Niche-specificity and the variable fraction of the Pectobacterium pan-genome.

We compare genome sequences of three closely related soft-rot pathogens that vary in host range and geographical distribution to identify genetic differences that could account for lifestyle differences. The isolates compared, Pectobacterium atrosepticum SCRI1043, P. carotovorum WPP14, and P. brasiliensis 1692, represent diverse lineages of the genus. P. carotovorum and P. brasiliensis genome contigs, generated by 454 pyrosequencing ordered by reference to the previously published complete circular chromosome of P. atrosepticum genome and each other, account for 96% of the predicted genome size. Orthologous proteins encoded by P. carotovorum and P. brasiliensis are approximately 95% identical to each other and 92% identical to P. atrosepticum. Multiple alignment using Mauve identified a core genome of 3.9 Mb conserved among these Pectobacterium spp. Each core genome is interrupted at many points by species-specific insertions or deletions (indels) that account for approximately 0.9 to 1.1 Mb. We demonstrate that the presence of a hrpK-like type III secretion system-dependent effector protein in P. carotovorum and P. brasiliensis and its absence from P. atrosepticum is insufficient to explain variability in their response to infection in a plant. Additional genes that vary among these species include those encoding peptide toxin production, enzyme production, secretion proteins, and antibiotic production, as well as differences in more general aspects of gene regulation and metabolism that may be relevant to pathogenicity.