RRS1 and RPS4 provide a dual Resistance-gene system against fungal and bacterial pathogens

Authors


*(fax +81 866 56 9453; e-mail yo_narusaka@bio-ribs.com).

Summary

Colletotrichum higginsianum is a fungal pathogen that infects a wide variety of cruciferous plants, causing important crop losses. We have used map-based cloning and natural variation analysis of 19 Arabidopsis ecotypes to identify a dominant resistance locus against C. higginsianum. This locus named RCH2 (for recognition of C. higginsianum) maps in an extensive cluster of disease-resistance loci known as MRC-J in the Arabidopsis ecotype Ws-0. By analyzing natural variations within the MRC-J region, we found that alleles of RRS1 (resistance to Ralstonia solanacearum 1) from susceptible ecotypes contain single nucleotide polymorphisms that may affect the encoded protein. Consistent with this finding, two susceptible mutants, rrs1-1 and rrs1-2, were identified by screening a T-DNA-tagged mutant library for the loss of resistance to C. higginsianum. The screening identified an additional susceptible mutant (rps4-21) that has a 5-bp deletion in the neighboring gene, RPS4-Ws, which is a well-characterized R gene that provides resistance to Pseudomonas syringae pv. tomato strain DC3000 expressing avrRps4 (Pst-avrRps4). The rps4-21/rrs1-1 double mutant exhibited similar levels of susceptibility to C. higginsianum as the single mutants. We also found that both RRS1 and RPS4 are required for resistance to R. solanacearum and Pst-avrRps4. Thus, RPS4-Ws and RRS1-Ws function as a dual resistance gene system that prevents infection by three distinct pathogens.

Introduction

Plants are exposed to various types of potentially invasive organisms, including viruses, bacteria, fungi, nematodes and protozoa, but are able to defend themselves by activating multiple defense mechanisms. A crucial event in the defense response is the recognition of the invading pathogen at an early stage of infection. The gene-for-gene hypothesis (Flor, 1971) provides a mechanism for specific recognition of the pathogen by the plant. This recognition is mediated by direct or indirect interactions between the product of a plant resistance (R) gene, and the corresponding effectors encoded by avirulence genes in the pathogen (Jones and Dangl, 2006). R genes have been isolated from a wide range of plant species (Hammond-Kosack and Parker, 2003). Most of them encode non-membrane proteins that contain a conserved nucleotide binding (NB) site and a C-terminal leucine-rich repeat (LRR) domain.

The Arabidopsis genome contains about 150 genes coding for NB-LRR-containing proteins (Meyers et al., 2003). This is far less than would be needed to respond individually, and specifically, to all of its potential pathogens. However, plants may have been able to limit the number of required NB-LRR-encoding genes, if host proteins are able to perceive sets of distinct pathogens (Van der Biezen and Jones, 1998). Recent studies on the model Arabidopsis–Pseudomonas system revealed that multiple pathogen effectors target the same host protein, and that its modification by effectors activates several distinct R proteins (Chisholm et al., 2006; Jones and Dangl, 2006). For example, AvrB, AvrRpm1 and AvrRpt2, three distinct effectors from Pseudomonas species target the host protein RIN4, the status of which is closely monitored by the NB-LRR proteins RPM1 and RPS2 (Axtell and Staskawicz, 2003; Mackey et al., 2003). However, RPM1 can detect RIN4 modification caused by AvrB and AvrRpm1, but not by AvrRpt2, and RPS2 is activated by the AvrRpt2-dependent cleavage of RIN4, but not by the AvrB- and AvrRpm1-dependent modification of RIN4. AvrRpm1 also elicits defense signaling through RPS2 (Kim et al., 2009). The precise molecular mechanism of how a particular modification activates certain R proteins remains unclear.

Colletotrichum higginsianum causes typical anthracnose lesions on the leaves, petioles and stems of cruciferous plants (Higgins, 1917). Inoculation of Arabidopsis ecotype Columbia (Col-0) leaves with C. higginsianum results in fungal growth and disease symptoms reminiscent of those induced in other cruciferous plants (Narusaka et al., 2004; O’Connell et al., 2004). During interactions with its hosts, C. higginsianum develops a series of specialized infection structures, including germ tubes, appressoria, biotrophic primary hyphae and secondary necrotrophic hyphae. Inoculation of a large number of accessions with isolates of C. higginsianum showed that Arabidopsis has at least two dominant resistance gene loci, designated RCH1 and RCH2, indicating that Arabidopsis resistance to C. higginsianum is controlled by a gene-for-gene interaction (Narusaka et al., 2004). In a previous study, we identified a single putative R locus, RCH1, on the top of chromosome 4, in the resistance Arabidopsis thaliana ecotype Eil-0 against C. higginsianum (Narusaka et al., 2004).

