Abstract
Variegation in Vitis hybrids was investigated to confirm the inheritance as a single, recessive gene as previously proposed and commonly observed in breeding programs. Variegated leaves have ornamental appeal, but the phenotype is sublethal in some environments. Twenty-nine grape families were characterized for variegation including F1, S1, and S2 populations. The majority segregated 3 wild type (WT):1 variegated and were supported by chi-square tests. Four populations had segregation ratios supporting 15:1 or 1:1 models, and a unique flecking phenotype was identified in a Landot 4511 S1 population that suggested the interaction of two recessive loci. A variegated parent was selfed to produce progeny with no WT offspring and was segregated 0:1. Marker trait associations including bulk segregant analysis (BSA), genome-wide association mapping, and quantitative trait loci (QTL) mapping was used on three populations. On chromosome 14, Lvar1 was identified and mapped to 24.5 to 29.5 Mb and associated closely with rhAmpSeq marker 14_27607541. Lvar2 was associated with rhAmpSeq marker 11_18433819 on chromosome 11 at 12.2 to 18.4 Mb. The identification of two loci and the segregation data in some populations suggest that grape breeding germplasm segregates for two recessive loci. The pedigree records suggest that ‘Frontenac’ inherited one of these loci, and that Landot 4511, an ancestor of many populations tested in this experiment, may carry two loci. A total of 252 candidate genes were identified at these loci, including a key target—adenosine triphosphate (ATP)-dependent zinc metalloprotease FtsH6, involved in photosystem II and similar to the var2 mutant in Arabidopsis. This knowledge can help breeders select for ornamental grapevines or eliminate variegation from their breeding programs.
Variegation is a common trait in many plant species and has been studied extensively to understand modes of inheritance and molecular mechanisms (Kirk and Tilney-Bassett 1978). In general, variegation is inherited in a Mendelian fashion, but variegation can be both heritable and nonheritable (Kirk and Tilney-Bassett 1978). Several different types of molecular mechanisms have been discovered among variegated plant species: nuclear recessive genes, chimerism, transposable element activity, RNA silencing, plastome mutators, plastome mutations, mitochondrial genome mutations, and plastid–nucleus incompatibility (Yu et al. 2007).
Variegation has been observed in grapevine for centuries. In 1808, Thomas Andrew Knight, who served as the second president of the Royal Horticultural Society, wrote a letter describing a variegated vine called ‘Aleppo’, which was crossed with ‘White Chasselas’ to produce variegated offspring (Downing 1857; Knight 1808). Later, in 1917, Rasmuson (as cited in Kirk and Tilney-Bassett [1978]) described another variegated grape cultivar Vitis vinifera ‘variegata’, in which the trait expressed a mosaic pattern and segregated 3:1 (WT:variegated) when selfed (Kirk and Tilney-Bassett 1978). Reisch and Watson (1984) later examined the inheritance of variegation in five S1 and five F1 V. vinifera and hybrid grape populations. Nine of the 10 populations segregated 3:1, providing strong evidence that variegation is a nuclear recessive trait under monogenic control. However, Filler et al. (1994) observed a 15:1 segregation in F2 grape populations made from crosses of V. riparia × V. riparia and V. riparia × Vitis spp. (French hybrid), suggesting the trait is controlled by two unlinked loci. One cross between V. riparia × V. vinifera segregated 3:1, in which they postulated one locus in the heterozygous state and the other in a homozygous recessive state.
Across most species tested, variegation was found to be inherited in a Mendelian fashion in the F2 generation as indexed by Kirk and Tilney-Bassett (1978). A few examples are Zea mays ‘argostripe’ (Eyster 1934), Arabidopsis thaliana ‘immutans’ (Rédei 1963; Röbbelen 1968), Barbarea vulgaris ‘variegata’ (Dahlgren 1921), and Oryza sativa ‘striped’ (Morinaga 1932). A few species show dihybrid segregation with a 15:1 ratio as in Phaseolus vulgaris ‘variegata’ (Zaumeyer 1942) and Plantago major ‘asiatica’ (Ikeno 1927).
Some types of variegation have been described as a chimera and genetic mosaic, possibly affecting both L1 and L2 cell layers (Kirk and Tilney-Bassett 1978). Bud sport mutations for variegation and albinism do occur in vineyards, including at the University of Minnesota (UMN; personal observation). Leaves, flowers, and fruit have been identified with variegated phenotypes from WT plants. Similarly, seedling populations segregating for the trait generally display variegation at the cotyledon stage, when the mutant seedlings are typically culled.
In grapevine, variegation is observed early in the grape breeding pipeline, and seedlings are often discarded as a result of the sublethal nature of the trait. In the UMN breeding program, very few variegated “escapees” survive to fruiting. Identifying the locus or loci controlling variegation would be especially useful for identifying parents that may be sources, and heterozygous carriers, of the trait to make more informative decisions in designing crosses. Because of the presence of variegation in the breeding program, understanding its association with other traits (in the heterozygous state), and its impact on plant growth and pleiotropic effects are important (Olson and Clark 2021).
