Virulence Profiles and Antibiotic Susceptibility of Escherichia coli Strains from Pet Reptiles

Exotic reptiles are increasingly being bred as pets in many countries around the world, including Poland. However, the close contact between reptiles and their owners provides favourable conditions for the transmission of zoonotic pathogens. In this work, we examined E. coli isolates from 67 captive reptiles regarding their virulence, antibiotic susceptibility, phylogenetic affiliation, and genetic diversity. The incidence of E. coli was highest in snakes (51.6%, 16 isolates/31 samples), and slightly lower in turtles (44.4%, 8/18) and lizards (44.4%, 8/18). Genes encoding virulence factors were confirmed in 50% of isolates and the most common were the traT (37.5%, n = 12), fyuA (21.87%, n = 7), and irp-2 (15.62%, n = 5). The majority (71.87%, n = 23) of E. coli isolates were susceptible to all of the antimicrobial substances used in the study. Streptomycin resistance (21.87%, n = 7) was the most frequent, while resistance to other antimicrobial substances was sporadic. One strain (3.12%) was classified as multidrug-resistant. The presence of resistance genes (aadA, tetA, tetB, tetM, and blaTEM) was confirmed in 12.5% (n = 4) of the isolates. The majority (65.6%, n = 21) of E. coli isolates represented the B1 phylogenetic group. (GTG)5-PCR fingerprinting showed considerable genetic variation in the pool of tested isolates. The frequency of E. coli in reptiles is much lower than in mammals or birds. Due to the presence of virulence genes, characteristic of both intestinal pathogenic E. coli (IPEC) and extraintestinal pathogenic E. coli (ExPEC), reptilian strains of E. coli have pathogenic potential, and therefore people in contact with these animals should follow good hygiene practices.


Introduction
In the last few years, exotic reptiles have risen in popularity as pets, with a population of over 9 million in European households. The estimated number of captive reptiles in Poland is 215,000 [1] (pp. 44, 50). The corn snake, ball python, steppe tortoise, Greek tortoise, and lizards such as the bearded agama, chameleon and geckos are very popular [2]. Pet reptiles are sometimes kept in terrariums, but often they are free to move about in homes and treated as companion animals. In addition, reptile exhibitions organized in public spaces provide opportunities not only to observe these animals, but also to touch or hold them. However, close contact between reptiles and humans poses a public health risk, as these animals may harbour and excrete potentially pathogenic microorganisms. Reptiles are well recognized as asymptomatic carriers of Salmonella, including serotypes which can cause infections in humans, known as reptile-associated salmonellosis (RAS). At the same time, data on the characteristics of reptilian E. coli strains (RepEC), which, similarly to Salmonella, belong to the Enterobacteriaceae family, are very limited [3,4].
Within the E. coli species, apart from commensals commonly colonizing the intestines of mammals and birds, there are also intestinal pathogenic E. coli (IPEC) and extraintestinal pathogenic E. coli (ExPEC) strains. The formers are diarrhoeagenic pathogens, and the latter colonize other parts of the host's body but can also exist as commensals. Each of these two groups contains a number of pathotypes that differ in their range of virulence factors and the type of disease they cause. However, E. coli has high genome plasticity, and hybrid strains carrying a combination of both IPEC and ExPEC virulence-associated genes are known as well [5]. ExpPEC strains, i.e., avian pathogenic E. coli (APEC), uropathogenic E. coli (UPEC), and neonatal meningitis E. coli (NMEC), are responsible for many diseases of humans and animals, such as urinary tract infections (UTIs), pneumonia, meningitis, diverse intra-abdominal infections, soft tissue infections, osteomyelitis, and sepsis [6]. Due to the high genetic similarity of virulent ExpPEC strains from animals and humans, it has been suggested that livestock or pets may be a reservoir for the transmission of E. coli infections [7,8]. Among IPECs, the enterohaemorrhagic E. coli (EHEC) strains that produce Shiga toxin are considered the most dangerous for humans. The symptoms of EHEC infection are bloody diarrhoea and severe colitis, and some patients develop haemolytic uraemic syndrome (HUS), which can be fatal [9]. The main reservoirs of Shiga-toxinproducing strains of E. coli are ruminants (cattle, sheep, and goats), but these strains have also been detected in reptiles. Moreover, lizards can be carriers of other IPECs as well, i.e., enteroaggregative E. coli (EAEC), enterotoxigenic E. coli (ETEC), and enteropathogenic E. coli (EPEC) [9,10].
Bearing in mind the above, as well as the growing interest in reptile breeding and the lack of information on the occurrence and characteristics of RepEC, in this study we aimed to determine the prevalence of E. coli strains in captive reptiles and their virulence and antibiotic sensitivity profiles, as well as their phylogenetic affiliation and genetic variation. Monitoring the presence of microorganisms that are dangerous to humans, such as pathogenic and drug-resistant E. coli, and identifying reservoirs of such strains is important for assessing the risk associated with their spread and possible infection.

