Antibiotic-Resistant Escherichia coli Strains Isolated from Captive Giant Pandas: A Reservoir of Antibiotic Resistance Genes and Virulence-Associated Genes

Simple Summary Escherichia coli (E. coli) carrying antibiotic resistance genes (ARGs) and virulence-associated genes (VAGs), which are considered as emerging environmental pollutants, undoubtedly causes adverse health effects on humans and animals. With the implementation of the Wild Release Plan of captive giant pandas, antibiotic-resistant E. coli carrying ARGs/VAGs in giant pandas’ feces that are discharged into the environment may become pollutants for nature. To evaluate the potential threat of E. coli strains from captive giant pandas, we analyzed the distribution and the association of ARGs and VAGs in antibiotic-resistant E. coli. Our results revealed a high prevalence of ARGs and a significant positive association among ARGs and VAGs, suggesting that the continuous monitoring of the impact of the spread of antibiotic-resistant E. coli (carried various ARGs and VAGs) is needed in future studies. Abstract Recent studies showed that Escherichia coli (E. coli) strains isolated from captive giant pandas have serious resistance to antibiotics and carry various antibiotic resistance genes (ARGs). ARGs or virulence-associated genes (VAGs) carried by antibiotic-resistant E. coli are considered as a potential health threat to giant pandas, humans, other animals and the environment. In this study, we screened ARGs and VAGs in 84 antibiotic-resistant E. coli strains isolated from clinically healthy captive giant pandas, identified the association between ARGs and VAGs and analyzed the phylogenetic clustering of E. coli isolates. Our results showed that the most prevalent ARG in E. coli strains isolated from giant pandas is blaTEM (100.00%, 84/84), while the most prevalent VAG is fimC (91.67%, 77/84). There was a significant positive association among 30 pairs of ARGs, of which the strongest was observed for sul1/tetC (OR, 133.33). A significant positive association was demonstrated among 14 pairs of VAGs, and the strongest was observed for fyuA/iroN (OR, 294.40). A positive association was also observed among 45 pairs of ARGs and VAGs, of which the strongest was sul1/eaeA (OR, 23.06). The association of ARGs and mobile gene elements (MGEs) was further analyzed, and the strongest was found for flor and intI1 (OR, 79.86). The result of phylogenetic clustering showed that the most prevalent group was group B2 (67.86%, 57/84), followed by group A (16.67%, 14/84), group D (9.52%, 8/84) and group B1 (5.95%, 5/84). This study implied that antibiotic-resistant E. coli isolated from captive giant pandas is a reservoir of ARGs and VAGs, and significant associations exist among ARGs, VAGs and MGEs. Monitoring ARGs, VAGs and MGEs carried by E. coli from giant pandas is beneficial for controlling the development of antimicrobial resistance.


Introduction
Escherichia coli (E. coli) is one of the most important intestinal floras of humans and animals [1,2]. The carrying of antibiotic resistance genes (ARGs) and virulence-associated genes (VAGs) among E. coli undoubtedly causes adverse health effects on humans and animals [3,4]. Antibiotic-resistant E. coli is also considered as an emerging potential environmental pollutant [5][6][7]. ARGs and VAGs carried by E. coli are widely studied due to the risk of transmission to the environment [8][9][10][11]. The giant panda (Ailuropoda melanoleuca) is one of the endangered animals in the world; the number of giant pandas (especially captive giant pandas) has increased with the implementation of protective projects [12]. However, giant pandas are still threatened by bacterial diseases, especially the diseases caused by intestinal pathogens (such as Escherichia coli, Klebsiella spp., Campylobacter jejuni, Pseudomonas aeruginosa, Yersinia enterocolitica and Clostridium welchii) [13]. Overusing antibiotics for captive animals promotes the resistance of microorganisms [14][15][16]. E. coli is one of the most common microorganisms and frequently occurs in nature and the gastrointestinal tract of animals. E. coli is usually considered as an indicator organism for antimicrobial resistance (AMR) [7,17]. Research indicated that E. coli strains from giant pandas are highly resistant to antibiotics such as tetracyclines, sulfonamides and β-lactams [12,18,19]. Wang et al. showed that streptomycin was the most prevalent resistance phenotype and strB was the most prevalent ARG in E. coli strains from wild giant pandas, depicting an association between phenotype resistance and ARGs [19]. Our previous study proved a significant association between AMR and mobile genetic elements (MGEs) in the E. coli of captive giant pandas, while the associations of ARGs with AMR and MGEs are still unknown [12].
Previous researchers showed that an association existed between ARGs and VAGs in E. coli isolated from animals (yaks, weaned piglets and waterfowls) [20][21][22]. The uropathogenic E. coli (UPEC) isolated from patients with urinary tract infections (UTIs) showed antibiotic resistance related to particular VAGs [23]. However, the association between ARGs and VAGs in E. coli isolated from giant pandas is unknown.
With the implementation of the Wild Release Plan of captive giant pandas [24,25], AMR E. coli carrying ARGs/VAGs in giant pandas' feces that are discharged into the environment may increase the risk of infections in animals and humans. To better understanding the characteristics of antibiotic-resistant E. coli, we detected ARGs and VAGs in 84 strains isolated from clinically heathy captive giant pandas and analyzed the associations of ARGs with AMR, VAGs and MGEs in this study.

