Genetic characterization of AmpC and extended-spectrum beta-lactamase (ESBL) phenotypes in Escherichia coli and Salmonella from Alberta poultry

Horizontal gene transfer is an important mechanism which facilitates bacterial populations in overcoming antimicrobial treatment. In this study, a total of 120 Escherichia coli and 62 Salmonella enterica subsp. enterica isolates were isolated from poultry farms in Alberta. Fourteen serovars were identified among Salmonella isolates. Thirty one percent of E. coli isolates were multiclass drug resistant (resistant to ≥ 3 drug classes), while only about 16% of Salmonella isolates were multiclass drug resistant. Among those, eight E. coli isolates had an AmpC-type phenotype, and one Salmonella isolate had an extended-spectrum beta-lactamase (ESBL)-type β-lactamase phenotype. We identified both AmpC-type (blaCMY-2) and ESBL-type (blaTEM) genes in both E. coli and Salmonella isolates. Plasmids from eight of nine E. coli and Salmonella isolates were transferred to recipient strain E. coli J53 through conjugation. Transferable plasmids in above total eight E. coli and Salmonella isolates were also transferred into a lab-made sodium azide-resistant Salmonella recipient through conjugation. The class 1 integrase gene, int1, was detected on plasmids from two E. coli isolates. Further investigation of class 1 integron cassette regions revealed the presence of an aadA gene encoding streptomycin 3”-adenylyltransferase, an aadA1a/aadA2 gene encoding aminoglycoside 3”-O-adenyltransferase, and a putative adenylyltransferase gene. This study provides some insight into potential horizontal gene transfer events of antimicrobial resistance genes between E. coli and Salmonella in poultry production.

≥ 3 drug classes), while only about 16% of Salmonella isolates were multiclass 26 drug resistant. Among those, eight E. coli isolates had an AmpC-type phenotype, and one Salmonella 27 isolate had an extended-spectrum beta-lactamase (ESBL)-type β -lactamase phenotype. We identified 28 both AmpC-type (bla CMY-2 ) and ESBL-type (bla TEM ) genes in both E. coli and Salmonella isolates. 29 Plasmids from eight of nine E. coli and Salmonella isolates were transferred to recipient strain E. coli 30 J53 through conjugation. Transferable plasmids in above total eight E. coli and Salmonella isolates 31 were also transferred into a lab-made sodium azide-resistant Salmonella recipient through 32 conjugation. The class 1 integrase gene, int1, was detected on plasmids from two E. coli isolates. 33 Further investigation of class 1 integron cassette regions revealed the presence of an aadA gene 34 encoding streptomycin 3''-adenylyltransferase, an aadA1a/aadA2 gene encoding aminoglycoside 3''-35 O-adenyltransferase, and a putative adenylyltransferase gene. This study provides some insight into 36 potential horizontal gene transfer events of antimicrobial resistance genes between E. coli and 37 Salmonella in poultry production.

Introduction 39
For decades, antimicrobial resistance (AMR) has been a global issue of grave concern. 40 Understanding potential mechanisms and driving forces for dissemination of genes encoding 41 antimicrobial resistance between bacteria will help reduce the prevalence of resistant bacteria and 42 thereby reduce risks to human and animal health. Acquisition of new resistance genes occurs 43 frequently and naturally among bacterial communities from humans, animals and environments as 44 outlined in the model known as 'the epidemiology of AMR' (John F. Prescott, 2006). However, the 45 mechanism of dissemination of resistance genes is not yet fully understood. sharing ~85% of their genomes in common at the nucleotide level (Mcclelland and Wilson, 1998;51 McClelland et al., 2000). 52 AmpC-type CMY β -lactamase genes (bla CMY ) have been found on both the chromosome and 53 plasmids of many gram negative bacteria such as, Klebsiella sp., Escherichia coli, and Salmonella sp. 54 CMY-2 is reported to be the most common plasmid-carried AmpC-type CMY in both E. coli and 55 Salmonella isolates from various global regions including Asia, North America and Europe (Guo et  56 al can be easily misidentified as ESBLs due to their overlapping activity against beta-lactam 60 antimicrobials. The inability to distinguish them could have significant treatment consequences 61 (Hanson, 2003). 62 Mobile genetic elements, such as plasmids or DNA transposons, are the main mechanisms 63 facilitating horizontal genetic transfer (HGT). Plasmid-mediated bla CMY-2 has been found to be the 64 most predominant among other acquired ampC genes (Mata et al., 2012

