Quantitative Trait Loci Analysis of Powdery Mildew Disease Resistance in the Arabidopsis thaliana Accession Kashmir-1

Corresponding author: Shauna C. Somerville, Department of Plant Biology, Carnegie Institution of Washington, 260 Panama St., Stanford, CA 94305., shauna{at}andrew2.stanford.edu (E-mail)

Communicating editor: A. H. D. BROWN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Powdery mildew diseases are economically important diseases, caused by obligate biotrophic fungi of the Erysiphales. To understand the complex inheritance of resistance to the powdery mildew disease in the model plant Arabidopsis thaliana, quantitative trait loci analysis was performed using a set of recombinant inbred lines derived from a cross between the resistant accession Kashmir-1 and the susceptible accession Columbia glabrous1. We identified and mapped three independent powdery mildew quantitative disease resistance loci, which act additively to confer disease resistance. The locus with the strongest effect on resistance was mapped to a 500-kbp interval on chromosome III.

ALTHOUGH commonly encountered, genetically complex or polygenic resistance that consists of multiple resistance loci against one pathogen race is poorly understood. This is primarily a result of the historic difficulty in studying genetically complex traits. However, with the inception of molecular DNA markers and quantitative trait locus (QTL) mapping, dissecting polygenic forms of disease resistance into component loci is now feasible (TANKSLEY 1993 ). Numerous studies to quantify and identify QTL affecting pathogen resistance in many crop species have been performed (for reviews see MICHELMORE 1995 ; YOUNG 1996 ; KOVER and CAICEDO 2001 ).

The small crucifer Arabidopsis thaliana (L.) Heynh. provides several unique advantages for cloning and characterizing plant disease resistance genes. Arabidopsis is a host for all types of phytopathogens including bacteria, fungi, viruses, plant parasites, and nematodes (KUNKEL 1996 ; BUELL 1998 ). This, combined with the multitude of publicly available molecular tools, including a complete genome sequence (ARABIDOPSIS GENOME INITIATIVE 2000), means that the cloning of disease resistance genes can proceed more quickly in Arabidopsis than in other plant species. Although there have been several QTL studies of natural variation in Arabidopsis (for a review see ALONSO-BLANCO and KOORNNEEF 2000 ), QTL analysis has not yet been applied to pathogen resistance in this model plant species.

The obligate pathogenic fungi belonging to the Erysiphales (Ascomycetes) are the causal agents of powdery mildew diseases (BRAUN 1987 ). Among susceptible hosts are several species of economic importance, including barley, wheat, pea, grape, cabbage, and tomato, and various ornamental bushes. Resistance to powdery mildew has been studied in many plant species with both mono- and oligogenic resistance reported (LUNDQVIST et al. 1991 ; LOHNES and BERNARD 1992 ; REDDY et al. 1994 ). Arabidopsis is also a host for powdery mildew with four isolates from three species reported to infect this plant (KOCH and SLUSARENKO 1990 ; ADAM and SOMERVILLE 1996 ; XIAO et al. 1997 ; PLOTNIKOVA et al. 1998 ). Mutational studies identified several Arabidopsis mutants that show significant qualitative resistance to this pathogen (FRYE and INNES 1998 ; VOGEL and SOMERVILLE 2000 ). A complementary approach is to study natural resistance in different accessions of Arabidopsis. At present, a minimum of eight loci controlling natural resistance to powdery mildew (designated RPW for recognition of powdery mildew) have been described, including both mono- and digenic resistance conferred by semidominant or recessive disease resistance genes (ADAM and SOMERVILLE 1996 ; XIAO et al. 1997 ). Recently two genes conferring resistance to powdery mildew were cloned from Arabidopsis, neither of which belongs to the abundant class of disease resistance genes with nucleotide binding site and leucine-rich repeat motifs (HAMMOND-KOSACK and JONES 1997 ). One encodes two related small novel proteins, designated RPW8.1 and RPW8.2 (XIAO et al. 2001 ), and the second encodes EDR1, a map kinase kinase kinase (FRYE et al. 2001 ).

Of the 93 Arabidopsis accessions found to be resistant to the powdery mildew Erysiphe cichoracearum UCSC1 (Salmon) in a previous study, Kashmir-1 (Kas-1) was highly resistant (ADAM and SOMERVILLE 1996 ; ADAM et al. 1999 ). Using QTL analysis, the genetic basis for the variation in powdery mildew resistance in a set of recombinant inbred lines (RIL) derived from a cross between Kas-1 and the susceptible Arabidopsis accession Columbia glabrous (Col-gl1) was found to be complex. Here, we describe the identification and mapping of three powdery mildew quantitative resistance loci and the genetic fine mapping of the major powdery mildew resistance QTL identified in this cross.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plant materials:
The seeds for Col-gl1 were obtained from Dr. C. R. Somerville (Carnegie Institution of Washington). The accession Kas-1 was originally obtained from the Arabidopsis Information Service seed bank (Frankfurt, Germany) and was taken through a generation of single seed descent to enhance genetic uniformity. Seeds from a single F₁ (Col-gl1 x Kas-1) plant of a cross, in which Col-gl1was used as the female, were grown and self fertilized. Three hundred F₂ plants from the (Col-gl1 x Kas-1) cross were self fertilized and advanced to the F₆ generation by single seed descent. A random sampling of 129 RIL was chosen from the 300 for genetic analysis. Seeds for these 129 F₆ (Col-gl1 x Kas-1) RIL are deposited at the Arabidopsis Biological Resource Center (Columbus, OH).

