Abstract
Pierce’s disease (PD), caused by the bacterium Xylella fastidiosa, is an important disease of grapevines in California, across the southern United States, and into South America. In regions where disease pressure is high, the cultivation of Vitis vinifera cultivars is difficult or impossible. This study reports on the introgression of PD resistance into elite wine, table, and raisin grape genetic backgrounds and on the reliability of PCR-based marker-assisted selection (MAS) to accelerate the breeding of PD-resistant grapes. This work documents the introgression of PD resistance from a homozygous resistant form of V. arizonica, b43-17. A total of 4,321 seedlings from 83 different crosses of resistant selections and high-quality V. vinifera cultivars from the F1 and first and second modified backcross generations (mBC1 and mBC2) were screened with two to three flanking microsatellite markers (VVIP26, ctg1026876 and VMC2a5) in the early spring 2006 and 2007. The alleles linked to resistance were unique in size and were not shared by susceptible V. vinifera selections. Based on the presence of unique resistant alleles, 1,683 seedlings from wine, table, and raisin grape background were selected. The distinctiveness of these resistant alleles allows the use of MAS to optimize the breeding of PD-resistant grape cultivars.
Pierce’s disease (PD) has restricted viticulture in the New World since colonists first introduced European (Vitis vinifera L.) grape cultivars into the warmer parts of North America. This disease is caused by Xylella fastidiosa (Wells et al. 1987), a xylem-limited bacterium that aggregates in vessel elements and induces tyloses and gums, preventing water movement and leading to desiccation and eventual vine death. Settlers to areas prone to PD found that intentional and accidental hybrids between native grape species and V. vinifera were often resistant to PD. The mechanisms by which grape species native to regions where PD exists resist the disease are not clearly understood. However, this lack of knowledge did not prevent breeders from attempting to introgress PD resistance from resistant native Vitis species into high-quality V. vinifera wine, table, and raisin grapes. Hundreds of PD-resistant varieties have been created within muscadine and bunch grape backgrounds (Overcash et al. 1981, Mortensen 1988), but their fruit and wine quality has not matched that of V. vinifera cultivars.
A more thorough understanding of the inheritance of PD resistance would greatly aid breeders in their efforts to produce high-quality PD resistant varieties. Mortensen (1968) first examined the inheritance of PD resistance in a Florida field trial planted with seedling populations derived from the PD-resistant species V. aestivalis ssp. smalliana (L.H. Bailey) Comeaux, V. simpsonii Munson, and V. shuttleworthii House. He concluded that resistance to PD is dominant to susceptibility and controlled by complementary gene action among three independent genes. More recently, it was reported that northern Mexico forms of V. arizonica Engelm. had strong resistance to X. fastidiosa (Krivanek and Walker 2005). Related studies screened V. arizonica hybrid selections under greenhouse conditions and concluded that a major gene with a dominant allele was responsible for V. arizonica-based X. fastidiosa resistance (Krivanek et al. 2005a). Researchers were later able to genetically map a primary locus for PD resistance, PdR1, on linkage group (LG) 14 from F8909-17, a V. rupestris (A. de Serres) × V. arizonica/candicans hybrid (b43-17) (Krivanek et al. 2006). Recent work with F8909-08 (sibling of F8909-17) and the homozygous resistant male parent, b43-17, indicated that the siblings (F8909-08 and F8909-17) had different alleles of PdR1 from the homozygous resistant b43-17 (Riaz et al. 2008). It was necessary to distinguish the two PD-resistant lines (F8909-17 and F8909-08) as they might represent different alleles of the same or a different gene. Thus, the resistance locus from F8909-17 was named PdR1a and the locus from F8909-08 was named PdR1b (Riaz et al. 2008). The genetic maps of F8909-17 and F8909-08 allowed us to identify alleles of simple sequence repeat (SSR) markers that are tightly linked to X. fastidiosa resistance for use in marker-assisted selection (MAS) and molecular breeding.
In recent years, there have been many reports that associate a trait of interest in grapes to molecular markers: seedlessness, berry and seed weight and size (Lahogue at al. 1998, Doligez et al. 2002, Cabezas et al. 2006, Mejia et al. 2007), muscat flavor (Doligez et al. 2006), fruit yield components (Fanizza et al. 2005), fungal disease resistance (Fischer et al. 2004, Welter et al. 2007), and nematode resistance (Xu et al. 2008). However, only few reports are available where the markers have been used in active MAS breeding (Barker et al. 2005, Eibach et al. 2007). In other crops, molecular markers have been used to assist in the backcrossing of major genes into elite cultivars; to select alleles with major effects on high-value traits when marker information is available across multiple populations; and to pyramid different disease resistance genes or alleles into a single cultivar (Chen et al. 2000, Arru et al. 2003).
