Abstract
Classical grapevine breeding is based on the controlled hybridization of elite genotypes and the selection of Fl hybrids. It is a slow process, especially for fruit traits, which requires a high investment in growing surface and culture management. Understanding the genetic control of trait variation and using linked molecular markers could allow the rapid identification of seedlings with high probability to express the desired characters. The present work is a preliminary result of a cooperative effort aimed at acquiring a broad and robust knowledge of the genetic determinism of key quality traits in table grape and tools to facilitate and accelerate the generation of new cultivars. Its starting goal was the selection of a common set of mapping markers to support the development and comparison of saturated linkage maps and quantitative trait loci (QTLs) in different grapevine genetic backgrounds. Two hundred ninety-eight simple sequence repeats (SSRs) were tested for polymorphism in 13 table grape cultivars involved as progenitors in seven intraspecific (Vitis vinifera L.) crosses. Based on polymorphism rate, amplification quality, and available mapping information, a common set of 86 highly polymorphic and well-distributed SSRs were chosen to match the homologous linkage groups in the different parental lines. Given its high level of polymorphism, this marker set is expected to be useful for rapid genotyping of additional progenies and comparison with the progenies analyzed here. The analysis of such a high number of loci led to some considerations about heterozygosity in table grape cultivars.
Detailed genetic maps based on molecular markers have been developed in a number of crop species and are being extensively used in studies of genome organization and evolution, for the dissection of complex traits and in practical plant breeding through marker-assisted selection. In addition, they facilitate map-based cloning and supply the genetic framework for physical map construction. Mapping at the multipopulation level has many advantages over that based on a single pedigree (Kianian and Quiros 1992). First, more loci can be mapped, which provides better genomic coverage. Second, marker orders and map distances are estimated more accurately and alterations in these values, because of possible chromosomal rearrangements affecting the parents involved in the crosses, are easily detected. Third, polymorphic loci common to all populations are disclosed and may be employed as consensus markers useful to align linkage groups and to assess the consistency of major genes and quantitative trait loci (QTLs) identified in different genetic backgrounds (Beavis and Grant 1991). Finally, markers common across mapping populations are particularly important when attempting to map specific genes of interest that are less likely to segregate within a specific progeny (Sewell et al. 1999).
Integrated maps have been constructed for a number of plants; among the various molecular markers currently used, simple sequence repeats (SSRs) seem to be especially suitable to anchor linkage groups as they are usually single-locus markers and provide an unambiguous means of defining linkage group homology across mapping populations (Cregan et al. 1999). For grape, several linkage maps have been published based on interspecific (Lodhi et al. 1995, Dalbó et al. 2000, Grando et al. 2003, Doucleff et al. 2004, Fischer et al. 2004) or intraspecific (Vitis vinifera L.) crosses (Doligez et al. 2002, Adam-Blondon et al. 2004, Riaz et al. 2004). They have been constructed using different classes of molecular markers (isozymes, RAPDs, RFLPs, AFLPs, and SSRs) and in several cases the presence of common SSRs, which are known to be highly polymorphic in grapevine (Thomas and Scott 1993, Bowers et al. 1996, 1999, Sefc et al. 1999), allowed the identification of homologous linkage groups.
We describe here the definition of a “common set” of SSR markers of general interest for the development of new genetic maps for grape. This is the first result produced by a cooperative effort among research institutions in Argentina, Chile, France, Italy, and Spain, which began in 2001 with the general objective to provide molecular tools for rapid and cost-efficient application of breeding programs in table grape.
Materials and Methods
Plant material and DNA extraction.
Thirteen table grape cultivars were used for this study. They were selected as progenitors in seven intraspecific (Vitis vinifera L.) crosses (Table 1⇓) to combine in progenies the adaptive traits of classical table grape cultivars (berry size and Muscat flavor) with the new traits demanded by the market, such as earliness and seedlessness. Total DNA was extracted from young leaves and shoot tips using standard procedures, slightly different depending on the laboratory.
Source, amplification, and analysis of SSR markers.