Phenotype and genotype confirmation of ESBL/AmpC genes 115
Two different ESBL/AmpC detection disc sets have been used to confirm ESBL/AmpC phenotypes. 116 The first set is a combination of 4 individual discs of Cefotaxime/ Cefotaxime + Clavulanic acid/ 117 Ceftazidime/ Ceftazidime + Clavulanic acid, purchased from either BD BBL TM or Oxoid company. 118 The second set is a combination of 4 individual discs of Cefpodoxime/ Cefpodoxime + ESBL 119 inhibitor/ Cefpodoxime + AmpC inhibitor/ Cefpodoxime + ESBL inhibitor + AmpC inhibitor, 120 purchased from Mast Group company (D68C set). 121 In addition, AmpC and ESBL β -lactamase genes were detected using PCR assays. A total of three 122 AmpC (bla CMY-2 , bla FOX , bla ACT-1/MIR-1 ) and ten ESBL (bla TEM , bla SHV bla CTX-M-1 , bla CTX-M-2 , bla CTX-M-β -lactamase genes were screened in AmpC/ESBL 124 positive isolates. Primers used in the PCR assays are listed in Table 6. 125 The MDR E. coli and Salmonella were selected for further experiments after being confirmed to 126 exhibit ESBL/AmpC phenotypes. 127

Plasmid characterization/ replicon typing 128
Plasmid miniprep was performed using an alkaline lysis method (Birnboim and Doly, 1984). 129 Replicon typing was performed using PCR assay as described previously, with primers listed in 130

Biofilm production assay in vitro 139
Isolates of interest were assessed for the ability to produce biofilm by culture in clear 96 well 140 microtiter plates as described below (Stepanovic et al., 2007). An overnight liquid culture was diluted 141 using a 1:100 ratio in tryptic soy broth (TSB) supplemented with casamino acids, and 200 µl of the 142 suspension was then aliquoted onto a 96-well plate (each sample was assayed in triplicate). The plate 143 was incubated as static culture at 37°C. After 24hr incubation, the culture medium was decanted and 144 plate wells were rinsed with distilled water three times. The plate was stained with 200 µl of 1% 145 crystal violet for 30 mins, rinsed under running tap water and then air-dried (by gently tapping then 146 placing up-side-down on paper towels). Subsequently, 200 µl of glacial acetic acid (33%) was added 147 into the wells to solubilize the dye, and absorbance was measured at 600 nm. The results were 148 interpreted as described previously (Stepanovic et al., 2007). 149

Detection of integrons/integrases 150
To further study other mobile genetic elements, different integrase classes were identified using PCR 151 assays (int1, int2 and int3) using primers listed in Table 6  Azi r ), an E. coli K-12 derivative strain, is resistant to sodium azide (63). Recipient and donor strains 158 were inoculated into LB broth and cultured overnight at 37°C. The next day, cells were harvested, 159 washed with saline, and mixed together in a ratio of 1:1, and spotted on to LB plates. They were also 160 spotted individually on LB plates as controls. After overnight incubation at 37°C, mating spots were 161 washed and resuspended in saline; and different dilutions were plated on LB media containing 162 sodium azide (0.2 gL -1 ) and ampicillin (100 µgml -1 ) to select transconjugants. Control spots were 163 transferred to the same selective media to make sure that no growth was observed. Conjugation 164 frequency was calculated by taking the ratio of the number of colonies counted on selective plates 165 (LB supplemented with sodium azide (0.2 gL -1 ) + ampicillin (100 µgml -1 )) for transconjugants over 166 the number of colonies on selective plates (LB supplemented with sodium azide (0.2 gL -1 )) for 167 recipients. If there were no transconjugants obtained, a helper strain (HB101/pRK600) was added 168 into the mating mix in the proportion of 1:1:0.5 (donor: recipient: helper strain) and spotted on LB 169 plates as described. If there was no growth on plates selected for recipients, we added trypsin to the 170 media to recover the recipients (42). 171 Salmonella isolate 112.2 were screened for spontaneously mutated colonies that resisted to sodium 172 azide (Azi R ) by plating on LB supplemented with sodium azide (0.2 gL-1). Then this Azi R 173 Salmonella was used as a recipient in conjugation with MDR isolates of interest as donors. 174 Conjugation protocol was performed as described above. 175