In perennial crops like grape, MAS is a very effective way of evaluating increased numbers of progeny just after seed germination, thus reducing the cost of time-consuming greenhouse screening and the maintenance of unwanted susceptible plants in the vineyard. Currently, the X. fastidiosa screening procedure takes six to eight months and includes the production and maintenance of four to five replicates of each seedling, inoculation with bacteria, symptom scoring, and ELISA testing to quantify bacterial populations (Krivanek et al. 2005b). Detection of resistant genotypes with molecular markers at the seedling stage greatly aid breeders by allowing backcrosses to be made to selected recurrent parents at least one year earlier.
This study reports on the introgression of the resistance locus PdR1a and PdR1b into elite V. vinifera selections. The main objective of this study was to validate the use of tightly linked SSR markers for MAS in multiple backcross populations derived from F8909-08 and F8909-17.
Materials and Methods
Plant material.
The source of X. fastidiosa resistance for PdR1 is b43-17, a form of V. arizonica that appears to have some V. candicans parentage as the result of natural hybridization (Riaz et al. 2007). Inheritance studies have found b43-17 to be homozygous resistant to X. fastidiosa (Riaz et al. 2008). It was collected by H.P. Olmo near Nuevo Leon, Monterrey, Mexico, and was crossed to the PD-susceptible V. rupestris A. de Serres, resulting in 13 resistant progeny (Krivanek et al. 2005a, Riaz et al. 2007). Two of the progeny, F8909-17 and F8909-08 (both staminate), were used as pollen parents in a number of crosses to V. vinifera wine, table, and raisin grapes to produce 50% V. vinifera progeny with PdR1a and PdR1b alleles, respectively. Selections from these populations were then crossed back to additional V. vinifera cultivars in a modified back-cross (mBC) strategy to produce 75% V. vinifera progeny. Selections from this mBC1 generation were also crossed back to V. vinifera cultivars to produce a mBC2 generation of 87.5% V. vinifera. Details of resistant and susceptible selections used in 83 different crosses are shown (Table 1⇓). Seven of the crosses contained PdR1a allele and 76 crosses carried PdR1b allele for PD resistance.
Resistant (R) and susceptible (S) progeny from 83 crosses to introgress the PdR1a and PdR1b locus into rootstock, wine, table, and raisin grape varieties. The first five resistant parents carry the PdR1a resistance locus and all others carry the PdR1b resistance locus. All plants were screened with two SSR markers. Only progeny that carry resistant alleles with both markers were planted in the field. Recombinant seedlings and plants without greenhouse screen data are not included.
All of the resistant and susceptible parental selections and breeding populations were maintained in the vineyards of the University of California (UC), Davis (Table 1⇑). All breeding selections from the F1 generation were planted in the field and tested with the greenhouse assay and with tightly linked markers to verify the linkage phase of the markers to the resistance locus. Progeny in the mBC1 and mBC2 generations were screened with tightly linked SSR markers to identify the resistant genotypes. Only PdR1a and PdR1b containing resistant progeny were planted in the field. Resistant selections with suitable morphological and horticultural traits to justify use as parents in later mBC generations were also evaluated with the greenhouse assay system to validate the marker-based screen.
Germination of seeds and plant maintenance.
Crosses were made in the spring, berries were harvested in late September, and seeds were removed from berries and stored at 1°C. A seed stratification/conditioning method modified from a published method (Ellis et al. 1983) was used before germination. Seeds were treated with a solution of 0.5 M hydrogen peroxide for 24 hr, rinsed three times, soaked in 350 ppm gibberellic acid (GA3) for another 24 hr, and then cold stratified at 1°C for up to 21 days. Seeds were planted in seedling trays and seedlings were transferred to 125 cm3 plastic pots about 5 weeks later and were maintained in the greenhouse. After PdR1b marker analysis, selected resistant seedlings were planted in the field in early May, protected with plastic grow tubes, and irrigated and fertilized as necessary to promote vigorous growth. Lateral shoots were removed on a regular basis to encourage the rapid development of a single strong apical shoot. When actively growing plants reached 1.4 m they were cut back to the 1-m high trellis wire and three shoots were allowed to grow from the upper lateral buds. In the dormant season the plants were pruned to one 6-node cane and two 1-bud spurs to encourage early flowering. Potential parents for future mBC generations were selected on the basis of the following criteria: presence of PdR1 flanking markers, vigor, V. vinifera-like leaf morphology and growth habit, productivity (number and size of fruit clusters), berry size, seed germination rates, and ease of rooting.
