Bacteria Broadly-Resistant to Last Resort Antibiotics Detected in Commercial Chicken Farms

Resistance to last resort antibiotics in bacteria is an emerging threat to human and animal health. It is important to identify the source of these antimicrobial resistant (AMR) bacteria that are resistant to clinically important antibiotics and evaluate their potential transfer among bacteria. The objectives of this study were to (i) detect bacteria resistant to colistin, carbapenems, and β-lactams in commercial poultry farms, (ii) characterize phylogenetic and virulence markers of E. coli isolates to potentiate virulence risk, and (iii) assess potential transfer of AMR from these isolates via conjugation. Ceca contents from laying hens from conventional cage (CC) and cage-free (CF) farms at three maturity stages were randomly sampled and screened for extended-spectrum β-lactamase (ESBL)-producing Enterobacteriaceae, carbapenem-resistant Acinetobacter (CRA), and colistin resistant Escherichia coli (CRE) using CHROMagar™ selective media. We found a wide-spread abundance of CRE in both CC and CF hens across all three maturity stages. Extraintestinal pathogenic Escherichia coli phylogenetic groups B2 and D, as well as plasmidic virulence markers iss and iutA, were widely associated with AMR E. coli isolates. ESBL-producing Enterobacteriaceae were uniquely detected in the early lay period of both CC and CF, while multidrug resistant (MDR) Acinetobacter were found in peak and late lay periods of both CC and CF. CRA was detected in CF hens only. blaCMY was detected in ESBL-producing E. coli in CC and CF and MDR Acinetobacter spp. in CC. Finally, the blaCMY was shown to be transferrable via an IncK/B plasmid in CC. The presence of MDR to the last-resort antibiotics that are transferable between bacteria in food-producing animals is alarming and warrants studies to develop strategies for their mitigation in the environment.


Introduction
The rise of antimicrobial resistant (AMR) bacteria is a serious threat to human and animal health, as increasing resistance to commonly used antibiotic therapies have created a burden on treatment options [1]. AMR bacteria can arise in nature and are commonly found in food producing animals like poultry [2]. For instance, the gastrointestinal tract (GIT) of chickens and the facilities that house these poultry serve as reservoirs for AMR resistant bacteria [3,4]. Of these AMR bacteria, extended-spectrum β-lactamase (ESBL) producing Enterobacteriaceae, carbapenem-resistant Acinetobacter (CRA), and colistin resistant E. coli (CRE) are emerging AMR bacteria found in the poultry environment [5][6][7]. The Centers for Disease Control and Prevention (CDC) lists ESBL-producing Enterobacteriaceae and CRA as serious and urgent threats, respectively [8]. Although not explicitly listed as an antibiotic resistant threat by the CDC, colistin resistance is clinically relevant given its use as a last resort antibiotic for treating multidrug resistant (MDR) infections [9]. As poultry is one of the most consumed meat sources globally [10], it is crucial to identify the presence and

Source Material
Ceca contents from hens in two commercial farms (CC and CF) located in Iowa were previously sampled, flash frozen, and stored at −80 • C between May and September of 2017 [24]. Samples were collected from three laying stages: (i) early lay (17-23 weeks), (ii) peak lay (25-39 weeks), and (iii) late lay (64-88 weeks) from the CC and CF commercial farms in Iowa. A total of 20 hens were randomly sampled for each maturity group in each facility (N = 20 × 3 maturities × 2 facilities = 120 total). The samples were analyzed as described in (Figure 1) and detailed below. Figure 1. Graphic description of sample analysis. Ceca contents from hens at different maturity stages (early, peak, and late) from conventional cage and cage-free poultry farms were analyzed for (A) resistance to clinically important antibiotics, (B) E. coli virulence genes and phylotype, and (C) plasmid profiles and transferability.

Identification of Last Resort Antibiotic-Resistant Isolates
Initially, 0.2 g ceca content samples were resuspended in 200 μL of sterile PBS, and 20 μL of suspensions were then added to CHROMagar COL-APSE™ (Paris, France) media for the selection of colistin-resistant bacteria. For enrichment of AMR bacteria, original suspensions were mixed with 500 μL Luria-Bertani broth (0.1% glucose) and incubated overnight at 37 °C. Thereafter, 20 μL of enriched suspensions were pipetted onto CHRO-Magar Orientation™ agar (Paris) to detect ESBL-producing Enterobacteriaceae and CHRO- Figure 1. Graphic description of sample analysis. Ceca contents from hens at different maturity stages (early, peak, and late) from conventional cage and cage-free poultry farms were analyzed for (A) resistance to clinically important antibiotics, (B) E. coli virulence genes and phylotype, and (C) plasmid profiles and transferability.

