Rice blast, caused by the fungus Magnaporthe oryzae, is one of the most devastating diseases of rice. To understand the molecular basis of Pi5-mediated resistance to M. oryzae, we cloned the resistance (R) gene at this locus using a map-based cloning strategy. Genetic and phenotypic analyses of 2014 F2 progeny from a mapping population derived from a cross between IR50, a susceptible rice cultivar, and the RIL260 line carrying Pi5 enabled us to narrow down the Pi5 locus to a 130-kb interval. Sequence analysis of this genomic region identified two candidate genes, Pi5-1 and Pi5-2, which encode proteins carrying three motifs characteristic of R genes: an N-terminal coiled-coil (CC) motif, a nucleotide-binding (NB) domain, and a leucine-rich repeat (LRR) motif. In genetic transformation experiments of a susceptible rice cultivar, neither the Pi5-1 nor the Pi5-2 gene was found to confer resistance to M. oryzae. In contrast, transgenic rice plants expressing both of these genes, generated by crossing transgenic lines carrying each gene individually, conferred Pi5-mediated resistance to M. oryzae. Gene expression analysis revealed that Pi5-1 transcripts accumulate after pathogen challenge, whereas the Pi5-2 gene is constitutively expressed. These results indicate that the presence of these two genes is required for rice Pi5-mediated resistance to M. oryzae.
THE innate immune response is critical to the survival of plants and animals (Asai et al. 2002; Martin et al. 2003; Nimchuk et al. 2003; Ausubel 2005; Lee et al. 2006). The response is mediated by the detection of pathogen-associated molecular patterns (PAMPs) (also referred to as microbe-associated molecular patterns) or avirulence (Avr) proteins by pathogen recognition receptors (PRRs; also called pattern recognition receptors or disease resistance proteins). In animals, a family of cytosolic PRRs that contain a nucleotide-binding oligomerization domain (NOD) mediates the apoptotic and inflammatory responses critical to protection from pathogen invasion. Plants also contain a set of intracellular PRR proteins, called nucleotide-binding and leucine-rich repeat (NB–LRR) R proteins, which are structurally similar to animal NOD proteins. These plant NB–LRR proteins are characterized by a tripartite domain architecture consisting of an N-terminal coiled-coil (CC) or Toll/interleukin-1 receptor (TIR) domain, a central NB domain, and a C-terminal LRR domain (Hammond-Kosack and Jones 1997; Martin et al. 2003; Ting and Davis 2005; McHale et al. 2006; Liu et al. 2007a) and typically recognize pathogen-derived Avr proteins (also called effectors) (Van der Biezen and Jones 1998; Dangl and Jones 2001; Martin et al. 2003; Innes 2004; Ausubel 2005; Chisholm et al. 2006; Jones and Dangl 2006).
In contrast to this intracellular-type recognition, plants and animals also respond to pathogen molecules present at the cell surface. In animals, the recognition of PAMPs in the extracellular compartment is largely mediated by the Toll-like receptor (TLR) family of proteins, which contain LRRs in the extracellular domain (Brennan and Anderson 2004; Ausubel 2005). The TLR proteins associate with kinases of the non-arginine-aspartic acid (non-RD) class to transduce the immune response (Dardick and Ronald 2006). On the basis of the currently available data for plants, cell surface recognition of PAMPs is mediated by receptor kinases that also fall into the non-RD class of kinases (Song et al. 1995; Zipfel et al. 2004, 2006; Lee et al. 2006).
It has been previously hypothesized that extracellular PRRs form homo- or heterodimers to transduce their function (Ronald 1997; Wang et al. 1998; Torii 2000; Chinchilla et al. 2007). For example, the rice XA21D resistance protein, which encodes a putative secreted LRR, is predicted to interact with an intact receptor kinase to transduce the associated resistance response (Wang et al. 1998). Arabidopsis FLS2 forms a complex with the BRI1-associated receptor kinase to transduce the innate immune response (Chinchilla et al. 2007). It has also been observed that cytoplasmically located NB-LRR R proteins recruit structurally similar proteins to transduce the response (Sinapidou et al. 2004; Peart et al. 2005; Ashikawa et al. 2008). For example, Arabidopsis RPP2-mediated resistance against Peronospora parasitica requires two TIR–NB–LRR proteins (Sinapidou et al. 2004). Similarly, the tobacco TIR–NB–LRR protein N requires tobacco N requirement gene1 (NRG1), encoding a CC–NB–LRR protein to mediate resistance to tobacco mosaic virus (Peart et al. 2005).
Rice blast is one of the most devastating diseases of rice and occurs in all areas of the world where rice is cultivated (Ou 1985). More than 70 blast R genes that confer resistance to geographically different sets of the rice blast pathogen Magnaporthe oryzae isolates have been identified to date (Ballini et al. 2008). For example, Pib confers robust resistance to a majority of the Japanese M. oryzae isolates (Wang et al. 1999). In contrast, Pi37 confers only partial resistance to Japanese isolates but complete resistance to Chinese isolates of the same pathogen (Chen et al. 2005). Hence, the isolation of multiple R genes is required to fully understand the molecular basis of the resistance to rice blast. Such characterization of these genes will facilitate development of agronomically useful rice cultivars through marker-assisted breeding or through transgenic approaches.
