IMP-68, a Novel IMP-Type Metallo-β-Lactamase in Imipenem-Susceptible Klebsiella pneumoniae

IMP-type metallo-β-lactamases comprise one group of the “Big 5” carbapenemases. Here, a novel blaIMP-68 gene encoding IMP-68 (harboring a Ser262Gly point mutant of IMP-11) was discovered from meropenem-resistant but imipenem-susceptible Klebsiella pneumoniae TA6363. The Ser262Gly substitution was previously identified as important for substrate specificity according to a study of other IMP variants, including IMP-6. We confirmed that IMP-68 exhibited weaker imipenem-hydrolyzing activity than that for other carbapenems, demonstrating that the antimicrobial susceptibility profile of TA6363 originated from IMP-68 substrate specificity, with this likely to affect treatment strategies using antibacterial agents in clinical settings. Notably, the carbapenem resistance conferred by IMP-68 was undetectable based on the MIC of imipenem as a carbapenem representative, which demonstrates a comparable antimicrobial susceptibility profile to IMP-6-producing Enterobacteriaceae that previously spread in Japan due to lack of awareness of its existence.

IMP-6, is weak (2)(3)(4). IMP-1 Ser262 is replaced with glycine in these variants, and Enterobacteriaceae encoding IMP-6 frequently show susceptibility to imipenem. This feature caused the spread of bla IMP-6 -harboring plasmids in Japan as determined after screening the antibiotic susceptibility of infectious bacteria to imipenem as a representative carbapenem (5)(6)(7). Here, we report a novel variant, IMP-68, which is a point mutant of IMP-11 corresponding to the Ser262Gly substitution.
Klebsiella pneumoniae TA6363 was isolated from ascites obtained from a hospitalized male patient suffering from peritonitis at the National Center of Global Health and Medicine in 2016. The patient was suspected to be domestically infected by this pathogen, because he had not been abroad from Japan within at least 90 days. This study was approved by the ethics committee of the Tokyo Metropolitan Institute of Public Health. TA6363 was initially determined to be resistant to meropenem by MicroScan walkaway 96 SI (Beckman Coulter, Brea, CA, USA) using the MicroScan Neg Combo EN 1T test card (Beckman Coulter), which employs meropenem as a carbapenem representative. However, this strain was subsequently found to be susceptible to imipenem using the dry strip method using Etest (bioMérieux, La Balme-Les-Grottes, France) (Table 1). TA6363 was positive for a modified carbapenem-inactivation method (8) and found to produce metallo-␤-lactamase using the double-disk synergy test involving meropenem and sodium mercaptoacetate (Eiken Chemical, Tokyo, Japan) disks. Pulsed-field gel electrophoresis using an S1-nuclease-digested DNA plug (S1-PFGE) showed that TA6363 carried three plasmids (Fig. 1), with the second larger one (pTMTA63632; ϳ90 kbp) harboring the novel bla IMP gene according to sequencing plasmids extracted from the S1-PFGE gel (MiSeq; Illumina, San Diego, CA, USA) (9, 10). We found that this bla IMP gene encoded a Ser262Gly point mutant of IMP-11 metallo-␤-lactamase, and we named this novel variant "bla IMP-68 " (see Fig. S1 in the supplemental material).
We tested the influence of IMP-68 on antibiotic resistance to ␤-lactams by transformation experiments. We cloned the open reading frame of bla IMP-68 into a chloramphenicol-resistant pHSG398 vector (Takara Bio, Shiga, Japan) at the EcoRI-KpnI site and used it to transform Escherichia coli DH5␣ cells (Takara Bio). Additionally, we tested bla IMP-11 for comparison, and conjugation transfer was tested using E. coli J53, as previously described (9). The MICs of ␤-lactams were determined by the dry strip method using Etest. To determine whether IMP-68 production was detectable by several phenotypic tests other than the modified carbapenem-inactivation method, TA6363 and the E. coli J53 transconjugant were applied to the modified Hodge (11), Carba NP (12), and Blue-Carba (13) tests.
