Introduction

Plant parasitic nematodes are among the most damaging agricultural pests. Ectoparasitic nematodes live and reproduce in soil and feed by puncturing the root surface with a stylet; in contrast, endoparasitic nematodes physically enter the roots to feed and reproduce. Economically, the most important are endoparasitic sedentary nematodes of the genera Meloidogyne, Heterodera and Globodera. Examples of plant genes that confer resistance to these three genera have been reported. For example, Hs1 pro−1 from sugar beet confers resistance to the beet cyst nematode H. schachtii (Cai et al. 1997). The tomato Mi-1.2 gene mediates resistance to three different root-knot nematode (Meloidogyne) species and to the potato aphid Macrosiphum euphorbiae (Milligan et al. 1998; Rossi et al. 1998). The Hero A gene from tomato confers resistance to various pathotypes of potato cyst nematodes G. rostochiensis and G. pallida (Ernst et al. 2002). Both Gpa2 (van der Vossen et al. 2000) and Gro1-4 (Paal et al. 2004) from potato confer resistance to G. pallida and G. rostochiensis, respectively. These genes have been cloned and characterized and are being used by breeders in their efforts to produce resistant plant material (Williamson and Kumar 2006). However, no plant genes that confer resistance to ectoparasitic nematodes have yet been physically mapped or identified due to the transient association between a mobile nematode and the site of feeding, which makes phenotypes more difficult to evaluate.

The grapevine fanleaf degeneration disease complex can reduce crop yields by 80% (Martelli and Savino 1990) and is a major worldwide viticultural problem (Andret-Link et al. 2004). Disease symptoms consist of disrupted fruit set caused by grapevine fanleaf virus (GFLV) and depressed plant growth caused by severe root damage from GFLV’s vector, the ectoparasitic nematode Xiphinema index (Thorne and Allen 1950). Xiphinema index feeds primarily on actively growing root tips and causes the root tip to swell and gall followed by decay and proliferation of lateral roots behind the root tip (Martelli and Savino 1990). GFLV is a nepovirus of the family Comoviridae (Wellink et al., 2000), and possesses a bipartite genome. The gene responsible for encoding the coat protein, 2CCP, resides on RNA 2 and has been found to dictate the specific transmission of GFLV by X. index (Andret-Link et al. 2004). The virus resides in the anterior region of the nematode stylet at the odontostyle–odontophore junction (Raski et al. 1973), and can be successfully transmitted after approximately 5 min of feeding (Alfaro and Goheen 1974).

Because of the very high value of vineyard land, fallow and crop rotation are not practiced as a means of controlling soil-borne pests and diseases. In addition, grape roots are known to survive and support nematodes for at least 7 years (Villate et al. 2008). The use of fumigants to control X. index in vineyards is no longer recommended because of their limited soil penetration, their high cost, and their detrimental environmental effects (Raski and Goheen 1988). Thus, the development of resistant rootstocks using Vitis sources of resistance has been a breeding goal for many years. Resistance to X. index exists in several Vitis species including V. arizonica, V. candicans, V. rufotomentosa, V. smalliana and V. solonis (Kunde et al. 1968). Genetic studies leading to the isolation and characterization of genes conferring resistance to X. index would further our understanding of resistance and assist molecular and classical breeding efforts to control X. index.

Grape (2n = 2x = 38) has a relatively small genome (C = 483 Mbp for V. vinifera) and an average genetic mapping unit (cM) corresponding to about 288 kb, given a grape genetic map of 1,676 cM (Di Gaspero et al. 2007), and is comparable to that of rice (~280 kb/cM) (International Rice Genome Sequencing Project, 2005). In addition, whole-genome sequences of V. vinifera are now publicly available in GenBank (http://www.ncbi.nlm.nih.gov/). These qualities and genome resources make grape a suitable woody fruit crop for molecular isolation of genes using map-based cloning strategies. In a previous report, a fast and reliable X. index resistance screening system based on gall numbers was developed and used to demonstrate that host resistance to X. index derived from V. arizonica is primarily controlled by a major quantitative trait locus (QTL) designated XiR1 on chromosome 19 (Xu et al. 2008). Combining these results with the recently available resources in grape genomics suggested that map-based cloning could lead to the identification and characterization of X. index resistance at the XiR1 locus. In an effort to further this goal, the research presented here details the development of a high-resolution genetic map around XiR1, construction of BAC contigs encompassing the XiR1 locus and sequence analysis of three BAC clones from the XiR1 region containing nucleotide binding/leucine-rich repeat (NB-LRR) genes XiR1.1 and XiR1.2, which are strong candidates for XiR1.