To date, a total of nine rice blast resistance genes have been cloned and characterized: Pib (Wang et al. 1999), Pita (Bryan et al. 2000), Pi9 (Qu et al. 2006), Pi2 and Piz-t (Zhou et al. 2006), Pi-d2 (Chen et al. 2006), Pi36 (Liu et al. 2007b), Pi37 (Lin et al. 2007a), and Pikm (Ashikawa et al. 2008). With the exception of Pi-d2, a non-RD receptor-like kinase (Chen et al. 2006; Dardick and Ronald 2006), these genes all encode NB–LRR-type proteins. Distinct features of these cloned rice blast resistance genes have been observed. The Pib protein contains a duplicated NB region (Wang et al. 1999). Pita lacks a classic LRR but contains a leucine-rich domain (LRD) consisting of imperfect repeats of various lengths. A single amino acid difference at the Pita LRD was found to distinguish resistant from susceptible alleles (Bryan et al. 2000). The allelic genes Pi2 and Piz-t show eight amino acid differences within three consecutive LRRs, and these residues are responsible for resistance specificity (Zhou et al. 2006). The Pi9 gene strongly resembles the Pi2 and Piz-t genes and is located within the same region on chromosome 6 (Qu et al. 2006; Zhou et al. 2006). The Pikm-mediated resistance requires two adjacent NB–LRR genes, Pikm1-TS and Pikm2-TS (Ashikawa et al. 2008). Among these cloned R genes, only Pita has been observed to interact with the corresponding M. oryzae avirulence protein, AvrPita (Jia et al. 2000). Thus, defense signaling mediated by NB–LRR-type proteins remains poorly characterized in rice.
It has been reported that Pi5 confers resistance to many M. oryzae isolates collected from Korea and the Philippines (Wang et al. 1994; Chen et al. 2000; Han 2001). To gain a further understanding of the molecular basis of Pi5-mediated rice blast resistance, we used a map-based method to isolate the Pi5 genomic region. We previously mapped Pi5 to a 170-kb interval on the short arm of chromosome 9 in the RIL260 rice cultivar (Jeon et al. 2003). In our study, Pi5 was more precisely mapped to a smaller physical interval using a new mapping population derived from a cross between RIL260 and IR50 lacking Pi5. Through sequence analysis of the Pi5 genomic region, two candidate blast resistance genes were identified on the basis of the presence of CC–NB–LRR domains in the predicted proteins. These two genes were designated Pi5-1 and Pi5-2. We subsequently carried out detailed genetic analysis to determine the function of each of these genes. Surprisingly, Pi5-mediated resistance required the presence of both Pi5-1 and Pi5-2 gene products. In response to pathogen inoculation, Pi5-1 transcripts accumulated. In contrast, the Pi5-2 gene was constitutively expressed.
MATERIALS AND METHODS
The RIL260 rice cultivar carrying the Pi5 allele and a rice blast-susceptible cultivar, IR50, were used as the parental lines. The RIL260 and IR50 cultivars were crossed to generate a mapping population for genetic linkage analysis. Self-pollinated seeds (F2) of the RIL260/IR50 F1 individuals were collected to obtain a sufficiently large mapping population. A japonica rice cultivar, Dongjin, was used as the susceptible control in the M. oryzae inoculation and rice transformation experiments. RIL260 and the monogenic rice line IRBL5-M carrying Pi5 (Tsunematsu et al. 2000; Yi et al. 2004) were used as the resistant control cultivars in the M. oryzae inoculation experiments. An additional eight monogenic rice lines, IRBLi-F5, IRBL9-W, IRBLb-B, IRBLta-K1, IRBLz-Fu, IRBLks-F5, IRBLkm-Ts, and IRBLsh-S, and the susceptible background cultivar of these monogenic lines, Lijiangxintuanheigu (LTH) (Tsunematsu et al. 2000; Yi et al. 2004), were also used in the inoculation experiments to determine the virulence pattern of M. oryzae isolates. Rice seedlings were grown in a greenhouse at 30° during the day and at 20° at night in a light/dark cycle of 14 hr/10 hr.
Pathogen inoculation and disease evaluation:
M. oryzae PO6-6, a Philippine isolate, which is incompatible with the Pi5 resistance locus, has been commonly used to detect this locus (Wang et al. 1994; Jeon et al. 2003; Yi et al. 2004). To analyze blast resistance in Pi5 transgenic rice plants, an additional five different Korean M. oryzae isolates, KJ105a, KJ107, KJ401, KI215, and R01-1, were used. All inoculations and disease evaluations were conducted in the greenhouse facilities at Kyung Hee University using a method that was slightly modified from Liu et al. (2002). Three-week-old plants of the F3 progeny of each of the identified recombinant lines and transgenic plants were used in the inoculation experiments. M. oryzae was grown on oatmeal agar medium for 2 weeks at 24° in the dark. Conidia were induced 4 days prior to collection by scratching the plate surface with a sterilized loop. The inoculated plants were placed in sealed containers to maintain humidity at 24° in darkness for 24 hr and then transferred to a growth chamber at 24° and 80% humidity under a 14-hr/10-hr (light/dark) photoperiod. Disease evaluation was carried out 7 days after inoculation.
Genotypic analysis of progeny from the RIL260/IR50 mapping population:
Cleaved amplified polymorphic sequence (CAPS) markers for C1454 (Jeon et al. 2003) and JJ817 (Kwon et al. 2008) and the sequence characterized amplified region (SCAR) marker JJ803 (corresponding to the previously reported dominant marker JJ80-T3) (Yi et al. 2004) were used for the analysis of the RIL260/IR50 segregating progeny (Table 1). The dominant markers JJ113-T3 and S04G03 were additionally utilized as needed (Jeon et al. 2003; Yi et al. 2004).