To our knowledge, no previous attempts have been made to map the loci controlling variegation in Vitis spp. Segregation ratios were tested in 29 grape breeding seedling families from the grape breeding programs at Cornell University and UMN. Several populations were used in three different genetic mapping strategies. Pedigrees were investigated to understand the possible inheritance of the variegation phenotypes observed. The objectives of this study were to map the regions associated with variegation in hybrid grape populations (Vitis spp.) using marker-assisted selection and to identify candidate genes located within the associated regions of interest for future study.
Materials and Methods
Description of variegated phenotype and segregation ratios.
Twenty-nine grape families were characterized for leaf variegation as young seedlings, and segregation was evaluated for the phenotypes observed. Among these populations were 23 F1 families, five S1 families, and one S2 family from the UMN and Cornell University breeding programs. The families were developed through controlled pollination in the field where hermaphrodite mother vines were emasculated and bagged to prevent selfing up to 2 d before anthesis. Pollen was collected separately from the selected male parent, applied to the receptive stigmas of the mother plant, and a crossing bag was applied. For self-pollination, inflorescences were checked to remove any open florets, and crossing bags were applied to inflorescences. Seed was collected at fruit maturity, cleaned, and stratified. Seeds were germinated in the greenhouse in flats containing germination mix. Populations ranged in size from 24 to 249 full-sib seedlings. A chi-square test for goodness of fit (α = 0.05) was used to determine whether a population fit the expected Mendelian segregation ratio of 3:1, or other typical ratios 1:1 and 15:1, to investigate other possible inheritance patterns (Table 1). A general description of variegated grapevine is based on observations made among the segregating hybrid grape families at the UMN grape breeding program, with a representative image provided in Fig. 1.
Representative images of the wild-type (A) and variegated (B–F) phenotypes observed in the mapping populations using the following descriptors: light green (B), yellow (C), white (D), flecking light green (E), and flecking white (F).
Citation: HortScience 57, 11; 10.21273/HORTSCI16763-22
Populations tested for segregation ratios of leaf variegation trait.
Mapping populations.
Nine of the 29 variegated hybrid grape families were genotyped and used in marker–trait association techniques and/or genetic mapping (Table 1). Three families contributed by Cornell University were F1 hybrids made from crosses made in 2013 and 2015: 13.0206 (n = 96), 15.0417 (n = 57), and 15.0418 (n = 38). Six families from UMN were F1 hybrids made from crosses in 3 successive years (2015–17): GE1642 (n = 165), GE1703 (n = 69), GE1818 (n = 46), GE1819 (n = 36), GE1838 (n = 43), and GE1848 (n = 55). The remaining family, GE1895 (n = 110), was an S1 family derived from MN1220. Phenotyping for variegation occurred in the greenhouse shortly after cotyledon development. Variegation was scored as a binary trait—either the plant was variegated (1) or WT and not variegated (0) (Fig. 2).
Developing grape cotyledons showing the wild-type (left; also observed in nonflecking variegation type) and mutant variegated seedlings (right).
Citation: HortScience 57, 11; 10.21273/HORTSCI16763-22
DNA isolation and genotyping.
For the UMN mapping families, leaf tissue was collected into paper coin envelopes from each individual and placed into a –80 °C freezer for 24 h. Samples were lyophilized for 48 to 72 h. After lyophilization, ∼0.1 mg of plant material was transferred to individual wells of a 96-well plate in preparation for DNA extraction. For the Cornell families, young leaves were excised from actively growing seedling shoot tips and placed into wells in 96-well plates. Each sample tube was capped and stored over ice until plates were taken to the laboratory. Caps were then removed and placed into an incubator set to 55 °C overnight. Dried samples were sealed with new caps and the lid was affixed to the sample box with a rubber band, and shipped for DNA extraction. Genomic DNA was extracted using a QIAGEN DNeasy Plant DNA kit with an addition of 2.8% PVP-40 in the lysis buffer (QIAGEN Inc., Hilden, Germany). PVP-40 was added for improved purity by removing polyphenol compounds from the DNA extracts. DNA extraction was performed on all individuals from UMN populations in-house, or for Cornell at Intertek AgriTech (Alnarp, Sweden) (Zou et al. 2020). Examination of DNA quality and concentration was conducted using the Quant-it Picogreen dsDNA Assay Kits (InVitrogen Life Technologies, Eugene, OR, USA). Dilutions were targeted to 30 to 60 ng of template DNA based on an average concentration of eight samples quantified per plate (Karn et al. 2021).
Two genotyping approaches were used. First, skim whole-genome resequencing (skim-seq) (Golicz et al. 2015), a low-coverage, alternative method to genotyping by sequencing, was conducted on nine mapping families. Genotype data generated from skim-seq were used in a BSA approach. Second, genotyping using the whole-genome rhAmpSeq marker platform with pair-end sequencing on an Illumina NextSEq 500 was conducted on 280 seedlings and the parents from three populations—GE1642 and GE1895 (S1), which are closely related, sharing MN1220 as a parent; and GE1703. These families were selected because of availability and larger population size while additional populations were being made. rhAmpSeq uses ∼2000 amplicon markers, targeting regions evenly distributed across the Vitis core genome (Zou et al. 2020). Markers within this platform represent multiallelic haplotypes at each locus, which may return up to four alleles per locus in two heterozygous parents. Genotype data generated from rhAmpSeq were used for genetic mapping, genome-wide association studies (GWAS) using single marker analysis, and QTL mapping.
