Abstract
Genetic relationships among important table-grape varieties were studied through nuclear and chloroplast microsatellite analysis. A total of 376 accessions were genotyped with 25 nuclear microsatellite loci using three multiplex PCRs. The average alleles per locus was 9.96, while the probability of identity was 1.66 x 10−21. A comparison among genotypes, morphology when needed, and literature data has defined synonyms, homonyms, sports, and mistakes among the plant material. In this study, groups of varieties such as Afus Ali, Ahmeur bou Ahmeur, Chasselas, and Muscat of Alexandria were clarified. A parentage analysis of table-grape cultivars was also carried out using a nonredundant genotype table, which included 273 accessions, accompanied by a thorough search in the literature and chloroplast haplotypes to define the direction of the crosses found. In most cases, the information available about the crosses originating these cultivars was correct, particularly in seedless cultivars. Nevertheless, some cultivars such as Cardinal could not originate from the crosses described or suggested in the literature and alternative parents have been proposed. In other cases, where previous information could not be found or was incomplete, some light has been shed on the genetic origin of some cultivars; for instance, the ancestors of Alphonse Lavallée, Admirable de Courtiller, and Calmeria have been suggested. A mathematical analysis, in the form of likelihood ratios, was done to determine the reliability of the suggested crosses.
Grapevine (Vitis vinifera L.) is an ancient crop with a complex taxonomic structure because of its vegetative propagation and wide distribution. It is represented by a high number of cultivars with a large proportion of synonyms (syn.; different names for the same variety) and homonyms (different varieties with the same name) (This et al. 2006). As an example, 155 different synonyms can be found for the variety Muscat of Alexandria at the Vitis International Variety Catalogue (VIVC: (www.vivc.bafz.de/index.php). It is important to shed light on this confusion to avoid legal problems concerning the accurate identification of cultivars with origin denomination and in the table-grape market, where there are numerous legally protected cultivars.
World production of table grapes has increased during the last years (OIV 2004). In breeding table grapes, numerous cultivars have been used, but a distinction between seeded and seedless cultivars must be made. Traditionally, more seeded cultivars had been available than seedless cultivars. During the past decades, seedlessness became a trait of increasing importance for the consumer, and new seedless cultivars were bred. Now there are many grapevine breeding centers throughout the world, and most of the new cultivars released to the market are seedless and subjected to plant breeders’ rights. The first generation of new seedless cultivars was obtained from crosses where Sultanina (syn. Thompson Seedless) was predominantly the male parent, as the in vitro embryo rescue technology was later developed (Cain et al. 1983). The female parent was variable but numerous cultivars were frequently used, such as Afus Ali (syn. Dattier de Beyrouth), Alphonse Lavallée, Muscat of Alexandria, Perle von Csaba, and Koenigin der Weingaerten (syn. Regina dei Vigneti). These cultivars were also preferred to produce new seeded cultivars, which were later used as parents for the next generations of seeded and seedless cultivars such as Cardinal, Italia, Sultana Moscata, and Muscat Hamburg (Wagner and Truel 1988). Although information about the origin of many elite table-grape cultivars would be of great value for present breeding programs, currently the origin of many cultivars remains unknown or, perhaps of greater concern, the information available may be misleading.
Microsatellite sites are excellent markers for grapevine characterization (Sefc et al. 2001). In general, all plants belonging to the same cultivar (i.e., from a monoclonal origin) show identical genotypes in all microsatellite loci. Although the origin of the present grapevine genome seems to arise from an ancestral hexaploidization (Jaillon et al. 2007), only two alleles are usually found in the microsatellite analyses published to date. Codominant Mendelian inheritance of microsatellites has been used previously to discover the genetic origin of some important wine cultivars (Bowers et al. 1999a). The use of chloroplast microsatellites allows for uncovering the direction of the crosses, as chloroplasts are only maternally inherited in grapevine (Arroyo-Garcia et al. 2002). In this paper we provide clarity on the confusion regarding synonyms and homonyms and reveal the genetic origins of some important table-grape cultivars through microsatellite analysis and a thorough literature review.
Materials and Methods
Plant material was obtained from the Vitis Germplasm Bank (BGV) at the Finca El Encín (IMIDRA, Alcalá de Henares, Spain), one of the largest collections in the world, and from CIDA (presently called IMIDA, Consejería de Agricultura, Agua y Medio Ambiente, Murcia, Spain). A total of 376 accessions of mostly table grapes were analyzed, although only 273 nonredundant genotypes were used for parentage analysis. DNA extractions were carried out from young leaves using the Qiagen DNeasy 96 Plant Kit (Hilden, Germany).
