Multiple Avirulence Loci and Allele-Specific Effector Recognition Control the Pm3 Race-Specific Resistance of Wheat to Powdery Mildew

Abstract

In cereals, several mildew resistance genes occur as large allelic series; for example, in wheat (Triticum aestivum and Triticum turgidum), 17 functional Pm3 alleles confer agronomically important race-specific resistance to powdery mildew (Blumeria graminis). The molecular basis of race specificity has been characterized in wheat, but little is known about the corresponding avirulence genes in powdery mildew. Here, we dissected the genetics of avirulence for six Pm3 alleles and found that three major Avr loci affect avirulence, with a common locus_1 involved in all AvrPm3-Pm3 interactions. We cloned the effector gene AvrPm3(a2/f2) from locus_2, which is recognized by the Pm3a and Pm3f alleles. Induction of a Pm3 allele-dependent hypersensitive response in transient assays in Nicotiana benthamiana and in wheat demonstrated specificity. Gene expression analysis of Bcg1 (encoded by locus_1) and AvrPm3 (a2/f2) revealed significant differences between isolates, indicating that in addition to protein polymorphisms, expression levels play a role in avirulence. We propose a model for race specificity involving three components: an allele-specific avirulence effector, a resistance gene allele, and a pathogen-encoded suppressor of avirulence. Thus, whereas a genetically simple allelic series controls specificity in the plant host, recognition on the pathogen side is more complex, allowing flexible evolutionary responses and adaptation to resistance genes.

Figure 1.

Interpretation of Genetic Segregation in the F1 Progeny of the Cross between the Wheat Powdery Mildew Isolates 96224 and 94202. The genetic loci segregating for avirulence in the population are color coded. Locus_1, involved in the six AvrPm3-Pm3 interactions, is indicated in red. Locus_2, involved in the AvrPm3f2-Pm3f interaction, is indicated in purple. Locus_3, involved in the AvrPm3b2-Pm3b, AvrPm3c2-Pm3c, and AvrPm3d2-Pm3d interactions, is indicated in blue. Finally, the unmapped AvrPm3b3 and AvrPm3d3 loci are indicated in orange and green, respectively. At each locus, the genotypes from the avirulent 96224 and virulent 94202 parents are indicated by darker and lighter shades, respectively. For simplicity, the AvrPm3 loci were abbreviated by a single letter indicating the genotype of the avirulent (uppercase) and virulent (lowercase) parent. Phenotypes are indicated by A (avirulence), I (intermediate avirulence), and V (virulence). Segregation ratios of A:I:V are indicated. Detailed description and interpretation of phenotype segregation ratios from (A) to (C) are provided in Results and Supplemental Methods. (A) Avirulence toward Pm3a, Pm3c, and Pm3e is determined by the parental genotype inherited at a single locus, which is commonly involved in the six AvrPm3-Pm3 interactions. (B) Avirulence toward Pm3f is determined by the parental genotype inherited at two loci. One out of the four possible genotype combinations determines avirulence on Pm3f, and three determine virulence. (C) Avirulence toward Pm3b and Pm3d is determined by the parental genotype inherited at three loci. Eight genotype combinations are possible. For Pm3b, five determine avirulence, one determines intermediate avirulence, and two determine virulence. For Pm3d, two combinations determine avirulence, one determines intermediate avirulence, and five determine virulence.

Figure 2.

Identification and Mapping of Three Major Loci Involved in the AvrPm3-Pm3 Interaction (A) Comparative analysis of the reaction of individual progeny isolates from an F1 segregating population of the cross 96224 × 94202 on the different Pm3 resistance alleles. Phenotypes on six Pm3 alleles are indicated for 143 progeny and sorted by avirulence (blue), intermediate avirulence (green), and virulence (red). The identity of the progeny is indicated according to phenotype sorting, by an identification number on the left side. A subset of 17 progeny (indicated with braces) corresponding to about one-eighth of the population is consistently avirulent on five Pm3 alleles. The subset of avirulent progeny on Pm3b includes all avirulent progeny on the five other alleles. The subset of progeny showing intermediate avirulence on Pm3b is different from the one showing this phenotype on Pm3d. (B) Linkage groups of consensus maps of the wheat powdery mildew genome containing loci involved in the AvrPm3-Pm3 interaction. The genotype of the avirulent parent at these loci is indicated. Three loci were mapped on three consensus linkage groups. The consensus map was produced by merging the linkage groups containing the AvrPm3/avrPm3 specificities, individually mapped in different progeny subsets, or across two populations (Supplemental Figures 3 to 6 and Supplemental Methods). Colocalization of different AvrPm3/avrPm3 loci was assigned based on the closest genetically linked KASP or AFLP marker and comparison of phenotype patterns in the progeny. Flanking markers inferred from consensus mapping are indicated. Genetic distance between the flanking markers as well as the size of the linkage group are indicated in centimorgans (cM).