Data visualization tools 176
Data visualization in this study was performed using following programs: Microsoft Excel 2013, R 177 programming (R version 3.4.1). 178

Sampling, isolation and identification of bacterial strains 180
Four E. coli isolates were obtained from each farm, resulting in 120 E. coli isolates from 30 farms. 181 Twenty-three of 30 farms Salmonella positive, with between one and four isolates identified per 182 farm. There were 14 different serovars identified among 120 Salmonella isolates (Table 1). 183

Antimicrobial susceptibility testing 184
Antimicrobial susceptibility in E. coli and Salmonella isolates is described. Isolates that were 185 resistant to three or more drug classes were considered multiclass drug resistant (MDR). Thirty-one 186 percentage of E. coli were MDR and 16% of Salmonella were MDR. About 4% of E. coli were 187 resistant to five drug classes, while none of Salmonella were resistant to five drug classes. 188 The majority of Salmonella isolates were resistant to streptomycin and tetracycline (Table 1). There 189 were 8 Salmonella serotypes that were sensitive to all tested drugs (Enteritidis, Typhimurium,190 Braenderup, Hartford, Infantis, Schwarzengrund, Senftenberg, Thompson). In addition to 191 streptomycin and tetracycline, the majority of MDR E. coli showed resistance to sulfasoxasole (Table  192 2a). Among E. coli that were bio-typed, those belonging to groups D or E had diverse AMR patterns 193 (Table 2b). Overall, E. coli isolates showed more diversity in resistance phenotype between farms 194 than did Salmonella (Figure 1)

Plasmid characterization 210
Nine isolates were found to carry I1 and A/C-type replicon plasmids (Table 3). The plasmids varied 211 in size from approximately 7kb to larger than 20kb. Only two E. coli isolates, 58.1 and 61.1, carried 212 small plasmids (<10kb), while the rest carried larger ones (≥20 kb ). 213

E. coli biotypes/Salmonella serovars of ESBL/AmpC-positive isolates 214
Both PCR methods confirmed that none of the E. coli isolates belonged to the group B2 (a group  215 with high potential for pathogenicity) (

Biofilm formation 219
In our biofilm assay, only one E. coli isolate lacked the ability to produce biofilm, while the other 220 eight isolates produced biofilm (Table 3). Two E. coli isolates, 12.1 and 89.1, demonstrated moderate 221 production, the highest amongst the isolates assayed. 222

Detection of integrons/integrases 223
Plasmids from two E. coli isolates, from 2 different farms, carried the class 1 integrase gene int1 224 (Table 3). The class 1 integron cassette region was also detected in two isolates by PCR (Table 3). 225 Sequencing these products identified the aminoglycoside resistance genes, aadA encoding 226 streptomycin 3''-adenylyltransferase, and aadA1a/aadA2 encoding aminoglycoside 3''-O-227 adenyltransferase. Class 1 integron cassettes were amplified from plasmids isolated from E. coli 228 isolate 82.1. The results were confirmed by blasting the sequence against NCBI database. The 229 sequences matched the sequence of a putative adenylyltransferase found on a plasmid isolated from 230 the Salmonella Heidelberg strain N418 (Accession no. CP009409). 231

Transfer of resistance genes by conjugation 232
Plasmids were mobilized from all but one E. coli isolate ( When using a lab-engineered sodium azide-resistant Salmonella as a recipient, we observed that 239 plasmids from eight E. coli isolates with the exception of the isolate mentioned above, were able to 240 move to Salmonella with variable conjugation frequency (Table 5). 241

