An LRR-truncated NLR natural variant confers dual resistance to drought and blast in rice Authors: Haifu Tu, Ying Ye, Huaijun Wang, Shumin Li, Zhiyue Feng, Xinhua Feng, Qian Liu, Xiaokai Li, Tiantian Ye, Yu Zhang, Dan Hu, Yu Yan, Yu Chen, Yongzhong Xing, Honghong Hu, Faming Dong, Xuelei Lai, Kabin Xie, Haiyan Xiong, Lizhong Xiong Published: 2026 ABSTRACT Climate change threatens global food security by imposing concurrent abiotic and biotic stresses on major crops such as rice. Genetic resources conferring dual resistance to drought and blast disease are rare, limiting the development of climate-resilient crops. Here, using a genome-wide association study, we identify NLR8-Hap2, a natural allele of the rice resistance gene NLR8 that enhances resistance to drought and blast disease. A single-base deletion in the NLR8-Hap2 coding sequence generates a premature stop codon, resulting in a protein lacking the C-terminal leucine-rich repeat (LRR) domain. The dual protective effects of NLR8-Hap2 arise from the enhanced association with and stabilization of OsbHLH148 and OsBIHD1, which regulate drought tolerance and blast defense, respectively. Introgression of NLR8-Hap2 into elite indica and japonica rice cultivars significantly improves dual-stress resistance without detectable yield penalties under normal field conditions. Together, these findings highlight NLR8-Hap2 as a genetic resource for developing multi-stress-resilient rice. Introduction Global climate change is increasing the frequency and severity of concurrent abiotic and biotic stresses, posing a significant threat to agricultural productivity and food security1,2. Rice (Oryza sativa L.), a staple food for over half the world’s population, is particularly susceptible to the combined impacts of drought and diseases such as blast, caused by Magnaporthe oryzae3,4. A major challenge in improving stress resilience in plants is the frequent genetic trade-off between biotic and abiotic stresses, in which enhancing tolerance to one often increases susceptibility to the other. This is exemplified by the ACQOS locus in Arabidopsis, where natural alleles that enhance disease resistance reduce osmotic stress tolerance5. Related studies in rice illustrate both antagonistic and coordinated outcomes: the OsSGS3–tasiRNA–OsARF3 module mediates a trade-off between thermotolerance and disease resistance, whereas the ACL1–ROC4/5 complex synergistically regulates drought tolerance and brown planthopper resistance6,7. Therefore, identifying genetic resources that can overcome this trade-off to confer multi-stress resilience without compromising yield is critical for developing “Green Super Rice” and ensuring future agricultural sustainability3,8. Intracellular nucleotide-binding leucine-rich repeat (NLR) proteins are pivotal immune receptors that detect pathogen effectors and trigger effector-triggered immunity (ETI)9,10. Canonical NLRs typically comprise an N-terminal signaling domain (e.g., coiled-coil [CC] or Toll/interleukin-1 receptor [TIR]), a central nucleotide-binding (NB-ARC) domain that binds ADP in the inactive state and undergoes nucleotide-dependent conformational changes upon activation, and a C-terminal leucine-rich repeat (LRR) domain often involved in effector recognition and auto-inhibition11–13. Structural studies of selected CNLs, including ZAR1 and the wheat NLR Sr35, further show that activation can involve interdomain rearrangements and oligomerization into resistosomes14,15. NLRs achieve functional diversity through complex molecular interactions16. Some operate as pairs, such as the RGA4/RGA5 sensor-executor system17, or as antagonistic pairs like PigmR/PigmS that balance immunity and yield18. Others extend their function by “guarding” key cellular hubs; for instance, the rice NLR PigmR protects the deubiquitinase PICI1 from effector-mediated degradation to sustain defense metabolism19. NLRs can also engage transcription regulators to trigger plant immunity. The rice NLR Pb1 stabilizes the TF WRKY45 to amplify defense signaling20, and PigmR interacts with the RRM-type TF PIBP1 to activate defense gene expression21. In Solanaceae, activated CC domains of several CNLs disrupt MED10b–MED7–JAZ-mediated repression and thereby activate jasmonate-responsive defense transcription22. The tomato NLR Sw-5b also directly engages the PP2C–SnRK2 ABA regulatory module to activate antiviral immunity23. However, a comprehensive