Bulked segregant analysis.
Skim-seq genotyping was conducted as 37 variegated pools and 36 WT pools composed of 248 individuals from the nine mapping families, including 13 pools that included parental sequences. Families had varying numbers of pools based on population size, with each pool containing five full-sib WT or variegated individuals. Variegated and WT pools were normalized for equal amounts of isolated DNA based on the individual with the lowest DNA concentration in the pool. Each of the pools was barcoded, multiplexed, and sequenced in a single flow cell on a Hiseq2000 Illumina sequencer. For BSA, variants were detected among pools using the DNAseq pipeline in Sentieon Genomics software (version 201808.07) with default options (Freed et al. 2017). This pipeline is a modification of GATK Best Practices recommendations (McKenna et al. 2010), which includes mapping reads to the PN40024 version 12X.v2 grape reference genome (Canaguier et al. 2017) by BWA MEM version 0.7.12 (Li 2013), removing duplicated reads, and variant detection with the HaplotypeCaller algorithm. Reads with a mapping quality of less than 60 were removed from the downstream analysis. The null hypothesis of the Fisher’s exact test is that the frequency of reference/alternative allele in the variegated and WT samples are the same. The variance that was enriched significantly in variegated samples was detected by the Fisher’s exact test using the NumPy (Oliphant 2006) package of Python v3.6. The physical position of the variants with P < 0.01 were plotted and defined as candidate sites associated with the variegation phenotype. The functional effect of candidate sites was further annotated using SNPeff (Cingolani et al. 2012).
GWAS of variegation.
The raw, converted VCF files for each grapevine family were imported in TASSEL 5.2.5.1 software (Bradbury et al. 2007). Parents were removed from the analysis, along with individuals that failed the multidimensional scaling or Mendelian error quality control analyses. Markers with more than 10% missing data and a minor allele frequency of less than 10% were filtered from the genotype dataset to minimize the influence of genotype calling artifacts. Next, genotypes were imputed using the LD-kNNi imputation plugin, which is also referred to as LinkImpute version 1.1.4 (Money et al. 2015) using the default parameters (high LD sites, 30; number of nearest neighbors, 10; and maximum distance between site to find LD, 10,000,000 bp).
GWAS were performed in TASSEL on three individual mapping families separately using the method described in Zou et al. (2020). Because of the allele calling procedure (up to four unique alleles per marker per family), the families could not be combined. Post hoc analysis of the mixed linear model output for each of the families was performed in R software (R Foundation for Statistical Computing, Vienna Austria) using the Bioconductor package (Morgan 2019). A Bonferroni-corrected threshold was determined for each association analysis using 1/N (α = 0.05), where N is the number of markers tested. Phenotypic variation was estimated by R2 values of each significant marker. Manhattan and quantile–quantile plots for each GWAS were visualized in R software (R Foundation for Statistical Computing) using the qqman package (Turner 2017).
Linkage map construction.
Genetic maps of GE1642 and GE1703 were constructed using the protocol described by Zou et al. (2020) in LEP-MAP3 software (version 0.2) (Rastas 2017). Quality control analyses were performed before map construction. Parents were removed from the analysis, along with individuals that failed the multidimensional scaling or Mendelian error quality control analyses. Consistency of the maps, genome organization, and structural variation were evaluated through correlation plots of genetic and physical distances of individual markers per chromosome for each family. A sex-averaged map, which combines and averages paternal and maternal maps, was generated for mapping in each family in LEP-MAP3 (Rastas 2017). Parental and sex-averaged maps were visualized in MapChart (version 2.32) (Voorrips 2002).
Interval mapping.
Interval mapping was performed in R software (R Foundation for Statistical Computing) using the scanone function with a binary model and four-way cross in the R/qtl package (Broman et al. 2003) on the sex-averaged genetic maps of GE1642 and GE1703, separately. A genome-wide logarithm of the odds (LOD) threshold was determined by performing 1000 permutations at α = 0.05. Associated loci were indicated by significant LOD peaks exceeding the threshold and the QTL interval identified with Bayesian intervals at a 95% confidence interval. The flower sex phenotype from a subset of WT individuals from GE1642 (n = 69) was used to map the well-characterized flower sex locus (Picq et al. 2014; Zou et al. 2021) as an additional method to validate it for producing accurate maps. GE1642 segregated for female and hermaphroditic individuals; thus, parents were both believed to be heterozygous.
Candidate gene identification.
Candidate genes within ∼5 kb of the associated region identified in the BSA, and genes tightly linked to single nucleotide polymorphisms (SNPs) in the GWAS were examined for their functional effects using VCost (version 3) annotations from URGI on the PN40024 version 12X.v2 reference genome. Nucleotide sequences were translated using the translate tool in the Bioinformatics Sequence Manipulation Suite (Stothard 2000) and subsequently put into National Center for Biotechnology Information BLASTp to determine the annotation for each gene. Genes within the QTL intervals detected in genetic mapping were examined for their functional effects using RefSeq annotations from the National Center for Biotechnology Information on the PN40024 version 12X.v2 reference genome. Some genes from GWAS and genetic mapping overlapped and were counted only once. Proteins encoded by the candidate genes were further organized by their functional annotation into several groups: catalysis, structural, transport, regulatory, hormone, storage, defense, and uncharacterized.