Figure 3.

High-Resolution Mapping of AvrPm3^f2 in locus_2. For each level, the most informative genetic or physical interval was chosen to be depicted and the 5′ to 3′ orientation is arbitrarily assigned to improve clarity in the absence of knowledge of the location of the centromere. (A) Linkage group 2 of the genetic map containing the AvrPm3f2 locus, the two closest KASP markers, M33RE (red) and M426MI (purple), the flanking marker on contig-52, 52M2 (blue), and a marker for contig-26, 26M4 (green). Distances are indicated in centimorgans (cM). (B) Fingerprinted contigs (FPCs) covering the AvrPm3f2 region. Contigs linked to AvrPm3f2 by the genome-wide SNP genotyping approach (contig-33 and -426) and by bulk segregant analysis (contig-52 and -26) are shown, and their estimated sizes and location of genetically informative markers are given. Red, blue, purple, and green boxes denote known sequence, while gray boxes indicate sequence gaps as derived from the wheat powdery mildew genome sequence (Wicker et al., 2013). Yellow is a sequence gap of unknown size in the FPC assembly. (C) Fine genetic mapping in the AvrPm3f2 region. The most informative CAPS markers and the two flanking KASP markers are shown, and their genetic distances from AvrPm3f2 are provided. The number of recombinants between the flanking and cosegregating markers is specified. (D) BAC minimum tiling path of LTC scaffold 10, spanning the AvrPm3f2 region. Markers are indicated by vertical bars and specific PCR amplifications from both CAPS and BAC-end markers are indicated by black diamonds. Horizontal bars indicate BACs from FPC contig-33 (red) and FPC contig-52 (blue). Only the informative BACs selected for sequencing are shown. (E) Annotated locus_2 sequence. Putative effector genes and their transcriptional orientation are given. Polymorphic genes within the candidate interval are indicated (asterisk). The functionally validated AvrPm3^f2 is indicated by a black arrow. An inverted duplication of 15 kb is highlighted (red overlay). Transposable elements are depicted as colored boxes.

Figure 4.

Description of the AvrPm3^a2/f2-Encoded Proteins and Protein Variants Used in This Study. (A) Alignment of the two protein sequences encoded by the avirulence (Pu_7^avr) and the virulence alleles (Pu_7^vir) of the Pu_7 effector gene. Amino acids are colored according to their biochemical properties. The predicted signal peptide (green box) and cleavage site (green asterisks) are indicated. Total protein size is 130 amino acids, with a predicted signal peptide of 24 residues and a predicted mature protein of 106 residues. The YxC motif and the second conserved cysteine are indicated (black boxes). The two amino acid polymorphisms between the parental alleles, one nonconservative (Gly/Glu) and one conservative (Ala/Val), are shown (red boxes). (B) Schematic diagram of the protein variants used in transient protein expression assays. At the top, the full-length protein encoded by AvrPm3^a2/f2 is shown (130 residues). The predicted signal peptide (SignalP version 4.1) is indicated in green. The location of the YxC motif is indicated in blue. Red bars indicate SNPs distinguishing between virulence and avirulence alleles (AvrPm3^a2/f2 and avrPm3^a2/f2) and the two variants generated by site-directed mutagenesis (AvrPm3^a2/f2_G84E and AvrPm3^a2/f2_A107V).

Figure 5.

Functional Validation of the AvrPm3^a2/f2-Pm3a/f Interaction (A) Agroinfiltration assays in N. benthamiana demonstrate induction of an HR response upon specific recognition of Pu_7^avr (AvrPm3^a2/f2) by Pm3a and Pm3f^L456P/Y458H. The avirulence (Pu_7^avr) and virulence (Pu_7^vir) alleles of Pu_7 without signal peptide were transiently expressed together with Pm3a and Pm3f^L456P/Y458H. Leaves of 4-week-old N. benthamiana plants were infiltrated with Agrobacterium tumefaciens cultures expressing each of the constructs indicated. An Avr:R ratio of 4:1 and OD₆₀₀ of 1.2 were used to test for HR induction. Results are consistent across replicates from at least three independent experiments where four to eight leaves were assayed. Photographs were taken 5 d after infiltration. (B) Cobombardment of AvrPm3^a2/f2 (Pu_7^avr) with Pm3a in wheat reduces the number of GUS-expressing cells. The AvrPm3^a2/f2 construct encoding for a protein version without signal peptide was transiently coexpressed in wheat leaf segments with Pm3a and the pUbiGUS reporter, using the particle bombardment assay. The number of GUS-expressing cells was assessed 48 h after bombardment and compared between leaves coexpressing (1) Pm3a + AvrPm3^a2/f2, (2) Pm3a + empty vector (pIPKb004) as a control, and (3) the cell death-inducing autoactive Pm3a^HR + empty vector (pIPKb004) as a second control. Values are given as the average number of GUS-stained cells per slide containing five leaves. Six slides were counted for each of four to six bombardment assays (see Methods). Statistical significance was assessed using the unpaired Student’s t test. Significance under the threshold P ≤ 0.001 (***) is indicated.

