High Prevalence of Colistin-Resistant Escherichia coli with Chromosomally Carried mcr-1 in Healthy Residents in Vietnam.

Elucidation of the mechanism of the wide dissemination of colistin-resistant bacteria in communities of developing countries is an urgent public health issue. In this study, we investigated the genetic background of the colistin resistance gene mcr in E. coli isolates from the fecal microbiota of healthy human residents living in a community in Vietnam with a high prevalence of colistin-resistant E. coli. Our study revealed for the first time, a surprisingly high percentage (36.8%) of colistin-resistant E. coli carrying chromosomal mcr-1, the emergence of which may have occurred recently, in the fecal microbiota of the community residents. The mcr-1 transposon on the chromosome may develop into a more stable genotype by the loss of insertion sequences (ISs). Our results are valuable in understanding the mechanism underlying the increasing prevalence of colistin-resistant bacteria within a community.

C olistin is recognized as the last-resort antibiotic for the treatment of infectious diseases caused by multidrug-resistant (MDR) Gram-negative bacteria, including carbapenem-resistant bacteria. However, the wide distribution of colistin-resistant (COR) bacteria threatens the effectiveness of treatments with colistin (1).
Current reports of a wide distribution of COR Escherichia coli with a mobile resistance gene, mcr, in a community highlight the importance of the stability and possible transfer of the resistance gene to the pathogen (2). In contrast to outbreaks in nosocomial settings, the wide distribution of COR bacteria that may be occurring over a long period in the community seems to be attributable to exposure to a persistent, low-concentration antibiotic pressure. This hypothesis is supported by the frequent use of colistin-containing feed in livestock (3) and the retention of COR E. coli in the fecal microbiota of livestock (4). However, in a previous study, we showed that the wide distribution of COR E. coli in residents was not due to the clonal distribution of a certain lineage of COR bacteria (2). Therefore, the transfer of the mcr-carrying plasmid is the most likely explanation for the wide distribution of the gene. In this regard, recent reports revealed that mcr-1 could be mobilized as an ISApl1-flanked composite transposon, Tn6330 (5,6). Therefore, the structure of the ISApl1 transposon with mcr potentially reflects the transposition of mcr in COR bacterial isolates in the community. However, further investigations are needed to confirm this hypothesis.
The long-term stability of COR bacteria in the community is another factor that affects their distribution. Chromosomal mcr may play an important role in their stability, but this has not been definitively established. To further understand the distribution mechanisms, in the present study, we investigated the mcr location, as well as the mcr transposon structure, of COR E. coli isolates obtained from healthy subjects in Vietnam residing in a community where COR bacteria are frequently detected.