Genomic DNA was isolated from young leaves of rice plants using a simple miniprep method (Chen and Ronald 1999). PCR analysis was performed in a final volume of 30 μl (100 pm of each primer, 200 μm each of dNTPs, 10 mm Tris–HCl, pH 9.0, 2 mm MgCl2, 50 mm KCl, 0.1% Triton X-100, and 0.5 units Taq polymerase) using 50 ng of genomic DNA as template. PCR products for the CAPS markers C1454 and JJ817 were subsequently digested with MluI and AseI, respectively, and were then size-fractionated on agarose gels.
DNA sequencing and gene prediction:
RIL260 binary BAC (BIBAC) clones spanning the Pi5 locus were selected for DNA sequencing analysis (Tsunoda et al. 2000; Jeon et al. 2003). Plasmids purified by a mini-preparation (Jeon and Ronald 2007) were partially digested with Sau3AI and separated by agarose gel electrophoresis. The 0.5- to 3.0-kb genomic DNA fragments were isolated using a commercial kit (gel extraction kit, Qiagen), subcloned into the BamHI site of pBluescriptII SK(−) (Clontech), and then transformed into Escherichia coli DH10B by electroporation. For DNA sequencing of each BIBAC clone with a 25-kb average insert size, ∼60 clones were selected and sequenced in one or both directions using the T3 and T7 primers.
Similarity searches against the NCBI database (http://www.ncbi.nlm.nih.gov/) were performed using BLAST (Basic Local Alignment Search Tool). To predict protein-coding gene regions, the Rice Genome Automated Annotation System (RiceGAAS) was utilized (Sakata et al. 2002; http://RiceGAAS.dna.affrc.go.jp/).
Vector construction for genetic complementation experiments:
Genomic DNA regions for Pi5-1 and Pi5-2 were reconstituted by subcloning from BIBAC clones (Jeon et al. 2003). To construct a clone carrying the entire Pi5-1 coding region, a 6.6-kb BamHI–SacI fragment of the JJ80 vector that includes the 0.5-kb predicted promoter was subcloned into the binary vector pC1300intC (GenBank accession no. AF294978). The resulting plasmid JJ104 was digested with BamHI and BstEII and fused to 7.3-kb HindIII–BstEII insert of JJ106 to construct JJ105 with a 5.2-kb promoter region. The 0.5-kb SacI–XhoI fragment was amplified by PCR using primers 5′-GTCCAAAGAGAAATGCGACAACAC-3′ and 5′-CGCTCGAGGTGGCATTTCATCCAATAGGCAAC-3′. The resulting product was inserted into the JJ105 to extend the terminator region, yielding the JJ204 construct carrying the 11,516-bp Pi5-1 genomic region.
The Pi5-2 gene was constructed by the multiple ligation of the following four fragments: a 4.2-kb EcoRI–BglII DNA fragment of JJ113, a 200-bp BglII–ClaI PCR product amplified using the primers 5′-GGATGATGTGATCTGCAGAGAAAC-3′ and 5′-CAGCCTCACTGAAATTGCGAAGCA-3′, a 4.2-kb ClaI–XbaI DNA fragment of JJ120, and an EcoRI–XbaI-digested pC1300intC vector fragment. In the resulting construct JJ117, the promoter region was extended by cloning the 3.7-kb NsiI–EcoRI fragment of JJ120. Finally, by inserting a 0.9-kb extended terminator sequence into the Eco065I site of the JJ142 plasmid, the 13,250-bp entire genomic sequence of Pi5-2 in JJ212 was constructed. The cloned genomic sequences in JJ204 and JJ212 were confirmed by DNA sequencing.
Production of transgenic rice plants:
Genomic clones for Pi5-1 and Pi5-2 were transformed into Agrobacterium tumefaciens EHA105 or LBA4404 by electroporation and introduced into the susceptible rice cultivar Dongjin via Agrobacterium mediation according to an established procedure (Jeon et al. 2000). The transgenic plants (T0) were self-pollinated and T1 seeds were collected. Homozygous Pi5-1 (Pi5-1-63) and Pi5-2 (Pi5-2-74) transgenic lines were then selected from T2 progeny resulting from self-pollination of the T1 lines on the basis of the segregation patterns of the transgenes. F1 plants carrying both Pi5-1 and Pi5-2 were produced from a cross between Pi5-1-63 and Pi5-2-74 lines and self-pollinated to produce F2 plants.
Isolation of Pi5-1 and Pi5-2 cDNAs:
Two preparations of total RNA were prepared from rice leaves collected at 24 and 48 hr after inoculation with M. oryzae PO6-6 using Trizol reagent (Invitrogen). Purified mRNAs were obtained using the PolyATtract mRNA isolation system (Promega) from each set of total RNA and mixed in a 1:1 ratio for cDNA synthesis. cDNAs larger than 0.5 kb were selected by size fractionation via gel filtration, and a cDNA library was constructed with the Uni-ZAP XR vector (Stratagene). This library was then screened via colony blot hybridizations using probes corresponding to the Pi5-1 and Pi5-2 coding regions, a 570-bp HindIII–KpnI fragment of JJ204 and a 589-bp EcoRV–SpeI fragment of JJ212, respectively. Isolated cDNA clones were analyzed by DNA sequencing.
