Background

The emergence of carbapenemase-producing Enterobacteriaceae (CPE) has become a serious problem in medical settings worldwide [1]. The most frequently detected and globally widespread carbapenemase produced by CPE between the Asian countries are the class B metallo-β-lactamases (MBLs), which include IMP-type, NDM-type, and VIM-type MBLs [2].

Because of the emergence of multidrug-resistant Gram-negative pathogens and the lack of new antibiotics with efficient activities, colistin, a polymyxin-type antibiotic, has been the last resort used to treat CPE infections [3, 4]. Bacteria acquire colistin resistance through chromosomal mutation(s) or plasmid transfer [5]. Chromosome-mediated colistin resistance results from mutation(s) or deletion(s) of two component systems, such as phoPQ and pmrAB, altering the structure of lipopolysaccharides [6, 7]. For example, colistin resistance has been associated with modification of the lipid A moiety in lipopolysaccharide, such as by the addition of 4-amino-4-deoxy-L-arabinose (L-Ara-4 N) and phosphoethanolamine (PEtN) to the anionic phosphate groups of lipid A. These additions reduce the anionic charges on lipid A and its affinity to the cationic colistin, inhibiting membrane destruction resulting from the binding of colistin to lipid A, followed by cell death [8].

To date, various types of plasmid-mediated mobilized colistin-resistance genes, mcr, have been identified, including mcr-1 to mcr-9, with several, including mcr-1, − 2, − 3, − 4, and − 6, shown to have PEtN transferase activity [6]. The mcr-1 gene was initially detected in isolates of Escherichia coli and Klebsiella pneumoniae obtained from humans and animals in 2015 in China [9], and mcr-9 was initially identified in a clinical isolate of the colistin-susceptible bacterium, Salmonella enterica serotype typhimurium. The amino acid and nucleotide sequences of mcr-9 are closest to those of mcr-3 with similarities of 64.5 and 99.5%, respectively [10]. In this study, mcr-9 was detected in 335 genomes in multiple genera of Enterobacteriaceae. The analysis of mcr-9 promoter region in these genomes showed conserved regions which is likely a recognition sequence for transcription regulator, suggesting that other factors might be involved in full-expression of mcr-9. Of the 335 genomes, 65 had at least one plasmid replicon indicating that mcr-9 can be found extrachromosomally in different species of Enterobacteriaceae [10].

Isolates of E. cloacae complex resistant to both carbapenem and colistin have been reported in several countries, including China [11, 12], France [13], India [14], the USA [15, 16] and Vietnam [17]. One of these, an isolate of E. cloacae complex (Enterobacter hormaechies) co-harboring blaVIM-4 and mcr-9, was first reported in the United States in 2019 [16]. In addition, a colistin-resistant E. hormaechei isolate producing both MCR-9 and NDM-1 was isolated from a patient in China with bloodstream infection in 2019 [11]. This emergence of colistin resistance, particularly in CPE, may result in significant clinical and public health concerns [18, 19].

The study describes three clinical isolates of E. cloacae complex that were resistant to carbapenem but susceptible to colistin. To our knowledge, this is the first report of isolates of E. cloacae complex harboring both blaIMP-1 and mcr-9 in Japan.

Methods

Bacterial strains

Thirty-two clinical isolates of E. cloacae complex, each with minimum inhibitory concentrations (MICs) of ≥2 μg/ml for meropenem and/or imipenem, had been obtained from individual patients at 14 hospitals in eight prefectures throughout Japan from July to October 2018 by BML Biomedical Laboratories R&D Center (Kawagoe, Saitama, Japan).

Drug susceptibility testing

The MICs of antibiotics were determined using a broth microdilution method according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI) [20]. The MICs of colistin were also determined by a broth microdilution using cation-adjusted Muller Hinton broth and 96-well microtiter plates (Kohjin Bio, Co., Ltd. Saitama, Japan) according to the guidelines of the European Union Committee for Antimicrobial Susceptibility Testing (EUCAST) [21].

