Use of microsatellite markers to assess the identity and genetic diversity of Vitis labrusca and Vitis rotundifolia cultivars

Ten grapevine cultivars were genotyped at eight microsatellite loci to characterize their identity and genetic diversity. Of these, nine cultivar profiles matched with those of databases and ‘Magoon’ matched with ‘Regale’ in the present study and ‘Regale’ in the University of California (Davis) database, implicating a likely error in planting. The number of alleles ranged from 5 (VVM5) to 9 (VVMD31), and the observed heterozygosity ranged from 37.14 (VVMD5) to 97.14% (VVMD27), with no significant differences in relation to the expected values for any of the loci, with the exception of VVMD5. The polymorphism information content values were observed to be above 0.25 in more than 85% of the loci analyzed, and VVMD31 was the most informative. The UPGMA analysis clustered the cultivars into two distinct groups. Within each group, the most divergent cultivars were ‘Bountiful’ (V. rotundifolia) and ‘Goethe’ (V. labrusca), also exhibiting the largest number of private alleles, 4 and 7, respectively. When comparing the two groups, the most divergent accessions were ‘Bountiful’ and ‘Bordo’, with the highest Nei distance. It was demonstrated that there is sufficient genetic variability in the cultivars used in this study to support breeding programs.


Introduction
Grapevine belongs to the Vitaceae family, comprising 19 genera.Vitis is the only genus that produces edible fruits and is divided into two subgenera or sections, namely Euvitis (bunch grapes) and Muscadinia (muscadine grapes).Muscadine grapes are genetically and morphologically distinct from the species within the subgenera Euvitis, and the most obvious genetic difference between these two taxa is the number of somatic chromosomes: Muscadinia species have 40, and Euvitis species have 38.Muscadinia species also differ from Euvitis species in their seeds, bark, tendrils and cluster morphology (OLIEN, 2001).
Euvitis species are divided into three geographical groups: American, Asian and European.Vitis vinifera, the main species currently cultivated worldwide, is the only member of the European group and has produced thousands of cultivars.The American group has approximately 30 species.Among these, V. labrusca, known for its 'foxy' fruit flavor, was widely used in the 18 and 19th centuries directly as rootstock for V. vinifera cultivars due to its phylloxera tolerance and incrossings originating the so-called 'direct producer' vines.These hybrids were planted ungrafted and were bred to incorporate the resistance to the foliar diseases of the North American grapes with the fruit characteristics of the European grapevines (POMMER, 2003).Muscadinia is endemic to the southeastern United States of America, and, among the three species known, only Vitis rotundifolia Michx. is of commercial value, possessing a very strong resistance to grape pests and diseases, and is employed worldwide in breeding programs (OLIEN, 2001).
The knowledge of the genetic diversity of species has two advantages in a breeding program: the first concerns the genetic heterogeneity that limits the vulnerability of a species to pests and diseases; and the second is related to providing a large supply of allelic variation, which can be used for creating new combinations of favorable genes.The feasibility of using genetic diversity as a criterion for the selection of the parents for hybridization has been demonstrated in several species, including grapevine (OLIVEIRA et al., 2005).
DNA-based molecular markers offer an advantage over morphological descriptors, as they are less prone to being affected by the environment or the developmental stage (CARIMI et al., 2010).Simple sequence repeat (SSR) markers have been increasingly used as molecular descriptors in grape.The usefulness of SSRs has been widely demonstrated in the identification and characterization of stock and rootstock varieties (DZHAMBAZOVA et al., 2007), evaluation of genetic variability (RIAZ et al., 2008;SCHUCK et al., 2009), pedigree studies (ORTIZ et al., 2004;VARGAS et al., 2009), and genetic mapping (RIAZ et al., 2004).The development of a standard set of microsatellite markers for the identification of grape cultivars was proposed by the European Project GENRES 081 using six microsatellite loci (THIS et al., 2004).This set of markers, VVS2 (THOMAS; SCOTT, 1993), VVMD5, VVMD7 (BOWERS et al., 1996), VVMD27 (BOWERS et al., 1999), VrZAG62 andVrZAG79 (SEFC et al., 1999), was used by this project to evaluate 13 reference cultivars in Europe and are widely recommended by the International Organization of Vine and Wine for grapevine genotyping (OIV, 2007).