DNA isolation.
Young leaves were obtained from greenhouse-grown seedlings when they had four to five leaves, and from field-grown plants when leaves were ~2-cm diam. DNA was extracted with a modified CTAB (hexadecyltrimethylammonium bromide) procedure (Lodhi et al. 1994).
Marker analysis of PD breeding material.
All F1 genotypes were screened with four SSR markers linked to the PdR1a and PdR1b locus: VMC6e1 (GenBank accession BV722733.1), VVIP26 (GenBank accession BV140645.1), VVIS70 (GenBank accession BV140769.1), and VMC2a5 (GenBank accession BV681663.1). The mBC1 and mBC2 breeding populations were screened with two markers (VVIP26 and ctg1026876), one on each side of the resistance locus spanning a genetic distance of 19cM (Riaz et al. 2006, 2008). The SSR primer sequences used have been reported in earlier studies: VVI series (Merdinoglu et al. 2005); VMC and VVI primer sequences (NCBI database uni-STS, www.ncbi.nlm.nih.gov/); and ctg primer sequence (EST-SSR database, UC Davis, http://cgf.ucdavis.edu/). DNA samples were amplified and run on denaturing polyacrylamide gels for the F1 populations according to a previous protocol (Riaz et al. 2004). The same protocol was followed for the mBC1 and mBC2 populations, with the exception that labeled forward primers were used and amplification reactions were cut down to a total volume of 10 μL. The PCR was carried out using a PTC-100 system (MJ Research, Waltham, MA), and the cycling program was performed as described previously (Riaz et al. 2004). Samples that did not amplify in the first run (because of poor DNA extraction or unsuccessful amplification) were only repeated again if the overall sample amplification failure rate for any particular cross was higher than 5%. The 5% cutoff level was established to accelerate the marker screening process. After successful amplification, labeled products of both markers were mixed and loaded into an ABI 377 DNA sequencer (PE/Applied Biosystems, Foster City, CA), using ROX-500 as an internal size standard. The PCR fragments were detected with GeneScan analysis software (version 3.1.2 and alleles were scored using the Genotyper DNA fragment analysis software (version 2.5.2) (PE/Applied Biosystems).
Disease evaluation.
A greenhouse-based screening technique was used to evaluate progeny and parents for PD resistance (Krivanek et al. 2005b). The X. fastidiosa inoculum was obtained from infected grapevines in the Stag’s Leap area of Napa Valley, California, and maintained in greenhouse-grown Chardonnay plants. To prepare for inoculation, actively growing bacteria were washed from petri plates with ddH2O, and the cell suspension was standardized to an absorbance of 0.25 at 600 nm (~6 × 108 colony forming units [cfu]/mL as determined by culture plating). Plants were needle-inoculated (Hopkins 1980) below the node within 5 to 10 cm from the base of each shoot with 10 μL bacterial suspension. Plants were reinoculated three days later above that node to ensure successful inoculation.
Plants were sampled 12 weeks postinoculation by taking 0.5 g sections of stem tissue from 30 cm above the point of inoculation. Samples were ground and tested by ELISA following published procedures (Krivanek and Walker 2005).
Results
MAS with selected PdR1a and PdR1b-linked markers.
Five markers linked to PdR1a and PdR1b, which mapped in the 9621 and 04190 populations, were selected for MAS (Riaz et al. 2008). Based on the analysis of the F1 populations and greenhouse ELISA screen data, two flanking markers, VVIP26 and VMC2a5, were used to screen additional mBC1 populations in the PdR1a background (Table 1⇑). The markers VVIP26 and ctg1026878 were selected to screen additional mBC1 and mBC2 populations within the PdR1b resistance background. All three markers produced clear, well-separated amplification products that were easily scored on the ABI 377 sequencer. The marker VVIP26 was very polymorphic with an observed heterozygosity of 0.82. Fourteen alleles were observed among 50 susceptible vinifera selections (data not shown). The allele in coupling to the PdR1a and PdR1b locus was 146 bp in size and susceptible selections did not carry it (Figure 1A⇓, Table 2⇓). The marker ctg1026878 was an EST-derived SSR marker and produced only three alleles (122, 126, and 132 bp). The resistant allele in coupling with the PdR1b locus was unique (126 bp) and present only in resistant selections. This marker could not be used with the PdR1a background as the resistant allele 122 was similar in size to the susceptible allele because of homoplasy. However, the presence of a unique allele with marker VMC2a5 resolved the analysis within the PdR1a background. The resistant allele with marker VMC2a5 was 168 bp in size and was unique to resistant genotypes and not present in six susceptible vinifera selections used for mBC1 crosses with F8909-17 background (Table 1⇑).
