The genetic basis of sex determination in grapes

Abstract

It remains a major challenge to identify the genes and mutations that lead to plant sexual differentiation. Here, we study the structure and evolution of the sex-determining region (SDR) in Vitis species. We report an improved, chromosome-scale Cabernet Sauvignon genome sequence and the phased assembly of nine wild and cultivated grape genomes. By resolving twenty Vitis SDR haplotypes, we compare male, female, and hermaphrodite haplotype structures and identify sex-linked regions. Coupled with gene expression data, we identify a candidate male-sterility mutation in the VviINP1 gene and potential female-sterility function associated with the transcription factor VviYABBY3. Our data suggest that dioecy has been lost during domestication through a rare recombination event between male and female haplotypes. This work significantly advances the understanding of the genetic basis of sex determination in Vitis and provides the information necessary to rapidly identify sex types in grape breeding programs.

Fig. 1. The morphology of flower sexes in grapes and a phylogenetic analysis of wild and cultivated species.

Side view (a) and top view (b) of dioecious Vv sylvestris female O34-16 (left), male DVIT3351.27 (middle), and of hermaphrodite Vv vinifera Chardonnay (right). Scale bar = 1 mm. c A phylogenetic tree predicted from whole-genome proteome orthology separates species by taxonomy and not by sex genotype. M. rotundifolia is an outgroup to the Vitis ingroup. For each individual, the genotype of the sex-determining region is indicated in parentheses followed by its corresponding sex type. The symbols ♀, ♂ and represent female, male, and hermaphrodite individuals, respectively. Numbers associated with nodes reflect bootstrap values (see “Methods”). Scale bar is in the unit of the number of substitutions per site.

Fig. 2. Sex-linked structural variants and their impact on gene content.

Whole-sequence alignments of the sex-determining region (SDR) in M (a), F (b), and H (c) haplotypes against the Vv vinifera Cabernet Sauvignon chromosome 2 hap1 (H) reference. The figures illustrate that M haplotypes (a) tend to be longer than either F (b) or H (c) haplotypes and there is a large inversion in the M haplotype of M. rotundifolia. d Schematic representations of the SDR in four of the 11 genomes analyzed for this study. From top to bottom, the figure illustrates the haplotypes of hermaphrodite Vv vinifera Cabernet Sauvignon (HF), male Vv sylvestris DVIT3351.27 (MF), female Vv sylvestris DVIT3603.07 (FF), male V. arizonica (MF), and male M. rotundifolia (MF). Each haplotype is annotated with arrows and rectangles to depict annotated genes. White arrows depict genes that are not sex-linked. Genes affected by nonsense mutations are indicated with an X. The scale above the haplotypes denotes the position on the Cabernet Sauvignon reference. The filled, black triangles on this scale mark the position of the sex-linked genetic markers VVIB23 and VSVV010; the white-filled triangle represents the amplicon VSVV011, which is not linked to the SDR. Source data are provided as a Source Data file.

Fig. 3. Sex-linked polymorphisms along the sex-determining region in Vitis spp.

The number of all (a) and nonsynonymous sex-linked SNPs (b) per kbp across the sex-determining region (SDR). SNPs were identified by aligning all Vitis haplotypes to Vv vinifera Cabernet Sauvignon hap1 (H). Only SNPs strictly (100%) linked to one sex type were plotted. c Linkage disequilibrium (LD) across the region represented as the median of the squared correlation coefficients (r²) between all pairs of SNPs calculated within 20 kbp windows. d TF-binding motif conservation per gene promoter detected in the SDR. The x-axis denotes the location on the Cabernet Sauvignon hap1 (H) SDR, and the two black triangles along this axis mark the position of the genetic markers VVIB23 and VSVV010 that are closely linked to the SDR. The potential position of the recombination event responsible for the reversion to hermaphroditism in domesticated Vv vinifera is also indicated. e Gene composition of the H haplotype of Vv vinifera Cabernet Sauvignon hap1. Genes are colored as in Fig. 2d and white-colored genes are not sex-linked. Genes affected by nonsense mutations are indicated with an X followed by the affected haplotype and the gene name in parentheses. f Neighbor-joining clustering of the protein sequences encoded by each gene of the SDR. Yellow, purple and gray colors represent F, M and H haplotypes, respectively. Source data underlying Figs. 3a–d, f are provided as a Source Data file.

Fig. 4. Mutations and segregation in VviINP1.

a Alignment of the first 100 bp of 20 VviINP1 coding sequences representing 12 F, 5 H and 3 M haplotypes along with two M. rotundifolia INP1 coding sequences from F and M haplotypes. Alignment revealed an F-linked 8 bp INDEL in VviINP1 throughout Vitis and shared with M. rotundifolia. Yellow, purple and gray colors represent F, M and H haplotypes, respectively. b Phylogenetic subtree of the INP1 coding sequences from Vitis spp. and M. rotundifolia. The tree was rooted with INP1 sequences from seven outgroups (see “Methods”). Tree branches are colored according to sex-determining region haplotype. Scale bar is in the unit of the number of substitutions per site. Sequence length in bp is indicated in parentheses. c Marker assay amplifying INP1 fragment without (top panel, 609 bp) or with (bottom panel, 605 bp) the 8 bp deletion. Actin was used as a PCR positive control (99 bp fragment). This assay was performed three times. Vvs Vv sylvestris, L Ladder, C− negative control, Vv vinifera: Bs seeded Black Corinth, BC seedlees Black Corinth, CF Cabernet Franc, CS Cabernet Sauvignon, Ca Carménère, Ch Chardonnay, GB Gouais blanc, Ri Riesling, SB Sauvignon blanc, Se Semillon, Zi Zinfandel, Vv sylvestris: 1, DVIT3351.27; 2, DVIT3603.07; 3, DVIT3603.16; 4, O34-16; Va V. arizonica, Vp V. piasezkii, Vr V. romanetii, Mr M. rotundifolia. Source data are provided as a Source Data file.

Fig. 5. Transcriptome and gene network analyses.

a Sex-determining region genes have sex-linked expression at each floral stage. Genes are classified in three groups based on their expression pattern. Only genes differentially expressed in one flower sex compared to the two other sex types are shown. The colors of the heat map depict the Z score of the normalized counts per gene. Gene coexpression networks of the module magenta (b), positively correlated with male sex, and red (c), negatively correlated with female sex. Node color indicates the degree of connectivity in the network. Positions of the gene encoding a PPR-containing protein, VviAPT3 and WRKY in their corresponding networks are highlighted. HP hypothetical protein. Source data underlying Fig. 5a are provided as a Source Data file.

Fig. 6. Model of the evolution of sex determination in grapevine.

a A graphical representation of the association between sex and the observed polymorphisms in promoter regions (top row) and encoded proteins (bottom row) of each gene present in the sex-determining region. Genes affected by nonsense mutations are indicated with an X with the affected haplotype. Sex linkage observed only in Vitis vinifera species are indicated with a Vv. b A potential model for the evolution of dioecy in Vitis and its relatives and the reversion to hermaphroditism in cultivated Vv vinifera. From left to right, a hermaphroditic ancestor gave rise to male-sterility mutation to produce gynodioecious individuals. Here we denote the recessive male-sterility mutation as sp, following the convention of Oberle. Next, male flowers originated as a consequence of a female-sterility mutation(s), labeled So, again following the convention of Oberle. According to our data, we hypothesize that a rare recombination event occurred in a Vv sylvestris male, leading to H haplotype and hermaphrodite individuals in domesticated Vv vinifera cultivars. The symbols ♀, ♂ and represent female, male and hermaphrodite individuals, respectively.