Materials and methods

Plant materials and screening for resistance to X. index

Three mapping populations with a total of 1,375 F1 genotypes were used to construct a high-resolution genetic map of the XiR1 region. The first population, 9621, was derived from a cross of D8909-15 × F8909-17, in which the XiR1 locus was initially identified (Xu et al. 2008). The 9621 population was expanded to 943 F1 individuals for this study. The second was the 0023 population consisting of 178 F1 individuals from a cross of D8909-15 × V. vinifera B90-116. The third population, 05384, consisted of 253 F1 individuals and derived from a cross of D8909-15 × V. vinifera Airen. The common female parent D8909-15 was the source of resistance to X. index, while all three male parents F8909-17, V. vinifera B90-116 and V. vinifera Airen were susceptible. D8909-15 is a selection derived from a cross of V. rupestris A. de Serres × V. arizonica b42-26. The maternal grandparent V. rupestris A. de Serres is susceptible to X. index while the paternal grandparent V. arizonica b42-26 is highly resistant. To distinguish the two haploid genomes as well as the genome origin for a specific homologous chromosome of D8909-15, the genomes and individual homologous chromosomes of b42-26 origin and those of A. de Serres origin were designated D8909-15-R and D8909-15-S genome/chromosome, or simply “R” and “S” genome/chromosome, respectively.

To minimize the number of plants required for the screening of X. index resistance, the three populations were analyzed with two previously identified markers M4F3F and VMCNg3a10 (Xu et al. 2008) that flank the XiR1 locus so that informative recombinant genotypes could be identified. Four plants of each of the resulting 99 recombinants along with the parents and the highly susceptible control V. rupestris St. George were propagated from herbaceous two-bud cuttings. The rooted plants were then evaluated for resistance to X. index based on the number of galls formed in 4–8 weeks after inoculation with 100 nematodes, as described previously (Xu et al. 2008).

BAC libraries

Three BAC libraries were used to identify BAC contigs encompassing the XiR1 locus. The first BAC library was constructed by Yimin Jin (M. A. Walker, unpublished data) from D8909-15 using a vector/enzyme combination of pECSBAC4/EcoRI. The D8909-15 BAC library has 33,792 clones with an average insert size of approximately 50 kb. The coverage of this library was estimated at 3.4× genome equivalents. A total of 88 DNA pools with one for each of the 384-well plates were prepared for the D8909-15 library, enabling PCR-based screening for clones of interest. The second BAC library was constructed from V. vinifera Cabernet Sauvignon by the E&J Gallo Company (Modesto, CA, USA) using two vector/enzyme combinations: pECBAC1/BamH I and pECBAC1/EcoR I and is now available from the laboratory of M. A. Walker at the University of California, Davis. This BAC library consists of 56,448 clones with a combined coverage of 13.9× genome equivalents. The number of clones, average insert size and genome equivalents are 27,648, 151 kb and 7.5× for the BamHI sub-library, and 28,800, 125 kb and 6.4× for the EcoRI sub-library, respectively. The third BAC library was constructed from V. arizonica b42-26 using a vector/enzyme combination of pCC1/HindIII (Amplicon Express, Pullman, WA, USA). The V. arizonica b42-26 library consists of 36,864 clones with a coverage of 6.1X genome equivalent with an average insert size of 140 kb. The Cabernet Sauvignon library and the V. arizonica b42-26 library were gridded onto a set of three and two 22.5 × 22.5 cm2 Hybond N+ filters, respectively. These filters were screened by hybridization using ECL Direct Nucleic Acid Labeling and Detection System (Amersham, GE Healthcare, Buckinghamshire, UK).

