Novel Subclone of Carbapenem-Resistant Klebsiella pneumoniae Sequence Type 11 with Enhanced Virulence and Transmissibility, China

We aimed to clarify the epidemiologic and clinical importance of evolutionary events that occurred in carbapenem-resistant Klebsiella pneumoniae (CRKP). We collected 203 CRKP causing bloodstream infections in a tertiary hospital in China during 2013–2017. We detected a subclonal shift in the dominant clone sequence type (ST) 11 CRKP in which the previously prevalent capsular loci (KL) 47 had been replaced by KL64 since 2016. Patients infected with ST11-KL64 CRKP had a significantly higher 30-day mortality rate than other CRKP-infected patients. Enhanced virulence was further evidenced by phenotypic tests. Phylogenetic reconstruction demonstrated that ST11-KL64 is derived from an ST11-KL47–like ancestor through recombination. We identified a pLVPK-like virulence plasmid carrying rmpA and peg-344 in ST11-KL64 exclusively from 2016 onward. The pLVPK-like–positive ST11-KL64 isolates exhibited enhanced environmental survival. Retrospective screening of a national collection identified ST11-KL64 in multiple regions. Targeted surveillance of this high-risk CRKP clone is urgently needed.

Single nucleotide polymorphisms (SNPs) were called using an in-house pipeline with SAMtools and bcftools, and a pseudo-genome alignment was generated. Recombined regions were detected using Gubbins (14). A phylogenetic tree was constructed using the SNPs outside of the recombination regions with RAxML using a GTR model and gamma correction (15). An ST258 genome was also included in a separate phylogenetic analysis, using the same methods, to establish the root of the ST11 tree. The phylogenetic tree and associated metadata were visualized together in Microreact (16) BactDating (17) was used to perform regression analysis of the root-to-tip genetic distance against sampling time for the 154 ST11 isolates sequenced in this study and to obtain a dated phylogeny based on a Bayesian approach. For both, the output from Gubbins was used as the input, together with the isolation dates. Default settings were used for the Bayesian function, except that the MCMC chain length was increased to 1 million. The MCMC convergence was tested by examining traces of the model parameters. The effective sample sizes of the model parameters were also determined using the R package, "coda," and these were >200.

Screening of ST11-KL47 and ST11-KL64 Isolates
We retrospectively screened 1,098 clinical BSI-KP strains to detect ST11-KL47 and according to a previous study (18). The positive isolates were further tested by MLST using the scheme of Institute Pasteur as described previously (19).

S1-PFGE and Southern Blotting
The plasmids and location of rmpA and rmpA2 genes were determined by S1-nuclease digestion and pulsed-field gel electrophoresis (S1-PFGE), followed with southern blotting hybridizations as previously described (20). Briefly, bacterial DNA was prepared in agarose blocks and was digested by XbaI nuclease. The digested DNA fragments were separated using PFGE with conditions of 14 h at 6 V/cm, 14°C, with a pulse angle of 120° and a switch time from 1 to 10s. The separated DNA fragments were transferred to nylon membranes (Hybond N, Amersham, UK), hybridized with digoxigenin-labeled rmpA or rmpA2 probes and detected using an NBT/BCIP color detection kit (Roche, Basel, Switzerland).

Detection of Virulence Plasmids
Detection of virulence plasmids was mainly dependent on BLASTn as described previously (21). The contigs of each genome were blasted against the reference plasmid and plotted by BLAST Ring Image Generator (BRIG). Two circularized virulence plasmids, pVir-KP16932 and pVir-KP47434, were used as references. If the alignment coverage was higher than 80% and nucleotide similarity higher than 90%, a reference-like virulence plasmid was supposed to be existing.

Statistical Analysis
The Student's t-test (for normally distributed variables) or Mann-Whitney U test (for variables that are not normally distributed) was performed to evaluate continuous variables. strength of all associations that emerged was determined using odds ratios (ORs) and 95% confidence intervals (CIs). Two-tailed tests were used to determine statistical significance.
Variables with a P-value ≤0.05 in the univariate analysis were used in cox regression for multivariate analysis to identify independent predictors. Kaplan-Meier product limit method was used to estimate the hazard ratios, and nonparametric (log rank and Wilcoxon) tests were used to compare hazard ratios in different groups. In all analyses, P-values ≤0.05 were considered significant. All statistical analyses were carried out by using the SPSS Version 23.0 (IBM Corporation, Armonk, NY, USA) and SAS v9.4 (SAS institute, Cary, NC, USA).