A phylogenetic tree was constructed that included Pi5-1, Pi5-2, and other cloned rice blast resistance proteins. Full-length protein sequences were aligned using Clustal W version 2.0 with default options (Larkin et al. 2007) and then corrected manually using the alignment editor software BioEdit Version 7.0.09 (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). Unrooted phylogenetic trees were generated in MEGA Version 4 (Tamura et al. 2007) by the neighbor-joining method, the Poisson distance method, and the pairwise deletion of gaps, with the default assumptions that the substitution patterns among lineages and substitution rates among sites were homogeneous. Because bootstrapping can provide an estimate of branch point confidence, we adopted 1000 bootstrap replicates to infer the statistical support for the tree.
To examine the changes in transcript accumulation in response to pathogen treatment, leaves from each of 10 RIL260, IRBL5-M, and transgenic rice plants inoculated with M. oryzae PO6-6 were collected at different time periods for RT–PCR analysis. Total RNA was prepared using Trizol reagent and reverse-transcribed with an oligo(dT) primer and a First Strand cDNA Synthesis kit (Roche) (Cho et al. 2006). First-strand cDNA was used in PCR reactions with gene-specific primers. Primers for the rice Actin1 gene and the pathogenesis-related probenazole-inducible (PBZ1) gene (Ryu et al. 2006) were used as internal controls (Table 1). PCR conditions were as follows: 94° for 5 min followed by 28–35 cycles of 94°, 1 min; 56°, 1 min; and 72°, 1 min, with a final extension at 72° for 5 min. Three independent amplifications were performed for each primer set.
Genetic characterization of a 130-kb chromosomal region carrying Pi5:
Previously, the Pi5 resistance gene was delimited to a 170-kb interval between the two flanking markers S04G03 and C1454 on rice chromosome 9. This finding was the result of our previous analysis of two populations generated by crosses between RIL260 carrying Pi5 and a susceptible cultivar, CO39, and between RIL260 and another susceptible cultivar, M202 (Jeon et al. 2003). To further delineate the Pi5 gene, in this study we generated a third mapping population derived from a cross between RIL260 and another susceptible cultivar, IR50. Through PCR screening we found that, among the susceptible cultivars tested, only IR50 contained the dominant marker JJ817, which was also found in the resistant cultivar RIL260 (data not shown). In contrast, we were not able to amplify a PCR product for JJ817 in other susceptible cultivars, including CO39 and M202. We selected IR50 as a mapping parent on the basis of the similarity between the genomic regions for RIL260 and IR50, which we speculated could facilitate recombination in the interval.
To identify rare recombinants within the 170-kb Pi5 locus, a prescreening strategy using the CAPS markers JJ817 and C1454 and the SCAR marker JJ803 (Jeon et al. 2003; Jeon and Ronald 2007) was employed in our analysis of the RIL260/IR50 F2 population. Of the 2014 F2 individuals analyzed, we identified eight recombinants between JJ817 and JJ803, but none between JJ803 and C1454 (Figure 1). Using the dominant markers JJ113-T3 and S04G03 (Jeon et al. 2003; Yi et al. 2004), we subsequently determined the breakage points of the eight recombinants that we isolated in their progeny (F3) plants, which enabled us to distinguish homozygous from heterozygous genotypes. In total, all eight lines were found to harbor recombination events between JJ113-T3 and JJ817.
The disease phenotypes resulting from M. oryzae PO6-6 infection of these eight identified lines were then determined in the F3 progeny in each case. These experiments further delimited the Pi5 gene to a 130-kb interval between the markers JJ817 and C1454 (Figure 1). Our previous and current results indicated that both the JJ803 and the JJ113-T3 markers cosegregate with Pi5-mediated resistance (Figure 1). We were unable to further fine-map the R gene at the Pi5 locus.
Genomic sequence analysis of the 130-kb chromosomal region containing the Pi5 locus:
To identify candidate R genes in the Pi5 locus, seven BIBAC clones, JJ80, JJ98, JJ106, JJ110, JJ113, JJ120, and JJ123, which covered the 130-kb Pi5 region (Jeon et al. 2003), were selected and sequenced. BLAST searches using these sequences against the public databases and also gene annotation analysis using the RiceGAAS program predicted a total of 18 open reading frames (ORFs) at the Pi5 locus in the RIL260 cultivar: seven hypothetical proteins, two NB–LRR proteins, two putative transposon proteins, a putative eukaryotic translation initiation factor, a putative GTP-binding protein, a putative tetrahydrofolate synthase, a putative aldose 1-epimerase, a putative histone H5, a putative cold-shock DEAD-box protein A, and an ankyrin-like protein (Figure 2 and supplementary Figure S1). From this genomic sequence analysis, two Pi5 candidate genes that showed homology with NB–LRR resistance genes were identified in RIL260 and designated Pi5-1 and Pi5-2.
The region of ∼90 kb, from JJ803 to JJ817, of the 130-kb RIL260 Pi5 interval was compared with the corresponding region of the japonica genome represented by the sequenced cultivar Nipponbare (International Rice Genome Sequencing Project 2005; Figure 2). The resulting sequence analysis showed that the Nipponbare Pi5 interval contains two NB–LRR genes, Os09g15840 (a Pi5-1 allelic gene) and a gene that was not identified in RIL260, Os09g15850, denoted Pi5-3. In contrast, Nipponbare lacks the corresponding allele of Pi5-2. Notably, the 5′ upstream sequences of the Pi5-1 allelic genes of RIL260 and Nipponbare were very different, indicating an extreme sequence divergence within the regulatory sequences of these alleles. In addition, we did not observe significant sequence similarity in any other part of the 90-kb Pi5 intervals in RIL260 and Nipponbare (Figure 2). These results suggest that the Pi5 resistance locus has significantly diverged between these resistant and susceptible rice cultivars.