BAC sequencing and marker development

Markers were developed based on combined information from BAC end sequences and the polymorphisms identified between the two homologous chromosomes of D8909-15 as well as between the parents from the three populations. BAC DNA was prepared from 500 ml LB overnight culture using a Qiagen Large Construct Isolation Kit (Valencia, CA, USA) and BAC end sequences were obtained by direct sequencing using BigDye terminator V3.1 chemistry (Applied Biosystems, Foster City, CA, USA). Primers were designed for each sequenced BAC end using Vector NTI program (Invitrogen, Carlsbad, CA, USA), and then used to: (1) directly screen the D8909-15 BAC library; (2) isolate the corresponding BAC ends for hybridization-based screening of the Cabernet Sauvignon BAC library; and (3) amplify, clone and sequence the corresponding genomic fragments from all the parents for marker development by identifying polymorphisms between the parental genomes. To estimate the size of inserts, BAC DNA was digested with restriction enzyme NotI (New England Biolabs, Ipswich, MA, USA) and analyzed with a Chef gel DRIII system (Bio-Rad Laboratories, Hercules, CA, USA) run for 16 h at 14°C with 6 V/cm and pulse 5–40 s. BAC clones S29J16 and M10J21 were sequenced using a shotgun strategy by Macrogen (Seoul, Korea) and Amplicon Express (Pullman, WA, USA), respectively. VA31P21 was sequenced by Macrogen (Seoul, Korea) using 454 technology. The DNA and putative protein sequences of XiR1.1 and XiR1.2 are available as GeneBank accession numbers GU471231 and GU471232, respectively.

Nested RT-PCR

Total RNA was isolated from 5 g of D8909-15 leaf and 10 g of root tissue using a cetyltrimethylammonium bromide (CTAB)-based RNA extraction protocol as described previously (Iandolino et al. 2004), with minor modifications. Poly (A)+ RNA was isolated using the Dynabeads mRNA purification kit (Dynal A.S, Oslo, Norway). First-strand cDNA synthesis with SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA, USA) was performed as described by the manual. A 2-μl aliquot of leaf and root first-strand cDNAs was used for the first-round PCR reactions, and 2 μl of the first-round PCR products were used as templates for the nested PCR reactions. The amplification of both PCR reactions were performed for 30 cycles (1 min at 94°C, 2 min at 55°C, 3 min at 72°C) followed by 7 min at 72°C. PCR products were separated on 1.2% agarose gels and visualized with ethidium bromide. Primers for nested RT-PCR had the following sequences: 1: 5′-GGCTGACCAAATTCCATTCAGTGTTGTGGAG-3′; 2: 5′-GCTGGATGACTATGCAACCCATTATCT-3′; 3: 5′-AGGATTTTCGTTCCAAACATCATCAAG-3′; 4: 5′-CAACTATGGATAGCACAAGGTTATATTCAA-3′; 5: 5′-ATTTCTCCCCTGGATACCAGTTCAACATCCCTCACGTT-3′; 6: 5′-AGTAAGTCCATCCTCCACAATTCCTTCAACTTTGGCAT-3′.

Annotation of BACs and sequence analyses

The editing and compiling of DNA sequence data as well as comparisons and analysis of DNA and deduced amino acid sequences were made using Vector NTI program (Invitrogen, Carlsbad, CA, USA). DNA and protein sequences of predicted genes were analyzed by NCBI databases Blast (http://www.ncbi.nlm.nih.gov/) and GeneMark.hmm (http://exon.biology.gatech.edu/), the latter using both Arabidopsis thaliana and Medicago truncatula settings. The CurLRR program at http://www.bioinf.manchester.ac.uk/curlrr/index.html, the COILS server at http://www.ch.embnet.org/software/COILS_form.html and the Motif Scan program at http://hits.isb-sib.ch/cgi-bin/motif_scan were used for the prediction of protein functional motifs.

