Whole-Genome Sequencing of Clinically Isolated Carbapenem-Resistant Enterobacterales Harboring mcr Genes in Thailand, 2016–2019

Mobile colistin-resistant genes (mcr) have become an increasing public health concern. Since the first report of mcr-1 in Thailand in 2016, perspective surveillance was conducted to explore the genomic characteristics of clinical carbapenem-resistant Enterobacterales (CRE) isolates harboring mcr in 2016–2019. Thirteen (0.28%) out of 4,516 CRE isolates were found to carry mcr genes, including 69.2% (9/13) of E. coli and 30.8% (4/13) of K. pneumoniae isolates. Individual mcr-1.1 was detected in eight E. coli (61.5%) isolates, whereas the co-occurrence of mcr-1.1 and mcr-3.5 was seen in only one E. coli isolate (7.7%). No CRE were detected carrying mcr-2, mcr-4, or mcr-5 through to mcr-9. Analysis of plasmid replicon types carrying mcr revealed that IncX4 was the most common (61.5%; 8/13), followed by IncI2 (15.4%; 2/13). The minimum inhibitory concentration values for colistin were in the range of 4–16 μg/ml for all CRE isolates harboring mcr, suggesting they have 100% colistin resistance. Clermont phylotyping of nine mcr-harboring carbapenem-resistant E. coli isolates demonstrated phylogroup C was predominant in ST410. In contrast, ST336 belonged to CC17, and the KL type 25 was predominant in carbapenem-resistant K. pneumoniae isolates. This report provides a comprehensive insight into the prevalence of mcr-carrying CRE from patients in Thailand. The information highlights the importance of strengthening official active surveillance efforts to detect, control, and prevent mcr-harboring CRE and the need for rational drug use in all sectors.


INTRODUCTION
The global spread of carbapenem-resistant Enterobacterales (CRE) has become a leading public health concern. The lack of accessible treatment has resulted in the use of colistin, an outmoded antibiotic, as a last-resort therapeutic drug for human infections by gram-negative bacteria. The widespread use of colistin in humans and animals has led to the emergence of colistin resistance in gram-negative bacteria, and the rates of resistance are continuously increasing (Kempf et al., 2016;Elbediwi et al., 2019). In 2017, the World Health Organization acknowledged that CRE are a serious priority in a published list of globally important of 12 antimicrobial-resistant pathogens (World Health Organization, 2017). Of particular concern is the spread of mcr genes into CRE, which would create strains that are potentially pan-drug resistant (PDR).
A common mechanism of colistin resistance is thought to be associated with chromosomal mediation (Meletis and Skoura, 2018). The discovery of the first plasmid-mediated colistin resistance gene mcr-1 in an Escherichia coli strain isolated from a pig in China led to an increasing number of reports on the identification of mcr in many bacterial species worldwide (Liu et al., 2016;Elbediwi et al., 2019). Thus far, 10 variants of mcr (mcr-1 through to mcr-10) have been reported (Gharaibeh and Shatnawi, 2019;Wang et al., 2020). The mcr gene has been shown to encode a phosphoethanolamine transferase that alters lipid A in the lipopolysaccharide of the bacterial outer membrane by adding a phosphoethanolamine (Gharaibeh and Shatnawi, 2019). This reduces the attachment of colistin to the bacterial outer membrane and, therefore, prevents cell lysis.
The presence of mcr has been reported in different gramnegative bacteria isolated worldwide from animal and human sources (Elbediwi et al., 2019;Gharaibeh and Shatnawi, 2019). In Thailand, the first reported mcr-1-harboring E. coli were isolated from the stool of an asymptomatic person in 2012 (Olaitan et al., 2016). There has since been documentation of mcr-1-, mcr-2-, and mcr-3-carrying E. coli and Klebsiella pneumoniae strains isolated from patients (Paveenkittiporn et al., 2017;Srijan et al., 2018;Kamjumphol et al., 2019;Malchione et al., 2019). The current study investigated the genomic characterization of antimicrobial resistance genes and antimicrobial susceptibility of CRE isolated from patients in Thailand during a laboratory-based surveillance program in 2016-2019.

