Occurrence and characteristics of Escherichia coli mcr‐1‐like in rabbits in Shandong, China

Abstract Polymyxin is regarded as the last retort to fight against multidrug‐resistant (MDR) Enterobacteriaceae. The emergency and spread of polymyxin‐associated resistance gene mcr‐1 evoked great panic of no medicine to cure the bacterial infection in society. mcr‐1 is widespread in domestic and wild animals. Therefore, continuous monitoring of its prevalence and characteristics is required. In this study, we used a polymerase chain reaction (PCR)‐based method to detect the mcr‐1 of Escherichia coli isolated from rabbits of Tai'an, China, and determined the characteristics of mcr‐1‐bearing plasmids. A total of 55 non‐duplicated E. coli was recovered from the swabs of rabbit faeces. Plasmid profiling, plasmid and chromosome PCR, complete genome sequencing, a conjugation experiment, lactose fermentation experiment, multilocus sequence typing and polymyxin resistance tests were performed to determine the characteristics of mcr‐1‐bearing plasmids. 14.6% (8/55) of the specimens were mcr‐1 positive. The mcr‐1‐positive E. coli harboured more drug‐resistant genes compared with the mcr‐1‐negative specimens, and results showed four sequence types. Overall, these findings suggested the possible threat of the transmission of mcr‐1 from rabbits to humans, especially since the gene is located on transferable plasmids making horizontal transfer relatively easy. Since food‐producing animals are necessary for our daily diet, worldwide cooperation is needed in fighting the spread of this drug resistance gene to avoid human infections with MDR pathogenic bacteria.


| INTRODUC TI ON
In China, antibiotics were very commonly used in fighting against rabbit bacterial inflammation, which was the main problem in Chinese rabbit breeding. The use of antibiotics can disrupt the balance of normal microbial colonies in the cecum of rabbits, and cause bacterial diseases and digestive disorders in rabbits. Escherichia coli in China has antibiotic resistance to most types of genes and the most prevalent genes are ESBLs and quinolones. Both ESBLs and PMQR genes were detected in faecal E. coli isolated from the non-human primates in six zoos in China, and the prevalence of ESBL-encoding genes was 32%, and the prevalence of PMQR genes was 33% (Yang et al., 2012).
In China, the output of rabbit meat is more than 400,000 tons, and the export volume is maintained between 20,000 tons and 30,000 tons every year (Agnoletti, Brunetta, Bano, Drigo, & Mazzolini, 2018). Colistin had been forbidden to use in human medicine for its nephrotoxicity, however, it was still useful in rabbits to prevent diseases like diarrhoea and promote growth, and drug-resistant genes are now widely distributed in the intestines of farm animals, which are continuously being identified (Briñas et al., 2002;Naseer & Sundsfjord, 2011;Bryan, Shapir, & Sadowsky, 2004). The emergence of mcr-1, a plasmid-mediated colistin resistance gene, has alerted the public health systems and led to changes in how resistance is perceived globally. Liu firstly reported that mcr-1 widely existed in E. coli, which was the most common host of mcr-1 in China (Liu et al., 2016). Until now, mcr-1 gene has been identified in ten diversified species of the Enterobacteriaceae, isolated from over 40 countries/regions (Sun, Zhang, Liu, & Feng, 2018). Following this pattern, it is likely that drug-resistant bacteria are present in rabbit faeces (Gao et al., 2015;Zhao, Ye, Chang, & Sun, 2017). Polymyxin was considered as a promising antimicrobial peptide, and very few bacteria showed polymyxin resistance. However, Chinese researchers identified mcr-1 as a gene conferring resistance to colistin and polymyxin (Liu et al., 2016). Although mcr-1 has been reported and detected worldwide, its global prevalence remains largely unknown. Liu et al. (2016) screened for mcr-1 in E. coli in raw pork and found that the gene was located on plasmids. The prevalence of E. coli mcr-1 in rabbits in China has not been reported. In these studies, the key methods to detect the location of genes were mainly based on Southern blotting. However, their detection methods were not based on polymerase chain reaction (PCR). Therefore, in this study, we employed a simpler method to determine the location and characteristics of E. coli mcr-1 among rabbits in China, that was, PCR combined with complete genome sequencing, which can help in estimating the existence, location and prevalence of mcr-1. We also developed a method of combination of conjugation, PCR and fermentation test to further prove that plasmid harboured mcr-1. For the final determination, we applied complete genome sequencing to the mcr-1-positive strains.