Results
Description of variegated phenotype.
The variegation phenotype in Vitis spp. was highly variable and presented as a light-green, yellow, or white phenotype, or a combination of all three with normal green. Distinct sectors are formed between different variegated phenotypes on the same leaf. A novel flecking appearance was observed in only a single population S1 from Landot 4511 (syn. ‘Landot noir’, GE1960; Fig. 1F). In general, the variegation trait was highly expressed in the first few mature leaves and becomes less pronounced further in development (Olson and Clark 2021). Some variegated plants showed strong expression in all leaves throughout development. Some populations expressed only light-green variegation, whereas others expressed a combination of variegated phenotypes. Mapping families used in GWAS and genetic mapping (GE1641, GE1703, and GE1895) expressed both light-green and white variegated phenotypes (Fig. 1). The GE1703 family displayed other morphological abnormalities among its progeny, such as enations, mottling, and dwarfing.
Segregation and inheritance of variegation trait.
Chi-square tests evaluating segregation (WT:variegated) were not significantly different from 3:1 (P > 0.05) for 24 of the 29 populations, indicating that the 3:1 hypothesis could not be rejected (Table 1). Twenty of these 24 were biparental F1 populations, three were S1 populations, and one was an S2 population (Table 1). This suggests that each of the parents of those 20 populations was heterozygous for the recessive allele, and within each F1 population the trait was determined by the same locus in both parents. Four populations did not segregate 3:1 (α = 0.05), yet resembled other Mendelian segregation ratios, and thus were tested for 15:1 and 1:1 ratios. A single F1 family from Cornell University (15.0417) segregated 1:1, whereas two of the three F1 populations (GE1940, GE1942, and GE1945), all with ‘Itasca’ as a common parent, segregated 15:1. A single S1 population generated from a self of a mature, variegated plant produced all variegated progeny resulting in a 0:1 ratio (GE1953). When considering just nonflecking variegation and green (WT) phenotypes, the GE1960 population (S1 Landot 4511) segregated 3:1 (P > 0.05), but when considering the flecking phenotype, the population segregated in a dihybrid epistatic fashion between a 9:3:4 ratio (green:flecking:white) and a 12:3:1 ratio (green:white:flecking), suggesting that Landot 4511 inherited recessive alleles at two loci for variegation that interact to produce the flecking.
BSA of variegation.
The BSA approach identified 69 markers associated significantly with variegation at P < 1 × 10–20 (Fig. 3). The low read depth resulted in a high amount of variance among pools, which resulted in many significant markers scattered across nearly all 19 chromosomes. Sixteen of the 19 chromosomes had at least one significant marker. Based on the previously mentioned single-locus hypothesis for these populations, we focused on chromosome 14 because it contained markers with the strongest association with variegation. Four SNP markers (of 97,407 total markers genome-wide) located on chromosome 14 were evaluated further. The Fisher’s exact test revealed that at each of these four markers, the variegated pools were homozygous (large difference between reference and alternative SNP alleles read count), whereas WT pools had strong evidence of both SNP alleles, suggesting heterozygosity (Table 2). The expected allele ratio across progeny (AA:AB:AB = WT, BB = variegated) would have been 4:2 A:B for WT, which matched most closely the read depth ratios for 14_21425734. The four significant markers are located within a transposable element (Vitvi14g01206), so instead we examined the closest candidate gene colocalized within ∼5 kb of the markers, which was a glyceraldehyde-3-phosophate dehydrogenase (GAPC1).
Manhattan plot of Fisher’s exact P values from the bulk segregant analysis of nine pooled variegated hybrid grape populations. Indicated by the significance line, markers were tested against the genome-wide threshold of P < 1 × 100–20. Legend shows marker density, as indicated by the bar along the x-axis above the label for each chromosome.
Citation: HortScience 57, 11; 10.21273/HORTSCI16763-22
Read counts reference and alternative alleles for four significantly associated loci on chromosome 14 identified using Fisher’s exact test between wild-type and variegated pools in bulked segregant analysis.
GWAS of leaf variegation.
GWAS were performed on three seedling families separately (GE1642, GE1895, and GE1703). A pedigree showing the interrelatedness of these populations is shown in Supplemental Fig. S1. GWAS of population GE1642 tested 962 filtered rhAmpSeq markers on 119 individuals. Five markers located between 27,199,005 and 30,126,457 bp on chromosome 14 were associated significantly with variegation (Fig. 4A). Marker 14_27199117 (chromosome 14 at position 27,199,117 bp) was the most significantly associated marker, with R2 = 0.452. GWAS of GE1895 (n = 98) tested 816 filtered markers, and 20 significantly associated markers were identified on chromosome 14 between the physical region of 24,462,111 and 29,574,449 bp (Supplemental Fig. S2). Marker 14_27607541 was the associated most significantly, with R2 = 0.756. In the third population, GE1703, GWAS tested 792 filtered markers and found eight markers located between 12,164,23 and 18,433,707 bp on chromosome 11 that were associated significantly with variegation (Fig. 5B). Marker 11_18433819 was the most significantly associated marker, with R2 = 0.658.