understanding of how NLRs mechanistically regulate transcriptional partners and their stability to fine-tune stress responses remains a key objective. While NLRs are primarily known for their roles in immunity, emerging evidence suggests their involvement in abiotic stress tolerance. For instance, PibH8 protects phenylalanine ammonia-lyase from E3 ligase-mediated degradation, thereby enhancing lignin and flavonoid biosynthesis and conferring both drought tolerance and blast resistance24. The specific NLR gene RGA4L (Disease Resistance Gene Analog 4-Like) confers chilling tolerance, likely by modulating the OsLEA5 pathway25. Previous studies have shown that truncation or derepression of NLR-related immune receptors can alter immune outputs. For example, truncated NLR-related proteins such as RBA1 and TN2 can activate defense responses, and derepression of the paired TIR-NLR receptor complex RPS4/RRS1 also triggers immunity26–28. Conversely, some truncated or non-canonical helper NLR-related proteins, such as NRG1C/NRG1.3 and ADR1-L family members, have distinct regulatory effects29,30. These findings indicate that structural diversification of NLR-related proteins can lead to different functional outcomes. However, it remains unknown whether a natural truncation can serve as an evolutionary innovation to coordinate two mechanistically distinct stress-signaling pathways by targeting different downstream TFs. The molecular principles governing how such a natural truncation enhances dual-stress adaptation are yet to be discovered. Here, we report the identification of NLR8-Hap2, a naturally occurring NLR allele that is significantly associated with drought resistance in rice through a genome-wide association study (GWAS). A single-base deletion in the coding region of NLR8 generates a premature stop codon, resulting in an NLR variant lacking the C-terminal LRR domain. Genetic, interaction, and protein-stability analyses support a model in which the LRR-truncated NLR8-Hap2 allele shows enhanced association with and stabilization of two functionally distinct transcription factors (TFs): OsbHLH148, a regulator of jasmonate-mediated drought response31, and OsBIHD1, a regulator of ethylene-mediated blast defense32. The genetic data support a model in which NLR8-Hap2-associated drought and blast resistance depend on downstream branches involving OsbHLH148 and OsBIHD1, respectively. Introgression of this allele into elite rice cultivars improves dual-stress resilience without detectable yield penalties under the tested normal field conditions, supporting its breeding potential. Together, these findings present NLR8-Hap2 as a specific example of a natural NLR allele encoding an LRR-truncated protein that enhances TF-associated stress responses and provides a potential genetic resource for engineering climate-resilient crops. Results NLR8 is associated with drought resistance To identify loci conferring drought resistance in rice, we performed a GWAS on a diverse panel of 240 rice accessions, using leaf rolling scores (LRS) under drought conditions at the panicle development stage as the primary trait. This analysis consistently identified a major locus on chromosome 8, designated qFDR8 (quantitative field drought resistance 8), across multiple consecutive time points during the progressive drought stress in the field (Fig. 1a, b; Supplementary Fig. 1 and Supplementary Data 1). The lead single nucleotide polymorphism (SNP) (Chr8: 27179494; A>G, P = 1.96 × 10−10) was located in the promoter region of NLR8, a coiled-coil nucleotide-binding leucine-rich repeat (CC-NBS-LRR) gene (Fig. 1c). Two additional variations in linkage disequilibrium with the lead SNP (r² > 0.8) were identified: a 25-bp deletion in the promoter (InDel 9656, P = 1.25 × 10−7) and a single-base deletion in the coding sequence (InDel 7708, TG > T, P = 1.31 × 10−9), the latter causing a premature stop codon. Sequence analysis revealed two predominant haplotypes: NLR8-Hap1 (e.g., NLR8Nip, NLR8ZH11) encodes a full-length NLR protein, while NLR8-Hap2 (e.g., NLR8ZS97) carries the single-base deletion, resulting in a premature stop codon and a truncated protein (Fig. 1d). Accessions carrying NLR8-Hap2 exhibited enhanced drought resistance compared to those with NLR8-Hap1, with lower LRS (Fig. 1e). To assess whether the haplotype-associated phenotypic variation might result from altered NLR8 expression, we first