We did not compare the Pi5 resistance locus with that of the publicly sequenced indica rice cultivar 93-11 due to a large gap at this locus (Yu et al. 2002). In an inoculation experiment, we found that both Nipponbare and 93-11 were susceptible to M. oryzae PO6-6 (data not shown), indicating that neither carries the Pi5 resistance gene.
Characterization of transgenic rice plants expressing Pi5 candidate genes:
To determine which one of the two candidate genes, Pi5-1 and Pi5-2, is responsible for the Pi5-mediated resistance to M. oryzae, we used the genomic clones JJ204 and JJ212 carrying Pi5-1 and Pi5-2, respectively, under the control of their native promoters to transform the susceptible japonica rice cultivar Dongjin using Agrobacterium-mediated transformation. RT–PCR analysis of the resulting transgenic lines revealed that 13 of 15 Pi5-1 and 12 of 13 Pi5-2 independently transformed lines expressed their transgenes upon M. oryzae PO6-6 inoculation (Figure 3A). The primary transgenic lines (T0) carrying either Pi5-1 or Pi5-2 were inoculated with M. oryzae PO6-6. Surprisingly, however, none of the 13 Pi5-1 or the 12 Pi5-2 transgenic plants showed resistance to the M. oryzae isolate PO6-6 (Figure 3B). To confirm these results, we inoculated T1 progeny from these T0 lines and found that all progeny were susceptible to the M. oryzae isolate to the same extent as the wild-type control Dongjin cultivar. This indicates that neither Pi5-1 nor Pi5-2 alone confers resistance to M. oryzae PO6-6.
Characterization of transgenic rice plants expressing both Pi5-1 and Pi5-2:
Because recent reports (Sinapidou et al. 2004; Peart et al. 2005; Ashikawa et al. 2008) have demonstrated that the presence of two R genes is required for resistance to pathogen infection, we decided to test plants expressing both candidate genes for blast resistance. We therefore generated transgenic plants carrying both Pi5-1 and Pi5-2 by crossing the highly susceptible homozygous Pi5-1 line #63 (Pi5-1-63) with the highly susceptible homozygous Pi5-2 line #74 (Pi5-2-74). Gene expression analysis revealed that the F1 plants resulting from the cross expressed both the Pi5-1 and the Pi5-2 genes upon M. oryzae PO6-6 inoculation (Figure 3A). Strikingly, the 23 of Pi5-1-63/Pi5-2-74 F1 plants tested all displayed complete resistance to M. oryzae PO6-6. Transgenic lines carrying either Pi5-1 or Pi5-2 were susceptible as previously determined (Figure 3B).
To confirm this finding, we inoculated the F2 progeny plants from the Pi5-1-63/Pi5-2-74 F1 lines with the M. oryzae isolate PO6-6. Of the 72 F2 progeny tested, 37 of these carried both transgenes and conferred resistance to M. oryzae PO6-6. In contrast, F2 progeny carrying either Pi5-1 or Pi5-2 only were susceptible (Figure 3C). RT–PCR analysis demonstrated that the Pi5-1-63/Pi5-2-74 lines expressed their transgenes at levels that were similar to RIL260 before and after M. oryzae PO6-6 inoculation (supplemental Figure S2). To test if Pi5-1 and Pi5-2 are required for resistance to other M. oryzae isolates, we inoculated the transgenic plants with four additional isolates incompatible with Pi5. These isolates displayed distinct virulence patterns on rice lines carrying different single R genes (supplemental Table S1), validating that these are indeed different M. oryzae isolates. We found that transgenic plants coexpressing Pi5-1 and Pi5-2 were resistant to all of the tested M. oryzae isolates. The resistance donor RIL260 and the monogenic line IRBL5-M carrying Pi5 were also resistant to these four isolates. In contrast, Dongjin and plants carrying either Pi5-1 or Pi5-2 only were susceptible to the tested M. oryzae isolates (Table 2). These results demonstrate that the two NB–LRR genes Pi5-1 and Pi5-2 are required for Pi5-mediated resistance to M. oryzae isolates.
The Pi5 monogenic line IRBL5-M is susceptible to M. oryzae KI215 (supplemental Table S1). Genomic sequence analysis indicated that the IRBL5-M genomic region carrying Pi5 is identical to that of RIL260 (data not shown). In addition, RT–PCR analysis further demonstrated that IRBL5-M expresses both Pi5-1 and Pi5-2 at levels similar to RIL260 either before or after M. oryzae PO6-6 inoculation (supplemental Figure S2). On the basis of these results, we hypothesized that transgenic plants expressing both Pi5-1 and Pi5-2 would also be susceptible to M. oryzae KI215. Indeed, our inoculation result showed that transgenic plants expressing both Pi5-1 and Pi5-2 are susceptible to M. oryzae KI215. In contrast, RIL260 was found to be resistant to M. oryzae KI215, indicating that it may contain an additional R gene that confers resistance to this isolate (Table 2).
