*Article* **ESBL-Producing** *Escherichia coli* **Carrying CTX-M Genes Circulating among Livestock, Dogs, and Wild Mammals in Small-Scale Farms of Central Chile**

**Julio A. Benavides 1,2,3,\* , Marília Salgado-Caxito 3,4 , Andrés Opazo-Capurro 3,5 , Paulina González Muñoz 3,5,6, Ana Piñeiro <sup>7</sup> , Macarena Otto Medina <sup>1</sup> , Lina Rivas 3,8, Jose Munita 3,8 and Javier Millán 1,9,10**


**Abstract:** Antibiotic-resistant bacteria of critical importance for global health such as extendedspectrum beta-lactamases-producing (ESBL)-*Escherichia coli* have been detected in livestock, dogs, and wildlife worldwide. However, the dynamics of ESBL-*E. coli* between these animals remains poorly understood, particularly in small-scale farms of low and middle-income countries where contact between species can be frequent. We compared the prevalence of fecal carriage of ESBL-*E. coli* among 332 livestock (207 cows, 15 pigs, 60 horses, 40 sheep, 6 goats, 4 chickens), 82 dogs, and wildlife including 131 European rabbits, 30 rodents, and 12 Andean foxes sharing territory in peri-urban localities of central Chile. The prevalence was lower in livestock (3.0%) and wildlife (0.5%) compared to dogs (24%). Among 47 ESBL-*E. coli* isolates recovered, CTX-M-group 1 was the main ESBL genotype identified, followed by CTX-M-groups 2, 9, 8, and 25. ERIC-PCR showed no cluster of *E. coli* clones by either host species nor locality. To our knowledge, this is the first report of ESBL-*E. coli* among sheep, cattle, dogs, and rodents of Chile, confirming their fecal carriage among domestic and wild animals in small-scale farms. The high prevalence of ESBL-*E. coli* in dogs encourages further investigation on their role as potential reservoirs of this bacteria in agricultural settings.

**Keywords:** antimicrobial resistance; *bla*CTX-M; Chile; domestic animals; *E. coli*; extended-spectrum beta-lactamases; wildlife

#### **1. Introduction**

The current increase of antimicrobial resistance (AMR) is considered a main global threat to human and animal health [1,2]. AMR is responsible for thousands of human fatalities annually [3] and large economic losses that could reduce global GDP in 1–4% by 2050 [2,4]. The intense use of antibiotics in livestock production and humans is the main

**Citation:** Benavides, J.A.; Salgado-Caxito, M.; Opazo-Capurro, A.; González Muñoz, P.; Piñeiro, A.; Otto Medina, M.; Rivas, L.; Munita, J.; Millán, J. ESBL-Producing *Escherichia coli* Carrying CTX-M Genes Circulating among Livestock, Dogs, and Wild Mammals in Small-Scale Farms of Central Chile. *Antibiotics* **2021**, *10*, 510. https://doi.org/ 10.3390/antibiotics10050510

Academic Editor: Piera Anna Martino

Received: 6 April 2021 Accepted: 24 April 2021 Published: 30 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

cause of the emergence and rapid spread of AMR [2,5]. In the last decade, the global growth of livestock has been associated with an increase in antibiotics use [2]. For example, 70% of antibiotics used in human medicine are consumed by animal production in the USA [6,7]. Extended-spectrum beta-lactamase-producing *Escherichia coli* (ESBL-*E. coli*) represent one of the highest burdens of AMR to public health and have globally spread in both hospital settings and the community [8]. ESBL-*E. coli* are commonly isolated from domestic animals such as cattle and dogs, but also wild animals [9–12]. Similar to humans, the misuse of third-generation cephalosporins in livestock generated a selective pressure resulting in the emergence and spread of ESBL-*E. coli* in this sector [9,13]. In contrast, the presence of ESBL-*E. coli* in wildlife is assumed to result from contamination in human-dominated environments [10,12,14].

The circulation of ESBL-*E. coli* across different animal populations requires an integrated One Health approach to better understand, predict, and prevent their dissemination [15]. However, most studies on ESBL-*E. coli* have focused on either one population (e.g., domestic or wild animals) or a large spatial scale (e.g., across cities or countries) [16–19]. For example, ESBL-*E. coli* have been detected worldwide in several livestock settings [13,20–22]. Likewise, ESBL-producing *Enterobacterales* have been found in at least 80 wildlife species since 2006 including rodents, bats, foxes, and wild birds [23–26]. Livestock or human proximity are often suggested as drivers of ESBL-*E. coli* in wildlife but, to our knowledge, no study has proven transmission from humans to wild animals [10,14,23]. Dogs living on farms could also contribute to the spread of ESBL-*E. coli* among agricultural settings because contact with livestock has been associated with an increased probability of ESBL-*E. coli* fecal carriage in dogs [27–30]. However, the circulation of ESBL-*E. coli* at the livestock and wildlife interface is still poorly understood [12,31,32].

Few studies on the circulation of ESBL-*E. coli* at the livestock and wildlife interface have been conducted in low- and middle-income countries (LMICs) [10,33–36]. Paradoxically, the consequences of AMR can be exacerbated in these countries by a higher number of bacterial infections and limited access to health facilities providing the appropriate antibiotic treatment [37,38]. Surveillance of AMR in livestock has been recommended by the World Health Organization (WHO), the Food and Agriculture Organization of the United Nations (FAO), and the World Organisation for Animal Health (OIE), but remains limited in LMICs [1,2,33]. Surveillance of AMR in wildlife and dogs is also mostly inexistent in LMICs. In this study, we use a One Health approach to compare the prevalence of ESBL-*E. coli* fecal carriage among livestock, dogs, and wild mammals located in small-scale agricultural settings of central Chile.

Chile, considered a high income economy but with an agricultural production more similar to LMICs, launched the 'National plan to combat antimicrobial resistance' in 2017, but no national surveillance has been implemented yet in the agricultural sector. ESBL-*E. coli* have not been detected in Chilean cattle herds [39,40], but have been isolated in feces from dogs [41], owls in rehabilitation centers [42], wild Andean condors (*Vultur gryphus*) [43] and gulls (*Leucophaeus pipixcan*) [36]. To our knowledge, no study has investigated the ESBL-*E. coli* fecal carriage of livestock nor simultaneously focused on dogs and wild mammals living closely to livestock. Central Chile hosts a large diversity of endemic terrestrial mammals including foxes and rodents [44,45] but also invasive species such as the European rabbit (*Oryctolagus cuniculus*) that has colonized most of the country [46–48]. Rodents and rabbits are commonly found living on farms and interacting with dogs and livestock [49,50]. Similarly, 85% of the territory of the Andean fox (*Lycalopex culpaeus*) overlaps with human-dominated habitat in central Chile [51]. This creates the potential for fecal-oral and environmental bacterial transmission between livestock and wild animals, which remains largely unknown. Previous studies focusing on foxes in the central region have identified the presence of *bla*CTX-M genes, but the bacteria carrying the gene was unknown [52]. The aims of this study were (i) to estimate and compare the prevalence of ESBL-*E. coli* fecal carriage between livestock, dogs, and wild mammals living in the same agricultural setting of central Chile, (ii) to detect the presence of the most common ESBL

genes including *bla*CTX-M, *bla*TEM, and *bla*SHV, and (iii) use high resolution molecular typing to assess potential ESBL-*E. coli* transmission within farms or between different species. **2. Materials and Methods**  *2.1. Sample Collection* 

ESBL-*E. coli* transmission within farms or between different species.