compared NLR8 transcript abundance between accessions carrying NLR8-Hap1 and NLR8-Hap2 under drought conditions. No significant difference was detected between the two haplotype groups (Fig. 1f). We then tested the activity of the NLR8-Hap1 promoter and two mutant derivatives carrying either the 25-bp deletion (InDel 9656) or the A-to-G substitution at the lead SNP. Neither mutation significantly altered promoter activity in rice protoplasts (Fig. 1g). These results suggest that the difference in drought resistance is unlikely to be explained by NLR8 transcript abundance under drought or by the effects of the promoter variants on reporter activity.Fig. 1Natural variations in NLR8are associated with drought resistance and improve drought and blast resistance in rice.a, b Manhattan plots from a genome-wide association study (GWAS) of leaf rolling scores (LRS) under drought at D20 and D22 (5:30–7:30 PM), identifying the significant qFDR8 locus on chromosome 8. The dashed line indicates the Bonferroni-adjusted suggestive threshold (P = 5.04 × 10−7). c Regional association and linkage disequilibrium (LD) plots for the NLR8 locus. The lead SNP (vg0827179494, P = 1.96 × 10−10, diamond) and causal single-base deletion (InDel 7708, P = 1.31 × 10−9, triangle) show strong LD (r2 > 0.8). d Schematic showing NLR8 haplotypes. NLR8-Hap1 encodes a full-length protein, while NLR8-Hap2 contains a premature stop codon. e LRS of accessions carrying NLR8-Hap1 (n = 154) and NLR8-Hap2 (n = 64). f Relative expression of NLR8 in accessions carrying either haplotype (n = 10 accessions per haplotype). g Activities of the NLR8-Hap1 promoter and its mutant derivatives carrying either the 25-bp deletion (InDel 9656) or the A-to-G substitution at the lead SNP, assessed using a dual-luciferase assay in rice protoplasts (n = 4 biological replicates). The empty reporter vector served as the control (CK). h Wild-type (WT), nlr8, NLR8-Hap1-COM, and NLR8-Hap2-COM plants under drought at the panicle stage. Scale bar, 20 cm. i–k LRS (i), seed-setting rate (j), and grain yield per plant (k) under drought (n = 6 plants). l Blast symptoms 7 days after inoculation with Magnaporthe oryzae isolate 99-20-2. Scale bar, 1 cm. Lesion length (m) and area (n), showing NLR8-Hap2’s enhanced blast resistance. Sample sizes: nlr8-#3, n = 15; WT, n = 13; NLR8-Hap1-COM-#1, n = 15; NLR8-Hap1-COM-#7, n = 14; NLR8-Hap2-COM-#7, n = 25; NLR8-Hap2-COM-#12, n = 19. For (m) and (n), n denotes the number of lesions measured on punch-inoculated leaves. Box plots show medians, interquartile ranges, and whiskers extending to the most extreme values within 1.5 times the interquartile range; dots show individual values. Statistical significance in e was determined using a two-sided Wilcoxon test; all other pairwise comparisons were performed using a two-sided Welch’s t-test. Source data are provided as a Source Data file. Natural variations in NLR8are associated with drought resistance and improve drought and blast resistance in rice. a, b Manhattan plots from a genome-wide association study (GWAS) of leaf rolling scores (LRS) under drought at D20 and D22 (5:30–7:30 PM), identifying the significant qFDR8 locus on chromosome 8. The dashed line indicates the Bonferroni-adjusted suggestive threshold (P = 5.04 × 10−7). c Regional association and linkage disequilibrium (LD) plots for the NLR8 locus. The lead SNP (vg0827179494, P = 1.96 × 10−10, diamond) and causal single-base deletion (InDel 7708, P = 1.31 × 10−9, triangle) show strong LD (r2 > 0.8). d Schematic showing NLR8 haplotypes. NLR8-Hap1 encodes a full-length protein, while NLR8-Hap2 contains a premature stop codon. e LRS of accessions carrying NLR8-Hap1 (n = 154) and NLR8-Hap2 (n = 64). f Relative expression of NLR8 in accessions carrying either haplotype (n = 10 accessions per haplotype). g Activities of the NLR8-Hap1 promoter and its mutant derivatives carrying either the 25-bp deletion (InDel 9656) or the A-to-G substitution at the lead SNP, assessed using a dual-luciferase assay in rice protoplasts (n = 4 biological replicates). The empty reporter vector served as the control (CK). h Wild-type (WT), nlr8, NLR8-Hap1-COM, and NLR8-Hap2-COM plants under drought at the panicle stage. Scale bar, 20 cm. i–k LRS (i), seed-setting rate (j), and grain yield per plant (k) under drought (n = 6 plants). l Blast symptoms 7 days after inoculation with Magnaporthe oryzae isolate 99-20-2. Scale bar, 