Excavation of Pid3 Orthologs with Differential Resistance Spectra to Magnaporthe oryzae in Rice Resource

Twenty-six orthologs of the rice blast resistance gene Pid3 from cultivated varieties and wild rice accessions distributed in different areas were cloned by allele mining. Sequence analysis showed that while each of the orthologous genes from indica varieties and most wild accessions encodes a complete NBS-LRR protein, each of the proteins encoded by those from japonica varieties and few wild rice accessions presents a premature termination. Eleven of the 26 orthologs were selected for blast resistance testing by transforming into the blast susceptible rice variety TP309, respectively. Inoculation of 23 M. oryzae strains collected from diverse regions of China to the respective transgenic plants revealed that 6 Pid3 orthologs showed susceptible to all the tested strains, while the other 5 orthologs showed differential resistance spectra in a gradually spectrum-widen order as Pid3-W3, Pid3-W4, Pid3-I3, Pid3-W5 and Pid3-I1. Amino acid sequences alignment of these orthologs indicated that the sequence diversities between the blast resistance orthologs were mostly located in the LRR domain such as the substitutions of Q694H,D856H,Q896R,D899E etc. However, the differences between the resistance orthologs and the susceptible ones were mostly located in the NBS domain. The present experiments provide an example of that the ortholog evaluation of plant R genes could be an efficient way to expand the rice blast resistance and some other plant disease resistance as well for breeding.


Introduction
Rice blast, caused by the filamentous ascomycete Mangnaporthe oryzae, is the most devastating fungal diseases affecting rice production [1]. Up to now, at least 78 major blast resistance (R) genes have been identified and mapped genetically on 11 rice chromosomes in exception of chromosome 3. Among the mapped blast R genes, 24 have been cloned [2][3][4][5] and functionally analyzed. With these known and cloned blast R genes, rice blast resistance breeding has become much more effective than before by using molecule markers linked to the known blast R genes [6].
Most of rice blast resistance genes conduct their reactions against a specific part of Mangnaporthe oryzae strains, thus pyramiding different rice blast R genes would facilitate rice breeding towards more durable and broader resistance to rice blast. For example, the plants pyramided with Pi1, Piz-5 and Pi-ta showed enhanced resistance as compared to those carrying one or two of them, and the combination of these three blast R genes could be then deployed into superior rice varieties by marker-aided selection (MAS) [7]. By crossing and backcrossing rice lines C101LAC, C101A51 and Jin 23B, which contain Pi1, Pi2 and Pi33 respectively, Chen et al. introgressed the three blast R genes into a receptor parent Jin 23B. The pyramided lines showed a wide blast resistance spectrum covering 96.7% of the tested blast strains, which was much higher than that of the initial single R gene lines respectively [8]. In 2010, Koide combined two major rice blast resistance genes, Pish and Pib, from two near isogenic lines, the resulted pyramided line expanded the resistance spectrum by 64.3% [9].
However, in the practice of resistance breeding, using a single R gene which has a broad resistance spectrum is more effective. Of the known rice blast R genes, Pi1 [10], Pi2 [11], Pib [12], Pi9 [13], Pi5 [14], Pi-ta [15], Pik [16] and Pik-p [17] have been characterised as R genes with a relatively broad blast resistance spectum, respectively. Though the most other blast R genes may have a relatively narrow resistance spectrum, several experiments have confirmed that their alleles or orhtologs showed varied blast resistance spectra. These allelic or ortholog genes would be born with abundant allelic variations in rice resources due to the coevolution of rice and blast pathogen in different rice growing environments, and could be utilized as more efficient and important gene resources in the improvement of rice blast resistance [18]. For example, Pi9,Pi2 and Piz-t are alleles from different rice resources while physically at the same gene locus on rice chromosome 6 with the sequence similarity up to 98.84%, but their resistance spectra are quite different, covering 93.7%, 92.2% and 54.5% of tested blast isolates, respectively [4,19,20]. The similar situation was also observed in Pik, Pik-m, Pik-p, Pi1, Pi54 and Pi54rh [3,5,16,17,21,22], which were all from the same chromosome locus but cloned independently from different varieties and showed differential resistance spectra [23].
Pid3 (GenBank accession no.: FJ745364.1), was initially identified in an indica variety Digu by performing a genome-wide comparison of '9311' (indica) and 'Nipponbare' (japonica) on the premise of the verification of obvious different resistance of indica and japonica varieties to M. oryzae strains collected from south and north China [24], and the Pid3 functional orthologs were found widely present in most tested indica varieties and wild rice species [25]. Later, Lv et al [2] isolated an ortholog of Pid3 from a common wild rice accession A4 (O. rufipogon) by sequencing-based allele mining, which confers a different resistance spectrum to a set of M. oryzae strains and was named Pid3-A4. The purpose here is to isolate more ortholog genes of Pid3 from cultivated varieties and wild rice accessions, and to evaluate their respective blast resistance by gene transformation and blast inoculation. Based on the respective blast resistance spectra of the cloned orthologs, comparative analysis were conducted between the amino acid polymorphic sites of the identified NBS-LRR proteins and their respective blast strain-specific resistances.