Identification of Last Resort Antibiotic-Resistant Isolates
Initially, 0.2 g ceca content samples were resuspended in 200 µL of sterile PBS, and 20 µL of suspensions were then added to CHROMagar COL-APSE™ (Paris, France) media for the selection of colistin-resistant bacteria. For enrichment of AMR bacteria, original suspensions were mixed with 500 µL Luria-Bertani broth (0.1% glucose) and incubated overnight at 37 • C. Thereafter, 20 µL of enriched suspensions were pipetted onto CHROMagar Orientation™ agar (Paris) to detect ESBL-producing Enterobacteriaceae and CHROMagar SuperCARBA™ (Paris) to detect carbapenem-resistant bacteria. Plates were incubated at 37 • C for up to 48 h to allow for the growth of slow growing bacteria like Acinetobacter spp. [27]. Initially, isolates were speculatively speciated by color change on the CHRO-Magar™ media according to manufacturer's instructions. Colistin resistant bacteria were enumerated, and cultures from individual colonies were stored in glycerol solution at −80 • C for future experiments. Colistin resistance was confirmed by growth on Mueller-Hinton agar (MHA) (4 mg/L colistin). Colonies from CHROMagar OrientationTM were later tested on CHROMagar COL-APSE™ and SuperCARBA™ to evaluate multiple resistance. Colonies from CHROMagar SuperCARBA™ were later tested on CHROMagar COL-APSE™ and Orientation™ respectively to evaluate multiple resistance. DNA was extracted from all colistin resistant and ESBL-producing bacteria as described [28] and were screened via polymerase chain reaction (PCR) for the E. coli housekeeping gene, uidA (Table S1). White/cream colonies isolated from CHROMagar Orientation™ and Super-CARBA™ media were speculatively designated as Acinetobacter spp., DNA was extracted, and isolates were confirmed by PCR amplification of the Acinetobacter spp. specific 16S rRNA gene (Table S1) [29]. All positively identified CRE, ESBL-producing E. coli and Acinetobacter, and CRA were thereafter screened for resistance to rifampicin (100 µg/mL), tetracycline (15 µg/mL), nalidixic acid (30 µg/mL), ampicillin (50 µg/mL), kanamycin (50 µg/mL), gentamicin (20 µg/mL), and chloramphenicol (20 µg/mL). β-lactamase production was confirmed through resistance to third generation cephalosporins (cefotaxime and ceftazidime) via Clinical and Laboratory Standards Institute (CLSI) guidelines [30]. Furthermore, minimum inhibitory concentration (MIC) of colistin, cefotaxime, and ceftazidime were obtained via agar dilution using CLSI breakpoints [28]. For all antibiotic resistance assays, E. coli K-12 strain MG1655 was used as a negative control. PCRs were performed to identify the presence of various mcr genes, multiple β-lactamase encoded genes, and carbapenemase encoded genes (Table S1). ESBL and mcr multiplexes were utilized as described previously [31,32]. NCBI Primer-blast was used to generate bla TEM primer pairs (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). Positive control strains used in this portion of the study were supplied from the CDC & FDA Antibiotic Resistance Bank (Table S2).

Plasmid Profiling and Typing
CRE, ESBL-producing E. coli and Acinetobacter, and CRA isolates were streaked onto LB agar plates and incubated overnight. Individual colonies from overnight plates were suspended in 3 mL LB broth (0.1% glucose) and shaken overnight at 37 • C. Plasmid extraction was performed using phenol: chloroform-based method as described previously [33]. The resulting plasmid extracts were loaded into 0.5% TAE agarose gels for agarose gel electrophoresis. Gels were run overnight at 40 V and 4 • C, stained with ethidium bromide, and imaged using a c300 imager (Azure Biosystems). Approximate band sizes were calculated using GelAnalyzer software (GelAnalyzer 19.1). All CRE and ESBL-producing E. coli were subjected to plasmid replicon typing for the detection of 18 common E. coli incompatibility groups (IncB/O, FIC, A/C, P, T, K/B, W, FIIA, FIA, FIB, Y, I1, Frep, X, HI1, N, HI2, and L/M) via multiplex PCR [34].