Considering that the knowledge about the genetic diversity of a specie movides subsidies to a breeding program, this study aimed to confirm the identity and estimate the genetic diversity among ten grapevine cultivars using eight microsatellite markers.

SSR analysis
The DNA of the Vitis labrusca cultivars was analyzed in a previous study by Schuck et al. (2009).
The Genomic DNA of the Vitis rotundifolia cultivars was extracted using a modified CTAB (hexadecyltrimethylammonium bromide) procedure (RIAZ et al., 2004).In the final step, the DNA pellets were suspended in 100 μL 1X Tris-EDTA buffer and stored at -20°C.The DNA quality was visualized on 1.2% agarose gels stained with ethidium bromide (10 mg mL -1 ), and the samples were then diluted in ultrapure water to a final concentration of 10 ng μL -1 .
The PCR reactions were performed in 10 μL reaction mixtures containing 5 pmol each primer, 2.5 mmol L -1 each dNTP, 1 μL 10X gold PCR buffer (Perkin Elmer Inc., Wellesley, MA, USA), 0.5 unit AmpliTaq Gold DNA polymerase (Perkin Elmer Inc., Wellesley, MA, USA), 2 mmol L -1 MgCl 2 and 10 ng genomic DNA.The temperature cycling for the PCR was performed using either a Peltier Thermal Cycler-200 (MJ Research, Inc., Waltham, MA, USA) or a Bio-Rad iCycler (Bio-Rad Laboratories, Hercules, CA, USA).The following cycling program was used: denaturation of DNA and activation of Taq DNA polymerase at 95°C for 10 min.; 35 cycles of amplification consisting of 45 s at 94°C, 45 s at 56°C and 1 min.at 72°C; a final extension of 10 min.at 72°C; and cooling at 4°C.To separate the amplification products, the PCR reactions were mixed with denaturing dye (98% formamide, 10 mmol L -1 EDTA, 0.05% bromophenol blue and xylene cyanol) and heated at 94°C for 2 min.before loading onto 5% polyacrylamide sequencing gels.The gels were electrophoresed at a constant 70 W for 2-3 hours, depending on the allele sizes.The samples were visualized by silver staining using a commercial kit (Promega, Madison, WI, USA).All of the gels were visually examined and scored on a light box and were then digitally scanned to preserve the images.
To determine the size of the DNA fragments obtained by the amplification of the microsatellite products of the V. rotundifolia samples, aliquots of the V. rotundifolia DNA already diluted and available in the laboratory were loaded on the gels.These samples are part of the Grape Germplasm Collection of the University of California -Department of Viticulture and Enology, and the genetic profile in base pairs (bp) is in the reference database for the identification of grapevine that is maintained by the Foundation Plant Service, University of California, Davis, CA, USA.The genetic profile of V. labrusca was obtained in a previous study by Schuck et al. (2009) using the same microsatellite markers used in the present study (VVS2, VVMD5, VVMD7, VVMD27, VrZag62 and VrZag79).

Data analysis
Various genetic parameters for the ten cultivars over the eight SSR loci were calculated.The GDA v1.0 program package software (LEWIS; ZAYKIN, 2001) was used to calculate the allelic frequencies, number of alleles per locus (N), percentage of polymorphic loci (P), observed heterozygosity (H o ) and expected heterozygosity (H e ).The H e values were estimated using the following formula (NEI, 1972): , where i p is the frequency of the ith allele.The GENES v2007.0.0 program package software (CRUZ, 2007) was used to determine the polymorphic information content (PIC).The PIC values were calculated according to Botstein et al. (1980), as follows: where i p and j p are the frequencies of the ith and jth allele.The probability of null alleles (r) was calculated for each microsatellite, as described by Brookfield (1996): , where H e is the expected heterozygosity and H o is the observed heterozygosity.