Plant growth procedures:
Seeds were sown in commercial compost (Pro-Mix HP; Premier) in 6-inch square "jumbo square" pots (Belden Plastics) and fertilized with Peters fertilizer 20-20-20 (NPK). The pots were covered with plastic lids to ensure high humidity and incubated at 4° for 48 hr. The pots were transferred to a controlled environment with a 14-hr photoperiod (100–150 µE/m^-2/sec^-1 in the 400- to 700-nm range) at 22°. After germination, the plastic lids were removed. The plants were thinned 10–12 days after germination to give an average of 20 plants, uniformly distributed per pot. The plants were grown for 3 weeks (until the third and fourth leaves were fully expanded) before inoculation.

Inoculation procedures:
E. cichoracearum isolate UCSC1 was used throughout this study as a source of inoculum (ADAM and SOMERVILLE 1996 ). Pure cultures of the isolate were created by repeated single colony subculturing. E. cichoracearum UCSC1 was grown and maintained on the secondary host plant squash (Cucurbita maxima cv Kuta; Park Seed, Greenwood, SC). Three-week-old squash plants were inoculated with the powdery mildew fungus and grown in a phytocell at 22°. Heavily infected squash leaves 10–11 days postinoculation (dpi) were used as the source of fungal inoculum. Inoculation of the Arabidopsis plants was performed by dusting conidia from squash leaves in a draft-free environment at a height of 1 m above the plants to achieve an even distribution of conidia. Inoculation densities were >100 conidia/mm². After inoculation, plants were moved to a humidified chamber (100% relative humidity, 22°) for 1 hr to stimulate spore germination. Plants were then moved back to normal growth conditions and were evaluated for their disease reaction 7 dpi.

Evaluation of disease reaction phenotypes:
The disease reaction (DR) phenotype scores were evaluated by visual inspection of the level of fungal growth observed on the adaxial surface of the inoculated third and fourth leaves of each individual plant. The scale of notation for the DR score ranged from 0, for resistant plants similar to Kas-1, to 4 for susceptible plants similar to Col-gl1. The DR score notation was as follows: 0, no fungal growth visible to the naked eye; 1, a limited amount of fungal hyphal growth (<25% leaf coverage) with no conidiophores; 2, an intermediate density of fungal growth (50% leaf coverage) accompanied by small restricted patches with conidiophores; 3, >75% leaf coverage by fungal growth with some conidiophores; 4, total coverage of the leaf by the fungi with abundant conidiophores conferring a "powdery" appearance.

Approximately 20 plants per RIL were grown with two Col-gl1 plants as positive controls for inoculation in the same pot. Plants were observed for disease symptoms 7 dpi. Data were recorded only from pots in which the two control Col-gl1plants possessed a highly susceptible phenotype (i.e., a DR score of 4). Each RIL was scored in this manner on two separate occasions. Each plant within the RIL was scored individually for fungal growth and the average DR score for 40 plants was then used as the DR score for that RIL.

Microscopy:
To ascertain whether microscopic lesions occurred on inoculated leaves, dead cells were stained with a lacto-phenol solution [250 µg/ml trypan blue in phenol, lactic acid, glycerol, and water (1:1:1:1)]. Leaves from Col-gl1 and Kas-1 plants were collected in tubes and vacuum-infiltrated twice in the lacto-phenol solution. Then the tubes were placed in a boiling water bath for 2 min and allowed to cool for 1 hr. The leaves were destained in the 1:1:1:1 solution for 1 hr and examined under bright field illumination (VOGEL and SOMERVILLE 2000 ). Observations of the third and fourth leaves were made at different time points from 6 hr postinoculation to 7 dpi.

The development of the fungus was assayed by the measurement of the total hyphal length of 10 colonies developing on the third and fourth leaves of both the parental lines at 1, 2, 3, and 4 dpi. This experiment was repeated three times. Very lightly inoculated leaves (10 conidia per leaf) were harvested and cleared in 95% ethanol. Then the hyphae were stained with 250 µg/ml trypan blue in a solution of lactic acid, glycerol, and water (1:1:1) for 15 min, rinsed in the same solution, and mounted (ADAM and SOMERVILLE 1996 ). The leaves were observed with a Leica microscope and individual colonies were photographed using a digital camera (Pixera, Los Gatos, CA). The total hyphal length per colony was measured and calculated with NIH IMAGE software (http://rsb.info.nih.gov/nih-image/).

RIL map generation:
For each F₆ line, small-scale DNA preparations from 30 to 50 seedlings grown in the greenhouse were performed as described by DELLAPORTA et al. 1983 . This method yielded 3–5 µg of genomic DNA per 0.5 g of fresh weight of tissue. To generate a map on the basis of the 129 F₆ (Col-gl1 x Kas-1) RIL, codominant polymerase chain reaction (PCR)-based molecular markers that were previously mapped in crosses between Columbia and Landsberg erecta (Ler) were screened using DNA from the parental plants (LISTER and DEAN 1993 ; BELL and ECKER 1994 ). DNA markers were tested for polymorphisms and informative markers that were evenly distributed roughly every 20 cM were selected (DARVASI and SOLLER 1994 ; YOUNG 1996 ). Seventeen simple sequence length polymorphism (SSLP) markers, 8 cleaved amplified polymorphic sequences (CAPS) markers, and the phenotypic marker GLABROUS1 were used in the generation of the map (Table 1). Only one marker, CD3-69(PCR), was specifically generated for this study (Table 2). Amplification conditions for PCR were described by BELL and ECKER 1994 for SSLP markers and by KONIECZNY and AUSUBEL 1993 for CAPS markers.

