Potential Mobilization of mcr-10 by an Integrative Mobile Element via Site-Specific Recombination in Cronobacter sakazakii

mcr-10 is a newly identified plasmid-borne colistin resistance gene, but its mobilization mechanism remains unclear. In this study, mcr-10 was found on an IncFIB plasmid carrying virulence genes mrkABCDFJ, iucABCD/iutA, and eitCBAD in a Cronobacter sakazakii isolate.

ABSTRACT

mcr-10 is a newly identified plasmid-borne colistin resistance gene, but its mobilization mechanism remains unclear. In this study, mcr-10 was found on an IncFIB plasmid carrying virulence genes mrkABCDFJ, iucABCD/iutA, and eitCBAD in a Cronobacter sakazakii isolate. By comparison with closely related plasmids, two recombination sites were identified flanking the genetic element containing mcr-10 and an integrase-encoding gene, suggesting that site-specific recombination mediated by an integrase of an integrative mobile element is a potential mechanism for mobilizing mcr-10.

INTRODUCTION

Plasmid-borne mobile colistin resistance genes (mcr) have become a common mechanism mediating resistance to colistin in Enterobacteriaceae (1, 2). Recently, we identified mcr-10, a novel mcr variant, on an IncFIA plasmid, pMCR10_090065, of an Enterobacter roggenkampii isolate (3). On pMCR10_090065, mcr-10 is adjacent to a tyrosine recombinase-like integrase-encoding gene (3). Typically, this type of integrases may recognize two 30- to 40-bp crossover sites (recombination sites) with inverted repeat symmetry and then realizes site-specific recombination, which results in integration, excision, or inversion of genetic components (4). However, only one recombination site was identified on pMCR10_090065, and therefore the mechanism for mobilizing mcr-10 has not been fully understood. In this study, we detected the presence of mcr-10 on an IncFIB plasmid carrying multiple virulence genes in a Cronobacter sakazakii isolate, and we found that the mobilization of mcr-10 was mediated by an integrative mobile genetic element.

Isolate 145005 was recovered from a fecal sample that was collected from a healthy volunteer in a microbiome project in 2019 at West China Hospital. This study was approved by the Ethical Committee of West China Hospital, with informed consent obtained. Short-read whole-genome sequencing of isolate 145005 using a HiSeq X10 platform (Illumina, San Diego, CA) generated 1.71 clean Gb (388.6× coverage), which were de novo assembled into 111 contigs (N₅₀, 137,377 bp) by using SPAdes v3.14.0 (5) invoked in Shovill v1.0.9 (https://github.com/tseemann/shovill). The draft genome was 4.4 Mb, with a 56.8% G+C content. Isolate 145005 belongs to C. sakazakii, as it has 99.63% pairwise average nucleotide identity (ANI) to ATCC 29544, the type strain of C. sakazakii (GenBank accession no. CP011047), as determined using JSpeciesWS (6). By querying the MLST multilocus sequence typing database (https://cge.cbs.dtu.dk/services/MLST/), this isolate was assigned to ST8, a sequence type known to cause neonatal infections (7, 8).

Isolate 145005 showed intermediate resistance to colistin (MIC, 2 mg/liter) but was susceptible to ampicillin (MIC, 1 mg/liter), aztreonam (MIC, 1 mg/liter), ceftazidime (MIC, 0.125 mg/liter), cefepime (MIC, 0.125 mg/liter), imipenem (MIC, 1 mg/liter), piperacillin-tazobactam (MIC, 4/4 mg/liter), and tigecycline (MIC, 1 mg/liter) as determined using the broth microdilution method of the Clinical and Laboratory Standards Institute (CLSI) (9). As there are no breakpoints of tigecycline from CLSI, those defined by EUCAST (http://www.eucast.org/) were applied. By querying the ResFinder database (https://cge.cbs.dtu.dk/services/ResFinder/), three known antimicrobial resistance genes, bla_CSA-1 (a chromosomal ampC gene intrinsic to Cronobacter spp. mediating resistance to first-generation cephalosporins) (10), fosA (a chromosomal gene mediating resistance to fosfomycin; identical to the NCBI reference sequence WP_007901855.1), and mcr-10 with 100% nucleotide identity to the original mcr-10 gene on pMCR10_090065 (3), were found in isolate 145005. mcr-10 is able to mediate reduced susceptibility to colistin, but such reduced susceptibility is below the >2-mg/liter breakpoint to define resistance (3). mcr-10 is mainly identified in Enterobacter strains (3) but had not been seen in Cronobacter spp. before, and such an identification expands the host spectrum of mcr-10.

