved and tylosin) from 7 antimicrobial classes were impli­ as a feed additive in swine. Sulfamethazine mediated the cated in the transfer of SGIL Tylosin and florfenicol horizontal transfer of SPil in vivo (Figure 3). Three of are 505 inhibitors, but that is the extent of the similari­ 4 SPil-free S enterica serovars (Betiocky, Seminole, and ties among these 7 antimicrobials. Because movement Litchfield) had transfer events in the presence of sulfa­ of SGil may be a phage-mediated event, 12 it is possible methazine (Table 1). Transfer of SPil was detected in the absence of an antimicrobial for 1 clone of S enterica that those 7 antimicrobials activated phage recrudes­ serovar Litchfield. Salmonella enterica serovar 5enften­ cense in 5Gll-bearing Salmonella organisms, similar to that recently reported for sulfamethazine, chlortet­ berg did not acquire SPil in this experiment. Results of 3 the PCR assays confirmed the presence of the sipB-sipC racycline, and penicillin. It is also possible that naked genomic segment in all clones that had invasion indis­ SGil DNA was transferred from Salmonella organisms tinguishable from that of strain SL1344. to Yersinia organisms, although to our knowledge this process has not been described in the literature. Analysis of results of the present study indicated Discussion that certain antimicrobials at subtherapeutic concentra­ The use of subtherapeutic concentrations of anti­ tions are more likely to mediate unwanted gene trans­ microbials in livestock feed is a controversial practice fer into pathogenic bacteria in ligated intestinal loops that is being scrutinized. Of concern are the unknown in swine. Sulfamethazine mediated 2 types of transfer, collateral effects on bacterial gene transcription, plas­ whereas apramycin, lincomycin, and neomycin exerted mid transconjugation from commensals to patho­ the greatest quantitative effect on a single transfer event gens, and viral-mediated transduction of genes from into 2 genera of Enterobacteriaceae. Bacitracin, carbadox, commensals to pathogens and from one pathogen to erythromycin, sulfathiazole, tiamulin, and virginiamycin another. Specifically, these collateral effects can acti­ did not significantly influence any of the 3 transfer events vate molecular processes culminating in gene transfer evaluated. No antimicrobial class-specific patterns were events that yield pathogens with multiple antimicrobial observed in the 3 transfer events. Transfer of the ESBL resistance or pathogens with new virulence capabilities. plasmid was detected at the highest frequency Protein The experiments in the present study involved the synthesis irregularities may underlie transfer of the E5BL use of a low concentration (1 µglmL) of antimicrobials plasmid, and activation ofbacteriophages may be involved

1082 AJVR, Vol 74, No. 8, August 2013 in SGil transfer or SPil transfer (or both). Regardless of 7. Wray C, Sojka WJ. Experimental Salmonella typhimurium in the mechanisms involved, subtherapeutic concentrations calves. Res Vet Sci 1978;25:139-143. of antimicrobials have potential ecologic impacts that in­ 8. Brewer M, Xiong N, Dier], et al. Comparisons of Salmonella conjugation and virulence gene hyperexpression mediated volve the dissemination of antimicrobial resistance and by rumen protozoa from domestic and exotic ruminants. Vet virulence genes. Microbial 2011;151:301-306. 