The RIL map was generated from the data obtained from the 129 F₆ lines using MAPMAKER version 3.0 computer software program for the PC (LANDER et al. 1987 ). The program option "RI lines obtained by selfing" was used to analyze the data. The two-point analysis command "group" (LOD = 3, maximum distance between markers = 40 cM) was first adopted to define the different linkage groups to which the molecular markers belonged. Multipoint analysis was then used with an error estimate of 1%. The best position for the unlinked markers relative to the determined order was assessed using the command "try." To test the map for ambiguity of neighboring markers, the command "ripple" was used. Finally, the best order for the markers in each linkage group was chosen using the option "compare." The recombination frequencies were converted to map distances in centimorgans with the Kosambi function (KOSAMBI 1944 ).

Genetic fine mapping of RPW10:
To genetically fine map the RPW10 locus, 120 plants derived from the F₆ RIL CK48 were infected and scored for powdery mildew disease as described above. The phenotype of the individual plants was confirmed in the next generation by infecting F₇ plants derived from the selfed F₆ plants. DNA was isolated from individual F₆ plants using the small scale DNA preparation method of EDWARDS et al. 1991 . New markers generated within the T04109-nga6 interval can be found in Table 2. The amplification conditions used were those of KONIECZNY and AUSUBEL 1993 for CAPS markers.

Statistical and QTL analyses:
The software program MQTL (TINKER and MATHER 1995A , TINKER and MATHER 1995B ) version 0.98 was used to look for QTL for resistance to powdery mildew in the RIL. This program uses the least-square methods (HALEY and KNOTT 1992 ) plus partial regression coefficients from background markers to control genetic variance from nontarget QTL. Only one environment and one trait, the DR score, were analyzed. The map distances between the markers generated by MAPMAKER were entered in MQTL with all markers being considered as background markers. Heterozygotes were treated as missing data points. The data were analyzed both with simple interval mapping (SIM) and simplified composite interval mapping (sCIM; TINKER and MATHER 1995A , TINKER and MATHER 1995B ). The thresholds for SIM and sCIM main effects were estimated by the software program with repeated "shuffling" of the whole data set using 10,000 random permutations, as advised by CHURCHILL and DOERGE 1994 , with a first error rate = 5% (BEAVIS 1998 ).

Test statistic values for SIM and for sCIM were calculated every 1 cM throughout the genome. To find the precise positions of the peaks, the command "find peak" was used. For all these locations, a confidence interval was fixed where the QTL had a 95% of chance of being present. We defined this confidence interval by the "1-LOD" support interval (LANDER and BOTSTEIN 1989 ). The total percentage of phenotypic variation (R_T²) was estimated with the option "make estimates." R_T² is the variance explained by the QTL detected divided by the total phenotypic variance of the DR score (TINKER and MATHER 1995B ). The percentage of phenotypic variation due to individual QTL (R²) can be related to the test statistic for SIM with this formula: R² = 1 - 1/exp(TS/n), with n = number of progeny (TINKER and MATHER 1995B ). Potential epistatic effects between a QTL and all the other points of the genome, including the other QTL, were also tested with MQTL.

The results of the QTL analysis obtained with MQTL were verified by analysis of variance (ANOVA) with the Statview 4.5 (ABACUS Concepts, Berkeley, CA) software program. Every marker was tested for the presence of a QTL using simple factor ANOVA (P < 0.05) and for its effects (R² = sum of squares explained by the marker divided by total sum of squares). All markers were further analyzed in pairs with a two-factor ANOVA to test for epistasis effects between every pair of markers throughout the genome and between the QTL, using the marker closest to the QTL and the rest of the markers (P < 0.005).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

DR phenotypes of Col-gl1, Kas-1, and the RIL:
Col-gl1 is susceptible to the powdery mildew pathogen E. cichoracearum UCSC1 (Fig 1). Diseased plants are characterized by the development of white powdery-like fungal growth on the surface of the leaf 7 dpi. Inoculated leaves of Kas-1, however, show little or no sign of fungal growth. Necrotic and chlorotic flecks developed 5–6 dpi on infected Kas-1 leaves. No difference in germination rates of the fungus on the different accessions was observed (data not shown). Hyphal lengths per colony at 2 dpi (P < 0.05), 3 dpi (P < 0.01), and 4 dpi (P < 0.01) were significantly shorter on Kas-1 than Col-gl1 (Fig 2). By 4 dpi, colonies on Kas-1 were only 45% the size of those on Col-gl1. The cell death observed with the naked eye in Kas-1 5–6 dpi was detected as early as 9 hr postinoculation (as observed by trypan blue staining) in Kas-1 but was absent in Col-gl1. Cell death, however, was infrequent (<5% of infection sites) and in many cases did not appear to result in the death of the fungal colony (data not shown).

View larger version (37K): In this window In a new window Download PPT slide   Figure 1. Parental phenotypes Col-gl1 and Kas-1, 7 dpi with Erysiphe cichoracearum UCSC1. Col-gl1 is susceptible and is characterized by total coverage of inoculated leaves by hyphae and abundant conidiophores, whereas Kas-1 is resistant and exhibits no fungal growth, although necrotic flecks, which develop late (5–6 dpi), can be observed.

View larger version (26K): In this window In a new window Download PPT slide   Figure 2. Hyphal length. Total hyphal length of individual powdery mildew colonies measured on Col-gl1 (solid circles) and Kas-1 leaves (open circles) 1, 2, 3, and 4 dpi. Each data point corresponds to the average hyphal length per colony (n = 30) with the SD.