Whole genome sequencing

DNA was extracted from each E. cloacae complex isolate using DNeasy Blood and Tissue kits (Qiagen, Tokyo, Japan). A Nextera XT DNA library was prepared from each extracted DNA sample. Each DNA library of was sequenced on the MiSeq system (Illumina) to obtain short reads with 300-bp paired-end reads. MiSeqRun was performed using Nextera XT Index Kit v2 and MiSeq Reagent Kit v3. DNA Libraries for MinION (Oxford Nanopore Technologies, Oxford, UK) were prepared from three isolates (A2483, A2504 and A2563) using Ligation Sequencing Kits 1D (SQK-LSK109) to yield long contigs. The long read generated by MinION were assembled using Canu v1.7.1 and polished with the short reads generated by MiSeq using Pilon v1.22. The nucleotide sequences of plasmids and chromosomes carrying blaIMP-1 and mcr-9 were compared with similar sequences using BLAST and visualized by In silico MolecularCloning. Ver.7 genomic edition (https://www.insilicobiology.co.jp/).

Bacterial species were identified by analyses of average nucleotide identity (ANI) [22] and digital DNA-DNA hybridization (dDDH) [23] of whole genome sequences. The seven type strains used as reference species included Enterobacter asburiae (ATCC35953T), E. cloacae (ATCC13047T), E. hormaechei (ATCC49162T), Enterobacter kobei (DSM13645T), Enterobacter ludwigii (EN-119T), Enterobacter nimipressuralis (DSM18955T) and Enterobacter xiangfangensis (LMG27195T). In silico multilocus sequence typing (MLST) was assigned by PUBMLST database (https://pubmlst.org/databases/). Acquired antibiotic resistance genes were identified using the ResFinder 3.2 tool (https://cge.cbs.dtu.dk/services/ResFinder/) from the Center for Genomic Epidemiology (CGE).

Phylogenetic analysis based on SNPs

Single nucleotide polymorphisms (SNPs) in the 32 isolates were identified by aligning whole-genome sequencing data of these isolates with the genomic sequences of the E. xiangfangensis reference isolate LMG27195 (GenBank accession no. CP017183.3), using the CSI Phylogeny 1.4 tool (https://cge.cbs.dtu.dk/services/CSIPhylogeny/) from CGE. A phylogenetic tree was constructed using Fig Tree (version 1.4.4) and a maximum likelihood phylogenetic tree (http://tree.bio.ed.ac.uk/software/figtree/).

Results

Phenotypic and genotypic properties of carbapenem-non-susceptible isolates

Drug susceptibility of carbapenem-non-susceptible isolates

The MICs of the 32 clinical isolates of E. cloacae complex are shown in Table 1. All were susceptible to amikacin and colistin, but resistant to ceftazidime. Of these 32 isolates, 25 were resistant to aztreonam, 15 were resistant to ciprofloxacin, and 12 were resistant to tigecycline. Of the all 32 isolates, 28 isolates were resistant to imipenem and/or meropenem with MICs ≥4 μg/ml, whereas the remaining 4 were intermediate to imipenem and/or meropenem with MICs ≥2 μg/ml (Table S1). There are no isolates susceptible to both imipenem and meropenem (Table S1).

Table 1 MIC values of 32 clinical E. cloacae complex isolates
Full size table

Whole genome sequences of carbapenem-non-susceptible isolates

Whole genome sequencing of the 32 isolates of E. cloacae complex showed that, based on ANI and dDDH analyses, 31 were E. xiangfangensis and one was E. asburiae. MLST analysis revealed that 13 isolates (40.6%) belonged to sequence type (ST) 78; 10 (31.2%) to ST133; two each (6.3%) to ST175 and ST1196; and one each (3.1%) to ST62, ST93, ST418, and ST484. The ST for one isolate could not be determined because its housekeeping genes did not match those of current STs. A phylogenetic tree of these 32 isolates revealed four major clades, with clades I, II, III and IV consisting of 14, 2, 10 and 6 isolates, respectively (Fig. 1). Clade I consisted of isolates belonging to ST78 and the non-typeable isolate, clade II of isolates belonging to ST418 and ST484, clade III of isolates belonging to ST133 and clade IV of isolates belonging to ST1196, ST175, ST93 and ST62. These isolates harbored various genes associated with drug resistance (additional file: Table S1). All 32 isolates harbored blaIMP-1, with three also harboring mcr-9 (Table S1).

Fig. 1
figure 1

Phylogenetic tree of 32 isolates of E. cloacae complex obtained from 14 hospitals in eight prefectures throughout Japan. The phylogenetic tree was constructed by the maximum-likelihood method based on single nucleotide polymorphisms (SNP) in the core genome and aligned with E. xiangfangensis reference isolate LMG27195 (GenBank accession no. CP017183.3) (Isolates No./Hospital No.)

