Comparative genomics of host-specific virulence in Pseudomonas syringae

Abstract

While much study has gone into characterizing virulence factors that play a general role in disease, less work has been directed at identifying pathogen factors that act in a host-specific manner. Understanding these factors will help reveal the variety of mechanisms used by pathogens to suppress or avoid host defenses. We identified candidate Pseudomonas syringae host-specific virulence genes by searching for genes whose distribution among natural P. syringae isolates was statistically associated with hosts of isolation. We analyzed 91 strains isolated from 39 plant hosts by DNA microarray-based comparative genomic hybridization against an array containing 353 virulence-associated (VA) genes, including 53 type III secretion system effectors (T3SEs). We identified individual genes and gene profiles that were significantly associated with strains isolated from cauliflower, Chinese cabbage, soybean, rice, and tomato. We also identified specific horizontal gene acquisition events associated with host shifts by mapping the array data onto the core genome phylogeny of the species. This study provides the largest suite of candidate host-specificity factors from any pathogen, suggests that there are multiple ways in which P. syringae isolates can adapt to the same host, and provides insight into the evolutionary mechanisms underlying host adaptation.

Figure 1.—

Comparative genomic microarray data for 91 P. syringae strains. Strains are arrayed by their evolutionary relationships as determined by MLST (Sarkar and Guttman 2004; Hwang et al. 2005) (neighbor-joining phylogenetic analysis and bootstrap scores from 1000 replicates are presented above each node). The host of each strain is presented after the strain designation. See Table 5 for strain details. The five major P. syringae phylogroups are designated g1–g5. PtoDC3000, which was used as the reference strain for the microarray design, falls into group 1. Numbers in boxes are clade numbers (see text). The solid and open grid represents the microarray data for genomic DNA hybridized against 353 VA genes. Solid, present; open, divergent; dark and light shading, present and divergent with lower confidence, respectively. See materials and methods for details. The VA genes are arranged by functional groups (Table 1). Supplemental Table 10 at http://www.genetics.org/supplemental/ presents the data in a more accessible form.

Figure 2.—

Presence/divergence data for T3SEs. Strains are arranged on the basis of their host of isolation and the family of that host. T3SEs are arranged by frequency among the strains. See Figure 1 for information on the shading.

Figure 3.—

Genetic variation and genome organization of VA genes ordered along the PtoDC3000 chromosome. The percentage of variation for each VA gene is plotted according to the genome location of its PtoDC3000 ortholog. The x-axis shows the genome location in megabases; the y-axis shows the variation among strains (0%, no variation and present in all strains). Triangles, effectors; circles, hrp/hrc genes; ×, all other VA genes. T3SS, fla, yer, gsp, and cor approximate locations of the genes encoding T3SS apparatus, flagella apparatus, yersiniabactin biosynthesis, general secretory pathway, and coronatine biosynthesis. pA and pB represent the two endogenous plasmids of PtoDC3000.