On the basis of extensive examination of the inheritance of powdery mildew resistance in F₃ (Col-gl1 x Kas-1) lines, the inheritance of disease resistance appeared to be complex (data not shown). In agreement with this conclusion, the distribution of the DR scores for the 129 RIL showed a multimodal distribution (Fig 3) with a mean DR score of 2.2 and a standard deviation (SD) of 1.3. This distribution is significantly different from the bimodal distribution expected if resistance was conferred by a single resistance gene, indicating that the inheritance of the resistance to powdery mildew in this cross is complex.

View larger version (75K): In this window In a new window Download PPT slide   Figure 3. Frequency distribution of DR scores of the 129 F6 (Col-gl1 x Kas-1) RIL. The DR score presented for each RIL is the average of the DR scores observed on 40 plants. The parental DR scores are 0 for Kas-1 (resistant) and 4 for Col-gl1 (susceptible).

Map construction:
A total of 26 markers, consisting of 17 SSLPs, 8 CAPS, and GL1, were analyzed (Table 1). The segregation data of these markers were used to obtain the linkage map shown in Fig 4. The raw segregation data can be obtained upon request or can be viewed at http://carnegiedpb.stanford.edu/shauna/kas.dataprint.htm and http://www.genetics.org/supplemental/. The 26 markers were assigned to five linkage groups with a total length of 378.3 cM and an average spacing of 14.6 cM (SD = 6.1 cM). The longest distance between markers was 33.2 cM, between Z30817 and GL1. The genetic length of each linkage group was comparable to the lengths reported for other mapping populations and all markers were located and ordered to similar positions as on previous Arabidopsis maps (LISTER and DEAN 1993 ; ALONSO-BLANCO et al. 1998 ).

View larger version (20K): In this window In a new window Download PPT slide   Figure 4. Genetic map of the Col-gl1 x Kas-1 cross. Twenty-five molecular markers (17 SSLP and 8 CAPS) and one morphological marker were mapped in 129 RIL and the map was generated with MAPMAKER (LANDER et al. 1987 ). The distances are displayed in centimorgans. The placement of the first and the last marker on every chromosome was done according to the distances available from the F8 (Ler x Col) RI map (LISTER and DEAN 1993 ). Significant deviations from the expected 1:1 distribution of Kas-1 and Col-gl1 alleles at each marker are indicated as follows: **, Col-gl1 alleles in excess (P < 0.01); ° °, P < 0.01; and °, P < 0.05, Kas-1 alleles in excess. Positions of the RPW loci are in boxes.

Analysis of the RIL:
The segregation ratio of the two homozygous classes at each marker was tested for the 1:1 expected proportion and the markers significantly distorted in their segregation ratios are indicated in Fig 4. It is clear that the segregation distortion occurred for a large portion of the genome. The single largest distortion of the segregation ratio was observed for nga8 (top of chromosome IV), where Col-gl1 alleles were 5.1-fold more abundant than Kas-1 alleles in the 129 RIL. The remaining distortions were of the ratio of 1:2.7 or lower.

RIL are expected to be essentially homozygous by the F₆ generation with the theoretical chance of a heterozygous RIL for a given marker being 3.1%. Analysis of the segregation data revealed a much higher rate of 9.1%, indicating that some bias for the selection of heterozygous plants via the method of single seed descent had occurred.

QTL analysis:
The software program MQTL was used to analyze the segregation data for the DR scores and molecular markers (TINKER and MATHER 1995A , TINKER and MATHER 1995B ). Graphs of the test statistics for SIM and sCIM main are shown in Fig 5. The thresholds computed were 10.3 for SIM and 26.7 for sCIM.

View larger version (17K): In this window In a new window Download PPT slide   Figure 5. QTL likelihood map for resistance to powdery mildew in the F6 (Col-gl1 x Kas-1) RIL. The abscissa corresponds to the genetic map in centimorgans; the limits of each chromosome are indicated by vertical dotted lines (ChrI to ChrV). The ordinal corresponds to the test statistics for SIM (simple interval mapping) and sCIM (simplified composite interval mapping) main effects generated by MQTL (TINKER and MATHER 1995A ). The test statistics are represented by a gray solid line for SIM and a black solid line for sCIM. The grey broken line indicates the treshold for SIM test statistics and black broken line indicates the threshold for sCIM test statistics. Three QTL were detected: RPW12 on chromosome II, RPW10 on chromosome III and RPW11 on chromosome V. The location of each QTL is symbolized by a horizontal black bar (), as determined form the "1-LOD" support interval (LANDER and BOTSTEIN 1989 ).

Three unlinked regions in the Arabidopsis genome were found to have test statistic values for both SIM and sCIM higher than the corresponding thresholds and so were designated as a QTL. Resistance alleles for all powdery mildew resistance loci were derived from the resistant parent Kas-1. RPW10 was mapped on the bottom of chromosome III to a confidence interval of only 6 cM. The second QTL, designated RPW11, occurred near the marker nga139 on the top of chromosome V with a confidence interval spanning 12 cM. The region spanning the lower part of chromosome II exceeded the SIM threshold. However, the curve representing the test statistic for sCIM refined the position of this QTL to one peak centered near nga1126. The confidence interval in which the QTL lies was localized to an 11-cM region (LANDER and BOTSTEIN 1989 ). This QTL was designated as RPW12 (Table 3). To further confirm the authenticity of the QTL identified by the program MQTL, the entire data set was reanalyzed by one-factor ANOVA. All three QTL were confirmed by ANOVA at the 95% confidence level (data not shown).

