Characterization of Grapevine Accessions from Ukraine Using Microsatellite Markers

Transcription

1 Characterization of Grapevine Accessions from Ukraine Using Microsatellite Markers Myriam Heuertz, 1,2 * Svitlana Goryslavets, 3 Jean-François Hausman, 1 and Valentina Risovanna 3 Abstract: Forty-seven wine and table grape accessions representing cultivars widely grown in Ukraine and 28 reference accessions were genotyped at six microsatellite loci (VVS2, VVMD5, VVMD7, VVMD27, VrZAG62, and VrZAG79) in order to solve accession labeling problems among Ukrainian accessions, characterize their genetic diversity, and establish genetic relationships with other European grapevine cultivars. Genotypes were standardized to meet the requirements of the European Vitis Database. Genetic diversity (H E = 0.845) and allelic richness (A S = 8.09 alleles/locus/random sample of 16 individuals) were found to be higher in the Ukrainian collection than in any other geographical region, based on previously published studies on grapevine accessions. No geographic structure was identified in the total data set. However, winegrapes were significantly differentiated from table grapes and a Bayesian method detected two genetic clusters, one containing mainly table and Muscat grapes and the other mainly winegrape accessions from throughout Europe, including Ukraine. These results highlight the contribution of Muscat and European winegrapes in the breeding design of Ukrainian varieties and suggest that selection and historical and contemporary movement of germplasm are major factors shaping the structure of the grapevine gene pool. The genetically rich and diversified gene pool of Ukrainian varieties represents valuable source material for future sustainable breeding and improvement of grapevine. Key words: grapevine germplasm, Ukraine, microsatellites, genetic diversity A major problem in the management of grapevine germ plasm collections is the occurrence of genetically identical accessions carrying different names (synonyms) and different accessions carrying the same names (homonyms) (Lopes et al. 1999, Martín et al. 2003). This issue is particularly relevant for germplasm collections of countries in economic transition, such as those of the Black Sea and Caucasus regions, where curation and conservation efforts have been difficult in the last decades. Nonetheless, the accessions of just these collections are of particular interest for research on grapevine domestication, breeding, and improvement of modern grapevine 1 Centre de Recherche Public-Gabriel Lippmann, rue du Brill 41, L-4422 Belvaux, Luxembourg; 2 Université Libre de Bruxelles, Behavioural and Evolutionary Ecology cp160/12, av. F.D Roosevelt 50, B-1050 Brussels, Belgium; and 3 Institute of Vine and Wine Magarach, Kirov Street 31, Yalta, Ukraine. *Corresponding author ( mheuertz@ulb.ac.be) Acknowledgments: The authors thank Didier Vares of INRA, Vassal, France, for kindly providing the reference material. Elodie Boland and Laurent Solinhac are acknowledged for help in the laboratory. This work was funded by the Luxembourg Ministry of Finance and by Bioversity International (formerly IPGRI, the International Plant Genetic Resources Institute). M. Heuertz is currently a postdoctoral researcher financed by the National Fund for Scientific Research of Belgium. The online version of this article contains freely available supplemental data. Manuscript submitted October 2007; revised January Publication costs of this article defrayed in part by page fees. Copyright 2008 by the American Society for Enology and Viticulture. All rights reserved. 169 cultivars, as well as in local viticulture and wine production. The region close to the Black and Caspian Seas is home to an extraordinary phenotypic diversity of wild and cultivated grapevines (Levadoux 1956, Maghradze 2005), and evidence of grapevine cultivation in the Caucasus, eastern Taurus, and northern Zagros mountains dates back 8000 to 6000 years (Unwin 1991, Zohary 1995, Olmo 1996, McGovern 2003). Similarly to germplasm management, breeding of new wine and table grape varieties with high-yielding capacity requires methods of accurate identification, which are free from subjective evaluation, in order to certify products and to protect the rights of breeders if appropriate. Nuclear microsatellite markers have recently proved to be useful in the identification of individual grapevine genotypes in the context of germplasm management (Lopes et al. 1999, Martín et al. 2003) and parentage analysis (Bowers et al. 1999a). They have also allowed characterizing diversity and genetic structure in the cultivated grapevine gene pool (Sefc et al. 2000, Aradhya et al. 2003). Only a relatively small fraction of diversity (up to 9 to 15%) was found between European wine-producing regions or groups of cultivars (Sefc et al. 2000, Aradhya et al. 2003), suggesting that cultivated grapevine represents a single complex gene pool, the structure of which is determined by strong artificial selection and vegetative mode of propagation. Several recent studies have analyzed grapevine accessions from the Black Sea and Caucasus areas (Lefort et al. 2003, Vouillamoz et al. 2006) and have identified a high number of unique genetic profiles within germplasm collections as well as