Bacterial Strains
Eighty-four E. coli strains were isolated and identified from the fecal samples of 84 clinically healthy captive giant pandas and preserved in the Key Laboratory of Animal Disease and Human Health of Sichuan (Chengdu, China) [12].

Antibiotic Sensitivity and Resistance
Antibiotics for the antimicrobial susceptibility test were selected based on the information provided by breeders at Chengdu Giant Panda Breeding & Research Base (CGPBRB) and other reports about antimicrobial susceptibility in wildlife. Antibiotic susceptibility results (22 antibiotics in six categories) of all the E. coli strains have been reported in our previous publication [12]. The antibiotic susceptibility results were interpreted according to the CLSI criteria (Clinical and Laboratory Standards Institute, 2022).

DNA Extraction and Screening for ARGs and VAGs
Total genomic DNA was extracted from isolates using the TIANamp Bacteria DNA kit (Tiangen Biotech, Beijing, China), according to the manufacturer's instructions. DNA samples were stored at −20 • C for subsequent PCR detection. We further amplified 16sRNA (16 S rRNA Primer: 5 -GAGTTTGATCCTGGCTCAG-3 ; 5 -AGAAAGGAGGTGATCCAGCC-3 ) [26] to confirm the quality of the extracted DNA before screening for related genes.

Association Analysis between ARGs and AMR/VAGs/MGEs
A χ2 test was used to compare the association between ARGs and AMR/VAGs/MGEs. Eleven MGEs have been identified, as reported in our previous study [12]. p-values < 0.05 were considered to be statistically significant. Odds ratios (OR) and 95% confidence intervals were also determined for statistically significant associations. An OR < 1 and an OR > 1 represent negative and positive associations, respectively. Data analysis was performed with SPSS Statistics (version 26.0).
Note: Only antibiotic resistance genes with a significant association (p < 0.05) are shown. OR, odds ratio for significant associations between antibiotic resistance genes (95% confidence interval in parentheses); NA, no results available or they could not be calculated; -indicates no significant associations (p ≥ 0.05); NS, no statistics were determined for the same gene.
Note: Only virulence-associated genes with a significant association (p < 0.05) are shown. OR, odds ratio for significant associations between virulence-associated genes (95% confidence interval in parentheses); the meaning of NA/NS/-is the same as that in Table 3. Note: Only antibiotic resistance genes vs. virulence-associated genes with a significant association (p < 0.05) are shown. OR, odds ratio for significant associations between antibiotic resistance genes and virulence-associated genes (95% confidence interval in parentheses); -indicates no significant associations (p ≥ 0.05).