The percentage of the total phenotypic variation (R_T²) explained by the three QTL was calculated at 63.0% (Table 3). RPW10 accounted for 45.0% of the total variation, RPW11 for 17.6%, and RPW12 for 10.4% (Table 3). The percentages of the total phenotypic variation explained by the individual QTL using ANOVA were similar to those obtained using MQTL (data not shown). Substitution of RPW10 susceptibility alleles for resistance alleles had the greatest impact on the DR score (i.e., 1.7 DR score points explained) compared to RPW11 and RPW12 (Table 3).

No epistatic effects were demonstrated by either MQTL or ANOVA. Therefore, all three QTL identified are additive in their effects on powdery mildew resistance.

Fine mapping of RPW10:
As previously mentioned the number of heterozygotes observed in the F₆ (Col-gl1 x Kas-1) RIL was approximately three times higher than expected. During scoring for powdery mildew resistance, a few RIL were observed to segregate for disease resistance in a simple Mendelian fashion. Examination of the genetic profile of several of these lines revealed that they were heterozygous for markers on the lower arm of chromosome III, including the marker R30025, which is near RPW10. As the line CK48 was homozygous for the Col-gl1 alleles at both RPW11 and RPW12, this line was likely to be segregating for only the RPW10 QTL. To confirm that this was the case, 120 plants from the F₆ line CK48 were planted and scored for powdery mildew resistance. Both resistant (DR score = 1) and susceptible plants (DR score = 4) as well as an intermediate class of plants (DR score = 2 or 3) were observed. These 120 plants were selfed and the resistance phenotypes of these plants confirmed in the F₇ generation. The segregation results obtained were compatible with a segregation ratio of resistant homozygote:heterozygote:susceptible homozygote of 1:2:1 (CK48: ² = 2.2, 0.5 < P < 0.25). Thus, in CK48, resistance is conferred by a single locus with the resistance allele being semidominant. To further define the location of RPW10, 120 CK48 lines were genotyped using the CAPS marker R30025 and the SSLP marker nga6. The marker T04109 was homozygous in line CK48. Results from these markers indicated the RPW10 locus is located between R30025 and T04109. The location of this locus was further refined using codominant PCR-based markers (Table 2). The RPW10 locus was genetically mapped to a 4-cM interval defined by the markers M005-S and CIC8E1RE, which is 500 kbp in size (Fig 6).

View larger version (21K): In this window In a new window Download PPT slide   Figure 6. Genetic mapping of the RPW10 locus. (A) The 37-cM region of chromosome III identified by QTL analysis as containing the powdery mildew disease resistance locus RPW10. The confidence interval for RPW10, using the "1-LOD" support interval (LANDER and BOTSTEIN 1989 ), is denoted by an open rectangle. (B) The location of RPW10 as determined by genetic mapping. A physical map of the region was constructed from published bacterial artificial chromosome and yeast artificial chromosome physical maps (CAMILLERI et al. 1998 ; SATO et al. 1998 ).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this study, we demonstrated that powdery mildew resistance in the Arabidopsis accession Kas-1 is inherited in a polygenic fashion. Using QTL analysis of a set of RIL, three unlinked resistance QTL were identified and, for each QTL, the resistance alleles were derived from Kas-1. The QTL, designated RPW10, RPW11, and RPW12, were found to act additively to confer resistance. The demonstration that RPW10 is allelic to the recently cloned gene RPW8 provides additional confirmation of the validity of RPW10 (XIAO et al. 2001 ). A genetic model based on all three QTL explained 63% of the total variation in powdery mildew resistance observed. The values for the percentage of variation explained by individual QTL and by the combined three-QTL model from these studies compare favorably to the values recorded in other studies dealing with QTL for pathogen resistance. For example, the percentage of total phenotypic variation explained ranged from 14 to 81%, with a mean of 50% in studies summarized in YOUNG 1996 .

The inheritance of powdery mildew resistance in the accession Kas-1 was studied preliminarily by ADAM and SOMERVILLE 1996 . It was reported that resistance to E. cichoracearum UCSC1 was conferred by a single locus, designated RPW1, based on a population of F₂ plants and genetic mapping data from a relatively small number (n = 54) of F₃ lines derived from a cross between Col-gl1 and Kas-1. One possible explanation for the discrepancy between the former and current studies is that Arabidopsis plants in the study by ADAM and SOMERVILLE 1996 were inoculated with a 10-fold lower density (5–10 conidia per mM²), which often led to patchy infections and made the scoring of disease symptoms difficult. Also, because the F₂ (Col-gl1 x Kas-1) population seemed to fit a simple model of resistance, DR scores derived from the F₃ families were rigidly categorized into three genotypic classes (i.e., homozygous resistant, homozygous susceptible, or heterozygous) to reflect a single gene model. These two factors, combined possibly with the segregation distortion observed for the bottom of chromosome II, resulted in the incorrect assignment of the powdery mildew resistance genotype for a number of F₃ families, which led to an incorrect understanding of the inheritance of powdery mildew resistance. Due to the different inoculation conditions and populations used in these two studies, all resistance QTL identified in this study were given new RPW designations. However, we cannot exclude the possibility that RPW1 may correspond to one of the RPW loci identified in this study.