Farm characteristics for nine isolates from which plasmids were mobilized 242
Seven 'conventional' (i.e., antimicrobials were used to some extent in all flocks) farms under the 243 veterinary care of one practice were represented by the nine ESBL/AmpC phenotyped isolates in this 244 study. All of the farms, excluding the farm providing isolate 89.1, received their chicks from the 245 same hatchery. All birds were Ross 308 strain. The most frequently used antimicrobials were 246 bacitracin and salinomycin administered in the poultry feed (n = 5), followed by the combination of 247 penicillin and streptomycin administered in the water (n = 3). Avilamycin (n = 1), decoquinate (n = 248 1), monensin (n = 1), tylosin (n = 2) and the combination narasin and nicarbazin (n = 2) were also 249 used as feed additives on these farms. 250 The number of chicks sampled per flock ranged from 14,790 to 55,000 within a single production 251 unit. Age on the day of sampling ranged from 30 to 35 days old with an average weight ranging from 252 1.7 kg to 2.2 kg. The recorded floor space in the barns ranged from 8000 ft 2 to 30550 ft 2 and stocking 253 density ranged from 0.54 to 0.67 ft 2 per bird. Reported mortality rates ranged from 2.47% to 7.19% 254 of the birds placed within the barn. 255 Hydrogen peroxide was used on three of the farms from which biofilm-producing bacteria were 256 isolated, for cleaning of water lines between flocks. Five of the farms also used chlorine for treatment 257 of their water lines during the production cycle. Footbaths (n = 3), dedicated farm clothes (n = 4) and 258 gloves (n =2) were methods of farm biosecurity utilized. Manure was stored onsite in the vicinity of 259 the barn on three farms. The most frequently reported method of cleaning the barns after each 260 production cycle was washing only (n = 6) and chlorine products were used for disinfection on four 261 farms. 262

4
Discussion 263 We examined the antimicrobial resistance phenotypes, genotypes, and mobile genetic elements of E. isolate. This is a of concern because in 2013-2014, a national outbreak of MDR Pseudomonas aeruginosa (Sanders, 1987;Philippon et al., 2002). Unlike these bacteria, E. coli and 322 Salmonella lack systems to produce inducible AmpC enzymes. Mutations in the ampC promoter have 323 increased the resistance to oxyimino-cephalosporins in E. coli (Caroff et al., 1999). 324 Plasmids are considered to be facilitators for disseminating β -lactamase genes between various 325 species such as P. mirabilis, Achromobacter, Salmonella and E. coli (Bobrowski et al., 1976;326 Levesque et al., 1982;Knothe et al., 1983;Bauernfeind et al., 1989). Molecular characterization of 327 MDR plasmids is essential, yet complicated, because these plasmids are very diverse and 328 promiscuous. The relatedness of plasmids can be analyzed using a PCR-based replicon typing 329 method or whole genome sequence analysis (Carattoli et al., 2005). In previous studies, bla  carrying plasmids found in either E. coli or S. enterica were most likely to belong to replicons I1 and 331 A/C (Carattoli, 2009). Our results are also in accordance with these findings. bla TEM genes have been 332 reported to be located on plasmids of various replicon types such as A/C, I1, K, ColE, H12, etc 333 (Carattoli, 2009 Salmonella. We do not know if the same holds true in the natural environment. 362 In conclusion, we identified MDR isolates of E. coli and Salmonella enterica with ESBL/AmpC 363 phenotypes, examined the sequences of the ESBL/AmpC genes in these isolates, and assayed their 364 ability to form biofilm in vitro. We isolated and identified MDR plasmids which were readily 365 transferred by conjugation between E. coli and Salmonella isolates. Results suggest the possibility of 366 natural HGT by conjugation between E. coli and Salmonella may readily occur in the poultry house 367 environment. 368 We would like to thank the poultry veterinarians and producers who voluntarily participated in the 370 CIPARS farm surveillance program and enabled data and sample collection. We are grateful to the 371 Chicken Farmers of Canada and the Alberta Chicken Producers for their valuable input to the 372 framework development and technical discussions. 373 This study was funded by Alberta Agriculture and Forestry (AAF) [grant number 2015R025R] with 374 significant in a kind support from PHAC and the AAF, Agri-Food Laboratories. 375

Conflict of Interest 376
The authors declare that the research was conducted in the absence of any commercial or financial 377 relationships that could be construed as a potential conflict of interest.       fox-F 5'-TGTGGACGGCATTATCCAG-3' fox-R 5'-AAAGCGCGTAACCGGATTG-3' ent-F 5'-AGTAAAACCTTCACCTTCACCG-3' ent-R 5'-ATGCGCCTCTTCCGCTTTC-3' tem-F 5'-ATGAGTATTCAACATTTCCGT-3' tem-R 5'-TTACCAATGCTTAATCAGTGA-3' x-axis is the number of drug classes that isolates showed resistance to. The y-axis is the code of individual farms participating in this project. 600 The size of each dot represents the number of isolates obtained in each farm. The color represents whether the isolates are E. coli or 601 Salmonella strains. 602