Abstract
The rootstock Freedom is the offspring of two open-pollinated parents (1613-59 and Dog Ridge 5), sources of root-knot nematode resistance. As phylloxera has increased in importance in vineyards, the risk of Vitis vinifera-derived phylloxera susceptibility from the unknown pollen donors is a concern. To clarify the pedigree of Freedom, genetic profiles of 31 potential progenitors were analyzed with 30 microsatellite markers. We confirm the reported parent (1613-59) and grandparent (1613C) of Freedom and identify 3306C (V. riparia x V. rupestris) as the pollen parent of 1613-59. We were unable to identify the pollen parent of Dog Ridge 5.
In winegrape cultivars, simple sequence repeats (SSRs) have been used to unravel the parentage of ancient varieties, such as Chardonnay, Cabernet Sauvignon, Petite Sirah, and Sangiovese (Bowers et al. 1999, Bowers and Meredith 1997, Meredith et al. 1999, Vouillamoz et al. 2007). The parentage of rootstock varieties is much less obscure; their breeding history extends back only to the 1860s, and partial pedigrees can be constructed from breeder’s records. In rootstocks, molecular markers have effectively been used to characterize collections and identify misclassified accessions (de Andres et al. 2007, Riaz et al. 2007), to analyze genetic relationships (This et al. 1997), and to correct errors in reported parentage (Laucou et al. 2008).
Although rootstock breeding records report the parents of a cross, the information is not always correct or sufficient to reconstruct pedigrees. Some parents used in crosses are themselves the offspring of open-pollination, that is, their maternal parent is known but the pollen donor is not. For example, in the fall of 1935, in Fresno, California, E. Snyder and F.N. Harmon collected seeds from open-pollinated blooms of Dog Ridge (Vitis x champinii) and 1613 Couderc (V. solonis x Othello), which were known to be resistant to the root-knot nematode (Meloidogyne incognita) (Snyder and Harmon 1952). Two of the open-pollinated seedlings, selected for their nematode resistance (1613-59 and Dog Ridge 5), were crossed in 1956 by J.H. Weinberger and F.N. Harmon. A seedling from the 1613-59 × Dog Ridge 5 cross (#91–102) was selected as the rootstock Freedom (Figure 1⇓) (Brooks and Olmo 1997). Freedom is a widely used rootstock, particularly in San Joaquin Valley (California) vineyards where root-knot nematodes are the key soil-borne pest. In the 40 years since its release, Freedom and its half-sibling rootstock Harmony have largely replaced 1613C and Dog Ridge in new plantings because of their nematode resistance and improved horticultural characteristics.
Phylloxera (Daktulosphaira vitifoliae) is an increasing concern in vineyards around the world. In general, all non-vinifera North American grape species are adequate for phylloxera control (Granett et al. 1987). Although Freedom has been reported to be immune to some biotypes of phylloxera (Granett et al. 1987), current recommendations advise against planting rootstocks with unknown but potential V. vinifera parentage in locations where phylloxera is the primary soil pest (Wolpert et al. 2002). It is known for certain that Freedom is at least 1/16th V. vinifera, because the grandparent 1613C is the progeny of V. solonis x Othello, and Othello is a cross of Clinton (V. riparia x V. labrusca) x Black Hamburg (V. vinifera), but it is possible that the pollen parents of 1613-59 and Dog Ridge 5 were also V. vinifera. When the parents of Freedom (1613-59 and Dog Ridge 5) were generated from open pollination, abundant grape pollen was likely available from hundreds of the V. vinifera and interspecific hybrid wine, table, juice, raisin and rootstock varieties recorded to be in the USDA collection at the time (Snyder 1937).
Some components of the pedigree of Freedom can be tested using extant accessions: 1613-59 (presumed parent), 1613C (presumed grandparent), and Dog Ridge (presumed grandparent). However, Freedom’s pollen parent Dog Ridge 5 no longer exists (Figure 1⇑). A variety of methods have been developed that can use genotypic data to reconstruct pedigrees and test for relationships among a population of individuals in organisms including grapevines (Blouin 2003, Jones and Ardren 2003). In order to identify potential pollen parents of Dog Ridge 5 and 1613-59, staminate and hermaphroditic plants from the USDA-ARS rootstock germplasm and breeding vineyard were genotyped with 30 microsatellite markers and analyzed with Kingroup and Cervus software (Kalinowski et al. 2007, Konovalov et al. 2004, Vouillamoz and Grando 2006).
Materials and Methods
Leaves from rootstocks and rootstock selections were collected in summer 2006 from the USDA-ARS Barstow Avenue Vineyard Block, Fresno, CA (Table 1⇓). DNA was extracted from young leaves using the DNeasy 96 Plant Kit (Qiagen, Valencia, CA). Extracted DNAs were quantified using PicoGreen (Invitrogen, Carlsbad, CA) and diluted to 20 ng/μL for SSR analysis.
