Comparative genomic insights into the epidemiology and virulence of plant pathogenic pseudomonads from Turkey

Abstract

Pseudomonas is a highly diverse genus that includes species that cause disease in both plants and animals. Recently, pathogenic pseudomonads from the Pseudomonas syringae and Pseudomonas fluorescens species complexes have caused significant outbreaks in several agronomically important crops in Turkey, including tomato, citrus, artichoke and melon. We characterized 169 pathogenic Pseudomonas strains associated with recent outbreaks in Turkey via multilocus sequence analysis and whole-genome sequencing, then used comparative and evolutionary genomics to characterize putative virulence mechanisms. Most of the isolates are closely related to other plant pathogens distributed among the primary phylogroups of P. syringae, although there are significant numbers of P. fluorescens isolates, which is a species better known as a rhizosphere-inhabiting plant-growth promoter. We found that all 39 citrus blast pathogens cluster in P. syringae phylogroup 2, although strains isolated from the same host do not cluster monophyletically, with lemon, mandarin orange and sweet orange isolates all being intermixed throughout the phylogroup. In contrast, 20 tomato pith pathogens are found in two independent lineages: one in the P. syringae secondary phylogroups, and the other from the P. fluorescens species complex. These divergent pith necrosis strains lack characteristic virulence factors like the canonical tripartite type III secretion system, large effector repertoires and the ability to synthesize multiple bacterial phytotoxins, suggesting they have alternative molecular mechanisms to cause disease. These findings highlight the complex nature of host specificity among plant pathogenic pseudomonads.

Keywords: Pseudomonas fluorescens; Pseudomonas syringae; bacterial diseases; phytotoxins; plant pathogens; type III secreted effectors.

Fig. 1.

Sampling sites and disease metadata for the 169 Pseudomonas outbreak strains collected in Turkey between 1996 and 2018. Six additional Pseudomonas outbreak strains were also included from Germany (four), Holland (one) and Switzerland (one). Pie charts are proportional to the number of isolates collected at each site and illustrate the distribution of isolates that cause different diseases in the corresponding regions.

Fig. 2.

Evolutionary relationships between the 175 Pseudomonas strains isolated from diseased hosts in this study based on concatenated MLSA sequences of the gapA, gltA, gyrB and rpoD genes. Phylogroups were assigned based on the clustering of strains with representatives from 11 P. syringae phylogroups and 9 P. fluorescens phylogroups (Figs S2 and S3). The tree was rooted at the base of the P. syringae species complex and the tree scale reflects the number of stubstitutions per site. All alignments were generated with muscle and the tree was generated using FastTree, with an SH-Test branch support cut-off of 50 %.

Fig. 3.

Population structure analysis conducted using structure v2.3.4 to assign all strains to population genetic clusters. For each analysis, the number of population clusters (k) was optimized using the Puechmaille method [93]. Bar plot panels indicate the clustering coefficients for replicate structure runs generated by clumpak [65] for three independent analyses comprising the unitig patterns of: the full pangenome, the core genome (100 % presence across strains) and the accessory genome (<100 % presence). The optimum k is six for the full pangenome, seven for the core genome and seven for the accessory genome. There is no relationship between the cluster colours of the three different structure analyses.

Fig. 4.

T3SS, T3SE and phytotoxin repertoires for each of the 58 representative Pseudomonas strains that were whole-genome sequenced. The phylogenetic tree was generated from a concatenated core-genome amino acid alignment using FastTree, with an SH-Test branch support cut-off of 50 %. The tree was rooted at the base of the P. syringae species complex and the tree scale reflects the number of stubstitutions per site. The six T3SSs analysed include the T-PAI from P. syringae PtoDC3000, the R-PAI from P. syringae Pph1448a, the S-PAI from P. syringae PchICMP3353, the A-PAI from P. syringae PcoICMP19117, the F-PAI from P. fluorescens PgeBBc6R8 and the C-PAI from P. corrugata PcoF113. The presence–absence of 70 established P. syringae T3SEs in each genome was used to quantify the collective effector repertoires in each strain. A phytotoxin was considered present if more than 50 % of the known protein sequences involved in the synthesis pathway had significant tblastn hits (1×10⁻⁵, >80 % identity) in the genome.

Fig. 5.

Genetic architecture of the T3SSs identified in this study. The genome architectures for each of the T3SSs was drawn from the following representative genomes: (a) T-PAI – P. syringae PtoDC3000; (b) R-PAI – P. syringae Pph1448a; (c) S-PAI – P. syringae PchICMP3353; (d) A-PAI – P. syringae PcoICMP19117; (e) F-PAI – P. fluorescens PgeBBc6R8; (f) C-PAI – P. corrugata PcoF113. All genes and non-coding regions are to scale.

Fig. 6.

Complete T3SE repertoires for each of the 58 Pseudomonas strains that were whole-genome sequenced in this study. The phylogenetic tree was generated from a concatenated core-genome amino acid alignment using FastTree, with an SH-Test branch support cut-off of 50 %. The tree was rooted at the base of the P. syringae species complex and the tree scale reflects the number of stubstitutions per site. The 70 established P. syringae T3SE families that we delimited in an earlier study are listed on the left of the plot. A filled box indicates that at least one T3SE from the family is present in the strain and an empty box indicates that the T3SE is absent.