Sir, The discovery of the plasmid-mediated colistin resistance gene mcr-1, first identified in 2015 among both clinical and animal isolates in China, raised concerns about pan-resistant bacteria.1 A plethora of publications quickly followed, suggesting mcr-1 was already established in many countries and across many continents, in various bacterial species, on a variety of plasmids types, and in bacteria isolated from diverse animal species, environmental sources and human health settings.2 Furthermore, several other mcr-like genes and gene variants have since been discovered.2,3 Colistin is among a diminishing group of antimicrobial agents available for the treatment of serious infections caused by extensively resistant Gram-negative bacteria. In April 2018 we conducted a small retrospective study comparing two commercial broth microdilution (BMD) methods [Sensititre (ThermoFisher) and SensiTest (Liofilchem)] for the determination of colistin susceptibility, against 23 Enterobacterales [19 colistin resistant (MIC ≥4 mg/L) by Phoenix™ (BD Diagnostics, USA) and 4 intrinsically resistant clinical isolates], randomly collected between 2015 and 2018. BMD according to EUCAST guidelines was used as the reference method.4 Detection of mcr-1 and -3 genes was performed by PCR.1,3 Here we describe the co-occurrence of mcr-1 and mcr-3 genes in a single Escherichia coli. The patient was a female in her late sixties with a spinal cord injury and a history of recurrent urinary tract infections (UTIs), who self-catheterized. She presented at a community healthcare facility, with symptoms of UTI, in October 2017. A urine sample was sent to Canterbury Health Laboratories, Christchurch, for analysis. The culture grew a coliform that was identified as E. coli using MALDI-TOF MS (Bruker Daltonics, USA). Susceptibilities were determined using Phoenix™ and EUCAST breakpoints. The isolate, designated CHL5009T, expressed resistance to multiple classes of antibiotics, but remained susceptible to fosfomycin, ciprofloxacin, carbapenems, tigecycline and piperacillin/tazobactam (Table 1). MIC values for and resistance genes found in E. coli CHL5009T NT, not tested. MIC determined by Phoenix with the NMIC-404 panel except for colistin reference BMD, which was determined by ESR by BMD according to ISO-20776. MIC values in bold indicate resistant or non-susceptible. The presence of other resistance mechanisms such as overexpression of TEM-1 and ampC, porin or efflux mutations can promote increased resistance to β-lactams. qnrS1 confers decreased susceptibility to quinolones; no gyr or par gene mutations associated with quinolone resistance were found. MIC values for and resistance genes found in E. coli CHL5009T NT, not tested. MIC determined by Phoenix with the NMIC-404 panel except for colistin reference BMD, which was determined by ESR by BMD according to ISO-20776. MIC values in bold indicate resistant or non-susceptible. The presence of other resistance mechanisms such as overexpression of TEM-1 and ampC, porin or efflux mutations can promote increased resistance to β-lactams. qnrS1 confers decreased susceptibility to quinolones; no gyr or par gene mutations associated with quinolone resistance were found. WGS was performed on E. coli CHL5009T using the Nextera DNA Flex microbial colony extraction procedure and library prepared using the Nextera DNA Flex library kit (Illumina, Australia), and sequenced using a micro flow cell (reagent kit v2, 300 cycles). Paired-end FASTQ reads were quality trimmed and assembled into contigs using Shovill v0.9.0 (https://github.com/tseemann/shovill). Bioinformatic analyses including MLST, plasmid replicon typing and detection of resistance genes were performed using mlst software v2.11 (https://github.com/tseemann/mlst) and ABRicate v0.8.2 (https://github.com/tseemann/abricate) with PlasmidFinder and ResFinder databases. To characterize the plasmids, DNA was further extracted using Pellet Paint (Merck) and long-range sequences were generated using a MinION SpotON flow cell (R9.4) (Oxford Nanopore). The FASTQ reads were extracted with Albacore v2.3.1. The assembled contigs were further corrected with Illumina short reads using Unicycler v0.4.4 and annotated using prokka v1.12.5 The plasmid contigs were matched to known plasmids in GenBank (NCBI) using BLAST and gene synteny was determined around the resistance genes using Mauve (build 2015–02-13), ISFinder (https://isfinder.biotoul.fr/), Galileo AMR (https://galileoamr.arcbio.com/mara/) and CLC genomic workbench v9.5 (QIAGEN). E. coli CHL5009T was determined to be ST101 by MLST with five replicons (IncFII, IncI1, IncI2, IncY and ColRNAI) detected by PlasmidFinder. The genomes of the E. coli and the seven plasmids determined by hybrid assembly (short and long reads) can be found in GenBank (accession numbers CP032936.1–CP032943.1); only pCHL5009T-94k (CP032938.1) was incomplete. The mcr-1 gene was located on a 61805 bp plasmid designated pCHL5009T-61.8k-mcr1 (GenBank accession number CP032940.1), containing an IncI2 replicon signature. The closest match in GenBank was pABC149-MCR-1 (KX013538.1). ISApl1-mcr-1-pap2 is inserted downstream of nikAB,6,7 with 2 bp DRs of GA, as described by Snesrud et al.8 for pMR0716-mcr1.6–8 The mcr-3 and blaCTX-M-55 genes were located on a 101806 bp plasmid designated pCHL5009T-102k-mcr3 (GenBank accession number CP032937.1), with IncFII replicon type and a blastn alignment match of 99% (15 differences) nucleotide identity in GenBank to 83 435 bp plasmid pHNHN21 (KX246267.1). There are two major regions of difference between the plasmids; additional 5.7 and 12.7 kb regions were found in pCHL5009T-102k, indicating significant plasmid rearrangement due to insertions of mobile genetic elements (Figure 1a). Annotation of the pCHL5009T plasmid. (a) Complete plasmid map showing location of the 5.7 and 12 kb regions. Blue, other CDS/genes; black, conserved plasmid genes (conjugation etc.); green, mobile genetic elements; yellow, antibiotic resistance genes; and red, 5 and 12 kb regions not present in pHNHN21. (b) Relationship between plasmids with the mcr-3 gene flanked by IS26. The pCHL5009T 12 kb region has integrated within Tn2 downstream of blaCTX-M-55. Mobile genetic elements were annotated by Galileo AMR software. The 5.7 kb region matches a similar segment from plasmid pH226B (KX129784.1, nucleotide positions 94667–100372) and includes a resolvase and a single copy of a transposase; the best match using ISFinder is 80% nucleotide identity to ISPsy42 (Tn3 family transposon). This 5.7 kb putative transposon has 35 bp terminal inverted repeats (IRs) (GGGGTTGTAAGCCGGAACCGCCGAAATTTCCGTCA) similar to IRs of ISPsy42 with 5 bp DRs (ATTAT) as shown in Figure 1(a). Comparatively, the 5.7 kb similar region in pH226B does not have complete copies of these repeats. The presence of the 5 bp DRs indicates that this 5.7 kb putative transposon has inserted into a pHNHN21-like plasmid to produce pCHL5009T-102k. The 12.7 kb region, in which mcr-3 is flanked by IS26 elements, is inserted downstream of blaCTX-M-55 within Tn2 (Figure 1b). It is closely related to most of a 13.5 kb region in pHN8 (MG780294.1), apart from lacking an 826 bp region that contains a bleomycin resistance gene and 7 nt differences. A qnrS1 gene is present in both plasmids, but in a different location relative to mcr-3, and the 12.7 kb region in pCHL5009T-102k-mcr3 includes a partial duplication of Tn2. This suggests incorporation of a circular molecule derived from a pHN8-like plasmid into a pHNHN21-like plasmid by recombination in Tn2. IS26, previously suggested to be involved in in mobilizing mcr-3, could have been responsible for creating such a circular molecule.9 Other antibiotic resistance genes include blaTEM-1B, blaCTX-M-55, aac(3)-IId, qnrS1 and sul3 (Table 1) and the class 1 integron cassette array dfrA12-gcuF-aadA2-cmlA1-aadA1a-qacl found on pCHL5009T-94k (CP032938.1, nucleotide positions 28986–33661). Subsequent inquiry revealed the patient had travelled to Thailand 4 months prior to her UTI. Both E. coli ST101 and CTX-M-55 are infrequently found in New Zealand. However, blaCTX-M-55 and mcr-3 have links to Thailand, with CTX-M-55 first described in clinical specimens from Thailand,10 an MCR-3-producing Klebsiella pneumoniae identified by Yin et al.3 from a human pus sample, and an E. coli harbouring pHN8 with mcr-3 described by Wang et al.9 from a healthy pig in China. Furthermore, a CTX-M-55-producing E. coli that co-produced mcr-3 was isolated from a bloodstream infection of a Danish patient who had recently travelled to Thailand.11 Interestingly, co-occurrence of mcr-1 and mcr-3 from clinical isolates has rarely been reported: one Salmonella isolate (also harbouring blaCTX-M-55) and one E. coli (also harbouring blaNDM-5).2 To the best of our knowledge, this is the first report of plasmid-mediated colistin-resistance genes in New Zealand, with both mcr-1 and mcr-3 co-occurring in a CTX-M-55-producing E. coli. We also describe the inclusion of mcr-3 within a uniquely arranged plasmid, pCHL5009T-102k-mcr3. The origin of this isolate is unknown, although it was most likely acquired by the patient in Thailand. The true prevalence of mcr-type genes in New Zealand remains to be determined. Thank you to the staff at the Antibiotic Reference Laboratory, ESR, for performing the molecular characterization of the study isolates and the BMD tests on them. This study was completed as part of our routine work. The work at the Institute for Environmental Science and Research (ESR) was funded by the New Zealand Ministry of Health. None to declare.
No takes yet. Share an insight, caveat, or question.
Creighton et al. (2019) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: