| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a New Zealand Institute for Crop & Food Research Ltd, P.O. Box 4704, Christchurch, New Zealand
b New Zealand Plant Breeding Ltd, P.O. Box 19, Lincoln, New Zealand
c Dep. of Agriculture Western Australia, 3 Baron-Hay Court, South Perth, Western Australia, Australia
d Ministry for Research, Science and Technology, Wellington, New Zealand
* Corresponding author (timmermang{at}crop.cri.nz)
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARYREFERENCES |
|---|
Abbreviations: A88, resistant breeding line 3148-A88 • AFLP, amplified fragment length polymorphism • CIM, composite interval mapping • cM, centimorgan • MAS, marker-assisted selection • QTL, quantitative trait locus/loci • RAPD, random amplified polymorphic DNA • RFLP, restriction fragment length polymorphism • R-gene, disease or pest resistance gene • RGA, resistance gene analog • SCAR, sequence characterized amplified region
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARYREFERENCES |
|---|
Resistance to Ascochyta blight has been demonstrated within pea germplasm. Pea genotypes show differences in resistance or susceptibility to M. pinodes and P. medicaginis var. pinodella that is independent of the virulence of the pathogen isolate (Onfroy et al., 1999; Wroth, 1998a; Xue et al., 1998). Complete resistance to infection by either pathogen has not been observed in pea. Using different germplasm, Wroth (1999) found that resistance to M. pinodes infection showed quantitative inheritance, while Clulow et al. (1991) suggested that resistance to M. pinodes showed major gene inheritance.
Molecular linkage maps and mapping of QTLs are valuable tools for characterizing the genetics of disease resistance, localizing resistance loci on linkage maps, and identifying linked polymorphic DNA sequences that might be used for marker-assisted selection (MAS) during plant breeding. QTL mapping has characterized multigene resistance to fungal pathogens in common bean (Phaseolus vulgaris L., Geffroy et al., 2000), barley (Hordeum vulgare L., Qi et al., 1998; Zhu et al., 1999), and pea (Dirlewanger et al., 1994).
QTL mapping of disease resistance loci facilitates the use of molecular approaches to understanding the nature of disease resistance QTLs and may eventually lead to the cloning of underlying genes. Major plant disease resistance genes (R-genes) have been cloned by various means including positional cloning based on molecular linkage maps (Hammond-Kosack and Jones, 1997). Disease response genes (DR-genes) have also been characterized, and encode proteins with a wide range of activities that are involved in defense-response pathways (Maleck and Dietrich, 1999). Colocalization of disease resistance QTLs with R-gene-like or DR-gene-like sequences has been demonstrated, for example in wheat (Triticum aestivum L., Faris et al., 1999), pepper (Capsicum annuum L., Pflieger et al., 1999), and rice (Oryza sativa L., Wang et al., 2001). In pea, NBS-LRR type R-gene analogues (RGAs) have been cloned and mapped on molecular linkage maps (Timmerman-Vaughan et al., 2000). A number of DR-gene or putative DR-gene sequences have been cloned from pea, and some of these have been mapped on molecular linkage maps (Gilpin et al., 1997; Weeden et al., 1998).
Slow progress in breeding pea for field resistance to Ascochyta blight can be attributed to the trailing nature of the pea plant, the difficulties in reliably establishing a uniform epidemic in a field situation, and the multigenic nature of the resistance. The development of pea cultivars with resistance to M. pinodes and P. medicaginis var. pinodella will contribute substantially to the control of this disease and help reduce losses. To identify molecular markers for use in MAS and to understand the genetics underlying resistance to Ascochyta blight, we have conducted a linkage mapping experiment to localize QTLs for field resistance to Ascochyta blight and for plant developmental stage on a molecular linkage map of the pea genome.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARYREFERENCES |
|---|
DNA Marker Methods and Linkage Mapping
DNA to reconstitute the genotype of each F2 line was extracted from leaves bulked from at least five F3 descendants, as described by Timmerman et al. (1993). RAPDs (random amplified polymorphic DNA) and RFLPs (restriction fragment length polymorphisms) were also carried out as described by Timmerman et al. (1993). AFLPs (amplified fragment length polymorphism) were analyzed by the methods described in Barrett and Kidwell (1998) and Pickering et al. (1995) by means of seven MseI:PstI primer combinations (Table 1)
. Using AFP2-P5e as an example, we detected this locus using selective primers MseP2 and PstP5; it was the fifth polymorphic band down from the top of the autoradiograph. AFLP designations are pedigree specific and do not necessarily relate to loci on other pea maps. RFLP probe c206 was obtained from Noel Ellis (John Innes Centre, Norwich, UK) and PeaCO was obtained from Richard MacKnight (Department of Biochemistry, Otago University, Dunedin, New Zealand). Other RFLP probes were described by Gilpin et al. (1997). RAPDs were amplified with primers from Operon Technologies (Alameda, CA). RAPD designations indicate the primer used followed by the approximate length of the amplified DNA fragment in base pairs.