PCR Screenings for ExPEC
PCRs were performed to identify the common E. coli phylogenetic groups (A, B1, B2, and D) and the presence of virulence factors associated with ExPEC plasmids (cvaC, iroN, iss, and iutA) (Table S1).

Statistics
Prism software version 6.0 (GraphPad, San Diego, CA, USA) was used to calculate significance for all statistical analyses. One-way ANOVA followed by Turkey's test for multiple comparisons of means was used to compare differences between groups. p values < 0.05 were considered significant.

Widespread Antibiotic Resistance Detected in Poultry Fecal Isolates
The breakdown of the total numbers of CRE, ESBL-producing E. coli and Acinetobacter, and CRA can be found in Table 2. Colistin resistant bacteria were detected, whereas ESBL-producing E. coli and Acinetobacter isolates were not detected until after enrichment. Colistin resistance was widely observed in bacteria from both CC and CF environments on CHROMagar COL-APSE™ plates ( Figure 2). There were no significant differences (p > 0.05) in colistin-resistant bacteria abundances between the three laying periods within either environment. Of the total colonies that were resistant to colistin, 3.56% in CC and 0.69% in CF were identified as CRE (Table 2). A total of 52 CRE were isolated from CC and 48 from CF for future experiments. A total of 29 (60%) CRE in CC and 27 (47%) CRE in CF were shown to be resistant to colistin at concentrations up to 64 mg/L (Table S3). Using CHROMagar Orientation™ plates post-enrichment, ESBL-production was relatively uncommon in both CC and CF isolates regardless of laying period. Of ESBL-producing isolates in the CC environment, 48% were identified as E. coli and 31% were identified as Acinetobacter spp. (Table 2). Of the ESBL-producing isolates from the CF environment, 3% were identified as E. coli and no Acinetobacter isolates were identified. Using CHROMagar SuperCARBA™ plates post-enrichment, carbapenem-resistant bacteria were solely identified in the CF environment. Of these isolates, 13% were identified as Acinetobacter. Importantly, the CRA isolates were shown to be resistant to both colistin and β-lactams, and were found in feces from CF hens in peak and late lay periods. All ESBL-producing E. coli and CRA were used for further experimentation. The minimum inhibitory concentration (MIC) of all CRE, ESBL-producing E. coli, and Acinetobacter to colistin, cefotaxime, and ceftazidime are listed in Supplementary Table S3. in CF were shown to be resistant to colistin at concentrations up to 64 mg/L (Table S3). Using CHROMagar Orientation™ plates post-enrichment, ESBL-production was relatively uncommon in both CC and CF isolates regardless of laying period. Of ESBL-producing isolates in the CC environment, 48% were identified as E. coli and 31% were identified as Acinetobacter spp. (Table 2). Of the ESBL-producing isolates from the CF environment, 3% were identified as E. coli and no Acinetobacter isolates were identified. Using CHROMagar SuperCARBA™ plates post-enrichment, carbapenem-resistant bacteria were solely identified in the CF environment. Of these isolates, 13% were identified as Acinetobacter. Importantly, the CRA isolates were shown to be resistant to both colistin and β-lactams, and were found in feces from CF hens in peak and late lay periods. All ESBL-producing E. coli and CRA were used for further experimentation. The minimum inhibitory concentration (MIC) of all CRE, ESBL-producing E. coli, and Acinetobacter to colistin, cefotaxime, and ceftazidime are listed in Supplementary Table S3.  Antibiotic resistance profiles for all CC and CF CRE, ESBL-producing E. coli, and Acinetobacter isolates are detailed in Figures 3 and 4, respectively. Tetracycline resistance was  Antibiotic resistance profiles for all CC and CF CRE, ESBL-producing E. coli, and Acinetobacter isolates are detailed in Figures 3 and 4, respectively. Tetracycline resistance was widely distributed in 45 (69%) and 31 (56%) of AMR bacteria in CC and CF, respectively. A total of 20 CC and 13 CF E. coli isolates were MDR against a combination of three or more antibiotics. CC and CF Acinetobacter were shown to be MDR as they were resistant to all antibiotics tested. CC Acinetobacter were not able to grow on CHROMagar SuperCARBA™ media like their CF counterparts (Figures 3 and 4). Using PCR to identify AMR genes corresponding to resistance mechanisms in these isolates, we found no isolates carrying any of the mcr genes tested. Additionally, all ESBL-producing CC and CF E. coli carried the bla CMY gene, which encodes the Class C β-lactamase AmpC ( Figure 5). Furthermore, only CC AmpC-producing E. coli carried the bla TEM gene, which encodes the Class A ESBL TEM. The MDR Acinetobacter from CC also carried the bla CMY gene. However, CRA isolates were tested negative for the β-lactamase and carbapenemase-expressing genes tested in this study. to all antibiotics tested. CC Acinetobacter were not able to grow on CHROMagar Super-CARBA™ media like their CF counterparts (Figures 3 and 4). Using PCR to identify AMR genes corresponding to resistance mechanisms in these isolates, we found no isolates carrying any of the mcr genes tested. Additionally, all ESBL-producing CC and CF E. coli carried the blaCMY gene, which encodes the Class C β-lactamase AmpC ( Figure 5). Furthermore, only CC AmpC-producing E. coli carried the blaTEM gene, which encodes the Class A ESBL TEM. The MDR Acinetobacter from CC also carried the blaCMY gene. However, CRA isolates were tested negative for the β-lactamase and carbapenemase-expressing genes tested in this study.