Results

Marker development and high-resolution mapping

Xu et al. (2008) found that the SSR marker VMC5a10 was tightly linked to the XiR1 locus on chromosome 19 (Fig. 1a) in the 9621 population, which then consisted of 188 progeny. To identify additional markers linked to the XiR1 locus, the D8909-15 BAC library was screened by PCR with VMC5a10. Markers developed from the ends of the identified BAC clones were then used to screen the D8909-15 and V. vinifera Cabernet Sauvignon BAC libraries. As a result, a total of seven markers (Table 1; Fig. 1b) were developed from the end sequences of 21 BAC clones (Fig. 1c, d) identified from these two BAC libraries. Because grape is a highly heterozygous crop and the two BAC libraries were constructed from two different sources of genomic DNA, it was expected that the D8909-15 BAC clones could be from either the D8909-15-R genome (one of the V. arizonica b42-26 genomes) or the D8909-15-S genome (one of the V. rupestris A. de Serres genomes). The Cabernet Sauvignon BACs are similarly from its two different haploid V. vinifera genomes. Detailed analysis based on specific alleles of markers suggested that seven D8909-15 BAC clones were derived from the D8909-15-R genome carrying the resistant allele of XiR1 (Fig. 1c). These seven BACs covered about 500 kb with three gaps in the XiR1 region of the “R” genome. Five BACs were derived from the D8909-15-S genome (Fig. 1c), and one M10J21 covered an estimated 50 kb gap between BAC clones S8O16 and S29J16. Among the nine BACs from the Cabernet Sauvignon library, four clones (23P24, 9I17, 1N2 and 34E12) were considered to originate from the first haploid genome origin of Cabernet Sauvignon, whereas the other five (5K4, 22H13, 2L17, 8N8 and 37P14) were from the second genome (Fig. 1d).

Fig. 1
figure 1

A high-resolution genetic and physical map of the XiR1 region on chromosome 19. a Genetic linkage map showing the relative position of XiR1 with SSR markers on chromosome 19. The numbers along the chromosome are genetic distance in centimorgans. b The horizontal bar represents a segment of the D8909-15-R genome, with BAC clones spanning the XiR1 region and positions of newly developed markers from this study. c The names, positions and sizes of twelve BAC clones derived from the D-8909 library. Seven of them are from the resistant background with three gaps, and five are from the susceptible genome. M10J21, which is from the susceptible genome that covers one of the gaps, is also shown. d The names, positions and sizes of nine BAC clones derived from the Cabernet Sauvignon library in two different genomes

Full size image
Table 1 Primer information and means of detection for DNA markers developed in the XiR1 region
Full size table

Screening for informative recombinants and evaluation of resistance to X. index

XiR1 was previously mapped within a 5.6 cM interval between two markers, VMCNg3a10 and M4F3F, on chromosome 19 (Fig. 1a; Xu et al. 2008). To fine-scale map the XiR1 region, these two flanking markers were used to identify informative recombinant genotypes from the three mapping populations 9621, 0023 and 05384. Genotypic data showed that the two markers segregated normally with an expected 1:1 ratio (χ2 = 1.34, P = 0.10–0.25 for VNG3a10, and χ2 = 0.16, P = 0.50–0.75 for M4F3F) in the three combined populations. There were 99 recombinants identified within this interval from the 1,375 individuals analyzed (Table 2). The overall percentage of recombined D8909-15 gametes in the three mapping population was 7.2%, and ranged from 5.6 to 7.9%. The two types of recombined D8909-15 gametes, type A (where a D8909-15-S allele at marker Vng3a10 recombined with a D8909-15-R allele at marker M4F3F) and type B (where a D8909-15-R allele at marker Vng3a10 recombined with a D8909-15-S allele at marker M4F3F), were significantly distorted (χ2 = 8.49, P < 0.005) from the expected ratio of 1:1 (Table 2). However, the overall distorted 1:1 ratio was due primarily to the 9621 population (χ2 = 11.09, P < 0.005) as the expected 1:1 ratio was observed in the 0023 (χ2 = 1.60, P = 0.10–0.25) and 05384 (χ2 = 0.00, P > 0.95) populations. These results suggested that male parents had significant effects on the ratio of the two recombinant types in their F1 progeny since the female parent D8909-15 was the same in the three populations.