Detection of Resistance Genes
The occurrence of the resistance genes was recorded for only 4 (12.5%) E. coli strains. Moreover, 1 strain (3.12%) was detected with the aadA gene encoding streptomycin/spectinomycin adenylyltransferase, 2 strains (6.25%) were confirmed with the tetA, tetB and tetM genes conferring resistance to tetracyclines, and 1 isolate (3.12%) carried the TEM-type beta-lactamase (blaTEM). For each of these strains, a correlation was found between the phenotype and the presence of resistance genes ( Table 2).

Determination of E. coli Phylogenetic Groups
Phylogenetic groups of E. coli were determined based on the electrophoretic profiles of multiplex PCR amplicons (yja, TspE4.C2, chuA, svg and uidA) ( Figure 1). Over 65% of the isolates (65.6%, n = 21) belonged to group B1, 18.7% (n = 6) were assigned to group A, 9.4% (n = 3) to B2, and 6.2% (n = 2) to group D. None of the isolates belonged to group B2 1 (no isolate contained the svg gene) ( Figure 1). It is worth noting that all the strains representing group B2 were collected from snakes ( Table 2).

Determination of E. coli Phylogenetic Groups
Phylogenetic groups of E. coli were determined based on the electrophoretic profiles of multiplex PCR amplicons (yja, TspE4.C2, chuA, svg and uidA) ( Figure 1). Over 65% of the isolates (65.6%, n = 21) belonged to group B1, 18.7% (n = 6) were assigned to group A, 9.4% (n = 3) to B2, and 6.2% (n = 2) to group D. None of the isolates belonged to group B21 (no isolate contained the svg gene) (Figure 1). It is worth noting that all the strains representing group B2 were collected from snakes ( Table 2).

Rep-PCR Fingerprinting
DNA amplification with the (GTG)5 primer revealed significant genotypic diversity among the E. coli isolates. The electrophoretic profiles were reproducible in the 2 separate PCR reactions, and the fingerprinting generated distinct amplification bands ranging in size from 350 bp to 2700 bp. The profiles obtained contained between 2 and 12 PCR prod-

Rep-PCR Fingerprinting
DNA amplification with the (GTG) 5 primer revealed significant genotypic diversity among the E. coli isolates. The electrophoretic profiles were reproducible in the 2 separate PCR reactions, and the fingerprinting generated distinct amplification bands ranging in size from 350 bp to 2700 bp. The profiles obtained contained between 2 and 12 PCR products and allowed the 32 strains to be grouped into 27 rep-types ( Figure 2). This indicates considerable genetic diversity of the E. coli isolates, with a maximum percentage of dissimilarity of 41%. There was no correlation between the electrophoretic profile and the phylogenetic group or the host species, although several strains representing the same phylogenetic group had identical fingerprints (Figure 3).

Rep-PCR Fingerprinting
DNA amplification with the (GTG)5 primer revealed significant genotypic diversity among the E. coli isolates. The electrophoretic profiles were reproducible in the 2 separate PCR reactions, and the fingerprinting generated distinct amplification bands ranging in size from 350 bp to 2700 bp. The profiles obtained contained between 2 and 12 PCR products and allowed the 32 strains to be grouped into 27 rep-types ( Figure 2). This indicates considerable genetic diversity of the E. coli isolates, with a maximum percentage of dissimilarity of 41%. There was no correlation between the electrophoretic profile and the phylogenetic group or the host species, although several strains representing the same phylogenetic group had identical fingerprints ( Figure 3).

