Koshu is a grape cultivar native to Japan and is one of the country’s most important cultivars for wine making. Koshu and other oriental grape cultivars are widely believed to belong to the European domesticated grape species Vitis vinifera. To verify the domesticated origin of Koshu and four other cultivars widely grown in China and Japan, we genotyped 48 ancestry informative single nucleotide polymorphisms (SNPs) and estimated wild and domesticated ancestry proportions. Our principal components analysis (PCA) based ancestry estimation revealed that Koshu is 70% V. vinifera, and that the remaining 30% of its ancestry is most likely derived from wild East Asian Vitis species. Partial sequencing of chloroplast DNA suggests that Koshu’s maternal line is derived from the Chinese wild species V. davidii or a closely related species. Our results suggest that many traditional East Asian grape cultivars such as Koshu were generated from hybridization events with wild grape species.
Citation: Goto-Yamamoto N, Sawler J, Myles S (2015) Genetic Analysis of East Asian Grape Cultivars Suggests Hybridization with Wild Vitis. PLoS ONE 10(10): e0140841. https://doi.org/10.1371/journal.pone.0140841
Editor: Dario Cantu, Univeristy of California Davis, UNITED STATES
Received: May 14, 2015; Accepted: September 29, 2015; Published: October 21, 2015
Copyright: © 2015 Goto-Yamamoto et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: Chloroplast sequences are available from the DNA Databank of Japan DDBJ[LC054004 - LC054031]. All other data are within the paper and its Supporting Information files.
Funding: This work was funded by a Grant-in Aid for Scientific Research (C) No. 24580063 from Japan Society for the Promotion of Science to NG. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Co-author Jason Sawler is employed by Anandia Labs. Anandia Labs provided support in the form of salaries for author JS, but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific role of this author is articulated in the ‘author contributions’ section.
Competing interests: The authors have the following interests: Co-author Jason Sawler is employed by Anandia Labs. There are no patents, products in development or marketed products to declare. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Grapes are one of the most important horticultural crops worldwide, and are typically consumed as fresh fruit or used in the production of wine, brandy, juice and raisins. The majority of wine grapes are cultivars of the domesticated grape V. vinifera, which originated in the Caucasus and spread from there to Europe and eventually to grape growing regions worldwide . Cultivars grown for sale as table grapes or juice are often interspecific hybrids of V. vinifera and wild species (e.g. V. labrusca). There are an estimated 60 species in the genus Vitis distributed broadly across the entire northern hemisphere, and many wild grapevines are used to breed new cultivars since they harbor desirable traits like disease resistance and tolerance to abiotic stress .
Several grape cultivars from China and Japan are widely grown for commercial purposes. For example, Koshu is one of Japan’s most widely planted and popular grape cultivars with approximately 1200 acres under vine in 1997 . It is known for its distinct pale purple skin containing both cyanidin-based and delphinidin-based anthocyanins , but not anthocyanin diglucosides . Wine produced from Koshu has aromatic characteristics similar to Sauvignon blanc  and has contributed significantly to the growth of the Japanese wine industry.
It is widely believed that Koshu is an oriental cultivar of V. vinifera  and previous analyses have supported this hypothesis. For example, a phylogeny generated using simple sequence repeats (SSRs) showed that Koshu clusters with well-known V. vinifera cultivars such as Sultanina and Muscat of Alexandria . Similar results were observed from analyses of AFLPs  and principal coordinate analysis of SSR data . In contrast, there is also evidence that Koshu may have ancestry from wild Vitis species. For example, Koshu has the SSR allele 4MG1 or n+4 at the VVS2 locus , which has only been found in rootstock cultivars, wild species, or cultivars crossed with wild Vitis (Boursiquot JM, personal communication). Similarly, Koshu and other East Asian cultivars possess unique SSR alleles that are not found in European V. vinifera cultivars [9, 10].
To verify the ancestry of popular East Asian cultivars on a genome-wide scale, we genotyped a panel of 48 ancestry informative markers (AIMs) in the Japanese cultivars Koshu and Koshu-sanjaku, and the Chinese cultivars Longan, Huotianhong, Baijixin. The AIMs are a subset of SNPs from the Vitis9kSNP array  that clearly differentiate wild Vitis species from V. vinifera. Our ancestry estimates are based on a method developed by Sawler et al. , who showed that genotyping fewer than 50 SNPs leads to accurate estimates of the genomic contributions of V. vinifera and North American wild Vitis in interspecific hybrid grape cultivars from the USDA grape germplasm collection. Here we apply the same method to quantify the potential genomic contribution of wild East-Asian Vitis to popular East Asian cultivars. In addition, the maternal ancestry of these cultivars was investigated by partially sequencing their chloroplast DNA and comparing them to published sequences.
Materials and Methods
Marker selection and genotyping
Forty-eight AIMs were selected as those with the highest loadings along the first principal component (PC1) from PCA performed on 6114 SNPs genotyped in 1031 samples labeled as V. vinifera and 786 labeled as wild Vitis grapevine accessions from the USDA grape germplasm collection . For each SNP, pairs of PCR primers were designed based on 100 bp flanking each side of the SNP. The amplified fragments were purified with a Nucleospin Gel and PCR Clean-up kit (Macherey-Nagel, Düren, Germany) and sequenced directly, or after gel electrophoresis and extraction from the gel when necessary. The genotyped SNPs and their primer sequences according to the 8x Pinot Noir reference genome  are shown in S1 Table.
The ten DNA samples genotyped for the AIMs in this study are Koshu, Pinot Noir, Koshu-sanjaku, Longan, Huotianhong, Baijixin, V. amurensis, V. coignetiae, V. ficifolia var. ganebu, and V. shiragai. The full genotype data are shown in S2 Table. Among them V. amurensis, V. coignetiae, V. ficifolia var. ganebu, and V. shiragai were extracted using standard methods from young leaves donated by Prof. Horiuchi and Dr. Mochioka of Osaka Prefecture University. All other DNA was extracted from young leaves of vines in an experimental vineyard in Higashi-Hiroshima.
For the analysis of chloroplast DNA, partial chloroplast sequences were obtained from the following six cultivars: Koshu, Koshu-sanjaku, Longan, Huotianhong, Baijixin, and Chardonnay.
Principal components analysis and ancestry estimation
PCA of the 48 AIMs was first performed using 33 V. vinifera cultivars and 33 accessions of East-Asian wild Vitis species. These two groups of accessions are considered ancestral populations for the purposes of this study. Except for four East Asian wild Vitis samples obtained for this study, the genotypes for these ancestral groups were obtained from previously published data from the Vitis9KSNP genotyping microarray [14, 15]. Genotype data from the 48 AIMs were obtained for this study by genotyping four East Asian wild species (V. amurensis, V. coignetiae, V. ficifolia var. ganebu, and V. shiragai), five oriental cultivars (Koshu, Koshu–sanajku, Longan, Huotianhong, Baijixin) and one V. vinifera (Pinot noir). After establishing the PC axes using the 33 V. vinifera and 33 wild accessions described above as ancestral populations, the remaining samples were subsequently projected onto the first two PCs. The proportion of V. vinifera ancestry in the samples genotyped as part of this study was then calculated as follows: ‘% V. vinifera = b/(a+b)’, where a and b are the chord distances along the first principal component from the centroids of the V. vinifera cultivars and wild species in PC space, respectively, according to .
Partial sequencing of chloroplast DNA
Following the method of Zecca et al.  with minor modification, four spacer regions of chloroplast DNA (trnH-psbA, trnK-rps16, trnF-ndhJ, and rpl32-trnL) were amplified and sequenced. The PCR and sequencing primers are shown in S3 Table. The sequences obtained from this have been submitted to DDBJ [LC054004—LC054031].
The reported sequences of wild Vitis species and those of V. vinifera subsp. sylvestris (HQ656029-HQ656577) were downloaded from GenBank and compared with the sequence obtained in this study.
Phylogenetic analysis of chloroplast sequence
The wild species for which sequence data were unobtainable from one or more of the four chloroplast genomic regions were omitted from the analysis. The downloaded sequences were trimmed to the same region of the sequence obtained in this study after alignment with ClustalW. The total resulting sequence length was 2081 bp including gaps. For each accession, the four regions were joined and a dendogram was constructed using the Neighbor-Joining (NJ) method in MEGA6 . Distances were calculated with Kimura’s 2 parameter model and substitutions included transitions and transversions. Gaps and missing data were treated as pairwise deletions. Bootstrap confidence values were calculated from 100 bootstrap replications. From this distance matrix, principal coordinate analysis (PCoA) was also carried out using GenAlEx  with standardization.
