Abstract
Over the past 15 years, the grape breeding program at the University of California, Davis, has been evaluating Vitis rupestris x Muscadinia rotundifolia selections for resistance to the dagger nematode, Xiphinema index, and to Pierce’s disease (PD). Selections from these crosses exhibit very strong resistance to X. index and PD. In addition to breeding efforts, populations from these crosses have been used to develop genetic maps and locate resistance loci. Genetic mapping efforts recently began incorporating SSR markers to refine and expand existing maps. The use of SSR markers revealed that the mapping population parents were not crosses of V. rupestris x M. rotundifolia. This discovery led to testing of the entire group of 161 V. rupestris x M. rotundifolia progeny. All possible male parents surrounding the V. rupestris female parents in the vineyard where the crosses were made were genetically fingerprinted with up to 15 SSR markers to determine the true male parents. Results indicated that most of the male parents were from collections of forms of V. arizonica gathered in Mexico in 1961. These now correctly identified selections represent novel sources of very strong resistance to X. index and PD.
In 1988, a series of crosses were made under the direction of H.P. Olmo, at the University of California, Davis (UCD). Olmo selected two pistillate cultivars of Vitis rupestris (A. de Serres and Wichita Refuge) and crossed them with pollen from six Muscadinia rotundifolia cultivars (Carlos, Cowart, Dixie, Magnolia, Southland, and Trayshed) in an effort to develop disease-resistant hybrids, which would combine the very high pest resistance of M. rotundifolia with the easy to root and graft nature of V. rupestris. The first six populations, 8908 to 8913, were crosses of V. rupestris A. de Serres by the six M. rotundifolia cultivars, in the above order. The second set of six populations, 8914 to 8919, were crosses of V. rupestris Wichita Refuge by the same ordered set of M. rotundifolia cultivars. The M. rotundifolia cultivars were all grown as potted plants under greenhouse conditions to force early bloom so that fresh pollen could be used in the crosses.
In the fall of 1988, berries were harvested and seeds extracted to produce 161 seedlings that were planted in 1989. Since then, these populations have been tested for resistance to root-knot nematode (Meloidogyne incognita), dagger nematode (Xiphinema index), and Xylella fastidiosa, the bacterial causal agent of Pierce’s disease (PD) (unpublished results). Two individuals were selected from the 8909 population (V. rupestris A. de Serres x M. rotundifolia Cowart): 8909-15, resistant to PD and X. index, and 8909-17, with PD resistance only. They were crossed to produce the 9621 genetic mapping population, in which X. index resistance segregated as a single dominant gene. Genetic mapping efforts commenced and a map based on amplified fragment length polymorphism (AFLP) DNA markers was published (Doucleff et al. 2004). Refined mapping efforts with simple sequence repeat (SSR) markers, which are capable of revealing parentage, indicated that M. rotundifolia Cowart was not the correct male parent for 8909-15 and 8909-17 (Krivanek et al. 2006). This discovery initiated an intensive effort to determine the true parents of these two selections and the entire collection of 161 progeny from the 12 crosses.
Testing of the V. rupestris x M. rotundifolia progeny has reveled very strong sources of resistance to PD and to root-knot and dagger nematode. Selections from these crosses are being used as parents to breed resistant root-stocks and scions. Given the strong pest and disease resistance and the use of these selections in breeding, it was important to determine the true parentage of these selections to understand the genetic basis of their resistance. This report details the parentage of the 12 crosses made to produce the “89 group” populations. The results of this study support the value of SSR markers at correctly discerning parentage, promote further research on pollen dispersal in grape, and have identified novel sources of pest and disease resistance.
Materials and Methods
Plant material.
In 1988, Prof. Olmo made the original crosses on two V. rupestris pistillate vines, A. de Serres and Wichita Refuge, in block M of the Department of Viticulture and Enology Armstrong vineyard, UCD. The 161 seedling progeny from these crosses were sampled for DNA extraction. In order to determine the identity of potential male parents, this block was surveyed and all staminate and hermaphroditic grapevines that were blooming at the same time as V. rupestris A. de Serres and Wichita Refuge were sampled for DNA extraction. To determine the rate of natural outcrossing, 200 seeds from open pollinated clusters of A. de Serres were collected for DNA extractions.
DNA extraction and SSR genotyping.
Small-scale DNA extractions were performed on young leaves, shoot tips, and seeds using a modified CTAB (cetyltrimethyl-ammonium bromide) procedure as described elsewhere (Lodhi et al. 1994). Samples were placed in grinding bags with 5-mL extraction buffer containing β-mercaptoethanol and ground with a Homes 6 mechanical homogenizer (Bioreba, Longmont, CO). The seeds were hammered gently in the grinding bags before using the homogenizer to avoid rupturing the bags. Two milliliters of each sample were used in the remaining steps of the DNA extraction protocol, and the remaining 3.0 mL of each sample was stored at −4°C in the freezer for possible reuse if needed.