In this study, we present the mapping and characterization of the RCH2 locus in Arabidopsis, and we demonstrate that a pair of neighboring genes, RRS1-Ws and RPS4-Ws, function cooperatively as a dual R-gene system against at least three distinct pathogens.

Results

Identification of RCH2, an R locus against Colletotrichum higginsianum

Three different strains of C. higginsianum (MAFF305635, MAFF305968 and MAFF305970) cause anthracnose disease symptoms in Arabidopsis Col-0 plants, similar to those incited on turnip plants (Brassica rapa L. and Brassica oleracea) (Narusaka et al., 2004). When challenged with these strains, Col-0 plants develop brown necrotic halo lesions, expanding from the inoculation site 4 days post inoculation (dpi), and coalescing over the entire leaf (Narusaka et al., 2004). In contrast, Arabidopsis ecotypes with a moderate level of resistance, such as Ws-0, initially develop brown necrotic lesions at the inoculation site, but lesion expansion is restricted (Narusaka et al., 2004). To identify the loci responsible for this resistance in Ws-0, we crossed Ws-0 and Col-0. All the Fl progenies exhibited the resistance phenotype to C. higginsianum strain MAFF305635 that is characteristic of the Ws-0 parent. In addition, the F2 progenies segregated at 1604:485 (3:1, χ2 = 3.54, P = 0.06) for resistance versus susceptibility to the pathogen (Table 1), suggesting that Ws-0 contains a single dominant resistance locus. This locus was designated RCH2, and was mapped to chromosome 5 between the molecular markers N01-K17O22 and N01-K11I1 (Figure 1; Table S1) by PCR-based microsatellite and amplified polymorphic sequence mapping in an F2 progeny population of 300. The RCH2 locus falls within an extensive cluster of disease-resistance loci known as the multiple resistance complex J (MRC-J) region (Botella et al., 1997). This cluster includes well-characterized R genes such as RPS4 (At5g45250), RRS1 (At5g45260), RPP8 (At5g43470) and at least eight additional NB-LRR-protein encoding genes. A tightly linked SSLP marker, N01-K9E15, was found 936 bp downstream of the stop codon of RPS4-Ws (Figure 1).

Table 1.   Distribution of interaction phenotypes among the F2 progeny of crosses between Col-0 and Ws-0 plants inoculated with Colletotrichum higginsianum (MAFF305635)
Arabidopsis thaliana cross (female × male) F2 segregation analysis Predicted ratioa
EcotypeGenerationSusceptibleResistant(S:R)χ2P
  1. Ratios are drawn from the observed behavior of selfed-progeny of F1 progeny of a cross between Col-0 and Ws-0.

  2. aFor use in a chi-square test of selected models for predicting the number of host genes associated with incompatibility.

Col-0Parent 50   0   
Ws-0Parent  0  50   
Col-0 × Ws-0F2 435 14031:31.740.19
Ws-0 × Col-0F2 50 2011:33.450.06
(Total of F2)(485)(1604)(1:3)(3.54)(0.06)
Figure 1.

 Map-based cloning of RRS1-Ws.
The RCH2 locus was mapped on chromosome 5 of Ws-0 by using PCR-based microsatellite and amplified polymorphic sequence genetic markers in 300 F2 plants. The numbers of recombinants between RCH2 and each marker are indicated in parenthesis. The RCH2 locus is between N01-K17O22 and N01-K11I1, and is tightly linked to N01-K9E15. Exons of RPS4-Ws and RRS1-Ws are indicated by black and gray boxes, respectively. The positions of the mutations in RPS4-Ws and RRS1-Ws are indicated by white triangles. The mutations in RPS4-Ws and RRS1-Ws were complemented with 6.3 kbp and the 8.2-kbp genomic fragments (shown as white boxes), respectively.

Natural variation analysis of the RCH2 locus

We investigated the susceptibility to C. higginsianum strain MAFF305635 among 19 Arabidopsis ecotypes, for which large-scale single nucleotide polymorphisms (SNPs) are available to define the association genetics for the RCH2 locus (Clark et al., 2007). Ecotypes Bay-0, Br-0, Est-1, Fei-0, Nfa-8, Rrs-7, Rrs-10, Sha, Tamm-2, Ts-1, Tsu-1, Van-0 and Ws-0 are resistant to C. higginsianum, whereas Bur-0, C24, Col-0, Cvi-0, Ler-1 and Lov-5 are susceptible (Figures 2, S1 and S2). To identify the RCH2 locus, natural sequence variations of amino acid residues from seven candidates (from At5g45200 to At5g45260) among 19 ecotypes were investigated by hierarchical cluster analysis (Clark et al., 2007). We found that RRS1 alleles from the susceptible ecotypes C24, Ler-1 and Lov-5 are clustered in one clade (Figures 2 and S1). RRS1 alleles from three other susceptible ecotypes, Bur-0, Col-0 and Cvi-0, contain a premature stop codon (Figure S1). These results suggest that the RRS1 allele is a candidate for RCH2.

