Pervasive, genome-wide positive selection leading to functional divergence in the bacterial genus Campylobacter

Abstract

An open question in bacterial genomics is the role that adaptive evolution of the core genome plays in diversification and adaptation of bacterial species, and how this might differ between groups of bacteria occupying different environmental circumstances. The genus Campylobacter encompasses several important human and animal enteric pathogens, with genome sequence data available for eight species. We estimate the Campylobacter core genome at 647 genes, with 92.5% of the nonrecombinant core genome loci under positive selection in at least one lineage and the same gene frequently under positive selection in multiple lineages. Tests are provided that reject recombination, saturation, and variation in codon usage bias as factors contributing to this high level of selection. We suggest this genome-wide adaptive evolution may result from a Red Queen macroevolutionary dynamic, in which species are involved in competition for resources within the mammalian and/or vertebrate gastrointestinal tract. Much reduced levels of positive selection evident in Streptococcus, as reported by the authors in an earlier work, may be a consequence of these taxa inhabiting less species-rich habitats, and more unique niches. Despite many common loci under positive selection in multiple Campylobacter lineages, we found no evidence for molecular adaptive convergence at the level of the same or adjacent codons, or even protein domains. Taken collectively, these results describe the diversification of a bacterial genus that involves pervasive natural selection pressure across virtually the entire genome, with this adaptation occurring in different ways in different lineages, despite the species tendency toward a common gastrointestinal habitat.

Figure 1.

Campylobacter species tree. The maximum likelihood tree was obtained after concatenating the 511 congruent genes or gene fragments. On the branches of the tree are reported the percentage of gene tree support (black) as well as the number of genes found under positive selection (gray). (s/s) Substitution per site.

Figure 2.

Distribution of the positively selected genes in the 14 tested lineages of Campylobacter. The genes and lineages were sorted following a correspondence analysis. (Black dots) Genes under positive selection, (gray circles of different diameters, bottom) number of lineages under positive selection (PS) for a specific gene. (CCF) C. consisus, C. curvus, and C. fetus ancestral lineage; (JCU) C. jejuni, C. coli, and C. upsaliensis ancestral lineage; (CC) C. consisus and C. curvus ancestral lineage; (CCFH) C. consisus, C. curvus, C. fetus, and C. hominis ancestral lineage; (JC) C. jejuni and C. coli ancestral lineage; (JCUL) C. jejuni, C. coli, C. upsaliensis, and C. lari ancestral lineage.

Figure 3.

Relative rate of positive selection in Streptococcus and Campylobacter lineages. The rate unit is percentage of the core genome under positive selection per substitution per site.

Figure 4.

Positive selection (PS) distribution pattern in Campylobacter (A,C) and Streptococcus (B,D). (A,B) Lineage pairwise similarity indexes between the random model of selection, a random model with a set of genes devoid of positive selection, and the observed indexes. (C,D) Global pattern of selection distribution, with the observed pattern (black and thick lines), as well as simulated patterns with different settings (dashed lines). The random distributions with two categories of genes were set with 150 genes, five times and seven times less likely to be under positive selection for Campylobacter and Streptococcus, respectively.

Figure 5.

Number of significant positive selection aggregation tests with different sliding window size, with and without FDR correction, and using different BEB posterior probability cutoffs.

Figure 6.

Genome-wide positive selection pipeline.