Occurrence and Molecular Characteristics of Mcr-1-Positive Escherichia coli from Healthy Meat Ducks in Shandong Province of China

Simple Summary Colistin has been used as a growth promotant in livestock feed for many years. To date, there are few reports about the prevalence and molecular characteristics of fecal Escherichia coli bearing mcr-1 in the meat ducks. In this study, among 120 fecal Escherichia coli strains isolated from healthy meat ducks, a total of nine mcr-1-containing E. coli strains were identified and two were identified as extra-intestinal pathogenic E. coli. The 9 mcr-1-bearing E. coli isolates were clonally unrelated, carried two different genetic contexts of mcr-1, and the colistin-resistant phenotype of them was successfully transferred to the recipient strains. These results highlight that healthy meat duck is a potential reservoir for multidrug resistant mcr-1-containing E. coli strains. Abstract Colistin has been used as a growth promotant in livestock feed for many years. In China, mcr-1-positive Escherichia coli strains have been isolated from humans, chickens, and pigs. To date, there are few reports about the prevalence and molecular characteristics of fecal E. coli bearing mcr-1 in the meat ducks. In this study, the prevalence of mcr-1 gene was investigated among 120 fecal E. coli strains isolated from healthy meat ducks in Shandong province of China between October 2017 and February 2018. A total of nine mcr-1-containing E. coli strains were identified and two were identified as extra-intestinal pathogenic E. coli (ExPEC) among them. The clonal relationship of the nine E. coli strains was determined by multilocus sequencing typing (MLST) and pulsed field gel electrophoresis (PFGE), and the results indicated that all mcr-1-carrying isolates were clonally unrelated. Two different genetic contexts of mcr-1 were identified among these isolates. Colistin-resistant phenotype of all the isolates was successfully transferred to the recipient strains by conjugation experiments and seven transconjugants carried a single plasmid. The mcr-1 was located on three replicon plasmids: IncI2 (n = 4), IncFII (n = 2) and IncN (n = 1). Complete sequence analysis of a representative plasmid pTA9 revealed that it was strikingly similar with plasmid pMCR1-IncI2 of E. coli, plasmid pHNSHP45 of E. coli, and plasmid pWF-5-19C of Cronobacter sakazakii, implying that pTA9-like plasmids may be epidemic plasmids that mediate the spread of mcr-1 among Enterobacteriaceae. These results highlight that healthy meat duck is a potential reservoir for multidrug resistant mcr-1-containing E. coli strains.


Introduction
Avian pathogenic Escherichia coli (APEC), a subgroup of extra-intestinal pathogenic E. coli (ExPEC), can cause severe disease characterized by perihepatitis, pericarditis, and airsacculitis, which results in economic and welfare costs in the poultry industry worldwide [1]. There are similar virulence genes between APEC strains and the ExPEC strains in humans [2]. Via the food chain, the multidrug resistant (MDR) APEC strains can transfer from poultry to man, which not only increases the difficulty of treating animal diseases, but also poses a serious threat to human health [3].
As a polymyxin antibacterial agent, colistin is considered as the last-resort drug with excellent bactericidal activity against multidrug-resistant Gram-negative pathogens in humans [4]. However, the recent emergence of mcr-like genes (mcr-1 to mcr-10) potentially threatens the clinical effectiveness of colistin [5][6][7]. These mcr genes have been disseminated to more than 40 countries across at least five continents in multiple ecosystems and traced to more than 11 bacterial species [8,9]. The worldwide distribution of mcr-1 gene strongly indicates a potential food-chain-based spread route [10]. Many studies showed that the prevalent dissemination of the mcr-1 gene relied on transfer by conjugative plasmids such as pHNSHP45, pECJS-B65-33, and pECJS-61-63 [8,9,11].
The intestinal flora of the food animals and humans is a reservoir for antibiotic resistance genes, and the resistant genes can spread from food animals to humans by commensal flora [12,13]. In China, mcr-1-positive E. coli strains have been isolated from humans, chickens, and pigs [14]. To date, prevalence and molecular characteristics of many viral and bacterial pathogens has been identified in Chinese duck flocks [15][16][17][18][19][20][21], but there are few reports about the prevalence and molecular characteristics of fecal E. coli bearing mcr-1 from the meat ducks [22][23][24]. In this study, we isolated E. coli strains from the feces of healthy meat ducks in Shandong province of China, and investigated the occurrence and molecular characteristics of the mcr-1-positive E. coli strains.

