Antimicrobial resistance and population genomics of extensively drug resistant Escherichia coli in pig farms in mainland China: a national wide epidemiological and microbiological study

Antimicrobial resistance (AMR) is one of the most urgent threats to the global public health, and the expanding use of antimicrobials in food animals is considered as a main reason for the worldwide rapid increasing of AMR. However, AMR in animals in many regions are poorly documented. China is the largest pig-rearing and pork consumption country in the world. In the present study, we identied AMR in pig farms from all provinces (including Tibet and Qinghai) of mainland China by investigation of a common indicator bacterium Escherichia coli from both pigs and the breeding environmental samples. A total of 2693 samples from pigs and environments in 67 pig farms in all 31 provinces of mainland China were collected between 1 October 2018 to 30 September 2019, and a total of 1871 E. coli strains were isolated. By testing the susceptibility of these 1871 E. coli isolates on 28 types of antibiotics that commonly used in both human and veterinary medicine, we found that resistance to tetracycline (96.26%), chloramphenicol (82.04%), moxioxacin (81.56%), and trimethoprim/sulfamethoxazole (80.38%) were the broad phenotypes among these E. coli isolates from pig farms in China. A proportion of E. coli isolates were resistant to colistin (3.79%), carbapenems (imipenem [2.62%], meropenem [2.30%], ertapenem [2.46%]), and broad-spectrum-cephalosporins (ceftriaxone [29.56%], cefepime [14.00%]). More than 70% of the isolates displayed multidrug-resistant (MDR), and/or extensively drug-resistant (XDR) phenotypes, and MDR/XDR-E. coli was observed in pig farms in all provinces of mainland China. We also systematically revealed the distribution of O-serogroups, sequence types, resistance genes, virulence factors encoding genes, and putative plasmids of MDR/XDR-E. coli in pig farms from different provinces of China, and partially characterized the pathotypes of certain MDR/XDR-E. coli strains. In addition, the genetic transmission basis of the bla NDM , mcr, ESBL-encoding, uoroquinolone-resistance, and tetX genes were addressed in this study. Most importantly, we suggested a very high genetic propensity of the pig farm-sourced MDR/XDR-E. coli in spreading into humans. To the best of our knowledge, this is the rst study on a national scale that the resistance phenotypes and population genomics of E. coli in pig farms in China are revealed. Our data presented herein will help understand the current prole of AMR in pigs and also provide reference for policy formulation of AMR control action in livestock in China.


Introduction
Antimicrobial resistance (AMR) is one of the most urgent threats to global public health. Recently, the emergence and rapid increase of multidrug-resistant (MDR), extensively drug-resistant (XDR), and even pan-drug-resistant (PDR) bacteria, particularly those resist the last-resort drugs (carbapenems, colistin, and tigecycline) have catalyzed the serious concern that the world may get its rst inkling of impending catastrophe of a return to the pre-antibiotic [1][2][3][4] . Several drug-resistant bacteria have received a worldwide great concern, of particular note is Escherichia coli 1,5,6 . This bacterial species is not only a leading cause of foodborne infections, but also represents a major reservoir of antimicrobial resistance genes (ARGs) due to its great capacity to accumulate ARGs, mostly through horizontal gene transfer 7 . A recent study showed the total economic cost of AMR in Staphylococcus aureus, E. coli, Klebsiella pneumoniae, Acinetobacter baumannii, and Pseudomonas aeruginosa reached $2.8 billion per year in the United States 8 . In addition, E. coli has been found to play important roles in disseminations of bla NDM−1 , mcr, and/or tet(X3)/(X4); these ARGs mediates resistance to the last-resort drugs (carbapenems, colistin, and tigecycline) for gram-negative bacteria which may lead to non-antimicrobials available in both human and veterinary medicine 3,4,9 . In addition, extended-spectrum beta-lactamase (ESBL)-producing E. coli is also a major global health problem 1,5 .
The large and expanding use of antimicrobials in food animals is considered as a main reason for the worldwide rapid increasing of AMR 10 . In the past two decades, meat production has plateaued in highincome countries but has largely grown in low-and middle-income countries particularly in Asia 11 . Under this background, antimicrobials are widely used in livestock either for growth promotion and/or for health maintain. Globally, 73% of all antimicrobials sold on the earth are used in animals raised for food 11 . The large and expanding use of antimicrobials can lead to the development of drug-resistant bacteria in the animal guts, and these resistant bacteria can spread to humans 1 . To date, AMR in animals in low-and middle-income countries are poorly documented, and this may be in part due to the absence of systematic surveillance systems and large epidemiological studies 11 .