Discussion
The scientific literature abounds with information on the drug susceptibility and virulence of E. coli isolated from humans, mammals, and birds, but reports on the occurrence and characteristics of E. coli in reptiles are scarce. The prevalence of the E. coli strains re-

Discussion
The scientific literature abounds with information on the drug susceptibility and virulence of E. coli isolated from humans, mammals, and birds, but reports on the occurrence and characteristics of E. coli in reptiles are scarce. The prevalence of the E. coli strains reported in these studies (47.8%) is in line with that reported by Gopee et al. [11] (52%) for reptiles from the Emperor Valley Zoo, Trinidad, Spain. The authors recorded the highest incidence of E. coli in Crocodylidae (83.3%), lower incidences in turtles (68.08%) and snakes (37.5-51.1%, depending on the family), and the lowest for lizards (33.3%). In contrast with our observations, a higher frequency of E. coli was recorded in herbivorous reptiles (66.7%) than in carnivorous reptiles (46.4%).
A significantly higher frequency of E. coli in captive and free-living reptiles was recorded by Ramos et al. [3] in Brazil (68%) and, as in our study, they reported a higher incidence of E. coli in snakes (84.4%) than in lizards (46.6%) and turtles (40-50%). Similar results were obtained by Sylvester et al. [12] and Bautista-Trujillo et al. [10], who showed the presence of E. coli in approximately 40% of green iguanas. Several other authors have also described the presence of E. coli in reptiles; however, they did not take into account the frequency of this bacterium in individual groups of reptiles. Książczyk et al. [4] isolated a total of 35 E. coli strains from 103 stool samples from reptiles at the zoo in Wrocław, Poland. These data indicate 34% prevalence, assuming each strain was isolated from a different sample. Much higher prevalence of E. coli in reptiles is suggested by the findings of Unger et al. [13], who isolated a total of 142 E. coli isolates from 150 faecal samples derived from >60 exotic reptile species crossing the veterinary border at Frankfurt airport.
Warm-blooded animals (mammals and birds) are already known to be reservoirs of both IPEC and ExPEC E. coli strains [14][15][16], but knowledge of the occurrence of pathogenic E. coli in cold-blooded animals is scant. In terms of the presence of virulence genes, our research is largely in line with the findings of Ramos et al. [3], who detected only 2 of the many virulence genes characteristic for ETEC, EHEC and NTEC, i.e., the cnf1 gene encoding cytotoxic necrotizing factor 1 (found in 9.2% of RepEC isolates) and the astA gene encoding EAST1 toxin (2.6%). The frequency described in this study (3.12%) of the astA gene, coding for heat-stable enterotoxin EAST1 characteristic of EAEC, is also consistent with results reported by Książczyk et al. [4]. In the case of virulence genes characteristic for IPEC, it should be emphasized that our research is the first report on the presence of the pic gene in RepEC, characteristic for the EHEC pathotype. Completely different results were presented by Bautista-Trujillo et al. [10], who classified 62 of 100 E. coli isolates from captive green iguanas in Mexico as diarrhoeagenic E. coli (DEC) strains. Among them, STEC strains (40.3%) carrying the stx1 (38.7%) or stx2 (1.6%) gene were the most prevalent pathotype, followed by EAEC (27.4%, aap-positive strains), ETEC (27.4% each, ltor st-positive) and EPEC (4.9%, bfpAor eaeA-positive). The frequency of tratT (37.5%, coding for complement resistance protein) and fyuA (21.87%, coding for the ferric yersiniabactin uptake receptor) genes reported in these studies was about twice as high as in the case of RepEC strains isolated in Poland (traT-16.7%, fyuA-8.3%) by another research team [4]. The fyuA gene is common in both UPEC and APEC strains, while the traT gene is characteristic of the UPEC pathotype [4]. The occasional presence of the irp-2 gene in RepEC (8.3%) was also demonstrated by Książczyk et al. [4]. However, in contrast to our results, they did not detect the iucD or vat gene in any strain, and in several strains, they confirmed the presence of other virulence genes characteristic of ExPEC, i.e., papC, iss and tsh.