Multiple Plasmid Types Found in AMR E. coli Isolates
AMR E. coli isolates from CC and CF environments contained a variety of plasmids, ranging in size from 6 kb to 150 kb ( Figure 6). However, we did not successfully extract plasmids from Acinetobacter isolates under any conditions. Microorganisms 2021, 9, x FOR PEER REVIEW 9 of 17

Multiple Plasmid Types Found in AMR E. coli Isolates
AMR E. coli isolates from CC and CF environments contained a variety of plasmids, ranging in size from 6 kb to 150 kb ( Figure 6). However, we did not successfully extract plasmids from Acinetobacter isolates under any conditions.  3 and 4). IncFIB plasmids were detected in 38 (58%) and 27 (49%) of CC and CF AMR E. coli respectively (Figures 3 and 4). Furthermore, ESBL-producing E. coli from the CC environment all carried similar plasmids identified as IncK/B plasmids ( Figure 3). However, ESBL-producing E. coli from CF did not yield any identified plasmids or replicon types ( Figure 4). Replicon types IncA/C, FIIA, T, Frep, X, W, and HI2 were not detected in any isolates in this study.

Various ExPEC Phylogenetic Groups and Virulence Markers Genes Identified in AMR E. coli Isolates
Using PCR to predict virulence potential of AMR E. coli isolates, we found the ExPEC virulence factors iutA and iss to be widespread in CC ( Figure 3) and CF (Figure 4) AMR E. coli. A total of 28 (48%) and 20 (34%) of CC AMR E. coli carried iutA and iss, respectively. Similarly, a total of 38 (69%) and 17 (31%) of CF AMR E. coli carried iutA and iss, respectively. Furthermore, E. coli phylogenetic typing yielded both B2 and D groups known for their virulence and relation to ExPEC (Figures 3 and 4) [39]. Specifically, 8 (14%) AMR E. coli were identified as B2 and 10 (17%) as D in the CC environment. Similarly, 12 (24%) AMR E. coli were identified as B2 and 4 (8%) as D in the CF environment.  (Figures 3 and 4). IncFIB plasmids were detected in 38 (58%) and 27 (49%) of CC and CF AMR E. coli respectively (Figures 3 and 4). Furthermore, ESBL-producing E. coli from the CC environment all carried similar plasmids identified as IncK/B plasmids (Figure 3). However, ESBL-producing E. coli from CF did not yield any identified plasmids or replicon types (Figure 4). Replicon types IncA/C, FIIA, T, Frep, X, W, and HI2 were not detected in any isolates in this study.