2 170 Heuertz et al. generally higher genetic diversity compared to Western European grapevine accessions, emphasizing the richness of these ampelographic germplasms. Genetic data from the maternally inherited chloroplast DNA support the Caucasus region as a center of domestication based on high within-population genetic diversity in wild grapevines (Grassi et al. 2006, Arroyo-García et al. 2006), but suggest at least one further important origin of cultivated grapevine germplasm in the western Mediterranean area (Imazio et al. 2006, Arroyo-García et al. 2006). Ukraine has a long tradition in viticulture and vine breeding. Fifty-nine Ukraine-bred and 12 autochthonous varieties are now broadly cultivated, as well as 46 foreign varieties. About 70% of the varieties are used for wine, and the remaining for table grape production. Approximately 3850 grapevine accessions are conserved in five ampelographic collections, the most important of which are the Magarach Institute for Vine and Wine, Yalta (FAO institute code UKR050, 3300 accessions) and the Tairov Institute of Viticulture and Wine, Odessa (UKR039, 433 accessions). Little information is available on the genetic diversity of Ukrainian varieties and their relationships with varieties of other countries, although the high genetic diversity in a collection of Crimean, Moldovan, and Russian grapevine varieties has been highlighted (Lefort et al. 2003). In this study, we characterize 47 accessions conserved at the Magarach Institute using six microsatellite loci recommended for grapevine genetic fingerprinting by Bioversity International ( formerly IPGRI) and the European Vitis Database ( This et al. 2004). These accessions are representative of grapevine varieties cultivated in Ukraine; most of them have been bred from the 1920s to the early 1980s, and 42 have been genotyped for the first time in this study. The aim of this paper is to contribute to labeling accessions in the Magarach collection, estimate the genetic diversity of Ukrainian grapevine varieties, and describe their genetic relationships to other European cultivars. Materials and Methods Plant material and DNA extraction. A total of 75 grapevine accessions were analyzed (Table 1; supplementary Tables 1 and 2), including 47 Ukrainian wine and table grape accessions from the Magarach Intitute, and 28 reference accessions from a collection of rootstock and grape accessions composed to contain a maximum allelic diversity to develop a standard set of microsatellite reference alleles (This et al. 2004). The Ukrainian collection included 37 Vitis vinifera and 10 accessions bearing crossings with other Vitis species, mostly V. rupestris and the highly frost-resistent V. amurensis, in their breeding information (Antei magarachskii, Atlant, Intervitis magaracha, Krymchanin, Nakhodka, Nimrang magaracha, Rubin Golodrigi, Seyve villard , Spartanets magaracha, and Tsitronnyi magaracha, see Table 1); hereafter these 10 accessions are referred to as hybrid accessions. Duplicates of some accessions were obtained from the Tairov Institute (supplementary Table 1). DNA from the Ukrainian material was extracted from leaves according to previous methods (Lefort and Roubelakis-Angelakis 2001). The reference varieties comprised 13 rootstocks representing different Vitis species and interspecific crosses (supplementary Table 2) and 15 wine or table varieties, of which 13 were V. vinifera and two were hybrid accessions (Jacquez and Salvador [Seibel128]). Cuttings were kindly provided by Didier Vares of the Institut National de la Recherche Agronomique, Vassal, France. DNA was extracted from buds with the NucleoSpin Plant Kit (Macherey Nagel, Düren, Germany) according to the manufacturer s instructions. Microsatellite analysis. Six grapevine microsatellite loci were assessed as recommended for compatibility with the European Vitis Database ( vitis/, This et al. 2004): VVS2 (Thomas and Scott 1993), VVMD5 and VVMD7 (Bowers et al. 1996), VVMD27 (Bowers et al. 1999b), and VrZAG62 and VrZAG79 (Sefc et al. 1999). Polymerase chain reaction (PCR) was performed separately for each locus using 0.05 μl of Fideli- Taq DNA polymerase (USB Corp., Cleveland, OH), 0.8 μl dntp mix (containing 25 mm of each dntp), 1 μl of each forward and reverse primer (4 μm), 1 μl DNA extract diluted 40 times, and 1X reaction buffer (USB Corp.) in a total volume of 10 μl. PCR conditions were as follows: 95 C for 4 min, 40 cycles of 94 C for 30 sec, 52 C for 20 sec, 72 C for 1 min, followed by a final extension of 7 min at 72 C and cooling to 4 C. The forward primer of each locus was labeled with a fluorescent dye (supplementary Table 1), allowing visualization of PCR products from all loci of an individual together in the same run on an ABI Prism 310 DNA sequencer (Applied Biosystems, Foster City, CA). Alleles were scored by eye using GeneMapper 3.0 (Applied Biosystems) and allele sizes were recorded in bp with two decimal precision. After binning of alleles with Microsoft Excel, allele sizes were standardized to the reference sizes of the European Vitis Database (This et al. 2004). Genotyping error rate. Since microsatellite genotyping errors might bias the estimation of population genetics parameters, a genotyping error rate should be assessed (Bonin et al. 2004). All data were read twice and double-checked to identify and eliminate errors that had occurred during scoring and data entry. Thirty individuals (29 diploids and one potential tetraploid) were genotyped twice. Allelic mismatches were counted by comparing the two genotypes, and an error rate was estimated as the number of errors per allele, that is, the number of incorrect alleles divided by the total number of alleles. Data analysis. The genotypes of 18 additional reference grapevine accessions were downloaded from the European Vitis Database and included into the data analysis (supplementary Table 1). The discrimination power D is an estimate of the probability that two randomly sampled accessions could be distinguished at their SSR genotypes.