Plant materials and rice blast strains
There are 10 cultivated rice varieties including 5 indica varieties and 5 japonica varieties used in this study, collected from south or north of China and other countries. The 10 common cultivated rice varieties are kept in our lab, which are used and planted in China widely. Seventeen wild rice accessions were presented by the professor Zhukuan Cheng's lab, Institute of Genetics and Developmental Biology and Lili Hao's lab, Beijing Institute of Genomics (Table S1 in File S1). The species were cultivated in an experimental field of the Institute of Genetics and Developmental Biology in Beijing under normal growing conditions. Twenty-three M. oryzae isolates used in this study are listed in Table 1. Of them, 20 isolates were collected from south of China and other 3 isolates from north of China. Zhong-10-8-14 was the initial isolate for the determination of the Pid3 gene [25].

Gene cloning
DNA was extracted from the fresh leaves of 27 cultivated and wild rice varieties. Forward primer Pid3F: 59 -TTTCTAGAAG-TAACACCCAAGGATAGGATAG -39 and reverse primer Pid3R: 59 -CTGTCGACGAACGACAAGTGCGACAT-GATTG -39 were designed and used to amplify the full coding sequence of Pid3 based on the published dada [2]. An XbaI and a SalI recognition site (underlined) with two protecting bases (TT and CT) were added to their 59 ends, respectively. PCR amplification was carried out using the following profile: initial DNA denaturation, 95uC for 4 min; followed by 30 cycles of denaturation, 95uC for 30 s; annealing, 58uC for 30 s; extension, 72uC for 3 min; and final extension at 72uC for 5 min. The PCR products were purified and sequenced respectively.

Rice transformation
No intron is present in the Pid3 gene. Therefore, the cloned gene fragments above were inserted into the binary vector pZH01 [26] through the XbaI and SalI cloning sites. After sequence verification, the final construct was introduced into Agrobacterium tumefaciens LBA4404. Agrobacterium-mediated transformation was performed using calli derived from mature embryos of susceptible rice variety TP309 according to Hiei et al [27].
Positive transgenic plants were detected through the amplification of the marker gene hygromycin (HYG) in vector using forward primer HYG-F: 59 -TGCGCCCAAGCTGCATCAT -39 and reverse primer HYG-R: 59 -TGAACTCACCGC-GACGTCTGT -39. PCR amplification was carried out using the following profile: initial DNA denaturation, 95uC for 4 min; followed by 35 cycles of denaturation, 95uC for 30 s; annealing, 58uC for 30 s; extension, 72uC for 30 s; and final extension at 72uC for 5 min. To confirm the positive transformants precisely, one cleaved amplified polymorphic sequence (CAPS) marker was designed based on the nonsense mutation locus of Pid3 in TP309 variety. A 658-bp fragment was amplified using the primer Pid3C (Pid3C-F: 59 -TACTACTCATGGAAGCTAGTTCTC -39 and Pid3C-R: 59 -ACGTCACAAATCATTCGCTC -39) and then digested with BamHI. The digested PCR products was resolved on 2% agarose gels as one band in TP309 variety and three premature termination Pid3 orthologs of Pid3-W9, Pid3-W14, and Pid3-J1 transformants while two bands in the other positive transgenic plants.
Expression analysis of Pid3 orthologs RNA were isolated from leaf sheath tissue with the TRIzol reagent (Invitrogen, Carlsbad, CA) and cDNA was synthesized from poly(A) + RNA using a cDNA synthesis kit (Transgen, Beijing). Semi-quantitative reverse-transcription (RT)-PCR was performed with the specific primer pair Pid3C for 30 cycles of amplification. Transcription of the Actin gene was used to normalize the cDNA levels with the primer pair 59 -AGCAACTGGGATGATATGGA -39 and 59 -CAGGGC-GATGTAGGAAAGC -39. Amplification of the Actin gene was conducted for 27 cycles. And the over-expression of Pid3 gene was detected by the primer pair Pid3D (Pid3D-F: 59-GAATG-CAAATGTTTGGTTCG-39 and reverse primer Pid3D-R: 59-CGCCACATCATAATTCCTTG-39). RT-PCR was initiated with one cycle at 95uC for 2 min followed by 30 or 27 cycles at 95uC for 30 s, 58uC for 30 s, and 72uC for 45 s, and the reaction was terminated with a final extension at 72uC for 5 min. The PCR products were resolved on 1% agarose gels.