Characterization and phylogenetic analysis of the proteins encoded by Pi5-1 and Pi5-2:
To isolate the cDNA clones corresponding to both Pi5 genes under study, a cDNA library for RIL260 was constructed with the Uni-ZAP XR vector using mRNA isolated from rice leaves collected at 24 and 48 hr after inoculation with M. oryzae PO6-6. This library was screened using a colony hybridization methodology using the gene-specific regions of Pi5-1 and Pi5-2 as probes. We identified seven and five cDNA clones for Pi5-1 and Pi5-2, respectively. Sequence analysis further revealed that three of the Pi5-1 cDNA clones contained an entire ORF, whereas the others lacked an N terminus encompassing an ATG translation initiation codon. Among the three full ORF clones, the longest clone (#1-7) was fully sequenced. These experiments revealed that Pi5-1 encodes a protein of 1025 amino acids and that the ORF is flanked by 5′- and 3′-untranslated regions of 70 and 220 bp, respectively (GenBank accession no. EU869185; Figure 4, A and B). Sequence analysis of the Pi5-2 clones revealed that three of the five clones contained an entire ORF. Among these, the longest clone (#2-4) was further characterized by sequencing. This analysis indicated that Pi5-2 encodes an ORF of 1063 amino acids and that this ORF is flanked by 5′- and 3′-untranslated regions of 73 and 164 bp, respectively (GenBank accession no. EU869186; Figure 4, A and C).
Comparison of their deduced amino acid sequences revealed that both Pi5-1 and Pi5-2 encode an N-terminal CC, a centrally located NB and LRR, and also C-terminal regions (Figure 4, B and C). A conserved domain search using the Pfam and SMART databases predicted that residue 109–576 of Pi5-1 and 109-567 of Pi5-2 contain an NB domain, which is a signaling motif shared by plant R-gene products (Hammond-Kosack and Jones 1997; Dangl and Jones 2001; Martin et al. 2003; Belkhadir et al. 2004; Liu et al. 2007a). The conserved internal domains characteristic of NB-containing R-gene products were also identified in Pi5-1 and Pi5-2, including the P-loop, kinase-2, RNBS-B, GLPL, RNBS-D, and MHDV domains (Meyers et al. 2003). Additional analysis using the Paircoil2 program (http://groups.csail.mit.edu/cb/paircoil2/) predicted a potential CC domain with a threshold of 0.1 between amino acids 31 and 67 in Pi5-1 and 26 and 87 in Pi5-2 (McDonnell et al. 2006), indicating that these proteins belong to the CC subset of the NB–LRR resistance proteins.
The LRR regions of Pi5-1 and Pi5-2 consist of 24.3 and 22.6% leucine residues, respectively, and contain a series of imperfect repeats (10–12) of various lengths (Figure 4, B and C). Of note, a few repeats of the Pi5-1 and Pi5-2 proteins matched the consensus sequence LxxLxxLxxLxLxxC/N/Sx(x)LxxLPxx observed in other cytoplasmic R proteins (Jones and Jones 1997). The first and third repeat regions of Pi5-1 and the first, third, and sixth repeat regions of Pi5-2 contained the xLDL motif that is conserved in the third LRR of many NB–LRR proteins (Axtell et al. 2001; Meyers et al. 2003; Figure 4, B and C). Notably also, the Pi5-1 and Pi5-2 proteins harbor a unique C terminus that is distinct from those of other NB–LRR proteins (Dodds et al. 2001) and that does not match any known protein motif.
Sequence comparisons between the cDNA and genomic sequences for these R genes revealed that Pi5-1 and Pi5-2 carry five and six exons, respectively (Figure 4A). The Pi5 genes have a larger number of introns within their coding regions compared with other cloned rice R genes that confer resistance to M. oryzae (Bryan et al. 2000; Qu et al. 2006; Zhou et al. 2006; Lin et al. 2007a; Liu et al. 2007b). Furthermore, the Pi5-1 and Pi5-2 genes contain an intron in both RNBS-D and MHDV domains.
Among the known plant NB–LRR proteins, the Pi5-1 and Pi5-2 proteins show relatively high levels of similarity with the wild potato species Solanum bulbocastanum gene Rpi-blb1, which confers broad-spectrum resistance to the oomycete pathogen Phytophthora infestans, the causal agent of late blight (supplemental Figure S3) (Van der Vossen et al. 2003). To further analyze the evolutionary relationship between the Pi5 genes and other rice NB–LRR genes, a phylogenetic tree was constructed for both the Pi5 proteins under study and the other cloned rice blast resistance proteins (supplemental Figure S3). The degree of similarity among these proteins was found to vary considerably and two heterogeneous groups could be recognized, indicating an early divergence in the evolution of rice blast resistance genes. Pi5 genes formed a clade with Pi37, which was separated from another clade containing the blast resistance genes Pib, Pi2/Piz-t, Pi9, Pita, Pi36, and Pikm. In addition, our results showed that Pi5-1 has a relatively close evolutionary relationship with Pi5-2.
Expression analysis of the Pi5-1 and Pi5-2 genes:
To examine whether the expression of the two identified R genes was altered upon pathogen treatment, we performed RT–PCR analysis of these two genes in RIL260, IRBL5-M, and Pi5-1-63/Pi5-2-74 transgenic plants infected with M. oryzae PO6-6 (Figure 5 and supplemental Figure S2). Total RNAs isolated from the leaves of 3-week-old plants harvested at different time points after M. oryzae PO6-6 inoculation were used for this purpose. The results revealed that Pi5-1 expression increased 12 hr after pathogen challenge, whereas the Pi5-2 gene is constitutively expressed at a low level in RIL260 both before and after infection (Figure 5). The IRBL5-M and Pi5-1-63/Pi5-2-74 lines also exhibited similar expression patterns of the Pi5 genes (supplemental Figure S2). These findings indicated that both Pi5-1 and Pi5-2 are expressed during pathogen infection, suggesting that the encoded proteins are also coexpressed. Transcripts of PBZ1, a pathogen-inducible gene, accumulated to high levels in M. oryzae-treated leaves (Figure 5).