Whole-Genome Sequencing and Analysis
DNA from 13 mcr-carrying CRE isolates was extracted from nutrient agar plate cultures using a DNeasy blood & tissue kit (Qiagen, Hilden, Germany) following the manufacturer's recommendations, and the concentration was determined using the Qubit dsDNA BR assay kit (Invitrogen, Oregon, United States). Sequence libraries were prepared using a Qiagen QIAseq FX DNA library kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol. The prepared libraries were sequenced on the Illumina Miseq platform with Illumina MiSeq 2X 250 base paired-end chemistry (Illumina, CA, United States) according to the manufacturer's instructions. The genomes for each strain were de novo assembled using CLC Genomics Workbench v12.0.2 (Qiagen, Aarhus, Denmark) with mostly default settings, except that the minimum contig size threshold was set to 500 bp.
Plasmid replicons were analyzed using PlasmidFinder (Carattoli et al., 2014) 4 . Phylogrouping for E. coli and the KL type of K. pneumoniae were based on analysis using ClermonTyping (Beghain et al., 2018) (Larsen et al., 2012) 9 . The genomic comparison of 13 mcrharboring CRE isolates was conducted using a modular Single Genome Analysis to search for the genetically closest relatives in the database following the single nucleotide polymorphism (SNP) approach with BacWGSTdb   10 . A phylogenetic tree was constructed using REALPHY and MEGA X via the neighbor-joining method with 500 bootstrap replicates by applying the Tamura three-parameter model (Bertels et al., 2014;Kumar et al., 2018). The tree was visualized and annotated using Interactive Tree of Life (ITOL) V4 (Letunic and Bork, 2016). E. coli K12 substrain MG1655 (accession no. U00096) and K. pneumoniae HS11286 (accession no. CP003200) were used as the reference sequences for SNP analysis. Details of the other genomes used for comparison with our isolates are shown in the Supplementary Materials.

Statistical Analysis
The associations between mcr genes, mcr-harboring E. coli isolates, and non-E. coli isolates were analyzed by calculating the odds ratios (OR) and p-values using the STATA version 14 software package, with p < 0.05 considered to be statistically significant.

Accession Number
The assembled genomic sequences were deposited under the BioProject accession number PRJNA525849 with BioSample accessions: SAMN15497997-SAMN15498009. The accession numbers for each mcr-1-harboring CRE isolate are provided in Table 2.
The proportions of E. coli and K. pneumoniae isolates showing mcr genes were 69.2% (9/13) and 30.8% (4/13), respectively. Individual mcr-1.1 genes were detected in eight E. coli (61.5%) from 13 mcr-carrying CRE isolates, whereas the co-occurrence of mcr-1.1 and mcr-3.5 was found in only one E. coli isolate (7.7%) ( Table 2). However, mcr-2, mcr-4, and mcr-5 through to   Frontiers in Microbiology | www.frontiersin.org mcr-9 were not detected. Analysis of the plasmid replicon types carrying mcr revealed that IncX4 was the most common (61.5%; 8/13), followed by IncI2 (15.4%; 2/13), as shown in Table 2. However, three isolates were unidentified. Six carbapenemresistant E. coli isolates harboring mcr carried IncX4, whereas two isolates contained IncI2. Of the four carbapenem-resistant K. pneumoniae isolates harboring mcr, two carried IncX4 and the other two had unknown plasmid replicon types. The genetic organization of the mcr genes in these 13 isolates is outlined in Figure 2. A common gene found downstream of mcr-1.1 in all isolates encoded the PAP2 family protein, whereas the upstream genes varied. However, a DUF2726-domain-containing proteinencoding gene was commonly found in 8 of the 13 isolates (seven E. coli and one K. pneumoniae). Furthermore, the upstream and downstream genetic organization of mcr-3.5 was quite different from that of mcr-1.1. Table 2, chromosomal-mediated colistin resistance gene mutations, including those in mgrB, pmrAB, phoPQ, crrB, and rpoN, were analyzed. We detected substitutions in pmrAB and rpoN in almost mcr-harboring CRE isolates, whereas phoPQ substitutions were found in three isolates, and no mutations were detected in mgrB. Substitutions were commonly found in the pmrA genes of 12 out of 13 mcrharboring isolates, while pmrB and rpoN substitutions were detected in 11 isolates. The isolates 54881 and 54882 contained more mutations in the chromosomal-mediated colistin resistance genes than other isolates. Substitution at S29G and E57G in pmrA was predominant in E. coli and K. pneumoniae, respectively. The pmrB substitution at D283G was commonly found in E. coli, whilst all four mcr-1-harboring K. pneumoniae contained a T246A substitution. Only one substitution in phoP (I44L) and one in phoQ (L343V) were detected in mcr-1carrying E. coli. Substitution at E150D and I165M in rpoN was predominant in E. coli isolates, in contrast to those at F304Y which was commonly found in K. pneumoniae isolates. In addition, one crrB substitution (L296Q) was found in all K. pneumoniae isolates. Insertion or deletion in those described genes was not detected in all mcr-1harboring CRE isolates.