3 Characterization of Grapevine Accessions from Ukraine 171 The discrimination power D T over all L = 6 loci was computed as D T = 1 Π P l 2,1 l=1 i=1 where P l,1 is the frequency of genotype I at locus l (Martín et al. 2003). The software Gi m l e t ver (Valière 2002) was used to identify identical multilocus genotypes among accessions, and a single copy of identical genotypes was retained for further population genetic analysis. Compatible first-degree parentage relationships between accessions were identified with Gi m l e t and compared to information on the breeding design of accessions. The polymorphism of groups of accessions was characterized as Nei s gene diversity H E and as allelic richness As, the number of alleles expected in a sample of given size, using Fs t a t ver (Goudet 1995). Allelic richness As allows comparing the number of alleles observed in samples of different sizes. The genetic structure of the data set was assessed with the model-based clustering algorithm implemented in St r u c t u r e ver. 2.1 (Pritchard et al. 2000, Falush et al. 2003). This program clusters individual accessions probabilistically according to the similarity of their multilocus genotypes. Ten runs with a burn-in of and a run length of iterations were performed for an imposed number of clusters ranging from K = 1 to K = 7. We used the admixture model, letting individual accessions have ancestry in more than one cluster and allowed correlation of allele frequencies among clusters. Differentiation between groups of accessions was estimated with F ST according to Weir and Cockerham (1984) using Spa g edi (Hardy and Vekemans 2002). Nested analyses of molecular variance (AMOVA) taking into account the origin and use (wine or table grapes), were performed with Arl e q u i n v. 3.1 ( In order to visualize the genetic distances between accessions, a factorial correspondence analysis (FCA) was carried out using g e n e t i x ver (www. genetix.univ-montp2.fr/genetix/intro.htm). The software p o p u l a t i o n s ver beta2 ( project/filelist.php?group_id=84) was used to construct a Neighbor joining cladogram of individuals based on pair-wise D A genetic distance (Nei et al. 1983); an efficient pair-wise distance to obtain an evolutionary correct topology (Takezaki and Nei 1996). One thousand bootstraps over loci were performed to assess significance of the tree topology. L Results In order to assess the genotyping error rate, 30 individuals (29 diploids and one potential tetraploid) were genotyped twice, which resulted in the comparison of 349 allele scores. Among these, two allelic dropouts were observed; the errors were confirmed by reamplification, which translates into an error rate of 0.57% per allele. n The total discrimination power over loci was D T = Three or four alleles per locus were observed in one potential tetraploid, Polivitis magaracha (L. Kireeva, personal communication, 2006; supplementary Table 1), which was excluded from further analysis. Among the remaining 92 accessions, 87 multilocus genotypes were identified. Identical genotypes were found in both Cabernet Sauvignon accessions (reference and Ukrainian), in both Sukholimanskii belyi accessions (from Magarach and from Tairov), in both Muskat tairovskii accessions (from Magarach and Tairov), in both Yantar oskhi accessions (from Magarach and Tairov), and in Danko and Sverkhrannii bessemyannyi. Lesya accessions from the Magarach and Tairov Institutes differed at five of six loci; both Zor ka accessions differed by one allele at each of two loci. For 10 Ukrainian accessions, microsatellite data identified one or more putative parent pairs in the data set. However, the only case where microsatellite geno types of offspring and both parent accessions suggested from the breeding design were available was Rubinovyi magaracha (= Cabernet Sauvignon x Saperavi), and the microsatellite data was compatible with the relationship suggested from the breeding design. The total number of alleles detected over the six loci was 110 in the overall data set, with 56 in the Ukrainian