China is the largest developing country in the world. Many provinces in China still belong to the low-and middle-income regions. China is also the number one country for antimicrobials producing, pig rearing and production, and pork consumption in the world 12,13 . Therefore, understanding the current pro le of AMR in pig farms in China has a great signi cance for global AMR surveillance. However, pig producing in China is very complex with different sizes of pig farms and production models in various regions. To date, there is still lack of systemic data re ecting the condition of AMR in the pork production chain in China. Here, we identi ed AMR in pig farms in most regions in China by investigation of E. coli, which is a commonly used biomarker of AMR in pig farms 14 . To the best of our knowledge, this is for the rst time that the pig farms in all regions in mainland China were included for the epidemiological investigation.
This study will contribute to understand the current pro le of AMR in pigs and also provide reference for policy formulation of AMR control action in livestock in China.

Results
Antimicrobial resistance phenotypes and microbiological characteristics of E. coli isolates from pig farms in China. Between 1 October 2018 to 30 September 2019, a total of 2693 samples from pigs and breeding environments in 67 pig farms in all 31 provinces of mainland China were collected for E. coli isolation, and a total of 1871 E. coli strains including 1108 strains from pig samples (anal swabs and/or diarrheal feces) and 763 strains from environmental samples (swabs of drinking and/or fecal slurry, oors, drinking and/or food troughs) were nally obtained (Fig. 1A). Initially, we attempted to include more farms in each of the provinces, but the sudden outbreak of African Swine Fever (ASF) in late 2018 and the coronavirus disease    (Fig. 1C). Many isolates showed resistance to TGC (37.31%, n = 698), but most of them with MIC values ranging from 0.5 µg/ml to 1 µg/ml (92.98%, n = 649), only a very small proportion of them showed high-level of resistance (MIC ≥ 4 µg/ml; 0.72%, n = 5) (Figs. 1B & C). When comparing isolates from different samples of collection, resistance rates of isolates from pigs against SXT, GEN, CIP and NOR were signi cantly higher than those of the isolates from environmental samples, while isolates from environmental samples had signi cantly higher resistance rates against TGC, AMS, and AMC compared to isolates from pigs (Supplementary materials Figure S1).
Genomic associations with antimicrobial resistance phenotypes. We rst searched for the presence of known acquired antimicrobial resistance genes (ARGs) in the genomic sequences. This approach led to the detection of 109 kinds of ARGs, including 28 aminoglycoside resistance genes, 35 β-lactam resistance genes, 7 phenicol resistance genes, ve tetracycline resistance genes, ten quinolone resistance genes, as well as three sulfonamide resistance genes, seven macrolide resistance genes, two rifampicin resistance genes, seven trimethoprim resistance genes, two fosfomycin resistance genes, and four lincosamide resistance genes (Supplementary materials Table S3). In particular, bla CTX−M−55 and bla TEM−1B were most-frequently determined ESBL genes in E. coli isolates (bla CTX−M−55 was carried in 209 sequenced isolates while bla TEM−1B was carried in 181 ones) from pig farms in China (Fig. 4A); while qnrS1, oqxB, and oqxA were most-frequently determined quinolone resistance genes (presence in 233, 114, and 109 sequenced isolates, respectively; Fig. 4B). Among the tetracycline-resistant isolates, tet(A) and tet(M) were most-frequently determined resistance genes (presence in 453 and 170 sequenced isolates, respectively; Fig. 4C). Notably, three carbapenem resistance genes (bla NDM−1 , bla NDM−5 , bla NDM−7 ), two colistin resistance genes (mcr-1.1, mcr-3.1), and one tetracycline resistance genes (tetX4) were determined, and they conferred resistance to carbapenems, colistin, and tigecycline, respectively (Figs. 4C ~ E). Over 15% NDM-producing isolates carried colistin resistance gene mcr-1, while only one isolates carried both colistin resistance gene mcr-1 and high-level tetracycline resistance gene tetX4. We also observed there was a strong correlation between the presence of ARGs and the expected resistance phenotypes. Only carbapenem resistance isolates harboring the three carbapenem resistance genes, while only isolates with high level resistance to tigecycline (MIC value ≥ 4 µg/ml) containing tetX4; in addition, the mcr genes were only found in isolates with resistance phenotypes to colistin (Fig. 4).
