Abundance of colistin-resistant Escherichia coli harbouring mcr-1 and extended-spectrum β-lactamase-producing E. coli co-harbouring blaCTX-M-55 or -65 with blaTEM isolates from chicken meat in Vietnam

Although the spread of plasmid-mediated antibiotic-resistant bacteria is a public health concern, food contamination with plasmid-mediated antibiotic-resistant Escherichia coli in Vietnam has not been well investigated. This study aimed to describe the prevalence of colistin-resistant, carbapenem-resistant, and endemic blaCTX-M in extended-spectrum β-lactamase (ESBL) producing E. coli isolates. Colistin and carbapenem-resistant ESBL-producing E. coli were isolated from chickens in Vietnam and Japan. Colistin-resistant and AmpC/ESBL-producing E. coli (52% and 93%, respectively) were detected in chickens from Vietnam, in comparison to 52.7%, AmpC/ESBL-producing E. coli found in chicken from Japan. Carbapenem-resistant E. coli has not been isolated in Vietnam and Japan. Genotyping revealed that colistin-resistant E. coli harboured mcr-1, and most of the AmpC/ESBL-related genes were blaCTX-M-55 and blaCTX-M-65 together with blaTEM in Vietnamese chickens and blaCMY-2 in Japanese chickens. Multi-drug resistance analysis showed that ESBL-producing E. coli isolates had greater resistance to quinolones, streptomycin, and chloramphenicol than colistin-resistant E. coli isolates from Vietnam, suggesting the selection of multiple antibiotic resistance genes in ESBL-producing E. coli. In conclusion, colistin-resistant E. coli was detected in approximately half of the chicken samples, the majority of which harboured mcr-1. The high prevalence of ESBL-producing E. coli has remained constant in the last 5 years. The predominant blaCTX-M in ESBL-producing E. coli was blaCTX-M-55 or blaCTX-M-65, with the coexistence of blaTEM in Vietnam. These results can be implemented in monitoring systems to overcome the development of antimicrobial resistance.


Introduction
The spread of antibiotic-resistant bacteria is a global public health concern (Giurazza et al. 2021). Plasmid-mediated antibiotic resistance is a critical issue that requires urgent attention owing to the development of resistance by horizontal gene transmission (Bevan et al. 2017).
In 2015, there was a report on the acquisition of colistin resistance by a new plasmid-encoded mcr gene (Liu et al. 2016). Colistin is expected to be a therapeutic agent for carbapenem-resistant bacteria. Therefore, the spread of Communicated by Erko Stackebrandt.
* Tatsuya Nakayama t-nakayama@hiroshima-u.ac.jp colistin-resistant bacteria in society is of great concern. In this study, we found that colistin is frequently used in agricultural and livestock farms in Vietnam . Yamamoto et al. (2019) found that approximately 70% of Vietnamese residents carried colistin-resistant bacteria. Although it is assumed that Vietnamese food can lead to human colonisation with colistin-resistant bacteria, the pathways of transmission are unclear. This is especially true in chickens, where plasmid-mediated antibiotic-resistant bacteria are frequently isolated. Extended-spectrum β-lactamases (ESBL) and carbapenem-resistant bacteria are often associated with plasmidmediated antibiotic resistance. Recent studies have focussed on ESBL-producing Escherichia coli, a bacterium that demonstrates plasmid-mediated antibiotic resistance (Bevan et al. 2017). Current knowledge suggests that ESBL-producing bacteria in food and humans are detected more frequently in Vietnam than in Japan (Nakayama et al. 2015;Le et al. 2015a, b). In an earlier study in Vietnam, it was found that ESBL-producing E. coli were frequently detected in chicken meat between 2013 and 2016. Most ESBL-producing E. coli isolates carried ESBL-related genes in the bla CTX-M-1 and M-9 groups (Nakayama et al. 2015).
Monitoring of antibiotic-resistant bacteria is currently underway in many countries around the world. Therefore, it is of utmost importance to identify not only the presence of ESBL-producing E. coli but also the bla CTX-M genotype, which is widespread. Carbapenem-resistant E. coli are also well known as plasmid-mediated antibiotic-resistant E. coli; however, there is a lack of research on its importance and the degree of contamination in chicken. Hence, as with ESBL-producing E. coli, caution is required for the spread of carbapenem-resistant E. coli. In Japan, food contamination with AmpC/ESBL-producing E. coli has been reported in multiple studies (Kameyama et al. 2013;Hiroi et al. 2012). Reports indicate that bla CMY-2 and bla CTX-M-2 are the most abundant ESBL-producing E. coli in chickens. There are no reports of carbapenem-resistant and colistin-resistant E. coli isolated from chickens in Vietnam or Japan. Therefore, this study aimed to determine and compare the prevalence of colistin, carbapenem-resistant, and ESBL-producing E. coli in chickens from Vietnam and Japan.

