Genome-wide patterns of genetic variation among wild and cultivated grapevines (V. vinifera L.)

Grapevine (V. vinifera L.) is one of the most important crops worldwide due to its global distribution and economic value. Two forms of grapevine still co-exist nowadays: the cultivated form V. vinifera subsp. sativa and the wild form V. vinifera subsp. sylvestris, which is considered the ancestor of present cultivars. Archeological and historical findings suggest that cultivated grapevines have been domesticated from wild populations of V. sylvestris circa 8,000 BP in the Near East. However, recent genetic analyses raised the outstanding question whether multiple domestication events occurred. During domestication the biology of grapes changed dramatically to guarantee greater yield, higher sugar content and more regular production. The changes in berry and bunch size as well as the transition from dioecious wild plants to hermaphrodite cultivated grapes were crucial. Additional studies on the genetic relationship between wild and cultivated grapevines are required in order to understand how this phenotypic evolution occurred and to clarify the process of adaptation to domestication in grapevine. This will be useful for the future genetic improvement of viticulture. In this regard, we investigated the genetic and phenotypic variation within a germplasm collection of wild and cultivated grapevine accessions. The whole population was first genotyped with the commercial GrapeReSeq Illumina 20K SNP array, yielding 16K good quality single nucleotide polymorphisms (SNPs). Afterwards, a novel Restriction Associated DNA-sequencing (RADseq) procedure was developed in order to further increase the density of molecular markers across the grapevine genome. By applying this novel RAD-seq protocol to the whole population, 37K SNPs were identified, which reflected a considerable level of genetic diversity between sativa and sylvestris accessions. The two merged SNP matrices were filtered for SNP loci with a missing rate > 0.2 and a minor allele frequency (MAF) < 0.05. The final panel of 27K SNPs evenly distributed along the grapevine genome was used to investigate the population structure by using both Principal Component Analysis (PCA) and the cluster algorithm implemented in fastSTRUCTURE software. In line with previous research, both analyses highlighted a low but clear differentiation between sativa and sylvestris individuals. Therefore, the extent of Linkage Disequilibrium (LD) was evaluated within the whole grapevine population and in the two subspecies separately. LD, as measured by the classical r2 correlation coefficient, decayed below 0.2 within 10 kb in the whole population. On the other hand, a slower LD decay was observed in the wild compartment, where r2 reached values below 0.2 within 20 kb. This result can be related with an elevated level of inbreeding among wild individuals, linked to a small effective population size and the missing gene-flow between wild populations. Population differentiation statistic (FST) was computed across the grapevine genomes looking for genomic regions with divergent allele frequencies between the two grapevine subspecies. An overall low level of genetic differentiation (FST = 0.12) was observed between cultivated and wild grapes, suggesting the occurrence of genetic exchange among the two subspecies. However, a non-random distribution of divergent sites was observed along the whole genome: over two thousands of SNP loci revealed a significant level of differentiation between sativa and sylvestris, validated empirically with a permutation test. 1,714 annotated genes were found in LD with these most significant SNPs, and showed an enrichment of predicted functions II related to the metabolic processes of nitrogen and carbohydrate as well as to the perception and adaptation to environmental stimuli. A slightly reduction of nucleotide diversity in the sylvestris (πsylvestris/ πsativa ~0.95) was observed in almost all the identified genes involved in stress responses, suggesting that a selection is likely acting in wild populations for adaptation to several environmental changes. Therefore, these results point the attention towards sylvestris grapevines as valuable resources of resilience genes or alleles, which may have been lost in cultivated grapevine during the domestication process. Genome-wide association study (GWAS) approach has been applied as an alternative strategy to identify the genes and mutations that have been targets of selection during crop domestication. Therefore, the germplasm collection of cultivated and wild grapevines has been evaluated in two years for single berry and single bunch weight, number of bunches per plant, yield and berry composition (sugar, organic acid and K+ concentrations, titratable acidity and pH). A great phenotypic variation was observed within and between the two grapevine subspecies, notably for berry size, pH, acid contents and titratable acidity. The association test, carried out accounting for confounding factors, identified significant genotype-phenotype correlations for all traits, except for single berry weight. Genes encoding proteins related to Ca2+ sequestration and signalling, transcription factors and enzymes involved in the metabolism of polyamines were identified in linkage with the SNPs significantly associated to yield and bunch weight. At the same time, genes with a central role in the control of berry flesh pH and acidity were detected, such as the isocitrate lyase and V-type proton ATPase subunit a3 genes. Therefore, the present research has proven for the first time the feasibility of population genetics and association mapping approaches for dissecting the genomic basis of phenotypic variation in a complex genetic system as grapevine. Moreover, further evidence of the relevance of wild grapevine as a model for understanding the mechanisms of adaptation to natural conditions has been provided. These results pave the way for understanding how wild and cultivated grapevines react to environmental stimuli, which will benefit the development of new breeding strategies to face the ongoing climate changes and the growing demand of a sustainable viticulture.