Sequence alignment also revealed that bla NDM−1 -carrying plasmid pXD33-05 was highly homologous to a bla NDM−1 -carrying plasmid pHNEC55 (GenBank accession no. KT879914) (Fig. 6B). The average nucleotide identity (ANI) between the backbones of pXD33-05 and pHNEC55 was higher than 99%. However, the MDR elements between the two plasmids were different. Structurally, the MDR elements of pXD33-05 consisted of two ARG cassettes, including a 7.6-kb cassette harboring a bleomycin resistance gene, a aminoglycoside resistance gene aph(3')-VI and bla NDM−1 as well as a 2.4-kb one harboring a aminoglycoside resistance gene rmtB and a ESBL-encoding gene bla TEM−1B (Fig. 6B). The 7.6-kb cassette was anked by an IS6 and an IS3 elements, while the 2.4-kb cassette was anked by a Tn3 and an IS6 elements. Sequence comparisons showed that the mcr-1-carrying plasmid pXD33-06 was highly homologous to plasmid pWI2-mcr (GenBank accession no. LT838201) (Fig. 6C). However, it displayed little homology to the high-impact mcr-bearing plasmid pHNSHP45 (GenBank accession no. KP347127) reported in China 4 (Fig. 6C). In addition to mcr-1, no other ARGs were found on pXD33-06 (Fig. 6C). We also analyzed the genetic environments of tetX4 which mediates resistance to high level tigecycline in E. coli isolates from pig farms in China. ONT sequencing on two high-level-TGC resistant isolates HB50 and SY36 revealed that tetX4 was carried by an IncX1 plasmid in both isolates, which was highly homologous to a previously reported E. coli tetX4-harboring plasmid pYY76-1-2 (GenBank accession no. CP040929) 16 (Fig. 6D). In all determined tetX4-carrying elements, tetX4 was adjacent to an ISCR2 element (Fig. 6D).
High genetic propensity of farm sourced XDR-E. coli in spreading into humans. To determine the genetic propensity of the MDR/XDR-E. coli isolates to spread into the human sector, the genetic relatedness of the 515 MDR/XDR-E. coli isolates from pig farms in China to 287 publicly available draft genomes of human commensal E. coli (Bioproject no. PRJNA4001047) were investigated 17 . The 802 E. coli isolates were phylogenetically divided into three lineages (Fig. 7A), and the 515 MDR/XDR-E. coli isolates from pig farms displayed a close relatedness to the 287 human E. coli strains (Fig. 7B). A large proportion of pigfarm originated MDR/XDR-E. coli isolates showed high genetic similarity (443/515, differed by only less than 1000 SNPs) to the human originated E. coli strains in China ( Fig. 7; Supplementary materials Table  S7). Of particularly concern is that 44.27% (228/515) of the MDR/XDR-E. coli isolates from pig farms differed by only less than 100 SNPs (as small as 3 SNPs) from the human E. coli isolated in China ( Fig. 7; Supplementary materials Table S7).

Discussion
In this study, we investigated AMR phenotypes of E. coli isolated from both pig samples and environmental samples in pig farms from all provincial regions in mainland China. As an important natural reservoir of ARGs and a commonly-used biomarker bacteria for monitoring AMR 7,14 , the resistance pro le of E. coli from pig farms may re ect the AMR condition in these farms. To the best of our knowledge, this is the rst time the AMR phenotypes of E. coli isolates from pig farms in all provinces of mainland China, including Tibet and Qinghai where there were little related data before 11 being reported.
Our determination of resistance phenotypes of E. coli isolates from pig farms in different provinces in China on 28 types of antimicrobials commonly used in both human and veterinary medicine indicated a worrisome condition of AMR in pig farms (Figs. 1&2). MDR and even XDR were the common phenotypes of E. coli isolates recovered in this study, and MDR/XDR-E. coli isolates were widely determined in pig farms in different provinces in whole mainland China, including Tibet, Xinjiang, and Qinghai (Fig. 2). This worrisome condition is widely accepted as the result of overuse and abuse of antibiotics in pig industry in the country 12,18 . During the past decades, along with a rapid increase in economy, the production of meat, eggs, and milk in China has rapidly increased, especially for pork, the main source of animal protein for most Chinese people 12,19 . To meet increasing demand for pork, both the number and the size of pig farms have grown markedly, and a massive amount of antibiotics are used in the country to support its rapid increase in pig production 12 . According to available reports 13,20,21 , major classes of antibiotics extensively used in China's pig farms include sulphonamides, tetracyclines, uoroquinolones, macrolides, and β-lactams, and in particular, uoroquinolones and β-lactams contributed more than half. Consistently, resistance to antibiotics belonging to those antibiotic classes were found to be the broad phenotypes for E. coli isolates from pig farms in China (Fig. 2).