The genetic distance between the cultivars were estimated as follows (NEI, 1972): where ln is the natural logarithm and (I) is the measure of the genetic identity, as calculated using the following formula: Acta Scientiarum.Agronomy Maringá, v. 36, n. 3, p. 301-308, July-Sept., 2014 considering i p and i q as the frequencies of the i allele in the P and Q cultivars.Based on the matrix of Nei's (1972) genetic distance, a dendrogram was constructed using the Unweighted Pair Group Method Arithmetic Average (UPGMA).To verify the adjustment between the distance matrix and the dendrogram, a cophenetic correlation coefficient (r) was applied (SOKAL; ROHLF, 1962) using the GENES program package software (CRUZ, 2007).

Results and discussion
Among the five V. rotundifolia cultivars analyzed based on eight SSR loci, four different SSR profiles were detected (Table 2).'Magoon' matched the same genetic profile of 'Regale' in the present study and the 'Regale' present in the Grape DNA Identification Reference Database, which is maintained by the Foundation Plant Service (University of California, Davis), indicating that the genotype was misnamed at the time of introduction and that the same genotype was planted under different names.
The occurrence of misidentification is common, particularly for such old clonal species as Vitis, and it has been observed that 5 to 10% of the grape cultivars maintained in grape collections are incorrectly annotated (ANDRÉS et al., 2007).The identification of the V. labrusca cultivars was part of a previous study (SCHUCK et al., 2009), and the SSR profiles of 'Bordo', 'Goethe', 'Isabel', 'Niagara Rosada' and 'Marta' were the same as those available in the database for the same cultivars using the same SSR markers used in the present study.Therefore, these cultivars were correctly identified (Table 2).
Table 3 shows the alleles and their frequencies obtained for the different loci, with the allele sizes ranging from 118 bp (VVS2) to 301 bp (VVMD32).There were found important differences in the allelic composition, reflecting the existence of different sets of alleles for V. rotundifolia and V. labrusca and the existence of private alleles in each set.There were 23 alleles specific for the V. rotundifolia cultivars, 32 specific to V. labrusca and three alleles (246, 249 and 259 bp) common to the two groups.Private alleles were found in the V. rotundifolia cultivars Bountiful (4) and Carlos (3) and in the V. labrusca cultivars Niagara Rosada (2), Goethe (7) and Marta (1).The most frequent alleles were  showing frequencies greater than 20%.In contrast, 18 alleles were relatively infrequent (a frequency less than 10%).It is also worth noting that, of the alleles with a frequency higher than 20%, seven were unique to V. rotundifolia, further emphasizing the differences in the allelic composition of this species.
The alleles found in this study and in another study (RIAZ et al., 2008) indicate that the V. rotundifolia alleles are very different in size and frequency from those of Euvitis species.Riaz et al. (2008) reported the presence of specific alleles in a genetic diversity study of 57 accessions of V. vinifera and V. rotundifolia and a VR hybrid (V.vinifera x V. rotundifolia), including the loci VVMD7-245,  also used in our study.These authors also found that, of the total 184 alleles of 14 microsatellites loci used, 88 were specific to V. rotundifolia.In the present study, there was a relatively high number of alleles (39%) specific to the V. rotundifolia cultivars (Table 4); however, these alleles were common among the group of V. rotundifolia cultivars, suggesting that the genetic base of these cultivars is limited.
The main genetic parameters, such as the number of alleles per locus (N), percentage of polymorphic loci (P), observed ( o H ) and expected ( e H ) heterozygosity, estimated frequencies of null alleles (r) and polymorphic information content (PIC) for the eight SSR loci are shown in Table 4   (1) low value; (2) high value.