Figure 2.

 Natural variation of RRS1.
The natural sequence variation of amino acid residues of RRS1 among 16 Arabidopsis ecotypes was investigated by hierarchical cluster analysis (DNASIS pro 2.8; Hitachi Software Engineering, http://www.hitachi-soft.com). The resistance and susceptibility to Colletotrichum higginsianum are indicated by ‘R’ or ‘S’, respectively. Note that the RRS1 sequences from the susceptible ecotypes C24, Lov-5 and Ler-1 are clustered.

Molecular cloning of RCH2

To test whether the Ws-0 allele of RRS1 (RRS1-Ws) is RCH2, we screened a T-DNA-tagged library of Arabidopsis Ws-0 seeds (ABRC CS22830) for loss of resistance to C. higginsianum strain MAFF305635. Of the 29 475 T3 plants screened, we identified three mutants that were more susceptible to C. higginsianum. The increase in susceptibility of these lines was confirmed in their progeny. We isolated a cosmid clone containing RRS1-Ws from a genomic library of Ws-2 (ABRC CD4-11), and determined the nucleotide sequence of a 20-kbp genomic DNA region covering the gene and part of its neighboring gene, RPS4-Ws. Using this sequence information, we found that two of the susceptible mutants contain independent T-DNA insertions within the fifth intron of RRS1-Ws (Figure 1) [4485 nt (rrs1-2) or 4934 nt (rrs1-1) downstream of the ATG start codon] in RRS1-Ws (Figure 1). In these mutants, no mutations in RPS4-Ws were found. The transcripts of the RRS1 gene in rrs1-1 and rrs1-2 lines were below the detection limit by RT-PCR analyses, and these mutant lines are likely to be transcriptionally null (data not shown). To confirm that RRS1-Ws confers resistance against C. higginsianum, we introduced an 8.2-kbp genomic RRS1-Ws fragment that includes, approximately, 1.8 kbp upstream and 176 bp downstream into the rrs1-1 mutant line. Resistance to C. higginsianum was fully restored in the transgenic plants (Figures 1 and 3a), confirming that RRS1-Ws is responsible for C. higginsianum resistance in Arabidopsis.

Figure 3.

Colletotrichum higginsianum resistance analysis in mutant and transgenic plant lines.
(a) Infection phenotypes of plant leaves inoculated with C. higginsianum.
Mature leaves of 28–30-day-old plants were inoculated by placing 5 μl of a spore suspension of C. higginsianum (5 × 105 spores ml−1) on each side of the leaf. Leaves were harvested at 6 days post inoculation (dpi), and were stained with trypan blue. rrs1-1/RRS1-Ws-#2 represents an independent transgenic rrs1-1 plant harboring the genomic RRS1-Ws fragment. rps4-21/RPS4-Ws-#9 and 11, represent independent transgenic rps4-21 plants harboring the genomic RPS4-Ws fragment. rps4-21/rrs1-1 represents a double mutant obtained by crossing rps4-21 and rrs1-1 mutants. Each picture shows a representative of three independent experiments.
(b) Quantification of C. higginsianum in planta by qRT-PCR.
Plants that were 28–30-days old were spray inoculated with C. higginsianum. Inoculated leaves were harvested at 4 dpi and total RNA was isolated. QRT-PCR was performed with Ch-ACT primers for each sample. Error bars indicate the SEs. The asterisks indicate the statistically significant difference from the wild-type (WT) controls (Dunnett’s method, < 0.05; Dunnett, 1955). This experiment was repeated three times with similar results.

RPS4-Ws is also required for resistance to Colletotrichum higginsianum

The third susceptible mutant did not have a mutation in RRS1-Ws, but contained a five-base deletion in the third exon (2140–2144 nt downstream of the ATG start codon) of RPS4-Ws, resulting in a frame shift in the LRR domain (Figure 1). The rps4-21 contains missense mutations at L639R and V640P in the LRR domain, and a stop mutation after 640P, and is thus likely to be null. To verify whether RPS4-Ws is also required for resistance to C. higginsianum, we introduced a 6.3-kbp genomic fragment containing RPS4-Ws into the rps4-21 mutant (Figure 1). All of the 28 transgenic T2 lines obtained showed resistance to C. higginsianum similar to wild-type (WT) Ws-0 (Figure 3a). Thus, RRS1-Ws and RPS4-Ws are both required for resistance to C. higginsianum in Arabisposis Ws-0.