Among the drug-resistant E. coli determined in pig farms in China, of particular note are the isolates with resistance phenotypes to carbapenems, CL, and TGC (Fig. 2). In particular, over 15% NDM-producing isolates carried CL resistance gene mcr-1. All of these antibiotics are proposed as the last-resort antibiotics for treating infections caused by MDR Gram-negative bacteria 4,22 . According to the recent o cial policy, the carbapenem antibiotics are not approved to be used in livestock in China, and it remains to be elucidated why E. coli isolates with resistance phenotypes to these antibiotics are recovered from pig farms. Notably, bacterial isolates with resistance phenotypes to carbapenems have also be recovered from poultry farms in China 23 . Although we cannot exclude the possibility that some pig farms might secretly use carbapenem antibiotics without receiving approval, a more preferrable possibility for the acquisition of these phenotypes is due to contaminated in-house environment 23 . The use of carbapenems in Chinese hospitals might cause the contamination of carbapenem-resistant bacteria or mobile carbapenem-resistant genes (e.g., bla NDM ) in environments (e.g., water, air, etc.). These resistant bacteria or ARGs may spread to livestock farms and in turn, lead to the emergence of carbapenem-resistant bacteria in animals. It was worth note that most of the carbapenem-resistant E. coli isolates recovered from pig farms in China in this study carried the NDM-producing gene bla NDM (Fig. 4D).
Continuous monitoring the persistence and spread of carbapenem-resistant bacteria in animals particularly food producing animals is essential, as these bacteria may represent high risks to human health.
The recovery of CL-resistant E. coli from pig farms might also due to the acquisition of CL-resistant mcr gene from the environment. Although the colistin withdrawal policy in 2017 and the decreasing use of colistin in agriculture have had a signi cant effect on reducing colistin resistance in both animals and humans in China 24 , plasmid-mediated CL resistance mcr genes might also persist in the environments of livestock farms 23,25 . The persistence of the mcr genes may contribute to the dissemination of CLresistant bacteria. In agreement with this hypothesis, the plasmid-mediated colistin resistance genes mcr-1 and mcr-3 were widely determined in E. coli isolates with resistance phenotypes to CL (Fig. 4E). Notably, CL-resistant E. coli were found in pig farms in 11 provinces of China in this study, suggesting a still worrisome condition for the persistence and dissemination of CL resistance. More active actions should be taken to solve the problem and continuous monitoring is still necessary 24 . Although a large number of E. coli isolates from pig farms in China showed resistance to TGC (37.31%, n = 698), most of them displayed low-level of resistance (MIC values ranging from 0.5 µg/ml to 1 µg/ml; 92.98%, n = 649), and only ve isolates displayed high-level of resistance (MIC ≥ 4 µg/ml; 0.72%) (Figs. 1B & C). However, the previously reported high-level TGC conferring gene tetX4 3 was only found in high-level TGC resistant isolates. Instead, several tet genes were widely detected in low-level TGC resistant isolates (Figs. 1B & C). Phenotypes of low-level TGC resistance might be conferred by these tet genes such as tetA, tetB, tetC, and/or tetM 26 .