largely unknown. Previous studies focusing on foxes in the central region have identified the presence of *bla*CTX-M genes, but the bacteria carrying the gene was unknown [52]. The aims of this study were (i) to estimate and compare the prevalence of ESBL-*E. coli* fecal carriage between livestock, dogs, and wild mammals living in the same agricultural setting of central Chile, (ii) to detect the presence of the most common ESBL genes including *bla*CTX-M, *bla*TEM, and *bla*SHV, and (iii) use high resolution molecular typing to assess potential

*Antibiotics* **2021**, *10*, x FOR PEER REVIEW 3 of 14

#### **2. Materials and Methods** Fresh fecal samples were collected between March 2019 and September 2019 from

#### *2.1. Sample Collection* livestock, dogs, and wildlife in and around 13 farming localities located in the municipal-

Fresh fecal samples were collected between March 2019 and September 2019 from livestock, dogs, and wildlife in and around 13 farming localities located in the municipalities of Colina (33.1045◦ S, 70.6159◦ W) and Lampa (33.2827◦ S, 70.8793◦ W) of the Chacabuco province in the Metropolitan Region of central Chile, in the peri-urban area of the Santiago Capital City (Figure 1). A farming locality was either a single private farm or an area where livestock from different owners grazed together and received the same health treatments. The province of Chacabuco includes mainly small- to medium-scale farmers, with an estimated livestock population of 10,662 cattle (mean: 38 animals/farm), 45,821 pigs (587/farm), 5490 goats (59/farm), 4441 sheep (42/farm), and 2897 horses (4/farm) [53]. Farms were randomly selected from a list provided by the Municipality's agrarian unit, accounting for areas overlapping with the known territory of wildlife as previously described [52]. Our sampling focused mainly on cattle because they had the highest potential of overlapping with wild mammals since they often free-ranged within wildlife habitat during our study period. ities of Colina (33.1045° S, 70.6159° W) and Lampa (33.2827° S, 70.8793° W) of the Chacabuco province in the Metropolitan Region of central Chile, in the peri-urban area of the Santiago Capital City (Figure 1). A farming locality was either a single private farm or an area where livestock from different owners grazed together and received the same health treatments. The province of Chacabuco includes mainly small- to medium-scale farmers, with an estimated livestock population of 10,662 cattle (mean: 38 animals/farm), 45,821 pigs (587/farm), 5490 goats (59/farm), 4441 sheep (42/farm), and 2897 horses (4/farm) [53]. Farms were randomly selected from a list provided by the Municipality's agrarian unit, accounting for areas overlapping with the known territory of wildlife as previously described [52]. Our sampling focused mainly on cattle because they had the highest potential of overlapping with wild mammals since they often free-ranged within wildlife habitat during our study period.

**Figure 1.** Study area. The inset figure shows the Chacabuco province within the Metropolitan region where farms and wildlife were sampled. Exact farm locations are not given to maintain our confidentiality agreement with farmers. Maps were obtained from the GADM **Figure 1.** Study area. The inset figure shows the Chacabuco province within the Metropolitan region where farms and wildlife were sampled. Exact farm locations are not given to maintain our confidentiality agreement with farmers. Maps were obtained from the GADM (http://www.gadm. org//, accessed on 15 April 2021) database using the *getData* function from the *raster* package of R.

We focused on sampling the most common wild mammals encountered in those farms including several species of endemic and invasive rodents, the invasive European wild rabbit and the Andean fox, who predates these herbivore species [54,55]. These species were previously determined by discussions with farmers and the municipality's agrarian unit during preliminary visits to the farms. Peri-urban and wild rodents were live captured, sampled, and released using Sherman traps. Fifty traps were placed in and around each sampled farm for at least 4 consecutive days and checked for captured rodents daily. Rectal swabs were collected from alive individuals immobilized, using gloves and protective equipment. Rodents were identified at the genus or species level based on morphological characteristics. Fresh fecal samples from European rabbits were collected early in the morning by identifying rabbit dens in areas where farmers commonly observed rabbits. To avoid sampling the same individual twice, we only collected fresh sample feces from the same den if they were more than 4 m apart, and only sampled each den once. Fresh fecal samples from foxes were collected by walking known paths where foxes were previously captured in the area [56]. Fresh samples from foxes were identified and differentiated from dog feces by their distinct 'fruit' seeds and morphology contained on the sample. To avoid sampling the same individual twice, we only collected a fresh sample in localities that were more than 5 km apart, considering 5 km<sup>2</sup> as the average home range size of foxes in this area [52]. Dogs were sampled by directly taking rectal swabs or waiting until the dog defecated, depending on whether the owner considered that the dog could be aggressive or not during sampling. For all samples taken from the ground, we only collected the portion that was not in contact with the ground to avoid bacterial contamination from the soil. This study was approved by the Ethical Committee of the Universidad Andrés Bello (permit number: 018/2018). The capture and sampling of rodents were also approved by the Servicio Agricola Ganadero (permit number: 2118/2019).

### *2.2. Sample Size and Prevalence Estimation*

The required sample size needed to estimate the prevalence of ESBL-*E. coli* in livestock (defined as the number of animals harboring at least one isolate of ESBL-*E. coli* over the total number of sampled animals) was calculated with the program Epi Info 7.2.2.6TM [57]. To our knowledge, no previous study has estimated the prevalence of fecal carriage of ESBL-*E. coli* among livestock in Chile. Thus, we assumed an expected prevalence of ESBL-*E. coli* of 30%, similar to a study conducted around the Lima capital in Peru with similar farm characteristics [12]. Based on this expected prevalence, a margin of acceptable error of 5% and a confidence interval of 95%, the minimum number of livestock to be sampled in the region was 323.

Based on previous studies on wildlife and dogs, we assumed an expected prevalence of 5% to estimate our sample size. In fact, 5% prevalence of ESLB-*E. coli* was found in wild rodents in China [34,58], no bacteria were found in a previous study conducted in European wild rabbit in Portugal [59], 4% prevalence was found in wild foxes of Portugal [60], and 8% was found in the only study conducted on dogs in Chile [41]. Based on an expected ESBL-*E. coli* prevalence of 5%, a margin of acceptable error of 5% and a confidence interval of 95%, the minimum number of animals to be sampled was 73. We aimed to collect 73 samples per wildlife group (e.g., foxes, rabbits, and rodents). However, giving the intrinsic lower density of foxes compared to small mammals and logistic constraints for finding foxes, we expected a much lower sample size for this species.

#### *2.3. Microbiology Analyses*

Fresh fecal samples were collected using Stuart Transport Medium (Deltalab®) and cultured within 3 days of sampling. Swabs were screened for cefotaxime non-susceptible *E. coli* by direct incubation in standard atmospheric conditions (100 kPa) at 37 ◦C for 24 h in a MacConkey medium containing 2 µg/mL of cefotaxime sodium salt (Sigma-Aldrich, St. Louis, MO, USA) [61]. Up to 3 isolates with different morphotypes compatible with *E. coli* per sample/plate were purified and then stored at −80 ◦C for further analyses. Bac-

terial species were confirmed by matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry (BioMérieux, Marcy l'Etoile, France) at the Genomics and Resistant Microbes (GeRM) Group of the Millennium Initiative for Collaborative Research on Bacterial Resistance (MICROB-R).

Cefotaxime non-susceptible *E. coli* isolates indicating ESBL were tested for antimicrobial susceptibility to 8 antibiotics from 6 classes including chloramphenicol (phenicol), ciprofloxacin (quinolone), sulfamethoxazole (sulfonamide), amikacin (aminoglycoside), tobramycin (aminoglycoside), ertapenem (carbapenem), tetracycline, and gentamicin (aminoglycoside). Multidrug resistance (MDR) was defined as resistance to at least 1 agent of 3 or more antibiotic classes [62]. The *E. coli* ATCC25922 strain was used for quality control and clinical breakpoints were in accordance with CLSI M100:28ED recommendations [61].