To assess if RRS1-Ws and RPS4-Ws function in concert, we generated an rps4-21/rrs1-1 double mutant by crossing rps4-21 and rrs1-1 mutants. We estimated the levels of C. higginsianum in the infected plant by quantifying the actin mRNA of the pathogen (Ch-ACT) using real-time quantitative RT-PCR (qRT-PCR). The Arabidopsis CBP20 (At-CBP20) gene is constitutively expressed, and was thus used for normalization. Colletotrichum higginsianum is quantified by calculating the number of copies of Ch-ACT mRNA per At-CBP20 mRNA prepared from infected A. thaliana leaves. Ch-ACT mRNA levels were significantly increased in the rrs1-1 and rps4-21 mutants, compared with Ws-0 at 4 dpi. No significant difference in Ch-ACT mRNA levels was observed between the rps4-21/rrs1-1 double mutant and the single allele mutants, suggesting that RRS1-Ws and RPS-4-Ws function cooperatively (Figure 3b).

Both RRS1-Ws and RPS4-Ws are required for resistance to Ralstonia solanacearum

The RRS1 allele from the Nd-1 ecotype (RRS1-Nd) was originally isolated as an R gene specifically against the R. solanacearum strain GMI1000 (Rs1000) (Deslandes et al., 2002). We found that Ws-0 is also resistant to strain 1002 (Rs1002), a Japanese isolate of this pathogen (Figure 4). To investigate whether RRS1-Ws and RPS4-Ws could provide resistance to this pathogen, we carried out a root inoculation assay. Wilting symptoms appeared after 3–4 days in rrs1-1, rrs1-2 and rps4-21 plants, resulting in plant death within 7 days of inoculation. On the contrary, the WT Ws-0 plants were healthy during the inoculation. As shown in Figure 4, bacterial growth was approximately 5-fold higher in the mutant lines than in WT Ws-0 and in complemented lines. These results indicate that both genes are required for resistance to R. solanacearum. To assess if RRS1-Ws and RPS4-Ws function in concert, the rps4-21/rrs1-1 double mutant was inoculated with R. solanacearum. No significant difference in bacterial growth was observed between the rps4-21/rrs1-1 double mutant and the single allele mutants, strongly suggesting that RRS1-Ws and RPS-4-Ws function cooperatively (Figure 4). Rs1002 contains a gene that is homologous with popP2 of Rs1000, encoding a YopJ/AvrRxv-type effector that can be recognized by RRS1-R (Deslandes et al., 2002). To determine if this gene plays a role in the infection, we used an Rs1002 strain in which the popP2 homolog was disrupted (Rs1002-ΔpopP2). We carried out a root inoculation assay with Rs1002-ΔpopP2. The WT Ws-0 plants were susceptible to the mutant strain, and the susceptibility level was similar to that in rrs1-1 or rps4-21 (Figure S3). These results show that, similar to Rs1000, PopP2 is the pathogen-determining factor for the RRS1/RPS4-based recognition of Rs1002 in Ws-0.

Figure 4.

Ralstonia solanacearum resistance analysis in mutant and transgenic plant lines, assessed by root inoculation assay.
Arabidopsis plants that 5–6-weeks old were inoculated with R. solanacearum strain 1002. Aerial plant parts were harvested 3 days after inoculation and analyzed. rrs1-1/RRS1-Ws-#2 represents an independent transgenic rrs1-1 plant harboring the RRS1-Ws genomic fragment. rps4-21/RPS4-Ws-#9 represents an independent transgenic rps4-21 plant harboring the genomic RPS4-Ws fragment. rps4-21/rrs1-1 represents a double mutant obtained by crossing rps4-21 and rrs1-1 mutants. The asterisks indicate statistically significant difference from the wild-type (WT) controls (Dunnett’s method, < 0.05). This experiment was repeated three times with similar results.