PCA based ancestry estimation
PCA of the two ancestral populations, wild East Asian species and V. vinifera, resulted in a clear separation of these two groups along PC1. Samples from the four East Asian wild species (V. amurensis, V. coignetiae, V. ficifolia var. ganebu, and V. shiragai) genotyped as part of this study all clustered with the East Asian wild ancestral population genotyped in . The Pinot Noir sample genotyped as part of this study also clustered as expected with the 33 V. vinifera cultivars genotyped in  and had an estimated V. vinifera ancestry proportion of 1.0. The V. vinifera ancestry proportions for the five East Asian cultivars genotyped as part of this study had values ranging from 0.715 to 1 (Fig 1).
Analysis of chloroplast diversity
We find that the partial chloroplast sequence of Chardonnay is identical to the reported sequence of V. vinifera cv. Maxxa (NC_007957.1) except for one site at position 205, as reported by Tabidze et al . Differences were observed between Koshu and Chardonnay at 8 sites (Table 1). All other oriental cultivars share an identical chloroplast sequence with Chardonnay (V. vinifera).
Nucleotide positions of Maxxa (NC_007957.1) are shown as the reference.
A PCoA plot based on a chloroplast genetic distance matrix enables the relationships among the accessions’ chloroplast sequences to be visualized in two dimensions (Fig 2). Fig 2 shows that Koshu is closest to the Asian wild species V. davidii and V. flexuosa. Because all other oriental cultivars share an identical chloroplast sequence with Chardonnay, they are represented in the plot as a single point. A neighbor-joining tree also supports the notion that Koshu derives its chloroplast DNA from an Asian wild species, as it clusters most closely with V. davidii (S1 Fig).
The ancestry of some of the most common East Asian grape cultivars is often claimed to be derived exclusively from the domesticated grape, V. vinifera, which originated in the Caucasus region of Central Asia. Studies investigating these claims have so far been limited in marker number and/or sample size. Here we use a set of genome-wide ancestry informative SNPs and chloroplast DNA sequences to investigate the potential contribution of wild grape species to the genomes of present-day East Asian grape cultivars.
Accessions of wild Asian grape species (V. amurensis, V. coignetiae, V. ficifolia var. ganebu, and V. shiragai) genotyped for the present study clustered closely in PC space with wild Asian species genotyped in a previous study using the Vitis9KSNP array  (Fig 1). In addition, the V. vinifera accession we genotyped here as a control (Pinot Noir) also clustered with the V. vinifera accessions genotyped previously . Thus, we are confident that the genotype data obtained from the AIMs we identified are concordant with genotype data generated from the Vitis9KSNP array and that the use of merged data for the purposes of ancestry assignments is justified.
Our analysis of 48 AIMs in the Japanese grapes Koshu and Koshu-sanjaku suggests that 20–30% of their genomes are derived from wild Vitis species (Fig 1). Thus, we conclude that Koshu and Koshu-sanjaku are not cultivars of V. vinifera, in contradiction to previous work [3, 7–9]. It is likely that hybridization with wild Vitis, intentional or by chance, occurred at some point in the history of these commercially successful Japanese wine grapes. We find that the chloroplast genome of Koshu appears most similar to V. davidii and not V. vinifera (Fig 2 and S1 Fig), indicating that at least one of its East Asian ancestors was a female plant.
Although we can safely conclude that Koshu is an interspecific hybrid, we cannot conclusively identify the type or number of wild species in its lineage as these species may be extinct, unidentified, or simply missing from the present analysis.
It is also worth noting that, based on our analysis of partial chloroplast DNA sequences, accessions from a single Vitis species did not always cluster together in genetic space: it was often the case that accessions from a single species were more genetically distant than accessions from different species (Fig 2 and S1 Fig). This may be due to curation error, or it may the case that there is incomplete lineage sorting: the partial chloroplast sequences studied here may segregate within and among Vitis species. The clear separation of North American and Asian wild Vitis species, together with the close relationship between V. vinifera and its wild ancestor V. vinifera subsp. sylvestris, provide support that the genetic distances based on chloroplast DNA sequences at least partially capture the evolutionary relationships among the samples studied here (Fig 2 and S1 Fig). A more comprehensive analysis of chloroplast DNA sequences is required to refine the relationship between Koshu and its chloroplast ancestor.