| Sample collection and identification of E. coli
The rabbits had been raised in large rabbit farms, free from thirst or starvation, and without signs of anxiety, fear or depression. The farms generally had 300~500 basic female rabbits. The formula for rabbit feed was 17% corn, 24% bran, 21% soybean meal, 5% imported fish meal, 3% active yeast and 30% grass powder. Meanwhile, polymyxin was added to feed with a proportion of 1:10,000. Faecal samples were randomly collected from the diarrhoea rabbits on three farms by rectal swabs, and the faeces were from individual animals. The three farms were separately selected in three administrative counties. Because the sampling process did not harm the rabbits, ethical approval was not required for the study.
Sixty rabbit faeces were collected in aseptic tubes (Agnoletti, 2012;Boullier & Milon, 2006), and plated on MacConkey agar to select and identify E. coli. The suspicious colonies were identified by bacterial biochemical tests. After biochemical identification, Gram stain and microscopic examination were performed to observe the morphology of the bacteria for confirmation. Positive colonies were then chosen for further biochemical identification using the automated API 20E system (Sysmex bioMérieux).
Recovered E. coli was cultivated in Luria-Bertani liquid medium containing 2 μg/mL polymyxin B, and positive specimens were selected as the PCR detection templates.

| PCR detection of mcr-1
We further attempted to amplify mcr-1 from extracted plasmids and bacterial chromosomes. The samples of mcr-1-positive strains were separated and the plasmids were extracted with the OMEGA plasmid kit (Omega Bio-Tek Co., Ltd.) and subjected to electrophoresis.
The extracted plasmid was used as the template for PCR.
The DNA isolated with the method of alkaline lysis from E. coli strains was amplified by PCR using mcr-1-specific primers, F: 5ʹ AGTAGGCGTTTATTTGATAAATACGGCA 3ʹ; R: 5ʹ TTATATCAGATAAATTGTACTGGATTTC 3ʹ. PCR systems included 25 μL PCR mix, 21 μL deionized water, 1 μL forward primer, 1 μL reverse primer and 2 μL template, for a total of 50 μL (Shanghai Sangon Co., Ltd.). The reaction programme was as follows: 94°C pre-denaturation for 5 min, followed by 28 cycles of 94°C denaturation for 30 s, annealing at 55°C for 30 s, and 72°C extension for 30 s. A final extension step was conducted at 72°C for 7 min. The PCR products of mcr-1 were then subjected to electrophoresis at 140 V for 30 min.
The positive specimens were sent to Sangon for direct sequencing for confirmation (White, Mclver, Deng, & Rawlinson, 2006), and the sequences of mcr-1-positive strains were compared by the BLAST of the National Center for Biotechnology Information website.
PCR was then carried out with the extracted plasmid as a template using primers specific to mcr-1 and other resistance genes including bla TEM under the same PCR conditions described earlier (Table S1).
Similarly, bacterial genome chromosomes were extracted and purified from the samples of mcr-1-positive strains, and PCR analysis was performed with mcr-1-specific primers as described earlier.

| Plasmid characterization and sequencing
One sample was randomly selected for sequencing. The concentration of the extracted genome was determined using the method of both Qubit Fluorometer and Nanodrop. Those meeting the requirements of sequencing were sent for sequencing, and the coding genes and structure were analysed by bioinformatics, such as spade, prokka and Pfam2go database. Through a comparative analysis of the extracted plasmids, the E. coli R45 strain carrying the mcr-1 gene was ultimately selected, and the extracted plasmid from this strain was designated pR45. A whole-genome shotgun strategy was used to construct libraries of different inserted fragments. Paired-end sequencing was performed on the Illumina MiSeq platform. SPAdes genome assembler (v 3.7.1) was used to construct contigs and the scaffold by the ab initio assembly of sequencing data, which were then removed and corrected.