|
Field Trials
A88 x Rovar progeny lines were evaluated in field trials conducted on an irrigated site at Medina, Western Australia, in 1997, 1998, and 1999. In all three trials, one plot was planted of each progeny line, because of a limit on seed, in an unreplicated design. Several plots of check lines (A88, Rovar, and ‘Dundale’) were distributed randomly throughout the trial to allow any field variation to be assessed and adjusted. Dundale is a susceptible dun type pea cultivar that is widely grown in Australia. Placement of check plots was based on a Latin-square-like arrangement to give approximately even distribution of these lines across the rows and columns of plots in the trial.
The 1997 trial was sown on 28 May, and included 225 F2:3 progenies in a 15- by 20-plot layout. There were 25 plots of each of three check lines. Resistant breeding line 3176-A26 was substituted for A88 in the 1997 trial because of limited A88 seed availability. The 1998 trial was planted on 13 May and included 157 F2:4 progenies in a 15- by 15-plot layout and included 23, 24, and 21 plots, respectively, of Dundale, Rovar, and A88. The 1999 trial, sown on 24 May, contained 174 F2:4 progeny plots in a 12- by 19-plot layout with 18 plots each of lines Dundale, Rovar, and A88.
The trials were machine planted in twin rows with a cone seeder. Each 2-m plot was sown with approximately 40 seeds. Twin rows in a plot were spaced 0.3 m apart, plots within a row were separated by 1 m, and plots between rows were 1 m apart. Standard agronomic practices used included application of fertilizers, as well as herbicides, insecticides, and methiocarb (4-methylthio-3,5-xylyl methylcarbamate) pellets to control weeds, insect pests, and slugs or snails, respectively.
Field trials were evaluated for Ascochyta blight twice for each trial. Trials were scored on the following dates: 4 to 6 and 13 to 15 Sept. 1997, 19 to 21 and 26 to 28 Sept. 1998, and 6 to 8 and 16 to 18 Sept. 1999. Scoring was timed so that most plots were at the pod-set to pod-swell reproductive stages (Knott, 1987), which is the time when disease incidence increases rapidly in the field. Separate rating scales were used to score disease development on stems, leaves, and pods in 1998 and 1999, and on stems and leaves only in 1997. Stems were scored by a 0-to-9 scale where 0 = no lesions, 1 = flecks, 3 = a few large lesions, 5 = many large lesions, 7 = stem girdled with lesions, 9 = plant dead. Disease development on leaves was scored by a 1-to-9 scale where 1 = no lesions, 3 = lesions up 1/4 of the plant height with only a trace of disease apparent, 5 = lesions up 1/2 of the plant height with several diseased areas, 7 = lesions up 3/4 of the plant height with several diseased areas, and 9 = lesions to the top of the plant with severe disease apparent. Disease development on pods was scored on the lowest pods apparent on the main stems by a 0-to-10 scale where 0 = no lesions, 2 = a few pinpoint lesions, 4 = many pinpoint lesions, 6 = many pinpoint lesions and a few coalesced and sunken lesions visible, 8 = large coalesced and sunken lesions present, 10 = pods nearly completely blackened. Intermediate scores were used for all three scales.
Plant developmental stages were scored on 6 Sept. 1998 and 8 Sept. 1999 by a rating scale based on the reproductive stages described by Knott (1987). This rating scale was used because it can be applied at the same time as disease scores are being collected. The scores used were 2 = visible buds, 3 = first open flower, 4 = pod set, 5 = flat pod, 6 = pod swell, 7 = pod fill, and 8 = green wrinkled pod. Developmental stages were scored on the lowest pods apparent on the main stems.