RESULTS AND DISCUSSION
Location of mcr-1 in COR E. coli isolates. The location of mcr-1 in 57 COR E. coli isolates that were originally obtained from 98 asymptomatic healthy residents of a rural community in Vietnam during a previous study (2) was assessed using S1/I-CeuI pulsed-field gel electrophoresis (PFGE) and Southern blot hybridization with mcr-1 probe analysis. All isolates used in this study were phylogenetically diverse, as determined by PFGE and multilocus sequence typing (MLST) analyses (Table 1). In addition, no clonal expansion in the community was observed, as determined in a previous study (2).
Southern blot analysis of the isolates after S1-PFGE (see Fig. S1 in the supplemental material) revealed that 36.8% (21/57) of the isolates carried mcr-1 on their chromosomes, whereas 63.2% (36/57) carried mcr-1 on only the plasmid. (Three of the original 60 mcr-1 ϩ E. coli isolates, from 98 residents, were excluded because of the difficulty in analysis.) In one isolate, mcr-1 was located on both the chromosome and the plasmid.
mcr-1 was originally discovered as a mobile resistance gene on the plasmid of COR bacteria (7), and it is recognized that chromosomally carried mcr-1 is very rare compared to that on the plasmids (8)(9)(10)(11)(12)(13). For instance, Li et al. found that only 4% of 200 mcr-1-positive E. coli isolates from animals, food, the environment, and human fecal samples collected in China carried chromosomal mcr-1 (11). In contrast to previous reports, the results from this study showed that 36.8% of the human fecal sample isolates had chromosomal mcr-1. All isolates examined in this study were obtained from different individuals, except two isolates that were obtained from the same person at a 1-year interval. The prevalence of chromosomal mcr-1 observed in this study is extremely high compared to that reported in previous studies. Although the reasons for the high prevalence of chromosomal mcr-1-carrying isolates in human fecal microbiota are not clear, it seems likely that the high prevalence in the area of sampling could be due to the frequent use of colistin as a livestock antibiotic and feed additive (3). Additionally, most domestic livestock in that community possessed COR E. coli carrying mcr-1 (4). Under the prevailing conditions, it is conceivable that the community from which the isolates were obtained could have experienced a constant colistin pressure on the microbes over a long period, which resulted in the microbes becoming intrinsically resistant due to chromosomally encoded resistance (11). This could have resulted in chromosomal mcr-1 becoming more prevalent and stable in the community.
The chromosomal mcr-1-carrying bacteria may contribute to the emergence of untreatable MDR bacteria when MDR genes, including carbapenem resistance genes on the plasmid (14), are transferred to bacteria possessing chromosomal mcr. In fact, this likelihood is supported by our current study, wherein several extended-spectrum ␤-lactamase (ESBL)-producing E. coli isolates that had MDR with mcr-1 were obtained from fecal microbiota of healthy human residents of the community (15).
Genetic structure of chromosomally and/or plasmid-carried mcr-1 transposon. Recently, it has been proposed that mcr-1 can be mobilized as an ISApl1-flanked composite transposon (Tn6330) (5,16). The study also reported that transmission is possible both upstream and downstream of ISApl1. The structure of this composite transposon is considered to be stabilized by the structure in which insertion sequences (ISs) have dropped out (6,17). Therefore, the structure of mcr-1 transposon Tn6330 in the bacterium is important not only for the transmission among microbes but also for the stability of mcr-1.
To elucidate the mcr-1 transposon structure of the isolates in this study, 14 representative strains (6 of 18 chromosomal mcr-1 isolates, one chromosomal/plasmid mcr-1 isolates, and 7 of 30 plasmid mcr-1 isolates) were subjected to genome analysis. Several strains were obtained from the same household (Table 2). In the case of household 2, two mcr-1 isolates were obtained from the same member at a 1-year interval. All these isolates were phylogenetically different (2). Table 2 shows the bacterial characteristics of chromosomally and/or plasmid-carried mcr-1 transposons for these 14 isolates.
Plasmid-carried mcr-1 transposon. Two isolates obtained from different residents showed different host bacterium MLST types but possessed very similar IncHI2 plasmids with the same mcr-1 transposon, mcr-1-PAP2. Similarly, three other isolates showed different E. coli MLST types but had the same IncX4 plasmid with the mcr-1-PAP2 transposon ( Table 2). These results indicate that the transfer of plasmids among bacteria may occur frequently in the human bacterial flora.