Fungal inoculation
Rice seedlings at the tillering stage were inoculated by injection of 0.1-0.2 ml of a 2.5610 5 conidia ml 21 spore suspension in the field. The disease reaction was evaluated seven days after inoculation with the susceptible transgenic line, Pid3-J1, as a control. Leaves lesion types were observed and scored as resistance (R), medium resistance (mR), medium susceptibility (mS), and susceptibility (S), respectively. MEGA 5 [30] was used to build phylogenetic tree of Pid3 orthologs. Sliding-window analysis, polymorphism and neutral tests of different regions of the Pid3 orthologs were conducted using DnaSP version 4.0 [31].

Sequence characteristics of the Pid3 orthologs
In addition to previously reported Pid3 (hereafter renamed as Pid3-I3) from the indica variety Digu [25] and Pid3-A4 (hereafter renamed as Pid3-W5) from the wild rice accession A4 (O. rufipogon) [2], a total of 28 Pid3 orthologs were cloned, including 19 from 16 wild rice accessions (the additional three orthologs were from the wild rice accession W1, W11 and W13, respectively), 4 from indica and 5 from japonica varieties by allele mining. The details for the varieties are listed in Table S1 in File S1. They are distinguished in geographic distributions.
The open reading frame (ORF) analysis revealed that 26 Pid3 orthologs each consist of 2775 nucleotides while the other two from W1-2 and W16 each consist of 2778 nucleotides. The pairwise alignment of all 30 Pid3 orthologs showed high homologous with an average identity from 99.0% to 100% at the DNA level. Of note, the ortholog sequences from W12 and I2 are identical, and those from J3 and J5 and from W11-1, J2 and J1 are the same too. Therefore, we have actually obtained 26 different Pid3 orthologs.

Nucleotide polymorphism of Pid3 orthologs
Pid3 genes encode a typical CC-NBS-LRR type protein with the CC, NBS and LRR domains coded in positions of 1-300, 481-1560 and 1621-2670, respectively, along the DNA sequence. By sliding window analysis of the 26 Pid3 orthologs, the nucleotide diversity distribution in the whole encoding sequences was revealed in Figure 1, showing two obvious characteristics: firstly, a variation peak exists in the NBS domain; secondly, the nucleotide diversities of Pid3 orthologs among the wild rice accessions are obviously greater than those among the cultivated rice varieties.
Further analysis was conducted in terms of different regions of Pid3 orthologs by the neutral tests (Table 2). Tajima's D values were mostly negative but not significant, suggesting that Pid3 orthologs may experience a little preference to positive selection but not balancing selection. The smaller p values (less than 0.5%) also showed the less variations for Pid3 orthologs, while the nucleotide diversity in LRR domain of Pid3 gene was larger than that in the CC and NBS domain. In Pid3 orthologs of wild rice  Amino acid sequence analysis of Pid3 orthologs A neighbor-joining tree was constructed with the encoded amino acid sequences to evaluate the phylogenetic relationships of Pid3 orthologs ( Figure 3). Indica, japonica and wild rice varieties were relatively clustered in groups separately. The orthologs of most wild rice accessions were far from those of the cultivated varieties, but few ones were present between indica and japonica  groups, which might reveal their possible genetic correlation in evolution.