As shown in
As shown in Table 2, of the 13 isolates, five and three mcr-harboring E. coli isolates carried bla NDM−1 and bla NDM−5 , respectively. Only one E. coli isolate contained bla OXA−48 . Three and one mcr-1.1-harboring K. pneumoniae isolates carried bla NDM−1 and bla OXA−181 , respectively. Among the β-lactamase genes, bla CTX−M , bla SHV , or bla TEM were detected in almost all isolates (92.3%, 12/13), and only one E. coli isolate had no β-lactamase genes. The predominant bla CTX−M was bla CTX−M−15 , which was detected in 53.8% (7/13) of all isolates. The gene bla SHV was found in all K. pneumoniae and one isolate of E. coli ( Table 2). Among the ampC β-lactamase genes, bla CMY−2 only was detected in four E. coli isolates (28.6%). The other antimicrobial resistance genes in the mcr-carrying CRE isolates, including those for fluoroquinolones, aminoglycosides, rifampicin, macrolides, chloramphenicol, sulfonamide, tetracycline, fosfomycin, and trimethoprim, are shown in Table 2.

Antimicrobial Susceptibility of CRE Isolates Harboring mcr Genes
As shown in Table 3, the colistin MIC values for the mcrharboring isolates were in the range of 4-16 µg/ml. According to the 2020 CLSI M100-S30 guidelines, a microbe with a colistin MIC value of ≥4 µg/ml is resistant, whereas MIC ≤ 2 µg/ml indicates intermediate resistance. Thus, the results indicate that 100% of the mcr-carrying CRE isolated from patients were resistant to colistin. The highest MIC value for colistin (16 µg/ml) was found in two carbapenem-resistant K. pneumoniae isolates. Most of the mcr-harboring E. coli had colistin MIC values of 4 µg/ml (8/9; 88.8%).
The genetic relationships based on the SNPs of these mcrharboring isolates are demonstrated in Figures 3, 4. Five E. coli ST410 isolates were widely distributed in several sub-clusters of the ST410 cluster. Strain no. 53360 was closely related to the strain AMA1167 from Denmark, whereas strain no. 56511 clustered with strains from Norway, India, Lebanon, China, and Denmark (Figure 3). Strains no. 58967 and no. 62122 were related to strain KBN10P04869 from South Korea. Strain no. 53037 clustered with strains from Norway, the United States, Brazil, and Germany. Two ST3052 isolates (nos. 54881 and 54882) were closely related to the strain WCHEC020028 from China; they had similar characteristics and were from different individuals in the same hospital ward, indicating that they were likely to have originated from the same source. Strain no. 54715 (ST1287) was related to a strain from the United States (ST617), and these were clustered together with strain no. 6000 (ST10) (Figure 3).    Frontiers in Microbiology | www.frontiersin.org FIGURE 2 | Comparison of the genetic organization of the mcr genes in 13 CRE isolates. The schematic shows the genes flanking the mcr genes in the isolates.
FIGURE 3 | Phylogenetic tree based on single nucleotide polymorphisms (SNP) in E. coli using the neighbor-joining method. The tree was visualized and annotated using Interactive Tree of Life (iToL). The tree is annotated based on sequence types (STs; color highlight), mcr (square), bla NDM (circle), bla KPC (right triangle), and bla OXA−48−like (star). Presentation of the genes is shown by filled symbols. STs 410, 1287, 617, 10, and 3052 are highlighted in yellow, green, brown, pink, and blue, respectively. The isolates used in this study are shown in bold.
Among the K. pneumoniae isolates, the ST336 isolates were closely related and clustered with other ST336 strains isolated in Thailand (Figure 4). Interestingly, the Thai-ST336 isolates were in a different cluster from that containing the ST336 isolates from other countries. Similarly, isolate no. 60220 (ST340) in this study was FIGURE 4 | Phylogenetic tree based on single nucleotide polymorphisms (SNP) in K. pneumoniae using the neighbor-joining method. The tree was visualized and annotated using Interactive Tree of Life (iToL). The tree is annotated based on sequence types (STs; color highlight), mcr (square), bla NDM (circle), bla KPC (right triangle), bla OXA−48−like (star), and bla IMP (left triangle). Presentation of the genes is revealed by filled symbols. STs 336, 340, 11, and 4008 are highlighted in yellow, pink, and brown, and green, respectively. The isolates used in this study are shown in bold.
closely related to strain 4300STDY6636946 circulating in Thailand (Figure 4).