The errors associated with the PCA-derived ancestry estimates presented here provide support to our conclusions and refine the possibilities of how wild introgression took place in the history of commercial East Asian grapes. Using simulated offspring from F1 and F2 hybrids between V. vinifera and wild Vitis species, Sawler et al.  used the same PCA-based approach to accurately estimate the simulated offsprings’ V. vinifera ancestry with 95% confidence intervals of approximately ± 0.04. Assuming similar variation in estimates in the present study, Koshu (71.5% V. vinifera) could therefore be F2 plants derived from an F1 hybrid parent (V. vinifera x East Asian wild species) and a V. vinifera parent. We cannot rule out the possibility, however, that this cultivar has a more complex pedigree including wild Asian Vitis and V. vinifera. Koshu-sanjaku (81.8% V. vinifera) has a higher estimated proportion of V. vinifera ancestry than Koshu and, according to the estimated error, is unlikely to be an F2 hybrid as described above. Thus, Koshu-sanjaku is more likely to be the result of a complex pedigree with V. vinifera and East Asian wild Vitis relatives. Longan (91.7% V. vinifera) and Huotianhong (90.4% V. vinifera) both cluster close to the V. vinifera population in Fig 1. They may be genetically distinct V. vinifera cultivars or possess a very small proportion of wild Vitis ancestry. Finally, we determined that the Chinese cultivar Baijixin (100% V. vinifera) is in fact most likely a V. vinifera cultivar. With sufficiently dense genotype data from a comprehensive sample of V. vinifera and oriental cultivars, it may be possible in the future to identify the V. vinifera cultivars that contributed to the ancestry of commercial oriental cultivars.
Taking into account the results of the SNP based ancestry estimation and observed genetic relatedness to wild Vitis based on chloroplast sequencing, the simplest hypothetical origin of Koshu is that it originated from a cross between an F1 hybrid (V. davidii or related x V. vinifera) and a second V. vinifera cultivar. V. davidii is a synonym of V. armata, and has a wide geographic distribution in China . Cultivars of V. davidii are grown in China, and are used for both wine making as well as breeding stock. The branches of V. davidii possess many spines, and it is sometimes referred to as “spiny Vitis”. The young shoots of Koshu have small spines that may be a result of V. davidii ancestry. In addition, V. davidii is highly resistant to disease and is tolerant to wet climates . These traits may have contributed to the success of Koshu given the wet and hot summers in Japan. However, nearly all anthocyanins of V. davidii are delphinidin-based and diglucosides , which differ from those produced by Koshu. As previously discussed, V. davidii is only one of many possible ancestors of Koshu, and further work examining more samples and markers will be required to confirm or refute the conjectures outlined here.
Studies of European grape cultivars have demonstrated introgression from local wild vines as humans brought grapes into Europe . While our data do not enable us to determine the timing of the hybridization events that led to the East Asian cultivars studied here, there was likely ample opportunity for wild introgression during the journey of the domesticated grape along the Silk Road trade route to Japan. Further studies are required to refine the timing and precise nature of the hybridization events that led to the grapes currently grown commercially in East Asia.
S1 Fig. NJ tree based on partial sequence of chloroplast DNA.
Labels on the branches are bootstrap confidence values. The letters following the species names refer to the IDs used in .
S1 Table. Primer sequence for the 48 ancestry informative SNPs genotyped.
The SNP ID is composed of the chromosome, followed after the colon by the position according to the 8x Pinot Noir (PN40024) genome sequence .
S2 Table. SNPs of 10 grape accessions for 48 AIMs.
The SNP ID is composed of the chromosome, followed after the colon by the position according to the 8x Pinot Noir (PN40024) genome sequence . When only a single allele appears, it indicates that the sample is homozygous for that allele.
The authors thank Shosaku Horiuchi, and Ryosuke Mochioka for providing wild grapevine samples for analysis. Special thanks for Mineyo Numata, Midori Joyo and Hiromi Kono for technical assistance. This work was partially funded by a Grant-in Aid for Scientific Research (C) No. 24580063 from Japan Society for the Promotion of Science to NG. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Conceived and designed the experiments: NG SM. Performed the experiments: NG. Analyzed the data: NG JS. Contributed reagents/materials/analysis tools: NG SM. Wrote the paper: NG JS SM.
- 1. McGovern PE. Ancient Wine: The Search for the Origins of Viniculture. Princeton, NJ: Princeton University Press; 2003.
- 2. Owens C. Grapes. In: Hancock J, editor. Breeding Temperate Fruit Crops: Germplasm to Genomics: Springer Netherlands; 2008. p. 197–233.
- 3. Morinaga K. Grape Production in Japan Bangkok, Thailand: Food and Agriculture Organization of the United Nations; http://www.fao.org/docrep/003/x6897e/x6897e07.htm; 2001 [cited 2015]. Available from: http://www.fao.org/docrep/003/x6897e/x6897e07.htm.