Manhattan plots visualizing log-transformed P values from the mixed linear model of the genome-wide association studies conducted on populations GE1642 (n = 119) (A), GE1895 (n = 98) (B), and GE1703 (n = 57) (C).
Citation: HortScience 57, 11; 10.21273/HORTSCI16763-22
Logarithm of the odds (LOD) plots visualizing results from interval mapping. (A) Genetic mapping of GE1642 identified a single major locus for leaf variegation Lvar1 on chromosome 14. (B) Genetic mapping of GE1703 identified a single major locus, Lvar2, for variegation on chromosome 11.
Citation: HortScience 57, 11; 10.21273/HORTSCI16763-22
Genetic map construction.
Biparental F1 populations GE1642 and GE1703 were subjected to genetic mapping. Filtering of seedlings using multidimensional scaling and Mendelian error check removed six individuals, reducing the number to 119 seedlings for GE1642 map construction. No individuals were removed from GE1703. Filtration of noninformative and distorted (threshold = 1 × 10–6) markers resulted in the removal of 833 markers from the original 1747 rhAmpSeq markers; the remaining 914 markers were used in map construction for GE1642 with 19 linkage groups (Supplemental Table S1). Similarly, 962 markers were removed from the original 1627, and the 665 remaining markers were used in map construction for GE1703 (Supplemental Table S2). Genetic maps built in Lep-MAP3 software conserve an equal number of markers (some positions imputed) between parental maps. Thus, the sex-averaged map also contains the same number of markers as the parental maps.
Interval mapping identifies two loci involved in variegation.
Interval mapping was performed using the genetic maps for populations GE1642 and GE1703 (Fig. 5). A single, major QTL (Lvar1) was detected in population GE1642 at the genetic position of 41.4 cM on chromosome 14 in the sex-averaged map. The marker at the peak of the QTL (14_27199005) had an LOD value of 12.7 and explained 38.8% of the phenotypic variance. A 95% Bayesian interval was determined between 41.39 and 43.10 cM, and was delimited by rhAmpSeq markers 14_26071477 and 14_27607465 (Supplemental Fig. S2).
Interval mapping for the flower sex phenotype mapped the locus to chromosome 2 between the interval 15.99 and 23.33 cM, with a peak position of 22.91 cM. The peak position of the interval aligns with the physical position of 4,928,051 bp, which is within the same genomic region reported previously (Picq et al. 2014; Zou et al. 2021).
Interval mapping of GE1703 identified a single, major QTL on chromosome 11 (Lvar2) at the genetic position of 71.9 cM (Fig. 5B). The detected locus interval is between 46.5 and 71.9 cM, and is delimited by markers 11_16272242 and 11_18433707. rhAmpSeq marker 11_18204562 was determined to be the peak marker, with an LOD score of 13.45. The interval explained 65.6% of the phenotypic variance.
Candidate gene identification.
A total of 252 candidate genes were found within the two loci, Lvar1 (136 genes) and Lvar2 (116 genes). Candidate genes and their respective functional proteins are listed in Supplemental Table S3 and S4. Organization of the genes’ functional proteins showed that catalysis comprised the largest percentage of genes (33.3%), whereas, on the other hand, zero storage genes were detected. Multiple proteins were found to belong to the regulatory (17.9%), transport (11.5%), and structural (5.2%) groups, whereas a small proportion was found in the hormone (0.8%) and defense (2.8%) groups. A large proportion of genes encoded uncharacterized proteins, which comprised 28.6% of candidate genes.
Thirteen of the identified candidate genes encode proteins localized to plastids. Genetic mapping of Lvar1 identified a gene encoding an ATP-dependent zinc metalloprotease FtsH6 protein, which is involved in the degradation of the light-harvesting complex of photosystem II during senescence or high light intensity (Zelisko et al. 2005). The FtsH6 protein belongs to the FtsH protein family, which similarly was found to be involved in the Arabidopsis thaliana var2 mutant with high similarity to an Escherichia coli FtsH (Chen et al. 2000).
Discussion
In this study, we identified two novel loci—Lvar1 and Lvar2—for leaf variegation in hybrid grape populations. Based on our observations of 3:1 segregation in 24 of 29 variegated populations, which supports previous research (Reisch and Watson 1984), we have strong evidence that variegation is most often controlled by one of two recessive nuclear genes.