Table 2 Screening for informative recombinants in the XiR1 region
Full size table

These 99 recombinants were further evaluated for X. index resistance based on the number of galls observed on their root systems, resulting in 61 resistant and 38 susceptible genotypes. In addition, all 99 recombinant plants were screened with BAC-derived genetic markers (Table 1), which allowed the identification of ten key recombinants (Fig. 2). These data permitted the positioning of the XiR1 locus within an interval of 0.51 cM (7 recombinants) between the two closest flanking markers 1N2R3b and M4F3R in the D8909-15-R genome (Figs. 1b, 2).

Fig. 2
figure 2

Schematic representation of ten key recombinant genotypes surrounding the XiR1 locus with nematode assay data. The solid bars represent chromosomal segments from D8909-15 that originated from the X. index resistant V. arizonica b42-26 (“R” genome) and the open bars represent segments from the susceptible V. rupestris A. de Serres (“S” genome). The crossover break-points are shown by the junctions between the solid and open bars

Full size image

Cloning and analysis of candidate XiR1 sequences

Sequence analysis of two overlapping BAC clones, one from the resistant source (S29J16, ~52 kb) and one from the susceptible parent (M10J21, ~75 kb) was completed. Sequencing data from BACs S29J16 and M10J21 revealed that both had a putative LRR region at the overlapping region, and an additional NB-LRR gene was identified in M10J21 only. There are four additional predicted genes in BAC S29J16 including a putative d-galactoside/l-rhamnose binding lectin, a reverse transcriptase and two hypothetical proteins which are unlikely to be the XiR1 candidates. These results suggest that the XiR1 locus in the D8909-15-S genome has been sequenced; however, a BAC clone spanning the corresponding region of the D8909-15-R was not identified in the D8909-15 BAC library, and the putative XiR1 genes most likely reside in the 50-kb gap between clones of S8O16 and S29J16.

Primers corresponding to the overlapping LRR regions in BAC clones of S29J16 and M10J21 were designed and used to re-screen the D8909-15 BAC library. Three new BAC clones were obtained; however, none of them covered the gap after detailed marker analysis (data not shown). PCR amplification using primers generated from the BAC ends of S29J16 and S8O16 was unsuccessful at “filling-in” the missing sequence data. These results prompted the construction of a new BAC library from V. arizonica b42-26, the homozygous X. index resistance source (Y. Jin and M. A. Walker, unpublished data). Eleven clones spanning the gap were isolated by screening this library with the end sequence from BAC clone S29J16. Database homology searches using sequence data from one of the shortest clones, VA31P21, which completely overlapped the BAC clone M10J21, revealed two tandem repeats of NB-LRR genes (XiR1.1 and XiR1.2). The deduced open reading frame was 1,268 amino acids for XiR1.1, 1,272 amino acids for XiR1.2 and 1,250 amino acids for the XiR1 susceptible homologue (XiR1sh) found in M10J21 (Fig. 3). The predicted polypeptides have 62.1% amino acid sequence identity. The major differences between these three predicted polypeptides are in the LRR domains. There are 12 LRRs in XiR1.1, 14 LRRs in XiR1.2 and 14 LRRs in XiR1sh (predicted by the CurLRR program). No apparent N-terminal signal sequence was found, indicating that the encoded proteins may be located in the cytoplasm. Because no Toll interleukin receptor (TIR) domain was found, the XiR1 genes are considered to belong to the non-TIR subfamily of R genes. XiR1.1, XiR1.2 and XiR1sh each contain a leucine zipper motif and an NB domain including the kinase-1a (P-loop), kinase-2 and kinase-3a motifs predicted by Motif Scan program (Fig. 3). Additionally, both VA31P21 and M10J21 contain two retrotransposon genes (GenBank accession #: ABA94354 and ABA98656), a putative intregase (GenBank accession #: ABH08430) and an ATP-dependent transporter (GenBank accession #: XP_001779695). Because plant NB-LRR genes are known for conferring resistance to diseases and pests, XiR1 and XiR1.2 are likely candidates for the XiR1 locus conferring resistance to X. index.