Various ExPEC Phylogenetic Groups and Virulence Markers Genes Identified in AMR E. coli Isolates
Using PCR to predict virulence potential of AMR E. coli isolates, we found the ExPEC virulence factors iutA and iss to be widespread in CC ( Figure 3) and CF (Figure 4) AMR E. coli. A total of 28 (48%) and 20 (34%) of CC AMR E. coli carried iutA and iss, respectively. Similarly, a total of 38 (69%) and 17 (31%) of CF AMR E. coli carried iutA and iss, respectively. Furthermore, E. coli phylogenetic typing yielded both B2 and D groups known for their virulence and relation to ExPEC (Figures 3 and 4) [39]. Specifically, 8 (14%) AMR E. coli were identified as B2 and 10 (17%) as D in the CC environment. Similarly, 12 (24%) AMR E. coli were identified as B2 and 4 (8%) as D in the CF environment.

bla CMY Can Be Exchanged between Virulent and Non-Virulent E. coli through an IncK/B Plasmid
Using in vitro conjugation assays to uncover a mechanism for β-lactam resistance dissemination, ESBL-producing E. coli (IA-EC-0010 and IA-EC-0018) isolates from CC were able to transfer the resistance of both cefotaxime and ceftazidime to both APEC strain χ7122 and its plasmid-cured derivative χ7368, but not to E. coli K-12 strain χ6092, MG1655, or human commensal HS-4. The AMR transfer was linked with the consistent transfer of a large 106 kb IncK/B plasmid (pIA-EC-0018-2) carrying the bla CMY gene ( Figure 5). However, using χ7122 (pIA-EC-0018-2) transconjugants from the previous assay as donors resulted in successful transfer of this large plasmid to χ6092 (Figure 7). The virulence markers iss and iutA that were observed in the donor strains were not transferred to the recipients. CF E. coli were unable to transfer bla CMY under any conditions tested in this study.

blaCMY Can Be Exchanged between Virulent and Non-Virulent E. coli through an IncK/B Plasmid
Using in vitro conjugation assays to uncover a mechanism for β-lactam resistance dissemination, ESBL-producing E. coli (IA-EC-0010 and IA-EC-0018) isolates from CC were able to transfer the resistance of both cefotaxime and ceftazidime to both APEC strain χ7122 and its plasmid-cured derivative χ7368, but not to E. coli K-12 strain χ6092, MG1655, or human commensal HS-4. The AMR transfer was linked with the consistent transfer of a large 106 kb IncK/B plasmid (pIA-EC-0018-2) carrying the blaCMY gene ( Figure 5). However, using χ7122 (pIA-EC-0018-2) transconjugants from the previous assay as donors resulted in successful transfer of this large plasmid to χ6092 (Figure 7). The virulence markers iss and iutA that were observed in the donor strains were not transferred to the recipients. CF E. coli were unable to transfer blaCMY under any conditions tested in this study.

Discussion
Commensal bacteria like Enterobacteriaceae and pathogens isolated from the chicken GIT are commonly resistant to multiple antibiotics [40]. Commonly, E. coli isolates sourced from chickens have shown resistance to tetracycline, chloramphenicol, ampicillin, and streptomycin due to the indiscriminate use of these antibiotics as growth promoters in poultry production [41,42]. Concurrent with these studies, our results show the extensive antibiotic resistance patterns associated with E. coli from both CC and CF farms in all maturity stages. Furthermore, the amount of CRE isolated from each maturity stage did not differ. Alarmingly, numerous CRE isolates from both CC and CF showed MDR to multiple antibiotics tested. Recent research has shown that the maturity of layer hens plays a significant role in the diversity of the ceca microbiota [26]. Although the detection of CRE is increasing in the poultry environment [15,43,44], there is little information on how ma-