Many of the mapping populations in this study share a common ancestor, ‘Frontenac’, which is likely heterozygous at one of the two detected loci based on segregation in an S1 population (GE1959; P > 0.05). A parent of ‘Frontenac’, Landot 4511, was tested for being heterozygous through selfing (GE1960), and segregated for variegation (3:1, P > 0.05). Coincidentally, Landot 4511 is a relative to the V. vinifera cultivar Aramon, which displayed a similar segregating phenotype in its S1 progeny (Galzy and Galzy 1964). We hypothesize that the allele for variegation was inherited in Landot 4511 through this ancestor. However, in the Landot 4511 S1 population, we observed a unique variegated phenotype (Fig. 1F), characterized as having a “flecking” appearance. Typically, we observed variegation in the cotyledons, but nearly half the seedlings that displayed the flecking variegation (GE1960) had WT cotyledons, and the variegation appeared in the first mature leaf. Additional analysis is needed to understand the nature of this type of variegation and to assess whether there is an interaction Lvar1 and Lvar2. Duplicate gene action has been proposed as the predominant inheritance pattern in variegated hybrid grapevine (Filler et al. 1994). After completing our genotyping experiments, we similarly observed two variegated populations (GE1940 and GE1945) segregating 15:1, supporting the idea of variegation of two recessive loci inheritance pattern in these populations. These populations could be genotyped for future genetic investigation. One population from Cornell University (15.4017) segregated 1:1, which may have been the result of the small population size or possible inclusion of seedlings that resulted from selfing. Last, a single S1 population (GE1953) of a variegated parent produced all variegated progeny (0:1), again reinforcing that the variegation locus or loci are in a homozygous, recessive state. The inheritance of variegation in Landot 4511 could be clarified by selfing parents ‘Landal’ and ‘Villard Blanc’, and observing segregation ratios in their progenies. If variegation were controlled by one gene, the expected segregation in these families would be 3:1. Alternatively, if variegation were controlled by two genes, the expected segregation in these families would be 15:1 or would resemble a two-gene ratio such as that observed in GE1960.
Using BSA, GWAS, and QTL mapping approaches, we identified two loci associated significantly with variegation on chromosomes 11 and 14 in separate populations. Lvar2 on chromosome 11 was detected in both GWAS and genetic mapping in population GE1703 (MN1256 × MN1327), whereas Lvar1 on chromosome 14 was detected in all three mapping approaches in GE1642 (MN1220 × MN1326) and GE1895 (MN1220 selfed). Genetic mapping validated our GWAS findings in GE1642 and GE1703, as both loci mapped to about the same region. It should be noted that initial attempts to construct a linkage map for GE1703 in LepMAP3 failed. Thus, the map was curated by creating each linkage group individually and combining all linkage groups at the end for a linkage map with all 19 chromosomes represented, resulting in a heavily inflated map (Supplemental Table S2). The total genetic distance of the map was ∼400 cM longer than the GE1642 genetic map. This is evident by the large interval identified in the QTL analysis (∼25-cM interval).
From the mapping approaches tested, 252 candidate genes were identified at Lvar1 and Lvar2. Genes encoding regulatory proteins would be of most interest based on the hypothesis that differential gene expression involved in chloroplast development may be the mechanism for variegation in grapevine. In variegated peach flowers, three genes involved in flavonoid biosynthesis were found to have significantly greater expression in red petals than in white petals, and 11 transcription factors were expressed differentially, suggesting the high expression levels may be regulated by one or more of the transcription factors (Chen et al. 2014). Regulatory proteins (17.9% of candidate genes) comprised the second largest functional protein group. Future studies measuring gene and RNA expression, such as reverse transcription–quantitative polymerase chain reaction or RNA-Seq, of regulatory genes identified here may provide some insight into whether one of these regulatory genes is involved in the mechanism of variegation.
A key candidate gene, ATP-dependent zinc metalloprotease FtsH6, on chromosome 14 is a target for future analysis as a result of its similarity in loss of chloroplast protein in the A. thaliana var2 mutant (Chen et al. 2000). Its role in photosystem II in high light intensity (Zelisko et al. 2005) is especially interesting. Future work should focus on gene expression analysis to elucidate the role of this gene in WT and variegated grapes with the Lvar1 mutation. Additional research is necessary to understand Lvar1 and potential epistatic interactions in the Landot 4511 genetic background.
Conclusion
Leaf variegation in grape is controlled by at least two major loci that segregate independently in breeding germplasm. Although variegation is first noted in cotyledons and can be culled easily by the breeder, a unique phenotype with flecking variegation was noted in a Landot 4511 selfed population. Landot 4511 is likely the donor for Lvar1 and Lvar2, and may trace back through the pedigree to the variegation observed in seedlings of ‘Aramon’. The maintenance of these recessive alleles through many generations of breeding suggests linkage to yet-to-be-mapped breeding targets such as fruit or wine quality, pest resistance, or viticultural attributes. DNA tests for Lvar1 and Lvar2 should be developed with rhAmpSeq haplotypes for use in marker-assisted parental selection to aid breeding decisions either to avoid crosses that might result in variegated seedlings or to focus on such phenotypes for their ornamental quality. However, additional research may be warranted to determine any potential favorable linkage.
