Investigation of the atypical type III secretion system in Pseudomonas syringae strains using long read sequencing technology

Abstract

Ubiquitous across many environments, Pseudomonas syringae is a phytopathogen that is spread through the water cycle. With a broad host range, Pseudomonas syringae has been reported to cause outbreaks across many agricultural crops including potato, mango, and kiwifruits. Although this species is highly studied, interest in agricultural impacts has meant that sequencing efforts have focused primarily on virulent strains isolated from diseased plants.

In this work, we explored the genetic differences of a recently diverged, closely related and putatively non-virulent (i.e., not visibly detected by host plant) subphylogroup of P. syringae strains, subphylogroup 2c. Strains in this monophyletic group are a part of the larger virulent phylogroup 2 and more broadly grouped with other agriculturally virulent strains in phylogroups 1 and 3. Virulence of this bacterial species complex has been attributed to the genetic mobility of genes encoding type III effector proteins that are secreted by the type III secretion system. Gain or loss of these genes can cause P. syringae strains to switch between virulent and avirulent (i.e., detected by host plant and rendered not virulent) phenotypes. Many other niche adaption genes such as phytotoxin biosynthesis genes can also contribute to a virulence phenotype. We propose that the mechanisms contributing to the putative non-virulence phenotype of subphylogroup 2c may be elucidated via identification and examination of the divergent gene clusters within P. syringae strains.

To capture the full complement of genes, we sequenced and assembled complete genomes of two subphylogroup 2c strains, a biocontrol strain P. syringae pv. syringae 508, and a strain isolated from a healthy leaf, P. syringae TLP2. These strains were sequenced using Oxford Nanopore Technology's MinION long read sequencer to generate complete genomes without additional sequencing. Using one strain per R9.4 flowcell, we were able to generate >300x coverage for each strain with maximum read lengths exceeding 500kb and 250kb respectively allowing easy assembly of the full genome lengths. However, the ease-of-use of nanopore sequencing is hampered somewhat by the relatively high sequencing error rate of the reads, requiring extensive testing and optimization of recently developed software to maximize assembly quality. Despite optimization, many genes in the initial assembly remained fragmented. To address this issue, we developed the software tool Kastor that identifies and corrects draft assembly errors through comparisons to reference genomes and/or protein sequence information. Users can reference related genomes for more precise error identification, or use reference proteins for a more generalized identification method that doesn't rely on data from closely related species. For both the Pss508 and PsyTLP2 sequenced genomes, fewer than 3000 errors (comprising <0.05% of the assemblies) were corrected, however assembly completeness was increased from <85% to >99.8% as benchmarked using single copy orthologs. The completed assemblies were then annotated using NCBI's PGAP.

The annotated gene sets of the two P. syringae strains were then compared to 428 P. syringae strains across the species complex. We identified several divergent gene clusters in both sequenced and subphylogroup 2c genomes, with the most prominent diverging gene cluster set encoding the atypical variant of the type III secretion system (T3SS). Using PIRATE, we identified a related gene cluster in the distantly related phylogroup 13. With the analysis of both gene sequence and gene content, we propose that the atypical variant is a result of horizontal transfer and not divergence from the canonical variant in closely related strains. The horizontal acquisition of this T3SS variant in the most recent common ancestor of subgroup 2c likely triggered the divergence of the subgroup from group 2, the loss of common type III effector genes, and a likely niche/virulence change, resulting in putatively weakly-virulent strains.