Comparative genomics of Pseudomonas syringae reveals convergent gene gain and loss associated with specialization onto cherry (Prunus avium)

Abstract

Genome-wide analyses of the effector- and toxin-encoding genes were used to examine the phylogenetics and evolution of pathogenicity amongst diverse strains of Pseudomonas syringae causing bacterial canker of cherry (Prunus avium), including pathovars P. syringae pv morsprunorum (Psm) races 1 and 2, P. syringae pv syringae (Pss) and P. syringae pv avii. Phylogenetic analyses revealed Psm races and P. syringae pv avii clades were distinct and were each monophyletic, whereas cherry-pathogenic strains of Pss were interspersed amongst strains from other host species. A maximum likelihood approach was used to predict effectors associated with pathogenicity on cherry. Pss possesses a smaller repertoire of type III effectors but has more toxin biosynthesis clusters than Psm and P. syringae pv avii. Evolution of cherry pathogenicity was correlated with gain of genes such as hopAR1 and hopBB1 through putative phage transfer and horizontal transfer respectively. By contrast, loss of the avrPto/hopAB redundant effector group was observed in cherry-pathogenic clades. Ectopic expression of hopAB and hopC1 triggered the hypersensitive reaction in cherry leaves, confirming computational predictions. Cherry canker provides a fascinating example of convergent evolution of pathogenicity that is explained by the mix of effector and toxin repertoires acting on a common host.

Understanding genomic diversity in Pseudomonas syringae throughout the forest and on the trees

“…While many previous studies have relied upon the metadata of pathovar designation as a metric for disease‐causing potential on a given host, Hulin et al . () build upon an additional recent publication from their group (Hulin et al ., ) in order to characterize virulence for all of the focal strains within one host under controlled conditions. They are therefore able to clearly designate strains as pathogenic or not, and utilize this information to identify genes whose presence and absence are statistically associated with virulence in cherry, while also incorporating close outgroup strains and statistically controlling for phylogenetic biases.…”

“…A recent publication from Hulin et al . (), in this issue of New Phytologist (pp. 672–696), provides excellent examples of the promise and complexities involved while performing genome comparisons of phytopathogens, but also provides a clear roadmap forward for designing experiments to overcome some of these challenges and link genomic changes to virulence phenotypes and host range evolution.

‘Just a few short years later, Hulin et al.

…”

“…672–696), provides excellent examples of the promise and complexities involved while performing genome comparisons of phytopathogens, but also provides a clear roadmap forward for designing experiments to overcome some of these challenges and link genomic changes to virulence phenotypes and host range evolution.

‘Just a few short years later, Hulin et al. () have beautifully demonstrated how to combine computational and wet‐lab based analyses to leverage genome sequencing to identify genes that contribute to host range changes in a model phytopathogen.’

…”

“…‘Just a few short years later, Hulin et al. () have beautifully demonstrated how to combine computational and wet‐lab based analyses to leverage genome sequencing to identify genes that contribute to host range changes in a model phytopathogen.’…”

“…Hulin et al . () start off by reporting the sequencing and assembly of numerous new genomes from strains that are pathogenic on sweet and wild cherry ( Prunus avium ). This collection samples multiple strains from within the pathovar morsprunorum (isolates of which can be differentially virulent on cherry and plum) but also incorporates comparisons of genomes from other pathovars ( avii and syringae ) that can cause disease on cherry (Fig.…”