Extended-spectrum beta-lactamase production was confirmed in all cefotaxime nonsusceptible *E. coli* isolates by the double-disk synergy test [30] on Müller Hinton agar (Difco, BD, Sparks, MD, USA) with and without the AmpC inhibitor phenylboronic acid (Sigma-Aldrich). Briefly, disks of ceftriaxone (30 µg), ceftazidime (30 µg), cefepime (30 µg), and aztreonam (30 µg) were used along with a disk of amoxicillin with clavulanic acid (30 µg) placed in the center of the plate at approximately 20 mm. Inhibition zones (ghost zones) observed around any of the cephalosporin disks towards the disk containing the clavulanic acid after 18–20 h of incubation at 37 ◦C aerobically were considered as a positive result to produce ESBL.

The presence of the most common ESBL-encoding genes in *E. coli* isolates including *bla*CTX-M, *bla*TEM, and *bla*SHV, was tested by a previously described multiplex PCR [63]. DNA samples from reference *bla*CTX-M, *bla*TEM, and *bla*SHV strains stored at the Universidad de Concepción's Laboratory of Research in Antimicrobial Agents were used as positive PCR controls. The specific group of each CTX-M alleles (CTX-M groups 1, 2, 8, 9, and 25) were detected by multiplex-PCR as described previously [64]. In order to explore the phylogenetic relationships between ESBL-*E. coli* isolates within and between host species or localities, isolates were fingerprinted by ERIC-PCR according to Bilung et al. [65].

#### *2.4. Statistical Analyses*

The prevalence of ESBL-*E. coli* was reported and 95% confidence intervals were calculated using the *binom.confint* function (Agresti-Coull method) in the *binom* package in R 3.6.1 [66]. Significant differences in prevalence between populations were tested using the Fisher's exact test in R, since the limited number of observations prevented the use of a Chi-Squared test. We constructed a dendrogram based on the ERIC-PCR electrophoretic patterns using the BioNumerics software v8.0 (Applied Maths, Belgium) and R [65,66]. An UMPGA dendrogram was built based on scaled densitometry curves from the ERIC-PCR obtained from BioNumeric using the *hclust* function of the *dendextended* R package.

#### **3. Results**

ESBL-*E. coli* fecal carriage was detected in chickens, cattle, pigs, sheep, goats, dogs, and one wild rodent (*Octodon degus*). The prevalence of ESBL-*E. coli* fecal carriage was significantly higher among dogs (24% [CI: 16–35%]; 20 out of 82) compared to livestock (3% [CI: 2–6%]; 10 out of 324, Fisher's exact test, Odds Ratio (OR) = 10.0, *p* < 0.0001) and wildlife (0.5% [CI: 0–3%]; 1 out of 186, Fisher's exact test, OR = 58.8, *p* < 0.0001) (Figure 2). The prevalence of ESBL-*E. coli* in livestock was also significantly higher than the prevalence in wildlife (Fisher's exact test, OR = 25.4, *p* < 0.0001). At least 1 animal carrying ESBL-*E. coli* was detected in 7 out of the 13 (54%) farm localities sampled. In all 3 farms where livestock carried ESBL-*E. coli* and dogs were sampled, at least 1 dog also carried ESBL-*E. coli*. Likewise, the wild rodent carrying ESBL-*E. coli* was detected in a farm where one cow also carried ESBL-*E. coli*.

also carried ESBL-*E. coli*.

*coli*. Likewise, the wild rodent carrying ESBL-*E. coli* was detected in a farm where one cow

**Figure 2.** Prevalence of ESBL-*E. coli* per species in small-scale farms of central Chile; 95% confidence intervals were esti-**Figure 2.** Prevalence of ESBL-*E. coli* per species in small-scale farms of central Chile; 95% confidence intervals were estimated using the *binom.confint* function (Agresti-Coull method) in the *binom* package in R.

mated using the *binom.confint* function (Agresti-Coull method) in the *binom* package in R. A total of 47 ESBL-*E. coli* isolates (confirmed by the double-disk synergy test) from 33 animals were analyzed. Fourteen ESBL-*E. coli* isolates were obtained from 10 livestock, 32 isolates from dogs and 1 isolate from a mouse. ESBL-*E. coli* isolates from livestock were resistant to a median (mean) of 1 (2.6) (range: 0–6) out of 8 antibiotics tested, while ESBL-*E. coli* isolates from dogs were resistant to a median (mean) of 1 antibiotic (1.9) (range: 0– 6) (Figure 3A). Overall, 21% of ESBL-*E. coli* isolates from livestock and 31% from dogs were susceptible to all antibiotics, 36% of ESBL-*E. coli* isolates from livestock and 21% from dogs were resistant to one antibiotic, and 43% of ESBL-*E. coli* isolates from livestock and 48% from dogs were resistant to two or more antibiotics. Additionally, 43% of ESBL-*E. coli* isolates from livestock, 47% from dogs and an isolate from one rodent were multidrug resistant (MDR). The ESBL-*E. coli* isolated from a rodent sample was resistant to chloramphenicol, sulfamethoxazole, and ciprofloxacin. More than 20% of ESBL isolates were resistant to ciprofloxacin, chloramphenicol, sulfamethoxazole, and tetracycline in both dogs and livestock. In contrast, no resistance was observed against ertapenem. Among ESBL A total of 47 ESBL-*E. coli* isolates (confirmed by the double-disk synergy test) from 33 animals were analyzed. Fourteen ESBL-*E. coli* isolates were obtained from 10 livestock, 32 isolates from dogs and 1 isolate from a mouse. ESBL-*E. coli* isolates from livestock were resistant to a median (mean) of 1 (2.6) (range: 0–6) out of 8 antibiotics tested, while ESBL-*E. coli* isolates from dogs were resistant to a median (mean) of 1 antibiotic (1.9) (range: 0–6) (Figure 3A). Overall, 21% of ESBL-*E. coli* isolates from livestock and 31% from dogs were susceptible to all antibiotics, 36% of ESBL-*E. coli* isolates from livestock and 21% from dogs were resistant to one antibiotic, and 43% of ESBL-*E. coli* isolates from livestock and 48% from dogs were resistant to two or more antibiotics. Additionally, 43% of ESBL-*E. coli* isolates from livestock, 47% from dogs and an isolate from one rodent were multidrug resistant (MDR). The ESBL-*E. coli* isolated from a rodent sample was resistant to chloramphenicol, sulfamethoxazole, and ciprofloxacin. More than 20% of ESBL isolates were resistant to ciprofloxacin, chloramphenicol, sulfamethoxazole, and tetracycline in both dogs and livestock. In contrast, no resistance was observed against ertapenem. Among ESBL isolates, the prevalence of resistance to each antibiotic was highly correlated between livestock and dogs (Spearman's test, Rho = 0.90, *p* < 0.0001), but livestock had a slightly higher prevalence than dogs for most antibiotics (Figure 3B).

prevalence than dogs for most antibiotics (Figure 3B).