Sampling locations and bacterial isolation
A total of 134 chicken samples (60 meat samples from retailers/supermarkets, Ho Chi Minh City, Vietnam and 33 meat samples from supermarkets and 41 faecal samples from laying hens in Japan) were investigated. Chicken meat (25 g) was added to 225 mL of buffered peptone water (BPW), and 1 g of faeces was added to 9 mL of BPW. After gentle shaking, 100 µL of the solution was spread on MacConkey agar (Eiken Chemical, Tochigi, Japan) containing 2 µg mL −1 of cefotaxime (CTX) for the isolation of AmpC/ESBL-producing E. coli. For the isolation of colistin-resistant and carbapenem-resistant E. coli, MacConkey agar (Eiken Chemical) containing either 2 µg mL −1 colistin or 0.25 µg mL −1 of meropenem (MEM), respectively, were used. The plates were incubated at 37 °C for 21 ± 3 h. After incubation, one to three colonies were selected for further analysis.

Bacterial identification and DNA extraction
Isolated bacteria were confirmed as E. coli using the following biochemical tests: the triple sugar iron (Eiken Chemical) and the lysin indole motility (Eiken Chemical) tests. Bacterial DNA was extracted by boiling suspensions of the isolates in tris (hydroxymethyl) aminomethane-EDTA buffer (Nakayama et al. 2015). Isolated bacteria were identified as E. coli using ECN amplification (Hoa et al 2020).

Multiplex PCR for phylogenetic group identification
The phylogenetic groups of isolated bacteria were determined using multiplex PCR amplification of a combination of two genes, chuA and yjaA, and a DNA fragment, TspE4C2 (Le et al. 2015a, b). The extracted DNA was amplified using a multiplex PCR kit (Qiagen, Hilden, Germany). Primer sets used for multiplex PCR are listed in Table S1. PCR was conducted under the following conditions: 35 cycles of denaturation at 98 °C for 10 s, annealing at 57 °C for 30 s, and extension at 72 °C for 30 s. The PCR-amplified products were visualised using 1.5% agarose gel and stained with midori-green (Nippon Genetics, Tokyo, Japan).

Multiplex PCR for mcr and AmpC/ESBL-related genes
The genotype of AmpC/ESBL-producing E. coli was determined using multiplex PCR, as described previously (Le et al. 2015a, b;Perez-Perez and Hanson 2002). The extracted DNA was amplified using a multiplex PCR kit (Qiagen). Multiplex PCR was conducted under the following conditions: 25 cycles of denaturation at 95 °C for 30 s, annealing at 60 °C for 90 s, and extension at 72 °C for 90 s; for AmpC, 25 cycles of denaturation at 95 °C for 30 s, annealing at 64 °C for 90 s, and extension at 72 °C for 60 s. Primers for multiplex PCR of both AmpC and ESBL-related genes are described in Table S1. The PCR-amplified products were visualised using a 3% agarose gel and stained with midorigreen (Nippon Genetics).

E. coli harbouring the mcr gene
Isolated E. coli were used to detect mcr using multiplex primers. The results showed that one strain was undetectable, and mcr-1 was detected in all the remaining E. coli isolates (Table 3).