Characterization of the Pi5 resistance locus:
From a total of 2014 F2 plants derived from an RIL260/IR50 cross generated in our study, we identified eight lines that had undergone recombination events in the 130-kb chromosomal region carrying Pi5. Each of these recombination events occurred close to the JJ817 marker, which shares similarity with the RIL260 and IR50 rice cultivars (Figure 1). In a previous study, we were unable to detect any recombination in the same Pi5 interval from >2100 segregating plants in the RIL260/CO39 and RIL260/M202 populations (Jeon et al. 2003). This suggests that the similarity in the genomic regions between RIL260 and IR50 reduces the suppression of recombination observed in the RIL260/CO39 and RIL260/M202 crosses.
Given that each screened recombinant was selected from 4028 meiotic events (2014 individuals), the eight recombination events in the 130-kb interval correspond to a genetic distance of ∼0.2 cM, giving a ratio of 650 kb/cM. This is much higher than the average physical/genetic ratio of 260–280 kb/cM estimated for the rice genome (Wu and Tanksley 1993). The most likely explanation for this is the lack of pairing and also subsequent strand-exchange events between homologous parental genomes at the Pi5 locus, which is supported by the significant differences between the DNA sequences of resistant and susceptible rice genomes (Figure 2). In further support of this hypothesis, the results of DNA gel-blot analysis confirmed the presence of Pi5-2 in the resistant RIL260 cultivar and the absence of the corresponding allele in the susceptible Nipponbare cultivar. Conversely, Pi5-3 hybridized with genomic DNA in Nipponbare but not in RIL260 (Figure 2; data not shown). Together with our finding that Pi5-1 is also polymorphic in resistant and susceptible rice cultivars, these data indicate that the Pi5 locus is highly divergent among rice cultivars that are resistant and susceptible to M. oryzae.
Structure of the Pi5-1 and Pi5-2 genes:
Pi5-1 and Pi5-2 belong to a family of CC–NB–LRR R genes and contain unique C-terminal regions consisting of 161 and 280 amino acids, respectively. Some plant NB–LRR genes such as RRS-1R (Deslandes et al. 2002), P2 (Dodds et al. 2001), RPS4 (Gassmann et al. 1999), RPP1-WsA, RPP1-WsB, and RPP1-WsC (Botella et al. 1998) encode proteins with additional domains after the LRR in their C terminus. For example, RRS-1R contains a WRKY motif in its C-terminal region whereas the other proteins listed contain a C-terminal non-LRR domain. The C termini of the Pi5-1 and Pi5-2 proteins did not match either of these known domains, nor have these domains been previously characterized. Future characterization of the functional role(s) of these novel C-terminal regions will provide valuable insights into the mechanism of Pi5-mediated resistance.
Our phylogenetic analysis indicated that the Pi5 genes form a distinct clade that can be separated from another clade containing cloned rice blast resistance genes (supplemental Figure S3). We examined the intron positions in the NB domain of the Pi5 genes to better interpret the phylogenetic relationship between Pi5 and other cloned rice blast resistance genes. Notably, Pi5-1 and Pi5-2 harbor an intron between their RNBS-D and MHDV domains. Pita (Bryan et al. 2000), Pi36 (Liu et al. 2007b), and Pikm (Ashikawa et al. 2008) carry an intron at the immediate N-terminal side of the kinase-2 motif, which is the most common intron position in cereals (Bai et al. 2002). Pib carries an intron between its RNBS-B and GLPL domains (Wang et al. 1999). Pi9, Pi2/Piz-t, and Pi37 do not contain any introns within the conserved NB motif region (Qu et al. 2006; Zhou et al. 2006; Lin et al. 2007a). In addition, Pi5-1 and Pi5-2 appear to have relatively many introns, four and five, respectively, compared with other identified rice blast resistance genes except for Pi36, which carries four introns in its coding regions (Liu et al. 2007b). Pi37 carries no intron within its coding region (Lin et al. 2007a). Pita and Pikm2-TS contain a single intron (Bryan et al. 2000; Ashikawa et al. 2008) and Pib, Pi9, Pi2/Piz-t, and Pikm1-TS have two introns in their coding regions (Wang et al. 1999; Qu et al. 2006; Zhou et al. 2006; Ashikawa et al. 2008). The distinctive number of introns and the genomic positions of Pi5-1 and Pi5-2 are thus consistent with the results of our phylogenetic analysis (supplemental Figure S3), indicating that they indeed belong to the same clade and are distinct from other NB–LRR genes.
Previous genetic studies have indicated that Pi5 and Pii are allelic (Inukai et al. 1996; Yi et al. 2004) and that Pi15 is located at the same locus (Lin et al. 2007b). Sequence analysis of the corresponding genomic DNA fragments in Pii- and Pi15-carrying cultivars and functional characterization of the candidate genes of Pii and Pi15 are underway in our laboratory to address whether these genes are allelic or not.
