Comparative phylogenomics and evolution of the Brucellae reveal a path to virulence

Abstract

Brucella species include important zoonotic pathogens that have a substantial impact on both agriculture and human health throughout the world. Brucellae are thought of as "stealth pathogens" that escape recognition by the host innate immune response, modulate the acquired immune response, and evade intracellular destruction. We analyzed the genome sequences of members of the family Brucellaceae to assess its evolutionary history from likely free-living soil-based progenitors into highly successful intracellular pathogens. Phylogenetic analysis split the genus into two groups: recently identified and early-dividing "atypical" strains and a highly conserved "classical" core clade containing the major pathogenic species. Lateral gene transfer events brought unique genomic regions into Brucella that differentiated them from Ochrobactrum and allowed the stepwise acquisition of virulence factors that include a type IV secretion system, a perosamine-based O antigen, and systems for sequestering metal ions that are absent in progenitors. Subsequent radiation within the core Brucella resulted in lineages that appear to have evolved within their preferred mammalian hosts, restricting their virulence to become stealth pathogens capable of causing long-term chronic infections.

FIG 1

Phylogenetic analysis of 42 Brucellaceae genomes. The maximum-parsimony tree is based on 193,760 SNPs. The early-diverging Brucella strains are clearly differentiated from the classic Brucella strains, with 2,672 SNPs unique to the classic Brucella strains and 1,172 SNPs unique to the outer clade (strains NF2653, 83/13, BO1, and BO2). The tree was rooted with Ochrobactrum spp. as outgroups. All branches have 100% support unless otherwise noted.

FIG 2

Phylogenetic analysis, based on maximum parsimony of the core Brucella genomes, showing a rapid radiation following the divergence of B. microti, resulting in six separate clades, with Brucella sp. strain NVSL07-0026 and B. suis bv. 5 as possible separate clades. The tree was rooted with B. microti as the outgroup based on results from Fig. 1. All branches have 100% support unless otherwise noted.

FIG 3

Phylogenetic analysis based on shared protein families in Brucella, showing the presence (colored circles) or absence (open circles) of specific genomic islands that are not universally shared across all 40 genomes (as in Table 1).

FIG 4

The first half of the Brucellaceae pan-proteome with protein families oriented by either B. microti (A) or O. anthropi (B), generated using the Protein Family Sorter tool at PATRIC. Black cells (columns, protein families; rows, genomes) indicate no annotated proteins, yellow indicates a single protein, and orange to red indicate increasing numbers of proteins annotated in that specific genome.

FIG 5

Location of genomic regions of interest across a “typical” Brucella strain of two chromosomes. Dark blue bands indicate that a genome or clade has all the genes in the region present. A lighter band indicates that a genome or clade is missing some or all of the genes in a region. Absence of a band shows that the genome or clade does not have this region. These data should be cross-referenced with column 2 of Table 1.

FIG 6

A model for the evolution of virulence in the genus Brucella. The phylogenetic tree is not drawn to scale.