Discussion
Commensal bacteria like Enterobacteriaceae and pathogens isolated from the chicken GIT are commonly resistant to multiple antibiotics [40]. Commonly, E. coli isolates sourced from chickens have shown resistance to tetracycline, chloramphenicol, ampicillin, and streptomycin due to the indiscriminate use of these antibiotics as growth promoters in poultry production [41,42]. Concurrent with these studies, our results show the extensive antibiotic resistance patterns associated with E. coli from both CC and CF farms in all maturity stages. Furthermore, the amount of CRE isolated from each maturity stage did not differ. Alarmingly, numerous CRE isolates from both CC and CF showed MDR to multiple antibiotics tested. Recent research has shown that the maturity of layer hens plays a significant role in the diversity of the ceca microbiota [26]. Although the detection of CRE is increasing in the poultry environment [15,43,44], there is little information on how maturity can impact the presence of these AMR bacteria. Factors such as disease and antibiotic intervention can drive the emergence of AMR bacteria in the poultry environment [45]. As the presence of MDR bacteria from chickens increases, the detection of specific resistance genes illustrates a relationship between resistant bacteria and the spread of resistance mechanisms. Future studies should investigate the role that the environment and possible intervention strategies that can limit the emergence of colistin resistant bacteria.
Our study highlights the importance of detection methods for colistin resistant bacteria using media like CHROMagar COL-APSE™. Notably, CHROMagar COL-APSE™ adheres to the European Committee on Antimicrobial Susceptibility Testing (EUCAST). The EUCAST breakpoint for colistin resistance is strictly set at 2 mg/L, whereas CLSI resistance cutoff in the United States is 4 mg/L. All isolates that were initially selected from CHROMagar COL-APSE™ were identified as resistant on MHA according to the EUCAST breakpoint of 2 mg/L [46]. The number of isolates initially selected from the CHROMagar COL-APSE™ media was greatly reduced after screening using MHA with 4 mg/L colistin. Using both CHROMagar COL-APSE™ media and further selection using CLSI guidelines can be useful for accurate detection of resistant isolates. Although use of CHROMagar COL-APSE™ and further confirmation with either CLSI or EUCAST guidelines can identify multiple resistant bacteria, these methods do not accurately detect the presence of mcr genes [47]. PCR investigation of all confirmed colistin resistant bacteria should be performed to confirm the presence of the emerging mcr genes.
The emergence of mcr genes in multiple settings have become an increasing global threat [14,48]. Notably, the principle mcr-1 gene has been reported in over 40 countries to date [49]. None of the mcr genes tested in our study (mcr-1, mcr-2, mcr-3, mcr-4, and mcr-5) were detected in any of our CRE isolates. Although our isolates did not carry the mcr genes that are often spread through conjugative plasmids [50], our study highlights the ability of AMR E. coli from chickens to exhibit spontaneous resistance to colistin. There are several mechanisms that can contribute to the ability of bacteria to become spontaneously resistant to colistin. For instance, the most common resistance strategy involves the modification of the bacterial outer membrane through alteration in the lipopolysaccharide (LPS) and reduction in its negative charge as this negative charge is the target of colistin [51]. Furthermore, overexpression of efflux-pump systems and overproduction of capsule polysaccharides can enhance the resistance to colistin in Enterobacteriaceae [52,53]. The possibility of threatening AMR bacteria like ESBL-producing E. coli and MDR Acinetobacter to exhibit resistance to colistin spontaneously is concerning because of colistin's use as a last resort antibiotic [54].
As evidenced in this study, the detection of ESBL-producing E. coli and MDR Acinetobacter can be detected in the poultry environment. In order to understand the genetic attributes that contributed to the resistance of these isolates, we investigated multiple βlactamase and carbapenemase producing genes in both groups. Specifically, we attempted to detect the bla TEM , bla SHV , bla CTX-M , and bla CMY β-lactamase genes as these β-lactamases are the most commonly identified [55]. Furthermore, we attempted to detect carbapenemase genes bla OXA , bla KPC , and bla VIM that are commonly found in CRA [56][57][58]. As none of the β-lactamase and carbapenemase genes were detected in the CRA isolates, other resistance mechanisms are likely responsible for the resistance. Carbapenem resistance in Acinetobacter can be mediated by reduced drug permeability through porin loss or modification and the overexpression of efflux pump [59,60]. Interestingly, we successfully identified the AmpC Class C β-lactamase encoding bla CMY gene in only the early lay period of both CC and CF environments. bla CMY has been increasingly reported in chickens as a source for the spread of AmpC to both human and avian pathogenic bacteria [61,62]. Although bla CMY is not classified as an extended-spectrum β-lactamase, the AmpC β-lactamase is resistant to otherwise useful β-lactamase inhibitors like clavulanic acid that is often paired with antibiotics, like ceftazidime, to limit the effects of β-lactamases [63,64]. Strikingly, in our study the ESBL gene bla TEM was detected in all cephalosporin resistant CC E. coli, illustrating the ability of AMR E. coli to harbor multiple β-lactamase producing genes.