In addition to using single environment trait scores for QTL mapping, we treated the three environments as replicates and trait scores were averaged across environments. Traits designated as Stem1 Mean, Stem2 Mean, Leaf1 Mean, and Leaf2 Mean are averages of scores from 1997, 1998, and 1999, while Pod1 Mean, Pod2 Mean, and Mat mean are averages of scores from 1998 and 1999.
Analysis of Trait Data for Field Trends
Trait data were examined independently for the presence of field trends. To visualize the range and distribution of the scores across the field, the scores and residuals from a null analysis were plotted in a 2-dimensional array to represent the trial, and each plot was colored according to value. To look for overall patterns in another way, the mean score or residual for each row (or column) was plotted against the row (or column) number. Residual maximum likelihood methods (REML; Patterson and Thompson, 1971) as implemented in Genstat (Genstat Committee, 1997, 2000) were used to fit models to the data to make adjustments for any field trends. The first model fitted, a null analysis, made no adjustment for trends. Models included general adjustments for row or columns patterns, auto-correlations between plots (Gilmour et al., 1997) or smooth-splines (Verbyla et al., 1999) across the rows or columns. Disease scores adjusted for the field trends could then be predicted from these models for each line, and the adjusted scores used in further analysis.
Linkage and QTL Mapping
A genetic linkage map of the A88 x Rovar cross containing RFLP, RAPD, AFLP, and SCAR markers was constructed by MAPMAKER/EXP ver. 3.0 (Lincoln et al., 1992) as described by Gilpin et al. (1997) for the 133 F2 progenies that were field trialed in all three years. Map construction was conducted by the MapMaker commands "two point," "group" (LOD = 4.0), "compare," and "build." The "ripple" command (LOD = 2.0) was used to test final orders. Data for 206 segregating marker loci were available. The linkage map was constructed with 96 loci. The loci were chosen to include all available codominant markers, to produce a map with average saturation of about 10 centimorgans (cM), and to minimize missing data. Three of the codominant loci (AFP3-P2bc, P3P8bM09, and AFP2-P2hl) were generated by joining haplotypes for three pairs of linked dominantly inherited markers that showed tight linkage when maps were constructed with the 206 segregating loci, that were in opposite phases, and that showed no crossovers between the A88/A88 and Rovar/Rovar genotypes. P3P8bM09 was generated by joining the haplotypes of AFP3-P8b and M09-2400. The resulting codominant loci mapped to the same intervals as the respective dominantly inherited markers and increased the support for the map orders obtained (data not presented).
QTLs were mapped with QTL Cartographer ver. 1.21 (Basten et al., 2001) using composite interval mapping (CIM). Cofactors were selected by the program using Model 6 with the genetic background controlled by up to five markers and the window size set at 10 cM. The percentage of variance (R2 under H3:H0) was estimated by QTL Cartographer. QTLs detected by CIM were declared as significant if they exceeded the chromosome-wise significance level at 
= 0.05 determined by conducting permutation tests (Churchill and Doerge, 1994) on individual linkage groups with QTL Cartographer and shuffling trait data 1000 times. The trait phenotypic means for the A88/A88 (A/A), A88/Rovar (A/B), and Rovar/Rovar (B/B) genotypic classes at marker loci linked to QTL peaks were calculated by Genstat software (Genstat Committee, 2000).
Culture and Characterization of Fungal Isolates
Infected pea plant material was collected from three field trial sites throughout Western Australia in September 1996, at the height of the Ascochyta blight epidemic. Leaves, stems, and tendrils containing discrete lesions were surface sterilized in household bleach diluted to 1% (v/v) hypochlorite then blotted on sterile filter paper. Individual lesions were dissected out, cut through the middle, placed on potato dextrose agar (PDA, Difco, Detroit, MI), then incubated overnight at 22°C. Fine hyphae growing from the lesions were excised and transferred to V8 (Campbell Soup Company, Camden, NJ) agar (300 mL V8 juice, 3 g CaCO3 and 15 g/L agar (Difco), pH 6.0) containing 200 mg/L Timentin (SmithKline Beecham, Philadelphia, PA) and incubated at room temperature under near UV light for up to 3 wk to permit formation of pycnidia–perithecia. Spore morphology (Bretag and Ramsey, 2001) was examined with light microscopy after samples were stained with lactophenol cotton blue.