| Conjugation experiments
To prove that the antibiotic resistance gene in E. coli has the ability to transfer in vitro, 55 mcr-1-harbouring E. coli strains were isolated, which were resistant to polymyxin but sensitive to sodium azide. E. coli J53 was resistant to sodium azide and sensitive to most antibiotics.
Conjugative testing was performed using the filter mating method, mixing at a ratio of 1:1 in broth culture, as previously described (Zhang, Zhou, Guo, & Chang, 2015). The transfer rate was determined, subsequently. The resulting transconjugants were selected on brain heart infusion agar plates supplemented with polymyxin B (2 mg/L) and sodium azide (Guardabassi, Schwarz, & Lloyd, 2004). The conjugated bacteria were also observed using plasmid extraction and electrophoresis analysis. Transfer of the resistance gene was considered to have taken place when the plasmids were transferred from the wild-type mcr-1-positive bacterium to the recipient bacterium.

| Multilocus sequence typing
According to http://bigsdb.paste ur.fr/ecoli/ prime rs_used.html, eight pairs of primers for housekeeping genes (dinB, icdA, pabB, polB, putP, trpA, trpB and uidA) were designed and used for PCR (Zhao, Yang, Ju, Chang, & Sun, 2018). The products of PCR amplification were then sequenced by Shanghai Sangon Biotech Co., Ltd. The results were amended using Chromas and DNAStar software and then submitted to the Pasteur online database for processing. The allele number of each housekeeping gene was obtained and the sequence type (ST) of each strain was acquired (Dotto, Giacomell, Grilli, Ferrazzi, & Carattoli, 2013).

| Phylogenetic analysis
The phylogenetic tree of the eight mcr-1 sequences and mcr-1 genes on GeneBank was constructed by a maximum likelihood method using Megalign 7.1.0 (DNAstar Co., Ltd) to determine the relationships among strains.

| Isolation and identification
A total of 55 E. coli strains were isolated from the overall 60 samples.

| Prevalence of mcr-1
Eight specimens were found to be mcr-1 positive, representing a positive rate of 14.6% (Figure 1). The accession number of the eight sequences were MH395740, and MH602237-MH602243.

| Plasmid sequencing results
The concentration of plasmid for analysis was 74.6 ng/μl. Complete genome sequencing was conducted on the mcr-1-positive strains.
BLAST showed that mcr-1 was located on the plasmid. The extracted plasmid, designated pR45, encoding 19 predicted genes including mcr-1 ( Table 2). The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive, under accession number CRA002525 that is publicly accessible at https:// bigd.big.ac.cn/gsa.

| Drug resistance test
The results of drug resistance phenotyping and resistance were consistent with the results of mcr-1 detection. (Table 3).
The results demonstrated that the E. coli resistance gene had transferability in vitro, and that the mobile plasmid played an important role in the process of drug resistance transmission in E. coli, among which ESBLs was included in this analysis as it is the most prevalent antimicrobial resistance genes in the samples (Table S2). Moreover, the plasmid profile of donor strain and recipient was the same, demonstrating that the plasmid of R45 transferred to J53 (Figure 2).

| Conjugation tes t s
To prove the transferability of mobile plasmids in vitro, 55 strains of E. coli were used as donor bacteria, and 39 transconjugants were obtained successfully with the transfer rate as high as 70%. The conjugation tests confirmed the horizontal transfer of mcr-1 in E. coli strains obtained from rabbit faeces, therefore proving that mcr-1 was located on plasmids ( Figure 3).

| Phylogenetic analysis
Phylogenetic tree showed the evolutionary relationships among the eight mcr-1 sequences, demonstrating that although the eight positive strains were non-duplicated E. coli, their mcr-1 sequences were identical (Figure 4).