**Figure 3.** (**A**) Prevalence of resistance to other antibiotic families among ESBL-*E. coli* isolates in dogs and livestock; (**B**) Correlation of the prevalence of resistance to each antibiotic between livestock and dogs; (**C**) Prevalence of *bla*TEM, *bla*SHV, and *bla*CTX-M in ESBL-*E. coli* isolated from livestock and dogs; (**D**) Prevalence of CTX-M groups identified in ESBL-*E. coli*  isolates from livestock and dogs. **Figure 3.** (**A**) Prevalence of resistance to other antibiotic families among ESBL-*E. coli* isolates in dogs and livestock; (**B**) Correlation of the prevalence of resistance to each antibiotic between livestock and dogs; (**C**) Prevalence of *bla*TEM, *bla*SHV, and *bla*CTX-M in ESBL-*E. coli* isolated from livestock and dogs; (**D**) Prevalence of CTX-M groups identified in ESBL-*E. coli* isolates from livestock and dogs.

ESBL-*E. coli* isolates from dogs were only encoded by the CTX-M genotype while all isolates from livestock carried CTX-M (100%), followed by TEM (14%), and SHV (7%) genotypes (Figure 3C). Among the most common CTX-M groups searched, 93% of ESBL-*E. coli* from livestock carried *bla*CTX-M-group 1 and 36% carried *bla*CTX-M-group 2 genes (Figure 3D). Isolates from dogs carried a more diverse pool of CTX-M genotypes with 78% carrying CTX-M from group 1, followed by group 2 (63%), group 9 (12.5%), group 8 (3%, one isolate), and group 25 (3%). The ESBL-*E. coli* isolate found on a wild mouse carried CTX-M ESBL-*E. coli* isolates from dogs were only encoded by the CTX-M genotype while all isolates from livestock carried CTX-M (100%), followed by TEM (14%), and SHV (7%) genotypes (Figure 3C). Among the most common CTX-M groups searched, 93% of ESBL-*E. coli* from livestock carried *bla*CTX-M-group 1 and 36% carried *bla*CTX-M-group 2 genes (Figure 3D). Isolates from dogs carried a more diverse pool of CTX-M genotypes with 78% carrying CTX-M from group 1, followed by group 2 (63%), group 9 (12.5%), group 8 (3%, one isolate), and group 25 (3%). The ESBL-*E. coli* isolate found on a wild mouse carried CTX-M from group 1.

isolates, the prevalence of resistance to each antibiotic was highly correlated between livestock and dogs (Spearman's test, Rho = 0.90, *p* < 0.0001), but livestock had a slightly higher

from group 1. The dendrogram analysis of the ERIC-PCR results showed a high diversity of ESBL-*E. coli* clones within species and farm localities. No visual clustering by species nor farm localities was observed (Figure 4). However, ESBL-*E. coli* isolates from a cow and a dog The dendrogram analysis of the ERIC-PCR results showed a high diversity of ESBL-*E. coli* clones within species and farm localities. No visual clustering by species nor farm localities was observed (Figure 4). However, ESBL-*E. coli* isolates from a cow and a dog from the same farm locality clustered together.

from the same farm locality clustered together.

**Figure 4.** Dendrogram produced by the analysis of the ERIC-PCR of ESBL-*E. coli* isolates from livestock and dogs using the UMPGA method in R. The colored column on the right side represents different farm localities where isolates were recovered. **Figure 4.** Dendrogram produced by the analysis of the ERIC-PCR of ESBL-*E. coli* isolates from livestock and dogs using the UMPGA method in R. The colored column on the right side represents different farm localities where isolates were recovered.

#### **4. Discussion**

**4. Discussion**

sistance genes.

The spread of AMR at the interface between domestic animals and wildlife remains poorly understood, particularly in low-income rural areas without specific barriers to limit the interaction between domestic and wild animals. In this study, we simultaneously estimated the prevalence of ESBL-*E. coli* fecal carriage among livestock, dogs, and wild mammals among small-scale agricultural localities of central Chile. The prevalence of ESBL-*E. coli* fecal carriage was lower in livestock (3%) and wildlife (less than 1%) compared to dogs (24%), suggesting that dogs can be an important carrier of these bacteria in agricultural settings. Dogs carried ESBL-*E. coli* in the three farms where ESBL-*E. coli* were detected in livestock, highlighting the potential sharing of these bacteria between dogs and livestock. Among ESBL-*E. coli* isolates, five CTX-M groups including groups 1, 2, 8, 9, and 25 were detected, with most isolates carrying CTX-M group 1. Molecular typing of ESBL-*E. coli* by ERIC-PCR showed no cluster of isolates by neither species nor locality, suggesting a wide range of ESBL-*E. coli* strains circulating on agricultural settings and highlighting the potential for cross-species transmission of either bacteria or antibiotic re-The spread of AMR at the interface between domestic animals and wildlife remains poorly understood, particularly in low-income rural areas without specific barriers to limit the interaction between domestic and wild animals. In this study, we simultaneously estimated the prevalence of ESBL-*E. coli* fecal carriage among livestock, dogs, and wild mammals among small-scale agricultural localities of central Chile. The prevalence of ESBL-*E. coli* fecal carriage was lower in livestock (3%) and wildlife (less than 1%) compared to dogs (24%), suggesting that dogs can be an important carrier of these bacteria in agricultural settings. Dogs carried ESBL-*E. coli* in the three farms where ESBL-*E. coli* were detected in livestock, highlighting the potential sharing of these bacteria between dogs and livestock. Among ESBL-*E. coli* isolates, five CTX-M groups including groups 1, 2, 8, 9, and 25 were detected, with most isolates carrying CTX-M group 1. Molecular typing of ESBL-*E. coli* by ERIC-PCR showed no cluster of isolates by neither species nor locality, suggesting a wide range of ESBL-*E. coli* strains circulating on agricultural settings and highlighting the potential for cross-species transmission of either bacteria or antibiotic resistance genes.

ESBL-*E. coli* have been detected across livestock in South America, with prevalence in cattle ranging from 18% in Brazil to 48% in Peru [12,67]. In this study, we detected ESBL-*E. coli* fecal carriage in cattle, swine, sheep, and chicken, showing the widespread ESBL-*E. coli* have been detected across livestock in South America, with prevalence in cattle ranging from 18% in Brazil to 48% in Peru [12,67]. In this study, we detected ESBL-*E. coli* fecal carriage in cattle, swine, sheep, and chicken, showing the widespread dissemination of these bacteria in agricultural settings. This is the first report of ESBL-*E. coli* in cattle in Chile, although their prevalence was low (3%) compared to a similar study in

Peru estimating a prevalence of 48% among small-scale farmers in the Lima region [12]. The observed prevalence in Chile is similar to farms in high-income countries such as France or Denmark, where the restriction of third-generation cephalosporins has been associated with a reduction in ESBL-*E. coli* [68,69]. The high prevalence of resistance to ciprofloxacin (over 60%) found in ESBL-*E. coli* isolated from domestic animals in this study is consistent with the high level of plasmid-mediated quinolone resistant found in 74% of ESBL-*E. coli* isolated from Chilean hospitals [70] and a high prevalence of resistance to ciprofloxacin (84%) in ESBL-*E. coli* recovered from intensive care units of Southern Chile [71]. The presence of ESBL-*E. coli* could result from low but existing selective pressure by the use of third generation cephalosporins in these farms, which requires further investigation. In a similar agricultural setting of Peru, the low use of cephalosporins [72] was associated to a high prevalence of ESBL-*E. coli* in livestock (50%) [12], suggesting that factors other than antibiotic use can influence AMR. For example, farm hygiene, herd size, contact with humans or other husbandry conditions such as storage of slurry in a pit have been associated with the presence of ESBL-*E. coli* in livestock [13,20,21].