Bacterial Isolate
From October 2017 to February 2018, a total of 120 cloacal swabs were collected from healthy meat ducks from 12 duck farms in Shandong province, China. The cloacal swabs were immediately put into Luria-Bertani (LB) broth and incubated for 24 h at 37 • C. All samples were seeded on selective MacConkey agar plates. Bright pink, round, and smooth surface E. coli colonies were picked on selective plates for further analysis. The E. coli isolates were identified through 16S rDNA sequence analysis, and the 16S rDNA primers were designed in this study (Table 1).

Antimicrobial Susceptibility Testing
The minimum inhibitory concentrations (MICs) of tetracycline, fosfomycin, colistin, gentamicin, imipenem, ciprofloxacin, cefotaxime, amikacin, and florfenicol for the E. coli isolates picked on the plates and transconjugants were tested by the broth dilution method and interpreted according to the Clinical and Laboratory Standards Institute [25,26]. The colistin breakpoint (≥2 µg/mL) was used according to the European Committee on Antimicrobial Susceptibility Testing guidelines [27]. E. coli ATCC 25,922 was used as the quality-control strain.

Molecular Typing
XbaI-PFGE was performed as described previously [33] using the CHEF-MAPPER System (Bio-Rad Laboratories, Hercules, CA, USA). Phylogenetic analysis of PFGE patterns was performed using the PyElph software version 1.4 [34]. The UPGMA method was used for clustering. Mcr-1-positive strains were studied by multilocus sequence typing (MLST) as previously described [35]. Phylogenetic classification was performed using a triplex PCR reaction [36].

Conjugation Assays
Conjugation experiments were performed using azide resistant E. coli J53 as the recipient [37]. Transconjugants were selected on agar containing 200 mg/L azide and 2 mg/L colistin and confirmed by enterobacterial repetitive intergenic consensus (ERIC)-PCR method [38].

Plasmid Characterization
Mcr-1-containing plasmids were sized by the S1 nuclease pulsed field gel electrophoresis (S1-PFGE) [33]. A single plasmid carried by transconjugants was used for plasmid analysis. The replicon types of plasmids were determined by PCR-based replicon typing (PBRT) [39]. A representative mcr-1-harboring plasmid, pTA9, was extracted using the Qiagen Large Construct kit (Qiagen, Hilden, Germany) and sequenced using the Illumina MisSeq system using prepared paired-end 2 × 300 bp libraries. The coverage of the plasmid is 200×. Raw data was assembled using the SPAdes Genome Assembler (http://cab.spbu.ru/software/spades/) and SSPACE (version 3.0). Gap was closed with PCR and Sanger sequencing. The plasmid was annotated using the RAST tool (http://rast.nmpdr.org/).

Ethics Statement
All animal experiments were carried out in accordance with guidelines issued by the Shandong Agricultural University Animal Care and Use Committee (approval number, SDAUA-2017-043).

Identification of Mcr-1-Carrying E. coli Isolates
In this study, a total of 120 fecal E. coli strains were isolated from healthy meat ducks from October 2017 to February 2018. Among them, only nine isolates (7.5%, 9/120) were resistant to colistin and identified as positive for mcr-1 gene by PCR amplification and sequencing. In China, high mcr-1 gene carriage rates (about 15% to 30%) were observed in E. coli isolates collected from poultry and pigs between 2011 to 2016 [14,40,41]. Colistin had been commonly used as a growth promotant in livestock feed for many years and had been banned from April 2017 in China. However, the samples in the above-mentioned studies were collected before the ban was issued [14,40,41]. The samples in this study were collected after the ban was issued. So, we speculated that the ban of colistin in animal feed might be the main reason why the low frequency of mcr-1 gene was found in fecal E. coli isolates in this study.

Antimicrobial Resistance Patterns and Resistance Genes
In this study, all of the 9 mcr-1-bearing E. coli isolates were MDR strains (resistance to antibiotics of at least three classes). Among them, 9, 8, 8, and 7 isolates were resistant to tetracycline, cefotaxime, ciprofloxacin, and florfenicol respectively, but all were susceptible to imipenem ( Table 2). Mcr-1 is usually found to coexist with other resistance genes (extended-spectrum β-lactam, floR, and tet(A)) in bacteria [42][43][44]. In this study, 6, 5, 5, and 2 of the nine mcr-1-bearing E. coli isolates harbored floR, bla CTX-M , bla TEM-1, and tet(A) genes, respectively ( Table 2). The association with other resistance genes is likely to favor the dissemination of mcr-1 by co-selection, since cephalosporins, florfenicol, and tetracycline are used extensively in animal husbandry in China. Table 2. Molecular characteristics of the 9 mcr-1-positive E. coli strains isolated from healthy meat ducks in this study.