For PCR, forward primers of publicly available microsatellite markers (Merdinoglu et al. 2005, Sefc et al. 1999, This et al. 2004) were labeled with a fluorescent dye (FAM, HEX, TAMRA, VIC, PET, or NED). Multiplex PCR was carried out in 20 μL reactions that included 20 ng genomic DNA, 1–16 pmol each forward and reverse primers, 200 uM dNTP, 1.6 × Taq polymerase buffer with 1.5 mM MgCl2 and one unit of Taq polymerase (Promega, Madison, WI). Markers were grouped into multiplexes as follows: (1) VMC5A1, VMC8G6, and VVIP10; (2) VMC6G1, VVIN40, and VVIN54; (3) VMC7G3 and VVIM43; (4) VVIB66 and VVIH01; (5) VVIP37 and VVIV16; (6) VRZAG21 and VVIS58; (7) VMC8A7 and VVIH54; (8) VRZAG62, VRZAG79, and VVMD7; (9) VVMD27 and VVS2; and (10) VVIP02 and VVIM11. Singleplex PCR using a lower concentration (1 x) of Taq polymerase buffer was used for the following markers: VVIC50, VVIN70, VVMD5, VMC1C10, VVIV69, VMC1E12, and VVIF52. An i-Cyler (Bio-Rad, Hercules, CA) and a PTC100 (Bio-Rad) were used to perform the following amplification protocol: 4 min at 94°C; 35 cycles of 1 min at 94°C, 30 sec at annealing temperature, 30 sec at 72°C; 1 hr at 72°C. The long final extension time allows Taq to add an additional adenine to all fragments. An annealing temperature of 57°C was used for all loci except VVIC50, where 50°C was used.
Following PCR amplification, 4 to 8 amplification products were pooled; 1 μL of the pooled products was combined with 10 μL Hi-Di formamide and 0.15 μL of the internal size standard GeneScan 500 LIZ (Applied Biosystems, Foster City, CA). Following denaturation for 4 min at 94°C, the samples were analyzed on an ABI3730 Genetic Analyzer (Applied Biosystems) at the Cornell Life Sciences Core Sequencing Lab in Ithaca, NY. Chromatogram analysis and allele-calling were performed with GeneMapper 3.7 software (Applied Biosystems).
Cervus software was used to calculate the likelihood of parentage, expressed as the LOD score (natural log of the likelihood ratio of true parent versus not true parent) of every candidate parent based on the allele frequencies (Kalinowski et al. 2007, Marshall et al. 1998). Settings used were as follows: candidate father = 30; prop. sampled = 0.80, prop. loci typed = 0.0915, prop. loci mistyped = 0.01. Kingroup software (Konovalov et al. 2004) was used to estimate the likelihood that pairs of individuals fit a particular relationship and to perform likelihood ratio tests for these hypotheses based on pairwise relatedness (r) (Pamilo 1990) and allelic identity by descent (R m, R p). Because the individuals in this study are not from a randomly breeding population, we estimated the background allele frequency from the whole population of individuals, simultaneously simulated with pedigrees (Konovalov and Heg 2008). We then tested the likelihood of the following hierarchical pairwise relationships between individuals, specifying the identity by descent of each allele (R m, R p): self (1,1); parent-offspring (1,0); full sib (0.5,0.5). We used both the standard descending ratio algorithm alone and with the Simpson assist (Konovalov et al. 2005).
Results and Discussion
To validate reported elements in the Freedom pedigree, initial parentage analysis was conducted with Cervus to test the hypotheses that 1613-59 is the parent of Freedom and 1613C is the parent of 1613-59. The pairwise LOD score for parent-offspring pair 1613-59 and 1613C was 3.80e01; for 1613-59 and Freedom it was 2.38e01. This is consistent with each parent sharing one allele with their putative offspring (Table 2⇓). Kingroup also confirmed that 1613-59 is a possible parent (versus full sib) of Freedom (r = 0.48; likelihood ratio 3.26) and that 1613C is a parent (versus full sib) of 1613-59 (r = 0.50; likelihood ratio 4.12) (Table 3⇓). The results were very similar when background allele frequencies were estimated from likely pollen donors or assigned equal frequencies per locus (data not shown).
Having established the maternal lineage, all of the extant potential pollen donors from the Barstow Avenue Block could be tested for paternity of 1613-59 using 1613C as the known mother. Analysis with Cervus identified rootstock 3306C as the most likely pollen donor with a pairwise LOD of 2.72e01; it was the only individual in the data set with a positive LOD score. Kingroup also identified the likelihood of 3306C as a parent of 1613-59 (versus the likelihood that it is a full sib) with r = 0.45 and a likelihood ratio of 3.93 (Table 3⇑). This is within the range of r values reported for established parent-offspring pairs from Pinot-Syrah crosses: 0.41 to 0.63 (Vouillamoz and Grando 2006). Locus VVIP10 is the only locus which is not consistent with paternity of 3306C: 1613-59 has a nonparental allele (182 bp, Table 2⇑). This inconsistency is not the result of genotyping error and is likely the result of mutation. Nonparental alleles have been observed in grapevine parent-offspring pairs and among clones, indicating that mutations are fairly common (Riaz et al. 2002, Vouillamoz and Grando 2006). An alternative hypothesis would be that the true parent is not 3306C, but rather one of its full siblings. However, historical records show that 3306C and 3309C were the only members of the Courderc family of V. riparia x V. rupestris rootstocks present in the vineyard in the 1930s (Snyder 1937). Our data set also included 3309C, a full sibling of 3306C, and in contrast to 3306C it had seven loci with mismatches to 1613-59 (Table 2⇑) and a LOD score of −1.65e01, indicating that this close relative of 3306C can be eliminated from parentage of 1613-59.