Isolates for DNA extractions were produced from sporulating cultures. Spores or asci scraped from plates were suspended in 100 µL sterile H2O then streaked out onto PDA containing 200 mg/L Timentin. Individual, well-separated colonies were then subcultured on PDA containing 200 mg/L Timentin. Mycelial cultures for DNA extractions were grown in 50 mL potato dextrose broth (Difco) at room temperature for 5 d with shaking. For DNA extractions, mycelia were harvested by vacuum filtration onto sterile filter paper and washed with sterile H2O. Mycelia were frozen in liquid N2 then ground to powder in a mortar and pestle. DNA was extracted with CTAB (hexadecyltrimethylammonium bromide) buffer containing 28 mM ß-mercaptoethanol and RAPD banding patterns were analyzed following the methods for plant DNA described by Timmerman et al. (1993).
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARYREFERENCES |
|---|
|
Linkage Map
A genetic linkage map (Fig. 2)
of 133 F2 progeny of the A88 x Rovar cross was constructed from 96 loci (28 RFLPs, 22 AFLPs, 43 RAPDs, and three SCARs) chosen from 206 segregating molecular markers. The map covers about 1050 cM of the pea genome on 11 linkage groups. The average interval between markers is 12.4 cM. There are 29 codominant markers. The linkage map was related to the consensus map of the pea genome (Weeden et al., 1998). Nine of the 11 linkage groups were identified by anchor markers (Fig. 2). The bottom half of linkage group V does not contain anchor loci but was related to the consensus map by mapping loci AFP2-P5d, Y16-1121 and AFP2-P2e between gp and Pgdc in the JI1794 x Slow (Weeden et al., 1993) reference population (data not presented). Linkage group A remains unassigned.
|
In these experiments, the progeny lines were not replicated within trials, primarily because of limited seed availability, but also because of limited field space in the disease nursery. However, the lines were replicated between three different trials conducted in consecutive years. All trials contained check plots of resistant and susceptible cultivars that were highly replicated throughout to assist in assessing spatial variation.
In a field trial containing a large number of plots, conditions that affect trait expression may vary significantly across the field; therefore, analysis of the effect of spatial variation can improve the estimation of genetic values (Gilmour et al., 1997). While the use of replication is one way to increase the precision of trait measurement, increased spatial variation may result from increased field size. Ideally, field trial designs would incorporate suitable replication along with analysis of spatial variation, although this could result in large and expensive trials. An alternative when there are insufficient resources to permit replication is to conduct unreplicated trials incorporating checks and to use analysis methods that compare the experimental genotypes (progeny lines in our study) with local checks or neighboring plots to control local variation and that analyze broad effects like row and column effects and gradient effects. The use of unreplicated designs that incorporate many checks has a long history in plant breeding. Kempton and Gleeson (1997) have presented methods for analyzing spatial variation in unreplicated trials and have reviewed some of the literature on the use of unreplicated trials with analysis of spatial variation in plant breeding. Cullis et al. (1989) also described a procedure for applying spatial analysis to unreplicated breeding trials. Unreplicated field trials with analysis of spatial heterogeneity have recently been used to develop a method for MAS that integrates molecular marker and spatial information to predict genetic values for grain yield in maize (Moreau et al., 1999). These authors showed that the accuracy of genetic value predictions was improved when appropriate statistical models were used to analyze spatial heterogeneity.
As discussed by Moreau et al. (1999), analysis of spatial variation is seldom used in QTL mapping experiments. Recently, two examples have been published where analysis of spatial variation by REML was combined with QTL analysis by regression (Eckermann et al., 2001, Smith et al., 2001). In these papers, standard mapping approaches were compared with a 1-step mixed model approach in which QTLs were detected by regression on pairs of linked marker loci simultaneously with modeling of field and/or laboratory trends. The results did not provide a clear demonstration that the 1-step mixed model approach is more powerful for detecting QTLs than standard approaches, and Eckerman et al. (2001) concluded that there are still difficulties that need to be resolved. Smith et al. (2001) did suggest, however, that the results indicate that standard approaches without spatial analysis may tend to overestimate the significance and effects of QTLs.
Phenotypic Variation
To assess whether the raw or adjusted trait data should be used for QTL detection, we conducted genome scans using composite interval mapping (CIM) for raw and adjusted trait values. The chromosome-wise thresholds ( = 0.05) for QTL peaks with LOD 
2.0 were determined by permutation testing of 1000 permutations of the trait data (Churchill and Doerge, 1994). When they were compared, the QTLs detected from raw versus adjusted scores had similar statistical support and the peak positions either coincided or overlapped. Since the trends observed in these trials were not large, this result was not surprising. Therefore, the raw scores were used for QTL mapping.