The 13 cultivars were genotyped for 298 SSR markers. One hundred forty-six (VMC) were developed by the Vitis Microsatellite Consortium, a cooperative international effort of 21 research groups coordinated by AgroGene (France). Twenty-three of these markers (VMC4D9.2, VMC4F8, VMC4F9.1, VMC4H9, VMC6B11, VMC6C10, VMC6D12, VMC6E10, VMC6G8, VMC7F2, VMC7G3, VMC7G5, VMC7H2, VMC7H3, VMC8A4, VMC8D1, VMC8G6, VMCNG2C2.1, VMCNG2E1, VMCNG2G7, VMCNG2H1, VMCNG2H2.2, and VMCNG2H7) are publicly accessible in the NCBI database of sequence tagged sites (http://www.ncbi.nlm.nih.gov/dbSTS/) and nine have been published. Ninety-five (VVI) were generated by INRA-France and have been recently published (Merdinoglu et al. 2005). The remaining 57 markers were already available from several grapevine research groups and were previously published (Thomas and Scott 1993, Thomas et al. 1994, Bowers et al. 1996, 1999, Sefc et al. 1999, Scott et al. 2000). For all but 13 SSRs (scu04, scu07, scu08, scu15, scu16, VMC4A1, VMCNG1D12, VMCNG2E1, VrZAG7, VrZAG12, VrZAG14, VrZAG15, and VrZAG82), genetic mapping information was available from grapevine maps (Adam-Blondon et al. 2004, Riaz et al. 2004). These maps comprise respectively 19 and 20 linkage groups (expected number is 19, 2n = 38 for subgenus Euvitis).
Amplification parameters varied according to the laboratory and the locus. PCRs were performed in a final volume of 10 to 20 μ L, including ~30 ng of DNA, 0.33 to 6.4 μ M of each primer, 200 μ M of each dNTP, 1x buffer, 0.375 to 0.5 U of DNA Taq polymerase, and 1.5 to 2.5 mM MgCl2. The amplification program consisted of 35 to 40 cycles and used 56°C as the standard annealing temperature. When necessary, conditions were optimized by changing annealing temperature.
Depending on the laboratory, PCR products were separated through PAGE or capillary electrophoresis. In the first case, standard 6 to 9% polyacrylamide sequencing silver-stained gels were used. In the second case, data were obtained with an ABI 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA) after the amplification with primers labeled with three different fluorochromes (6-FAM, HEX, and NED). The pedigree of each offspring was checked following the transmission of alleles at some SSR loci to verify their hybrid origin from the two parents. Polymorphisms for all 298 SSRs within each parent pair were noted following the codes required for JoinMap software for a CP population (aaxab, abxaa, abxab, abxac, abxcd). Comments about PCR conditions and quality of electrophoretic patterns were also recorded.
Selection of the mapping SSR set.
A total of 298 SSRs were tested for polymorphism in 13 table grape cultivars used as the progenitors of seven mapping progenies (Table 1⇑). For each locus the number of segregating crosses (number of crosses out of seven with at least one heterozygous parent) and the number of heterozygous parents out of 13 were registered. A common set of markers that could be mapped in most of the seven crosses was selected on the basis of polymorphism, amplification quality, and map position, according to the following criteria: (1) in most cases at least one parent of each cross was heterozygous (whenever necessary because of a lack of polymorphism, two markers were chosen, each one segregating in complementary crosses and as close as possible to each other instead of a single one segregating in all seven crosses); (2) each primer pair amplified no more than two fragments per genotype; (3) the amplified pattern was easily scorable; (4) the position on reference genetic maps was not redundant with other proposed markers; and (5) the number of selected SSR markers per linkage group ranged from three to seven.
Heterozygosity.
The percentage of heterozygosity (Ho) was calculated for each cultivar as the ratio between the number of heterozygous SSR loci and the total number of loci scored (minus multilocus and missing data). The percentage of heterozygosity was estimated both genomewide and for each linkage group, following the SSR genetic map positions in reference maps (Adam-Blondon et al. 2004, Riaz et al. 2004).
Results
Selection of the mapping SSR set.
Of the 298 SSR markers analyzed, 15 were discarded because they amplified more than one locus (VMC2C7, VMC2E9, VMC3B12, VMC4A5, VMC9F4, VVIB19, VVIH02, VVIP09, VVIP11, VVIP25.2, VVIP34, VVIT65, VVIU09, VVIV35, and VVMD37) and 25 because of the poor repeatability of amplification (scu01, scu02, scu03, scu05, scu09, scu12, scu13, VMC1D10, VMC3C7, VMC3D8, VMC6C3, VMC8H10, VMC9A3.1, VMCNG1A9, VMCNG1D12, VMCNG2D11, VMCNG2E1, VMCNG4E10.1, VrZAG12, VrZAG14, VrZAG26, VrZAG30, VVMD8, VVIB59, and VVS5). Eight additional markers did not show any polymorphism among the 13 genotypes tested (scu04, scu07, scu08, scu16, VMC6E4, VVIM42.2, VVIN68, and VVIQ66). From the remaining 250 SSRs, 114 were polymorphic in all the segregating progenies, whereas 136 segregated in fewer progenies (Table 2⇓). Map position was known for 245 markers (Table 3⇓), among which a set of 86 mapping markers was selected to provide the widest possible genome coverage and the highest possible polymorphism in terms of number of segregating crosses and parents (percentage instead of absolute number of heterozygous parents was considered because of amplification failure of some loci for one or more genotypes).