accessions, 67 in the (deliberately allele-rich) reference grape accessions (52 and 62, respectively, when hybrid accessions were excluded), and 72 alleles in the rootstocks. Thirty-eight alleles were specific to the reference rootstocks and 10 (eight excluding hybrids) to the reference grape accessions. Among the Ukrainian accessions, three new alleles were found compared with previous reference alleles (This et al. 2004): allele 224 (n+2) at VVMD5 and 192 (n+17) at VVMD27 of Vitis vinifera origin and 181 (n+7) at VrZAG62 potentially of interspecific origin (new Ukrainian alleles are underlined in supplementary Table 1). Ukrainian grapevine accessions displayed an allelic richness of As = 8.09 alleles per locus in a random sample of 16 accessions (Table 2). In the reference varieties, As = 9.10 among cultivated grape accessions and As = among rootstocks. The gene diversity was similarly high in Ukrainian accessions, H E = 0.845, and reference grape accessions, H E = 0.831, and slightly higher in rootstocks, H E = The St r u c t u r e algorithm (Pritchard et al. 2000) identified the highest posterior probability of the data for K = 3 clusters [Ln P(D) = ± (SD)] and separated clearly the reference rootstock accessions from all others (cluster 3, Figure 1). St r u c t u r e allowed confirming the interspecific hybrid origin of four out of 10 hybrid accessions, as suggested from breeding information [criterion of >10% ancestry in the rootstock gene pool using K = 2 clusters, Ln P(D) = ± 9.68 (SD)]: Jacquez and Salvador (Seibel128) from the reference accessions and Intervitis magaracha and Krymchanin from the Ukrainian accessions. When rootstocks were removed, St r u c t u r e identified K = 2 clusters [Ln P(D) = ± 4.86 (SD)],

4 172 Heuertz et al. Table 1 Ukrainian grapevine accessions genetically characterized in this study and their parentage information. Hybrid means that non-vinifera material was used in the breeding design; it is indicated in the parentage information. Genus/Species Variety (Ukrainian) Variety (English) All possible synonyms Berry color b Use c Local institute accession number Position in the UKR050 collection Parentage Year of breeding 1 a Vitis hybrid Антей магарачский Antey magarachskii Magarach B TG,WG IVM00002 XXII-7 Rubinovyi Magaracha x Magarach ([Seyve villard 20347: Phylloxera-resistant interspecific hybrid selected by Seyve Villard, France] x pollen mix of V. vinifera) Vitis hybrid Атлант Atlant Magarach W TG IVM00019 XXVIII-90 Katta Kurgan x Magarach (interspecific hybrid) Vitis vinifera L. Бессемянный Магарача Bessemyanniy Magaracha Magarach W TG IVM00036 I-119 (Katta Kurgan x Kirovabadskii stolovyi) x Sverkhrannii bessemyannyi Vitis vinifera L. Бессемянный Мельника Bessemyannyi Mel nika W TG IVM00303 IV-119 Chaush x Kishmish chernyi Vitis vinifera L. Каберне Совиньон Cabernet Sauvignon B WG IVM02711 III-IV-11 Cabernet franc x Sauvignon blanc 6 Vitis vinifera L. Данко Danko B WG IVM00005 XIX-8 Limberger x Rkatsiteli Vitis hybrid Интервитис Магарача Intervitis Magaracha Magarach B TG IVM00015 XXVIII-107 Magarach P (Katta Kurgan x Shabash krupnoyagodnyi) x Magarach (V. rotundifolia x pollen mix of V. vinifera) Vitis vinifera L. Изумрудный Izumrudnyi W WG IVM00659 IV-96 Bastardo x Muskat de Hambourg Vitis vinifera L. Кастель черный Kastel chernyi B WG IVM00747 I-65 unknown 10 Vitis vinifera L. Кирмизи сап судакский Kirmizi Sap Sudakskii W TG IVM00786 I-70 unknown 11 Vitis vinifera L. Кишмыш Магарача Kishmish Magaracha Magarach W TG IVM00038 I-118 Magarach (Kata-Kurgan x Kirovobadskii stolovuii) x Sverchrannii bessemyannyi Vitis hybrid Крымчанин Krymchanin B WG IVM00016 XXVIII-109 Seyve Villard (Phylloxera-resistant interspecific hybrid selected by Seyve Villard, France) x Dzhalita Vitis vinifera L. Леся Lesya Леся Украинка, Украинка P TG - - Nimrang x Zhemchug saba Vitis vinifera L. Леся Lesya Леся Украинка, Украинка P TG IVM01649 XIX-32 Nimrang x Zhemchug saba Vitis vinifera L. Мечта Mechta Nadezhda P TG IVM01048 XXIV-27 Chaush rozovyi x Kishmish Chernyi Vitis vinifera L. Мускат ОСХИ Muskat oskhi W TG