ESBL-producing E. coli co-harbouring bla CTX-M and bla TEM
In 115 ESBL-producing E. coli isolates from Vietnamese chickens, bla CTX-M-1 (67%) tended to be more prevalent than the bla CTX-M-9 (28.7%), which was also the case in the 39 Japanese isolates ( (Fig. 1b). Multiple ESBL-related genes were detected in Vietnamese samples. In particular, the combination of bla CTX-M and bla TEM was present in more than 70% of the strains (Table 4). This result was significantly different between ESBL-producing E. coli harbouring CTX-M and TEM isolates in Vietnam and Japan (p < 0.01).

Antibiotic susceptibility and multidrug-resistant E. coli
The susceptibility assay showed that colistin-resistant E. coli harbouring mcr differed significantly from the ESBL strains derived from Vietnamese chicken in third-generation cephalosporins and quinolones (Fig. 2a, b), and the result of multidrug resistance showed that the highest numbers of drug resistance were 4 and 6, with an average of 6.7 (Fig. 3). Compared to strains detected in Japanese chickens, ESBL-producing E. coli isolates from Vietnamese chickens had a significantly higher percentage of resistance to all  treatments except β-lactams (p < 0.01) (Fig. 2b, c). The percentage of resistance to quinolones, SXT, and chloramphenicol was particularly high in ESBL-producing E. coli isolates from Vietnam (p < 0.01). ESBL-producing E. coli (8) showed the highest number of drug-resistant strains (average number of drug-resistant strains was 9.1), followed by 11 drug-resistant strains. In comparison, in ESBL-producing E. coli isolates from Japanese chickens, four and three kinds of drug resistance were the most common, with an average drug resistance of 4.7 (Fig. 3).