The complete plasmid sequence carrying mcr-10 was obtained by long-read whole-genome sequencing with a MinION sequencer (Nanopore, Oxford, United Kingdom) and subsequent de novo hybrid assembly of both short (Illumina) and long reads with Unicycler v0.4.3 (11) and Pilon v1.22 (12). In isolate 145005, mcr-10 was found on a 120-kb IncFIB(pCTU1) plasmid, designated pMCR10_145005 here, with the replicon type being assigned by querying PlasmidFinder 2.0 (https://cge.cbs.dtu.dk/services/PlasmidFinder/). No genes encoding conjugation-related proteins were found in pMCR10_145005, suggesting that the plasmid is non-self-transmissible. A previous study demonstrated that all RepFIB plasmids of Cronobacter carry two iron acquisition loci: the iucABCD/iutA operon encoding a siderophore (13) and the eitCBAD operon encoding an ATP-binding cassette transport-mediated system (14). Both loci are also present on pMCR10_145005 (Fig. 1). In addition, an mrkABCDFJ locus was also identified on pMCR10_145005. The mrkABCDFJ genes encode the mannose-resistant Klebsiella-like hemagglutinins to form type 3 fimbriae for attachment to surfaces and therefore form biofilms, which are associated with virulence in Enterobacteriaceae (15, 16). The presence of the aforementioned genes suggests that pMCR10_145005 carries known virulence genes and the colocation of virulence genes and mcr-10 in the single plasmid is of clinical importance.

FIG 1.

Comparison of pMCR10_145005 with pCsaCS931b. pCsaCS931b (GenBank accession no. CP049260) was recovered from C. sakazakii isolate CS-931 in Mexico. The genes encoding plasmid key functions (replication and partition) and virulence and coding and insertion sequences in the two extra regions on pMCR10_145005 are shown. A truncated gene or insertion sequence is indicated by Δ. parA and parB encode a plasmid partitioning system, while repB encodes the replication initiation protein of the IncFIB(pCTU1) replicon. int2, which encodes a site-specific integrase, Int2, with 69% amino acid identity to Int_MCR-10 encoded by int, and its adjacent attB site (sequence shown in Fig. 2) on pCsaCS931b are shown in gray. The figure was generated using CGView (19).

In addition, there is a new insertion sequence on pMCR10_145005 that is 1,613 bp in length and is most closely related to ISPa31 of the IS3 family, with 44% coverage and up to 67.31% identity. The insertion sequence has 20/29-bp imperfect inverted repeats and generates 4-bp (CGTA) direct target repeats (DR) on insertion. The new insertion sequence has been designated ISCrsa1 by the ISfinder database (https://www-is.biotoul.fr/).

The most closely related plasmid of pMCR10_145005 was pCsaCS931b (GenBank accession no. CP049260) of C. sakazakii isolate CS-931, with 60% coverage and up to 99.62% identity (Fig. 1). Isolate CS-931 was recovered from feces of a 6-year-old boy in Mexico in 2016 (BioSample no. SAMN07540307). Compared to pCsaCS931b, pMCR10_145005 has two extra regions: one of 33.5 kb and one of 11 kb (Fig. 1). The 33.5-kb extra region on pMCR10_145005 contains mcr-10, the type 3 fimbria-encoding mrkABCDFJ gene locus, and various insertion sequences, including 6 complete sequences and 2 truncated remnants of IS26, 3 complete sequences and a truncated remnant of ISKpn26, an ISEc36, an IS1, a truncated remnant ISEcl1-like element (91% nucleotide identity to ISEcl1), and a truncated remnant of IS903 (Fig. 1). The 11-kb extra region contains genes encoding a kinase, a phosphatase 2C domain-containing protein, and a ribbon-helix-helix protein of the CopG family, as well as multiple insertion sequences, including ISCrsa1, a truncated IS1, and two complete copies of IS26 (Fig. 1). By BLAST, both of the 33.5- and 11-kb extra regions have multiple fragmented matches, suggesting multiple and complicated origins formed by the action of the insertion sequences in the regions.

Both of the 33.5- and 11-kb extra regions on pMCR10_145005 are separated by a 14-kb intervening region, and both regions are bounded by IS26 at one side, while the two IS26 copies are in the opposite orientation (Fig. 2A). IS26 generates characteristic 8-bp DRs on insertion. It has been known that when a single copy of IS26 targets a site in the same replicon (chromosomes or plasmids) and attacks the opposite strand (trans attack), the genetic component between the IS26 and the target site is inverted and bounded by the original IS26 and a new copy of IS26 in the opposite orientation (17). Such IS26-mediated intramolecular replicative reaction also results in the duplication of the target site, which is, however, in an inverted orientation (17). We carefully examined the sequences and found that the 8-bp flanking sequence (GTATAAAC) of the IS26 that bounded the 33.5-kb region is the exact inverted repeat of that (GTTTATAC) of the IS26 that bounded the 11-kb region (Fig. 2A). By artificially inverting the sequence between the two IS26 copies, it became clear that the IS26 that bounded the 11-kb region was inserted into a glutathione S-transferase family protein-encoding gene, which is also present on pCsaCS931b, with generation of the characteristic 8-bp DRs. Therefore, the 33.5- and the 11-kb extra regions on pMCR10_145005 were once clustered together but were separated due to inversion caused by an IS26-mediated intramolecular replicative reaction (Fig. 2A).