Both RPS4-Ws and RRS1-Ws are required for resistance to Pst-avrRps4

RPS4-Col was previously shown to confer resistance to Pst-avrRps4 (Gassmann et al., 1999). To determine whether RPS4-Ws is required for resistance to Pst-avrRps4, bacterial infection levels were monitored in Ws-0 and in the rps4-21 mutant at 0 and 3 dpi (Figure 5a). The Pst-avrRps4 count was about five times higher in rps4-21 than in WT Ws-0, and the RPS4-Ws complemented lines showed similar results as WT Ws-0, indicating that RPS4-Ws confers resistance to Pst-avrRps4 in Ws-0. In rrs1-1 and rrs1-2, the growth of Pst-avrRps4 was approximately 10-fold higher than in Ws-0 (Figure 5b), whereas the complemented RRS1-Ws line showed a bacterial count similar to the WT, indicating that the RRS1-Ws is required for resistance to Pst-avrRps4. Furthermore, no significant difference in bacterial growth was observed between the rps4-21/rrs1-1 double mutant and the single allele mutants, suggesting that RRS1-Ws and RPS-4-Ws function cooperatively (Figure 5). In contrast, Pst without the avrRps4 effector gene grew to similar levels in WT Ws-0, rps4-21, rrs1-1 and rrs1-2 (Figure 5c). The susceptibility in these mutants is not as high as that attained in plants with Pst without avrRps4, suggesting that there may be other genes independently contributing to the resistance. Thus, both RPS4-Ws and RRS1-Ws are required for specific recognition of avrRps4 for plant resistance against Pst-avrRps4.

Figure 5.

Pst-avrRps4 resistance in mutant and transgenic plant lines.
(a) Growth of Pst-avrRps4 in rps4-21, its transgenic lines and rps4-21/rrs1-1 double mutant after inoculation. Leaves of 7-week-old plants were infiltrated with bacterial suspensions (5 × 104 cfu ml−1). Leaves were harvested at 0 (white columns) and 3 days (black columns) after inoculation. The rps4-21/RPS4-Ws-#9 and 11 lines represent independent transgenic rps4-21 plants harboring the RPS4-Ws genomic fragment. rps4-21/rrs1-1 represents a double mutant obtained by crossing rps4-21 and rrs1-1 mutants. The asterisks indicate statistically significant difference from the wild-type (WT) controls (Dunnett’s method, < 0.05). Error bars indicate SEs. This experiment was repeated three times with similar results.
(b) Growth of Pst-avrRps4 in rrs1 alleles and a transgenic line after inoculation. Leaves of 7-week-old plants were infiltrated with bacterial suspensions (5 × 104 cfu ml−1). Leaves were harvested at 0 (white columns) and 3 days (black columns) after inoculation. rrs1-1/RRS1-Ws-#2 represents transgenic rrs1-1 plants harboring the RRS1-Ws genomic fragment. Asterisks indicate a statistically significant difference from WT controls (Dunnett’s method, < 0.05). Error bars indicate SEs. This experiment was repeated three times with similar results.
(c) Growth of Pst in rps4 and rrs1 mutants and WT plants after inoculation. Leaves were harvested at 0 (white columns) and 3 days (black columns) after inoculation. Error bars indicate SEs. The results of statistical test (Dunnett’s method, < 0.05) indicated no significance of differences among means. This experiment was repeated three times with similar results.

Evolutionary conservation of R gene pairs

The function of RRS1-Ws and its neighboring gene RPS4-Ws in tandem is also supported by the evolutionary conservation of the gene pair. Close homologs of RPS4 are often physically paired with homologs of RRS1 in a head-to-head (inverted) tandem arrangement (Gassmann et al., 1999). Using the blast algorithm, we found nine occurrences of this combination in the Arabidopsis genome, indicating the potential selective importance of the inverted tandem arrangement (Figure 6). RRS1 belongs to the Toll interleukin 1 receptor (TIR)-type NB-LRR (TNL)-A clade, and RPS4 belongs to the TNL-B clade (Meyers et al., 2003). Similarly, Arabidopsis RPP2A and RPP2B, which cooperate to specify resistance to the Hyaloperonospora arabidopsis (Ha) Cala2 strain, are also classified into TNL-A and TNL-B clades, respectively (Sinapidou et al., 2004). Therefore, it appears that a paired R-gene system may be a common mechanism in pathogen recognition by plants.

Figure 6.

 Conserved R-gene pairs similar to RPS4 and RRS1 in Arabidopsis-ecotype genomes.
Putative gene products displaying the highest sequence similarity to RPS4 and RRS1. The homologs of RPS4 were physically paired with the corresponding homologs of RRS1, which are arranged head-to-head. The orientation of the genes is indicated by arrows. The distance between start codons is indicated in brackets. The position of the WRKY domain, if applicable, is shown by ‘W’.