The low prevalence of ESBL-*E. coli* in wildlife (less than 1%) is similar to other studies focusing on ESBL-*E. coli* among wildlife in Latin America and other LMICs [12,73]. For example, a previous study estimated a 4% prevalence of ESBL-*E. coli* among vampire bats (*Desmodus rotundus*) in Peru using a similar methodology for screening [12]. Previous studies conducted in Chile and Latin America have detected the presence of ESBL-*E. coli* on wild birds including gulls [36], Andean condors [43], and three species of owls [42]. Likewise, *bla*CTX-M genes have been previously detected using qPCR methods from feces in Andean foxes [52] and the guiña (*Leopardus guigna*) [74], although the bacteria species carrying the genes, and whether it was expressed or not, remains unknown. To our knowledge, this is the first study to report *E. coli* carrying CTX-M group 1 on wild mammals in Chile. The origin of ESBL-*E. coli* found in a rodent remains to be clarified. Given the presence of similar *bla*CTX-M genes among a nearby farm and a wide variety of ESBL-*E. coli* strains circulating, one potential explanation is the transmission of *bla*CTX-M from domestic animals, although other potential contamination sources (e.g., humans, water contamination) cannot be discarded.

The high prevalence of ESBL-*E. coli* found in dogs (24%) highlights their role as either passive 'receivers' or reservoirs of ESBL-*E. coli* in agricultural settings. Although there are only a limited number of studies estimating the prevalence of ESBL-*E. coli* among dogs, previous studies have shown a prevalence in Latin American dogs ranging from 9–30%, and a global prevalence of 7% [30,75–79]. The detection of ESBL-*E. coli* in dogs has been associated with previous antibiotic treatment, but also close contact with livestock, implying the potential transmission of these bacteria between livestock and dogs [29,30,80]. The latest is also suggested by our study, as the three farms where we detected ESBL-*E. coli* in livestock also had a dog carrying ESBL-*E. coli*. Molecular typing by ERIC-PCR showed no cluster of ESBL-*E. coli* by host species, while isolates sampled from a cow and a dog at the same farm clustered together. These results suggest that bacterial strains or ESBL genes such as *bla*CTX-M could be exchanged between host populations. Overall, the circulation of ESBL-*E. coli* among dogs highlights the potential public health risk for domestic animals but also for dog owners, given the potential spillover of bacteria from dogs to humans [28,29,81]. Moreover, the higher prevalence observed in dogs compared to livestock suggests that ESBL-*E. coli* could be spreading from dogs to livestock, and not necessarily in the other direction, as most previous studies have assumed.

Our study constitutes one of the first One Health approaches to simultaneously address the circulation of ESBL-*E. coli* among livestock, dogs, and wildlife in a rural setting. However, several future research can complement our findings and provide further insight into the selection and spread of AMR among these compartments. First, the limited sample size of foxes prevented a more accurate estimation of ESBL-*E. coli* prevalence in this species. Thus, we could not conclude whether predators or preys are more likely to carry ESBL-*E. coli* in this setting. Secondly, the low selective pressure for ESBL-*E. coli* should be

confirmed by studies on antibiotic use among farmers in these agricultural settings [72], which are currently lacking in Chile. Although the use of antibiotics in Chilean terrestrial livestock remains unknown, the national health authority (Servicio Agricola Ganadero) advises the use of fluroquinolones and cephalosporins as a last resource antibiotic in livestock, following a susceptibility test [82]. Antibiotic residues of tetracyclines, betalactams, aminoglycosides, and macrolides have been found in eggs from backyard poultry production [83]. Thirdly, although the ERIC-PCR technique used has a high resolution and allows us to differentiate among *E. coli* strains from the same locality and host species [65], several other molecular techniques can improve our understanding of the transmission dynamics of resistance genes and *E. coli*. For example, future work could determine the pathogenic potential of these strains using whole genome sequencing, or whether *bla*CTX-M genes are carried by specific mobile elements such as plasmids. Finally, future research should identify associated factors to ESBL-*E. coli* fecal carriage in each animal population (e.g., individual characteristics of dogs and cattle).

**Author Contributions:** Conceptualization, J.A.B.; Data curation, J.A.B.; Formal analysis, J.A.B. and A.O.-C.; Funding acquisition, J.A.B.; Investigation, J.A.B., M.S.-C. and A.O.-C.; Methodology, J.A.B., M.S.-C., A.O.-C., P.G.M., A.P., M.O.M., L.R., J.M. (Jose Munita) and J.M. (Javier Millán); Project administration, J.A.B.; Resources, J.A.B.; Software, J.A.B.; Supervision, J.A.B.; Validation, J.A.B.; Visualization, J.A.B.; Writing—original draft, J.A.B.; Writing—review and editing, J.A.B., M.S.-C. and J.M. (Javier Millán). All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by the National Agency for Research and Development (ANID) FONDECYT Iniciación 11181017, awarded to J.A.B. Jose M. Munita was supported by Comisión Nacional de Investigación Científica y Tecnológica (CONICYT) (grant number: FONDECYT 1171805) and the ANID Millennium Science Initiative, MICROB-R, NCN17\_081, Government of Chile.

**Institutional Review Board Statement:** This study was approved by the Ethical Committee of the Universidad Andrés Bello (permit number: 018/2018). Capture and sampling of rodents were also approved by Servicio Agricola Ganadero (permit number: 2118/2019).

**Informed Consent Statement:** Informed consent was obtained from all farmers for the inclusion of their dogs and/or livestock.

**Data Availability Statement:** The data presented in this study are available within this article.

**Acknowledgments:** We thank all farmers involved in this study for their cooperation and help with livestock and dog sampling. We also thank the personnel of the Municipalidad de Colina (particularly Carlos Telleria y Maximiliano Larrain) for their great help contacting farmers and helping us accessing farms. We thank all the staff members of @themonkey\_lab for their assistance in the laboratory. We thank Gabriel Carrasco for participating in the collection of rodents.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Poultry and Wild Birds as a Reservoir of CMY-2 Producing** *Escherichia coli***: The First Large-Scale Study in Greece**

**Zoi Athanasakopoulou <sup>1</sup> , Katerina Tsilipounidaki <sup>2</sup> , Marina Sofia <sup>1</sup> , Dimitris C. Chatzopoulos <sup>1</sup> , Alexios Giannakopoulos <sup>1</sup> , Ioannis Karakousis <sup>3</sup> , Vassilios Giannakis <sup>3</sup> , Vassiliki Spyrou <sup>4</sup> , Antonia Touloudi <sup>1</sup> , Maria Satra <sup>5</sup> , Dimitrios Galamatis <sup>6</sup> , Vassilis Diamantopoulos <sup>7</sup> , Spyridoula Mpellou <sup>8</sup> , Efthymia Petinaki <sup>2</sup> and Charalambos Billinis 1,5,\***


**Abstract:** Resistance mediated by β-lactamases is a globally spread menace. The aim of the present study was to determine the occurrence of *Escherichia coli* producing plasmid-encoded AmpC βlactamases (pAmpC) in animals. Fecal samples from chickens (n = 159), cattle (n = 104), pigs (n = 214), and various wild bird species (n = 168), collected from different Greek regions during 2018–2020, were screened for the presence of pAmpC-encoding genes. Thirteen *E. coli* displaying resistance to third-generation cephalosporins and a positive AmpC confirmation test were detected. *bla*CMY-2 was the sole pAmpC gene identified in 12 chickens' and 1 wild bird (Eurasian magpie) isolates and was in all cases linked to an upstream ISE*cp1*-like element. The isolates were classified into five different sequence types: ST131, ST117, ST155, ST429, and ST1415. Four chickens' stains were assigned to ST131, while five chickens' strains and the one from the Eurasian magpie belonged to ST117. Seven pAmpC isolates co-harbored genes conferring resistance to tetracyclines (*tetM, tetB, tetC, tetD*), 3 carried sulfonamide resistance genes (*sul*I and *sul*II), and 10 displayed mutations in the quinolone resistance-determining regions of *gyrA* (S83L+D87N) and *parC* (S80I+E84V). This report provides evidence of pAmpC dissemination, describing for the first time the presence of CMY-2 in chickens and wild birds from Greece.