Twenty-five nuclear microsatellites were studied in three independent PCRs labeled S, A, and B. The multiplex PCR S included nine microsatellites: VVS2 (Thomas and Scott 1993), VVMD5, VVMD27, VVMD28 (Bowers et al. 1996, 1999b), ssrVrZAG29, ssrVrZAG62, ssrVrZAG67, ssrVrZAG83, and ssrVrZAG112 (Sefc et al. 1999). The multiplex PCR A included 11 microsatellites: VVS2 (Thomas and Scott 1993), VVMD7, VVMD24, VVMD25 (Bowers et al. 1996, 1999b), VVIB01, VVIH54, VVIN73, VVIP31, VVIP60, VVIQ52 (Merdinoglu et al. 2005), and VMC1B11 (Zyprian and Topfer 2005). The multiplex PCR B also included nine markers: VVMD5, VVMD21, VVMD27, VVMD28, VVMD32 (Bowers et al. 1996, 1999b), VVIN16, VVIV37, VVIV67 (Merdinoglu et al. 2005), and VMC4F3.1 (Di Gaspero et al. 2000). Reactions mixes and thermal cycler conditions are given in Supplementary Tables 1 and 2. The markers VVS2, VVMD5, VVMD27, and VVMD28 were analyzed twice. Five chloroplast microsatellites were also analyzed in the 67 varieties involved with the 43 pedigrees selected: Ccmp3, Ccmp5, Ccmp10 (Weising and Gardner 1999), ccSSR9, and ccSSR14 (Chung and Staub 2003). A multiplex PCR (C) was performed including the five markers (Supplementary Tables 1 and 2).
The separation of fragments and data analysis was carried out in several ways. The multiplex PCR S was analyzed in an ABI PRISM 310 Genetic Analyzer (Applied Biosystems, Foster City, CA) and the fragments were sized with GeneScan software using TAMRA 500 as an internal marker (Applied Biosystems). The multiplex PCRs A, B, and C were analyzed in an ABI 3130 Genetic Analyzer, and the fragments were sized with GeneMapper 3.7 using LIZ 500 as an internal marker (Applied Biosystems). Chloroplast haplotypes were named using published nomenclature (Arroyo-Garcia et al. 2006).
Diversity analysis was carried out using Identity 1.0 (Wagner and Sefc 1999) and the Excel application Microsatellite Toolkit (Park 2001). Pedigree analysis was conducted using Identity 1.0 (Wagner and Sefc 1999) and Cruces 2.0 (J. Ibáñez, author’s unpublished data, 2005) software. Statistical analysis was carried out by means of likelihood ratios (Hagelberg et al. 1991) in the cases where new parent/offspring combinations were found. The likelihood ratio is the quotient of the probability of the progeny genotype if it had the presumptive parents, and the probability of that genotype if it had random unrelated parents.
Results and Discussion
Diversity analysis.
Three multiplex PCRs including 9, 11, and 9 nuclear markers, respectively, were used to genotype 376 grapevine accessions. Equimolar primer concentrations in these PCRs showed uneven amplification, with some of the products scarcely amplified. Similar amplification of each marker could be obtained by decreasing the quantity of primer for the strongly amplified fragments, increasing the amount of primers for the poorly amplified fragments, and adjusting the concentrations of the remaining PCR reagents accordingly. As far as we know, this is the first time that multiplex PCRs including such a number of markers have been used in grapevine.
A total of 273 nonredundant genotypes were obtained from the analysis of 376 mostly table-grape accessions using 25 nuclear microsatellites. Some wine cultivars were introduced because they were involved in the pedigrees. The total number of alleles per locus varied between three (ssrVrZAG29) and 18 (VMC4F3), with an average of 9.96. The total probability of identity was 1.66 x 10−21, which is a low value when compared with other published data (Ibáñez et al. 2003, Lefort and Roubelakis-Angelakis 2001, Sefc et al. 2000), undoubtedly because of the higher number of microsatellites used in this study. Observed and expected heterozygosities were very high (Supplementary Table 3), with only two markers below 0.6. The PIC average of 0.7 supports the suitability of these markers for grapevine characterization.