Full size image

Phenotypic and genotypic properties of isolates harboring both blaIMP-1 and mcr-9

Bacterial identification and drug susceptibility

Of three isolates co-harboring blaIMP-1 and mcr-9, two, A2483 and A2504, were E. xiangfangensis and one, A2563, was E. asburiae (Table S1). The A2483 and A2504 strains obtained in a hospital belonged to ST1199, whereas the A2563 strain obtained in another hospital belonged to ST484. The two hospitals located in the same prefecture in Japan. The drug susceptibility profiles of the two E. xiangfangensis isolates were identical to each other, with both A2483 and A2504 being resistant to aztreonam, ceftazidime, imipenem, meropenem and tigecycline, and susceptible to amikacin, ciprofloxacin and colistin (Table 2). The E. asburiae isolate was resistant to ceftazidime and imipenem, had intermediate resistance to meropenem, but was susceptible to the other drugs tested including colistin (Table 2).

Table 2 Drug susceptibility profile of E. cloacae complex isolates co-harboring mcr-9 and blaIMP-1
Full size table

Whole genome sequences of isolates harboring both blaIMP-1 and mcr-9

As shown in Table 3, E. xiangfangensis A2483 contained a chromosome of 5,024,985 bp with a GC content of 55.24% and two plasmids of 61,594 bp and 288,696 bp, respectively. E. xiangfangensis A2504 contained a chromosome of 4,934,510 bp with a GC content of 54.70% and two plasmids of 61,594 bp and 276,927 bp, respectively. The whole genome sequences of A2483 were very close to those of A2504 with similarities of 100% (98% query coverage) on the chromosome, 100% for the 62-kbp plasmid and 100% (96% query coverage) for the 289-kbp plasmid. E. asburiae A2563 contained a chromosome of 4,934,510 bp with a GC content of 55.80% and one plasmid 115,246 bp in size (Table 3). In addition to blaIMP-1 and mcr-9, these isolates harbored several other genes associated with drug resistance, including aac (6′)-IIc, blaACT-6, blaACT-7, fosA and sul1 (Table 3). blaACT genes are the intrinsic AmpC encoding genes of Enterobacter cloacae complex species.

Table 3 Genetic characterization of carbapenem-resistant and colistin non-resistant E.cloacae complex isolates coharboring blaIMP-1 and mcr-9
Full size table

Location of mcr-9 and its genetic environments

The mcr-9 gene was present on the 289-kpb IncHI2 plasmid of A2483 and the 277-kpb IncHI2 plasmid of A2504, but was present on the chromosome of A2563 (Table 3). The two plasmids harboring mcr-9, pA2483mcr-9 on A2483 and pA2504mcr-9 on A2504, had the same GC content of 46.30%, and contained open reading frames (ORFs) of 360 and 358, respectively (Fig. 2. (a)). The nucleotide sequences of these plasmids were identical to each other, except for a genetic region with 11,770 bp, from nucleotide (nt) 146,310 to nt 158,080, in the 277-kbp plasmid. The mcr-9 gene on the chromosome of A2563 was detected at nt ~ 129 Mb.

Fig. 2
figure 2

Circular structures of chromosome and plasmids harboring blaIMP-1 and mcr-9. a Map of the plasmids pA2483mcr-9 and pA2504mcr-9, showing that, relative to pA2483mcr-9, a pA2504-mcr-9 had a deletion of 11,770 bp from nt 146,310 to nt 158,080. b Map of the plasmids pA2483imp-1 and pA2504imp-1, showing that their structures were identical. c Map of the chromosome of A2563 coharboring blaIMP-1 and mcr-9. The two outermost circles show forward (light blue) and reverse (dark blue) genes and the innermost circle show GC content, with dark orange indicating above average and light orange indicating below average. Colored arrows indicate the positions and directions of antibiotic resistance genes (pink), surrounding genes (light grey) and insertion sequences (dark grey). Truncated genes are indicated by ∆

