| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
a Dep. of Agronomy, Lilly Hall, Purdue Univ., 915 W. State St., West Lafayette, IN 47907
b Dep. of Agriculture and Food Western Australia, Locked Bag 4, Bentley Delivery Centre, WA, Australia. M. Francki also Value Added Wheat Cooperative Research Centre, North Ryde, NSW, 2113 Australia. Contribution from Purdue Univ. Agricultural Research Programs as Journal Article no. 2005-17714
* Corresponding author (hohm{at}purdue.edu).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: CIM, composite interval mapping • cM, centimorgan • RI, recombinant inbred • QTL, quantitative trait locus • SNB, stagonospora nodorum blotch • SSD, single-seed descent
a Dep. of Agronomy, Lilly Hall, Purdue Univ., 915 W. State St., West Lafayette, IN 47907
b Dep. of Agriculture and Food Western Australia, Locked Bag 4, Bentley Delivery Centre, WA, Australia. M. Francki also Value Added Wheat Cooperative Research Centre, North Ryde, NSW, 2113 Australia. Contribution from Purdue Univ. Agricultural Research Programs as Journal Article no. 2005-17714
* Corresponding author (hohm{at}purdue.edu).
Resistance to stagonospora nodorum blotch (SNB) in glumes of hexaploid wheat (Triticum aestivum L.), caused by Phaeosphaeria (Stagonospora anamorph) nodorum was investigated in a recombinant-inbred (RI) population. The Purdue University winter wheat breeding lines P91193D1 and P92201D5, unrelated by parentage but both exhibiting partial SNB resistance, were crossed to develop 254 RI lines by single-seed descent (SSD) from a random population of F2 plants, to identify quantitative trait loci (QTLs) controlling SNB resistance in wheat glumes. The RI population, together with parent lines, was phenotyped for glume resistance to SNB under field conditions in F8:10 at Evansville, Vincennes, and Lafayette, IN, in 2003; in F7:9 at South Perth, Australia, in 2004; and in F8:10 in greenhouse-grown inoculated tests at Lafayette in 2003 and 2004. Two QTLs for resistance to SNB in glumes were identified: QSng.pur-2DL.1 from P91193D1 and QSng.pur-2DL.2 from P92201D5. The QTL QSng.pur-2DL.1 explained from 12.3% of the phenotypic variation for resistance in southern Indiana (Evansville and Vincennes) to 38.1% at South Perth; QSng.pur-2DL.2 accounted for 6.9 and 11.2% of the phenotypic variation in Indiana and South Perth, respectively. This study is the first report of SNB glume blotch resistance in which the same QTLs were identified in tests on different continents where Stagonospora nodorum populations are probably genetically diverse.
Abbreviations: CIM, composite interval mapping • cM, centimorgan • RI, recombinant inbred • QTL, quantitative trait locus • SNB, stagonospora nodorum blotch • SSD, single-seed descent
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Durable resistance to SNB is necessary since S. nodorum has both a mixed mating system and a high degree of genetic diversity (McDonald and Linde, 2002). There is significant variability within populations of S. nodorum; Krupinsky et al. (1972) reported variation for mycelial color, pycnidia density, and number of conidia per pycnidium; pathogenicity was identified against seedlings (Krupinsky, 1997) and adult wheat plants (Arseniuk and Czembor, 1999).
Resistance to SNB in wheat has been effective for controlling yield losses in Western Australia (Loughman et al., 1999) and the southeastern USA (Cunfer and Johnson 1999), where multiple sources of resistance have been identified. Consequently, host resistance appears to be an effective method of reducing yield losses to SNB. There have been a few reports of single dominant genes conferring SNB resistance (Ma and Hughes, 1995; Murphy et al., 2000b) and host genotype-specific toxins conferring sensitivity to S. nodorum (Friesen et al., 2006; Liu et al., 2004). Most studies indicate, however, that SNB resistance is quantitatively inherited with additive genetic effects (Bostwick et al., 1993; Wicki et al., 1999; Wilkinson et al., 1990). Disease severity may be influenced by plant height because the flag leaves and glumes, important in grain filling, are infected with rain-splashed pycnidiospores earlier on short plants than tall plants (Eyal, 1981; Scott et al., 1982). Also, late infection of the spike causes less disease than early infection due to a shorter time between infection and onset of natural senescence before the disease has a chance to develop (Williams and Jones, 1973). Significant environmental and cultivar differences affect disease development (Wainshilbaum and Lipps, 1991), increasing the difficulty of breeding for SNB resistance. McDonald and Linde (2002) suggested that quantitative resistance is an effective breeding strategy to reduce crop losses caused by SNB and that marker-assisted selection provides an effective technology to combine multiple SNB resistance quantitative trait loci (QTLs) in wheat.