The set of F₆ (Col-gl1 x Kas-1) RIL represents the seventh set of Arabidopsis RIL generated and the fourth set that has been characterized in detail genetically (REITER et al. 1992 ; LISTER and DEAN 1993 ; HOLUB and BEYNON 1997 ; ALONSO-BLANCO et al. 1998 ; DESLANDES et al. 1998 ; C. L. SCHIFF, I. W. WILSON and S. C. SOMERVILLE, personal communication). Although significant distortions in the segregation ratios of markers were detected in large portions of the genome, the magnitude and extent of the distortions were comparable to those observed for the other genetically characterized RIL (REITER et al. 1992 ; LISTER and DEAN 1993 ; ALONSO-BLANCO et al. 1998 ; C. L. SCHIFF, I. W. WILSON and S. C. SOMERVILLE, personal communication). For the telomeric region of the lower arm of chromosome II, the distortion is present in three sets of RIL. The origin and possible sources of this segregation distortion in the F₆ (Col-gl1 x Kas-1) RIL are presently unknown and present an interesting avenue for understanding complex gene interactions. The threefold higher level of heterozygosity observed in our set of RIL may be due to some bias for the involuntary selection of heterozygous plants during single seed descent and is probably a reflection of heterosis (MITCHELL-OLDS 1995 ).

Although most Arabidopsis accessions possess a relatively short life cycle, creating RIL is still a lengthy and laborious process, especially with Kas-1, as this accession requires a vernalization treatment to flower. However, the RIL offer several advantages, such as the possibility of permanently propagating the population without further genotyping and the advantage of studying a trait on several sibling plants per line to minimize the environmental variation (ALONSO-BLANCO and KOORNNEEF 2000 ). Overall, the 129 F₆ (Col-gl1 x Kas-1) RIL constructed in this study will provide useful material for the analysis of other traits that vary between these two accessions (e.g., size, vernalization).

Another major application for RIL is to use the residual heterozygosity of some lines for map-based cloning of QTL using standard map-based cloning methods (TUINSTRA et al. 1997 ; LUKOWITZ et al. 2000 ). In this study, using a single RIL (CK48) that segregated for powdery mildew resistance in a Mendelian fashion, we were able to genetically map the RPW10 locus to an interval of 500 kbp. This region agrees well with the confidence interval calculated by MQTL, confirming the accuracy of our QTL analysis.

The value of QTL analysis in Arabidopsis lies in the ability of this system to address fundamental questions concerning the nature of QTL for disease resistance. Disease resistance QTL have been variously proposed to be weak alleles of race-specific disease resistance genes or to be a class distinct from known disease resistance genes (e.g., genes encoding elements of signal transduction pathways or genes encoding defense response components; PFLIEGER et al. 1999 ). In this regard, the demonstration that RPW10 is allelic to RPW7, which confers resistance to E. cruciferarum (Junnel), lends support to the hypothesis that this locus encodes a broad-spectrum resistance mechanism (XIAO et al. 1997 , XIAO et al. 2001 ). The other feature that distinguishes RPW10 from typical disease resistance genes is that RPW10-mediated resistance does not lead to the arrest of fungal growth at a specific step in the infection nor is this resistance associated with a typical rapid-acting hypersensitive necrosis response. These observations together with the novel nature of the cloned RPW8 genes (allelic to RPW10 and RPW7) support the hypothesis that resistance QTL are distinct from classical race-specific resistance genes (XIAO et al. 2001 ). Additional studies of natural sources of resistance will likely yield new insights into the nature of powdery mildew resistance.

	FOOTNOTES

¹ Present address: C.S.I.R.O., Division of Plant Industry, Canberra 2601, Australia.
² Present address: Northwestern University, Evanston, IL 60208.

	ACKNOWLEDGMENTS

We thank Drs. N. A. Tinker and A. Charcosset for their helpful comments and advice and Eriko Miura and Beverly Jow Fang for their technical assistance. We also thank S. Xiao and J. G. Turner for access to markers and data prior to publication. The financial support of the Carnegie Institution of Washington and the U.S. Department of Energy is gratefully acknowledged. This is publication number 1421 of the Carnegie Institution of Washington.

Manuscript received June 26, 2000; Accepted for publication April 24, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADAM, L. and S. C. SOMERVILLE, 1996  Genetic characterization of five powdery mildew disease resistance loci in Arabidopsis thaliana. Plant J. 9:341-356[Medline].

ADAM, L., S. ELLWOOD, I. WILSON, G. SAENZ, and S. XIAO et al., 1999  Comparison of Erysiphe cichoracearum and E. cruciferarum and a survey of 360 Arabidopsis thaliana accessions for resistance to these two powdery mildew pathogens. Mol. Plant-Microbe Interact. 12:1031-1043[Medline].

ALONSO-BLANCO, C. and M. KOORNNEEF, 2000  Naturally occurring variation in Arabidopsis: an under exploited resource for plant genetics. Trends Plant Sci. 5:22-29[Medline].

ALONSO-BLANCO, C., A. J. M. PEETERS, M. KOORNNEEF, C. LISTER, and C. DEAN et al., 1998  Development of an AFLP based linkage map of Ler, Col and Cvi Arabidopsis thaliana ecotypes and construction of a Ler/Cvi recombinant inbred line population. Plant J. 14:259-271[Medline].

Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. (2000) Nature 408:796-815[Medline].

BEAVIS, W. D., 1998 QTL analysis: power, precision and accuracy, pp. 145–162 in Molecular Dissection of Complex Traits, edited by A. H. PATERSON. CRC Press, Boca Raton, FL.

BELL, C. J. and J. R. ECKER, 1994  Assignment of 30 microsatellite loci to the linkage map of Arabidopsis. Genomics 19:137-144[Medline].

BRAUN, U., 1987 A Monograph of the Erysiphales (Powdery Mildews). J. Cramer, Berlin.

BUELL, C. R., 1998  A weed leading the field of plant-pathogen interactions. Plant Physiol. Biochem. 36:177-186.