9. Tosini F, Visca P, Luzzi I, et al. Class 1 integron-borne multiple­ a. Fon Dodge Animal Health, Fort Dodge, Iowa. antibiotic resistance carried by lncFI and lncVM plasmids in b. Ethicon Inc, Somerville, NJ. Salmonella enterica serotype Typhimurium. Antimicrob Agents c. Sigma Chemical Co, St Louis, Mo. Chemother 1998;42:3053-3058. cl. Yersinia enterocolitica 9610, ATCC, Manassas, Va. 10. Gay K, Robicsek A, Strahilevitz J , el al. Plasmid-mediated qui­ e. Shigella flexn eri 9199, ATCC, Manassas, Va. nolone resistance in non-Typhi serotypes of Salmonella enterica. f. Proteus mirabilis 4630, ATCC, Manassas, Va. Clin Infect Dis 2006;43:297-304. g. Becton Dickinson, Franklin Lakes, NJ. 11. Briggs CE, Fratamico PM. Molecular characterization of an anti­ h. Thermo Fisher Scientific, Pittsburgh, Pa. biotic resistance gene cluster of Salmonella typhimurium DT104. i. Pfizer Inc, Kalamazoo, Mich. Antimicrob Agents Chemother 1999;43:846-849. j. Bayer Corp, Shawnee Mission, Kan. 12. Schmieger H, Schicklmaier P. Transduction of multiple drug k. Invitrogen Corp, Carlsbad, Calif. resistance of Salmonella enterica serovar typhimurium DT104. I. National Veterinary Service Laboratories, Ames, Iowa. FEMS Microbiol Lett 1999;170:251-256. m. StatView, version 2, SAS Institute Inc, Cary, NC. 13. Soto S, Lobato M, Mendoza M. Class 1 integron-borne gene cassettes in multidrug-resistant Yersinia enterocolitica strains of different phenotypic and genetic types. Antimicrob Agents Che­ References mother 2003;4 7:421-426. 1. Funk], Lej eune ] , Wittum T, et al. The effect of subtherapeutic 14. Mciver C, White P, Jones L, et al. Epidemic strains of Shi­ chlonetracycline on antimicrobial resistance in the fecal flora of gella sonnei biotype g carrying integrons. J Clin Microbial swine. Microb Drug Resist 2006;12:210-218. 2002;40: 1538-1540. 2. Poole T, Hume M, Genovese K, et al. Persistence of a vancomycin­ 15. Ahmed A, Hussein A, Shimamo:o T. Proteus mirabilis clinical resistant Enterococcus faecium in an anaerobic continuous-flow cul­ isolate harbouring a new variant of Salmonella genomic island ture of porcine microflora in the presence of subtherapeutic con­ 1 containing the multiple antibiotic resistance region.] Antimi­ centrations of vancomycin. Microb Drug Resist 2001;7:343-348. crob Chemother 2007;59: 184-190. 3. Allen H, Looft T, Bayles D, et al. Antibiotics in feed induce 16. Carlson SA, Bolton LF, Briggs CE, et al. Detection of Salmonella prophages in swine fecal microbiomes. mBio 2011;2:e00260- typhimurium DT104 using multiplex and fluorogenic PCR Mal e00211. Cellular Probes 1999;13:213-222. 4. Wu MT, Carlson SA, Meyerholz DK. Cytopathic effects ob­ 17. Ginocchio CC, Rahn K, Clarke RC, et al. Naturally occurring dele­ served upon expression of a repressed collagenase gene pres­ tions in the centisome 63 pathogenicity island of environmental ent in Salmonella and related pathogens: mimicry of a cyto­ isolates of Salm onella spp. Infect Immun 1997;65:1267-1272. toxin from multiple antibiotic-resistant Salmonella enterica 18. Hensel M, Shea j, Baumler A, et al. Analysis of the boundar­ serotype Typhimurium phagetype DT104. Microb Pathogen ies of Salmonella pathogenicity island 2 and the correspond­ 2002;32:1-9. ing chromosomal region of Escherichia coli K-12. ] Bacteria l 5. Clinical and Laboratory Standards Institute. MIC Interpretive Stan­ l 997;179:1105-1111. dards (µglmL) for Enterobacteriaceac. In: Pe1formance standards for 19. Frana T, Carlson S. Development and use of a plasmid encoding antimicrobial disk and dilution susceptibility tests. 2nd eel. Wayne, green fluorescent protein in multiple antibiotic-resistant Salmo­ Pa: Clinical and Laboratory Standards Institute, 2008;16-19. nella. Bio Techniques 2001;30:28-32. 6. Lyczak JB. Commensal bacteria increase invasion of intestinal 20. Gianella RA, Washington 0, Gemski P, et al. Invasion of HeLa epithelium by Salmonella enterica serovar Typhi. Infect Immun cells by Salmonella typhimurium: a model for study of invasive­ 2003;71:6610-6614. ness of Salmonella . .f lnfect Dis 1973;128:69-75.

AJVR, Vol 74, No. 8, August 2013 1083