Discussion

Colletotrichum species cause devastating anthracnose diseases on a large number of agronomically important crops. Anthracnose diseases can often be controlled by the introduction of genetic resistance traits, but the molecular components of the resistance have remained unknown. Here, we report that the Arabidopsis RCH2 locus encodes two NB-LRR proteins that are both required for resistance to C. higginsianum. Both proteins are well-characterized R proteins involved in resistance against bacterial pathogens: RRS1 confers resistance to strain Rs1000 of R. solanacearum, and RPS4 confers resistance to Pst-avrRps4. Furthermore, we found that both genes RRS1-Ws and RPS4-Ws are required for resistance to Pst-avrRps4, and to strain Rs1002 R. solanacearum. Thus, these two adjacent R genes confer resistance, in tandem or individually, to three distinct pathogens with very different infection strategies and virulence mechanisms.

Several examples of two NB-LRR genes acting cooperatively to confer resistance against a pathogen have been reported. For example, Arabidopsis RPP2A and RPP2B reside adjacently in the RPP2 locus (Sinapidou et al., 2004). Similarly, the tobacco NB-LRR encoding gene N functions with N requirement gene 1 encoding another NB-LRR protein to mediate resistance against a virus (Peart et al., 2005). The blast resistance in Pikm-containing rice is conferred by a combination of two NB-LRR-encoding genes: Pikm1-TS and Pikm2-TS (Ashikawa et al., 2008). Pi5-mediated resistance against the rice blast requires two NB-LRR-encoding genes (Lee et al., 2009). It is not known whether these NB-LRR genes function dependently or independently. However, because of the structural similarity with the RRS1/RPS4 genes, it is possible that resistance to the pathogens is also conferred by cooperation between the two NB-LRR genes.

There are several reports showing that a single R gene/locus can confer resistance to multiple pathogens. For instance, tomato Mi mediates resistance against three distinct types of pests, including root-knot nematodes, potato aphids and sweet potato whitefly (Nombela et al., 2003). It is also known that different alleles of a single R locus can confer resistance to distinct pathogens. For example, RCY1, an R gene against cucumber mosaic virus in Arabidopsis ecotype C24, is allelic to RPP8 against Peronospora parasitica in the Ler ecotype, and HRT against tobacco cauliflower virus in the Dijon-17 ecotype (Takahashi et al., 2002). The Gpa2 locus, which confers specific potato resistance to a small population of the potato cyst nematode Globodera pallida, is tightly linked to the Rx gene responsible for resistance to potato virus X (Van der Vossen et al., 2000). The Gpa2 and Rx proteins share an overall amino acid identity of 88%. In this study, we suggest that two distinct R genes located in a conserved head-to-head organization confer resistance to three distinct pathogen species by acting cooperatively.

Recent studies exploiting natural variations led to the identification and functional analysis of genes underlying ecologically relevant processes and complex traits (Mitchell-Olds and Schmitt, 2006). Our natural variation analysis of RRS1 revealed distinct single nucleotide polymorphisms that are likely to be responsible for the loss of resistance to C. higginsianum. RRS1 alleles from Bur-0, Col-0 and Cvi-0 contain a premature stop codon at S1291. The stop at S1291 results in an 83-amino-acid deletion after the WRKY DNA binding domain, as compared with RRS1-Ws (Figure S1). These C-terminal amino acids have been shown to be important for resistance against R. solanacearum (Deslandes et al., 2002). However, this region is not required for resistance to Pst-avrRps4, as Col-0, which lacks this region, is also resistant (Gassmann et al., 1999). Thus, the recognition mechanism of Pst-avrRps4 is likely to be different from those of C. higginsianum and R. solanacearum.

