Comparative Genome Analysis of Livestock and Human Colistin-Resistant Escherichia coli Isolates from the Same Household

Introduction

Colistin, one of the few treatment options for intractable infectious diseases caused by multidrug-resistant Gram-negative bacteria,¹ is a popular drug in livestock breeding; it is not only used to treat infections but also as a livestock growth promoter and infection protective agent.² The overuse and misuse of colistin is considered responsible for the emergence and dissemination of colistin-resistant (COR) bacteria in animal husbandry. In addition, since colistin resistance can be transferred by a mobile colistin resistance gene, mcr, wide dissemination of COR bacteria can be expected in livestock as well as humans. In fact, livestock is considered a reservoir for COR bacteria.³

Our recent studies have reported wide dissemination of COR bacteria in human residents of communities where livestock carried COR bacteria.^4–6 The results of these studies suggest that horizontal transmission of resistance may occur between livestock and human residents in the same community, including transmission of specific resistant isolates. However, there is a lack of direct evidence supporting the transmission of COR bacteria and/or the resistance gene, mcr, in the community.

A recent study on mcr-1-harboring plasmids obtained from COR Escherichia coli isolates from humans and livestock has reported the details of mcr-1-carrying IncI2 plasmid,⁷ which is one of the major replicons for the mcr-1-carrying plasmids.⁸ However, these studies analyzed clinical isolates from patients and livestock isolates retrospectively. In this study, we aimed to study the transmission of mcr-1 between COR E. coli isolates from humans and domestic animals in the same household by performing a comparative genomic analysis of these isolates.

Materials and Methods

Bacterial Strains

COR E. coli isolates, strains 2018–18.25PC and 2018–18.25CC, were previously isolated from rectal swab specimens of pig and chicken, respectively, in the backyard of a household in a rural community in Vietnam.⁴ Human COR E. coli isolates, 2017.02.01CC, 2017.02.02CC, and 2018–02.2CC, were also previously isolated from the fecal specimens of two family members of the same household as the livestock.⁵ The study was approved by the Ethics Committees of Osaka University (Osaka, Japan) and Thai Binh University of Medicine and Pharmacy (Thai Binh, Vietnam). The characteristics of these isolates are summarized in Table 1.

Table 1 Characterization of Colistin-Resistant E. coli Isolates from Human and Livestock

Pulsed-Field Gel Electrophoresis (PFGE)

Xbal-digested genomic DNA samples from the COR E. coli isolates were analyzed on a CHEF-DR III System (Bio-Rad, Hercules, CA, USA) following the methods reported previously.⁹

Genome Sequencing

In this study, we analyzed the strains 2018–18.25PC and 2018–18.25CC obtained from a pig and a chicken, respectively, by genome sequencing. Whole-genome sequencing and assembly of the isolates as well as of the plasmid harboring mcr-1, were performed using the Illumina MiSeq (Illumina Inc., CA, USA) and MinION (Oxford Nanopore Technologies, London, UK) sequencers, as described previously.¹⁰ The genome sequences were automatically annotated using the DDBJ Fast Annotation and Submission Tool Pipeline (https://dfast.nig.ac.jp). Genome analysis was performed using the Geneious R11 software (Biomatters, Ltd., NZ). Antibiotic resistance genes and replicon typing were conducted in ResFinder 4.0¹¹ and PlasmidFinder 1.3 on the Center for Genomic Epidemiology website (http://www.genomicepidemiology.org). Multilocus sequence type (MLST) analysis was performed using a MLST software (https://github.com/tseemann/mlst). The mcr-1-bearing IncHI2 plasmids were compared with the reference plasmid pECJS-59-244¹¹ using BRIG¹² and visualized with Easyfig.¹³

Genome analysis of the human isolates was performed previously, and the analyzed data were used in this study.¹⁴

Results

Comparison of Livestock- and Human-Derived Strains Isolated from the Same Household