Several reports have identified QTLs for SNB resistance in either seedlings (Czembor et al., 2003), leaves (Arseniuk et al., 2004), or glumes (Schnurbusch et al., 2003; Aguilar et al., 2005). These studies have not determined, however, whether the same QTLs are effective on different continents, where fungal isolates are likely to be genetically diverse. The objectives of this study were to identify QTLs associated with SNB glume blotch resistance in a recombinant-inbred (RI) population derived from two winter wheat inbred lines, both having partial but different resistance genes, and to compare QTLs identified in the USA and Australia, representing different S. nodorum populations. Given the high variability within S. nodorum populations, this knowledge will determine the feasibility of exploiting QTLs from international sources or whether breeding should focus on genes for glume resistance within local environments.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Disease Screening in the USA
In 2003, the RI population and its two parent lines were grown at three locations in Indiana: six replications at Lafayette and two replications each at Evansville and Vincennes. Plots, in a randomized complete block design, were sown as single 1-m rows (containing 100 seeds) and spaced 30 cm apart. The Lafayette site was mist irrigated daily for 2 h in the morning and evening from approximately 3 wk before to 2 wk after flowering to enhance SNB infection. Field disease evaluation at Evansville, Vincennes, and Lafayette was based on natural infection. Disease severity was visually assessed as the percentage of diseased glume tissue of the entire row when early-, mid-, and late-maturing RI lines were at the early- to soft-dough stage (Feekes 11.2). Maturity (days from 1 January) and plant height at physiological maturity were recorded for three replications at Lafayette.
The RI population and its parents were evaluated in two experiments in a greenhouse at Purdue University in 2003 and 2004. Some RI lines were lost during vernalization and before flowering; therefore, 226 RI lines and the parents were evaluated in 2004. The RI lines and parents were planted in 30- by 60- by 10-cm depth trays and vernalized at 3°C for 75 d. After vernalization, seedlings were transplanted to 10-cm-diameter plastic pots, one seedling per pot, and placed on benches in a greenhouse in a randomized complete block design, with six replications in 2003 and five replications in 2004 in which the experimental units were single plants. Day length was 10 h for 14 d (22/20°C day/night), then increased to 12 h for 7 d (28/24°C day/night), and increased to 16 h (28/24°C day/night) until maturity. Plants were well watered as needed, and were fertilized with Miracle-Gro (Miracle-Gro Corp., Marysville, OH) weekly. The S. nodorum inoculum was prepared from a field isolate collected from infected spikes of the susceptible cultivar Caldwell at Lafayette, IN, and cultured on V-8 medium. The primary spike of each plant was mist inoculated after spike emergence (Feekes 10.3) with approximately 1 mmol/L of inoculum containing 1.6 x 106 conidia/mL deionized H2O using a Solo Spraystar 460 (Solo Corp., Newport News, VA) hand sprayer. Inoculated spikes were covered with a 5- by 10-cm plastic bag for 48 h. After inoculation, the greenhouse benches were flooded to 2-cm depth during the evening to increase humidity and enhance SNB disease development. Disease severity of the single inoculated spike of each plant was assessed as the percentage of infected glume tissue 21 d after inoculation.
Disease Screening in Australia
Seeds of the parents and 239 RI lines were germinated for 24 h in petri dishes lined with moist filter paper and vernalized for 8 wk at 3°C in a cold room with 8 h of light per day. They were then incubated at 10°C for 48 h before transplanting to the field in three blocks.