RRS1 specifically confers resistance to Rs1000 and Rs1002 via the effector protein PopP2 (Deslandes et al., 2003; this study). Deslandes et al. showed that the recognition of the pathogen by the plant is mediated by a direct interaction between RRS1 and PopP2. PopP2 is localized in the plant nucleus, where RRS1 accumulates when PopP2 is expressed. PopP2 also interacts with host RD19, an Arabidopsis cysteine protease that is required for resistance (Bernoux et al., 2008). Because RRS1 is likely to be a receptor of PopP2, and because RPS4-Ws is also required for resistance to R. solanacearum, the simplest model predicts that RPS4 is a downstream component of PopP2-triggered RRS1 signaling, or that RPS4 functions in concert with RRS1. The latter hypothesis is reinforced by the fact that RRS1-Ws is also required for the specific recognition of Pst-avrRps4. Recognition of the pathogen is not likely to be mediated by a PopP2 homolog (Ma et al., 2006), because no YopJ family (Orth et al., 2000; Orth, 2002) members are found in Pst. Instead, RRS1-Ws specifically recognizes AvrRps4, as Pst without avrRps4 grows to about the same level in WT Ws-0, and in the mutants rrs1 and rps4. Similar to the PopP2 case, it is possible that AvrRps4 is also directly recognized by RRS1-Ws, and that RPS4 functions downstream of, or together with, RRS1-Ws in the signaling pathway. Alternatively, the direct or indirect action of AvrRps4 on RPS4 may be recognized by RRS1-Ws. The causative effector from C. higginsianum is likely to be distinct from PopP2 or AvrRps4, as these proteins are not found in ascomycetes. Nonetheless, RRS1-Ws and RPS4-Ws are both required for the resistance to C. higginsianum, and no additive effects were observed in rrs1/rps4, strongly suggesting that these two NB-LRR proteins function closely together (Figure 3b). This is markedly different from the case of the recently identified TAO1 in Arabidopsis, an NBS-LRR gene that additively contributed, with RPM1, to the disease resistance response against Pseudomonas syringae expressing AvrB (Eitas et al., 2008).

The evolutionary conservation of homologous gene pairs in a head-to-head arrangement also supports the idea that the cooperative function of two R genes could be a common mechanism of defense against pathogens. As the two open reading frames are only 264 bp apart, the promoter regions of the gene pairs are likely to overlap one another, leading to the co-regulation of the genes. Interestingly, 10 of the 11 pairs of TNL-A and TNL-B type genes show a head-to-head configuration (Meyers et al., 2003). The duplication of the TNL-A and TNL-B tandem is likely to have taken place after an ancestral duplication event of a TNL-like gene: each of these genes form a monophyletic group (Meyers et al., 2003). The head-to-head configuration may assure balanced levels of the protein pair to meet a strict stoichiometric requirement to act together, possibly in a complex. As a practical application, this finding may provide a new strategy for creating transgenic plants that express R genes from other plants. The introduction of two R genes in a head-to-head orientation may be necessary for effective of pathogen resistance.

Experimental procedures

Plant material

Arabidopsis ecotypes Bay-0, Br-0, Bur-0, C24, Cvi-0, Est-1, Fei-0, Ler-1, Lov-5, Nfa-8, Rrs-7, Rrs-10, Sha, Tamm-2, Ts-1, Tsu-1 and Van-0, a library of T-DNA insertion mutants of Arabidopsis Ws-0 (CS22830) and the cosmid-based Ws-2 (CD4-11) genomic library were obtained from the Arabidopsis Biological Resource Center (ABRC, http://www.biosci.ohio-state.edu/~plantbio/Facilities/abrc/abrchome.htm). Col-0 and Ws-0 were obtained from the RIKEN BRC (http://www.brc.riken.go.jp/lab/epd/Eng/). Arabidopsis plants were grown in professional potting soil No. 2 for cutting (Dio Chemicals, Ltd., http://www.dionet.jp) in a growth chamber at 22°C, until plants were inoculated with a pathogen.

Constructs and Arabidopsis transformation

The 8.2-kbp genomic RRS1-Ws fragment, including the approximately 1.8 kbp upstream and 176 bp downstream region, was amplified from a cosmid clone containing RRS1-Ws by PCR using two specific primers (5′-ACCTCTGACTTGAGCGTCG-3′ and 5′-TGATGGGTTTACAGTTTGGGGAGGACTGGTAATTG-3′), and was cloned into pCR8GW-TOPO (Invitrogen, http://www.invitrogen.com). The resultant plasmid, pCR8GW-RRS1-Ws, was subcloned into the destination vector pGWB1 (a gift from Dr Nakagawa) using the LR cloning reaction.

The binary vector pBI101 (Clontech, http://www.clontech.com) was modified by inserting a multiple-cloning site (HindIII–SacI) from pBluescript SK+ (Stratagene, http://www.stratagene.com) at the HindIII and SacI sites to produce pBI101-SK+. A 6.3-kbp genomic RPS4-Ws fragment, including the approximately 2.1 kbp upstream and 109 bp downstream region, was amplified from genomic DNA of Ws-0 by PCR, using specific primers (5′-TGGGAAATTGATTACAGGATGGACTTCAGG-3′ and 5′-ACTAGAGTCATACATACAAGG-3′). This fragment was cloned into the pBI101-SK+ SmaI site. The transformation of Arabidopsis was performed according to the vacuum infiltration method using Agrobacterium tumefaciens strain GV3101 (Bechtold et al., 1993).