References
Bradbury, P., Zhang, Z., Kroon, D.E., Casstevens, T.M., Ramdoss, Y. & Buckler, E.S. 2007 TASSEL: Software for association mapping of complex traits in diverse samples Bioinformatics 23 2633 2635 https://doi.org/10.1093/bioinformatics/btm308
Broman, K.W., Wu, H., Sen, Ś. & Churchill, G.A. 2003 R/qtl: QTL mapping in experimental crosses Bioinformatics 19 889 890 https://doi.org/10.1093/bioinformatics/btg112
Canaguier, A., Grimplet, J., Di, G.G., Scalabrin, S., Duchêne, E., Choisne, N., Mohellibi, N., Guichard, C., Rombauts, S., Le, C.I., Bérard, A., Chauveau, A., Bounon, R., Rustenholz, C., Morgante, M., Le, P.M., Brunel, D. & Adam-Blondon, A. 2017 A new version of the grapevine reference genome assembly (12X.v2) and of its annotation (VCost.v3) Genom. Data 14 56 62 https://doi.org/10.1016/j.gdata.2017.09.002
Chen, M., Choi, Y., Voytas, D.F. & Rodermel, S. 2000 Mutations in the Arabidopsis VAR2 locus cause leaf variegation due to the loss of a chloroplast FtsH protease Plant J. 22 4 303 313 https://doi.org/10.1046/j.1365-313x.2000.00738.x
Chen, Y., Mao, Y., Liu, H., Yu, F. & Li, S. 2014 Transcriptome analysis of differentially expressed genes relevant to variegation in peach flowers PLoS One 9 3 e90842 https://doi.org/10.1371/journal.pone.0090842
Cingolani, P., Platts, A., Coon, M., Nguyen, T., Wang, L., Land, S.J., Lu, X. & Ruden, D.M. 2012 A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3 Fly (Austin) 6 2 80 92 https://doi.org/10.4161/fly.19695
Dahlgren, K.V.O 1921 Vererbungsversuche mit einer buntblättrigen Barbarea vulgaris Hereditas 2 88 98
Downing, A.J 1857 Fruits and fruit-trees of America 328 329 Wiley and Halsted New York, NY, USA
Eyster, W.H 1934 Genetics of Zea mays Bib. Genet. 11 187 392
Filler, D.M., Luby, J.J. & Ascher, P.D. 1994 Incongruity in the interspecific crosses of Vitis L.: Morphological abnormalities in the F2 progeny Euphytica 78 3 227 237 https://doi.org/10.1007/BF00027521
Freed, D., Aldana, R., Weber, J.A. & Edwards, J.S. 2017 The Sentieon Genomics Tools: A fast and accurate solution to variant calling from next-generation sequence data bioRxiv 115717 https://doi.org/10.1101/115717
Galzy, R. & Galzy, P. 1964 Action de la temperature sur la panachure blanche de la vigne Prog. Agric. Vitic. 162 271 280
Golicz, A.A., Bayer, P.E. & Edwards, D. 2015 Skim-based genotyping by sequencing Methods Mol. Biol. 1245 257 270 https://doi.org/10.1007/978-1-4939-1966-6_19
Ikeno, S 1927 Eine Monographie über die Erblichkeitsforschungen bei der Reispflanze Bib. Genet. 3 245 312
Karn, A., Zou, C., Brooks, S., Fresnedo-Ramírez, J., Gabler, F., Sun, Q., Ramming, D., Naegele, R., Ledbetter, C. & Cadle-Davidson, L. 2021 Discovery of the REN11 locus from Vitis aestivalis for stable resistance to grapevine powdery mildew in a family segregating for several unstable and tissue-specific quantitative resistance loci Front. Plant. Sci. 12 733899
Kirk, J.T.O. & Tilney-Bassett, R.A.E. 1978 The plastids: Their chemistry, structure, growth and inheritance Elsevier/North-Holland Amsterdam, Netherlands
Knight, T.A 1808 On the variegation of plants: In a letter to Richard Anthony Salisbury, Esq., F.R.S. and L.S., by Thomas Andrew Knight, Esq., F.R.S. and L.S Philos. Mag. Lett. 32 127 306 308 https://doi.org/10.1080/14786440808562800
Li, H 2013 Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM arXiv 1303 3997 https://doi.org/10.48550/arXiv.1303.3997
McKenna, A., Hanna, M., Banks, E., Sivachenko, A., Cibulskis, K., Kernytsky, A., Garimella, K., Altshuler, D., Gabriel, S., Daly, M. & DePristo, M. 2010 The genome analysis toolkit: A MapReduce framework for analyzing next-generation DNA sequencing data Genome Res. 20 1297 1303 https://doi.org/10.1101/gr.107524.110
Money, D., Gardner, K.M., Migicovsky, Z., Schwaninger, H., Zhong, G.Y. & Myles, S. 2015 LinkImpute – fast and accurate genotype imputation for non-model organisms G3: Genes, Genomes, Genetics 5 2383 https://doi.org/10.1534/g3.115.021667
Morgan, M 2019 BiocManager: Access the bioconductor project package repository R package version 1.30.10. https://CRAN.R-project.org/package=BiocManager. [accessed 1 Jul 2022]
Morinaga, T 1932 The chlorophyll deficiencies in rice Bot. Mag. Tokyo 46 202 207 https://doi.org/10.15281/jplantres1887.46.202
Oliphant, T.E 2006 A guide to NumPy Vol. 1 Trelgol Publishing USA