A total of 14 SSR markers were used to analyze the progeny of A. de Serres (population codes 8908 to 8913), and 15 SSR markers were used to analyze the progeny of Wichita Refuge (8914 to 8917). The majority of the markers were used for both progeny sets (Table 1⇓). Ten of these SSR markers were used to analyze the A. de Serres open-pollinated seed. Simple sequence repeat primer sequences were obtained from different sources, including previously published primers (VrZAG62, VrZAG79; Sefc et al. 1999) and the Vitis Microsatellite Consortium (VMC) coordinated by AgroGene S.A. Moissy Cramayel, France (now Eurofins; www.eurofins.com), these latter primer sequences are now publicly available from the NCBI uni-STS database for Vitis (www.ncbi.nlm.nih.gov/). Amplification conditions for the SSR markers were previously described (Riaz et al. 2004). The amplified products of SSR markers were separated on denaturing 6% polyacrylamide sequencing gels and visualized by silver staining with a commercial kit (Promega, Madison, WI). Each marker was amplified and scored twice. Allele sizes in base pairs were determined by direct comparison to a sequencing reaction present on each gel. A set of six samples was used as standards on each gel to ensure consistent scoring across gels. All gels were visually examined and scored on a light box and were then scanned to preserve them as a digital image. Parentage analysis was performed by visually comparing the allele sizes of progeny plants to the list of potential male parents for all markers.
Results and Discussion
In 1988, Prof. Olmo used pollen from six different M. rotundifolia cultivars (Carlos, Cowart, Dixie, Magnolia, Southland, and Trayshed) to pollinate the pistillate V. rupestris selections, A. de Serres and Wichita Refuge. The success rate of Vitis (2n = 38) x Muscadinia (2n = 40) crosses is low because of several factors, the first being the different chromosome numbers of the parents, which normally results in sterile (2n = 39) progeny when the two species are crossed. Phenological factors also greatly limit the success of a V. rupestris x M. rotundifolia cross; V. rupestris blooms early in the season (generally mid-April in Davis, CA) while M. rotundifolia blooms late (late May through June in Davis). Olmo obtained the pollen for these crosses from M. rotundifolia plants that were forced into an early bloom in the greenhouse so that fresh pollen could be used in the crosses as opposed to storing M. rotundifolia pollen from the previous year. Grape pollen can be stored dry at 4°C for a year and still retain some viability, but fresh pollen is far more successful (Olmo 1942).
Although fresh pollen was used to make these 12 crosses, analysis of the SSR marker allele sizes determined that only five seedlings out of 161 were true V. rupestris x M. rotundifolia hybrids and that these five seedlings originated from three different crosses. The vineyard block where the outcrossing occurred was carefully inspected when the V. rupestris female parents were blooming, and 40 staminate or hermaphroditic accessions were selected as possible pollen parents based on the overlap of their flowering times with V. rupestris. Parentage analysis revealed that 14 different staminate or hermaphroditic plants, consisting of breeding selections and Vitis species collected by Olmo in Mexico in 1961 fathered 146 of the seedlings. The male parent for 10 of the seedlings, two from the A. de Serres populations and eight from the Wichita Refuge populations, could not be identified (Table 2⇓). It is probable that the pollen donor parents for these 10 progeny no longer exist, given that the crosses were made in 1988 and that the vineyard block had not been cultivated since then.
Once the parentage analysis was complete, the progeny were organized into new full sibling populations by adding the letters A through S as a prefix to the original 8908 through 8919 population numbers (Table 2⇑). Most of the seedlings (55) in the A. de Serres group were crosses with the staminate V. rupestris Pillans (population C), which was adjacent to (1-m spacing), and had shoots intertwined with A. de Serres. The next largest population (R) consisted of 46 individuals and was a cross of Wichita Refuge x b40-14, a form of V. arizonica from northern Mexico. b40-14 was growing about 7.5 m from Wichita Refuge (Figure 1⇓).