Heterozygosity.
Heterozygosity values (Ho) were calculated for each cultivar on a linkage group basis and for the whole genome. Values ranged from 0.568 to 0.727, with 0.638 as mean value in this table grape sample (Table 4⇓). At the linkage group level, heterozygosity values averaged over the 13 cultivars ranged from 0.454 to 0.777, with a mean of 0.645. These results suggest the existence of important differences in level of heterozygosity both among cultivars and among linkage groups. The most polymorphic cultivars were 2121-30 (Ho = 0.727) and Sultanine (Ho = 0.722), whereas Ruby Seedless (Ho = 0.568) was least polymorphic. The most polymorphic linkage groups were LG10 (Ho = 0.777) and LG5 (Ho = 0.772), whereas LG11 (Ho = 0.454) and LG18 (Ho = 0.532) were the least polymorphic. Complete homozygosity was found for linkage group 6 in Autumn Seedless (Ho = 0.000).
Discussion
The potential usefulness of a common set of easy-to-score, single-locus, highly polymorphic, and well-distributed markers, such as SSRs, has already been discussed for other plant species (Macaulay et al. 2001). In grapevine this is the first time that such a tool is developed. It will allow the integration of the maps produced from different populations, thereby yielding improved precision in molecular marker positions for applications in breeding and genome functional analyses. The direct comparison of QTL map positions and effects in different genetic backgrounds based on this set of common markers will permit the identification of common as well as cultivar specific QTLs. Common QTLs that affect the same character and that have been validated in different environments and population structures represent ideal targets for marker-assisted selection, fine mapping, and map-based cloning (Grandillo and Tanksley 1996).
The data generated in this work yielded a realistic estimate of heterozygosity in table grape (0.638). Mean heterozygosity values ranging from 0.680 to 0.861 have been reported (Sefc et al. 1998, Crespan et al. 1999, Sanchez-Escribano et al. 1999, Zulini et al. 2002, Aradhya et al. 2003) for the analysis of 10 to 43 table grape cultivars at 6 to 11 SSRs. The lower values we obtained for table grape could result from the use of a larger sample of SSR loci, since in the above studies heterozygosity values were estimated based on a limited set of highly polymorphic markers, commonly used for grapevine diversity studies. However, the lower heterozygosity values obtained here may also be due to an overall more inbred sample than in previous reports. The same genetic source on both parental pedigree sides might actually have been used during the development of the 13 cultivars analyzed in this study to fix at the homozygous state the favorable alleles for target traits, such as seedlessness, muscat aroma, and berry size.
Differences in the heterozygosity values found among the 13 cultivars could be due to differences in inbreeding during their development. Apart from a few traditional table grape cultivars of uncertain origin (such as Dominga, Sultanine, and Muscat of Alexandria), most of these 13 table grape cultivars are F1 hybrids, in some cases sharing a genotype between paternal and maternal pedigree sides. The differences in heterozygosity observed among linkage groups could suggest presence of chromosomal regions that tended to be fixed during selection in table grape cultivars. However, understanding the basis of these differences will require further research in the identification of the genetic determinants for relevant traits in table grape. The presence of a completely homozygous linkage group (LG6) in Autumn Seedless, a commercial hybrid derived hybrid from the cross between cultivar Calmeria and a F1 from Muscat of Alexandria x Sultanine (Ledbetter and Ramning 1989), cannot be explained based on chromosomal segregation if Calmeria originated from the open pollination of Ohanez, as reported by Loomis and Weinberger (1979). However, the genotype of Calmeria at 18 microsatellite loci is compatible with it being derived from a cross between Ohanez and Sultanine (J. Ibañez et al. 2005, unpublished data), which would explain the extensive homozygosity found at LG6.
Conclusion
A set of 86 highly polymorphic and well-distributed SSRs has been defined to allow the direct comparison of the homologous linkage groups identified in mapping experiments using different grapevine genetic backgrounds. These markers are expected to be useful to integrate grapevine genetic maps and for rapid and efficient genotyping of additional populations in further mapping studies. This SSR mapping set could also be a potent tool to select unlinked polymorphic loci to be used in diversity studies, which should reduce the bias in estimates of diversity and genetic distances among grapevines.
Footnotes
Acknowledgments: This research was supported by the European Community in the framework of the project “MASTER” (Marker Assisted Selection for Table Grape), contract number ICA4-CT-2001-10065.
- Received January 2006.
- Revision received May 2006.
- Revision received July 2006.
- Copyright © 2007 by the American Society for Enology and Viticulture