Discussion
The contamination of Vietnamese food with colistin-resistant bacteria has not been adequately studied. Nguyen et al. (2021) investigated colistin-resistant E. coli isolated from 15 chickens and found that colistin resistance was found in 66.7% of the samples tested. In the analysis of chicken faeces from backyard chickens in Vietnam, Kawahara et al. (2019) found that 97.2% of chickens carried E. coli harbouring the mcr gene. Le et al. (2021) investigated colistin resistance in ESBL-producing E. coli isolated from chickens. They found that 53.2% of ESBL-producing E. coli isolates were colistin-resistant, and all of these colistin-resistant isolates harboured mcr-1 . In this study, 43.3% of the food products contained colistinresistant E. coli, and almost all of them harboured mcr-1, suggesting that mcr-1 may be spreading between chickens in Vietnam.
To easily isolate colistin-resistant E. coli, we analysed chicken faeces from Japan, however, there was no trace of the bacteria. Therefore, colistin-resistant E. coli did not increase in Japanese chickens. In this study, carbapenemresistant E. coli strains were not isolated from either Vietnamese or Japanese chickens. Although it can be isolated in hospitals and the environment (Mathys et al. 2019;Zhong et al. 2021), carbapenem-resistant E. coli is not generally transmitted to foods in Vietnam or Japan.
In this study, the phylogenetic group was determined for colistin-resistant and ESBL-producing E. coli isolates. Colistin-resistant E. coli showed a pattern similar to that of the ESBL-producing Vietnamese strains. The prevalence of B2 type E. coli was much lower than that of the other types. Although there are no studies that have investigated the pattern of phylogenetic grouping of E. coli isolates from chickens, the absolute number of E. coli type B2 bacteria contamination may be low. The phylogenetic grouping of ESBL-producing E. coli in Japanese strains showed that B1 and D types were very high. This result may be due to regional differences.
To monitor antibiotic resistance, it is necessary to clarify the actual conditions of antibiotic resistance genes in chicken meat that are frequently contaminated. In Vietnam, the bla CTX-M gene in ESBL-producing E. coli is the most prevalent plasmid antibiotic resistance gene. In this study, ESBL-producing E. coli was detected in 90% (54/60) of chicken meat samples from Vietnam. Similar to our study, Le et al. (2015a, b) and Nguyen (2016) found ESBL-producing E. coli isolates in 88.3% (Le et al. 2015b), 58.7% (Le et al. 2015a), and 92.7% (Nguyen et al. 2016) of Vietnamese chicken samples. The Vietnamese government is motivated to research antibiotic resistance and plans to limit the use of antibiotics in chicken farms. However, the prevalence of ESBL-producing E. coli has not changed, 5 years later, and is considered to have become chronic due to its prevalence.
Previous studies by Le et al. (2015a, b) and Nguyen et al. (2016) examined the CTX-M sub-group and found that bla CTX-M-1 group was the most common, followed by bla CTX-M-9 group. No actual identification was performed. In this study, the bla CTX-M-1 group was the most common, followed by bla CTX-M-9 . Among the bla CTX-M-1 group, bla CTX-M-55 was the most common, and among the bla CTX-M-9 a ESBL-producing Escherichia coli strains in Vietnam b ESBL-producing Escherichia coli strains in Japan Fig. 1 Different CTX-M patterns of ESBL-producing Escherichia coli isolates between Vietnam and Japan. a ESBL-producing E. coli isolates from chicken in Vietnam. b ESBL-producing E. coli isolates from chicken in Japan group, bla CTX-M-65 was the most common. The results of a previous study of ESBL-producing E. coli from pork in Vietnam showed that bla CTX-M-55 was abundant, and the plasmid carrying bla CTX-M-55 was also found in workers and patients with urinary tract infection (Hoang et al. 2017), suggesting plasmid transmission within the community. Based on these findings, bla CTX-M-55 is abundant in pork and chicken.
Multiple studies have reported the genotyping of AmpC/ ESBL-producing E. coli isolated from Japanese chickens. Kameyama et al. (2013) reported that the AmpC β-lactamase gene bla CMY-2 (66%) was most prevalent, followed by bla CTX-M-1 (26%) and bla CTX-M-55 (26%). Hiroi et al. (2012) reported that ESBL-producing E. coli harbouring bla CTX-M-2 were most prevalent in chickens. Nahar et al. (2018) also reported that the bla CTX-M-2 and bla CTX-M-1 groups represented 45% and 34% of ESBL-producing E. coli, respectively. Our study found clear differences in the prevalence of bla CTX-M in Vietnamese and Japanese chicken samples. Hence, the pattern of bla CTX-M varied between the regions. In Vietnam, the co-harbouring genes bla CTX-M-55 or bla CTX-M-65 with bla TEM were most prevalent in ESBL-producing E. coli. These bla CTX-M genes could be used to monitor antibiotic resistance genes to improve healthcare in Vietnam. The prevalence of antibiotic resistance genes is known to change annually (Bevan et al. 2017). The bla CTX-M types of ESBL-producing E. coli isolated from humans are identified as bla CTX-M-15 (bla CTX-M-1 group) and bla CTX-M-14 (bla CTX-M-9 group), which are prevalent in many parts of the world. Accounting for prominent global trends in bla CTX-M epidemiology, there is increasing evidence suggesting that bla CTX-M-27 has started outcompeting other bla CTX-M genotypes (Bevan et al. 2017). In Vietnam, a study of ESBL-producing E. coli isolates from healthy Vietnamese people from 2013 to 2015 reported that bla CTX-M-27 was the most common type, followed by bla CTX-M-55 , bla CTX-M-15 , and bla CTX-M-65 (Hoang et al. 2017). Our results show that in 2019, the overwhelming majority of bla CTX-M isolates from chickens had ESBL-producing E. coli harbouring bla CTX-M-55 and bla CTX-M-65 . It has not been verified whether the bla CTX-M gene is transferred from chickens to humans, however, it is possible that bla CTX-M-55 or bla CTX-M-65 will be predominant in human carriers in the future. Therefore, it is necessary to carefully monitor the presence of bla CTX-M-27 in ESBL-producing E. coli in the future.