- 4. Kobayashi H, Suzuki S, Tanzawa F, Takayanagi T. Low Expression of Flavonoid 3′,5′-Hydroxylase (F3′,5′H) Associated with Cyanidin-based Anthocyanins in Grape Leaf. American Journal of Enology and Viticulture. 2009;60(3):362–7.
- 5. Yokotsuka K, Nozaki K, Kushida T. Comparison of Phenolic Compounds Including Anthocyanin Pigments between koshu and Ryugan Grapes: Chemical Stidues on Coloring and Flavoring Substances in Japanese Grapes and Wines (XXI). Journal of fermentation technology. 1984;62(5):477–86.
- 6. Kobayashi H, Takase H, Kaneko K, Tanzawa F, Takata R, Suzuki S, et al. Analysis of S-3-(Hexan-1-ol)-Glutathione and S-3-(Hexan-1-ol)-l-Cysteine in Vitis vinifera L. cv. Koshu for Aromatic Wines. American Journal of Enology and Viticulture. 2010;61(2):176–85.
- 7. Goto-Yamamoto N, Mouri H, Azumi M, Edwards KJ. Development of Grape Microsatellite Markers and Microsatellite Analysis Including Oriental Cultivars. American Journal of Enology and Viticulture. 2006;57(1):105–8.
- 8. Goto-Yamamoto N. Phenetic clustering of grapes (Vitis spp.) by APLP analysis. Breeding Science. 2000;50(1):53–8.
- 9. Goto-Yamamoto N, Azuma A, Mitani N, Kobayashi S. SSR Genotyping of Wild Grape Species and Grape Cultivars of Vitis vinifera and V. vinifera × V. labrusca. J Japan Soc Hort Sci. 2013;82(2):125–30.
- 10. Goto-Yamamoto N, Numata M, Wan G-H, Shimamoto T, Hashizume K. Characterization of oriental cultivars of grapevine using a reference allele system of microsatellite data and assignment test. J Japan Soc Hort Sci. 2009;78(2):175–9.
- 11. Myles S, Chia J-M, Hurwitz B, Simon C, Zhong GY, Buckler E, et al. Rapid Genomic Characterization of the Genus Vitis. PLoS ONE. 2010;5(1):e8219. pmid:20084295
- 12. Sawler J, Reisch B, Aradhya MK, Prins B, Zhong G-Y, Schwaninger H, et al. Genomics Assisted Ancestry Deconvolution in Grape. PLoS ONE. 2013;8(11):e80791. pmid:24244717
- 13. The French-Italian Public Consortium for Grapevine Genome Characterization. The grapevine genome sequence suggests ancestral hexaploidization in major angiosperm phyla. Nature. 2007;449(7161):463–7. pmid:17721507
- 14. Miller A, Matasci N, Aradhya MK, Prins BH, Zhong GY, Schwaninger H, et al. Vitis phylogenomics: hybridization intensities from a SNP array outperform genotype calls. PLoS ONE. 2013;8:e78680. pmid:24236035
- 15. Myles S, Boyko AR, Owens CL, Brown PJ, Grassi F, Aradhya MK, et al. Genetic structure and domestication history of the grape. PNAS. 2011;108(9):3530–5. pmid:21245334
- 16. Zecca G, Abbott JR, Sun W-B, Spada A, Sala F, Grassi F. The timing and the mode of evolution of wild grapes (Vitis). Molecular Phylogenetics and Evolution. 2012;62(2):736–47. pmid:22138159
- 17. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. MEGA6: Molecular Evolutionary Genetics Analysis version 6.0. Molecular Biology and Evolution. 2013.
- 18. Peakall R, Smouse PE. GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research—an update. Bioinformatics. 2012;28(19):2537–9. pmid:22820204
- 19. Tabidze V, Baramidze G, Pipia I, Gogniashvili M, Ujmajuridze L, Beridze T, et al. The complete chloroplast DNA sequence of eleven grape cultivars. Simultaneous resequencing methodology. J Int Sci Vigne Vin. 2014;48:99–109.
- 20. Wan Y, Schwaninger H, Li D, Simon C, Wang Y, He P. The eco-geographic distribution of wild grape germplasm in China. Vitis. 2008;47(2):77–80.
- 21. Liang N-N, Pan Q-H, He F, Wang J, Reeves MJ, Duan C-Q. Phenolic profiles of Vitis davidii and Vitis quinquangularis species native to China. J Agric Food Chem. 2013;61(25):6016–27. pmid:23721215