DISCUSSION
The discovery that plasmid-mediated colistin resistance is encoded by mcr genes and the high prevalence of human isolates harboring these genes are of global concern. A recent report revealed the overall average prevalence of mcr genes to be 4.7% (0.1-9.3%) in 47 countries across six continents (Elbediwi et al., 2019), and as many as 10 mcr genes (mcr-1 through to mcr-10) have been reported (Gharaibeh and Shatnawi, 2019;Wang et al., 2020). A study of the global prevalence of mcr genes revealed that mcr-1 (4917/5191; 94.7%) is a common gene and has a wider distribution compared with mcr-2 through to mcr-8 (Elbediwi et al., 2019). Human infections with both CRE and non-CRE isolates carrying mcr-1 have been widely reported (Liu et al., 2016;Mediavilla et al., 2016;Paveenkittiporn et al., 2017;Quan et al., 2017;Mendes et al., 2018;Srijan et al., 2018;Zhong et al., 2018;Elbediwi et al., 2019).
The coexistence of mcr and carbapenemase genes, such as bla NDM , bla OXA−48−like , and bla IMP , in CRE isolates has been described in countries worldwide (Mediavilla et al., 2016;Arabaci et al., 2019;Huang et al., 2020;Kananizadeh et al., 2020). The current study found the predominant mcr gene to be mcr-1, which more frequently coexists with bla NDM than bla OXA48−like , highlighting the potential dissemination of mcr-1 and bla NDM among CRE isolates in Thailand. This concurs with a study in China (Huang et al., 2020), where mcr-1 and bla NDM−5 were predominant (78. 6%, 11/14). In this study, mcr-1 and bla NDM−1 were the most prevalent resistance genes (61.5%, 8/13). A previous study showed that mcr-3 had a wide distribution in water, animals, food, and human isolates (Elbediwi et al., 2019). We found a 7.7% (1/13) prevalence for mcr-3.5, which co-occurred with mcr-1 and bla OXA−48 in E. coli. The phenomenon of double mcr genes has been reported in isolates from humans, with K. pneumoniae harboring mcr-3 and mcr-8 being isolated from the stool of a healthy individual in Laos (Hadjadj et al., 2019).
Our study revealed that the most common type of plasmid replicon carrying mcr was IncX4. Previous reports have shown IncX4, IncI2, and IncHI2 to be the major plasmid types driving the global dissemination of mcr-1 (Wu et al., 2018). A study in Thailand revealed two predominant plasmid types (IncX4 and IncI2) carrying mcr-1 in CRE (Shanmugakani et al., 2019). This suggests that IncX4 bearing mcr-1 mediates the transmission of CRE and may promote its circulation throughout Thailand. IncX4 and IncI2 acting as vehicles for mcr-1 propagation enhance host fitness and provide a competitive advantage over strains with other plasmid replicon types, resulting in greater plasmid stability (Wu et al., 2018).
On the basis of the Clermont phylotyping scheme, E. coli species can be divided into eight main phylogroups, termed A, B1, B2, C, D, E, F, and G (Clermont et al., 2019). The nine carbapenem-resistant E. coli isolates carrying mcr (55.6%) in this study belonged to phylogroup C, whereas the rest belonged to phylogroups A and D (22.2% each). The E. coli strains responsible for extra-intestinal infection were more likely to be members of phylogroups B2 or D, which show greater pathogenesis than A, B1, or C (Clermont et al., 2013). Strains belonging to phylogroups A, B1, and C are commonly commensal, suggesting that more than half of the E. coli harboring mcr isolated from patients in this study were commensal strains.