| Char ac teris tic s of mcr-1
Structure of plasmid pR45 showed that the whole length of the plasmid was 237,728 bp, and mcr-1 ranged from 49,992 to 51,617 bp ( Figure 5).
The structure of plasmid finally determined that mcr-1 located on the plasmid.

| D ISCUSS I ON
The prevalence of mcr-1 (8/55, 14.6%) detected in E. coli strains obtained from rabbits is similar to that reported in a study conducted in Italy (50/320, 15.6%) (Fabrizio, Romina, Luca, Ilenia, & Elena, 2018), and is markedly higher than that reported for humans (1~2%) (Yi, Liu, Wu, Liang, & Liu, 2017). This high rate may be due to the greater use of polymyxin in farms than in clinical practice.
Most importantly, all of the mcr-1-positive strains obtained in this study were isolated from a single farm among the three sampled farms which was perhaps because the amount of polymyxin use varied across the different farms, which would impose different selection pressures on mcr-1.
Because of the limitation of the total amount of specimens, it is difficult to generalize the results overall. Nevertheless, the antibiotic PCR tests demonstrated that the mcr-1-positive plasmids were more likely to harbour other resistant genes than mcr-1-negative plasmid. Accordingly, the mcr-1-positive E. coli had a greater probability of being MDR than mcr-1-negative E. coli (p < .05).
Bacteria without plasmids readily gained donor bacterium plasmids and the mcr-1 gene along with the ability for lactose fermentation and polymyxin resistance at the same time. Therefore, these results strongly suggest the high horizontal dissemination potential of mcr-1.
Moreover, the low diversity of mcr-1 sequences among the E. coli strains indicated that the mcr-1 gene was most likely derived from The bold words indicate the topic of this case study. PB is the drug that mcr-1-positive bacteria are resistant to. The article aimed to investigate the prevalence of mcr-1. We highlight mcr-1 and PB to illustrate the link between the drug resistant phenotype and .PB resistance gene mcr-1. This may be related to the fact that this region is relatively isolated, far from the city, with minimal flow of people.
Although mcr-1 gene was very conservative, they have diverse STs, demonstrating that mcr-1 had different origins. The resistance gene mcr-1 was found in eight strains of bacteria, which showed that the presence of plasmids for bacteria made it possible to produce drug resistance and survive in adversity.
Resistance genes not only transfer from one bacterium to another or from one bacterium species to other species but also move geographically consequently (Kun et al., 2006). Therefore, the threat of drug resistance is not localized to a given animal farm or region, but represents a worldwide concern requiring global cooperation.
Indeed, the fact that the bacterial resistant gene is located on the plasmid makes it potentially more difficult to control than a chromosomal gene. Plasmid transmission makes the spread of drug resistance genes easier and faster, and since the same plasmid can carry a variety of resistance genes, the recipient can immediately become resistant to multiple drugs. This finding suggested that it would be very difficult to cure humans infected with MDR pathogenic bacteria (Valat et al., 2016).

| CON CLUS ION
The conjugation test and complete genome sequence analysis of the ligated plasmid demonstrated that the E. coli resistance gene mcr-1 was circulating in rabbits of Eastern China, with the ability for horizontal transfer in vitro, indicating that the mobile plasmid played an important role in the process of antibiotic resistance of E. coli (Silva et al., 2010). As the antimicrobial resistance-positive bacterial strains can survive in the presence of antibiotics, the bacteria can readily acquire additional drug resistance genes, resulting in a new MDR phenotype for the bacteria. Therefore, continuous selective pressure of antibiotics in farms will result in the production of new drug resistance genes that can readily circulate among domestic and wild animals, and even humans. To prevent the impact of mcr-1 on humans, we should first reduce the probability of mcr-1-harbouring strains in humans for proliferation and infection. Governments should carefully monitor and report the use of antibiotics in their jurisdictions.
It is possible to effectively control the further spread of mcr-1 in humans and animals and to curb the development of polymyxin resistance. The whole world should cooperate to deal with the problem of drug resistance.

CO N FLI C T O F I NTE R E S T
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1002/vms3.340.