Olson, J. & Clark, M.D. 2021 Characterization of anatomical and physiological effects of variegation mutation on grapevine HortScience 56 1251 1257 https://doi.org/10.21273/HORTSCI15929-21
Picq, S., Santoni, S., Lacombe, T., Latreille, M., Weber, A., Ardisson, M., Ivorra, S., Maghradze, D., Arroyo-Garcia, R., Chatelet, P., This, P., Terral, J.-F. & Bacilieri, R. 2014 A small XY chromosomal region explains sex determination in wild dioecious V. vinifera and the reversal to hermaphroditism in domesticated grapevines BMC Plant Biol. 14 1 229 https://doi.org/10.1186/s12870-014-0229-z
Rastas, P 2017 Lep-MAP3: Robust linkage mapping even for low-coverage whole genome sequencing data Bioinformatics 33 3726 3732 https://doi.org/10.1093/bioinformatics/btx494
Rédei, G.P 1963 Somatic instability caused by a cysteine-sensitive gene in Arabidopsis Science 139 767 769
Reisch, B.I. & Watson, J.P. 1984 Inheritance of leaf variegation in Vitis species J. Hered. 75 5 417 418 https://doi.org/10.1093/oxfordjournals.jhered.a109973
Röbbelen, G 1968 Genbedingte Rotlicht-Empfindlichkeit der Chloroplastendifferenzierung bei Arabidopsis Planta 80 3 237 254 https://doi.org/10.1007/BF00392394
Stothard, P 2000 The sequence manipulation suite: JavaScript programs for analyzing and formatting protein and DNA sequences Biotechniques 28 1102 1104 https://doi.org/10.2144/00286ir01
Turner, S 2017 qqman: Q-Q and Manhattan plots for GWAS data R package version 0.1.4. https://CRAN.R-project.org/package=qqman [accessed 1 Jul 2022]
Voorrips, R.E 2002 MapChart: Software for the graphical presentation of linkage maps and QTLs J. Hered. 93 77 78 https://doi.org/10.1093/jhered/93.1.77
Yu, F., Fu, A., Aluru, M., Park, S., Xu, Y., Liu, H., Liu, X., Foudree, A., Nambogga, M. & Rodermel, S. 2007 Variegation mutants and mechanisms of chloroplast biogenesis Plant Cell Environ. 30 3 350 365 https://doi.org/10.1111/j.1365-3040.2006.01630.x
Zaumeyer, W.J 1942 The inheritance of a leaf variegation in beans J. Agric. Res. 64 119 127
Zelisko, A., García-Lorenzo, M., Jackowski, G., Jansson, S. & Funk, C. 2005 AtFtsH6 is involved in the degradation of the light-harvesting complex II during high-light acclimation and senescence Proc. Natl. Acad. Sci. USA 102 38 13699 13704 https://doi.org/10.1073/pnas.0503472102
Zou, C., Karn, A., Reisch, B., Nguyen, A., Sun, Y., Bao, Y., Campbell, M.S., Church, D., Williams, S., Xu, X., Ledbetter, C.A., Patel, S., Fennell, A., Glaubitz, J.C., Clark, M., Ware, D., Londo, J.P., Sun, Q. & Cadle-Davidson, L. 2020 Haplotyping the Vitis collinear core genome with rhAmpSeq improves marker transferability in a diverse genus Nat. Commun. 11 1 413 https://doi.org/10.1038/s41467-019-14280-1
Zou, C., Massonnet, M., Minio, A., Patel, S., Llaca, V., Karn, A., Gouker, F., Cadle-Davidson, L., Reisch, B., Fennell, A., Cantu, D., Sun, Q. & Londo, J.P. 2021 Multiple independent recombinations led to hermaphroditism in grapevine Proc. Natl. Acad. Sci. USA 118 15 e2023548118 https://doi.org/10.1073/pnas.2023548118
Pedigree showing the interrelatedness of the three populations (green boxes) used in genetic mapping and genome-wide association studies for the leaf variegation trait. These populations were also included in the bulked segregant analysis. All three populations share ‘Frontenac’ and Landot 4511 as ancestors (orange boxes), which are believed to be heterozygous for variegation.
Citation: HortScience 57, 11; 10.21273/HORTSCI16763-22
Chromosome 14 linkage group from genetic mapping of GE1642 displaying the associated locus, Lvar1, between ∼40 and 50 cM. Flanking markers are highlighted in bold, and the peak marker in red. The locus region is indicated by a green bar on the chromosome. The logarithm of the odds plot is included to show marker association across the chromosome and to highlight marker association within the locus.
Citation: HortScience 57, 11; 10.21273/HORTSCI16763-22
Distribution of single nucleotide polymorphism markers and total genetic length in parental and sex-averaged genetic maps of GE1642 (MN1200 × MN1326).
Distribution of single nucleotide polymorphism haplotype markers and total genetic length in parental and sex-averaged genetic maps of GE1703 (MN1256 × MN1327).
Candidate genes at Lvar1 based on bulk segregant analysis (BSA), genetic mapping, and significantly associated markers on chromosome 14. Some genes were found to have more than one copy and thus the number of genes is provided.