By performing whole genome sequencing, the population genomics of MDR/XDR-E. coli in pig farms in China was revealed. During the analyses, we characterized several O-serogroups that are frequently associated with E. coli pathotypes including Enterotoxigenic E. coli (ETEC), Enteropathogenic E. coli (EPEC), Enterohaemorrhagic E. coli (EHEC), Enteroaggregative E. coli (EAEC), Enteroinvasive E. coli (EIEC), and Diffusely adherent E. coli (DAEC) 27,28 (Fig. 3). Importantly, several marker VFGs of E. coli pathotypes (e.g., astA, eae, east1, stb, stx 2eB , toxB, etc. 27,28 ) were also determined in E. coli isolates belonging to these O-serogroups (Fig. 3). The wide presence of these MDR/XDR-E. coli in pig farms represents a great health risks and should receive more attentions. Determination of ARGs identi ed mobile genes conferring resistance to the tested antibiotic classes in E. coli isolates in pig farms in China (Fig. 3; Supplementary materials Table S3). In addition, a large number of macrolide-resistance genes were also determined. These ndings suggest that the extensive use of sulphonamides, tetracyclines, uoroquinolones, macrolides, and β-lactams in China's pig farms facilitates the acquisition of related resistance genes in E. coli in farms. Along with the detection of multiple ARGs, we also detected many putative plasmids in E. coli isolates from pig farms in China (Supplementary materials Table S4). Several groups of these plasmids (e.g., IncA/C, IncB/O, IncFIB, IncX1, IncFIIA, IncX2, IncY, etc.) are capable of carrying transfer, MDR, and virulence functions in broad-host-range 29 . The presence of those plasmids may accelerate the dissemination of ARGs and even VFGs as well. We also determined several groups of plasmids that were rarely associated with the spread of speci c ARGs. For example, the IncX3 plasmid has been reported to be account for majority of bla NDM carriage in livestock farms 23,30 . However, we determined an IncFII-typebla NDM -carrying plasmid (Fig. 6B), and this type of plasmid was observed in all E. coli isolates coproducing NDM and MCR (Supplementary materials Figure S2). In addition, the mcr-carrying plasmid determined E. coli isolates co-producing NDM and MCR in this study was also different from the plasmid mediating the dissemination of mcr in China reported recently 4 (Fig. 6C). The presence of these plasmids makes the dissemination of bla NDM−1 and/or mcr more heterogeneous. It was worth note that the elements mediating the spread of high-level TGC resistance gene tetX4 in E. coli isolates from pig farms in this study shared highly homologous to those reported previously 3,17,31 (Fig. 6D), suggesting the dissemination of tetX4 might be not as heterogeneous as bla NDM and mcr. However, continuous monitoring should be taken in the future. Most notably, the MDR/XDR-E. coli isolates from pig farms displayed a very close relatedness to the E. coli strains from humans in China (Fig. 7). Most pig farmorigin E. coli isolates in this study differed by only less than 1000 SNPs to the human-originated E. coli strains, and many differed by only less than 10 SNPs (Fig. 7; Supplementary materials Table S7). These ndings suggest a very high genetic propensity of farm sourced MDR/XDR-E. coli in spreading into humans in China.

Conclusions
Although this work has a limitation that we could not include more pig farms in different provinces in China for sample collection due to the outbreak of ASF and COVID-19, our sample collection still covers pig farms in all provinces of China. On a national scale for the rst time, we characterized the resistance phenotypes as well as population genomics of E. coli in pig farms in China. Our results revealed a worrisome condition of AMR in pig farms in China and there is still a long way for China to take actions to reduce AMR in livestock. Fortunately, Chinese government has taken a series of active actions to solve the urgent AMR conditions in animal husbandry. A noteworthy action is the Ministry of Agriculture and Rural Affairs (MARA) has issued a policy to ban the addition of antibiotics for promoting animal growth in feed from July 1, 2020 (MARA Announcement No.194, 07-10-2019). In this study, we also systematically revealed the distribution of O-serogroups, sequence types, ARGs, VFGs, as well as putative plasmids of MDR/XDR-E. coli in pig farms in different provinces of China. These data will provide comprehensive insights to help understand AMR in pig farms, and may also be bene cial for the government to make policies for reducing AMR in pig industry in the country. Notably, we also determined many MDR/XDR-E. coli with potential pathogenicity to humans and most importantly, we found there was a very high genetic propensity of pig farm-sourced MDR/XDR-E. coli in spreading into humans. The persistence and dissemination of these isolates represent important health risks and should receive more attentions.

Materials And Methods
Sample collection, identi cation, and antimicrobial susceptibility testing Afterwards, swab cultures were streaked on MacConkey agars and were incubated at 37°C for 16 hours. E. coli isolates were con rmed by PCR ampli cation of the 16S rRNA gene with primers (F: 5'-GAAGCTTGCTTCTTTGCT-3', R: 5'-GAGCCCGGGGATTTCACAT-3') documented previously 32 . PCR assays amplifying the presence of seven house-keeping genes of E. coli (adk, fumC, gyrB, icd, mdh, purA, and recA) 33 were set for double con rmation. using the CLSI breakpoints (CLSI M100, 28th Edition). If CLSI breakpoint was not available, the European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoint (v 8.1) was alternatively used for the interpretation. Each antibiotic was tested with three duplicates. E. coli ATCC 25922 was used as quality control.