Fig. 3
figure 3

Alignment of the deduced amino acid sequences of XiR1.1, XiR1.2 and XiR1sh. Solid background indicates identity and conservative substitutions. Putative leucine zipper, kinase-1a, kinase-2 and kinase-3a motifs are underlined. The boundaries of LRR domains are indicated. Sequence gaps inserted to maintain alignment are shown as dashes. The GeneDoc program (http://www.psc.edu/biomed/genedoc/) was used for formatting the alignment

Full size image

XiR1.1, XiR1.2 and XiR1sh are transcribed

The GeneMark.hmm program predicted that XiR1.1 contains five exons and four introns, and that XiR1sh contains three exons and two introns; however, XiR1.2 was predicted to contain a single exon with no introns (Fig. 4a). To determine whether the XiR1 candidate genes were expressed, poly(A)+ RNA was isolated from both leaf and root tissues of D8909-15. Genomic DNA contamination was tested by using these RNA samples as templates with primers 1/3 (Fig. 4a) for PCR amplification and no bands were observed (data not shown). The poly(A)+ RNA samples were then subjected to nested RT-PCR analysis. The amplified fragments were designed to include both N-terminus (NT) and LRR regions (Fig. 4a). PCR using primer sets (1/3 and 4/6) yielded an NT-840 bp band and an LRR-1,150 bp band, respectively. These PCR products were further used as templates for PCR using primer sets (2/3 and 4/5) to generate a 600-bp band from the NT region and an 800 bp from the LRR region (Fig. 4b). The populations of NT-nested PCR products (600 bp) derived from both leaf and root tissue mRNAs were subcloned into the vector pGEM-T Easy (Promega, Madison, WI, USA) and transformed into E. coli. The nucleotide sequence analysis of 15 colonies from each of the resulting transformations confirmed the expression of cDNA sequences corresponding to XiR1.1, XiR1.2 and XiR1sh from both leaf and root tissues of D8909-15.

Fig. 4
figure 4

Prediction of intron–exon organization and expression of XiR1 homologues in both root and leaf tissues of D8909-15. a Schematic overview of the putative XiR1 gene. The positions of PCR primers (1–6) are indicated by arrows. Exons containing coding regions are indicated as black boxes. Introns are shown as black lines. Numbers given below indicate the sizes of exons and introns in base pairs. Thin lines before ATG are the 5′ untranslated regions. b Agarose gel electrophoretic analysis of PCR products, obtained using primer pairs (1 + 3 and 4 + 6) and (2 + 3 and 4 + 5). The cDNA samples used in lanes of 1, 5, 9 and 13 were from leaf mRNA of D8909-15; cDNA samples used in lanes of 2, 6, 10 and 14 were from root mRNA of D8909-15. Lanes 3, 7, 11 and 15 are positive controls that amplified PCR fragments from BAC clone VA31P21. Lanes 4, 8, 12 and 16 are negative controls that contain PCR reagents only without adding cDNAs. Lane M indicates DNA size markers

Full size image

Discussion

XiR1 fine-scale map

The ability to produce novel grapevine cultivars by conventional breeding is frequently hampered by a high degree of heterozygosity, the polygenic inheritance of desired characters and the long juvenile period before grape seedlings produce fruit. Molecular markers tightly linked to agriculturally important traits expedite the grape breeding process by allowing very early screening for single and preferably multiple traits, and therefore the earlier selection for additional traits and for next generation parents. Using 188 F1 progeny derived from a cross between D8909-15 and F8909-17, the XiR1 locus was mapped near marker VMC5a10 on chromosome 19 (Xu et al. 2008). Then, the positional cloning strategy was employed to isolate the putative XiR1 QTL using the tightly linked marker VMC5a10 to identify BAC clones covering the XiR1 region. The BAC end sequences of these clones were used to develop PCR-based markers to re-screen BAC libraries and to analyze recombinant plants in order to fine-scale map the XiR1 locus. The positions of the new markers were determined to refine the XiR1 vicinity to an interval size of 115 kb between M4F3R and 1N2R3b at seven recombination events using 1,375 F1 plants in three populations. This data further led to the isolation of the XiR1 locus, which contains members of NB-LRR type plant resistance genes.