CAMILLERI, C., J. LAFLEURIEL, C. MACADRE, F. VAROQUAUX, and Y. PARMENTIER et al., 1998  A YAC contig map of Arabidopsis thaliana chromosome 3. Plant J. 14:633-642[Medline].

CHURCHILL, G. A. and R. W. DOERGE, 1994  Empirical threshold values for quantitative trait mapping. Genetics 138:963-971[Abstract].

DARVASI, A. and M. SOLLER, 1994  Optimum spacing of genetic markers for determining linkage between marker loci and quantitative loci. Theor. Appl. Genet. 89:351-357.

DELLAPORTA, S. L., J. WOOD, and J. B. HICKS, 1983  A plant DNA minipreparation. Version II. Plant Mol. Biol. Rep. 1:19-21.

DESLANDES, L., F. PILEUR, L. LIAUBET, S. CAMUT, and C. CAN et al., 1998  Genetic characterization of RRS1, a recessive locus in Arabidopsis thaliana that confers resistance to the bacterial soilborne pathogen Ralstonia solanacearum. Mol. Plant-Microbe Interact. 11:659-667[Medline].

EDWARDS, K., C. JOHNSTONE, and C. THOMPSON, 1991  A simple and rapid method for the preparation of plant genomic DNA for PCR analysis. Nucleic Acids Res. 19:1349[Free Full Text].

FABRI, C. O. and A. R. SCHAFFNER, 1994  An Arabidopsis thaliana RFLP mapping set to localize mutations to chromosomal regions. Plant J. 5:149-156.

FRYE, C. A. and R. W. INNES, 1998  An Arabidopsis mutant with enhanced resistance to powdery mildew. Plant Cell 10:947-956[Abstract/Free Full Text].

FRYE, C. A., D. Z. TANG, and R. W. INNES, 2001  Negative regulation of defense responses in plants by a conserved MAPKK kinase. Proc. Natl. Acad. Sci. USA 98:373-378[Abstract/Free Full Text].

HALEY, C. S. and S. A. KNOTT, 1992  A simple regression method for mapping quantitative trait loci in line crosses using flanking markers. Heredity 69:315-324[Medline].

HAMMOND-KOSACK, K. E. and J. D. G. JONES, 1997  Plant disease resistance genes. Annu. Rev. Plant Physiol. Mol. Biol. 48:575-607.

HOLUB, E. B. and J. L. BEYNON, 1997  Symbiology of mouse-ear cress (Arabidopsis thaliana) and Oomycetes. Adv. Bot. Res. 24:227-273.

KOCH, E. and A. J. SLUSARENKO, 1990  Fungal pathogens of Arabidopsis thaliana. Bot. Helv. 100:257-268.

KONIECZNY, A. and F. M. AUSUBEL, 1993  A procedure for mapping Arabidopsis mutations using codominant ecotype-specific PCR-based markers. Plant J. 4:403-410[Medline].

KOSAMBI, D. D., 1944  The estimation of map distance from recombination values. Ann. Eugen 12:172-175.

KOVER, P. X. and A. L. CAICEDO, 2001  The genetic architecture of disease resistance in plants and the maintenance of recombination by parasites. Mol. Ecol. 10:1-16[Medline].

KUNKEL, B. N., 1996  A useful weed put to work: genetic analysis of disease resistance in Arabidopsis thaliana. Trends Genet. 12:63-69[Medline].

LANDER, E. S. and D. BOTSTEIN, 1989  Mapping Mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121:185-199[Abstract/Free Full Text].

LANDER, E. S., P. GREEN, J. ABRAHAMSON, A. BARLOW, and A. DALY et al., 1987  MAPMAKER: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1:174-181[Medline].

LISTER, C. and C. DEAN, 1993  Recombinant inbred lines for mapping RFLP and phenotypic markers in Arabidopsis thaliana. Plant J. 4:421-429.

LOHNES, D. G. and R. L. BERNARD, 1992  Inheritance of resistance to powdery mildew in soybeans. Plant Dis. 76:964-965.

LUKOWITZ, W., C. S. GILMORE, and W. R. SCHIEBLE, 2000  Positional cloning in Arabidopsis. Why it feels good to have a genome initiative working for you. Plant Physiol. 123:795-805[Abstract/Free Full Text].

LUNDQVIST, U., J. MEYER, and A. LUNDQVIST, 1991  Mutagen specificity for 71 lines resistant to barley powdery mildew race-D1 and isolated in 4 highbred barley varieties. Hereditas 115:227-239[Medline].

MICHELMORE, R., 1995  Molecular approaches to manipulation of disease resistance genes. Annu. Rev. Phytopathol. 33:393-427.

MITCHELL-OLDS, T., 1995  Interval mapping of viability loci causing heterosis in Arabidopsis. Genetics 140:1105-1109[Abstract].

PFLIEGER, S., V. LEFEBVRE, C. CARANTA, A. BLATTES B, and A. BLATTES BGOFFINET ET AL., 1999  Disease resistance gene analogs as candidates for QTLs involved in pepper-pathogen interactions. Genome 42:1100-1110[Medline].

PLOTNIKOVA, J. M., T. L. REUBER, and F. M. AUSUBEL, 1998  Powdery mildew pathogenesis of Arabidopsis thaliana. Mycologia 90:1009-1016.

REDDY, K. S., S. E. PAWAR, and C. R. BHATIA, 1994  Inheritance of powdery mildew (Erysiphe polygoni Dc) resistance in mungbean (Vigna radiata L Wilczek). Theor. Appl. Genet. 88:945-948.