In the pedigree analysis, a number of compatible crosses were chosen among those found, and the 104 accessions involved with such selected crosses were further studied. Many synonyms and sports were detected among these accessions, as well as several homonyms and errors (Table 1⇓). Prime names for varieties were taken from VIVC (www.vivc.bafz.de/index.php), except in the case of Listan Prieto (Tapia et al. 2007, Zerolo et al. 2006). In some cases, full matching for microsatellite genotypes was obtained but the information available did not allow us to decide if there was a nonpreviously described synonym or a mistake involved.
Some of the synonyms and sports detected are described in the literature (Branas and Truel 1965, Cosmo 1975, Galet 2000, OIV 1987) (VIVC: www.vivc.bafz.de/index.php). This is the case of synonyms found for Alphonse Lavallée: Ribier, Royal, Royal Terheyden, Almería Negra, and Leopold III (an autotetraploid from Alphonse Lavallée). There is another accession named Royal in the collection that is a homonym, as it showed a different genotype.
Cannon Hall has been described as a possible tetraploid from Muscat of Alexandria (Galet 2000) and both cultivars have the same genotype, in which case it may be an autotetraploid. In addition, five synonyms were detected for Muscat of Alexandria: Moscatel, Moscatel Blanco, Moscatel de Málaga, Moscatel Gordo, and Moscatel Romano. Three color sports were also found for this variety: Moscatel Negro, Moscatel Negro de Valencia, and Muscat Flame. Moscatel de Encinacorba and Vizaca showed identical genotypes, but no information was found to determine if they are synonyms or mistakes. In contrast, accessions with the same genotype that correspond to errors (because they are generally considered different varieties) are Garnacho Rojo (syn. Garnacha Tinta) and Pedro Ximenez. Romé is also considered a different variety, but its morphology description is almost identical to Muscat of Alexandria (García de Luján et al. 1990); thus, it could be an unknown synonymy for this variety. Zibibbo is a well-known synonym for Muscat of Alexandria; however, the Zibibbo accession in the collection of El Encín has an identical genotype to Principessa di Piemonte, and thus it must be an error.
The group of Chasselas also comprises different synonyms and sports as described in the literature (Branas and Truel 1965, Galet 2000): Chasselas Doré, Chasselas Blanca, Chasselas de Montauban, and Albillo are synonyms. Nevertheless, there are two homonymous Albillo corresponding to two different varieties. Chasselas Apyrene is a seedless sport, while Chasselas Rose, Chasselas Rouge, and Chasselas Violet are color sports. Chasselas Musqué is a f lavor sport, and Chasselas Cioutat presents a different shape in the leaf. Chasselas Gros Coulard is a tetraploid mutation (Branas and Truel 1965). Corneille (syn. Chasselas Corneille) appears in the literature as a different variety, although it seems similar to Chasselas Doré. All have identical genotypes for all the microsatellites studied, with the exception of several possible mutations in different cultivars, including the presence of three alleles for some microsatellites and cultivars (Supplementary Table 4).
Actoni Maceron, Hafiz Ali, Kérino, Pepita de Oro, Razaki Isla de Creta, Rosaki, and Roseti are previously described within the group of synonyms for Afus Ali, but not so Princes Chasselas. Marawi, Pensal Blanco, Professor Aberson, and Sabalkanskoi also have identical profiles. They are described as different varieties, and therefore are mistakes in the collection.
There are three accessions in the collection named Teta de Vaca with different genotypes, which are homonyms. One is identical to De Cuerno and Pizzutello Bianco, and all are thus synonyms or sports for Cornichon Blanc (Cosmo 1975, Galet 2000). Ahmeur bou Ahmeur, Argelina, and Royal Gordo are synonyms and presented the same genotype as a second Teta de Vaca. Parras de la Casa also matched, but no information was found on this variety. Boto de Gall and Corazón de Cabrito are described as synonyms for the first Teta de Vaca (Cornichon Blanc), but showed the same genotypes as the second (Ahmeur bou Ahmeur). No synonym was found for the third accession called Teta de Vaca.
Regina is a name used for different varieties (Afus Ali, Koenigin der Weingaerten, Regina). It showed identical genotype to Italia but a different shape of the leaves; hence, it can be a sport. In contrast, the Opale accession studied is probably incorrect, as it matched with Italia and cannot be the offspring of a Pirovano 57 x Muscat of Alexandria cross, as described.