The frequency distributions obtained for the disease severity and plant reproductive stage traits assessed in 1997, 1998, and 1999 on 133 progeny are shown (Fig. 3) . Parental means are shown for 1998 and 1999 only since A88 was not randomized throughout the 1997 trial because of limited seed availability. Small numbers of transgressive segregants were observed for the following traits: Stem1 and Stem2 (1998), Stem1 (1999), Pod1 and Pod2 (1998), Maturity (1998), and Maturity (1999). A larger number of transgressive segregants was observed for Pod1 (1999). The Rovar parental mean for this trait was 4.42 and there were 44 (33%) progeny lines with scores between 6 and 9. However, the Pod2 (1999) scores did not display this extent of transgressive segregation. Transgressive segregation may suggest that alleles conditioning trait values are contributed by both parental lines or that genotype x genotype (G x G) interactions may occur.
|
= 0.05 chromosome-wise threshold. The QTL mapping results indicate that resistance to Ascochyta blight is complex in the A88 x Rovar population. Putative QTLs for resistance to Ascochyta blight field epidemics were discovered on linkage groups I, II, III, IV, V, VII, and Group A. Eight of the QTLs were detected in two or more environments or by more than one measure of disease resistance. These QTLs have been given locus designations (Fig. 4). The remaining QTLs were detected in only one set of single environment disease scores.
|
|
The resistance alleles of most of the QTLs are associated with marker alleles contributed by the resistant parent A88 (Table 2). In contrast, the resistance alleles of three QTLs are associated with Rovar marker alleles: the linkage group I QTL associated with Q407, the linkage group V QTL associated with locus sAFP2P2c, and the group A QTL associated with P202. A number of the QTLs displayed overdominant inheritance modes. In these cases, the progeny that are heterozygous at the marker locus associated with the QTL are on average more susceptible (i.e., have a higher disease score) than progeny that are A88/A88 or Rovar/Rovar homozygotes at the associated marker locus. The observations that Rovar contributes resistance alleles and of overdominance may account for the transgressive segregation observed in the progeny (Fig. 3).
The phenotypic variation (R2) explained by the individual resistance QTLs ranged from 8 to 35% (Table 2). Individual traits only detected a subset of the 13 QTLs that were identified from all trait scores across all environments. Pod2 (1999) and Leaf1 (1998) detected four QTLs each, the most detected by a single environment disease score. Simulation and validation studies have shown that QTL mapping studies involving small progeny numbers and relatively low heritabilities often overestimate the magnitude of QTL effects (Beavis, 1994; Melchinger et al., 1998; Utz and Melchinger, 1994). Furthermore, Melchinger et al. (1998) showed that the relative bias in R2 is greater when there are a large number of minor QTLs than when a small number of major QTLs is involved. In this paper, we have discovered a large number of QTLs, most of which explain only a small fraction of the phenotypic variation, using a population of 133 progeny lines. Therefore, the R2 values that we have estimated may be biased. Further studies are underway to confirm the QTLs and their contributions to phenotypic variation using an additional F2 family and a population of >300 recombinant inbred lines that have been developed from related resistant germplasm.
QTLs for Plant Maturity
In field epidemics, Ascochyta blight severity increases as pea matures; therefore, later maturing plants or cultivars may appear to have enhanced resistance when lines in field trials are assessed on the same date (Kraft et al., 1998). This may be explained if genes for late maturity are linked to genes for Ascochyta blight resistance or have a pleiotropic effect on resistance. Even though A88 and Rovar were chosen for their similarities in reproductive maturity, variation in maturity was noticed among the progeny of the A88 x Rovar cross in the 1997 field trial. Consequently, we mapped QTLs for maturity using scores for reproductive stage from the 1998 and 1999 field trials.