Full size image

The genetic environments of mcr-9 in the A2483 and A2504 plasmids were identical to each other, with mcr-9 located in a ~ 30 kb region surrounded by two insertion sequences encoding an IS5-like element (IS903 family transposase; Fig. 3). The region upstream of mcr-9 included rcnR (encoding a Ni/Co-binding transcriptional repressor), pcoS (encoding a two-component sensor histidine kinase) and pcoE (encoding a copper-binding protein). The region downstream of mcr-9 included wbuC (encoding a cupin fold metalloprotein) but no genes encoding the two-component system qseC-qseB, which has been associated with the expression of mcr-9 [24]. Insertion sequences were not detected in the region downstream of mcr-9 on the A2563 chromosome. The region upstream of mcr-9 was rcnR-pcoS-ΔpcoE, whereas the region downstream of mcr-9 was wbuC-qseC-qseB (Fig. 3). The A2483 and A2504 plasmids showed 83% query coverage and 99.97% identity to the IncHI2 plasmid, pME-1a (GenBank accession no. NZ_CP041734.1), in E. hormaechei, a strain isolated in 2019 from a pediatric inpatient in the USA (Fig. 3) [16]. Three IncHI2 plasmids were identified with similar sequences, pCTXM9_020038 from E. hormaechei isolated in China in 2018 (83% query and 99.97% identity; GenBank accession no. CP031724), pRH-R27 from Salmonella enterica Infantis in Germany in 2015 (82% query coverage and 99.99% identity; GenBank accession no. LN555650), and pMCR-SCNJ07 from E. hormaechei isolated in China in 2019 (80% query and 99.99% identity; GenBank accession no. MK933279) (Fig. 3).

Fig. 3
figure 3

Linearized genetic environments of mcr-9 harbored by IncHI2-type pA2483mcr-9, pA2504mcr-9 (this study), pME-1a (GenBank accession no. NZ_CP041734.1), pCTXM9_020038 (GenBank accession no. CP031724), pRH-R27 (GenBank accession no. LN555650), p707804-NDM (GenBank accession no. MH909331 and by the A2563 chromosome (this study). Colored arrows indicate positions and directions of antibiotic resistance genes (pink), surrounding genes (light grey) and insertion sequences (dark grey). Truncated genes are indicated by ∆

Full size image

Location of blaIMP-1 and its genetic environment

The blaIMP-1 gene was present on the 62-kbp plasmids of A2483 and A2504 and on the chromosome of A2563 (Table 3). The two plasmids harboring blaIMP-1, pA2483imp-1 from A2483 and pA2504imp-1 from A2504, had the same GC content of 47.40% and identical nucleotide sequences (Fig. 2 (b)). In these plasmids, blaIMP-1 was located in a class I integron containing intl-blaIMP-1-aac (6′)-IIc-qacE∆1-sul1. The blaIMP-1 gene on the chromosome of A2563 was present in the same class I integron (Fig. 2 (c)). The same class I integron containing blaIMP-1 and aac (6′)-IIc was detected in the bacteria E. asburiae NUH15_ECL035_1 (GenBank accession no AP019388.1), Enterobacter cloacae NUH15_ECL020 (GenBank accession no AP019386.1) and Enterobacter asburiae NUH12_ECL030 (GenBank accession no AP019383.1), all of which were isolated in Japan in 2019. The two pA2483imp-1 and pA2504imp-1 showed 84% query coverage and 97.75% identity with the plasmid pJJ1886_4 (GenBank accession no CP006788.1), which was detected in the USA and did not contain a class I integron or any other resistance genes [25].

Discussion

The mcr-9 gene may be silently spreading in Enterobacteriaceae throughout the world. The prevalence of mcr-9 is unclear because this gene is not actually related to colistin resistance, as it may be silent or inducible in clinical isolates of Enterobacteriaceae. For example, an isolate of E. hormechei harboring mcr-9 did not express its gene product [16]. This isolate was susceptible to colistin, likely because the two-component system genes qseCB were lacking from the region downstream of mcr-9. In contrast, another isolate of E. hormechei that harbored and expressed mcr-9 was found to be resistant to colistin and to have the two-component system genes in the region downstream of mcr-9 [11]. The expression of mcr-9 is mediated by the two-component system QseCB and can be induced by subinhibitory concentrations of colistin [24]. At least 11 mcr-9-positive IncHI2 plasmids have been detected by Blast, with six having and five lacking the two-component system genes [11].