FIG 2.

Inversion due to IS26-mediated intramolecular replicative reaction and the alignment of att sites. (A) Inversion due to IS26-mediated intramolecular replicative reaction. In the 33.5-kb extra region (shown as a red line), int, int2, and mcr-10 are shown as a white arrow, a gray arrow, and a dark yellow arrow (not scaled), respectively. Attachment sites attL and attR of the integrative element are also indicated. At the top, on a proposed progenitor of pMCR10_145005, the single copy of IS26 targets and attacks a site (GTTTATAC) of a glutathione S-transferase family protein-encoding gene (shown as a yellow arrow) in the opposite strand (the dotted circle). The trans attack is shown by two dotted brown arrows. At the bottom, pMCR10_145005 is formed by the trans attack, which results in the inversion of the genetic component between the IS26 and the target site (the 11-kb extra region and the 14-kb intervening region here) and also generates a new copy of IS26 in the opposite orientation. The detailed scheme is depicted in Fig. 1 of reference 17. On pMCR10_145005, the target site is duplicated, but in an inverted orientation (17), as shown by the two yellow arrows. The 8-bp sequence GTATAAAC is the exact inverted repeat of the target site (GTTTATAC). Of note, the size is not scaled. (B) The alignment of att sites. The 16-bp sequence on the right is shown in lowercase. The unmatched nucleotides are depicted in red. The attB (present on pCsaCS931b) and attP (present on pMCR10_090065) sites are deduced from attL and attR.

On pMCR10_145005, mcr-10 was located upstream of insertion sequence ISEc36 and downstream of a site-specific integrase-encoding gene. The integrase-encoding gene, designated int here, encodes a phage-related integrase (Protein Analysis Through Evolutionary Relationships [PANTHER] Classification System no. PTHR30349), which consist of a core-binding (CB) domain and a catalytic (CAT) domain. The integrase is designated Int_MCR-10 here. Of note, there is another site-specific integrase-encoding gene, designated int2 here, on pMCR10_145005, which is located 43.3 kb downstream of mcr-10 (Fig. 2) and is also present on pCsaCS931b. The integrase Int2 encoded by int2 has 69% amino acid identity to Int_MCR-10.

By carefully comparing the sequences of pMCR10_145005 and pCsaCS931b, two 32-bp recombination sites were identified on pMCR10_145005: one at 5 bp upstream of the int start codon and the other at 85 bp upstream of the int2 start codon in the intervening sequence between the two extra regions (Fig. 2A). The two 32-bp sites are highly similar, with 29 identical nucleotides, and both consist of two imperfect 16-bp inverted repeats, with 12 of the 16 nucleotides being matched (Fig. 2B). Therefore, the two recombination sites are the attachment sites attL and attR of the site-specific integrase Int_MCR-10 or Int2. Phage integrases recognize the specific attachment site in the bacterial chromosome or plasmids, the attB site, and then recombine with the phage attachment site, attP, to generate the hybrid sites attL and attR containing half of the attB site and half of attP (18). Therefore, we were able to deduce the sequences of attB and attP based on attL and attR (Fig. 2B). The deduced attB sequence is identical to a 32-bp sequence present on pCsaCS931b, and the deduced attP sequence is identical to the recombination site of Int_MCR-10 on pMCR10_090065, the plasmid from which mcr-10 was discovered in the same hospital. The sequence comparison therefore suggests that mcr-10 might have been mobilized from a pMCR10_090065-like plasmid into a pCsaCS931b-like Cronobacter plasmid by site-specific recombination. Such a site-specific recombination could be mediated by either the integrase Int_MCR-10 or Int2. However, as the attB site (the recipient site) is on pCsaCS931b and the large region containing Int_MCR-10-encoding int was mobilized, the site-specific recombination is more likely due to the action of Int_MCR-10 rather than that of Int2. The element encoding the site-specific integrase Int_MCR-10 is therefore likely an integrative mobile element able to mobilize other genetic components, including antimicrobial resistance genes such as mcr-10. This warrants further verification by experiments. On pMCR10_145005, the integrative mobile element appears to consist of both the 33.5- and 11-kb extra regions. As there are multiple insertion sequences in the regions, whether genetic components such as the mrkABCDFJ locus were originally part of the integrative mobile element or were introduced by insertion sequences independently is unable to be determined.

In conclusion, we identified the presence of mcr-10 on a 120-kb IncFIB plasmid in a Cronobacter isolate, which has not been found before. The plasmid carries virulence genes, and the colocation of virulence genes and an antimicrobial resistance gene (mcr-10 in this case) is of clinical importance. mcr-10 was mobilized by an integrative mobile element via site-specific recombination mediated by an integrase.

Accession number(s).

The draft genome sequence of isolate 145005 and the complete sequence of pMCR10_145005 have been deposited in GenBank under accession no. JABTXY000000000 and JABTXY010000030, respectively.