Blast resistance of Pid3 orthologs
Eleven Pid3 orthologs, including 3 prematurely terminated Pid3 orthologs from wild rice accessions of W9, W14 and japonica variety of J1 and 8 complete Pid3 orthologs from wild rice accessions of W3, W4, W5, W8, W10, W16 and indica varieties I1 and I3, were selected for evaluation of their respective resistance to a group of rice blast strains based on their geographic distribution. To ensure the consistency, each of the Pid3 orthologs were inserted into the binary vector pZH01 under Cauliflower mosaic virus (CAMV)35S promoter control and transformed into the susceptible rice variety TP309, which was the same recipient used for pCaMV35S::Pid3 [25] and pCaMV35S::Pid3-A4 [2] in our previous study. The independent primary transgenic lines (T0) were obtained and determined by the transgene CAPS marker ( Figure 4A) and the transgene transcripts ( Figure 4B). The results showed that all the candidate transformants were transgenepositive and over-expressed. The number of the positive transgenic plants in T0 generations was shown in Table S2 in File S1. Further, the transgenic lines, in which the respective transgenes were expressed at the approximately same levels ( Figure 5), were selected for co-segregation analysis in their respective T1 (the selfed progeny of T0 line) generations (there was an example in Figure S1 in File S1) and then homozygous transgnic T2 (the selfed progeny of T1 line) lines were identified. The homozygous T2 lines from each of the 11 Pid3 orthologs were applied to subsequent blast inoculation assays. As we expected, all of the Pid3 orthologs transgenic plants did not present any observable side effects on phenotypes.
Next, all Pid3 orthologs homozygous were inoculated with 23 M. oryzae strains respectively, which were collected from south and north of China. These blast inoculation experiments repeated 3 times on average and the results were listed in Table 1. Five completely encoded Pid3 orthologs, Pid3-W3, Pid3-W4, Pid3-W5, Pid3-I1 and Pid3-I3, were resistant to Zhong-10-8-14, the evaluating strain of the Pid3 gene [2], while the other three completely encoded Pid3 orthologs (Pid3-W8, Pid3-W10 and Pid3-W16) and three premature terminated Pid3 orthologs (Pid3-W9, Pid3-W14 and Pid3-J1) were susceptible ( Figure 6). In the further inoculation test, the 6 Zhong-10-8-14 susceptible Pid3 orthologs also showed susceptibility to the other 22 M. oryzae strains, whereas the 5 Zhong-10-8-14 resistant Pid3 orthologs displayed differential blast resistance spectra of the remaining strains: Pid3-I1, Pid3-W5, Pid3-I3, Pid3-W4 and Pid3-W3 were resistance to 19, 17, 13, 10 and 5 M. oryzae strains, respectively.  Of the 8 completely encoded Pid3 orthologs, 5 showed resistance to rice blast but with different spectra while the other 3 ones showed completely susceptible to all strains tested. In comparison of the protein sequences encoded by the two-group genes (resistance and susceptibility), the polymorphic loci were found mostly located in the NBS domain (Table 3), such as the G204C substitution in the Kinase1a motif of Pid3-W16 and the P441L substitution in the RNBS-D motif of Pid3-W10, indicating that the varied amino acids are important to the NBS domain's function.
Among the 5 resistance Pid3 orthologs with different resistance spectra, the polymorphic loci were mostly present in the LRR domain, implying that the region may function in recognizing the special blast strains. By comparison of the LRR domain sequences between the resistance orthologs, some amino acids could be recognized as important to the blast resistance spectra. For example, Q694H substitution that was only present in Pid3-W3, which showed the narrowest resistance spectrum; the D899E substitution between Pid3-I1 and Pid3-W4 led to a decrease of the resistance spectrum from 86.95% (Pid3-I1) to 47.82% (Pid3-W4); the substitutions, Q694H of Pid3-I3 and Q896R of Pid3-W3, altered their resistance spectrum from 60.86% (Pid3-I3) to 26.08% (Pid3-W3); and notably, the D856H substitution in either of Pid3-I3 and Pid3-W3, could result in a relatively narrow resistance spectrum.