Phylogenetic Groups and Virulence Genes
All of the nine mcr-1-bearing E. coli isolates contained virulence genes, and the iutA (aerobactin acquisition) gene was identified in 6 ones ( Table 2). Two of the nine E. coli isolates, namely TA9 and TA103 carrying both iutA and papC genes were identified as ExPEC according to the standard [32] ( Table 2). The presence of mcr-1-harboring ExPEC isolates in healthy meat ducks posed a serious health threat to consumers. Fortunately, no virulence gene was co-transferred with mcr-1 gene to the recipient (Table 3). To the best of our knowledge, this is the first report about mcr-1-positive ExPEC isolates identified from healthy meat animals. Phylogenetic group analysis revealed that seven (77.8%) of the nine mcr-1-bearing E. coli isolates belonged to group A and the other two isolates were classed into group D and B1, respectively (Table 2). Similar results were found in the fecal E. coli isolates from chickens in Australia, which were classed into group A, D, B1, and B2, and group A was dominant [45]. The two ExPEC isolates (TA9 and TA103) respectively belonged to groups A and D, which was similar to the result that ExPEC isolates from retail chicken meat products and eggs belonged mainly to group A and D [46].

Molecular Typing
Based on XbaI-PFGE analysis, we found that the nine mcr-1-bearing E. coli isolates were highly diverse ( Figure 1). These data suggested that the spread of mcr-1 gene among E. coli isolates was not due to clonally expansion. MLST analysis result showed that the nine mcr-1-bearing E. coli isolates belonged to nine STs: ST457, ST69, ST2973, ST469, ST10, ST354, ST3170, ST345, and ST410 (Table 2), which also revealed the high genetic diversity among the nine mcr-1-bearing E. coli isolates. As the most common mcr-1-containing E. coli, ST10 was often found in China [47,48]. The E. coli ST410 was widely disseminated in the environment, food animals, humans, and wildlife [49]. The high genetic diversity of the mcr-1-bearing E. coli isolates in this study indicates that the molecular type of E. coli isolates from healthy meat ducks is very complicated.
Animals 2020, 10, x 5 of 11 diversity of the mcr-1-bearing E. coli isolates in this study indicates that the molecular type of E. coli isolates from healthy meat ducks is very complicated.

Genetic Environment of Mcr-1 Gene
Two different genetic contexts of mcr-1 (0 or 1 copy of ISApl1 was present beside mcr-1) were identified among the nine mcr-1 positive E. coli strains (Figure 2 and Table 3). The type I genetic context of mcr-1 (one copy of ISApl1 was present beside mcr-1) was identified in seven mcr-1containing E. coli isolates. The type II genetic context of mcr-1 (ISApI1 was absent) was found in two mcr-1-bearing E. coli strains. All mcr-1 positive E. coli strains included the conserved mcr-1-pap2 segment, which might be horizontally transferred into various plasmids [50]. An ISApl1 element was located upstream of the mcr-1 gene on seven mcr-1-positive isolates. The absence of ISApl1 in mcr-1bearing plasmids could be explained by the mobilization of an ISApl1 composite transposon to conjugative plasmids, which subsequently lost ISApl1 copies [51].

Genetic Environment of Mcr-1 Gene
Two different genetic contexts of mcr-1 (0 or 1 copy of ISApl1 was present beside mcr-1) were identified among the nine mcr-1 positive E. coli strains (Figure 2 and Table 3). The type I genetic context of mcr-1 (one copy of ISApl1 was present beside mcr-1) was identified in seven mcr-1-containing E. coli isolates. The type II genetic context of mcr-1 (ISApI1 was absent) was found in two mcr-1-bearing E. coli strains. All mcr-1 positive E. coli strains included the conserved mcr-1-pap2 segment, which might be horizontally transferred into various plasmids [50]. An ISApl1 element was located upstream of the mcr-1 gene on seven mcr-1-positive isolates. The absence of ISApl1 in mcr-1-bearing plasmids could be explained by the mobilization of an ISApl1 composite transposon to conjugative plasmids, which subsequently lost ISApl1 copies [51].
containing E. coli isolates. The type II genetic context of mcr-1 (ISApI1 was absent) was found in two mcr-1-bearing E. coli strains. All mcr-1 positive E. coli strains included the conserved mcr-1-pap2 segment, which might be horizontally transferred into various plasmids [50]. An ISApl1 element was located upstream of the mcr-1 gene on seven mcr-1-positive isolates. The absence of ISApl1 in mcr-1bearing plasmids could be explained by the mobilization of an ISApl1 composite transposon to conjugative plasmids, which subsequently lost ISApl1 copies [51].