Figure 6.

Functional Analysis of the AvrPm3^a2/f2-Pm3a/f Interaction in the N. benthamiana System. (A) Agroinfiltration assays in N. benthamiana with AvrPm3^a2/f2 constructs encoding the protein versions with and without signal peptide, coinfiltrated with Pm3a or Pm3f^L456P/Y458H. Stronger HR was observed with the AvrPm3^a2/f2construct expressing the version without signal peptide. Photographs were taken 5 d after infiltration. (B) Both polymorphic residues in the protein encoded by AvrPm3^a2/f2are necessary for HR induction by Pm3a and Pm3f^L456P/Y458H. Constructs expressing the AvrPm3^a2/f2_G84E and AvrPm3^a2/f2_A107V variants were coinfiltrated with those expressing Pm3a and Pm3f^L456P/Y458H, and HR was assessed after 5 d. No HR was observed for either variant with a single amino acid exchange, in comparison with coinfiltration of AvrPm3^a2/f2with Pm3a or Pm3f^L456P/Y458H. (C) Suppression of the AvrPm3^a2/f2-Pm3a/f-mediated HR by Pm3b and Pm3c in transient assays in N. benthamiana. The construct expressing AvrPm3^a2/f2 was coinfiltrated with constructs expressing Pm3a or Pm3f^L456P/Y458H together with those expressing Pm3b, Pm3c, or GUS reporter. HR was assessed after 5 d. Suppression of HR was observed with all the combinations including Pm3b or Pm3c in presence of AvrPm3^a2/f2 and Pm3a, and with AvrPm3^a2/f2 and Pm3f^L456P/Y458H. By contrast, no suppression of HR was observed in a positive control where Pm3b and Pm3c were replaced by GUS reporter or when AvrPm3^a2/f2, Pm3a, and Pm3f were combined. Results in (A) to (C) were consistent across replicates from at least three independent experiments where four to eight leaves were assayed.

Figure 7.

Gene Expression Analysis of the SvrPm3^a1/f1 Candidate Suppressor Bcg1, and the AvrPm3^a2/f2 Avirulence Effector Pu_7, in Parental Isolates and F1 Progeny of Powdery Mildew. (A) The mean normalized expression of Pu_7 (AvrPm3^a2/f2, black line) and Bcg1 (svrPm3^a1/f1, red line) in the avirulent parental isolate 96224, 1 to 5 d after infection of the susceptible wheat line ‘Chinese Spring.’ (B) The mean normalized expression of Bcg1 at 2 d after infection of the susceptible wheat line ‘Chinese Spring’ with the parental isolates 94202, JIW2, and 96224. The phenotypes on Pm3a and Pm3f are indicated. The genotypes of Bcg1 and Pu_7 are indicated according to the proposed Avr/Svr nomenclature (see text). Statistical significance was assessed using the unpaired Student’s t test. Significance under the threshold P ≤ 0.01 (**) is indicated. (C) The mean normalized expression of Bcg1 (red histograms) and Pu_7 (black histograms) at 2 d after infection. Leaf segments from the susceptible backcrossing line ‘Chancellor’ were infected with (1) the parental isolates (94202 and 96224), (2) four progeny selected from each allelic combination of the parental genotypes for Bcg1 and Pu_7 (#104, #13, #45, and #90), (3) the Pm3f gain of virulence mutant EXP3 (UV mutagenized 96224), and (4) the Pm3f intermediate progeny #43. Progeny are labeled with their identification number as assigned in Figure 2A. The phenotypes on Pm3f are indicated. The parental genotypes of Bcg1 and Pu_7 are indicated according to the proposed Avr/Svr nomenclature (see text). Gene expression levels were compared using the Tukey-Kramer mean separation procedure (α = 0.05). Significantly different patterns are indicated by different letters on top of the histograms. In all these assays, total RNA was extracted from infected leaf tissue flash-frozen at the indicated time points. The values represent the average expression from three independent biological replicates. Standard error of the mean is indicated.

Figure 8.

The Avr/R/Svr Genetic Model. We propose that the Pm3 multiallelic race-specific resistance follows a three component interaction model involving (1) an allele of the R gene; (2) an avirulence effector (Avr) specifically recognized by this allele that can also occur as a virulence form (avr) that is not recognized; and (3) a pathogen-encoded suppressor of the Avr-R interaction (Svr) that can also occur as a nonsuppressor form (svr). Resistance is mediated only by the Avr/R/svr combination where the resistance gene recognizes its cognate avirulence effector in absence of the suppressor. In absence of the cognate Avr or in presence of the Svr suppressor, the interaction results in susceptibility.