Search

Grasslands are crucial ecosystems that provide various ecosystem services and are a key component in livestock production. In these systems, Italian ryegrass (Lolium multiflorum Lam.), is among the most widely used species due to its high yield, fast ground cover, and great nutritional quality. Its high adaptation to mowing makes it a species of choice for fodder in the form of hay and silage. One of the main diseases that affect L. multiflorum is bacterial wilt, caused by Xanthomonas translucens pv. graminis (Xtg), which can lead to severe yield losses. To date, breeding for resistant cultivars is the only feasible way for an efficient and durable control of the disease. Cultivars with increased resistance have been obtained by recurrent phenotypic selection, but due to the outbreeding nature of ryegrasses, cultivars are highly heterozygous, and susceptibility still occurs. Precisely identifying genetic loci involved in Xtg resistance in order to specifically target them in genomics-assisted breeding would be crucial to develop new cultivars that are highly resistant to Xtg. Furthermore, identifying molecular components that are involved in the host-pathogen interaction is essential to better understand the disease and develop new ways to control infection. Therefore, this thesis aimed at identifying resistance genes in L. multiflorum and genes involved in pathogenicity in Xtg by using advanced genomic-based analyses. Chapter 1 starts with a general introduction about the L. multiflorum – Xtg pathosystem and the state of the art about their host-pathogen interaction. It then details the currently available means to control the disease and the importance of genomics-based tools to better understand this interaction and identify sources of resistance in order to breed for resistant cultivars. Chapter 2 describes a first comparative analysis within X. translucens aimed at characterizing the species and identifying pathovar-specific features. Indeed, like most Xanthomonas species, X. translucens is divided into pathovars, defined by their host range. The species contains pathovars that infect cereals, as well as pathovars that infect forage grasses, like Xtg. In this study, we produced complete genome sequences for all eleven pathotype strains of X. translucens using long read sequencing technology, enabling a comprehensive comparative genomic analysis within the species. This revealed that the species is genetically separated into three distinct clades, with Xtg being closely related to the four other forage grasses-infecting pathovars. A total of 2,181 genes were found to be shared by all pathovars, while 190, 588 and 168 genes were exclusive to each clade, respectively. The sequenced Xtg strain was found to have the largest set of genes in the species, and a high number of pathovar-specific genes. Furthermore, the Xtg strain lacked many virulence features commonly present in other strains of the species, highlighting the particularity of Xtg. This study laid the foundations for further in-depth analyses within the X. translucens species. Chapter 3 reports the production of four additional complete genome sequences for Xtg strains, allowing a comprehensive comparative genomic analysis of Xtg. This revealed that while Xtg strains were highly related, with 99.9 to 100% average nucleotide identity, they displayed a high genome plasticity, with many chromosomal rearrangements between strains. This plasticity is likely due to the presence of many mobile genetic elements, with 413 to 457 insertion/excision elements identified per Xtg strain compared to 24 to 123 in other pathovars of the species. Genes that were specific to Xtg were identified, including XopE and XopX class effectors, a unique set of minor pilins of the type IV pilus, 17 TonB-dependent receptors, and gumP, a gene associated with the gum xanthan synthesis gene cluster. These constitute a set of potential virulence factors that could be involved in the pathogenicity of Xtg. Chapter 4 describes a bulked segregant analysis approach aimed at fine-mapping a previously identified quantitative trait locus for resistance to Xtg. To this end, a large (n = 7,484) population was screened and the 750 most resistant and 761 most susceptible individuals were selected to form two pools. These pools were sequenced by whole-genome deep sequencing, allowing a high resolution for fine-mapping. This revealed a 300 kbp region that was highly associated with Xtg resistance and contained many genes with a potential role in disease resistance, including genes encoding for Pik2-, RGA4- and RGA5-like disease resistance proteins, cysteine-rich kinases, and serine protease inhibitors. This region and the underlying genes and genetic markers hold great potential to breed cultivars with increased resistance. Finally, Chapter 5 discusses the results obtained in this work and expands on the future perspectives to validate the identified genes for resistance to Xtg, define markers for markerassisted selection, and identify new sources of resistance for a more efficient and durable control of Xtg.