A total of 58 alleles were amplified, ranging from five (VVM5) to nine (VVMD31), with an average of 7.25 alleles per locus.These SSR loci have previously been used in the genetic characterization of other grape cultivars (LEÃO et al., 2009).Moreno-Sanz et al. (2008) analyzed the diversity among 46 accessions of V. vinifera and obtained a total of 37 alleles, with an average of 6.2 alleles per locus using six of the eight loci used in the present study.A larger number of alleles per SSR loci were observed in the V. vinifera cultivars (11) (MARTÍN et al., 2003), French-American hybrid cultivars (11) (POLLEFEYS; BOUSQUET, 2003) and table-grape cultivars (18) (IBÁÑEZ et al., 2009).In the case of the V. vinifera and table-grape cultivars, it is likely that the large number of cultivars analyzed (176 and 376, respectively) could account for the marginally higher number of alleles detected.
The observed heterozygosity varied between 37.14% (VVMD5) to 97.14% (VVMD27), with a mean value of 79.44%, slightly higher than the corresponding expected heterozygosity for all of the loci, with the exception of the VVS2, VVMD5 and VVMD32 loci.The estimated frequency of null alleles (r) was negative for five out of eight loci and positive for each of those loci that showed o H < e H and only high for VVMD5.For the VVMD5 locus, the e H was high (71.47%),as with the other loci, but a very low o H was observed (37.14%).For the other loci (VVS2 and VVMD32), the e H value was slightly higher than the o H value, and the frequency of null alleles was very low (Table 4).Nevertheless, the e H and o H values were high for these two loci.Such null alleles can arise when mutations prevent the primers from binding to the region.Cipriani et al. (2008) reported that a locus with a positive null allele value indicates an excess of homozygotes, but this does not necessarily imply that a null allele is present.Hence, the assumption of homozygosity rather than heterozygosity for a null allele could be considered adequate in the present study.Indeed, the highest number of homozygous cultivars in relation to the others was found for the VVMD5 locus (Table 2), a trend also observed by Martín et al. (2003).In the present study, the average value of observed heterozygosity (79.44%) was higher than that obtained for the V. vinifera (70.7%) (IBÁÑEZ et al., 2003) and V. rotundifolia (60%) (RIAZ et al., 2008) cultivars but is in the same range as that obtained when rootstock cultivars are included in the analyses (ANDRÉS et al., 2007;THIS et al., 2004).Additionally, it was observed in the present work that the values of the expected and observed heterozygosity were greater than 70% for the majority of the loci (93.75%), indicating a high genetic variability of the markers analyzed.A high heterozygosity is commonly found in vegetatively propagated perennial species, such as grape (ARADHYA et al., 2003).
To characterize the usefulness of the eight SSR markers, the polymorphic information content (PIC) of each locus was assessed (Table 4).The PIC values were higher than 0.25 for more than 85% of the loci analyzed.The loci VVS2, VVMD27 and VrZAG79, which exhibited eight, seven and seven alleles, respectively, were also the most informative loci in other studies (MARTÍN et al., 2003).The number of alleles per locus, expected heterozygosity and polymorphic information content indicate VVMD31 as the locus with the highest information content (nine alleles, e H = 85.59% and PIC = 0.34), whereas VVMD5 (PIC = 0,19) is the least informative.According to Botstein et al. (1980), markers with PIC values between 0.25 and 0.50 are considered moderately informative, and the markers chosen in this study are consistent with this criterion.The mean value of PIC (0,2825) and the number of alleles per locus (7,25) obtained in our study were lower than in previous studies: 0,782 and 11 (MARTÍN et al., 2003) and0,777 and7,8 (GONZÁLEZ-ANDRÉS et al., 2007), respectively.The authors of each study used six microsatellite loci, five (VVS2, VVMD5, VVMD7, VrZAG62 and VrZAG79) equal to those used in our study, to characterize the genetic diversity of 318 and 79 Spanish cultivars, respectively.The lowest number of alleles per locus and the PIC value for the loci used in this study can be attributed to the smaller number of cultivars analyzed (9).
A dendrogram depicting the genetic distance among the ten cultivars analyzed at eight loci was constructed using the UPGMA method (Figure 1).The cophenetic correlation coefficient (r) was 0.97, revealing a correlation between the dendrogram and the distance matrix.According to Sokal and Rohlf (1962), values of r ≥ 0.8 indicate a good fit between the original matrix and the cophenetic matrix values.
