UC Davis

The European grape (Vitis vinifera) is used worldwide for wine making. However, its production is constantly challenged by a range of pests and pathogens that cause substantial losses in yield and fruit quality. Among these, Pierce’s disease (PD) and powdery mildew (PM) pose a serious threat to the wine industry in California. The management of both diseases currently relies on intensive pesticide applications, which raises concerns regarding environmental sustainability, application costs and the emergence of resistant pathogen populations. Wild North-American Vitis species are not suitable for winemaking but represent a valuable reservoir of genetic diversity and natural disease resistance that can be leveraged to improve susceptible V. vinifera cultivars. Breeding efforts have focused on introducing some of these favorable traits into susceptible varieties, but the process is lengthy and can result in linkage drag. Advancements in grapevine genomics and genome editing offer the unprecedented opportunity to identify, characterize and utilize these traits to develop disease-resistant cultivars. This dissertation aims to characterize the resistance locus associated with PD in wild Vitis species and target susceptibility genes associated with PM using genome editing. Together, these approaches provide complementary strategies to understand the molecular basis of disease resistance and contribute to the development of improved vines.Chapter 1 provides a comprehensive overview of recent advancements in grapevine and pathogen genomics aimed at achieving durable disease resistance. This research emphasizes the critical role of diploid genome assemblies in dissecting disease resistance loci, enabling the precise identification of candidate genes suitable for gene stacking and precision breeding. Additionally, it highlights the importance of understanding pathogen evolution, including mechanisms of adaptation, effector discovery, and resistance breakdown. This knowledge is essential not only for informed resistance breeding strategies, but also for identifying novel gene targets for genome editing.Chapter 2 focuses on the identification of candidate PD-resistance genes in a wild grape species, Vitis arizonica b40-14, that segregates for resistance using genome-wide association studies (GWAS). Significant associations between genotype and phenotype were identified across multiple genomic regions. However, one of the strongest association peaks was located on chromosome 14, between the SSR markers defining the Pierce’s disease resistance locus 1 (PdR1) locus, further supporting this region as a candidate resistance locus. The integration of gene expression data revealed four expressed candidate genes encoding for extracellular receptors potentially involved in PD resistance. These included two genes coding for leucin-reach repeat receptor-like proteins (RLPs), one coding for a lysin motif receptor-like kinase (LysM-RLK) and one coding for a leucin-reach repeat receptor-like kinase (LRR-RLK).Chapter 3 explores variation within the PdR1 locus across different V. arizonica hybrids accessions. This study integrated multiple layers of information to refine the PdR1 region in V. arizonica b40-14 and identified candidate genes associated with PD. Genome assemblies generated from HiFi reads for four V. arizonica hybrids enabled the generation of highly contiguous and complete haplotypes. Among these, four haplotypes were associated with PD resistance and three were not associated with the resistant phenotype. One resistant haplotype was sequenced from a PD-resistant backcross individual that inherited resistance from V. arizonica b40-14, the wild grape accession used in Chapter 2. To investigate sequence diversity and identify variants specific to PdR1, sequence graph approaches were employed. Results showed a substantial sequence and structural divergence between PD-resistant haplotypes and haplotypes not associated with PD resistance, as well as variation among the resistant haplotypes. By integrating these findings with gene expression data, it was possible to narrow down the candidate genes in the PdR1 refined region to two genes coding for LRR-RLPs, which were advanced to functional validation.Chapter 4 builds upon the findings from Chapter 2 and 3 and aims to validate the function of PD candidate resistance genes. To achieve this, a dual complementary platform was developed. The candidate genes were overexpressed in the susceptible V. vinifera cv. Thompson Seedless (TS), and the same genes were knocked-out in the resistant V. arizonica b40-14. For overexpression in TS, Agrobacterium-mediated transformation was used to generate transgenic plants. Gene knock-out in V. arizonica was achieved using through CRISPR/Cas9 editing of protoplasts, which is an extremely challenging yet valuable approach.Transformation was successful for two out of the four PD candidate genes. Ten putatively transgenic TS lines were obtained, five for each of the two candidate genes that could be transformed. For the other two genes, transformation was unsuccessful, suggesting that constitutive overexpression under the CaMV35S promoter may have detrimental effects. For the knockout experiment, preliminary results showed that the sgRNAs designed to target the candidate genes display a wide range of editing efficiencies and that the type, frequency and pattern of mutation is strongly guide-dependent. Putatively edited V. arizonica embryos were regenerated from protoplasts for all the genes of interest and currently at plantlets stage.Chapter 5 explores an alternative strategy for inducing resistance by targeting a family of susceptibility genes associated with PM, known as Mildew Resistance Locus O (MLO) genes. Using DNA-free genome editing, we generated MLO-knockout lines carrying different mutation patterns across three MLO genes: VvMLO3, VvMLO7 and VvMLO11. Edited plants regenerated from protoplasts were genotyped and the absence of CRISPR components was verified confirming that the plants are not transgenic. Eight edited lines were evaluated for susceptibility to PM and carried diverse combination of mutations across the three MLO loci. Phenotypic assessment was conducted both visually, by scoring pathogen growth, and quantitatively, by measuring pathogen accumulation in the edited plants at multiple time points. Among the edited lines, one showed overall partial reduced pathogen accumulation compared to the non-edited susceptible control.