IVM01098 IV-121 Chaush rozovyi x Kishmish belyi oval nyi 17 Vitis vinifera L. Мускат таировский Muskat tairovskii B TG - - Odesskii rannii x Muskat de Hambourg 18 Vitis vinifera L. Мускат таировский Muskat tairovskii B TG IVM01136 V-35 Odesskii rannii x Muskat de Hambourg 19 Vitis vinifera L. Мускат жемчужный Muskat zhemchuzhnyi W TG IVM01121 XX-93 Chaush rozovyi x Zhemchug sabo 20 Vitis hybrid Находка Nakhodka Ovidiopol skii W WG IVM01178 VI-2 Severnyi (Seyanec Malengra x V. amurensis) x Odesskii ustoychivyi (Sereksiya x V. rupestris du Lot) Vitis hybrid Нимрант Магарача Nimrang Magaracha Nimrang ustoichivyi Magaracha, Magarach W TG IVM00012 XXIV-21 Nimrang x SV (Phylloxera-resistant interspecific hybrid selected by Seyve Villard, Fr) Vitis vinifera L. Одесский черный Odesskii chernyi B WG IVM01225 III-95 Alicante bouchet x Cabernet Sauvignon 23 Vitis vinifera L. Одесский ранний Odesskii rannii W TG IVM01221 XVII-60 Chaush x Muscat de Hambourg 1946

5 Characterization of Grapevine Accessions from Ukraine Vitis vinifera L. Октябренок Oktyabr onok B TG IVM01240 IV-61 Nimrang x Alphonse Lavallee Vitis vinifera L. Первенец Донбасса Pervenets Donbassa W TG IVM01285 XIX-52 Shasla muskatnaya x Anzhevin Oberlena 26 Vitis vinifera L. Поливитис Магарача Polivitis Magaracha (3n or 4n) Magarach P W TG IVM00041 VI-115 Katta Kurgan (2n) x Shasla gro kulvar belaya (4n) Vitis vinifera L. Радуга Raduga Magarach 510, Nimrang oboepolyi P TG IVM00045 IV-33 Nimrang x Kishmish Chernyi Vitis vinifera L. Рислинг мускатный Risling muskatnyi Magarach 376 W WG IVM00048 III-78 Risling x Muskat belyi Vitis hybrid Рубин Голодриги Rubin Golodrigi B WG IVM02633 XXVIII-128 Rubinovyi Magaracha x Magarach (Ravat [complex hybrid of Seibel, Fr, containing V. labrusca, V. riparia, V. rupestris, V. lincecumii, V. vinifera, V. cinerea] x V. vinifera or open pollinated) Vitis vinifera L. Рубиновый Магарача Rubinovyi Magaracha Magarach n o 56 B WG IVM00042 III-47 Cabernet Sauvignon x Saperavi Vitis vinifera L. Саперави Saperavi B WG IVM01430 X-46 unknown (Georgian autochthonous variety) 32 Vitis hybrid Сейв Виллард Seyve villard Perle Noire B WG IVM01292 XX-37 Panse x Seyve Villard (Phylloxeraresistant interspecific hybrid selected by Seyve Villard, Fr) 33 Vitis hybrid Спартанец Магарача Spartanets Magaracha W WG IVM00009 XX-70 Seibel x Saperavi severnyi (Seyanec Malengra x V. amurensis) Vitis vinifera L. Сухолиманский белый Sucholimanskii belyi W WG - - Chardonnay x Plavai Vitis vinifera L. Сухолиманский белый Sucholimanskii belyi W WG IVM01560 III-76 Chardonnay x Plavai Vitis vinifera L. Сверхранний бессемянный Sverkhrannii bessemyannyi Magarach W TG IVM00033 V-87 Magarach 417 (Muskat krasnyi de Madeyra x Khalili belyi) x Magarach 653 (Madeleine angevine x Ak Yakdona) Vitis vinifera L. Ташлы Tashly Tashly misket, izyum W TG,WG IVM01597 I-77 Unknown 38 Vitis vinifera L. Таврида Tavrida Magarach 171 B WG IVM00044 III-69 Mourvedre x Muskat chernyi Vitis vinifera L. Таврия Tavriya Magarach B TG IVM00034 XIX-49 Rannii kabraiskii x Koroleva vinogradnikov Vitis hybrid Цитронный Магарача Tsitronnyi Magaracha W WG IVM00014 XVIII-86 Madeleine angevine x pollen mix Novoukrainskii rannii and Magarach (Rkaciteli x Magarach : complex hybrid) Vitis vinifera L. Украинский ранний Ukrainskii rannii Magarach 339 P TG IVM01650 V-84 Madeleine angevine x Shakar Angur Vitis vinifera L. Восток Vostok B TG IVM00419 IV-65 Nimrang x Matyash yanosh Vitis vinifera L. Янтарь ОСХИ Yantar oskhi W TG IVM01915 III-36 Koroleva vinogradnikov x Oktyabr skii (Nimrang x Karmannuii) Vitis vinifera L. Янтарь ОСХИ Yantar oskhi W TG - - Koroleva vinogradnikov x Oktyabr skii (Nimrang x Karmannuii) Vitis vinifera L. Южанка Yuzhanka W TG IVM01896 III-35 Chaush rozovyi x Arakseni belyi Vitis vinifera L. Зорька Zor ka P TG - - Chaush rozovyi x mixturi of pollen (Zhemchug saba + Irshai oliver) Vitis vinifera L. Зорька Zor ka P TG IVM00654 V-34 Chaush rozovyi x mixturi of pollen (Zhemchug saba + Irshai oliver) 1958 a Number in supplementary Table 1. b W: white; B: black; P: pink. c WG: winegrape; TG: table grape.

