Core genome multilocus sequence typing (cgMLST) applicable to the monophyletic Klebsiella oxytoca species complex

Abstract

The environmental bacterium Klebsiella oxytoca displays an alarming increase of antibiotic-resistant strains that frequently cause outbreaks in intensive care units. Due to its prevalence in the environment and opportunistic presence in humans, molecular surveillance (including resistance marker screening) and high-resolution cluster analysis are of high relevance. Furthermore, K. oxytoca previously described in studies is rather a species complex (KoSC) than a single species comprising at least six closely related species that are not easily differentiated by standard typing methods. To reach a discriminatory power high enough to identify and resolve clusters within these species, whole genome sequencing is necessary. The resolution is achievable with core genome multilocus sequence typing (cgMLST) extending typing of a few housekeeping genes to thousands of core genome genes. CgMLST is highly standardized and provides a nomenclature enabling cross laboratory reproducibility and data exchange for routine diagnostics. Here, we established a cgMLST scheme not only capable of resolving the KoSC species but also producing reliable and consistent results for published outbreaks. Our cgMLST scheme consists of 2,536 core genome and 2,693 accessory genome targets, with a percentage of good cgMLST targets of 98.31% in 880 KoSC genomes downloaded from the National Center for Biotechnology Information (NCBI). We also validated resistance markers against known resistance gene patterns and successfully linked genetic results to phenotypically confirmed toxic strains carrying the til gene cluster. In conclusion, our novel cgMLST enables highly reproducible typing of four different clinically relevant species of the KoSC and thus facilitates molecular surveillance and cluster investigations.

Keywords: Klebsiella oxytoca; cgMLST; molecular surveillance; nosocomial infections; resistance markers; species complex; virulence markers.

Fig 1

MST of 28 isolates from a nosocomial outbreak investigation (30). The distance is based on 2,536 target genes with pairwise ignore missing values. Isolates are colored according to the species. Clustering of outbreak strains is visible (MST cluster 1). Also, two environmental isolates from the same location cluster together (cluster 2) and the remaining environmental and reference strains of this study are distant to all other strains, but still cluster on species level (as indicated by the color of nodes).

Fig 2

Minimum spanning tree of 21 isolates from a neonatal intensive care unit in Graz from 2010 to 2013 (4, 7), and controls based on 2,536 target genes with pairwise ignore missing values. The close relatedness of all outbreak strains in MST cluster 1 is visible. Strains mentioned in the first outbreak investigation are colored cyan, isolates from the second one are colored blue. Control strains from either investigation are marked with the respective faded color.