The three replications of the experiment were grown in an irrigated field nursery in a split-plot design at South Perth, Western Australia, during 2004. The plots were split into control and treatment subplots and replicated three times to compare treatment effects. Infection was established at full spike emergence (Feekes 10.3) by spraying spikes to runoff with a conidial suspension of 10 isolates composited of S. nodorum (106 conidia/mL with 0.5% gelatin) produced from grain cultures (Fried, 1989) obtained from the culture collection maintained by the Department of Agriculture and Food Western Australia. High humidity was created by watering the site just before inoculation and covering individual subplots with plastic bags secured over polyvinyl chloride rings (15 cm high, 30-cm diameter) for 48 h after inoculation. Before covering the plots, plastic bags were misted internally with water. Inoculated plants were shaded from direct sunlight during moist incubation by further covering the plastic bags with shade cloth bags (84–90% cover factor). The percentage of glume infection was scored 340°C thermal days after inoculation on individual plants and then averaged for each row. Rating scales for SNB assessment followed James (1971).
Data Assessment and Analysis
The SNB disease severity data were tested for assumptions and variances were homogeneous. These data significantly violated the normality assumption (P < 0.010); appropriate transformations were evaluated but did not significantly improve the normality of these data. Untransformed data were used for ANOVA using the SAS GLM procedure (Version 8.2, SAS Institute, Cary, NC) and QTL analysis. Homogeneity of experimental location variances was analyzed using Bartlett's test for equality of variances (Little and Hills, 1978). The Evansville and Vincennes locations were similar (Evansville is 70 km south of Vincennes), disease severity at Evansville and Vincennes was homogeneous, and QTL analysis results were similar across locations; the mean data values were pooled for QTL analysis and presented as southern Indiana. The disease severity variances among the remaining tests were heterogeneous and not pooled, thus QTL analysis was conducted on five independent experiments. Differences between entries at each test location were determined using Fisher's LSD at 
= 0.05. Narrow-sense heritability estimates of the genotypic means and corresponding confidence interval estimates were calculated according to Knapp et al. (1985).
Genotypic Analysis
Genetic Map Construction and Quantitative Trait Locus Analysis
A genetic linkage map of the RI population was constructed based on markers polymorphic to the parents, P91193D1 and P92201D5, and their distribution across the wheat genome. A total of 382 SSR and DArT (Jaccoud et al., 2001; Semagn et al., 2006) markers was included and the linkage map constructed using Mapmanager QTXb20 (Manly et al., 2001) using a linkage criterion of P = 0.001. A Kosambi function was used to calculate genetic distances from recombination fractions. The "Ripple" command was used to find the best order of markers within linkage groups. Linked markers had a minimum logarithm of odds (LOD) score of 2.4 and their order was confirmed using consensus genetic maps (USDA-ARS, 1999) and bin map locations to specific regions of the wheat genome (Sourdille et al., 2004). Fifty-two markers were not assigned to linkage groups and were removed from the mapping data set. There were 42 occasions where at least two markers were completely linked as a haplotype. The haplotypes were merged to a single marker locus for mapping and QTL analysis. Thus, the genetic map for QTL analysis was constructed from 274 markers. The genetic framework map comprised an average of 13 markers per chromosome with an average of 10.4 centimorgans (cM) between marker loci. The genetic map distance was 2859 cM with 1025.6, 1090.8, and 744.6 cM in the A, B, and D genomes, respectively. Twenty-seven markers exhibited significant (0.05 
P > 0.01) segregation distortion, and 116 were highly significant (P 
0.01).
The disease severity data from each location were used in a whole-genome survey for identifying QTLs. Composite interval mapping (CIM; Zeng, 1994) Model 1 of Windows QTL Cartographer Version 2.5 was used with conditional settings of 10-cM control intervals, five control markers (determined by QTL Cartographer to account for the genetic background variation), and forward regression (Wang et al., 2005). The QTLs for plant height and maturity from the Lafayette, IN, and South Perth, Australia, locations were also evaluated for possible co-location of QTLs for SNB glume resistance. Experimentwise critical thresholds for significance of potential QTLs at each location were determined using CIM to conduct permutation tests as described by Churchill and Doerge (1994). Significance at the = 0.05 level was determined from 1000 permutations. Highly significant QTL thresholds at = 0.01 level were determined from 10,000 permutations.