REITER, R. S., J. G. WILLIAMS, K. A. FELDMANN, J. A. RAFALSKI, and S. V. TINGEY et al., 1992  Global and local genome mapping in Arabidopsis thaliana by using recombinant inbred lines and random amplified polymorphic DNAs. Proc. Natl. Acad. Sci. USA 89:1477-1481[Abstract/Free Full Text].

SATO, S., H. KOTANI, R. HAYASHI, Y. G. LIU, and D. SHIBATA et al., 1998  A physical map of Arabidopsis thaliana chromosome 3 represented by two contigs of CIC YAC, P1, TAC and BAC clones. DNA Res. 5:163-168[Abstract].

TANKSLEY, S. D., 1993  Mapping polygenes. Annu. Rev. Genet. 27:205-233[Medline].

TINKER, N. A. and D. E. MATHER, 1995  Methods for QTL analysis with progeny replicated in multiple environments. J. Agric. Genom. 1, a http://www.ncgr.org/research/jag/papers95/paper295/indexp295.html.

TINKER, N. A. and D. E. MATHER, 1995  MQTL: software for simplified composite interval mapping of QTL in multiple environments. J Agric. Genom. 1, b http://www.ncgr.org/research/jag/papers95/paper195/indexp195.html.

TUINSTRA, M. R., G. EJETA, and P. B. GOLDSBROUGH, 1997  Heterogenous inbred family (HIF) analysis: a method for developing near-isogenic lines that differ at quantitative trait loci. Theor. Appl. Genet. 95:1005-1011.

VOGEL, J. and S. SOMERVILLE, 2000  Isolation and characterization of powdery mildew-resistant Arabidopsis mutants. Proc. Natl. Acad. Sci. USA 97:1897-1902[Abstract/Free Full Text].

XIAO, X. E. S., K. FINDLAY, R. P. OLIVER, and J.G. TURNER, 1997  Characterization of three loci controlling resistance of Arabidopsis thaliana accession Ms-0 to two powdery mildew diseases. Plant J. 12:757-768[Medline].

XIAO, S., S. ELLWOOD, O. CALIS, E. PATRICK, and T. LI et al., 2001  Broad-spectrum mildew resistance in Arabidopsis thaliana mediated by RPW8. Science 291:118-120[Abstract/Free Full Text].

YOUNG, N. D., 1996  QTL mapping and quantitative disease resistance in plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 34:479-501.

This article has been cited by other articles:

				J. Wan, X.-C. Zhang, D. Neece, K. M. Ramonell, S. Clough, S.-y. Kim, M. G. Stacey, and G. Stacey A LysM Receptor-Like Kinase Plays a Critical Role in Chitin Signaling and Fungal Resistance in Arabidopsis PLANT CELL, February 1, 2008; 20(2): 471 - 481. [Abstract] [Full Text] [PDF]

				A. C. Diener and F. M. Ausubel RESISTANCE TO FUSARIUM OXYSPORUM 1, a Dominant Arabidopsis Disease-Resistance Gene, Is Not Race Specific Genetics, September 1, 2005; 171(1): 305 - 321. [Abstract] [Full Text] [PDF]

				K. Ramonell, M. Berrocal-Lobo, S. Koh, J. Wan, H. Edwards, G. Stacey, and S. Somerville Loss-of-Function Mutations in Chitin Responsive Genes Show Increased Susceptibility to the Powdery Mildew Pathogen Erysiphe cichoracearum Plant Physiology, June 1, 2005; 138(2): 1027 - 1036. [Abstract] [Full Text] [PDF]

				S. Xiao, B. Emerson, K. Ratanasut, E. Patrick, C. O'Neill, I. Bancroft, and J. G. Turner Origin and Maintenance of a Broad-Spectrum Disease Resistance Locus in Arabidopsis Mol. Biol. Evol., September 1, 2004; 21(9): 1661 - 1672. [Abstract] [Full Text] [PDF]

				D. J. Wolyn, J. O. Borevitz, O. Loudet, C. Schwartz, J. Maloof, J. R. Ecker, C. C. Berry, and J. Chory Light-Response Quantitative Trait Loci Identified with Composite Interval and eXtreme Array Mapping in Arabidopsis thaliana Genetics, June 1, 2004; 167(2): 907 - 917. [Abstract] [Full Text] [PDF]

				E. J.M. Clerkx, M. E. El-Lithy, E. Vierling, G. J. Ruys, H. Blankestijn-De Vries, S. P.C. Groot, D. Vreugdenhil, and M. Koornneef Analysis of Natural Allelic Variation of Arabidopsis Seed Germination and Seed Longevity Traits between the Accessions Landsberg erecta and Shakdara, Using a New Recombinant Inbred Line Population Plant Physiology, May 1, 2004; 135(1): 432 - 443. [Abstract] [Full Text] [PDF]

				C. Weinig, L. A. Dorn, N. C. Kane, Z. M. German, S. S. Halldorsdottir, M. C. Ungerer, Y. Toyonaga, T. F. C. Mackay, M. D. Purugganan, and J. Schmitt Heterogeneous Selection at Specific Loci in Natural Environments in Arabidopsis thaliana Genetics, September 1, 2003; 165(1): 321 - 329. [Abstract] [Full Text] [PDF]

				O. Loudet, S. Chaillou, P. Merigout, J. Talbotec, and F. Daniel-Vedele Quantitative Trait Loci Analysis of Nitrogen Use Efficiency in Arabidopsis Plant Physiology, January 1, 2003; 131(1): 345 - 358. [Abstract] [Full Text] [PDF]

				N. A. Eckardt Plant Disease Susceptibility Genes? PLANT CELL, September 1, 2002; 14(9): 1983 - 1986. [Full Text] [PDF]