A similar case occurred with Molinera (syn. Red Malaga) and Castellano Morado, which were thought to be different varieties. In fact, they matched at all the microsatellite loci, even though the leaves are different, and therefore Castellano Morado could be a sport of Molinera or vice versa. Listán Prieto is the suggested correct de-nomination for an old variety cultivated both in America (Criolla, Mission) and in the Canary Islands (Tapia et al. 2007, Zerolo et al. 2006).
Pedigree analysis.
Two hundred and seventy-three nonredundant genotypes were analyzed. A total of 437 possible crosses were obtained when as many as three incompatible alleles were tolerated: 81 compatible crosses were found allowing incompatibility in 0 alleles; 37 crosses were detected allowing incompatibility in one allele; 85 crosses were detected allowing incompatibility in two alleles; and 234 crosses were found allowing incompatibility in three alleles. These results provided information about possible mistakes in the genotype table. A total of 43 compatible crosses were further studied. Chloroplast microsatellites were useful to support the compatibility of the selected crosses and allowed for determining its direction in many cases (Table 2⇓, Table 3⇓) (Supplementary Table 5). Pedigrees confirmed for seedless and seeded table grape cultivars as previously described in the literature (Branas and Truel 1965, Cosmo 1975, Galet 2000, OIV 1987, Wagner and Truel 1988) (VIVC) are shown (Table 2⇓). New pedigrees for varieties with ancestors that were unknown or mistaken and the likelihood ratios calculated for the crosses suggested are also shown (Table 3⇓).
Seedless cultivars.
Pedigree analysis confirmed 20 crosses producing seedless grape cultivars as described in the literature (Table 2⇑). Chloroplast haplotypes allowed for establishing the female parent in most cases. Sultanina (syn. Thompson Seedless) is a compatible parent in many of those crosses. As expected, it acted usually as the male parent, since it is a “seedless” variety. The Sultanina genotype is noncompatible at the locus VVIV67 in some previously described crosses (Pasiga, Rodi, and Sultana Moscata). A mutation in Sultanina of the 401 bp allele to the 399 bp allele would explain the noncompatibility at this locus. This type of mutation has been described in grapevine, especially in old varieties (Ibáñez et al. 2000). In the case of Beauty Seedless, one of the parents described is Black Kishmish (which is not available), but the cross is compatible using Black Seedless instead. The name Kishmish is used in Eastern countries as a synonym for Sultanina, the main seedless variety.
Some apparent mistakes were detected. The origin of the cultivars named Italia x Sultanina V-6 and Italia x Sultanina VI-4 cannot be crosses between Italia and Sultanina as their names clearly indicate. These cultivars came to the BGV El Encín from Bulgaria, where they were obtained (Pàstena 1972). Italia can be a parent but Sultanina cannot: it is incompatible in eight loci with Italia x Sultanina V-6 and in 11 loci with Italia x Sultanina VI-4 (Supplementary Table 4).
A cross between Delizia di Vaprio and Black Monucca can be the origin of Pirovano 166A, for which no previous information could be found. A very high likelihood ratio supports this finding (Table 3⇑).
Seeded cultivars.
Fifteen pedigrees described in the literature that have been confirmed by means of the analysis of 25 microsatellite loci and their chloroplast microsatellite haplotypes are shown (Table 2⇑). Perle von Csaba is a probable parent of Koenigin der Weingaerten. The other assumed parent is Souvenir Reine Elisabeth, which has not been studied. A similar named cultivar, Regina Elisabetta, which is a supposed synonymy (Galet 2000) (VIVC), has identical genotype to Queen but cannot be the parent of Koenigin der Weingaerten. Researchers have explained that Souvenir Reine Elisabeth is identical to Dattier de Beyrouth (prime name Afus Ali) (Branas and Truel 1965), while Pirovano and Longo (cited in Galet 2000) described it as a similar but slightly more rustic cultivar. Effectively, microsatellite genotypes showed that a cross between Afus Ali and Perle von Csaba could originate Koenigin der Weingaerten, with Afus Ali the female parent.
Several mistakes have been detected in the literature. The origin of Ohanes was traditionally attributed to a natural hybridization between Jaen and Ragol (Rueda Ferrer 1932, cited in García de Luján and Lara 1998) because the vineyards of Jaen were usually bordered by Ragol and because Ohanes characteristics can be considered the blend of those two cultivars. Microsatellite analysis indicates that neither Jaen nor Ragol can directly be the parent of Ohanes, although they share many alleles with the cultivar; neither could the progeny of a cross between Ragol and Jaen be one of the parents of Ohanes.