Whole genome sequencing and data availability All E. coli strains with resistant phenotypes to one of the carbapenems tested, broad-spectrumcephalosporins (ceftriaxone and cefepime), tigecycline, or colistin, were selected for NGS. Bacterial genomic DNA was extracted by using a commercial DNA Kit (TIANGEN, Beijing, China). The quality and concentration of the bacterial genomic DNA was evaluated via electrophoresis on a 1% agarose gel as well as NanoDrop2000 (Thermo Scienti c, Waltham, MA, USA) and Qubit 4 Fluorometer (Thermo Scienti c, Waltham, USA). Libraries were constructed based on the quali ed DNA by using a NEBNext Ultra™ II DNA Library Prep Kit (New England BioLabs, Ipswich, USA), and were sequenced on a NovaSeq 6000 platform using the pair-end 150-bp sequencing protocol (Novogene, Beijing, China). Raw reads with low quality were removed as described previously 34 . High-quality reads were de novo assembled via SPAdesv3.9.0 to generate genome contigs.
The complete genome sequences of bla NDM− 1 -, mcr-1-, and/or tetX4-carrying plasmids were generated by ONT sequencing in combination with the Illumina technology. Plasmid DNA was extracted using the phenol-chloroform protocol combined with Phase Lock Gel tubes (Qiagen GmbH) and was detected by the agarose gel electrophoresis as well as quanti ed by Qubit® 2.0 (Thermo Scienti c, Waltham, USA).
Libraries for ONT and Illumina sequencing were prepared using an SQK-LSK109 kit and a NEBNext® Ultra™ DNA Library Prep kit, respectively. Prepared DNA libraries were sequenced using Nanopore PromethION platform and Illumina NovaSeq PE150 at Novogene Co. LTD (Tianjin, China), respectively.
ONT and Illumina short reads were nally assembled and combined using the Unicycler v0.4.4 software with default parameters.
Whole genome sequences (WGSs) of E. coli isolates have been deposited into GenBank (BioProject accession no. PRJNA688628). GenBank accession numbers are given in Supplementary materials Table   S1. Bioinformatical analysis Sequence alignments were performed by using the MAFFT software version 7.471 39 . RAST Sever was used for sequence annotations 40 . Average nucleotide identities between two genome sequences were calculated by ANI calculator 41 . A comparative genome analysis was performed and visualized using the BRIG package 42 and/or the EasyFig package 43 . Phylogenetic trees based on whole-genome single nucleotide polymorphisms (WG-SNPs) were generated using Parsnp (version 1.2) software 44 and were visualized using Interactive Tree Of Life (iTOL v.5) 45 . The draft genomes of 287 human commensal E. coli (PRJNA400107) were downloaded from NCBI and included for phylogenetical analysis in this study 17 .

Plasmid conjugation experiments
Plasmid conjugation experiments between carbapenem-resistant E. coli, colistin-resistant E. coli, and/or tigecycline-resistant E. coli (donors) and rifampin-resistant E. coli C600 (recipient) were performed as described previously 25 . Brie y, bacterial donor and recipient strains at mid-log phase (OD 600 = 0.5 ~ 0.6) were mixed at a ratio of 1:3 (v/v). Bacterial mixture was spotted on nitrocellulose membranes that were pre-plated on LB agars. An incubation at 37°C for 12 h was given to each of the plates, and bacteria on the membrane were washed using LB broth followed by being shaken at 37°C for 4 h. Transconjugants were selected on LB agar plates with rifampin (1000 mg/L) plus imipenem (20 mg/L) [to screen carbapenem-resistant transconjugants], or rifampin (1000 mg/L) plus colistin (2 mg/L) [to screen colistin-resistant transconjugants], or rifampin (1000 mg/L) plus tigecycline (4 mg/L) [to screen tigecyclineresistant transconjugants]. Antimicrobial susceptibility of the transconjugants was determined using broth microdilution method as mentioned above.

Declarations Supporting Information
All authors declare no competing interests.  territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.    SNP differences between E. coli isolates from humans and/or pig farms. Five pairs of farm-originated E. coli isolates and human isolates that shared less than 10 SNPs are displayed. Numbers of SNPs between the farm originated XDR-E. coli isolates and the human isolates are also shown in a column chart.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.