Recovery of FRI-5 carbapenemase at a Japanese hospital where FRI-4 carbapenemase was discovered

Sir, Carbapenemase-producing Enterobacteriaceae is a major global concern. French imipenemases (FRIs) are recently discovered class A carbapenemases, first isolated in Europe.1,2 We recently reported a novel carbapenemase variant of FRI, FRI-4,3 in an Enterobacter cloacae complex isolated from a patient in Japan, the first of its kind outside Europe. Here we report the isolation of an E. cloacae complex producing a new FRI variant, FRI-5, from a patient at the hospital where the FRI-4 carrier had been admitted.3 This study was part of a clinical study approved by the Keio University School of Medicine Ethics Committee (20160300); informed consent was waived owing to the retrospective nature of this study. In September 2015, the E. cloacae complex strain Keio-CL-MIC-002 was isolated from a faecal sample of a haemodialysis patient admitted to Keio University Hospital, where the patient with FRI-4-producing E. cloacae complex TA6166 (isolated in April 2016) was admitted to a different ward from 2015 to 2016. The strain was resistant to imipenem and meropenem (Table S1, available as Supplementary data at JAC Online), had a positive result for the Carba NP test and modified carbapenem inactivation method (mCIM), and a negative result for the modified Hodge test.4 Further, the strain was negative for all five major carbapenemase genes (blaKPC, blaNDM, blaVIM, blaOXA-48-like and blaIMP-1) using the Xpert Carba-R (Cepheid, Sunnyvale, CA, USA) assay. Genomic characterization of the E. cloacae complex Keio-CL-MIC-002 was performed by WGS on a MiSeq DNA sequencer (Illumina, San Diego, CA, USA). Assembled contigs obtained by SPAdes3.95 from the Center for Genomic Epidemiology server (https://cge.cbs.dtu.dk/services/) were analysed by ANI calculator (http://enve-omics.ce.gatech.edu/ani/index), which identified Keio-CL-MIC-002 as Enterobacter asburiae. The MLST identified in silico6 was ST252, which is different from other FRI-producing strains (STs of FRI-2- and FRI-4-producing strains were ST8292 and ST250, respectively, where the ST for the FRI-4-producing strain was our own unpublished data3). In addition, the acquired antimicrobial resistance genes analysed by Resfinder7 at 90% nucleotide identity as threshold value were blaACT-1, qnrE1 and fosA. However, the carbapenemase genes were not detected. Thus, we compared assembled contigs with the reference E. asburiae L1 using GenomeMatcher.8 A BLAST search was performed for contigs unmatched against the reference sequence (cover rate <0.1), and an array resembling blaFRI-1 was identified. The sequence termed blaFRI-5, encoding FRI-5 (Figure 1a), was 88.8% identical to blaFRI-1 and 87.5% identical to blaFRI-4. The closest FRI variant to FRI-5 was FRI-4 (84.9% amino acid identity). (a) Comparison of the amino acid sequences of French imipenemase (FRI)-type carbapenemases. The signal peptide cleavage site is indicated by an arrowhead. Consensus residues are shown by asterisks. Conserved motifs among class A β-lactamases are shown in bold. Conserved residues among class A carbapenemases are shaded. (b) Circular map of pKeio-CL-MIC-002-3. The ORFs are shown with the direction of strands. Two inner blocks indicate the shared regions with pTMTA616613 or pJF-5872 with ≥80% nucleotide identity. PFGE using an S1 nuclease-digested DNA plug of Keio-CL-MIC-002 was performed,3 and the sizes of three detected plasmids were ∼110, 170 and 230 kb (Figure S1). Because we could not obtain the complete sequences of these plasmids directly from Keio-CL-MIC-002 by the hybrid assembly of MiSeq and MinION (Oxford Nanopore, Oxford, UK) data, Escherichia coli DH5α was transformed with plasmids extracted from Keio-CL-MIC-002 under selection by 100 mg/L ampicillin in order to separate the FRI-5-carrying plasmid. We obtained E. coli DH5α transformed with the smallest plasmid (100 kb) termed pKeio-CL-MIC-002-3 (Figure S1). pKeio-CL-MIC-002-3 was then completely sequenced by the hybrid assembly of MiSeq and MinION data of the transformed cell using Unicycler (v. 0.4.7).9 ResFinder7 showed that this plasmid harboured only the blaFRI-5 gene among antimicrobial resistance genes found in WGS (the ResFinder search was performed after GenBank accession of blaFRI-5). PlasmidFinder 2.010 did not find any replicon in pKeio-CL-MIC-002-3 over 98% nucleotide identity, whereas a sequence which was only 87.7% identical to the replicon corresponding to IncFII(Yp) was detected. The upstream region of blaFRI-5 encodes the transcriptional regulator FriR (Figure 1b), similar to that observed with other FRI carbapenemase-encoding plasmids.1–3 However, the nucleotide sequence of pKeio-CL-MIC-002-3 was completely different from those of pJF-587 (FRI-2)2 and pTMTA61661 (FRI-4)3 (Figure 1b). The most similar sequence on the NCBI BLAST site was of E. cloacae strain M12X01451 plasmid pM12X01451 (accession number CP017473). However, the query coverage was 15% (the region encoding chromosome partitioning proteins only), suggesting that pKeio-CL-MIC-002-3 is a rare type of plasmid. The genes involved in conjugative transfer were not detected. Additionally, the antimicrobial resistance to β-lactams conferred by blaFRI-5 was examined. In addition to the transformation with pKeio-CL-MIC-002-3, E. coli DH5α was also transformed by a recombinant pHSG398 vector (Takara Bio, Shiga, Japan) with an insert of the ORF of blaFRI-5, as described previously.3 MICs of β-lactams for the two transformants were measured by Etest (BioMérieux, La Balme-Les-Grottes, France). MICs for both transformants were significantly higher than those for the recipient (Table S1). The higher MIC of imipenem compared with other carbapenems, including meropenem, for transformants was similar to the previous reports on FRI-1,1 FRI-22 and FRI-4.3 This is the second report of a novel FRI carbapenemase detected outside Europe. The fact that two novel FRI carbapenemase-producing organisms were discovered from the same hospital and during the same period does not suggest a carry-over transmission in the hospital because the nucleotide sequences of FRI-4- and FRI-5-encoding plasmids were completely different. However, these observations suggest that FRI variants are more widespread and might be missed during detection. In conclusion, a novel FRI-type carbapenemase, FRI-5-producing E. cloacae complex, was discovered from a hospital in Japan. The nucleotide sequence of blaFRI-5 was deposited in GenBank (MH208723), and the WGS data of Keio-CL-MIC-002 (DRX171180) and the nucleotide sequence of pKeio-CL-MIC-002-3 (LC467982) were deposited in the DNA Data Bank of Japan. This work was supported by the Japanese Ministry of Education, Culture, Sports, Science, and Technology (KAKENHI grant; 16K19618) and Precision System Science Co. Ltd. None to declare.