The QTLs that were identified from the CIM analysis were used as the preliminary model for multiple interval mapping (MIM) in Windows QTL Cartographer 2.5 according to the protocol outlined in Robertson-Hoyt et al. (2006). Single markers that appeared to be closest to the QTLs identified by MIM were tested for QTL x environment interactions using PROC GLM in SAS (Version 8.2) according to the model: Yijk = µ + Ei + Lj + G(L)jk + LEjk + 
ijk, where µ = overall mean, Ei = effect of environment i, Lj = effect of marker locus j, G(L)jk = effect of line k within locus j, LEjk = the interaction of locus j with environment i, and ijk = error.
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
Quantitative Trait Locus Analysis
The QTL QSng.pur-2DL.1 for glume resistance was identified on chromosome 2DL in the region containing the molecular marker interval Xgwm526.1–Xcfd50.2. The QTL was identified at all experimental locations (Fig. 2
) using CIM (Table 3
) and was highly significant ( = 0.01) in all experiments. The LOD peak of this QTL is distal to marker Xgwm526.1. The phenotypic variation explained by this QTL ranged from 12.3% at southern Indiana to 38.1% at South Perth. The same QTL was also identified in the greenhouse experiments, explaining 17.4 and 20.7% of the phenotypic variation during the 2003 and 2004 evaluations, respectively. The additive effect of this QTL ranged from 4.4% at southern Indiana to 6.9% in the 2004 greenhouse experiment.
|
|
|
= 0.05) in the 2004 greenhouse, with a LOD score of 3.6. This QTL appears to have been most effective in the South Perth and southern Indiana experiments, explaining 10.2 and 11.2% of the phenotypic variation, respectively, at this locus. The average disease score of the RI lines with both QSng.pur-2DL.1 and QSng.pur-2DL.2 was 34.3%, compared with 51.1% (LSD0.05 of 2.5%) for RI lines having both alternative alleles. Combined, the QTLs accounted for 33.5% of the phenotypic variation. The QTLs for SNB resistance in the glumes were tested for epistatic interactions and QTL parameters were estimated simultaneously. No epistatic interactions among QTLs were detected. Simultaneously estimated additive effects and effects of allelic substitutions were consistent with the parameters identified using CIM (Table 3). Quantitative trait locus x environment interactions were not significant.
Two chromosomal regions associated with plant height were identified in the South Perth experiment (Table 3). The region associated with the greatest phenotypic effect was on chromosome 2B in the marker region wPt4701–Xgwm630 and was inherited from the P91193D1 parent having an additive effect of 5.6 cm. The second chromosomal region associated with plant height was also inherited from the P91193D1 parent and had an additive effect of 2.2 cm. This region is located on chromosome 2DL between marker loci Xgwm320 and wPt3757.
Two chromosomal regions associated with late maturity were identified in the South Perth experiment only (Table 3). Neither QTL was co-located with SNB glume resistance. The region associated with the greatest phenotypic effect was on chromosome 2D associated with the markers Xbarc095–wPt6329 and was contributed by P92201D5 having an additive effect of 4.6 d later emergence of the spike. The second chromosomal region associated with later maturity was on chromosome 2A associated with the markers Xgwm512–wPt1224 contributed by P92201D5. That QTL had an additive effect of 4.5 d later emergence of the spike.
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
An important consideration for evaluating glume diseases is morphological characteristics that may have confounding effects on disease assessment. This has previously been reported for septoria tritici blotch (Simón et al., 2004) and Type II resistance to fusarium head blight in barley (Hordeum vulgare L.; Zhu et al., 1999). Due to the effects of significant variation in plant architecture in those studies, it was difficult to determine whether pleiotropic effects of plant height, time to ear emergence, or linkage confounded the identification of genetically resistant and susceptible phenotypes. Therefore, variation in morphological characteristics that may influence disease assessments were evaluated in this study. Genetic resistance could be discriminated from plant maturity in the P9819RB1 population despite their correlation with glume resistance at some test sites; however, correlation of phenotypes and co-location of QTLs for plant height and glume resistance within the marker interval Xgwm320–wPt3757 indicates that either an unidentified height-controlling gene having pleiotropic effects on disease assessment or linkage to a resistance gene exists at this locus. Given that a plant height QTL was detected only at one location whereas a QTL for resistance was detected at most locations, however, it is likely that a minor gene controlling plant height, where its expression is influenced by environmental conditions, is linked to a resistance gene at this QTL.