A significant QTL affecting plant reproductive stage was detected on linkage group II (Mat2.1, Fig. 4 and Table 2) in a genomic region containing Ascochyta blight resistance QTL Asc2.1. The LOD plot peaks for Mat2.1 and Asc2.1 overlap, and the late maturity and low disease score alleles of both QTLs are associated with A88 marker alleles. Asc2.1 and Mat2.1 may be linked loci or Asc2.1 may detect a pleiotropic effect of Mat2.1. To attempt to resolve these two possibilities, we examined the inheritance modes for the QTLs. Calculation of degree of dominance (d/a, Paterson et al., 1991) suggested recessive/additive inheritance for the Mat2.1 QTL on the basis of Maturity (1999) and Maturity Mean traits which had d/a ratios of -0.73 and -0.61, respectively. In contrast, the disease traits associated with Asc2.1 showed dominant/additive inheritance modes, with the following d/a ratios: Leaf2 (1997), d/a = 0.55; Leaf1 (1998), d/a = 0.89; Leaf2 (1998), d/a = 0.88; Pod2 (1998), d/a = 0.66, Leaf2 Mean, d/a = 0.90; and Leaf1 Mean, d/a = 2.99. The d/a ratio for Leaf1 Mean suggests overdominant inheritance, but this inheritance mode is not obvious from examining the Asc2.1 genotypic means for this trait (Table 2). However, examination of the distribution of disease scores reveals that the heterozygous (A88/Rovar) class contains the progeny lines with the highest (most susceptible) disease scores (data not presented).
Mat2.1 maps in the vicinity of PeaCO, the pea homolog of the Arabidopsis thaliana CONSTANS (CO) gene. CO is a zinc-finger transcription factor which in A. thaliana (L.) Heynh. controls flowering in response to long days (Robson et al., 2001). In rice, the QTL Hd1 controls heading date and was recently shown to be a homolog of CO (Yano et al., 2000). Although the Mat2.1 QTL peak does not coincide with PeaCO, the QTL peak is closely linked to PeaCO; therefore, the pea CO homolog should be considered as a candidate gene for the Mat2.1 QTL. Flowering time in pea is controlled by a number of major genes including Lf, E, Hr, Sn, and Dne and may also be controlled by quantitative gene systems (Murfet, 1990). Lf (late flowering) determines the minimum length of the vegetative period, and also maps to linkage group II, approximately 10 cM from A locus (Murfet, 1971). The consensus linkage map for pea (Weeden et al., 1998) places Lf on the other side of LgJ from the genomic region where we have mapped Mat2.1.
Our observation that QTLs for plant maturity and resistance coincide are not unique. Other studies of multigenic resistance to diseases that affect adult or mature plants have also shown colocalization of QTLs for resistance and maturity (Collins et al., 1999; Ewing et al., 2000; Kim and Diers, 2000; Qi et al., 1998). In our case, only Asc2.1 maps in a genomic region containing a QTL affecting plant reproductive stage. Therefore, the genetic basis for field resistance to Ascochyta blight in the A88 x Rovar population is largely independent of the effects of plant reproductive stage.
Mechanisms for Ascochyta Blight Resistance
The terms "vertical resistance" and "horizontal resistance" are used to describe the interactions between hosts and pathogens (Nelson, 1972; Simmonds, 1979; Van der Plank, 1968). Vertical resistance is hypersensitive, race, or isolate specific and involves "gene-for-gene" interactions between pathogen avirulence genes and host resistance genes. Horizontal resistance is quantitatively inherited, assumed to be race or isolate nonspecific, and multigenic. Linkage mapping of quantitative resistance has revealed examples where minor genes combine with major QTLs or where QTLs are colocalized with isolate specific resistance genes (Li et al., 1999; Messmer et al., 2000; Pernet et al., 1999; Wang et al., 2000). Two models could explain the genetic basis of Ascochyta blight resistance in pea. Resistance could be (i) race or strain nonspecific and due to minor genes, or (ii) due to a combination of major or isolate specific genes and minor genes. We cannot currently choose between the two models, first because the response of the pathogens to the resistant germplasm has not been characterized and second because we do not have clear evidence for the existence of major QTLs. Furthermore, our studies are based on the response of progeny lines to field epidemics where complex pathogen populations exist; therefore, we have not examined race or isolate specificity of the resistance.
Studies to demonstrate whether pathotypes exist among M. pinodes isolates have produced contradictory results. Pathotype groups have been identified among isolates from Canada (Xue et al., 1998), Germany (Nasir and Hoppe, 1991), and the UK (Clulow et al., 1991). In contrast, studies on isolates from Western Australia (Wroth, 1998b) and France (Onfroy et al., 1999) did not define pathotype groups among isolates although they did detect variation in virulence. Onfroy et al. (1999) also could not define pathotype groups among P. medicaginis var. pinodella isolates. All the above studies were based on seedling tests and may not be relevant to adult plant resistance (Ali et al., 1978). To assess whether multigenic resistance to Ascochyta blight in A88 and Rovar involves isolate specific genes, further studies are needed to determine whether isolates of M. pinodes and P. medicaginis var. pinodella show differences in virulence on A88 and related resistant breeding lines versus more susceptible germplasm.