Allele mining is an important way to expand the rice blast resistance
With the proceeding of the rice genome sequencing projects and the map-based cloning technology, over 20 rice blast resistance  genes have been identified, but, in recent years, the number of new blast resistance loci have not increased as much (Figure 7). Some of the newly found genes were finally determined as an allele or an ortholog to one of the previously cloned genes [3][4][5]32]. The blast resistance alleles or orthologs showing different resistance spectra from each other have thus greatly expanded the rice blast resistance resource for breading. This fact indicates that sequencing-based allele mining, as initially termed by Kumar et al [33], should be an efficient and economical way to expand the roles of each blast resistance loci in rice genome.
Some studies involving the silencing or over-expression of plant R genes have shown that disease resistance is correlated with Rgene expression, thus we have to admit that the acquisition of promoter sequences may play a crucial role to the functions of the involved R-genes in evolution. For example, a recent reported rice blast resistance gene, Pit, may be a result of the refunctionalization of some 'sleeping' rice blast R genes by transposon-mediated transcriptional activation in rice genome [34]. Here in current study, our research purpose was to evaluate and compare differential resistance spectra of Pid3 ortholog sequences cloned directly by allele mining. Since the cloned ortholog sequences only involved their respective ortholog coding sequences, all the Pid3 ortholog coding sequences were driven by the same CaMV35S promoter rather than their respective native promoters and were respectively expressed in transgenic rice plants of TP309 with much the same expression levels. This approach is an easy and effective way to explore possible available Pid3 orthologs in spite of that some of the orthologs might be not functional in their own rice sources. Of course, we are unable to exclude the possibility that stronger expression of a Pid3 ortholog in other rice plants might change its inherent blast resistance spectrum though such unusual situation did not occur in our previous studies [2,25].
Thus, by this way, a total of 26 Pid3 orthologs were cloned respectively from wild rice accessions and cultivated rice varieties distributed in different geographical areas, respectively. Eleven of them were tested for the rice blast resistance and showed different blast resistance spectra to these Chinese M. oryzae strains. The blast resistance spectra of Pid3-I1 from Kasalath (indica) and Pid3-W5 from O. rufipogon (A4) increased up to 86.95% and 78.26% respectively, as compared with the corresponding spectrum (60.86%) of Pid3-I3, the original Pid3 gene from Digu. Noteworthily, Pid3-I1, which is from an Indian variety Kasalath, showed the best resistance to Chinese M. oryzae strains.