Plasmids Analysis
Conjugation experiments and ERIC-PCR analysis results showed that the colistin-resistant phenotype was successfully transferred from donors to azide-resistant E. coli J53 at conjugation frequencies 1.13 × 10 −2 -4.35 × 10 −7 (transconjugants/recipients) ( Table 3). The mcr-1 gene was identified in 9 transconjugants. S1-PFGE analysis showed that seven transconjugants carried a single plasmid used for plasmid analysis (Figure 3). Transconjugant harbored a single mcr-1-associated plasmid, which ranged in size between 65 and 102 kb and was assigned to IncI2 (n = 4), IncFII (n = 2) and IncN (n = 1) replicon types (Table 3), which have been reported by recent studies to be associated with mcr-1 [14,52,53]. Resistant gene bla CTX-M-55 was co-transferred with mcr-1 on pTA59 plasmid, while no other resistant gene was found to coexist with mcr-1 on the other six plasmids. In this study, two IncI2 plasmids were obtained from the same farm, whereas the other five plasmids were respectively recovered from different farms. As a common mcr-disseminator, IncI2 plasmid was identified in isolates from animals, vegetables, and humans [49,54,55]. These results suggest that diversified conjugative plasmids, especially IncI2 plasmid, may be the key vectors that mediate the dissemination of the mcr-1 among Enterobacteriaceae [56].