As expected, mistakes in the breeding process are normally due to the pollen donor, which seems to be the case for Cardinal and Bogni 17. Cardinal is one of the most important table-grape varieties, and, according to the literature, comes from the cross between Flame Tokay and Alphonse Lavallée. Nevertheless, it is not compatible with Flame Tokay in several loci as has been recently published (Akkak et al. 2007, Ibáñez et al. 2006). We found that it could be a descendant of an Alphonse Lavallée (female) x Koenigin der Weingaerten cross, if the existence of a null allele for the locus VVIN16 in Cardinal and Alphonse Lavallée is accepted. This locus was eliminated in the calculation of the likelihood ratio for this cross, which was still above 1012 (Table 3⇑).
Bogni 17 was considered the progeny of Bicane x Bonarda. Bicane is compatible, but the Bonarda we have available is not. Bogni 17 probably would not be a descendant of the cross described, since it has a Muscat flavor (Galet 2000). There are other varieties called Bonarda that should be checked, but our data indicated that Bicane (female) x Muscat Hamburg can be the cross originating this variety. The male parent would explain the presence of the Muscat f lavor in Bogni 17.
Pirovano bred Primiera from a cross between Delizia di Vaprio and Madeleine Royale, but the latter was incompatible as a parent at two loci. Instead, a cross between Delizia di Vaprio and Madeleine Angevine was fully compatible with Primiera.
In addition, some pedigrees for which previous information was incomplete or did not exist have been found. Though no pedigree was found in the literature for Admirable de Courtiller, two interesting facts are described (Branas and Truel 1965): it is similar to Chasselas Cioutat and it has a known synonym, Chasselas de Courtiller. Results suggested that it could have its origin from breeding Bicane (female) and Chasselas (perhaps Chasselas Cioutat).
Alphonse Lavallée is another historical table-grape variety whose 19th-century genetic origin is unknown. As Galet (2000) pointed out, Kharistvala Kolkhuri (syn. Gros Colman) could be one of the parents, as their berries are similar in shape. Effectively, microsatellite genotypes indicate that this variety could be the female parent, while Muscat Hamburg, also bred in the 19th century, could be the other parent.
A seed of Ohanes gave rise to Calmeria. Microsatellite analysis demonstrated that Ohanes can be one parent but not both, as expected since the Ohanes f lower is female. Ahmeur bou Ahmeur and Molinera were used as pollinators in Ohanes plantations. Nevertheless, these cultivars cannot be parents of Calmeria. In contrast, a cross between Ohanes and Sultanina could give rise to the microsatellite genotype present in Calmeria.
Pinot has been shown as the parent of many wine varieties cultivated in Europe (Bowers et al. 1999a). Madeleine Royale is probably another progeny of Pinot, where Frankenthal is the male parent.
Likelihood ratios for the crosses suggested in this study are shown (Table 3⇑). All are between 1012 and 1020, in the order of other likelihood ratios published in grapevine: 1012 to 1017 using 32 microsatellite markers (Bowers et al. 1999a), which indicates a high reliability for the suggested crosses.
Conclusion
Three multiplex PCRs including 29 nuclear microsatellites were used to genotype grape varieties. Numerous synonyms and sports were detected for the main table-grape varieties, and several homonyms and errors were also found. Certain information present in the literature about the crosses that originated different cultivars were confirmed by the analysis of 25 nuclear and five chloroplast microsatellite loci. In addition, the analysis has brought new insight on the origin of varieties for which there was either misleading or no previous information. Future studies must be carried out to know the origin of other important cultivars, which will contribute to the progress in grapevine breeding.
Footnotes
We thank Juan Carreño for sending plant material from CIDA (Murcia).
Supplementary tables are freely available with the online version of this article at www.ajevonline.org
Acknowledgments: This work was financially supported by the projects RF99- 009 (INIA, Agriculture Ministry of Spain) and GrapeGen (a joint venture between Genome Canada and Genoma España). A.M. Vargas was funded by a predoctoral fellowship from IMIDRA.
- Received April 2008.
- Revision received August 2008.
- Accepted September 2008.
- Published online March 2009
- Copyright © 2009 by the American Society for Enology and Viticulture