The dendrogram analysis identified two main groups: group I, containing the five V. rotundifolia cultivars ('Bountiful', 'Carlos', 'Magnolia', 'Regale' and 'Magoon'), and group II, with the five V. labrusca cultivars ('Bordo', 'Goethe', 'Isabel', 'Martha' and 'Niagara Rosada') (Figure 1).This result was expected because, in addition to the chromosomal differences, there are important distinguishing anatomical and morphological features between these species.For example, Euvitis species, such as V. labrusca, have a haploid number of 19 chromosomes, longitudinally shredding bark, lack lenticels on the stems, and possess a non-continuous pith that is segmented by diaphragms.In contrast, V. rotundifolia has 20 haploid chromosomes, adherent bark, prominent lenticels on the shoots and fruit, and a continuous pith lacking diaphragms (POMMER, 2003).
As expected, 'Regale' and 'Magoon' in the V. rotundifolia group presented a distance equal to zero (Figure 1).'Magoon' showed the same genetic profile as the 'Regale' in this study and the 'Regale' present in the grape reference database (Table 2).The distance from 'Magnolia' to 'Regale' was 26%, and the distance from 'Carlos' to this group was 38.12% (Figure 1).'Magnolia' shared 74% of its alleles with 'Regale', whereas 'Carlos' shared 67 and 62% of its alleles with 'Magnolia' and 'Regale', respectively (Table 2).These three cultivars have their geographic origin in North Carolina, and 'Magnolia' is also reported to be in the pedigree of 'Regale', showing the genetic similarity between these cultivars (RIAZ et al., 2008).'Bountiful' was the most distant cultivar from all of the others in the same group, with a distance of 85.76% (Figure 1).Moreover, 'Bountiful' showed the largest number of private alleles within the V. rotundifolia group (Table 3).In the V. labrusca group, 'Bordo' and 'Marta' were separated by a distance of 20.8% (Figure 1) and differed genetically only at four alleles (Table 2).'Niagara Rosada' was separated from 'Bordo' and 'Marta' by a distance of 41.58%, and 'Isabel' was separated from this group by 48.51%.'Goethe' was the cultivar most distant from all of the others of the same group, with a distance of 74.1% (Figure 1) and, within the V. labrusca group, was the most divergent cultivar and with the largest number of private alleles (Table 3).
The knowledge of the most divergent cultivars is essential to guide breeders in choosing the best combination for a cross.However, when selecting the parents, the agronomic characteristics of the genotypes and the purpose of the breeding program should be considered in addition to the genetic divergence.In a breeding program aimed at creating rootstocks, it is essential to choose cultivars with resistance and/or tolerance to biotic and abiotic stress.Regarding the cultivars characterized in this study, it is known that the V. rotundifolia cultivars are resistant to almost all of the pests and diseases of grapevine.It is worth highlighting the cultivars with resistance ('Bountiful' and 'Magnolia') and tolerance ('Regale') to the main soil pest of southern Brazil, ground pearl (BOTTON; COLLETA, 2010).The cultivars belonging to the V. labrusca group are well adapted to the vineyards of southern Brazil and display tolerance to Fusarium oxysporum f.sp.herbemontis (GARRIDO et al., 2004), making them an important genetic source for grapevine breeding programs.

Conclusion
SSR markers are efficient tools to confirm the identity and to assess the genetic diversity of Vitis cultivars.
The dendrogram clustered the ten grapevine cultivars in two groups, proving to be efficient in the separation of the cultivars based on their genealogy and botanical classification, showing relationships among cultivars of the same group.
The V. rotundifolia and V. labrusca groups have a high genetic diversity and can be used to support breeding programs.

Table 1 .
Description of the ten grapevine cultivars, type of flower, fruit color, species or parentage and geographical origin.

Table 4 .
Genetic parameters at the eight microsatellite loci, number of alleles per locus (N), percentage of polymorphic loci (P), observed (H o ) and expected ( e H ) heterozygosity, estimated frequencies of null alleles (r) and polymorphic information content (PIC), calculated by grouping all of the cultivars into one group.