Discussion
Recently, antibiotic resistance E. coli as well as antibiotic resistance genes (ARGs) and virulence-associated genes (VAGs) carried by E. coli strains have been extensively studied as environmental pollutants, which may cause adverse effects to ecology, human and animal health [21,[32][33][34]. To evaluate the potential threat of E. coli strains from captive giant pandas, the distribution and association of ARGs and VAGs in antibiotic-resistant E. coli were analyzed. Our results showed that many ARGs (especially β-lactam resistance genes, 100.00% for bla TEM ) and VAGs (especially adhesion-related virulence genes, 91.67% for fimC) were detected from the 84 E. coli strains. Fourteen ARGs were detected, and the top five ARGs ranked by detection rates were bla TEM (100.00%), tetA (86.90%), sul3 (61.90%), cmlA (48.81%) and flor (47.62%). Another study from the Dujiangyan Giant Panda Conservation and Research Center and the Wolong Nature Reserve showed that the top five ARGs detected were sul1 (45%), bla CTX-M (44%), ant(3 )-Ia (38%), tetA (37%) and qnrB (35%) [35]. Those data showed a high prevalence of ARGs (especially for β-lactam, tetracyclines and sulfonamides resistance genes) in E. coli strains from giant pandas, and the carrying rate of the ARGs in the E. coli isolated in our study is higher than that in strains from Dujiangyan and Wolong. Notably, all of the 84 strains were positive for bla TEM in our study. E. coli strains containing bla TEM were reported in other wild animals (such as owls, foxes, wild rabbits, genets, etc.) [36]. Another study showed that 5 out of 13 ARGs were detected in E. coli isolated from wild giant pandas, but bla TEM was not detected [13]. The different detection rates of ARGs (especially for bla TEM ) in captive and wild giant pandas may result from their different living environments [37]. The results from non-human primates have also shown that the carrying rates of ARGs in E. coli are affected by living environments [10,38]. As far as we know, this is the first report of a high prevalence of bla TEM detected in E. coli strains from captive giant pandas. Some studies showed that bla CTX-M and sul1 are prevalent in E. coli from captive giant pandas and wild birds [35,39]. The most prevalent ARGs in E. coli isolated from captive giant pandas in Dujiangyan and Wolong were bla CTX-M (44%) and sul1 (45%), which is consistent with our results (bla CTX-M was 42.86% and sul1 was 39.29%) [35]. E. coli carrying various ARGs was considered as an environmental pollutant, since the spread of antibiotic-resistant bacteria may pose a potential threat to human health [40][41][42]. Referring to the previous report on wild gulls, the antibiotic-resistant E. coli strains from giant pandas may serve as a pool of ARGs [43]. Moreover, VAGs related to diarrheagenic Escherichia coli (DEC) or extraintestinal pathogenic Escherichia coli (ExPEC) were screened [44][45][46]. In the 84 E. coli strains, 9 out of 17 VAGs were detected, of which the top 5 resistance genes were fimC (91.67%), ompT (63.10%), sitA (55.95%), iroN (44.05%) and fyuA (39.29%). None of the isolates were positive regarding tsh, hlyA, hlyF, iss, astA, estA, estB and elt. Only eaeA (which is a marker of enteropathogenic E. coli, EPEC) was detected in 18 strains. Most strains (91.67%, 77/84) carried at least one VAG related to ExPEC, which raises a health concern. Similarly, ExPEC-associated VAGs were observed in a higher proportion than those associated with DEC in E. coli from waterfowl and yaks [20,21]. VAGs related to a high-pathogenicity island (fyuA and irp2) were observed in 33 strains in our study. The detection rate of the VAGs in our study (ranging from 2.38% to 91.67%) is higher than the result of Wang et.al (all detection rates below 25%) [19]. Notably, the most prevalent VAG in our study is fimC, which belongs to adhesion-related virulence genes. Deng et al. showed that an E. coli strain from a urine sample of a giant panda carried the largest number (42.45%, 239/563) of adhesion-and invasion-related virulence genes [47]. Our present study showed E. coli strains from clinically healthy captive giant pandas carrying a large number of VAGs, which was also observed in healthy chickens and pigs [48,49]. As the phylogenetic grouping is closely related to virulence genes (group A and group B1 carry fewer virulence genes, while group B2 and group D carry more virulence genes [31]), we further analyzed the phylogenetic grouping and the number of virulence genes carried by each E. coli strain from giant pandas. Our results showed that the B2 group is the most prevalent group (67.86%, 57/84) which carries the highest number of virulence genes. Another study on E. coli strains of wild giant pandas showed that the B2 group was the most prevalent (38%, 31/82) [13], which is consistent with our study. In summary, our results showed that the antibiotic-resistant E. coli derived from giant pandas is a natural reservoir of ARGs and VAGs. In our previous study, we detected a high prevalence of MGEs in E. coli strains isolated from giant pandas [12]. Another study also proved that the release of wild animals may spread ARGs through MGEs [50]. With the release program of captive giant pandas being implemented in China [24,25], E. coli strains carrying various ARGs and VAGs may be harmful when spreading to the surrounding environment, breeders and tourists, posing a higher risk to public health.
In general, the antibiotic resistance phenotype is determined by the carrying and expression of ARGs [51][52][53]. ARGs carried by E. coli strains are closely related to the antibiotic resistance phenotype [54]. Thus, we further analyzed the relations between antibiotic resistance phenotypes and ARGs carried by E. coli from captive giant pandas. As shown in Table 2, for amide alcohols, aminoglycosides and quinolones, the antibiotic resistance phenotypes were all matched with the detection rates of ARGs. However, for tetracyclines, β-lactams and sulfonamides, the detection rates of related ARGs fluctuated (such as the resistance rate of sulfonamides being 17.86%, while the detection rate of sul ranged from 9.52% to 61.90%). The detection rates of ARGs are not strictly correlated with the corresponding antibiotic resistance phenotypes [10]. This might be due to the fact that the expression of ARGs is just one of the mechanisms that mediate bacterial resistance or due to the abnormal expression or low expression of ARGs that cannot cause the drug resistance [51,55,56]. Previous studies have shown an association between ARGs and VAGs carried by E. coli in wild waterfowls and weaned piglets [21,22]. We further analyzed the association among 13 ARGs and 9 VAGs, respectively. We observed 30 pairs of positive associations and 1 negative association among 13 ARGs, of which the strongest positive association (p < 0.05, OR > 1) was observed for sul1 and tetC. Similarly, positive associations among different ARGs were observed in another study on wild birds [21]. For the VAGs, 14 pairs of VAGs showed positive associations, and none of the VAGs pairs were negatively associated, which is consistent with the results of E. coli strains from Qinghai Plateau yaks [20]. The strongest positive association (p < 0.05, OR > 1) was observed for fyuA and iroN within VAGs. Associations among ARGs and VAGs in E. coli from captive giant pandas were further analyzed. Our results showed that positive associations were more common than negative ones (45 pairs were positive and only 1 was negative). The strongest positive association was observed for sul1 and eaeA between ARGs and VAGs. Interestingly, a study on waterfowls showed that negative associations were more common [21]. According to the available information, this is the first report about the association among ARGs and VAGs carried by E. coli from captive giant pandas. Previous studies have shown that specific microorganisms (such as endophytes or pathogenic E. coli) are enriched for ARGs [57][58][59], and significant associations exist between VAGs and AMR [60]. However, the direction of associations (positive or negative) among ARGs and VAGs is unknown. We speculate that carrying VAGs increases the probability of strains carrying ARGs, and vice versa. Considering that MGEs mediate the horizontal transmission of ARGs and that associations exist among MGEs and ARGs [61][62][63], we further analyzed the association between ARGs and MGEs. Only significant positive associations (p < 0.05, OR > 1) were observed between ARGs and MGEs. Positive associations for ARGs and MGEs were also reported in E. coli strains from wastewaters and activated sludge reactors [64,65]. Our present results showed that E. coli strains from captive giant pandas are a reservoir of ARGs, VAGs and MGEs; ARGs are significantly associated (p < 0.05) with VAGs and MGEs. As captive giant pandas are released into the wild, various ARGs, VAGs and MGEs carried by E. coli strains from captive giant pandas may have an impact on the environment.