6 174 Heuertz et al. Table 2 Diversity estimates per locus and population: As, allelic richness based on a minimum sample size of 16 diploid individuals; H S and H E, gene diversity. Reference rootstocks Reference grape Reference grape excluding hybrids Accessions Ukrainian Ukrainian excluding hybrids VVS2 As VVMD5 As VVMD7 As VVMD27 As VRZAG62 As VRZAG79 As Average As VVS2 H S VVMD5 H S VVMD7 H S VVMD27 H S VRZAG62 H S VRZAG79 H S Multilocus estimate H E differentiated from Ukrainian accessions (pair wise F ST = 0.026, p < 0.001). All results on pair-wise differentiation between populations remained similar when hybrid accessions were excluded (Table 3). Vitis vinifera winegrapes were significantly differentiated from table grapes (F ST = 0.056, p < 0.001), irrespective of geographic origin. A nested AMOVA contrasting reference versus Ukrainian V. vinifera revealed a significant differentiation of wine versus table grapes within origins (F SC = 0.038, p < 0.05; Table 4). However, when the upper level in the AMOVA was wine versus table V. vinifera, Ukrainian accessions were not significantly differentiated from reference accessions within groups (F SC = 0.012, ns; Table 4), suggesting that the use of grape accessions was a stronger determinant of differentiation than their geographic origin. This is reflected in the factorial correspondence analysis (FCA; Figure 3), where the first factor (axis 1) essentially separates table from wine grapes (from left to right). Finally, all individuals having >50% ancestry in St r u c t u r e cluster 2 (shaded bar in Figure 2) have a negative abscissa in the FCA (Figure 3), and the differentiation between clusters was F ST = (when assigning individuals to the cluster in which they had >50% ancestry). Figure 1 Clustering of the three groups of grapevine accessions with the Structure algorithm assuming K = 3 clusters. with cluster 1 (unshaded bar in Figure 2) predominating among the reference grape accessions while Ukrainian accessions had their highest percentage of ancestry in either of both clusters. Muscat-type vines were the only reference vines to have over 80% ancestry in cluster 2 (shaded bar in Figure 2). In agreement with these findings, differentiation of the rootstocks was relatively marked from both reference grape (pair wise F ST = 0.098, p < 0.001) and Ukrainian accessions (pair wise F ST = 0.084, p < 0.001; Table 3). Reference grape accessions were weakly but significantly Discussion This study contributed to solving labeling issues among Ukrainian grapevine accessions by confirming the identity of Sukholimanskii belyi, Muskat tairovskii, and Yantar oskhi accessions from the Tairov Institute, Odessa, and the Magarach Institute, Yalta. An identical genotype at six microsatellites was identified for Danko and Sverkhrannii bessemyannyi. These accessions should be synonyms since genotypes are unlikely to be the same by chance considering the high discrimination power of the loci (combined power of D T = ). These accessions display, however, important morphological differences, so that this result requires confirmation. Lesya accessions from both institutes differed at five of six loci, suggesting that this name is a homonym for different accessions. Both Zor ka accessions differed by one allele at each of two loci, which could be explained by the combination of, for instance, one allelic dropout and one mistyping, although the probability of two errors occurring in the same genotype is low [p = 0.002, binomial distribution probability of two errors in 12 trials with error rate per allele P(error) = ]. Allele differences

7 Characterization of Grapevine Accessions from Ukraine 175 Figure 2 Posterior probability of ancestry of grapevine accessions (excluding rootstocks) in each of K = 2 clusters (cluster 1 unshaded, cluster 2 shaded) and standard deviations obtained from 10 runs of the St r u c t u r e algorithm (Pritchard et al. 2000). in accessions from the same variety can alternatively be due to the accumulation of somatic mutations that are transmitted through vegetative propagation (Hocquigny et al. 2004). Within the cultivated grape gene pool, the use as table or wine grapes was found to be a more important criterion for genetic differentiation than geographic origin. This finding confirms the lack of clear geographic structure in a large collection of cultivated grapes as noted elsewhere (Aradhya et al. 2003) and suggests that genetic structure in grapevine is rather determined by selection and historically early (Rivera Núñez and Walker 1989, Aradhya et al. 2003) as well as contemporary movement of germ plasm (e.g., European varieties entered as parents of recently bred Ukrainian varieties, see below). The strongest differentiation in the data set was found between the two St r u c t u r e (Pritchard et al. 2000) clusters. Cluster 2, strongly represented among the Ukrainian accessions, contained essentially table grapes and Muscat types. According to information from the breeding design, ~80% of the studied Ukrainian accessions have Muscat genotypes in their parentages (Table 1). For instance, the varieties Mechta, Muskat oskhi, Muskat zhemchuzhnii, Yuzhanka, Bessemyannyi mel nika, and Odesskii rannii have the Muscat grape Chaush or its pink version Chaush rozovii as female parent. Further, the variety Muskat Ottonel entered as a grandparent of the accessions Muskat zhemchuzhnii, Lesya, and Vostok. St r u c t u r e cluster 1 contains, on the other hand, a majority of Western as well as Eastern European