Patterns of genome evolution that have accompanied host adaptation in Salmonella

Abstract

Many bacterial pathogens are specialized, infecting one or few hosts, and this is often associated with more acute disease presentation. Specific genomes show markers of this specialization, which often reflect a balance between gene acquisition and functional gene loss. Within Salmonella enterica subspecies enterica, a single lineage exists that includes human and animal pathogens adapted to cause infection in different hosts, including S. enterica serovar Enteritidis (multiple hosts), S. Gallinarum (birds), and S. Dublin (cattle). This provides an excellent evolutionary context in which differences between these pathogen genomes can be related to host range. Genome sequences were obtained from ∼ 60 isolates selected to represent the known diversity of this lineage. Examination and comparison of the clades within the phylogeny of this lineage revealed signs of host restriction as well as evolutionary events that mark a path to host generalism. We have identified the nature and order of events for both evolutionary trajectories. The impact of functional gene loss was predicted based upon position within metabolic pathways and confirmed with phenotyping assays. The structure of S. Enteritidis is more complex than previously known, as a second clade of S. Enteritidis was revealed that is distinct from those commonly seen to cause disease in humans or animals, and that is more closely related to S. Gallinarum. Isolates from this second clade were tested in a chick model of infection and exhibited a reduced colonization phenotype, which we postulate represents an intermediate stage in pathogen-host adaptation.

Keywords: Salmonella; host adaptation; metabolism; pseudogene.

Fig. 1.

Chromosome-based phylogenetic relationships. Midpoint-rooted maximum-likelihood phylogenetic tree based upon the chromosome. Branch lengths are determined by number of SNPs. Numbers on the tree indicate how many pseudogenes were identified on that branch. Colored strain names show serotype: S. Enteritidis, blue; S. Pullorum, gold; S. Gallinarum, red; S. Dublin, green. Metadata columns: Clade, phylogenetic grouping of S. Enteritidis isolates; ST, sequence types; pVIR, virulence plasmid; stf–sth, fimbriae; SPI-6 and -19, Salmonella pathogenicity islands; i, ii, SPI-6 partial hits represent two distinct versions at this locus; iii, SPI-19 partial presence is different from classic S. Enteritidis in this strain. Partial presence indicates one or more genes have been lost.

Fig. 2.

Functions lost through pseudogene formation. (A) Functional classification of pseudogene sets. All, pseudogenes shared by all strains in the phylogeny (98); Gal/Pul, shared by S. Gallinarum and S. Pullorum (80); Gal, Pul, Dub, remaining pseudogenes present in all strains of S. Gallinarum (151), S. Pullorum (132), and S. Dublin (95), respectively. (B) Venn diagram showing the distribution of metabolic pathway and transport loss between S. Gallinarum, S. Pullorum, and S. Dublin, irrespective of causative pseudogenes. (C) Schematic depicting interconnectivity of pseudogene-affected pathways and transport systems. Processes inside the dotted line only occur anaerobically. Operons involved are shown in red.

Fig. 3.

Infection of the avian host by nonclassic S. Enteritidis isolates. Invasion of Salmonella strains into the (A) spleen, (B) liver, and (C) colonization of chick ceca 7 d (except S. Gallinarum 287/91, 5 d, separated by dashed line) postinfection (n = 5). Solid lines represent means. *P < 0.05, **P < 0.01. classic, classic clade isolate; Ent, S. Enteritidis; GAL, S. Gallinarum; second, second clade isolate.