Our study revealed seven carbapenem-resistant STs that carry mcr, of which E. coli ST410 (35.7%) and K. pneumoniae ST336 (21.4%) isolates were predominant. Elbediwi et al. (2019) reported that E. coli ST101 carrying mcr-1 have been found in environmental samples, animals, and humans. However, ST10 was the most globally common ST of E. coli carrying mcr-1 (Elbediwi et al., 2019). In Asia, ST116 was found to be the predominant ST that carries mcr-1 isolated from humans, followed by ST117, ST10, ST38, ST101, and ST156 (Elbediwi et al., 2019). E. coli ST410 is internationally considered a new high-risk clone that can cause several types of infection; it is highly resistant and has a global distribution (Roer et al., 2018). This ST has been described in Southeast Asia following multiple introductions through several independent events and differs from clones detected in Europe and North America (Nadimpalli et al., 2019). SNP phylogenetic analysis in this study revealed that the ST410 isolates were diverse or not closely related to other strains. Instead, they were shown to be related to strains from countries other than Thailand. This adds support to the assumption that there have been multiple dissemination events into this area.
Klebsiella pneumoniae ST336 (CC17) is considered an international clone (Rodrigues et al., 2014;Novović et al., 2017;Palmieri et al., 2020) and has been frequently associated with the worldwide spread of bla CTX−M−15 and bla OXA−48−like (Rodrigues et al., 2014;Novović et al., 2017;Palmieri et al., 2020). Interestingly, all ST336 isolates in the current study carried bla NDM−1 , but not any bla CTX−M genes, suggesting that they may be from different lineages. To the best of our knowledge, the carbapenem-resistant ST336 isolates in this study are the first to be described as having mcr-1. Previous studies revealed that colistin-resistant ST336 resulted from an mgrB mutation, and no mcr genes have been detected in this ST (Novović et al., 2017;Palmieri et al., 2020). SNP phylogenetic analysis allocated the ST336 isolates to the same cluster as other Thai-ST336 isolates, and this cluster was independent from another ST336 cluster consisting of isolates from other countries. This suggests that Thai-ST336 isolates circulate throughout our country by clonal expansion.
The colistin MIC values (4-16 µg/ml) for our isolates indicated they have 100% resistance. Combinations of mcr and chromosome-mediated colistin resistant genes (pmrAB, phoPQ, crrB, or recN) contributed to the colistin resistance of our isolates. It is interesting that substitutions in pmrAB and crrB in mcr-1-carrying K. pneumoniae are quite different from previous reports (Olaitan et al., 2014a,b;Wright et al., 2015;Cheng et al., 2016). In addition, substitution of phoPQ was not found in our K. pneumoniae isolates comparing to those studies (Olaitan et al., 2014a,b;Wright et al., 2015;Cheng et al., 2016). On the other hand, substitutions of pmrA (G144S), pmrB (H2R, D283G, Y358N), and phoP (I44L) in our mcr-1-harboring E. coli are similar to a study previously described (Choi et al., 2020). In case of rpoN, inactivation of this gene resulting in polymyxin resistance has been observed in Salmonella enterica (Barchiesi et al., 2009). However, polymyxin resistance via rpoN inactivation or substitution is not yet reported in either E. coli or K. pneumoniae. Although our study detected substitution of rpoN, but its role on polymyxin resistance remain to be investigated.