Wild rice is an important resource for blast resistance
Sun et al [35,36] reported that during rice domestication the number of alleles of cultivated rice was only 60% that of wild rice, leading to lower genetic diversity of the cultivated rice [37]. The two well-known rice R genes, Xa21 for resistance to Xanthomonas oryzae and Pi9 for resistance to Mangnaporthe oryzae were originated from the wild rice O. longistaminat [38] and O. minuta [39], respectively. They both showed broader-spectrum resistances to pathogens and have been widely applied to rice breeding. Thus the two R genes have set good examples for exploring more disease resistance genes in wild rice resources.
In plant, R genes were classified into four types according to their polymorphism level (p value): conserved (type I; p,0.5%), intermediate-diversified (type II; 0.5%,p,5%), highly diversified (type III; p.5%), and present/absent genes (type IV; P/A) [40,41]. The higher p value means the evolution of the gene responding to the pathogen more rapidly. R locus with a low level of polymorphism usually has a relatively slow evolutionary rate. Pid3 orthologs widely exist in the genomes of either cultivated rice subspecies or wild rice species. The present study showed that the p value, the nucleotide diversity, of Pid3 orthologs in wild rice accessions is 0.651% (intermediate-diversified class) while 0.372% (conserved type) in cultivated rice varieties. Obviously, the diversity of wild rice is larger than that in cultivated rice, which supports that Pid3 gene encoding sequences gradually tend to unity  Table 3. Comparison of Pid3 proteins. Pid3-I1 during the domestication of rice species. It also indicates that wild rice accessions are worthwhile for further allele mining of broadspectrum Pid3 gene orthologs.
In this study, we compared the blast resistance spectra of 5 Pid3 orthologs, of which 3 were from wild rice accessions. As mentioned above, Pid3-W5 had a blast resistance spectrum 17% wider than the original Pid3-I3 from Digu. And Pid3-W4 showed a resistance spectrum which is complement to that of Pid3-I3. Therefore, allelic mining from the wild rice is an effective way to expand rice blast resistance spectra in breeding.
The amino acids variations in Pid3 protein domains relative to blast resistance and resistance spectra In the neighbor-joining tree constructed with the amino acid sequences of all Pid3 proteins (Figure 3), the identified 5 resistant Pid3 orthologs (Pid3-W3, Pid3-W4, Pid3-W5, Pid3-I1, Pid3-I3) are aligned in a branch, suggesting that some conservation in Pid3 orthologs sequences is necessary for their resistance function. Based on such conservation, we may predict that in this group ( Figure 3) the other two members, Pid3-W7 and Pid3-W15, would have a blast resistance function too, though this prediction need to be further confirmed.
Many evidences showed that NBS domain in the resistance NBS-LRR proteins functions as a molecular switch in processing the resistance reaction, which decides the resistance level of the R genes [42]. In this study, of the 8 Pid3 orthologs, each encoding a complete NBS-LRR protein, five were tested to be blast resistant while the other three were blast susceptible. When the susceptible protein sequences were compared with the resistance ones, the altered amino acids were found mainly in the NBS domain (as shown in Table 3, of total 13 substituted amino acids in the susceptible gene group, 11 were in NBS domain). Notably, only two substitution sites, G204 and, P441L are located in the conserved Kinase1a motif and RNBS-D motif, respectively, which suggest that amino acid substitution in other region of the NBS domain, besides the two conserved motifs, may also have the lossof-function effects on the blast resistance of NBS-LRR proteins.
Many studies suggested that the LRR domain of R proteins is involved in avirulence (AVR) recognition or in physical interaction with AVR proteins, such as the recognition between barley Mla1 and powdery mildew AvrMla1 [43], potato Rx and PVX [44]; Pita and blast AVR-Pita [45]; flax L and rust AVR-L567 [46]. However, in some cases, the CC domain was also found to play a major role in the interactions between R and AVR proteins. For example, in rice the Pik and AVR-Pik interaction involves the CC domain, but not the LRR domain, of the R protein [47]. Besides, Chen et al. [48] reported that the potato RB (also known as Rpi-blb1) gene encoded NBS-LRR protein binds to Phytophthora infestans AVR protein IPI-O through its CC domain.
In this study, we found that the amino acid variation in the LRR domain of the allelic Pid3 proteins greatly affects their resistance spectra. Within the 5 resistance Pid3 orthologs, a total of 10 amino acid variations exist in the LRR domain, and accordantly their resistance spectra range from 86.95% to 26.08% of 23 tested blast isolations. Compared with the LRR domain of the original Pid3 allele from Digu (Pid3-I3), which showed a resistance spectrum of 60.86% (14/23), two substitutions of Q694H and R896Q in Pid3-W3 decreased the resistance spectrum to 26.08% (6/23); the substitutions of F815Y, H856D, V894I, R896Q, and E899D in Pid3-W4 decreased the resistance spectrum to 47.82% (11/23). Meanwhile, some mutation sites increased the recognition to M. oryzae avirulence (AVR) elements and broadened the resistance spectrum, such as the substitutions of F815Y, H856D, V894I, and R896Q in Pid3-I1 increased the resistance spectrum to 86.95% (20/23), the substitutions of S589L, N799S, V851M, H856D, R896Q, and R900Q in Pid3-W5 increased the resistance spectrum to 78.26% (18/23). The data here may provide important clues for further exploring the core AVR recognition elements in the Pid3 proteins and understanding the interaction mechanism. Nevertheless, our study is preliminary, and the present data cannot exclude the possible role of the CC domain in the AVR recognition by Pid3 proteins.
Up to now, researches on the defense mechanism of rice blast caused by Magnaporthe oryzae have made great progress, resulting in identification and cloning of many defense related genes, which had been expected to facilitate the rice disease resistance breeding program via marker assisted selection and transgenic approaches [49]. Nevertheless, the actual progress in application of these defense related genes to breeding is rather limited. After evaluating the modification of the expression of more than 60 defense-related genes, Delteil et al [50] concluded that altered expression of genes involved in resistance signal transduction and transcription could lead to many unwanted side effects, like lesion mimic phenotypes, in contrast to R genes which could confer broad spectrum and high levels of resistance without obvious side effects. In this research, all Pid3 orthologs transgenic plants though with the character of an over-expressed transgene did not present any observable side effects (data not shown). So far most of rice blast R genes have been cloned by map-based cloning approach, but this approach is labor-cost and time-consuming. The study here demonstrates that sequence-based allele mining (or coding sequence-based mining) is a more effective way to expand the rice blast R gene resource for controlling the devastating blast disease. This approach can be applied to control of other crop diseases too.

Supporting Information
File S1 Supporting information file containing Tables S1, S2 and Figure S1. Table S1. Rice varieties and wild rice accessions used in this experiment. (DOC). Table S2. The number of the positive transgenic plants in T0 generations. (DOC). Figure  S1. Co-segregation of the Pid3-I1 gene with the resistant phenotype. The T1 generations (the selfed progeny of T0 line) carrying Pid3-I1 were inoculated with M. oryzae Zhong-10-8-14, and the genotypes were analyzed using the HYG gene and marker Pid3C. Susceptible TP309 (J1) was used as the control. (TIF). (RAR)