The results of this study on drug resistance in E. coli most closely match the findings of Sylvester et al. [12], who showed that among E. coli strains from wild and pet green iguanas (from Grenada), resistance to streptomycin (12%) and amoxicillin-clavulanic acid (12%) was the most common, and only 7% of strains were multi-drug resistant. A much higher prevalence of resistant E. coli strains was recorded in reptiles housed at zoo in Trinidad [11], and resistance to ampicillin (66.7%), tetracycline (57.6%) and gentamicin (12.1%) was the most frequent. Interestingly, the prevalence of streptomycin resistance (36.4%) demonstrated by these studies was somewhat similar to our data (21.9%). An agreement was also noted in the lack of resistance to fluoroquinolones. A high percentage of resistant E. coli strains was also demonstrated by Bautista-Trujillo et al. [10], who classified as many as 82.3% of DEC isolates (from captive green iguanas in Mexico) as MDR, and most frequently noted resistance to amikacin (~40%) and ampicillin (~30%).
The higher incidence of drug-resistant strains in captive reptiles from Trinidad and Mexico may be due to different veterinary practices and the different prevalence of drugresistant E. coli in the reptile diet and environment. Details on the type and frequency of antibiotic use in reptiles in these countries are not known, but Mexico was among the top ten countries in the world with the highest use of antibiotics in veterinary medicine in 2017 [16].
The higher frequency of resistance to streptomycin (~22%) compared to resistance to other antimicrobial substances (0-6.25%) reported in these studies may be due to the widespread use of this antibiotic in veterinary medicine as well as in horticulture as a plant protection product. The resulting streptomycin-resistant strains may spread in the environment and eventually be acquired by the reptiles via their feed (plants, rodents, meat). Moreover, the results obtained in this study and some literature data suggest that the reptilian Enterobacteriaceae may have a reduced susceptibility to streptomycin. Streptomycin resistance was the most frequently reported antibiotic resistance phenotype in reptile-derived Salmonella strains in Poland (25%) [17] and Indonesia (75%) [18]. Similarly, Bertolini et al. [19] showed a high percentage (79%) of Salmonella strains intermediately susceptible to streptomycin in reptiles.
According to our best knowledge, this work is the first full report of the occurrence of resistance genes in E. coli from reptiles. The aadA (aminoglycoside (3") (9) adenylyltransferase), tetA (tetracycline efflux protein TetA), tetB (tetracycline efflux protein TetB) and blaTEM genes detected in several RepEC strains are commonly found in drug-resistant Enterobacteriaceae isolated from humans and farm animals [20,21]. The tetM gene coding for tetracycline ribosomal protection protein has thus far been found sporadically in E. coli, mainly in pig and chicken strains [21]. Its presence confirmed in this work in one RepEC isolate may be the result of genetic transfer from Enterococcus or Lactobacillaceae, common carriers of tetM [22,23]. The TEM family of β-lactamases covers a large group of enzymes and determining which TEM lactamase has been detected in these studies requires more detailed sequencing-based analyzes. In E. coli strains, the most common is the prototype narrow-spectrum blaTEM-1 lactamase, from which the broad-spectrum lactamases (ESBL) have evolved by mutation [24,25]. The blaTEM-1 gene was also found in two strains of E. coli from reptiles transported across the veterinary border in Frankfurt, where it coexisted with many other resistance genes [13].
Phylogenetic group B1, which dominated (65.6%) among the RepEC isolates, is characteristic of non-pathogenic E. coli strains, while groups A and B2, recorded in this study for a total of 9 isolates, is characteristic of ExPEC, i.e., UPEC and APEC [4,26]. It is therefore worth noting that 3 of 4 RepEC strains, containing 4 virulence genes simultaneously, belonged to the B2 phylogenetic group. The results presented in this study are consistent with the reports of Gordon and Cowling [27], who assigned nearly 70% of E. coli isolates from snakes and lizards to phylogenetic group B1, while strains of groups A, B2 and D accounted for 15.1%, 6.1% and 9.1%, respectively. A higher percentage (25%) of strains representing the B2 phylogenetic group was recorded in turtles [28]. A higher frequency of strains of the B1 group in RepEC was reported by Ramos et al. [3] (88.4%) and Książczyk et al. [4] (95.8%).