Allele size information for elite Vitis vinifera cultivars and table grape selections using two flanking makers in MAS.
Gel images of PCR products amplified with SSR markers VVIP26 and ctg1026876. (A) Resistant selections (with 146 bp allele) and select susceptible V. vinifera parents used in F1, mBC1, and mBC2 crosses. (B) 045554, mBC2 family at the 87.5% V. vinifera level. Arrows note resistant alleles. Progeny labeled -19, -42, and -51 are ELISA-resistant recombinant genotypes with resistant alleles for marker VVIP26, but they lack the resistant allele for marker ctg1026876; the recombination event occurred between PdR1b and marker ctg1026876. The recombination event for progeny -13 occurred between PdR1b and marker VVIP26.
A total of 4,321 plants were screened with markers in early spring 2006 and 2007. The presence of unique alleles with flanking markers helped to greatly expedite the genotypic analysis. The PdR1a and PdR1b containing individuals that were selected as parents were also evaluated with the greenhouse-based ELISA technique. In every case, when the plants had the resistant alleles with both flanking markers, they had suppressed bacterial populations in their stems (Table 3⇓, Table 4⇓). Marker data was usually more reliable than the greenhouse screen data. For example, there were a few rare examples where an individual was designated as resistant based on ELISA results, but the resistant marker alleles were absent. When the greenhouse-based screen was repeated, these plants tested as susceptible.
ELISA-transformed bacterial levels and SSR marker alleles of the inoculated resistant and select susceptible (V. vinifera) parents. All tested plants have a minimum of four replicates. Only the resistant plants have SSR alleles linked to resistance.
Bacterial levels of resistant (R) and susceptible (S) selections with 50%, 75%, and 87.5% V. vinifera in their parentage. All resistant selections have SSR alleles that are linked to the PdR1b locus.
The 1,683 seedlings that had the resistance allele at both markers were planted in the field; 1,863 seedlings were discarded because they did not have resistance alleles with both markers. Data could not be obtained for ~5% of the seedlings (data not shown). Recombinant seedling plants (resistant based on the greenhouse screen, but possessing only one of the two markers) were also preserved to aid in the fine-scale mapping of PD-resistance genes for both backgrounds.
Discussion
Markers have been widely used in plant breeding to assist in introgression of major genes into elite cultivars in many other important crops (Chen et al. 2000, Singh et al. 2001, Shi et al. 2008). In grapes also, there are many examples of marker association to trait of interest (Doligez et al. 2002, 2006, Fischer et al. 2004, Fanizza et al. 2005, Welter et al. 2007, Xu et al. 2008); however, very few examples of the use of markers for active breeding are reported (Barker et al. 2005, Eibach et al. 2007). Marker-assisted selection is most effective when the marker allele for a favorable trait is consistent across breeding populations. Thus, marker alleles tightly linked to a resistance trait in the original F1 mapping population should also be linked to resistance in later generation populations derived from the same resistance source (Cregan et al. 1999). The allele sizes for the flanking markers VVIP26 and VMC2a5, linked to PdR1a, and for the flanking markers VVIP26 and ctg1026878, linked to PdR1b, were unique and not shared by the susceptible V. vinifera selections being used for breeding wine, table, and raisin grape cultivars (Figure 1A⇑, Table 2⇑). All of the resistant selections from the F1, mBC1, and mBC2 progeny had the resistant allele from both markers and had very low bacterial populations in stem tissue (Table 3⇑, Figure 1B⇑). The uniqueness of alleles is a key feature because it allowed these markers to be used as effective diagnostic tools for the identification of target alleles in crosses among many diverse parents (Figure 1⇑, Table 2⇑, Table 3⇑). The effective use of MAS in breeding programs requires the use of markers that flank both sides of the resistance locus to clearly distinguish recombinants. The use of markers closer to the target gene is another approach to minimize losses through recombination. In the case of the PD-resistance locus, the markers used for MAS are tightly linked to the resistance gene; however, the distance between the locus and flanking markers is large enough for recombination to occur. A recombination frequency of ~10% has been observed in all tested plants (data not shown). To make MAS most effective and to minimize further cost, the need to develop markers that are in complete linkage to the PD-resistance gene is necessary. The process of developing markers that are closer to the resistance locus to minimize the number of recombinants is underway.