The vertical versus horizontal resistance models may also be distinguished if we determine whether major QTLs are involved. The most significant QTLs detected were Asc1.1, Asc2.1, and Asc3.1. These had large LODs (5.67, 5.40, and 5.13, respectively) and accounted for up to 35, 16, and 16% of the observed phenotypic variation, respectively (Table 2). However, the R2 values may overestimate the phenotypic variation explained, as discussed above. Further studies are underway to map QTLs for field resistance to Ascochyta blight using additional populations developed with resistant breeding lines related to A88. These studies will provide data to validate the number of QTLs, their genomic locations and the sizes of the QTL effects, and may help determine whether major QTLs are involved in specifying the resistance in these lines.
Candidate Disease Resistance Genes
QTL mapping permits candidate genes to be identified as a first step toward understanding the molecular basis of multigenic disease resistance. To facilitate candidate gene identification, we previously mapped cloned genes of known function (Gilpin et al., 1997) and RGAs (Timmerman-Vaughan et al., 2000) on pea linkage maps. In the A88 x Rovar population, only two candidate disease response sequences have been mapped, PI39 which encodes DRR230-b, a defensin-like molecule (Chiang and Hadwiger, 1991), and DCCHIT which encodes inducible chitinase (Chang et al., 1995). Asc3.1 is linked to PI39. A Medicago sativa L. protein with sequence similarity to DRR230-b was shown to enhance resistance to a fungal pathogen in transgenic potato (Gao et al., 2000). The QTL on unidentified group A shows linkage to the DCCHIT locus.
Linkage of disease resistance QTLs and known R-genes or RGAs has been shown in other plant species (Geffroy et al., 2000; Pflieger et al., 1999; Wang et al., 2001). To determine whether pea RGAs occur in genomic regions containing Ascochyta blight resistance QTLs, the A88 x Rovar map was compared with the pea linkage maps containing RGAs (Timmerman-Vaughan et al., 2000). RGAs were not mapped directly in the A88 x Rovar population because most were not polymorphic for these parental lines when parental Southern blots were probed (data not presented). The map comparison showed that three regions containing RGAs also contain Ascochyta blight resistance QTLs. RGA1.1, which is linked to M27 on linkage group III, is in the vicinity of Asc3.1; RGA2.97 on linkage group VII is linked to N96AB and therefore in the vicinity of the QTL detected from the Leaf1 (1998) scores; and RGA-G3A maps to linkage group VII between N96AB and D8B therefore is in the vicinity of the QTL detected from the Stem2 (1999) scores.
|
SUMMARY |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARYREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARYREFERENCES |
|---|
Received for publication November 29, 2001.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARYREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. X. Zhang and B. D. Gossen Heritability Estimates and Response to Selection for Resistance to Mycosphaerella Blight in Pea Crop Sci., November 7, 2007; 47(6): 2303 - 2307. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Zhang, S.-F. Hwang, B. D. Gossen, K.-F. Chang, and G. D. Turnbull A Quantitative Analysis of Resistance to Mycosphaerella Blight in Field Pea Crop Sci., January 22, 2007; 47(1): 162 - 167. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Zhang, S.-F. Hwang, K.-F. Chang, B. D. Gossen, S. E. Strelkov, G. D. Turnbull, and S. F. Blade Genetic Resistance to Mycosphaerella pinodes in 558 Field Pea Accessions Crop Sci., October 2, 2006; 46(6): 2409 - 2414. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. M. Timmerman-Vaughan, A. Mills, C. Whitfield, T. Frew, R. Butler, S. Murray, M. Lakeman, J. McCallum, A. Russell, and D. Wilson Linkage Mapping of QTL for Seed Yield, Yield Components, and Developmental Traits in Pea Crop Sci., May 27, 2005; 45(4): 1336 - 1344. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Hecht, F. Foucher, C. Ferrandiz, R. Macknight, C. Navarro, J. Morin, M. E. Vardy, N. Ellis, J. P. Beltran, C. Rameau, et al. Conservation of Arabidopsis Flowering Genes in Model Legumes Plant Physiology, April 1, 2005; 137(4): 1420 - 1434. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||
