Given a shared toolbox of pathogenicity-related genes among a set of species, why is one species a biotroph and specialist while another is a necrotroph and generalist? Is it the result of selection on primary sequence, or on proteins, or alternatively, differences in the timing and magnitude of gene expression? The Sclerotiniaceae (Ascomycota, Leotiomycetes, Helotiales) is a relatively recently evolved family of fungi whose members include host generalists and host specialists, and the spectrum of trophic types. Based on a phylogeny inferred from three, presumably evolutionarily conserved housekeeping genes, the common ancestor of the Sclerotiniaceae was necrotrophic, with at least two shifts from necrotrophy to biotrophy. Phylogenies inferred from eight pathogenicity-related genes, involved in cell wall degradation and the oxalic acid pathway, were incongruent with the presumably neutral phylogeny. Site-specific likelihood analyses, which estimate the rate of nonsynonymous to synonymous substitutions (dN/dS), showed evidence for purifying selection acting on all pathogenicity-related genes, and positive selection on sites within five of eight genes. Rate-specific likelihood analyses showed no differences in dN/dS rates between necrotrophs and biotrophs, and between host generalists and host specialists, indicating that selection acting on the genes does not drive divergence toward changes in trophic type or host association. In vitro screens for oxalic acid production demonstrated that all necrotrophic generalists produce oxalic acid by 72 hours, while production was either absent or delayed among biotrophs and host specialists. This pattern was also observed during the course of Arabidopsis thaliana infection, in which large spikes of expression were seen in the oxalic acid pathway-related gene, oah, within eight hours of inoculation among necrotrophic generalists only. Results suggest that necrotrophic generalists can be distinguished from biotrophs and host specialists in the Sclerotiniaceae by the ability to produce abundant amounts of oxalic acid early in infection and to cause large proliferating lesions on A. thaliana.

Given a shared toolbox of pathogenicity-related genes among a set of species, why is one species a biotroph and specialist while another is a necrotroph and generalist? Is it the result of selection on primary sequence, or on proteins, or alternatively, differences in the timing and magnitude of gene expression? The Sclerotiniaceae (Ascomycota, Leotiomycetes, Helotiales) is a relatively recently evolved family of fungi whose members include host generalists and host specialists, and the spectrum of trophic types. Based on a phylogeny inferred from three, presumably evolutionarily conserved housekeeping genes, the common ancestor of the Sclerotiniaceae was necrotrophic, with at least two shifts from necrotrophy to biotrophy. Phylogenies inferred from eight pathogenicity-related genes, involved in cell wall degradation and the oxalic acid pathway, were incongruent with the presumably neutral phylogeny. Site-specific likelihood analyses, which estimate the rate of nonsynonymous to synonymous substitutions (dN/dS), showed evidence for purifying selection acting on all pathogenicity-related genes, and positive selection on sites within five of eight genes. Rate-specific likelihood analyses showed no differences in dN/dS rates between necrotrophs and biotrophs, and between host generalists and host specialists, indicating that selection acting on the genes does not drive divergence toward changes in trophic type or host association. In vitro screens for oxalic acid production demonstrated that all necrotrophic generalists produce oxalic acid by 72 hours, while production was either absent or delayed among biotrophs and host specialists. This pattern was also observed during the course of Arabidopsis thaliana infection, in which large spikes of expression were seen in the oxalic acid pathway-related gene, oah, within eight hours of inoculation among necrotrophic generalists only. Results suggest that necrotrophic generalists can be distinguished from biotrophs and host specialists in the Sclerotiniaceae by the ability to produce abundant amounts of oxalic acid early in infection and to cause large proliferating lesions on A. thaliana.

Les champignons phytopathogènes sont des parasites majeurs des espèces végétales, autant naturelles que domestiquées. Botrytis cinerea, l’agent de la pourriture grise, en infecte plus de 1400 et est ainsi considéré comme un pathogène généraliste. Pourtant, des données récentes ont mis en évidence une structure des populations liée à leur hôte d’origine. Cette observation soulève l’hypothèse d’une spécialisation à l’hôte, à l’œuvre chez une espèce généraliste. Ce modèle d’étude pourrait permettre de faire avancer la connaissance des mécanismes évolutifs en jeu dans la divergence précoce des populations et la formation de nouvelles espèces fongiques. Cette thèse de doctorat a pour objectifs : (i) de démontrer formellement la spécialisation à l’hôte dans les populations de B. cinerea et d’en déterminer la magnitude et (ii) d’identifier les déterminants génomiques de cette spécialisation. Ainsi j’ai étudié la structure des populations par l’analyse de microsatellites d’un échantillon de 683 isolats, que nous avons corrélé à des tests de pathogénicité croisés sur différents hôtes. Ces travaux ont permis de mettre en évidence la spécialisation de B. cinerea aux hôtes tomate et vigne. En complément de ces lignées sélectionnées, l’espère Botrytis cinerea est constituée d’individus généralistes capables de coloniser les autres hôtes. Des méthodes d’inférence de structure et de généalogies sur des données de polymorphisme issues du séquençage de 32 individus nous ont permis de mieux définir la structure des populations ainsi que d’identifier une lignée spécialisée à la tomate. Enfin, des tests de McDonald-Kreitman et des méthodes de scans génomiques pour détecter des balayages sélectifs ont permis de mettre en évidence des gènes soumis à la sélection divergente entre les populations spécialisées, révélant de possibles déterminants de cette spécialisation. Ces travaux sont ainsi une base pour la validation de plusieurs gènes impliqués dans la pathogénicité hôte-spécifique de B. cinerea, et ouvrent la voie à des améliorations de la gestion du réservoir d’hôtes et des pratiques culturales contre la pourriture grise. / Phytopathogenic fungi are major parasites to wild or domesticated plant species. The grey mold fungus, Botrytis cinerea, infects more than 1400 plant species and thus is considered a broad generalist. However, recent data have revealed population structure correlated to the host of origin of isolates. This observation raises the hypothesis of ongoing host specialization in a generalist species. Studying this question could greatly deepen our theoretical knowledge of the evolutionary mechanisms involved in the early stages of population divergence and subsequent speciation. This thesis aims (i) to formally demonstrate the host specialization in B. cinerea’s populations and determine its magnitude, and (ii) to identify the genomic determinants of this specialization. Thus, I studied population structure based on 683 isolates characterized using microsatellite markers. We compared the inferred genetic structure with variations in aggressiveness measured through cross-pathogenicity tests on multiple hosts. These experiments and analyses confirmed the specialization of B. cinerea to tomato and grapevine hosts. Besides these specialized lineages, the species B. cinerea is composed of generalist individuals capable of infecting multiple hosts. I sequenced the whole genome of 32 individuals and characterized nucleotide polymorphism. Structure inference and genomic genealogy methods allowed us to more accurately define the population structure and identify a lineage specialized on tomato. Lastly, McDonald-Kreitman tests and genomic scans methods allowed the identification of genes under divergent natural selection between populations, revealing possible genomic determinants of specialization. This work can serve as foundation for the validation of multiple genes involved in host-specific pathogenicity of B. cinerea, and pave the way for the implementation of efficient strategies for managing pathogen reservoirs and new agricultural practices for controlling grey mold.