To evaluate whether CRE and ESBL-producing E. coli detected in this study can be pathogenic, we tested our isolates using E. coli phylogenetic typing, a common technique utilized to sort E. coli isolates into groups that differ in ecological niches, life-history characteristics, and propensity to cause disease [65]. Namely, four main E. coli phylotypes (A, B1, B2, and D) are extensively used for classification [66][67][68]. We identified A, B2, and D groups in all maturity stages in both environments. Historically, phylogenetic group A is more commonly associated with commensal E. coli, whereas phylogenetic groups B2 and D are associated with virulent ExPEC infections [69,70]. For instance, phylogenetic groups B2 and D have been associated with E. coli isolates that cause urinary tract infections (UTI) and show increased presence of virulence [71] factors associated with UTI when compared to the B1 and A phylogenetic groups.
Additionally, because of their plasmidic location, we attempted to detect four ExPECassociated virulence factors (cvaC, iroN, iss, and iutA) among our isolates [33]. cvaC and iroN were not identified in any of our isolates. However, there was wide distribution of both iss and iutA in isolated from both environments. Interestingly, the late lay period of the CC isolates exhibited markedly less iss and iutA than the early and peak lay periods. Encoding the receptor for the siderophore aerobactin, iutA is an important factor in urinary pathogenic E. coli (UPEC) infections that allow the bacteria to competitively acquire iroN that would otherwise be acquired by the host [72]. Furthermore, the iss virulence factor encodes a specific outer membrane protein that increases serum survival for ExPEC during extraintestinal infection [73]. In addition to their role in the pathogenesis, these virulence factors could confer a competitive advantage to isolates that express them compared to their counterparts in the GIT.
The ability for E. coli to spread AMR genes through plasmids via HGT plays a significant role on the distribution of resistance in different bacteria. Several factors play a role in the ability of plasmids to be transferred between bacteria. For instance, incompatibility restriction, host genetics, and strain-specific factors can influence the transfer of plasmids [38,74,75]. In this study, we investigated a total of 18 plasmid incompatibility groups in all E. coli isolates. We were able to identify several isolates that carry IncN, P, HI1, and I1 incompatibility groups that have been shown to be associated with a broad-host range [76][77][78]. Interestingly, in our study, all ESBL-producing E. coli in the CC environment carried IncK/B type plasmids. Due to their multiple plasmids visualized in plasmid extraction, and presence of the common IncK/B plasmids, the CC ESBL-producing E. coli were selected for conjugation assays. CF E. coli that carried the bla CMY gene did not yield any of the plasmid replicon types investigated in this study, and conjugation was not observed under any conditions. Furthermore, CRA isolates did not yield any plasmid replicon types investigated in this study, and no plasmids were visualized using the method described or using IBI-Scientific (Dubuque, IA, USA) and Qiagen (Hilden, Germany) commercially available plasmid extraction kits. This suggests that the AMR genes of these isolates could be chromosomal rather than plasmidic.
In this study, we demonstrated the ability of the ESBL-producing E. coli from CC to transfer AmpC β-lactamase via a large 106 kb IncK/B plasmid (pAmpC) to a variety of hosts, including the APEC strain χ7122, its avirulent plasmid-cured derivative χ7368, and E. coli K-12 χ6092, indicating the ability of the plasmid to transfer to both virulent and non-virulent bacteria. Interestingly, numerous studies have highlighted the presence of pAmpC associated with both IncK plasmids and the poultry environment [79][80][81][82]. Although transfer of AMR genes and plasmidic virulence factors on the same plasmid has been shown [83,84], the ESBL-producing donors in our study did not transfer the virulence factors iss and iutA to the avirulent strains. This suggests that the plasmidic virulence factors are located on different plasmids than the pAmpC in the donor strains. Nonetheless, the presence of pAmpC in populations of MDR bacteria with the propensity to cause disease is concerning. Finally, increased detection of pAmpC can cause an eventual decline in the efficacy of cephalosporin antibiotics in both humans and poultry.

Conclusions
Overall, this study identified wide-spread colistin-resistant E. coli and MDR E. coli in all three lay periods in CC and CF hens. AmpC β-lactamase mediated bla CMY was identified in E. coli and Acinetobacter of CC and E. coli of CF. Furthermore, MDR Acinetobacter were detected in peak and late lay periods of both CC and CF, with CF isolates being also resistant to carbapenems. Finally, the AmpC β-lactamase mediated bla CMY was shown to be plasmidborne in the CC environment only. Future experimental studies will seek environmental and host factors that trigger dissemination of novel and threatening antibiotic resistant patterns to enable comparisons between CC and CF hens, as well as further sequencing of resistant isolates and transferable plasmids to understand the prevalent resistance genes in laying hens of different maturity stages.