Three mapping populations were used to develop a high-resolution map of the XiR1 region. Since the D8909-15 resistance source was used as the female parent in the three populations, it was expected that similar recombination frequencies would be seen in the interval between markers VNG3a10 and M4F3F. The recombination frequency ranged from 5.6 to 7.9% across the three populations, confirming the expected limited variation (Table 2). However, the number and frequency of recombinants identified in the XiR1 region suggested that there were large differences among the three populations. Nine recombinants were found within this region in the 9621 population of 943 genotypes; a frequency of 0.95% in a cross of two half-siblings. There were a total of 431 genotypes in the 0023 and 05384 populations, which shared V. rupestris × V. arizonica parent D8909-15 as the female parent, but had different V. vinifera male parents. The number and frequency of recombinants in the XiR1 region in these populations were 1 and 0.23%, respectively, and much lower than that observed in the 9621 population, suggesting that the male parents had a large impact on the frequency of XiR1 region recombinants. The cause for this observation is unknown, but these data suggest that the male parent F8909-17 of the 9621 population was genetically superior for producing informative recombinants in the XiR1 region.

The accuracy of positional cloning of the XiR1 locus

Many plant QTLs that have been mapped, including those with large and small effects, and have subsequently been verified using tagged or cloned genes. Examples of verified QTLs with large effect include fw2.2 for fruit size in tomato (Frary et al. 2000), Gpc-B1 for wheat grain protein, zinc, and iron content (Uauy et al. 2006), FT1 for flowering time in soybean (Yamanaka et al. 2005), and Sub1 for submergence tolerance in rice (Xu et al. 2006). Examples of verified QTLs with small effects include tb1 for shoot morphology in maize × teosinte (Doebley et al. 1997), Pup1 for P uptake in rice (Wissuwa et al. 2002) and inv/GE for sugar content in potato (Li et al. 2005). The XiR1 QTL was previously mapped near marker VMC5a10 at the telomeric end of chromosome 19 (Xu et al. 2008). In this study, XiR1 was physically located within 115 kb between markers 1N2R3b and M4F3R, which are within 1 LOD from the peak of the QTL, confirming that the position of XiR1 was very accurate. Although the XiR1 region physically spans 115 kb in the D8909-15-S genome, the corresponding region in the D8909-15-R genome was 40–70 kb with BACs S8O16 and S29J16 of the “R” genome origin, thus there is a gap estimated at 50 kb between the two BACs. This initial speculation was later confirmed using sequence data from BAC clone V31P21. The inability to completely cover the XiR1 region in the “R” genome may have been caused by the limited coverage of the D8909-15 library, which has 3.4× genome equivalents.

The major differences in the genetic and physical location of marker VMC5a10 were observed between the D8909-15 “S” and “R” genomes. This marker mapped with marker 1N2R3b on one side of XiR1 in the “R” genome, while it was physically mapped to the other side of XiR1 in the “S” genome. When the two V. vinifera Cabernet Sauvignon genomes were examined, a large insertion that carried a partially duplicated XiR1 region was inserted near marker 9I17R in one genome, resulting in a much larger XiR1 region than that in the other V. vinifera genome (data not shown). However, the 115 kb XiR1 region defined by the two markers 1N2R3b and M4F3R was largely conserved across genomes from the three Vitis species: the D8909-15-R (V. arizonica) genome, the D8909-15-S (V. rupestris) genome and the V. vinifera Cabernet Sauvignon genome without the large insertion. Thus, the XiR1 region provides a unique opportunity to examine evolutionary impacts on the microstructure of a given region specific to each of the genomes.