Conclusions
The present study showed that antibiotic-resistant E. coli from captive giant pandas is a pool of ARGs, VAGs and MGEs, and significant associations exist among them. Positive associations are more prevalent than negative ones among ARGs and VAGs, and only positive associations were observed between ARGs and MGEs in E. coli from captive giant pandas. Antibiotic-resistant E. coli strains that carry various ARGs, VAGs and MGEs are a potential threat not only to the health of giant pandas but also to the environment. We strongly recommend implementing surveillance programs for the use of antibiotics in giant pandas and monitoring the impact of antibiotic-resistant E. coli (carrying various ARGs, VAGs and MGEs) on humans, other animals and the environment in future studies.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/vetsci9120705/s1, Table S1: PCR primer and conditions for the antibiotic resistance genes used in this study; Table S2: PCR primer and conditions for the virulenceassociated genes used in this study; Table S3: Antibiotic resistance genes and virulence-associated genes pattern for Escherichia coli isolates from captive giant pandas; Table S4: Odds ratios (OR) of antibiotic resistance genes (ARGs) associated with antimicrobial resistance (AMR) in Escherichia coli isolates from captive giant pandas (n = 84); Table S5: Odds ratios (OR) of antibiotic resistance genes (ARGs) associated with mobile genetic elements (MGEs) in Escherichia coli isolates from captive giant pandas (n = 84).

Data Availability Statement:
The data presented in this study are available in the article/supplementary material; further inquiries can be directed to the corresponding authors.