The other important factor for successful use of DNA-based MAS is whether the process is economically justified, which depends on the relative cost and ease of genotype versus phenotype screening, the time saved through the use of MAS; and the benefits associated with accelerated release of resistant accessions. The addition of MAS to the PD-resistance breeding program has allowed rapid selection and advancement of superior resistant parents, thus saving time and money. In the early years of the breeding program, selection of resistant plants was solely based on the greenhouse/ELISA method and three or more years were needed to identify resistant parents. In the first year, plants were grown from seed, transplanted to the field, and trained to promote rapid upright growth, which encourages cluster initiation and development. In the second year, cuttings were taken to produce replicates from each seedling for the 6- to 8-month-long greenhouse screening for X. fastidiosa resistance. In some cases, the greenhouse screening had to be repeated to verify the resistance status of an individual, which took an additional 6 to 8 months. Under optimal conditions, a new resistant donor parent with desired horticultural characteristics could be selected in three years, although it was often longer, and then backcrossed to elite V. vinifera cultivars. With the addition of MAS, the time frame for the selection of resistant genotypes has been reduced by at least one year. In addition, about half as many plants are planted and maintained in the field, and the cost of growing and maintaining numerous plants in the greenhouse for screening has also been reduced. Moreover, MAS for PD resistance proved to be more reliable than the greenhouse/ELISA screening method.
There are many ways in which the greenhouse/ELISA screening method could result in erroneous decisions regarding the resistance status of a test plant. It is a biological assay that involves many steps, each one prone to potential human error-based mistakes, from plant propagation, to inoculation, to ELISA technique errors. In addition, the age or morphology of a plant’s vascular system also has major influence on whether it can prevent X. fastidiosa development and movement (unpublished data). On the other hand, MAS has the potential to be a much more robust screening technique and is capable of establishing the resistance status of a plant at a very young stage. The use of MAS proved to be a more consistent and dependable technique of screening for PD-resistance breeding. In addition, given increased resources for PdR1 screening, seed population sizes could be doubled, allowing the production of more plants with X. fastidiosa resistance and improved viticultural characteristics. The use of markers to select putatively resistant lines with optimal horticultural characteristics followed by phenotypic evaluation of resistance allows the breeder to maintain the linkage phase in selected progeny. Thus, MAS can be self-reinforcing and ensure that the same set of markers will be effective in future crosses (Cregan et al. 1999).
Conclusions
Because grape cultivars are clonally propagated, resistance is fixed after new resistant genotypes with optimal horticultural characteristics are established. The PD-resistance breeding program is now evaluating the fruit quality of PdR1a and PdR1b containing wine, table, and raisin grape populations with up to 87.5% V. vinifera parentage, and 93.75% V. vinifera containing progeny with PdR1b were being tested in 2008. The next steps are to develop reliable markers for a range of viticultural traits. Markers for some of these traits are easily measured and can be selected in one growing season. However, other traits such as wine quality will be more difficult because of the need for wine production, aging, and sensory analysis. Preliminary testing of reliability and accuracy of small-scale winemaking techniques is underway and could lead to marker development in a few years. If these tests can identify which quality parameters are useful and comparable at micro- and commercial-scale winemaking levels, the breeding of PD-resistant winegrapes will be accelerated to an even greater extent.
Footnotes
-
Acknowledgments: Research funding from the California Department of Food and Agriculture’s Pierce’s Disease Board and the Louis P. Martini Endowed Chair funds is gratefully acknowledged.
-
The authors are grateful for the technical assistance of Rong Hu, Kurt Kabica, Geoff Dervishian, and Dan Ng.
- Received May 2008.
- Revision received December 2008.
- Accepted February 2009.
- Copyright © 2009 by the American Society for Enology and Viticulture
Literature Cited
Vol 60 Issue 2
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
Advertisement