**Keywords:** *Escherichia coli*; AmpC β-lactamases; antimicrobial resistance; CMY-2 type; ISE*cp1*; chickens; wild birds; livestock; Greece

### **1. Introduction**

Antimicrobial resistance (AMR) is a globally emergent, constantly evolving threat affecting humans, animals, and the environment, thus today constituting one of the greatest One Health challenges. Bacterial resistance to cephalosporins is mainly mediated by the production of extended-spectrum β-lactamases (ESBL) and AmpC β-lactamases. AmpC enzymes confer resistance to β-lactams, with the exception of fourth-generation cephalosporins and carbapenems, and subsequently render this essential class of antibiotics ineffective [1,2]. The presence of an AmpC combined with loss of outer membrane porins can, notably, further mediate resistance to carbapenems [2,3]. Hence, although

**Citation:** Athanasakopoulou, Z.; Tsilipounidaki, K.; Sofia, M.; Chatzopoulos, D.C.; Giannakopoulos, A.; Karakousis, I.; Giannakis, V.; Spyrou, V.; Touloudi, A.; Satra, M.; et al. Poultry and Wild Birds as a Reservoir of CMY-2 Producing *Escherichia coli*: The First Large-Scale Study in Greece. *Antibiotics* **2021**, *10*, 235. https://doi.org/10.3390/ antibiotics10030235

Academic Editor: Piera Anna Marti-no

Received: 14 February 2021 Accepted: 23 February 2021 Published: 26 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

plasmid-encoded AmpC enzymes (pAmpC) are less prevalent than ESBL in most parts of the world, they may lead to resistance of a broader spectrum, while additionally being harder to detect [2].

The most common pAmpC β-lactamase reported in *Escherichia coli* (*E. coli*) isolates of both human and animal origin globally is CMY-2 [4]. The zoonotic potential of this resistance determinant is illustrated by the detection of *bla*CMY-2 on related plasmids and *E. coli* clones in various hosts [5–7]. Insertion sequences, such as ISE*cp1*, are known to play an important role in the mobilization and thus, the spread of this gene [8,9]. Among animals, poultry have been described as the most frequent *bla*CMY-2 carrier that can also act as an important infection source for humans, especially through meat and meat products [10,11]. On the contrary, cattle and pigs are less frequently detected to harbor this gene [12]. Alarmingly, the worldwide spread of pAmpC has additionally been evidenced in wildlife and the environment [13,14]. Wild birds play an important role as vectors of AMR and have been suggested as sentinels of circulating resistance genes within a certain geographic region [15,16]. Omnivorous, synanthropic birds are more likely to carry and disseminate resistant strains due to their vicinity to human activities and their feeding habits [17]. Despite the well documented role of animals as reservoirs and spreaders of pAmpC, their ability to directly transmit resistant bacteria to humans remains debatable [10,18].

AMR constitutes a serious threat for Greek public health. According to the surveillance report of the European Centre for Disease Prevention and Control (ECDC), Greece is classified among the countries confronting AMR the most [19], while native consumption of anti-infectives for systematic use is the highest in Europe [20]. pAmpC variants of the CMY family seem to circulate among human isolates in the country [21], while there is evidence to support that this case applies for companion animal isolates as well [22,23]. In livestock and poultry, the presence of pAmpC strains has also been ascertained [12,24]. However, there is hitherto paucity of knowledge regarding the molecular characteristics of pAmpC strains isolated from farmed and wild animals, as well as their possible relationship to human hosts.

Considering the emergence of AMR and the lack of detailed data in Greece, this study aimed to evaluate the presence of pAmpC-producing *E. coli* from poultry, cattle, pigs, and wild birds, to detect the responsible pAmpC genes and to identify the *E. coli* sequence types (ST). All pAmpC-producing *E. coli* isolates that were phenotypically resistant to antimicrobials other than β-lactams, including tetracyclines, sulfonamides, and quinolones, were further tested for the respective resistance determinants.

#### **2. Results**

#### *2.1. Detection of pAmpC Genes in E. coli Isolates*

Among the 646 animal samples, 168 were derived from wild bird species, 104 from cattle, 214 from pigs, and the remaining 159 from chickens. A total of 13 *E. coli*, 12 from chickens (12/159, 7.5%) and 1 from a Eurasian magpie (1/168, 0.6%), was found to be resistant to third-generation cephalosporins (3GC) and had a positive pAmpC-confirmation test. Molecular screening for pAmpC encoding genes revealed that all isolates carried the CMY-2 type and no other pAmpC gene type was detected in any isolate.

All strains were positive in the PCR targeting ISE*cp1 – CMY,* and sequencing analysis confirmed that *bla*CMY-2 genes were linked to an upstream ISE*cp1*-like element.

#### *2.2. Molecular Typing*

Molecular typing of the 13 isolates classified them into five different STs. ST117 *E. coli* was recovered from the wild bird as well as from five chickens. Among the remaining seven chicken strains, four were assigned to ST131 and three were identified as either ST155 or ST429 or ST1415.

#### *2.3. Detection of Additional Resistance Genes*

According to susceptibility testing, 12 of the 13 CMY-2-positive *E. coli* strains, including the one from the wild bird, exhibited concurrent resistance to at least three classes of antibiotics. ESBL production, by phenotypic testing, was not observed for any strain. Six strains from chickens and the one from a wild bird exhibited resistance to tetracycline (TETR). Out of the seven tetracycline-resistant strains, six carried *tetM,* while co-occurrence of *tetB, tetC,* and *tetD* was observed in the remaining one. Resistance to sulphonamides was expressed in two strains from chickens as well as in the one from the Eurasian magpie, which all harbored both *sul*I and *sul*II genes. Ten strains showed resistance to quinolones and fluoroquinolones (QN/FQNR), although none carried *qnrA, qnrB,* or *qnrS*. Sequencing analysis of the QRDRs of *gyrA* and *parC*, performed on the resistant isolates, revealed that all strains displayed a mutation of serine-83 to leucine and a mutation of aspartic acid-87 to asparagine in *gyrA*. In addition, ST131 strains also had alterations of serine-80 to isoleucine and glutamic acid-84 to valine in the QRDR of *parC*.

The antimicrobial resistance and molecular typing results of the strains are summarized in Table 1.

**Table 1.** Characteristics of the plasmid-encoded AmpC β-lactamase (pAmpC)-producing *E. coli* isolates.


AMP—ampicillin, AMC—amoxicillin/clavulanic acid, TZP—piperacillin/tazobactam, CEX—cefalexin, CF—cefalotin, CEF ceftiofur, CFIX—cefixime, CTX—cefotaxime, CAZ—ceftazidime, CTRX—ceftriaxone, FLU—flumequine, TET—tetracycline, SXT trimethoprim/sulfamethoxazole.

#### **3. Discussion**

In this study, pAmpC-producing *E. coli* strains were detected in 7.5% of chickens and 0.6% of wild birds, while they were not identified in cattle and pig samples. The higher frequency of pAmpC isolates among poultry, compared to other species, was in accordance with previously published data [10,12]. Their absence in cattle and pigs was expected, considering the European Union Summary Report on Antimicrobial Resistance for the years 2017 and 2018 that described low detection among fattening pigs and zero occurrence in bovine meat from Greece [12].