Structure and expression of genes in the XiR1 locus

Sequencing analysis of two BAC clones, one resistant (VA31P21) and one susceptible (M10J21), revealed several similar genes in both clones including two retrotransposon genes. Among the predicted candidate genes, the NB-LRR encoding gene is of particular interest because four out of the five characterized endoparasitic nematode resistance (R) genes are also NB-LRR genes (Williamson and Kumar 2006). It is intriguing to find that a NB-LRR gene may also confer resistance to an ectoparasitic nematode; perhaps, employing a similar strategy as tomato Mi-1.2 gene-based resistance to potato aphid. In both cases, the pest uses a stylet to penetrate and obtain nutrients while feeding from outside the host. There are two R gene homologues, Mi-1.1 and Mi-1.2, at the Mi-1 locus, only Mi-1.2 confers resistance to both nematodes and aphids (Milligan et al. 1998; Rossi et al. 1998). The predicted proteins from XiR1.1 and XiR1.2 have 70.3% identical amino acids. However, the alignment of the deduced amino acid sequences of XiR1.1, XiR1.2 and Mi-1.2 revealed 26% identity with structural similarity only to the LZ and NB domains. This is primarily due to the lack of a long N-terminus in XiR1.1 and XiR1.2, which is present in Mi-1.2 and only exists in Solanaceous plants. A search in GenBank found that these predicted proteins had highest similarity with the NB-LRR protein RGH1 (GenBank accession # AAO37645.1) from cassava (Manihot esculenta), an accession that currently has no known function (López et al. 2003). The finding that XiR1.1, XiR1.2 and XiR1sh transcripts were expressed in both root and leaf tissues of D8909-15 suggests that biological functions other than nematode resistance may be encoded by these three genes.

Plant R genes are usually found in tandem arrays of paralogues (Hulbert et al. 2001) clustered together with retroelememts (Michelmore 2000). It has been proposed that intragenic recombination is important in R gene evolution by combining parts of different alleles to create new specificities (Van der Hoorn et al. 2001). Sequence comparisons between our data and the French/Italian Pinot noir genome sequencing database (http://www.genoscope.cns.fr/externe/GenomeBrowser/Vitis/) indicate that the XiR1 region at the telomeric end of chromosome 19 contains both NB-LRR-encoding and retrotransposon genes. The majority of plant R genes cloned to date belong to the NB-LRR family. These genes can be subdivided into two broad classes based on whether the N-terminal region contains a TIR domain or a predicted coiled coil (CC) that contains the leucine zipper domain. XiR1.1, XiR1.2 and XiR1sh each contain a possible CC structure with the predicted leucine zipper motif based on the COILS program and thus fit well into the CC–NB–LRR subclass. In addition, these three genes contain the kinase-1a (P-loop), kinase-2 and kinase-3a motifs in their conserved NB domains. The variable C-terminal LRR domain is involved in pathogen recognition specificity and in regulating the signalling activity of the R protein and mutations or single amino acid changes in this domain can lead to constitutively active proteins (Hwang et al. 2000; Bendahmane et al. 2002; Hwang and Williamson 2003). XiR1.1, XiR1.2 and XiR1sh are predicted to have 12, 14 and 14 LRRs, respectively.

To test whether XiR1.1 or XiR1.2 is responsible for conferring X. index resistance, several nematode-susceptible grape lines will be transformed with constructs containing each of these genes using Agrobacterium-mediated (A. tumefaciens and A. rhizogenes) transformation. Although the efficiency of Agrobacterium-mediated genetic transformation for grape has improved, the confirmation of XiR1 function will take at least 1 year. Nevertheless, the map-based cloning approach for grape resistance genes has directly aided the grape rootstock breeding program by providing DNA markers that have accelerated screening for resistance saving time and cost in the greenhouse and field. This study has identified the first locus responsible for ectoparasitic nematode resistance and serves as a model for additional molecular genetic studies on the mapping, cloning and characterization of key grape traits.