Other plasmids showed different mcr-1 transposons, including ISApl1-mcr-1-PAP2, which was carried by different Inc types, such as IncP1, IncI2, and IncY. These results confirmed that mcr-1 was retained in various Inc-type plasmids in bacteria from fecal microbiota of residents in the community, as previously shown by the diversity of mcr-1-carrying plasmids (9). In addition, the majority (7 of 8) of plasmids that were assessed had the simplest mcr-1 transposon structure, mcr-1-PAP2. One remaining plasmid had the ISApl1-mcr-1-PAP2 transposon. Our results indicate that mcr-1 on the plasmid of COR E. coli strains in the fecal microbiota of humans may be stabilized by the loss of ISApl1 in many cases, as described by Wang et al. (17).
Chromosomally carried mcr-1 transposon Tn6330. The mcr-1 transposon structure in the chromosome of the seven COR E. coli isolates assessed in this study showed that the majority (6 of 7) of the isolates possessed a complete ancestral mcr-1 transposon, Tn6330 (6), without the loss of ISApl1 in their chromosomes (Fig. 1). The Tn6330 insertion sites on the chromosomes of these isolates were random. Besides, the AT-and CG-rich sequences at either end of Tn6330 were not found on the chromosomes of the isolates (data not shown). Because the insertion sites of Tn6330 are notable for AT-and CG-rich regions (6), it is presumable that the absence of such 2-bp target site duplications in these isolates may be caused by mutations after the transposition of Tn6330, even though there was no evidence to support it. The one remaining isolate lost the ISApl1 present downstream of Tn6330, which may contribute to stabilizing the mcr-1 gene (16). The genome analysis of these isolates revealed that there was only one Tn6330 in the chromosome of the isolates tested. In addition, there was no independent mcr-1 without an IS in the chromosomes. Thus, the results from the genome assessment of COR E. coli isolates of fecal microbiota from the community residents support a previous finding that ISApl1 facilitates mcr-1 transmission (11).
The dominant fully intact Tn6330, ISApl1-mcr-1-PAP2-ISApl1, on the chromosome among COR E. coli isolates from fecal microbiota of community residents may indicate that this transposon insertion into the chromosome occurred recently. If the high prevalence of chromosomally carried mcr-1 has occurred recently, it may indicate that the wide distribution of COR bacteria in the community has progressed to a stable state. Furthermore, although it was found in one isolate, the loss of ISApl1 from the mcr-1 transposon Tn6330 indicates that the resistance gene on the chromosome has shifted to a more stable state.
One isolate was found to harbor mcr-1 on both the chromosome and plasmid. Because the origin of the mcr-1 transposon is unclear from the transposon elements, the relationship between mcr-1 on the chromosome and the plasmid in this isolate is unknown. In this regard, the following can be speculated. Because truncated Tn6330 was found in all plasmids assessed in this study, it is likely that COR E. coli isolates with plasmids harboring Tn6330 may have prevailed in the community for a long time. Moreover, it can be presumed that the transposition of Tn6330 from the plasmid to the chromosome occurred, followed by the loss of the IS during the process of stabilizing mcr-1. Therefore, this isolate may be a transitional intermediate-type strain. The finding of the intermediate-type isolate suggests that the process for stabilizing mcr-1 is in progress.
The assessment of mcr-1 location in COR E. coli isolates in this study showed that only one isolate carried it both on the chromosome and on the plasmid, whereas the other isolates carried the mcr-1 either on the chromosome or on the plasmid. It is not clear why the Tn6330 was carried on either the plasmid or the chromosome and only rarely on both. Because large antibiotic-resistant plasmids may be lost during their multiplication in an antibiotic-free environment due to their significant metabolic burden on the host strain, it can be speculated that after the transposition of Tn6330 from the plasmid to the chromosome, the plasmid may no longer need to carry Tn6330 or the plasmid itself may not be needed (18).

MATERIALS AND METHODS
Sample collection for COR E. coli isolates. A total of 57 COR E. coli isolates with mcr-1 were assessed in this study. All the COR E. coli isolates were initially obtained from healthy residents in a rural community in Vietnam between November 2017 and February 2018. One isolate was obtained from each resident. It was found that a high percentage (70.4%) of the residents were carrying COR E. coli with mcr-1 in their stool, as reported in a previous study (2).
Assessment of mcr-1. Digoxigenin (DIG)-labeled DNA probes used for the detection of mcr-1 and 16S rRNA genes in Southern blot hybridization were prepared using the PCR DIG probe synthesis kit (Sigma-Aldrich, St. Louis, MO) with the primers shown in Table S1 in the supplemental material. The location of mcr-1 in the examined isolates was determined by S1 nuclease PFGE and Southern blot hybridization, as previously described (19). For some bacterial isolates, I-CeuI PFGE and Southern blot hybridization were also performed to determine mcr-1 locations according to the methods described in previous studies (20,21).

ACKNOWLEDGMENT
This study was supported by the Japan Society for the Promotion of Science KAKENHI (grant 17H01687).