The two-component QseCB system, consisting of a sensor (qseC) and a response regulator (qseB), plays an essential role in the expression of mcr-9 [24]. Our finding, that the isolate A2563 harbored mcr-9 along with the two-component system genes qseCB but was susceptible to colistin suggests that other, as yet undetermined, genes or molecules may regulate mcr-9 expression. The pA2480mcr-9 and pA2504mcr-9 had similar structures to those of pME-1a and pCTXM9_020038, as they lacked qseCB. This two-component system was transcribed as an operon, with the QseB promoter binding to low- and high-affinity binding sites located − 500 to − 10 bp at upstream of qseB [26]. The nucleotide sequence of this region in A2563 was 100% identical to that of the QseB promoter (− 500 to + 1 bp) in pMCR-SCNJ07, which confers resistance to colistin [11], suggesting that the QseB promoter in A2563 may be repressed by an as yet undetermined mechanism [26]. Four plasmids, pME-1a, pCTXM9_020038, pRH-R27 and pMCR-SCNJ07, had the conserved gene structure, rcnR-pcoS-pcoE-IS-5, upstream of mcr-9. Whereas, the chromosome of A2563 had the same conserved gene structure, but with a 53-bp deletion in pcoE (ΔpcoE), suggesting that the deleted region may be associated with mcr-9 expression. Further studies are necessary to determine the mechanism for regulation of mcr-9 expression in Enterobacteriaceae.

To our knowledge, it is the first report describing a bacterial isolate harboring mcr-9 on its chromosome, indicating that mcr-9 may have been inserted into the chromosome by mobile elements. Several Enterobacteriaceae isolates from animals and humans have reported the chromosomal location of mcr-1 and mcr-2 [27,28,29,30,31,32,33]. The mcr-1 was detected on the chromosomes of two colistin-resistant E. coli strains isolated from swine in 2012 in China [28], and on the chromosome of an E. coli ST410 strain harboring blaCTX-M-15 isolated from a sample of turkey meat in 2013 in Germany [27]. The chromosomal integration of mcr-1 was also detected in a clinical strain of E. coli ST156 harboring blaNDM-5 isolated from a bile sample in 2015 in China [29], in E. coli isolated from food production animals in 2011–2016 in Poland [32], and in E. coli isolated from veal calves in 2016 in the Netherlands [30]. Chromosomes carrying mcr-1 were detected in Enterobacteriaceae from environmental water sources in 2017 in China [33]. Moreover, the mcr-2 gene (mcr-6.1) was detected on the chromosome of a strain of Moraxella isolated from a pig in 2014–2015 in Great Britain [31]. These studies support the mobility characterization of mcr genes across different genetic elements and insertion of the plasmid-variant of mcr into chromosome could lead to higher prevalence of colistin resistance among Enterobacteriaceae specious.

The direct origin of the mcr-9 on the chromosome of A2563 is unclear. However, the genetic environments of the mcr-9 and qseCB genes in A2563 are similar to those of pMCR-SCNJ07 from E. hormaechei in China in 2019 (GenBank accession no. MK933279), pRH-R27 from Salmonella enterica Infantis in Germany in 2015 (GenBank accession no. LN555650), pT5282-mphA from E. cloacae in China in 2012 (GenBank accession no. KY270852), pN1863-HI2 from E. cloacae in China in 2017 (GenBank accession no. MF344583), pSE15-SA01028 from S. enterica subsp. enterica in Germany in 2018 (GenBank accession no. NZ_CP026661) and p707804-NDM from Leclercia adecarboxylata in China in 2018 (GenBank accession no. MH909331). These 7 strains carrying plasmids with mcr-9 in China and Germany did not harbored blaIMPs, but blaNDMs or blaVIMs [11].

The plasmids pA2483imp-1 and pA2504imp-1 had the same backbone as the plasmid pJJ1886_4 (GenBank accession no CP006788.1), which had been isolated in the USA. The 55,956 bp plasmid pJJ1886_4, which was smaller in length than the 61,594 bp plasmids pA2483imp-1 and pA2504imp-1, lacked a class I integron carrying blaIMP-1 (intl-blaIMP-1-aac (6′)-IIc-qacE ∆1-sul1). The E. cloacae EN3600 plasmid (GenBank accession no CP035638.1) carrying blaIMP-8 also had the same backbone as pJJ1886_4, with 83% coverage and 96.8% identity. These findings indicate that pJJ1886_4 has spread globally and captured drug-resistance genes and that this plasmid functions as a carrier of acquired drug-resistance genes.

Conclusion

In conclusion, this study describes the characterization of the complete genomes of three clinically obtained isolates of carbapenem-resistant and colistin-susceptible E. cloacae complex harboring both blaIMP-1 and mcr-9 from different hospitals in Japan. Enterobacteriaceae harboring both blaIMP-1 and mcr-9 may become a healthcare problem, suggesting the need for steps to prevent their further dissemination.