To purify IMP-11 and IMP-68, the nucleotide sequence for the restriction site of the PreScission protease (GE Healthcare, Chicago, IL, USA) was added to the N terminus of bla IMP-11 and bla IMP-68 after the signal peptide region (first 19 amino acids) by PCR, followed by cloning into the pETBK vector (BioDynamics, Tokyo, Japan) for expression in E. coli Rosetta II cells (Merck-Millipore, Darmstadt, Germany). His-tagged proteins were purified using a nickel-nitrilotriacetic acid (Ni-NTA) resin (Qiagen, Hilden, Germany) and digested by PreScission protease to remove the His tag. Kinetics experiments were performed by spectrophotometry (14)(15)(16), with initial hydrolysis rates for each ␤-lactam measured using a Biomate 3 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) at 298 nm for imipenem and meropenem, 260 nm for ceftazidime and cefotaxime, and 235 nm for ampicillin at 24 Ϯ 1°C in 50 mM sodium phosphate buffer (pH 7.0) supplemented with 50 M ZnCl 2 . K m and k cat values were determined by the Michaelis-Menten equation. Experiments were performed in triplicate. Table 1 shows the MICs for TA6363 and E. coli transformants. Although TA6363 was resistant to meropenem, doripenem, and ertapenem, the MIC of imipenem for TA6363 was classified as susceptible (S) according to Clinical and Laboratory Standards Institute criteria (12). E. coli transformation by bla IMP-68 alone significantly increased the MICs of ␤-lactams, and the substrate specificity of IMP-68 and IMP-11 differed. The lower MIC of imipenem for IMP-68 than that for IMP-11 was similar to the relationship between IMP-6 and IMP-1, where the MIC of imipenem for IMP-6 is lower than that for IMP-1 (2). Furthermore, the bla IMP-68 -carrying native plasmid pTMTA63632 was transferred by conjugation from TA6363 to E. coli J53, resulting in enhanced antibiotic resistance via pTMTA63632. TA6363 and the transconjugant were positive for all the abovementioned tested phenotypic methods. Despite the use of imipenem, the Carba NP and Blue-Carba tests were sensitive enough to detect IMP-68 production.
Several differences in substrate specificity between IMP-11 and IMP-68 identified during antibiotic susceptibility testing were consistent with the results of kinetics experiments ( Table 2). The k cat /K m values of IMP-68 against imipenem were lower than those of IMP-11, whereas IMP-68 exhibited higher meropenem-hydrolyzing activity. Additionally, the k cat /K m values of IMP-68 against ceftazidime were lower than those of IMP-11. These differences in substrate specificity between IMP-68 and IMP-11 correlated with those between IMP-6 and IMP-1 (2).
Additionally, genomic DNA extracted from TA6363 was sequenced on MiSeq and  (20), with the threshold of nucleotide coverage and identity used for Inc-type identification established at 96% and 98%, respectively (21). We obtained four circular sequences (Table S1) corresponding to a chromosome and three plasmids ( Fig. 1 and Fig. S2). TA6363 was classified as ST268, and pTMTA63632 was determined to be an IncL/M plasmid. Chromosomal mutations corresponding to antimicrobial resistance to carbapenems (ompK36, ompK37, and acrR) and fluoroquinolones (gyrA and parC) were not found (Table S1). Class 1-integron-harboring bla IMP-68 in pTMTA63632 (Fig. 1), which was assigned to In1702 in the INTEGRALL database (22), was similar to previously reported IMP genes harboring class 1 integrons, as the gene cassettes were located between intI1 and qacE⌬1-sul1 (10,23,24). Specifically, In1702 was compared with the class 1 integrons harbored by pNUH14_ECL028_1 and pIMP-A2015-49 plasmids previously found in Japan (24), which carried IMP-1 and IMP-11, respectively (Fig. S3). The inclusion of conjugation transfer genes by pTMTA63632 was consistent with the conjugation experiment performed using E. coli J53 ( Table 1). The higher MICs of ampicillin and piperacillin for the transconjugant than for bla IMP-68 alone ( Table 1) likely originated from the effect of bla TEM-1 in pTMTA63632. TA6363 was resistant to both amikacin and gentamicin, whereas the transconjugant was resistant only to amikacin. This difference was attributable to the carriage of aminoglycoside resistance genes by pTMTA63632 [aac(6=)-Ia] and pTMTA63633 [aac(3)-IId, aph(3=)-Ib, and aph(6)-Id] (Table S1) (25); namely, pTMTA63632 conferred antimicrobial resistance to amikacin, whereas pTMTA63633 was additionally necessary for the gentamicin resistance.
In conclusion, IMP-68 should be noted as a novel carbapenemase that does not sufficiently confer resistance to imipenem on Enterobacteriaceae due to the Ser262Gly substitution from IMP-11. IMP-68 production would have been missed if the MIC of imipenem had been used to investigate the carbapenemase-producing Enterobacteriaceae. The diversity of substrate specificity among IMP-type enzymes might affect treatment plans using antibacterial agents in clinical settings.