The phylogenetic relation, mcr-1 location, and mcr transposon structure were assessed for five isolates. Two isolates were obtained from a chicken and a pig each, which were maintained in the backyard of the household; one isolate was obtained from a human member of the same household at the time of the isolation of the livestock strains, and two isolates were obtained from the human members of the household 10 months prior to the isolation of the livestock strains. The phylogenetic likelihood of these five isolates was assessed using PFGE, and the results are shown in Figure 1. None of the five isolates belonged to the same clone. However, the pig isolate 2018–18.25PC and the human isolate 2018–02.2CC belonged to sequence type (ST) 10, and the chicken isolate 2018–18.25CC and the human isolate 2017.02.01CC to ST48.

Figure 1 Dendrogram of pulsed-field gel patterns of isolates from human and livestock.

Comparative genomic analysis was performed to assess the genetic details of mcr-1 in these isolates. Similar to the human isolates, the livestock isolates carried mcr-1 on either the chromosome or a plasmid (Table 1). The chicken isolate 2018–18.25CC harbored mcr-1 on plasmid IncHI2 as observed in the human isolate 2017.02.01CC. Although both isolates belonged to ST48, phylogenetically, they were different. The pig isolate 2018–18.25PC belonged to ST10 and harbored mcr-1 on the chromosome similar to the human isolate 2018–02.2CC; however, these two isolates were different (Figure 1). The human isolate 2017.02.02CC was also different from the livestock isolates.

Comparative Details of mcr-1 in Human and Livestock Isolates

The results of the comparative analysis of the mcr gene showed that both the human strain 2018–02-2CC and the pig strain 2018–18.25PC retained the mcr-1 transposon on the chromosome; however, the human strain had an incomplete transposon, ISApl1-mcr-1-PAP2, whereas the pig strain possessed the complete transposon, ISApl1-mcr-1-PAP2-ISApl1 (Table 1). Furthermore, the location of mcr-1 on both chromosomes was genetically related despite the strains being different.

Sequence of mcr-1-plasmid IncHI2 found in the human strain 2017.02.01CC and the chicken strain 2018–18.25CC showed that both plasmids (p2017.02.01CC, and p2018-18.25CC) were similar (query cover; 88%, identity; 99.99%, Figure 2) except for some drug resistance genes in the multidrug resistance (MDR) region of p2018-18.25CC, which were inserted in the reverse direction (Figure 3). IncHI2 plasmid carries a large MDR region containing antibiotic resistance genes other than mcr-1.¹⁵ Therefore, a comparative study with the plasmids from the isolated strains and the reference plasmid pECJS-59-244 (KX084394) was conducted. As shown in Figure 3, plasmids from the human and chicken isolates were similar in terms of plasmid organization to the reference plasmid. However, the MDR region of the human and chicken isolates lacked certain resistance genes present in the reference plasmid. Interestingly, IncHI2 isolated from the human and chicken isolates harbored floR, a gene coding for resistance against the veterinary drug florfenicol; this was similar to the reference IncHI2 (KX084394).

Figure 2 Sequence alignment of IncHI2 plasmids with mcr-1 in isolates obtained from a human and a chicken of the same household. Plasmids 2018–18.25CC and p2017.02.01CC were present in the chicken and human isolates, respectively. The reference IncHI2 plasmid pECJS-59-244 was isolated from a pig in a previous study. The circle indicates the putative MDR region.

Figure 3 Sequence alignment of the putative MDR region in plasmids. Plasmids 2018–18.25CC and p2017.02.01CC were present in the chicken and human isolates, respectively. The reference IncHI2 plasmid pECJS-59-244 was isolated from a pig in a previous study.

Discussion

The results of this study showed that human members of the household harbored COR E. coli strains with the same ST as that of the livestock. In particular, the COR E. coli ST48 strains were isolated from chicken as well as humans. These isolates were genetically similar but were different clones, as assessed by PFGE analysis. This indicates that these isolates were derived from the same clone and evolved by long-term adaptation to the hosts (humans and livestock) as well as environmental conditions. Furthermore, since this study was not conducted during a short-term epidemic of COR bacteria spreading in the community, it was highly unlikely to observe the same clone in both humans and livestock. Therefore, a clonal distribution of COR strains among livestock and humans was anticipated. In addition, identification of the same IncHI2 plasmid carrying mcr-1 in these isolates might provide the initial evidence for COR bacteria dissemination and subsequent spread of plasmids in the household.

The phylogenetic similarity of human and pig ST10 COR isolates was low. The COR E. coli ST10 isolate from the pig did not possess an mcr-1-carrying plasmid but carried a mcr-1 on the chromosome akin to the human ST10 isolate. However, unlike the human ST10 isolate, the structure of the mcr-1 transposon found in the pig isolate was well conserved. Since the insertion sequences (ISs) of the mcr-1 transposon TN6330 (ISApl1-mcr-1-PAP2-ISApl1) have dropped out during stabilization of the genes across generations,^16,17 it is likely that transmission of TN6330 between plasmids, chromosomes, and strains may have occurred in the household over a long period. Therefore, the ST10 clone may have spread between humans and livestock like the ST48 clone. Furthermore, the presence of floR on plasmids of both livestock and human isolates indicates the direction of this transmission. The occurrence of floR is commonly associated with mcr-1 and suggests animal origin of the isolate.^8,17 However, more isolates need to be assessed to confirm this observation.

Since the discovery of mcr on plasmids, many studies have been conducted on mcr-carrying plasmids of COR bacteria isolated from various sources.^7,8,18,19 Outcomes of a metadata analysis revealed that the mcr-carrying plasmids are diverse.¹⁹ Among these mcr-carrying plasmids, IncI2, IncX4, and IncHI2 plasmids are found most frequently, even in livestock isolates. The outcomes of the present study regarding mcr-1-harboring plasmid IncHI2 of the livestock COR isolate are consistent with those of previous studies.¹⁵

The IncHI2 plasmid is also known to actively acquire other foreign resistance genes.¹⁵ BLASTN and ResFinder analysis on the resistance gene database showed that the plasmids of the chicken isolate 2018–18.25CC and the human isolate 2017.02.01CC comprised MDR determinants, including floR, bla_TEM, tetM, aaC3, dfrA5, and sul1. However, genes such as bla_CTX-M, fosA, and aadA, reported in the reference pig isolate obtained from farms in China, were not present in our isolates.¹⁵ Addition of MDR genes to mcr-carrying plasmids and the subsequent spread of the plasmid in the microbiomes of humans and livestock poses a threat to public health worldwide. The results of the present study showed that accumulation of resistance genes was limited in the isolates. However, accumulation of resistance genes can be expected in the future. Therefore, monitoring of MDR genes besides mcr in the plasmids of human and livestock microbiomes is urgently required.

The chromosomally encoded mcr-1 is important due to its stability.^13,20 The presence of chromosomal mcr-1 has been reported, but its frequency was low (3.5%).²⁰ In this study, although only two isolates from livestock were assessed genetically, the finding of chromosomal mcr-1 in both of them is consistent with a previous study reporting the occurrence of chromosomal mcr-1 in 36.8% of the isolates from healthy residents living in the same area.¹³ Therefore, the monitoring of chromosomal mcr in COR bacteria it is also important.

Conclusion

To the best of our knowledge, this is the first report of a direct comparative analysis of COR E. coli isolates from livestock and humans of the same household. The initial findings of this study suggest distribution of COR E. coli and subsequent transmission of the colistin resistance gene mcr-1 transposon between livestock and humans of the same household. The findings also support that dissemination of COR bacteria in communities occurs through domestic livestock. The study warrants further genetic analysis of COR bacteria isolates in communities where COR bacteria are disseminated.