winegrapes. Indeed, Western European varieties entered as parents for a number of Ukrainian cultivated accessions. For instance, Cabernet Sauvignon is, respectively, female and male parent of Rubinovyi magaracha and Odesski chernyi, Chardonnay is female parent of Sukholimanskii belyi, and Mourvèdre is female parent of Tavrida (although our microsatellite data were in agreement with the latter parentage at only five out of six loci; one allelic dropout could have occurred at locus VVMD7). Autochthonous varieties of different Eastern European origins also entered in the parentage of Ukrainian accessions, such as the Moldavian grape Plavai (parent of Sukholimanskii belyi) and the Georgian grapes Saperavi and Rkatsiteli (Table 1). Our genetic structure results also confirm UPGMA analysis on predefined groups (Aradhya et al. 2003), which distinguished essentially French winegrapes from Eastern European table and dual-use grapes. However, we highlight that our UPGMA analysis on individual accessions lacked bootstrap support of all but a few tip nodes and hence did not produce any conclusive grouping of accessions (supplementary Figure 1), suggesting the need to apply more statistically powerful approaches, such as the St r u c t u r e algorithm (Pritchard et al. 2000) to detect any reliable structure in the cultivated gene pool. The inclusion of varieties from very diverse origins into the breeding design of Ukrainian accessions and the analysis of both wine and table grape accessions contributes

8 176 Heuertz et al. Table 3 Differentiation among grapevine populations estimated by F ST. Population differentiation was tested by 10,000 permutations of individuals among populations [P(1-sided test, H1: obs>exp)] using SPAGeDi. A sequential Bonferroni correction was applied to correct for multiple testing. Reference grape accessions Reference grape accessions excluding hybrids Ukrainian accessions Ukrainian accessions excluding hybrids Reference rootstocks *** a *** *** *** Reference grape accessions ns *** *** Reference grape accessions excluding hybrids *** ** Ukrainian accessions ns a **, ***, and ns indicate significance as p < 0.01, p < 0.001, and not significant, respectively. Table 4 Nested AMOVA analysis. Significance tests are based on 10,000 permutations. Source of variation df a Sum of squares Variance components % of variation Reference vs. Ukrainian V. vinifera Among groups (reference vs. Ukrainian V. vinifera) Va 1.12 Among populations (wine vs. table grapes) within groups Vb 3.80 Within populations Vc Total F ST = 0.049, Vc and F ST : P = F SC = 0.038, Vb and F SC : P = F CT = 0.011, Va and F CT : P = Wine vs. table V. vinifera Among groups (wine vs. table V. vinifera) Va 4.75 Among populations (reference vs. Ukrainian V. vinifera) Vb 1.12 within groups Within populations Vc Total F ST = 0.059, Vc and F ST : P = F SC = 0.012, Vb and F SC : P = F CT = 0.048, Va and F CT : P = a df: degrees of freedom. Figure 3 Factorial correspondence analysis visualizing genetic relationships between Vitis accessions excluding rootstocks. The two first factors are plotted, explaining 7.1% (horizontal axis) and 5.3% (vertical axis) of the total variation, respectively.

9 Characterization of Grapevine Accessions from Ukraine 177 to explaining the high genetic diversity encountered in the Ukrainian material: H E = (H E = excluding hybrid accessions) was slightly higher than H E = in Moldovan, Crimean autochthonous and Russian accessions (Lefort et al. 2003), H E = in Armenian, Georgian, and Turkish accessions (Vouillamoz et al. 2006), and higher than the average H E = in seven European wine-producing regions (Sefc et al. 2000) and the average H E =0.771 in 16 groups of European, Tunisian, and Near Eastern grapevine accessions (Aradhya et al. 2003). Genetic diversity in Ukraine was very similar to the reference collection. We highlight that this collection is not a random sample of (Western) European vine varieties, but was specifically composed to habor a maximal allelic diversity for the description of microsatellite reference alleles (This et al. 2004). The number of alleles was high among Ukrainian grapevine accessions with an allelic richness of 8.09 alleles in a random sample of 16 individuals, compared to the 6.3 alleles on average in random samples of 13 individuals from European regions (Sefc et al. 2000). Aradhya et al. (2003) did not detect more than an average of 6.9 alleles in groups of up to 20 accessions. It seems that the high diversity identified in wild grapevine from the Black Sea and Caucasus area (Levadoux 1956, Grassi et al. 2006) is paralleled in autochthonous and recently bred cultivated varieties (this study, Lefort et al. 2003, Vouillamoz et al. 2006), adding support to an early domestication and conservation of a broad genetic pool in domesticated varieties. Conclusions Our results contribute to characterizing grapevine genetic resources of Ukraine and solve labeling issues among accessions. The analysis of genetic structure highlights the contribution of Muscat and European winegrapes in the breeding design of Ukrainian varieties and suggests that selection and historical and contemporary movement of germplasm are major factors shaping the structure of the grapevine gene pool. Our results let us conclude that Ukrainian grapevine varieties represent a genetically rich and diversified genetic resource that can enter as valuable material for future sustainable breeding and improvement of grapevine. Literature Cited Aradhya, M.K., G.S. Dangl, B.H. Prins, J.M. Boursiquot, M.A. Walker, C.P. Meredith, and C.J. Simon Genetic structure and differentiation in cultivated grape, Vitis vinifera L. Genet. Res. 81: Arroyo-García, R., et al Multiple origins of cultivated grapevine (Vitis vinifera L. ssp. Sativa) based on chloroplast DNA polymorphisms. Mol. Ecol. 15: Bonin, A., E. Bellemain, P. Bronken Eidesen, F. Pompanon, C. Brochmann, and P. Taberlet How to track and assess genotyping errors in population genetics studies. Mol. Ecol. 13: Bowers, J.E., G.S. Dangl, R. Vignani, and C.P. Meredith Isolation and characterization of new polymorphic simple sequence repeat loci in grape (Vitis vinifera L.). Genome 39: Bowers, J.E., J.M. Boursiquot, P. This, K. Chu, H. Johansson, and C. Meredith. 