Collection of Faecal Samples
The research material consisted of faecal samples collected between 2017 and 2020 from 67 reptiles, i.e., snakes (n = 31), lizards (n = 18) and turtles (n = 18) living in captivity or as pets. The animals did not show any symptoms of disease. They were obtained from private owners (n = 53) and from pet shops (n = 14) in the Lubelskie Province, Poland.

E. coli Isolation
Stool specimens collected with a swab were suspended in peptone water and incubated for approximately 18 hours, inoculated onto MacConkey agar (Oxoid Ltd., Altrincham, UK), and incubated at 37 • C for 24 h under aerobic conditions. Single pink colonies were inoculated on TSB (trypticase soy broth) (Oxoid Ltd., Altrincham, UK), and pure cultures supplemented with 20% glycerol were stored at −80 • C for further analysis. To determine the ability of the strains to produce glucuronidase, they were additionally plated on TXB agar and incubated~24 h at 44 • C.

Identification of E. coli
Presumptive E. coli isolates were identified based on detection of the uidA glucuronidase gene and the E. coli-specific flanking region of the uspA gene (coding for universal stress protein). PCR was performed using DreamTaq polymerase (Thermo Fisher Scientific, Vilnius, Lithuania) and the following two pairs of primers: uidAF-TATGGAATTTCGCCGATTTT and uidAR-TGTTTGCCTCCCTGCTGCGG (uidA amplicon size 166 bp) and uspAF-CCGATACGC TGCCAATCAGT and uspAR-ACGCAGACCGTAGGCCAGAT (uspA amlicon 884 bp) [25]. The PCR reaction were performed using Dream Taq polymerase and the following thermalcycling program: initial denaturation at 94 • C for 5 min; 30 cycles of 94 • C for 45 s, 57 • C for 45 s and 72 • C for 1 min; final extension step at 72 • C for 8 min. PCR products (8 µL, 600 ng) were separated by electrophoresis (100 V, 1 h) on 1.8% agarose gels and visualized by SimplySafe (Eurx, Gdańsk, PL) staining. PCR product sizes were determined by comparison with a Nova 100 bp DNA ladder (Novazym, Poznań, Poland) using Image Lab 6.1. Software (Bio-Rad, Hercules, CA, USA).

DNA Extraction
Whole-genome DNA was extracted from all of the tested E. coli isolates (n = 32) using the Gene MATRIX Bacterial & Yeast Genomic DNA Purification Kit (Eurx, Gdańsk, Popland) following the manufacturer's instructions and stored at −20 • C. DNA concentration was determined by measuring the absorbance of 2 µL of the sample at 260/280 nm using the NanoDrop Lite spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), and the quality of DNA was analyzed by agarose (1.5% w/v) gel electrophoresis. No signs of DNA degradation (smearing) were observed in any of the samples, and its concentration was 20 ng/µL.

Antimicrobial Susceptibility Testing
Antibiotic susceptibility testing was performed by the agar disk diffusion method according to Clinical and Laboratory Standards Institute (CLSI) recommendations [31].  [31]; only in the case of nitrofurantoin was the interpretation of the results based on EUCAST guidelines [32]. E. coli ATCC 25922 reference strain was used as a quality control.

Determination of E. coli Phylogenetic Groups
To determine the phylogenetic groups of the E. coli isolates, five sets of primers for the five sets of primers for the genes yja, TspE4.C2, chuA, svg and uidA (Table S3) were used in a multiplex PCR, as previously described [33]. PCR products were separated by electrophoresis in 2.5% (wt/vol) high resolution agarose (Blirt, Gdańsk, Poland). The amplicons sizes were determined by comparison with a Nova 100 bp DNA ladder (Novazym, Poznań, Poland) using Image Lab 6.1. Software (Bio-Rad, Herkules, CA, USA). The phylogenetic groups were determined based on the PCR gel pattern [33].
The presence of a given band was coded as 1 and the absence of a given band was coded as 0 in a data matrix (Excel, Microsoft Office), and the unweighted pair group method with arithmetic averages (UPGMA) was used to generate a dendrogram of dissimilarity in Statistica software (Ver. 13.1, Tulsa, OK, USA).

Statistical Analysis
Pearson's chi-squared test was used to determine the relationship between diet (carnivores, omnivores, herbivores) and the frequency of E. coli, as well as between the type of reptile (snakes, lizards, turtles) and the frequency of E. coli isolation. The level of significance was set as p < 0.05. The statistical analysis was performed using Microsoft Excel 2019.

Conclusions
The results presented in this study contribute to knowledge of the occurrence and phenotypic and genotypic features of RepEC strains. Although the incidence of drug resistance in E. coli strains from reptiles in Poland is low, the possibility of transmission of resistance genes to other members of Enterobacteriaceae, including Salmonella, which commonly inhabits reptile intestines, should be taken into account. In addition, RepEC strains contain virulence genes characteristic of ExPEC, and less often of IPEC. IPEC strains are diarrhoeagenic pathogens, while ExPEC strains may exist as commensals, but under favourable conditions can cause infections. Due to the pathogenic potential of RepEC strains, they may pose a health risk to humans, and reptile keepers should use good hygiene practices when handling these animals.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/ 10.3390/pathogens11020127/s1, Table S1: Sequences of primers used to detect virulence genes, amplicon sizes and references [14,29,30]; Table S2: Sequences of primers used to detect resistance genes [14,23,[35][36][37]; Table S3: Primer sequences used to determine phylogenetic groups in E. coli strains [35]. Data Availability Statement: All the data presented in the study are included in the article. Further enquires can be directed to the corresponding author.