Comment on: MDR carbapenemase-producing Klebsiella pneumoniae of the hypervirulence-associated ST23 clone in Poland, 2009–19

Biedrzycka and co-authors recently reported a novel MDR lineage of hypervirulent Klebsiella pneumoniae ST23 detected in Poland. The article highlighted features that are atypical compared with the hypervirulent ST23 clone, including resistance to last-line antimicrobials and distinct polysaccharide loci (KL57 versus KL1 capsule loci; O2v2 versus O1v2 O antigen loci). ST23 with KL57 (ST23-KL57) and AMR have been reported elsewhere in Europe, notably in a 2021 ECDC report describing an urgent public health inquiry. We would like to clarify the evolutionary history of ST23-KL57 strains, and highlight that despite sharing the same 7-locus MLST profile, ST23-KL57 and ST23-KL1 are actually two distinct lineages. Importantly, while ST23-KL1 are well known as the dominant lineage causing community-acquired hypervirulent infections globally (i.e. highly invasive infections that can occur in otherwise healthy hosts, commonly liver abscess with metastatic spread), there is currently no evidence to suggest that ST23-KL57 causes ‘hypervirulent’ disease. Whole-genome comparisons by Biedrzycka et al. demonstrated a clear segregation of ST23-KL1 from ST23-KL57 (nucleotide divergence of ∼0.42%). This divergence far exceeds the typical range of ST23 (mean 0.0045% divergence reported for n = 97 clonal group 23 genomes) and is similar to that reported between genomes belonging to distinct phylogenetic lineages of K. pneumoniae (∼0.5%). The separation becomes clearer when contextualized against a species-wide tree including 442 diverse genomes from the European Surveillance of Carbapenemase-Producing Enterobacteriaceae (EuSCAPE) study and 13 ST23-KL57 from Biedrzycka et al. ST23-KL57 are clearly clustered with ST218-KL57 (as noted by Biedrzycka et al.) and neither are clustered with ST23-KL1 [Figure 1(a)]. The distant relationship between these two ST23 lineages was also noted in the ECDC report, where one of the key take-home messages was the importance of ‘reliably differentiating these two clades’. We would like to take this opportunity to alert your readership to an existing core-genome MLST (cgMLST) scheme and associated sublineage (SL) and ‘cgLIN’ (Life Identification Number) code nomenclatures that readily distinguish ST23-KL57 from hypervirulent ST23-KL1. The cgMLST scheme encompasses 629 loci. Genomes differing by >190 cgMLST loci generally correspond to distinct phylogenetic lineages, are assigned as discrete SLs and can be distinguished by the third digit in the cgLIN code. We retrieved the cgMLST and cgLIN code information for the ST23-KL1 reference genome (SGH10, GCA_002813595.1) and a representative ST23-KL57 genome (Kp_Goe_154414, GCA_001902335.1), which identified them as cgST3913 and cgST3381, respectively. These cgSTs differ by 518 alleles and are assigned to distinct sublineages, SL23 and SL218. Their cgLIN codes are 0_0_429_0_29_0_0_0_0_0 and 0_0_115_0_0_0_0_1_0_0, respectively; notably the uniqueness of the third digit in the cgLIN codes indicates they are as distinct from one another as are any two phylogenetic lineages of K. pneumoniae. This highlights that 7-locus MLST alone should not be used to conclude two K. pneumoniae are clonally related. Notably, a surprising number of alleles in the 7-locus MLST scheme appear in multiple STs. Using the profile of ST23 as an example, the alleles gapA_2, infB_1, mdh_1, pgi_1, phoE_9, rpoB_4 and tonB_12 appear in 276 to 2756 STs each (as of November 2022). Given the prevalence of these alleles and frequency of recombination that occurs within the species, there is a non-trivial possibility of a lineage evolving independently to share the same 7-locus ST profile of another unrelated lineage, as observed here with ST23. To clarify the evolutionary history of ST23-KL57 (cgST3381, SL218) we used the publicly available cgMLST database to explore its relationship to other lineages. We found that cgST3381 (SL218) shares 502 of 629 alleles with cgST1872 (also SL218); this corresponds to ST218 in the 7-locus scheme, differing from ST23 at a single locus, infB. Of the 127 loci that differentiate the SL218 strains cgST3381 and cgST1872, cgST3381 shares 111 loci with cgST2547 (SL395; ST395 in the 7-locus scheme, sharing infB_1 with ST23). We identified a representative genome for SL395 (GCA_900508075.1, EuSCAPE_TR085) and plotted the distribution of pairwise nucleotide variants along the chromosome [Figure 1(b)]. This analysis supports that cgST3381 (SL218, ST23-KL57) is likely the recombinant progeny of a cgST1872-like (SL218) and cgST2547-like (SL395) K. pneumoniae [Figure 1(b)]. Neither SL218 nor SL395 are closely related to ST23-KL1 (SL23) in