A population genetics-based and phylogenetic approach to understanding the evolution of virulence in the genus Listeria

Abstract

The genus Listeria includes (i) the opportunistic pathogens L. monocytogenes and L. ivanovii, (ii) the saprotrophs L. innocua, L. marthii, and L. welshimeri, and (iii) L. seeligeri, an apparent saprotroph that nevertheless typically contains the prfA virulence gene cluster. A novel 10-loci multilocus sequence typing scheme was developed and used to characterize 67 isolates representing six Listeria spp. (excluding L. grayi) in order to (i) provide an improved understanding of the phylogeny and evolution of the genus Listeria and (ii) use Listeria as a model to study the evolution of pathogenicity in opportunistic environmental pathogens. Phylogenetic analyses identified six well-supported Listeria species that group into two main subdivisions, with each subdivision containing strains with and without the prfA virulence gene cluster. Stochastic character mapping and phylogenetic analysis of hly, a gene in the prfA cluster, suggest that the common ancestor of the genus Listeria contained the prfA virulence gene cluster and that this cluster was lost at least five times during the evolution of Listeria, yielding multiple distinct saprotrophic clades. L. welshimeri, which appears to represent the most ancient clade that arose from an ancestor with a prfA cluster deletion, shows a considerably lower average sequence divergence than other Listeria species, suggesting a population bottleneck and a putatively different ecology than other saprotrophic Listeria species. Overall, our data suggest that, for some pathogens, loss of virulence genes may represent a selective advantage, possibly by facilitating adaptation to a specific ecological niche.

FIG. 1.

Position of 10 MLST loci on the L. monocytogenes EGD-e chromosome (GenBank accession number AL591824).

FIG. 2.

Mixture of ancestry as inferred by the program STRUCTURE. (A) Proportions of ancestry from ancestral L. monocytogenes lineages I and II population (green), ancestral L. monocytogenes lineages III and IV population (yellow), ancestral L. marthii population (purple), the ancestral L. innocua population (red), the ancestral L. welshimeri population (light blue), the ancestral L. seeligeri population (orange), and the ancestral L. ivanovii population) as inferred by STRUCTURE assuming K = 8 ancestral populations. The asterisks mark hemolytic L. innocua isolates. Each vertical column represents an isolate and is colored according to the inferred proportion of single-nucleotide alleles that were derived from one of the ancestral subpopulations. (B to E) Posterior probabilities that an individual SNP allele is derived from one of the ancestral subpopulations within an individual isolate, including L. marthii FSL S4-120 (B), L. innocua FSL J1-023 (C), L. seeligeri FSL S4-015 (D), and L. ivanovii subsp. londoniensis FSL F6-596 (E). Colors of the columns indicate the ancestral subpopulation of a given SNP (see the color legend, above); the height of the color indicates the posterior probability (indicated on the y axis) that a given SNP is derived from a given ancestor. The numbers on the x axis represent the loci where these SNP alleles are found (1, addB; 2, ldh; 3, lmo0490; 4, lmo1555; 5, lmo2763; 6, pbpA; 7, polC; 8, prs; 9, rarA; 10, sigB).

FIG. 3.

Phylogram inferred by the maximum likelihood based on 10 concatenated loci. Bootstrap values of >60% are indicated above the branches. The bars summarize Bayesian support for the individual branches as found with the BEAST analysis of the individual loci, depicted as individual compartments of the bar (1, prs; 2, ldh; 3, lmo490; 4, sigB; 5, polC; 6, rarA; 7, lmo1555; 8, pbpA; 9, addB; 10, lmo2763); white indicates <95% PP and black indiates >95% PP.

FIG. 4.

Stochastic character mapping of gain/loss of the prfA cluster over the evolutionary history of Listeria. Solid branches indicate the presence of the prfA cluster, and white branches indicate the absence of the prfA cluster. This analysis shows that the most likely scenario for Listeria is one in which the MRCA of Listeria had the prfA cluster, and the cluster was lost on five separate occasions, including (i) once in the ancestral lineage of L. welshimeri, (ii) once in the ancestral lineage of L. marthii, (iii) twice within L. innocua, and (iv) a recent loss in L. seeligeri.

FIG. 5.

Comparison of hly (left) and 10 concatenated loci (right) ML phylograms of a subset of isolates of species with the prfA cluster. Values above the individual branches are bootstrap values based on 100 ML bootstrap replicates.