Although we only tested a subset of the hundreds of grapevine varieties held by the UDSA in the 1930s, the statistical support for 3306C provides conclusive evidence of its role in Freedom’s pedigree. 3306C is a V. riparia x V. rupestris hybrid made by George Couderc in France in 1881 under selection for phylloxera resistance and lime tolerance. It was present in the breeding materials of the USDA collection in Fresno at the time the open-pollinated seeds of 1613C and Dog Ridge were collected (Snyder 1937). 3306C has moderate vigor, resistance to the root-knot nematode, and has been shown to be resistant to type B phylloxera (Granett et al. 1987). Freedom, 1613-59, and 3306C share the morphological character of dense pubescence on petioles and young stems. In contrast, this trait is not observed on 1613C, 3309C, Dog Ridge, or 3306C full siblings 3308C and 3310C, which is further confirmation of the pedigree relationships among Freedom, 1613-59, and 3306C (Galet 1988). With 3306C as a known non-vinifera grandparent of Freedom, the contribution of V. vinifera in the pedigree is limited to the known great-great-grand-parent Black Hamburg and possibly the unknown pollen parent of Dog Ridge 5.
There are some inherent disadvantages to reconstructing pedigrees among breeding materials. While Kingroup and other likelihood-based methods are designed to robustly reconstruct pedigrees in randomly mating populations, small samples from related populations can be problematic (Konovalov and Heg 2008). In addition, genotyping errors can affect likelihood-based methods. Typical error rates range from 2.0 to 2.5% in microsatellite screens, and for most relationships, this will not affect likelihood estimates very much. The exception is self and parental relationships between individuals, where a single mismatch can drop the estimated relationship to zero (Blouin 2003). Some likelihood-based methods such as Cervus incorporate an estimated error rate (in our case 0.01), that allows them to accommodate genotyping errors between self and parent-offspring relationships.
Other relationships emerged from this data set (Table 3⇑). Ramsey and Rotundifolia are so similar that their genetic relatedness was estimated at 1 (self); they are also similar morphologically (P. Cousins, personal observation, 2006). This may represent an historical misidentification in the collection. Oppenheim 4 (also known by the acronym SO4) is a selection from crosses between V. berlandieri and V. riparia (Galet 1988). Teleki 5G (also known as Geisenheim 56) was reported to be a cross between V. vinifera cv. Trollinger and V. riparia, but it had an identical genetic fingerprint to Oppenheim 4, indicating a historical misidentification in the collection, which has been seen in other collections (de Andres et al. 2007, Walker and Boursiquot 1992). The pistillate parent of 1613C is reported to be a member of the species V. solonis, but the representative V. solonis accession tested was not the parent of 1613C.
We also attempted to identify the pollen parent of Dog Ridge 5 to determine the other grandparent of Freedom. Dog Ridge 5 itself no longer exists, and with the genotypes of Dog Ridge, Freedom, and Freedom’s half sib Harmony, we were unable to conclusively assign this relationship among the samples listed in Table 1⇑. Even using simultaneous simulation of background allele frequencies while determining pedigree relationships (Konovalov and Heg 2008), there was not sufficient information to uniquely identify this individual among the available samples. Comparison of the genotypic data for Dog Ridge and Freedom shows shared alleles at 58% of loci screened (Table 2⇑), which is consistent with their reported relationship (Figure 1⇑).
Conclusions
Through parentage analysis we have identified 3306C as the parent of 1613-59 and grandparent of Freedom. Complete analysis of Freedom ancestry and assessment of the risk that it harbors V. vinifera-derived recessive alleles for phylloxera susceptibility still requires identification of the pollen donor that is the male parent of Dog Ridge 5. In addition, identification of phylloxera resistance genes will give a direct assessment of the susceptibility of Freedom and other rootstocks.
Footnotes
Acknowledgments: This research was financially supported by the USDA-ARS (projects 1910-21220-004-00 and 1910-21220-003-00).
The authors thank Warren Lamboy for initial assistance with computational analysis; Kathleen Deys, Rick Emershad, Hong Lin, and Parminder Sahota for technical assistance; Dave Matthews and Gayle Volk for comments and suggestions on the manuscript; and two anonymous reviewers for valuable comments.
- Received September 2008.
- Revision received January 2009.
- Accepted February 2009.
- Published online September 2009
- Copyright © 2009 by the American Society for Enology and Viticulture