To the best of our knowledge, this is the first time that CMY-2 type is identified from *E. coli* isolates of farmed chickens in Greece and *bla*CMY-2 was the sole pAmpC gene detected, which is in agreement with previous studies [25–27]. Carriage was relatively low (7.5%), compared to recent reports from neighboring countries such as Turkey [28], Romania [29], and Italy [25]. Our finding may be indicative of CMY-2 type low occurrence in Greek poultry but, given the lack of previous screening studies, further investigations would be helpful to verify the aforementioned low prevalence. Considering the European prohibition of cephalosporins' use in poultry, the emergence of ESBL/pAmpC-producing Enterobacteriaceae may be attributed to the treatment of eggs and/or one-day-old chickens in grandparent and parent flocks, along with the current management practices [30,31]. It has been shown that broilers can maintain pAmpC *E. coli* imported to the flock via one-dayold chicks or breeding animals even in the absence of selective antibiotic pressure [32,33]. This can be reflected in poultry meat, raising concern about the zoonotic capacity of pAmpC isolates.

We additionally detected a pAmpC-producing *E. coli* harbored by a Eurasian magpie (*Pica pica*) and, as far as we know, this is the first identification of CMY-2 type gene in a wild bird species from Greece. CMY-2 prevails among pAmpC *E. coli* isolates of corvids from The Czech Republic, Poland [34], Austria [16], Canada [17], and The USA [35,36], and of aquatic birds from The Netherlands [13], Spain [37], and Florida, USA [38]. We found a relatively low pAmpC carriage (0.6%) and our results are comparable with those of *Alcala* et al. [37] who reported 1.0% detection in Spain. Although higher pAmpC carriage has been published previously, varying from 3.4% in The Netherlands [13] to 26.9% in Florida [38], the low detection reported in our study could be attributed to the wide variety of the sampled wild bird species. Sampling and testing were performed, for screening purposes, not only in corvids and aquatic birds, but additionally in "low-risk" wild bird species, which are neither migratory nor omnivorous or aquatic-associated. Eurasian magpie is an omnivore and opportunistic scavenger, highly adapted to human environments and one of the most abundant corvids in Europe. Its diet and ecology, frequently interacting with humans and domestic animals, could explain the detection of a pAmpC-producing strain, as previously described for corvid populations [17]. Eurasian magpies are also known to form large communal roosts outside the breeding season, which could contribute to CMY-2 persistence and dissemination by bird-to-bird transmission during winter.

ISE*cp1* was found in the upstream region of *bla*CMY-2 in all our isolates. Co-existence of ISE*cp1* with ESBL/pAmpC genes in *E. coli* strains is well documented and has been associated with their efficient capture, expression, and mobilization [39,40]. Being responsible for *bla*CMY-2 transposition to different plasmids, ISE*cp1* probably has an important role in the dissemination of this beta-lactamase and subsequently the enhancement of its zoonotic potential [41].

MLST analysis demonstrated that the CMY-2-producing *E. coli* isolates of chickens were distributed in five different STs. Four chickens' strains were assigned to ST131, a clone with a worldwide distribution that has contributed to the dissemination of the ubiquitous ESBL variant CTX-M-15, as well as other resistance genes [42,43]. This finding highlights the potential of acquired AmpC enzymes to arise as an important zoonotic issue. Further supporting this claim, we also detected *bla*CMY-2 type in a chicken *E. coli* ST155, a clone commonly reported in poultry but additionally significant for public health [44,45]. On the contrary, ST429 that was detected to express CMY-2, is a predominant avian pathogenic lineage, related only to incidental human infections [46,47]. In Greece, CMY-2-producing *E. coli* ST429 has previously been isolated from a healthy household dog [23], which could imply inter-species circulation of the clone in the country. The CMY-2 type-producing *E. coli* isolated from the Eurasian magpie (*Pica pica*) belonged to ST117, previously reported in

corvids both in Europe and in Canada [17,34]. Five chickens' isolates were also assigned to this clinically important multiresistant ST, suggesting possible strain transmission among different animal hosts in the country. Detection of ST117 in poultry and a wild bird raises concern, given its frequent association to hospital-based and community-acquired human infections worldwide [48–50]. Finally, an *E. coli* of chicken origin was classified as ST1415, a rather rare ST that, to our knowledge, has not been previously related to CMY-2.

Tetracycline resistance genes were identified in 6 out of the 12 CMY-2-producing poultry isolates, as well as in the Eurasian magpie isolate. Five chickens' strains carried *tetM*, while *tetB*, *tetC,* and *tetD* were detected in the remaining one. The high frequency of tetracycline resistance among chicken pAmpC-producing isolates probably depicts the widespread use of this antibiotic in poultry husbandry all over the world [51]. Cooccurrence of *bla*CMY-2 and *tet* genes has formerly been reported in *E. coli* isolates from chicken carcasses in South Brazil [41], retail chicken meat in Canada [52], as well as in avian pathogenic *E. coli* from septicemic broilers in Egypt [53]. Additionally, the Eurasian magpie CMY-2 type-positive isolate displayed tetracycline resistance mediated by *tetM* and our finding complies with *Sen* et al. [35], who detected co-occurrence of *tetM* and *bla*CMY-2 in crow isolates.

Resistance to sulfonamides was detected in three strains, two from chickens and the one from the Eurasian magpie, which all harbored *sul*I and *sul*II sulfonamide resistance genes. In the past, sulfonamides were extensively used in traditional poultry production systems in order to achieve higher population densities and increased production. Overconsumption of this antimicrobial class resulted in the development of high resistance rates, reducing significantly its role in the poultry production nowadays [54,55]. As far as the Eurasian magpie isolate is concerned, resistance against chemically synthesized antibiotic classes such as sulphonamides has been reported in wild fauna, even though these antimicrobials are not expected to be widespread in the environment [56]. Co-occurrence of ESBL/pAmpC and sulfonamide resistance determinants on the same plasmid could probably explain the latter's detection in the wild bird isolate [57].

Quinolone resistance was also reported in CMY-2 *E. coli* strains from nine chickens and the Eurasian magpie. Mutations were responsible for the QN/FQN<sup>R</sup> phenotype and all isolates possessed the same amino acid substitution pattern in *gyrA* gene. ST131 *E. coli* possessed the S83L + D87N in *gyrA* combined with S80I + E84V in *parC*. Notably, the same mutations have been found in a collection of ST131 *E. coli* isolated from humans in Central Greece [58]. That study suggested that fluoroquinolone resistance in humans could be related to the use of these antimicrobials in the veterinary practice and the poultry production of the area. Our results verify that this specific substitutional pattern exists in *E. coli* strains of poultry origin. However, no isolate in our study co-harbored *bla*CMY-2 and plasmid mediated quinolone resistance (PMQR) genes, as has previously been described for ESBL/pAmpC-producing *E. coli* of poultry and wild bird origin [13,59,60].

#### **4. Materials and Methods**

#### *4.1. Sample Collection*

During 2018–2020, a total of 646 non duplicated fecal samples of clinically healthy animals were collected from different regions of Greece. In particular, 159 stool samples were collected from chickens, 104 from cattle, 214 from pigs, and 168 from thirty different wild bird species (Table 2). Samples were obtained by inserting a sterile cotton swab (Transwab® Amies, UK) into the rectum or the cloaca and gently rotating the tip against the mucosa.

Regarding sampling of different wild bird species, Larsen and Australian type traps as well as modified bird catching nets were used, located in a variety of habitats. The sampling site of each wild bird was recorded using handheld Global Positioning System (GPS) units. All wild birds were released immediately following sampling, according to the prerequisites of the Greek Legislation.

