Prevalence and Distribution of mcr Genotypes in a Large Retrospective Collection of Clinical Carbapenemase-Producing Enterobacterales, Singapore

LETTER

Carbapenem-resistant Gram-negative pathogens have become a major health care burden, limiting treatment options to agents such as colistin and tigecycline in combination with other antibiotics (1). Following the discovery of mobile colistin resistance (mcr) encoded by mcr-1 in 2016, other variants, mcr-2 to mcr-10 (2–9), have since emerged, although mcr-1 is still by far the most prevalent globally. Cocarriage of mcr and carbapenemases tends to be infrequent (10, 11), but such bacteria are of significant concern due to their multidrug resistance traits. Here, we determined the prevalence of mcr genotypes within the genomes of a large clinical carbapenemase-bearing Enterobacterales collection.

The assembled genomes (12) of 1,251 carbapenemase-bearing Enterobacterales were screened for all known mcr gene variants. These characterized genomes were from a retrospective collection (years 2011 to 2015) of phenotypically carbapenem-resistant clinical Enterobacteriaceae deposited at the reference National Public Health Laboratory, Singapore (12), and had been previously screened for the five major carbapenemases genes encoding KPC, NDM, VIM, IMP, and OXA-48-like (12). Colistin broth microdilution was performed in accordance with the CLSI reference method (13), and MIC was determined based on EUCAST guidelines (EUCAST 2016) (14).

mcr variants were detected in 50/1,251 (4%) isolates. The most common mcr-positive specimen type was rectal swabs, 35/50 (70%) (see the supplemental material). bla_KPC was the carbapenemase genotype most commonly associated with mcr (25/50, 50%). The mcr genes were distributed across 6 species, viz., Escherichia coli, Escherichia fergusonii, Enterobacter cloacae, Citrobacter freundii, Klebsiella quasipneumoniae subsp. similipneumoniae, and Klebsiella quasipneumoniae subsp. quasipneumoniae. E. coli was found to be significantly associated with mcr-1 positivity (P < 0.05). This observation has also been noted in another mcr-1 prevalence study (15). Interestingly, within the Klebsiella genus, mcr was not identified in K. pneumoniae sensu stricto; instead mcr-9 and mcr-10 were detected in K. quasipneumoniae subsp. quasipneumoniae and K. quasipneumoniae subsp. similipneumoniae. A cluster of eight K. quasipneumoniae subsp. quasipneumoniae strains with the sequence type 256 (ST256) bearing mcr-10 was identified (see the supplemental material). The isolates were from rectal swab specimens with the exception of one specimen from abdominal fluid. All ST256 isolates originated from the same hospital, suggesting nosocomial transmission, although we did not further investigate epidemiological linkage.

mcr-1.1 was the most prevalent (15/50, 30%), followed by mcr-10.1 (13/50, 26%) (supplemental material). Dual mcr genes were detected as combinations of mcr-4.3 and mcr-9.1 (3/50, 6%) as well as mcr-9.1 and mcr-10.1 (2/50, 4%). The cooccurrence of two different mcr variants, though uncommon, has been described previously (16).

mcr-2, mcr-5, mcr-6, mcr-7, and mcr-8 were noticeably absent. Intensive animal farming and livestock are reservoirs of colistin-resistant organisms and mcr genes (17). In Singapore, there is almost no agricultural activity, and this may account for the lack of certain mcr genotypes. mcr-6 has been identified only in Moraxella pluranimalium and has not disseminated into the Enterobacterales family (18).

Of the 50 mcr-bearing isolates, 12 (24%) were phenotypically colistin susceptible with colistin MICs of ≤2 mg/L. This included all the eight mcr-9 isolates in the study, three mcr-4.3 isolates (GenBank accession no. MG026621.1), and one mcr-3.4 isolate (GenBank accession no. MG026622). mcr-3.4 and mcr-4.3 have been previously reported as nonfunctional due to the presence of a premature stop codon and missense mutations, respectively (19). The remaining isolates exhibited colistin resistance, with MICs of >2 mg/L. No chromosome-mediated colistin resistance mechanisms were detected (20).

MCR-9 confers only reduced colistin susceptibility, not resistance (8). Expression of mcr-9 comes under the regulation of QseB/QseC two-component regulators. We found that most of our mcr-9.1 genes (85%) did not associate with QseB/QseC regulators, possibly accounting for the lack of phenotypic colistin resistance.

“Classic” mobile elements associated with mcr dissemination were identified. These included an ISApl1-flanked classic transposon, Tn6330, in mcr-1 (5/15) isolates, although most of the mcr-1 isolates (66.7%, 10/15) appeared to follow the natural evolution of globally described mcr-1 structures, with deletions of ISApl1 from the ancestral Tn6330 (21). mcr-4.3 was present on a 12,808-bp nonconjugative ColE plasmid identical to the mcr-4.3 plasmid described in Enterobacter kobei strain IB2020 (GenBank accession no. CP059482.1). mcr-4 tends be carried on a conserved prototypical ColE plasmid (19, 22). mcr-9.1 was found in variable genetic contexts typically bracketed by two insertion sequences and accompanied downstream by wbuC, qseC, qseB, and an ATPase gene (8), while mcr-10 was found adjacent to a XerC-type tyrosine recombinase which was likely to mediate the mobilization of mcr-10 via site-specific recombination.

Although our in silico screening indicated a higher mcr positivity rate than that in reports from China at 0.78% (11) and Thailand at 0.28% (10), our data suggest that the presence of mcr does not always confer colistin resistance, as a quarter of our mcr-positive isolates were not phenotypically colistin resistant. The clinical microbiology laboratory may find molecular detection of the entire repertoire of mcr variants unnecessary.