1999a. Historical genetics: The parentage of Chardonnay, Gamay and other wine grapes of northeastern France. Science 285: Bowers, J.E., G.S. Dangl, and C.P. Meredith. 1999b. Development and characterization of additional microsatellite DNA markers for grape. Am. J. Enol. Vitic. 50: Falush, D., M. Stephens, and J.K. Pritchard Inference of population structure using multilocus genotype data: Linked loci and correlated allele frequencies. Genetics 164: Goudet, J FSTAT (version 1.2): A computer program to calculate F-statistics. J. Heredity 86: Grassi, F., M. Labra, S. Imazio, R. Ocete Rubio, O. Failla, A. Scienza, and F. Sala Phylogeographical structure and conservation genetics of wild grapevine. Cons. Genet. 7: Hardy, O.J., and X. Vekemans SPAGeDi: A versatile compute program to analyse spatial genetic structure at the individual or population levels. Mol. Ecol. Notes 2: Hocquigny, S., F. Pelsy, V. Dumas, S. Kindt, M.C. Heloir, and D. Merdinoglu Diversification within grapevine cultivars goes through chimeric states. Genome 47: Imazio, S., M. Labra, F. Grassi, A. Scienza, and O. Failla Chloroplast microsatellites to investigate the origin of grapevine. Genet. Res. Crop Evol. 53: Lefort, F., and K.K.A. Roubelakis-Angelakis Genetic comparison of Greek cultivars of Vitis vinifera L. by nuclear microsatellite profiling. Am. J. Enol. Vitic. 52: Lefort, F., S. Gorislavets, V. Risovannay, and L. Troshin Genetic profiling of Moldavian, Crimean and Russian cultivars of Vitis vinifera L. with nuclear microsatellite markers. In Oenologie: 7e symposium international d oenologie. A. Lonvaud-Funel et al. (Eds.), pp Editions Tec & Doc, Paris. Levadoux, L.D Wild and cultivated populations of Vitis vinifera L. Ann. Amélior. Plantes 6: Lopes, M.S., K.M. Sefc, E. Eiras Dias, H. Steinkellner, M. Laimer da Câmara Machado, and A. da Câmara Machado The use of microsatellites for germplasm management in a Portuguese grapevine collection. Theor. Appl. Genet. 99: Maghradze, D Conservation of local grapevine varieties in the Caucasus and northern Black Sea region: An update. IPGRI Newsl. Europe 30:14. Martín, J.P., J. Borrego, F. Cabello, and J.M. Ortíz Characterization of Spanish grapevine cultivar diversity using sequencetagged microsatellite site markers. Genome 46: McGovern, P.E Ancient Wine: The Search for the Origins of Viticulture. Princeton University Press, Princeton, NJ. Nei, M., F. Tajima, and Y. Tateno Accuracy of estimated phylogenetic trees from molecular data. J. Mol. Evol. 19: Olmo, H.P The origin and domestication of the vinifera grape. In The Origins and Ancient History of Wine. P.E. McGovern et al. (Eds.), pp Gordon and Breach, Amsterdam. Pritchard, J.K., M. Stephens, and P. Donnelly Inference of population structure using multilocus genotype data. Genetics 155:

10 178 Heuertz et al. Rivera Núñez, D., and M.J. Walker A review of palaeobotanical findings of early Vitis in the Mediterranean and of the origins of cultivated grape-vines, with special reference to new pointers to prehistoric exploitation in the western Mediterranean. Rev. Palaeobot. Palyn. 61: Sefc, K.M., F. Regner, E. Turetschek, J. Glössl, and H. Steinkellner Identification of microsatellite sequences in Vitis riparia and their applicability for genotyping of different Vitis species. Genome 42: Sefc, K.M., M.S. Lopes, F. Lefort, R. Botta, K.A. Roubelakis- Angelakis, J. Ibáñez, I. Pejic, H.W. Wagner, J. Glössl, and H. Steinkellner Microsatellite variability in grapevine cultivars from different European regions and evaluation of assignment testing to assess the geographic origin of cultivars. Theor. Appl. Genet. 100: Takezaki, N., and M. Nei Genetic distances and reconstruction of phylogenetic trees from microsatellite DNA. Genetics 144: This, P., et al Development of a standard set of microsatellite reference alleles for identification of grape cultivars. Theor. Appl. Genet. 109: Thomas, M.R., and N.S. Scott Microsatellite repeats in grapevine reveal DNA polymorphisms when analysed as sequencetagged sites (STSs). Theor. Appl. Genet. 86: Unwin, T Wine and the Vine: An Historical Geography of Viticulture and the Wine Trade. Routledge, London. Valière, N GIMLET: A computer program for analysing genetic individual identification data. Mol. Ecol. Notes 2: Vouillamoz, J.F., P.E. McGovern, A. Ergul, G. Söylemezoğlu, G. Tevadze, C.P. Meredith, and M.S. Grando Genetic characterization and relationships of traditional grape cultivars from Transcaucasia and Anatolia. Plant Genet. Res. 4: Weir, B.S., and C.C. Cockerham Estimating F-statistics for the analysis of population structure. Evolution 38: Zohary, D The domestication of the grapevine Vitis vinifera L. in the Near East. In The Origins and Ancient History of Vine. P.E. McGovern et al. (Eds.), pp Gordon and Breach, Amsterdam.