Colletotrichum higginsianum strain and inoculations

Colletotrichum higginsianum isolates (MAFF305635, MAFF305968 and MAFF305970) were obtained from the MAFF Genebank project, Japan (http://www.gene.affrc.go.jp/). After single-spore isolation, cultures of the isolates were maintained on potato dextrose agar (Difco, http://www.bd.com) at 24°C in the dark (Narusaka et al., 2004). Plants that were 28–30-days old (12-h light cycle) were inoculated by spraying the leaves with a spore suspension (5 × 105 spores ml−1 in distilled water) of Chigginsianum. Alternatively, two 5-μl drops of the spore suspension were placed on each leaf. Inoculated plants were then placed in a growth chamber at 100% relative humidity at 22°C (12-h light cycle). Control plants were mock-inoculated with distilled water. Fungal hyphae within the resulting lesions were stained with lactophenol-trypan blue, as described previously (Bowling et al., 1997).

Quantification of Colletotrichum higginsianum ACT mRNA

Plants sprayed with C. higginsianum were harvested at 4 dpi for qRT-PCR analysis. Several leaves were frozen in liquid nitrogen and then ground using an SH-48 grinding apparatus (Kurabo, http://www.kurabo.co.jp). Total RNA was purified using the Trizol reagent and PureLink RNA Mini Kit (Invitrogen). To completely eliminate DNA contaminantion, we treated the total RNA preparation with DNase 1 (Takara, http://www.takara-bio.com) before the washing step of the purification procedure. First-strand cDNA was synthesized using 500 ng of total RNA by using the PrimeScript RT reagent Kit (Takara). QRT-PCR was performed using the Opticon real-time detection system (Bio-Rad, http://www.bio-rad.com) using the SYBR Green PCR Master Mix (Takara). The nucleotide sequences of gene-specific primers for each gene are as follows: At-CBP20 (At5g44200: forward primer, 5′-TGTTTCGTCCTGTTCTACTC-3′; reverse primer, 5′-ACACGAATAGGCCGGTCATC-3′), and Ch-ACT (forward primer, 5′-CTCGTTATCGACAATGGTTC-3′; reverse primer, 5′-GAGTCCTTCTGGCCCATAC-3′). qRT-PCR data for Ch-ACT expression from C. higginsianum and At-CBP20 expression from Arabidopsis were collected as log(copy number), obtained from a standard curve of cycle times as a function of log(copy number). The abundance of Ch-ACT was normalized with that of At-CBP20 in infected samples.

Ralstonia solanacearum strain and inoculations

The R. solanacearum strain 1002 used in this study has been described previously (Mukaihara et al., 2004). Arabidopsis plants were grown in rock wool for 5–6 weeks (12-h light cycle) at 22°C. Root inoculations were performed by cutting approximately 2 cm from the bottom of the rock wool, and the exposed roots of the plants were immersed in a bacterial suspension at a concentration of 5 × 108 cfu (colony forming unit) ml−1 for 24 h. The plants were then transferred to a growth chamber at 25°C (16-h light cycle). Bacterial growth was estimated at 3 days after inoculation, as described by Deslandes et al. (1998). The aerial parts of three plants inoculated with Rs1002 were used for this analysis.

Pst-avrRps4 infections

Pst-avrRps4 has been described previously (Gassmann et al., 1999). These strains were grown in a liquid King’s Broth medium containing kanamycin (25 μg ml−1) and rifampicin (25 μg ml−1). Bacteria were harvested by centrifugation, and the cell pellet was washed once with 10 mm MgSO4, and resuspended in 10 mm MgSO4 to a concentration of 5 × 104 cfu ml−1 for in planta growth assays. Arabidopsis plants were grown in professional potting soil No. 2 for 7 weeks in a growth chamber at 22°C, under an 8-h light cycle. Abaxial leaf surfaces were infiltrated with the bacterial suspension with a needle-less syringe. Bacterial growth (cfu cm−2) was determined at 0 and 3 dpi, using five leaf disks, as previously described (Weigel and Glazebrook, 2002).

Acknowledgements

We thank Y. Hase (Japan Atomic Energy Agency) for providing the F2 progeny from the Ws-0 and Col-0 cross, K. Nakagawa (Shimane University) for providing pGWB1, and N. Hosaka and M. Kanagawa for excellent technical assistance. This work was supported in part by an Industrial Technology Research Grant Program from the New Energy and Industrial Technology Development Organization (NEDO) of Japan, by the Program for Promotion of Basic Research Activities for Innovative Biosciences (PROBRAIN) to YN, by KAKENHI (19580053 to YN, 18780028 to MN and 19678001 to KS), and by The Sumitomo Foundation to MN.

Ancillary