Diversity and evolution of surface polysaccharide synthesis loci in Enterobacteriales

Kathryn E Holt,Rafał J Mostowy,Florent Lassalle,Ryan Wick,Kelly L Wyres

doi:10.1038/s41396-020-0628-0

Abstract

Bacterial capsules and lipopolysaccharides are diverse surface polysaccharides (SPs) that serve as the frontline for interactions with the outside world. While SPs can evolve rapidly, their diversity and evolutionary dynamics across different taxonomic scales has not been investigated in detail. Here, we focused on the bacterial order Enterobacteriales (including the medically relevant Enterobacteriaceae), to carry out comparative genomics of two SP locus synthesis regions, cps and kps, using 27,334 genomes from 45 genera. We identified high-quality cps loci in 22 genera and kps in 11 genera, around 4% of which were detected in multiple species. We found SP loci to be highly dynamic genetic entities: their evolution was driven by high rates of horizontal gene transfer (HGT), both of whole loci and component genes, and relaxed purifying selection, yielding large repertoires of SP diversity. In spite of that, we found the presence of (near-)identical locus structures in distant taxonomic backgrounds that could not be explained by recent exchange, pointing to long-term selective preservation of locus structures in some populations. Our results reveal differences in evolutionary dynamics driving SP diversity within different bacterial species, with lineages of Escherichia coli, Enterobacter hormaechei and Klebsiella aerogenes most likely to share SP loci via recent exchange; and lineages of Salmonella enterica, Cronobacter sakazakii and Serratia marcescens most likely to share SP loci via other mechanisms such as long-term preservation. Overall, the evolution of SP loci in Enterobacteriales is driven by a range of evolutionary forces and their dynamics and relative importance varies between different species.

Highlights

Polysaccharide capsules and lipopolysaccharides (LPS) with an O antigen, here broadly called surface polysaccharides (SPs), are the most diverse bacterial cell surface structures
Most locus types and SP locus gene families (LGFs, homology groups identified in SP loci and clustered at 50% amino acid identity; see “Methods”) were species specific: in the cps region, 90% of locus types and 61% of LGFs were found in a single species, while in the kps region it was 93% and 78%, respectively
Across the Enterobacteriales population, SP locus genes are overall less conserved than other genes—see Supplementary Fig. S8— this approach is unlikely to result in false-positive calls of recombinational exchange.) We considered locus-type-sharing lineage pairs in which the SP loci rank in the top 5% most similar CDS for that pair as likely resulting from horizontal exchange of the SP locus between lineages relatively recently; and those locus-type-sharing pairs in which the SP loci rank in the bottom 60% most similar CDS for that pair as evidence of absence of recent horizontal exchange; the remaining cases we considered unresolved