Several studies have shown mcr-1-carrying CRE isolated from humans to have high frequencies of colistin resistance, such as 71.4% in China (Huang et al., 2020), 100% in the United States (Mediavilla et al., 2016), and 100% in Turkey (Arabaci et al., 2019). A previous study in Thailand reported colistin resistance rates of 75.0 and 79.1% in carbapenemresistant E. coli and K. pneumoniae isolates, respectively (Eiamphungporn et al., 2018). The 13 carbapenem-resistant mcr-1-harboring CRE isolates described here showed a high susceptibility (>50%) to the antibiotics amikacin, fosfomycin, and nitrofurantoin. In contrast, the human mcr-1-harboring CRE isolate from China was reported to be highly susceptible only to tigecycline, amikacin, and aztreonam (Huang et al., 2020), whereas the isolate from the United States was susceptible to more antibiotics, including amikacin, aztreonam, gentamicin, nitrofurantoin, tigecycline, and trimethoprim-sulfamethoxazole (Mediavilla et al., 2016).
Polymyxins, including colistin, were reintroduced into human medical practice by the WHO in 2012 (World Health Organization, 2012). In China, this antibiotic was approved for clinical use in the treatment of bacterial infections in 2017. Since then, the relative prevalence of CRE carrying mcr genes increased from 0.41 to 1.38% (Huang et al., 2020). A study in Singapore revealed that prior exposure to polymyxin (adjusted OR, 21.31; 95% CI, 3.04-150.96) and carbapenem (OR 3.74; CI 1.13-12.44) were independent risk factors for polymyxin-resistant CRE among hospitalized patients (Teo et al., 2019). A study in Thailand demonstrated that chronic kidney diseases (OR 3.95; CI 1.26-12.32) and exposure to antimicrobials for less than 3 months (OR 2.29; CI 0.29-18.21) were risk factors associated with infections by mcr-1-carrying Enterobacteriaceae (Shanmugakani et al., 2019). In patients infected with polymyxin-resistant CRE, the 30-day all-cause in-hospital mortality was 50.0% compared with a 38.1% mortality in patients with polymyxin-susceptible CRE (Teo et al., 2019). Therefore, minimizing the use of polymyxin and carbapenem is strongly recommended.
The findings of the current study provide comprehensive insights into the prevalence of mcr-carrying CRE in patients in Thailand. In general, mcr-1 was present in E. coli and K. pneumoniae isolates. The co-occurrence of two mcr genes was also demonstrated in CRE isolated from patients. To slow the emergence of XDR or PDR strains, priority should be given to strengthening official surveillance, active control, and prevention efforts as well as minimizing the dissemination of mcr genes among CRE isolates in humans.

DATA AVAILABILITY STATEMENT
The datasets generated for this study can be found in the Raw sequencing data were deposited in the Sequence Read Archive (SRA) of NCBI under the BioProject ID PRJNA380676.

ETHICS STATEMENT
The Human Research Ethics Committee of Department of Medical Sciences, Ministry of Public Health, reviewed this study and judged that the protocol constituted routine public health activities and therefore did not involve human subject research. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

AUTHOR CONTRIBUTIONS
WP and AK conceived and designed the study, performed the data analysis, drafted the manuscript, and performed critical revisions of the manuscript for intellectual content. WK performed the laboratory experiments and analyzed the data. RU performed the statistical analysis and critical revision of the manuscript for intellectual content. All the authors read, edited, and approved the final manuscript.