It is intriguing that the QTL identified by marker locus Xcfd50.3 on chromosome 2DL was detected at locations in the USA and Australia but the significance was reduced in the 2004 and not detected in the 2003 greenhouse environments. It has previously been reported that SNB infection is enhanced in the field due to microclimate effects in the canopy increasing humidity for the pathogen to proliferate (Scharen, 1964; Scott et al., 1982). It is reasonable to assume that the controlled greenhouse environments used in this study did not provide sufficient canopy to induce sufficient spore production or an environment that failed to encourage pycnidiospores to infect glumes by water-splashing effects, providing further evidence for the environmental conditions influencing disease severity and ratings. The fact that QSng.pur-2DL.2 was detected as having a significant effect in the 2004 greenhouse test is a good indication, however, that the QTL is not associated with plant height. Mean SNB disease severity of the RI lines having both QSng.pur-2DL.1 and QSng.pur-2DL.2 had lower disease severity than the lines with QSng.pur-2DL.1 or QSng.pur-2DL.2 alone (35.9, 43.3, and 48.0%, respectively; LSD0.05 = 3.3%). This significant difference between the RI lines with both QTLs compared with the RI lines with one or the other of the two QTLs probably accounts for a large proportion of the transgressive segregation that was observed in this population.
The correlations between the greenhouse and field experiments were low. Factors like extended cloud cover, rain duration and intensity, and wind cannot be accurately simulated in large-scale field testing or greenhouse evaluations. Nevertheless, the QSng.pur-2DL.1 QTL was consistently expressed regardless of field or greenhouse evaluations, indicating the robustness of resistance at this locus. Additionally, the QSng.pur-2DL.2 QTL was detected in all but the 2003 greenhouse evaluation. The fact that these QTLs have been identified in very different regions gives credence to their durability.
Given the previous reports of S. nodorum populations having genetic variability (Krupinsky et al., 1972; Murphy et al., 2000a; McDonald and Linde, 2002) and the environmental influence on disease development (Wainshilbaum and Lipps, 1991), the objective of this study was to identify whether resistance QTLs are effective in very different environmental conditions. This is the first report of a QTL study using the same genetic population in different countries, providing valuable information on the durability of SNB glume resistance. By evaluating the RI population at several locations, including two continents, we took advantage of the probable variation among the S. nodorum populations to evaluate the effects of favorable genetic combinations in individuals of the population and extreme environmental conditions on SNB QTL durability. Our results indicate that the virulence of pathogen populations was similar at most locations, supporting the ideas of Keller et al. (1997) that population differentiation is not expected to be as strong in systems of quantitative resistance as in systems with qualitative resistance. Phenotypic correlations between locations were moderate, which indicates that allelic differences were detected in different locations, supported by the identification of QTLs for resistance in one location but not in others; however, we elected to not include data of QTLs that were not consistently identified at multiple locations. The QTLs not consistently identified at all locations require further investigation. The QTLs for glume resistance identified on chromosome 2DL and their significant effects across very different environments on two continents indicates that these QTLs are potentially durable for wheat breeding. Therefore, collaborative evaluations in very different environments similar to this study provides a means whereby durable QTLs can be identified for deploying resistance in international wheat improvement programs to develop wheat cultivars with enhanced SNB glume resistance.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Received for publication November 22, 2006.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. L. Friesen, Z. Zhang, P. S. Solomon, R. P. Oliver, and J. D. Faris Characterization of the Interaction of a Novel Stagonospora nodorum Host-Selective Toxin with a Wheat Susceptibility Gene Plant Physiology, February 1, 2008; 146(2): 682 - 693. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