Plasmids Analysis
Conjugation experiments and ERIC-PCR analysis results showed that the colistin-resistant phenotype was successfully transferred from donors to azide-resistant E. coli J53 at conjugation frequencies 1.13 × 10 −2 -4.35 × 10 −7 (transconjugants/recipients) ( Table 3). The mcr-1 gene was identified in 9 transconjugants. S1-PFGE analysis showed that seven transconjugants carried a single plasmid used for plasmid analysis (Figure 3). Transconjugant harbored a single mcr-1-associated plasmid, which ranged in size between 65 and 102 kb and was assigned to IncI2 (n = 4), IncFII (n = 2) and IncN (n = 1) replicon types (Table 3), which have been reported by recent studies to be associated with mcr-1 [14,52,53]. Resistant gene blaCTX-M-55 was co-transferred with mcr-1 on pTA59 plasmid, while no other resistant gene was found to coexist with mcr-1 on the other six plasmids. In this study, two IncI2 plasmids were obtained from the same farm, whereas the other five plasmids were respectively recovered from different farms. As a common mcr-disseminator, IncI2 plasmid was identified in isolates from animals, vegetables, and humans [49,54,55]. These results suggest that diversified conjugative plasmids, especially IncI2 plasmid, may be the key vectors that mediate the dissemination of the mcr-1 among Enterobacteriaceae [56].  The nucleotide sequence of plasmid pTA9 from strain TA9 has been deposited in GenBank with accession number MN106912. The plasmid size of pTA9 was 66.603 kb, whose GC% was 41.3%, encoding 72 ORFs (Figure 4). The plasmid pTA9 featured an IncI2 plasmid backbone encoding plasmid transfer, stability, and replication. Two conjugative genes (pil and tra) were predicted on pTA9, which were responsible for the transfer of plasmid between intra-and interspecies bacteria. BLASTn analysis showed that pTA9 was highly similar (the query coverage of 85-97% and the identities 99%) with other mcr-1-bearing plasmids, such as pMCR1-IncI2 of E. coli (isolated from human in Jiangsu province of China, KU761326.1) [50], pWF-5-19C of Cronobacter sakazakii (isolated from chicken in Shandong province of China, KX505142.1) [57], and the first identified mcr-1-bearing plasmid pHNSHP45 of E. coli (isolated from pig in Shanghai of China, KX505142.1) [14] (Figure 5). TnpA and tnpB were identified in pTA9, pMCR1-IncI2, and pWF-5-19C. In addition, ISApl1 was identified in pTA9, pWF-5-19C, and pHNSHP45. An mcr-1-pap2 element was identified in pTA9 and pMCR1-IncI2. This suggests that pTA9-like plasmids may be epidemic plasmids that mediate mcr-1 dissemination between distinct host bacteria in China.
Animals 2020, 10, x 7 of 11 The nucleotide sequence of plasmid pTA9 from strain TA9 has been deposited in GenBank with accession number MN106912. The plasmid size of pTA9 was 66.603 kb, whose GC% was 41.3%, encoding 72 ORFs (Figure 4). The plasmid pTA9 featured an IncI2 plasmid backbone encoding plasmid transfer, stability, and replication. Two conjugative genes (pil and tra) were predicted on pTA9, which were responsible for the transfer of plasmid between intra-and interspecies bacteria. BLASTn analysis showed that pTA9 was highly similar (the query coverage of 85-97% and the identities 99%) with other mcr-1-bearing plasmids, such as pMCR1-IncI2 of E. coli (isolated from human in Jiangsu province of China, KU761326.1) [50], pWF-5-19C of Cronobacter sakazakii (isolated from chicken in Shandong province of China, KX505142.1) [57], and the first identified mcr-1-bearing plasmid pHNSHP45 of E. coli (isolated from pig in Shanghai of China, KX505142.1) [14] (Figure 5). TnpA and tnpB were identified in pTA9, pMCR1-IncI2, and pWF-5-19C. In addition, ISApl1 was identified in pTA9, pWF-5-19C, and pHNSHP45. An mcr-1-pap2 element was identified in pTA9 and pMCR1-IncI2. This suggests that pTA9-like plasmids may be epidemic plasmids that mediate mcr-1 dissemination between distinct host bacteria in China.   The nucleotide sequence of plasmid pTA9 from strain TA9 has been deposited in GenBank with accession number MN106912. The plasmid size of pTA9 was 66.603 kb, whose GC% was 41.3%, encoding 72 ORFs (Figure 4). The plasmid pTA9 featured an IncI2 plasmid backbone encoding plasmid transfer, stability, and replication. Two conjugative genes (pil and tra) were predicted on pTA9, which were responsible for the transfer of plasmid between intra-and interspecies bacteria. BLASTn analysis showed that pTA9 was highly similar (the query coverage of 85-97% and the identities 99%) with other mcr-1-bearing plasmids, such as pMCR1-IncI2 of E. coli (isolated from human in Jiangsu province of China, KU761326.1) [50], pWF-5-19C of Cronobacter sakazakii (isolated from chicken in Shandong province of China, KX505142.1) [57], and the first identified mcr-1-bearing plasmid pHNSHP45 of E. coli (isolated from pig in Shanghai of China, KX505142.1) [14] (Figure 5). TnpA and tnpB were identified in pTA9, pMCR1-IncI2, and pWF-5-19C. In addition, ISApl1 was identified in pTA9, pWF-5-19C, and pHNSHP45. An mcr-1-pap2 element was identified in pTA9 and pMCR1-IncI2. This suggests that pTA9-like plasmids may be epidemic plasmids that mediate mcr-1 dissemination between distinct host bacteria in China.   In this study, pTA9 could be transferred to E. coli J53 isolates in vitro. This suggests that the mcr-1 gene present in gut flora of meat duck can be horizontally transferred by bacterial conjugation among distinct bacterial hosts. Similar scenarios have already been observed in the human intestinal Animals 2020, 10, 1299 8 of 11 flora [58,59]. So mcr-1-bearing fecal E. coli in healthy meat ducks could be a source for the transfer of mcr-1 through contaminated food to humans.

Conclusions
This study revealed the carriage rate of mcr-1 among fecal E. coli isolates obtained from healthy meat ducks in China. PFGE and MLST results indicated that mcr-1-bearing E. coli isolates were clonally unrelated. This suggested that the horizontal transfer of plasmids was the main mechanism for the dissemination of mcr-1 gene in meat duck farms. The pTA9-like plasmids have been isolated from different bacterial hosts across distinct regions of China, implying that pTA9-like plasmids are likely to be the epidemic mcr-1-bearing plasmids that mediate the dissemination of mcr-1 in China. Since China is the biggest exporter of meat duck products in the world, the spread of pTA9-like conjugative plasmids across other regions and countries should attract attention. In addition, the mcr-1-bearing E. coli usually carry bla CTX-M and floR, conferring resistance to cephalosporins and florfenicol, which made coselection possible when these drugs were used. Restrictive/rational use of antibiotics in animal husbandry, especially in food animals in China may help to limit the spread of mcr-1 gene.

Conflicts of Interest:
The authors declare no conflict of interest.