Swabs were transported under refrigeration and laboratory analysis was initiated 24–48 h from the samples' collection day.

**Table 2.** Number of samples per wild bird species included in the study.


#### *4.2. Isolation, Identification and Antimicrobial Susceptibility Testing of pAmpC-producing E. coli*

For the isolation of pAmpC-producing Enterobacterales, swabs were directly streaked on ESBL selective media (CHROMID® ESBL, BioMérieux, Marcy l'Etoile, France) (a medium able to detect both ESBLs and high-level expressed AmpC cephalosporinases) and then the plates were incubated aerobically at 37 ◦C for 48 h in order to increase sensitivity [61]. Each morphologically different pink colony, corresponding to *E. coli* grown on the plates, was sub-cultured on MacConkey agar. Identification of the isolated bacteria and antimicrobial susceptibility testing were carried out using the automated Vitek-2 system (BioMérieux, Marcy l'Etoile, France), according to the manufacturer's instructions. The antimicrobial agents tested, using the AST-GN96 card, were ampicillin, amoxicillin/clavulanic acid, ticar-

cillin/clavulanic acid, cefalexin, cefalotin, cefoperazone, ceftiofur, cefquinome, imipenem, gentamicin, neomycin, flumequine, enrofloxacin, marbofloxacin, tetracycline, florfenicol, polymyxin B, and trimethoprim/sulfamethoxazole. Interpretation of the antimicrobial susceptibility testing was performed automatically by the Vitek-2 software (BioMérieux, system version 8.02). Susceptibility to piperacillin/tazobactam, cefixime, cefotaxime, ceftazidime, and ceftriaxone was also tested by Etest, according to EUCAST guidelines [62].

All *E. coli* isolates that were resistant to 3GC were further tested for phenotypic AmpC production using Etest strips containing cefotetan and cefotetan plus cloxacillin (Liofilchem). Isolates that had a ratio cefotetan/cefotetan + cloxacillin ≥8 were selected for molecular detection of AmpC genes and molecular typing. Additionally, these isolates were phenotypically screened for ESBL production using Etest strips containing cefotaxime +/- clavulanic acid and Ceftazidime +/- clavulanic acid (Liofilchem). An MIC ratio ≥8 or the presence of a deformed ellipse were considered indicative of ESBL production.

#### *4.3. DNA Extraction of the AmpC-Producing E.coli*

Bacterial DNA was extracted from overnight cultures of the selected isolates using the PureLinkTM Genomic DNA Mini Kit (Invitrogen, Darmstadt, Germany), according to the manufacturer's instructions for Gram-negative bacteria.

#### *4.4. Molecular Confirmation of PAmpC Production and Screening of Insertion Sequence*

In all isolates, simplex PCRs were performed for amplification of genes for the most common types of plasmid mediated AmpC β-lactamases using the primers described by Pérez-Pérez and Hanson [63] (Table 3). Post-amplification products were visualized on 2% agarose gel electrophoresis. The PCR products were purified and were analyzed by sequencing (3730xl DNA Analyzer, Applied Biosystems).


**Table 3.** Primer sequences, amplicon sizes, and optimal annealing temperatures of each simplex PCR performed for the amplification of pAmpC and other resistance genes.


**Table 3.** *Cont.*

The presence of ISE*cp1* insertion element upstream of the *bla*CMY-2 was investigated by PCR, using a forward primer targeting the ISE*cp1* element and a reverse primer targeting the *bla*CMY, as described previously [41] (Table 3).

#### *4.5. Molecular Typing of Isolates*

Molecular typing of isolates was based on Multilocus Sequence Typing (MLST) in which amplification of seven gene loci (*adk, fumC, gyrB, icd, mdh, purA, recA*) was performed by PCR (Table 3). PCR products were purified using PureLinkTM PCR Purification Kit (Thermo Fisher Scientific), according to the manufacturer's instructions. Purified products were sequenced (3730xl DNA Analyzer, Applied Biosystems) and analysis of the alleles was conducted using an online available database (https://pubmlst.org/bigsdb?db=pubmlst\_ ecoli\_achtman\_seqdef) (accessed date: 5 February 2021).

#### *4.6. Molecular Detection of Additional Resistance Genes*

Strains in which the presence of a pAmpC gene was confirmed and were phenotypically resistant to tetracyclines, sulfonamides, and/or quinolones were additionally tested for the respective resistance genes. In detail, genes conferring resistance to tetracycline (*tetA*, *tetB, tetC, tetD, tetM*), to sulfonamides (*sul*I, *sul*II), and the PMQR determinants (*qnrA, qnrB, qnrS*) were investigated by PCR. Quinolone-resistant isolates were also screened for mutations in the quinolone resistance-determining regions (QRDRs) of *gyrA* and *parC* by PCR and sequencing of the amplicons was performed (3730xl DNA Analyzer, Applied Biosystems) (Table 3).

#### **5. Conclusions**

In this study, we investigated, for the first time, the occurrence of pAmpC-producing *E. coli* from various hosts in Greece. Chicken and wild bird strains harbored *bla*CMY-2 type in a low prevalence, while pAmpC were not detected in cattle and pigs. ST117 and ST131 were the predominant circulating CMY-2 *E. coli* clones. Tetracycline, sulfonamide, and quinolone resistance were also identified in the CMY-2 strains, revealing the presence of *tet* genes, *sul* genes, and of mutations in the QRDRs, respectively.

**Author Contributions:** Conceptualization, Z.A., K.T., V.S., E.P., and C.B.; methodology, Z.A., K.T., M.S. (Marina Sofia), D.C.C., A.G., I.K., V.G., V.S., E.P., and C.B.; validation, Z.A., E.P., and C.B.; formal analysis, Z.A. and K.T.; investigation, Z.A., K.T., M.S. (Marina Sofia), D.C.C., A.T., and M.S. (Maria Satra); resources, A.G., I.K., V.G., D.G., V.D., and S.M.; data curation, Z.A., K.T., M.S., D.C.C., A.T., and M.S. (Maria Satra); writing—original draft preparation, Z.A., K.T., and M.S.; writing—review and editing, Z.A., V.S., E.P., and C.B.; supervision, V.S., E.P., and C.B.; project administration, V.S., E.P., and C.B.; funding acquisition, M.S. (Marina Sofia), D.C.C., A.G., I.K., V.G., V.S., E.P., and C.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work has been co-funded by the European Union and the General Secretariat for Research and Innovation, Ministry of Development & Investments, under the project «Novel technologies for surveillance and characterization of Extended-spectrum β-lactamase and Carbapenemase producing Enterobacteriaceae, in humans and animals (CARBATECH)» T2DGE-0944, of the Bilateral S&T Cooperation Program Greece–Germany 2017. This support is gratefully acknowledged.

**Institutional Review Board Statement:** All samples were obtained by noninvasive rectal or cloacal swabs and no research on animals, as defined in the EU Ethics for Researchers document (European Commission, 2013, Ethics for Researchers-Facilitating Research Excellence in FP7, Luxembourg: Office for Official Publications of the European Communities, ISBN 978-92-79-28854-8), was carried out for this study. Official permissions for capturing and sampling crows, migratory and epidemic wild birds were provided by the Hellenic Ministry of Environment and Energy (159469/1920/21- 7-2017), (181997/1000/10-5-2019). Capturing, handling and sampling wild birds complied with European and national legislation.

**Data Availability Statement:** Most data for this study are presented within the manuscript. The remaining data are available on request from the corresponding author. The data are not publicly available as they are part of the PhD thesis of the first author, which has not yet been examined, approved and uploaded in the official depository of PhD theses from Greek Universities.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**