Olmo certainly placed paper bags over the clusters of both female vines to prevent contamination of the crosses by other than the applied M. rotundifolia pollen. Wild grapevines are dioecious and are pollinated by wind and insects. The flowers have nectaries that produce a strong odor, presumably to attract insect pollinators such as bees and wasps. Insects are also attracted to the staminate flowers by the copious amounts of pollen they produce. The role of wind as a pollination agent cannot be ignored. Flowering of the early species usually begins in mid-April in Davis. Warm temperatures during the first two weeks of April 1988 would have accelerated the flowering of the early blooming species and selections evaluated in this study. During the first 12 days of April 1988, the average high temperature was 26.6°C and the high daily temperature on April 10 and 11 was 33°C (California Irrigation Management Information Service data, www.cimis.water.ca.gov/cimis/data.jsp). These warm temperatures could have induced early anthesis in the male accessions, promoted pollen drying, and thus increased pollen movement. The majority of the male plants were in a row 11 m to the north of the pistillate A. de Serres and Wichita Refuge vines (Figure 1⇑). The pollen from four other male plants traveled an even greater distance (14.6 to 30 m). It is possible that the warm weather induced earlier bloom and that gusty winds carried the drier pollen farther distances where it was deposited on the flower clusters of female V. rupestris plants before they bloomed. Wind could have also deposited pollen on the female flower clusters when the bags were removed from the clusters for pollination. Grape pollen is relatively heavy and has limited ornamentation, which limits its dispersal by wind (di Collalto et al. 1982). However, whether by insects or wind, the results of this study demonstrate that grape pollen can be easily transported.
A successful cross depends on a number of factors: proper isolation of male and female parents; pollen quality; careful emasculation of hermaphroditic female parents; stigma receptivity; preventing pollen contamination of clusters, brushes, and forceps; and the dexterity of the person making the cross. It is not unusual for a small percentage of progeny in a cross with an emasculated hermaphroditic female parent to be off-types as a result of self-pollination. However, in a cross to a properly isolated pistillate female parent, off-types are rare.
One hundred and ninety-two seeds from an open-pollinated cluster of A. de Serres collected in fall 2005 were analyzed with 10 SSR markers (Table 3⇓). The neighboring staminate V. rupestris Pillans was the male parent for 154 of the open-pollinated seeds. Shoots from the Pillans vine were intertwined with the pistillate A. de Serres vine and the Pillans pollen had ample opportunity to outcompete other pollen sources. Yet, 38 seeds (~20%) from the population were found to be progeny of A. de Serres and five other male parents, from a row 11 m to the north of A. de Serres. Eleven of the seeds were from the rootstock AXR#1, on which Olmo had grafted many of Mexican Vitis collection that resided in rows 1 through 4 (Figure 1⇑). A few AXR#1 shoots were found in this section of the vineyard, but it was not possible to determine which specific root-stock sucker was involved in the pollination. Because these A. de Serres clusters were not covered, pollination could have been due to wind or insects.
The progeny from the crosses produced by Olmo have been under evaluation since 1991. They were considered to be V. rupestris x M. rotundifolia hybrids until SSR-based mapping of selections from this group commenced in 2004 (Krivanek et al. 2006, Riaz et al. 2006). About one-third of the A. de Serres progeny were V. rupestris A. de Serres x V. rupestris Pillans, and these plants are very similar to the few true V. rupestris x M. rotundifolia progeny. Both types of progeny have glabrous leaves and stems, leaf blades that fold closed on themselves, and broad teeth on the leaf margins. However, many of the “89 group” progeny had stems and leaves with sparsely scattered hairs, which was unexpected given the glabrous nature of the V. rupestris and M. rotundifolia parents. In addition, the “89 group” progeny were mostly fertile with 2n = 38 chromosome counts (data not presented). These morphological and fertility data cast doubt on the parentage of these selections, but their remarkable resistance to the root-knot nematode, Meloidogyne incognita (data not presented), the dagger nematode, Xiphinema index (Walker and Jin 1998), and Pierce’s disease (Krivanek and Walker 2005), supported M. rotundifolia as the correct parent given its strong and broad resistance to the above pests (Walker et al. 1994, Bouquet 1981, Mortensen et al. 1977). The discovery of V. arizonica and its relatives as the true parents of many of the “89 group” progeny revealed a new and exceptional source of pest resistance and a form of resistance that can be easily introgressed into other Vitis species, selections, cultivars, and rootstocks.
Conclusions
This study presents the results of an extensive parentage analysis that discovered the true parents of a series of crosses among V. rupestris and M. rotundifolia selections. These selections have been intensively tested for pest resistance, and their corrected parentage revealed surprising sources of nematode and Pierce’s disease resistance from Mexican forms of V. arizonica. Selections from these crosses have been used to breed PD-resistant wine, table, and raisin grapes, and to breed nematode-resistant rootstocks.
Footnotes
Acknowledgments: The authors gratefully acknowledge research funding from the California Grape Rootstock Improvement Commission, the California Department of Food and Agriculture’s Pierce’s Disease Board, and the Louis P. Martini Endowed Chair in Viticulture.
The breeding efforts of Prof. Harold Olmo in providing the materials for this study are gratefully acknowledged.
- Received March 2007.
- Revision received July 2007.
- Copyright © 2007 by the American Society for Enology and Viticulture