Disease resistance to M. oryzae requires the presence of both the Pi5-1 and Pi5-2 genes:
To further elucidate the mechanism of Pi5-mediated resistance in rice, we investigated the expression patterns of our two identified Pi5 genes by RT–PCR. There have been some previous reports on induced R-gene expression in response to pathogen inoculation in rice. For example, transcripts of rice Xa1 are detected in leaves 5 days after wounding or inoculation with compatible or incompatible strains of Xanthomonas oryzae pv. oryzae (Xoo) (Yoshimura et al. 1998). Pib transcripts accumulate in response to M. oryzae infection and also in response to altered temperature and darkness (Wang et al. 1999). In contrast, other cloned R genes, including Pita, Pi2/Piz-t, Pi9, Pi36, and Pi37, are expressed constitutively in the absence of pathogen challenges (Bryan et al. 2000; Chen et al. 2006; Qu et al. 2006; Zhou et al. 2006; Lin et al. 2007a; Liu et al. 2007b). Pi5-1 transcripts accumulate in response to pathogen infection, whereas Pi5-2 expression is constitutive (Figure 5). These results suggest that the increased expression of Pi5-1 in response to pathogen challenge in conjunction with the constitutive expression of Pi5-2 is an important aspect of Pi5-mediated resistance in rice.
Our present data indicate that Pi5-mediated resistance to rice blast is conferred by two CC–NB–LRR genes, Pi5-1 and Pi5-2. It was previously reported in a study using genetic complementation experiments in Arabidopsis that two adjacent TIR–NB–LRR genes, RPP2A and RPP2B, are essential for the resistance to the P. parasitica isolate Cala2 (Sinapidou et al. 2004). RPP2A has an unusual structure as it harbors two incomplete TIR–NB domains and a short LRR motif, whereas RPP2B has a complete TIR–NB–LRR structure. Similarly, using a virus-induced gene-silencing system, the CC–NB–LRR protein NRG1 was found to be an essential component of N-mediated resistance against tobacco mosaic virus (Peart et al. 2005). In the absence of NRG1, N-mediated resistance is affected both in the N transgenic Nicotiana benthamiana plants and in N. edwardsonii carrying the N gene. It has been recently found that rice Pikm-mediated resistance is also conferred by cooperation of two independent proteins, Pikm1-TS, a CC–NB–LRR protein, and Pikm2-TS, an NB–LRR protein lacking a CC domain (Ashikawa et al. 2008). These earlier results together with the data presented here suggest that a requirement for the presence of two NB–LRR proteins is more common than has been previously recognized.
The gene-for-gene hypothesis predicts that a single plant R-gene product recognizes a single bacterial Avr gene product (Flor 1971). In support of this hypothesis, it has been shown that the R-gene products rice Pita and flax L directly bind to their cognate Avr proteins (Jia et al. 2000; Dodds et al. 2006; Ellis et al. 2007). In contrast, in several other species it has been shown that R proteins do not directly interact with Avr gene products (Gabriel and Rolfe 1990; Van der Biezen and Jones 1998; Innes 2004; Jones and Dangl 2006). For example, in Arabidopsis, the NB–LRR R proteins RPM1 and RPS2 do not appear to interact directly with the cognate Pseudomonas syringae Avr proteins, AvrRpm1 and AvrB, and AvrRpt2, respectively. Instead, RPM1 and RPS2 are hypothesized to guard the host protein RIN4. Upon pathogen attack, these effectors modify RIN4 (Mackey et al. 2002, 2003; Axtell and Staskawicz 2003; Kim et al. 2005; Jones and Dangl 2006). Similarly, Arabidopsis RPS5 appears to guard the serine/threonine kinase protein PBS1, which is targeted for proteolysis by AvrRphB (Ade et al. 2007). Moreover, the silencing of NRG1 does not block the N-protein oligomerization that has been observed as an early response to the p50 elicitor, suggesting that NRG1 may not be necessary for elicitor recognition.
We speculate that it is unlikely that Pi5-1 and Pi5-2 control independent resistant pathways that act in an additive fashion because our current data clearly demonstrate that transgenic plants carrying either Pi5-1 or Pi5-2 remain highly susceptible to M. oryzae. We instead propose several possible models for the mechanism underlying Pi5-1-mediated resistance. First, Pi5-1 and Pi5-2 may interact with each other directly or indirectly. In this scenario, it is possible that both Pi5-1 and Pi5-2 proteins interact with the corresponding Avr effector. Alternatively, the presence of the corresponding Avr effector may be required to trigger their interaction. Second, either Pi5-1 or Pi5-2 alone may interact with the Avr effector while the other serves as a “guard.” Third, both the Pi5-1 and Pi5-2 proteins may guard a third host protein, which is targeted by the Avr effector.
Our findings demonstrate that two CC–NB–LRR genes, called Pi5-1 and Pi5-2, are required to confer Pi5-mediated resistance to M. oryzae. The future successful cloning of the AvrPi5 effector and investigations of Pi5-1/Pi5-2 interactions will contribute to a more complete understanding of the mechanism of rice Pi5-mediated resistance.
We are thankful to Dahu Chen, Matt Campbell, Kangle Zheng, and Dave Mackill for helpful discussions. We thank Peter Ouwerkerk for providing the pC1300intC vector. This work was supported, in part, by grants from the Crop Functional Genomic Center (CG2111-2); the 21 Century Frontier Program; the Plant Metabolism Research Center; the Korea Science and Engineering Foundation; the World Class University program, the Korean Ministry of Education, Science and Technology; and the U. S. Department of Agriculture. M.-Y.S. was supported by the Korea Research Foundation (KRF-2007-511-F00009).
- Received December 8, 2008.
- Accepted January 13, 2009.
- Copyright © 2009 by the Genetics Society of America