9 July 2018
Mobile genetic elements (MGEs) that frequently transfer within and between bacterial species play a critical role in bacterial evolution, and often carry key accessory genes that associate with a bacteria’s ability to cause disease. MGEs carrying antimicrobial resistance (AMR) and/or virulence determinants are common in the opportunistic pathogen Klebsiella pneumoniae, which is a leading cause of highly drug-resistant infections in hospitals. Well-characterised virulence determinants in K. pneumoniae include the polyketide synthesis loci ybt and clb (also known as pks), encoding the iron-scavenging siderophore yersiniabactin and genotoxin colibactin, respectively. These loci are located within an MGE called ICE Kp, which is the most common virulence-associated MGE of K. pneumoniae, providing a mechanism for these virulence factors to spread within the population. Here we apply population genomics to investigate the prevalence, evolution and mobility of ybt and clb in K. pneumoniae populations through comparative analysis of 2498 whole-genome sequences. The ybt locus was detected in 40 % of K. pneumoniae genomes, particularly amongst those associated with invasive infections. We identified 17 distinct ybt lineages and 3 clb lineages, each associated with one of 14 different structural variants of ICE Kp. Comparison with the wider population of the family Enterobacteriaceae revealed occasional ICE Kp acquisition by other members. The clb locus was present in 14 % of all K. pneumoniae and 38.4 % of ybt+ genomes. Hundreds of independent ICE Kp integration events were detected affecting hundreds of phylogenetically distinct K. pneumoniae lineages, including at least 19 in the globally-disseminated carbapenem-resistant clone CG258. A novel plasmid-encoded form of ybt was also identified, representing a new mechanism for ybt dispersal in K. pneumoniae populations. These data indicate that MGEs carrying ybt and clb circulate freely in the K. pneumoniae population, including among multidrug-resistant strains, and should be considered a target for genomic surveillance along with AMR determinants.