1 cm. Lesion length (m) and area (n), showing NLR8-Hap2’s enhanced blast resistance. Sample sizes: nlr8-#3, n = 15; WT, n = 13; NLR8-Hap1-COM-#1, n = 15; NLR8-Hap1-COM-#7, n = 14; NLR8-Hap2-COM-#7, n = 25; NLR8-Hap2-COM-#12, n = 19. For (m) and (n), n denotes the number of lesions measured on punch-inoculated leaves. Box plots show medians, interquartile ranges, and whiskers extending to the most extreme values within 1.5 times the interquartile range; dots show individual values. Statistical significance in e was determined using a two-sided Wilcoxon test; all other pairwise comparisons were performed using a two-sided Welch’s t-test. Source data are provided as a Source Data file. NLR8-Hap2 confers dual resistance to drought and blast To genetically validate the function of NLR8 and directly compare the effects of the two haplotypes, we generated knockout mutants (nlr8-#1, nlr8-#2, and nlr8-#3) in the japonica/geng rice variety Zhonghua 11 (ZH11), which carries the NLR8-Hap1 allele and was used as the wild-type background (Supplementary Fig. 2a). We also generated complementation lines expressing either NLR8-Hap1 (NLR8-Hap1-COM) or NLR8-Hap2 (NLR8-Hap2-COM) in the nlr8-#3 background (Supplementary Fig. 2b). Under drought stress at the panicle development stage, nlr8 mutants showed increased drought sensitivity compared to wild-type (WT) plants, confirming that NLR8 is a positive regulator of drought resistance (Fig. 1h; Supplementary Fig. 3). While both complementation lines rescued the drought-sensitive phenotype, the NLR8-Hap2-COM lines exhibited significantly stronger drought resistance than the NLR8-Hap1-COM lines, as evidenced by lower LRS, higher seed-setting rates, and increased grain yields (Fig. 1h–k). Given that NLR proteins are canonical immune receptors in plants, we hypothesized that NLR8 might also function in disease resistance. To test this, we inoculated the same set of genetic materials with the blast fungus (M. oryzae). The nlr8 mutants displayed increased susceptibility to blast disease compared to WT plants, with larger infection lesions (Fig. 1l; Supplementary Fig. 4a–c). Consistent with the drought phenotype, both complementation lines restored blast resistance, whereas the NLR8-Hap2-COM lines showed significantly enhanced blast resistance compared to both WT and NLR8-Hap1-COM lines, with markedly reduced lesion development (Fig. 1l–n). Because enhanced NLR activity can sometimes be associated with constitutive defense activation or growth defects, we further examined whether NLR8-Hap2 caused obvious autoimmunity-related phenotypes. Under normal growth conditions, OsPR1, OsPR1a, and OsPR1b were not constitutively upregulated in NLR8-Hap2-COM plants compared with WT (Supplementary Fig. 4d), and no obvious difference in growth performance was observed (Fig. 1h), indicating an absence of constitutive autoimmunity or growth defects under the tested conditions. These results provide genetic evidence that NLR8 positively contributes to both drought and blast resistance, and that the naturally truncated NLR8-Hap2 allele confers enhanced resistance to both stresses. NLR8 interacts with two stress-response TFs To investigate the molecular mechanism underlying NLR8-mediated dual resistance, we explored its interactions with potential regulatory partners. NLR8 exists in two variants: the full-length Hap1, comprising the canonical CC, NB-ARC, and LRR domains, and the truncated Hap2, which lacks the LRR domain due to a premature stop codon (Fig. 2a). Both NLR8 variants localized to the nucleus and cytoplasm (Fig. 2b; Supplementary Fig. 5a). They were then used as baits to screen a yeast two-hybrid (Y2H) library. The screen identified multiple putative interacting proteins, including several transcription factors (Supplementary Data 2). Among these candidates, we prioritized OsbHLH148 and OsBIHD1 for further analysis, because OsbHLH148 has been reported to function in drought tolerance and OsBIHD1 has been implicated in blast defense-related signaling31,32. However, whether they function as direct partners of a single NLR to coordinate dual-stress resistance remained unknown. We therefore examined their full-length interactions with NLR8 using a Y2H assay (Fig. 2c). The interactions of the two transcription factors with NLR8 were further validated by a series of assays, including bimolecular fluorescence complementation (BiFC) assays in Nicotiana benthamiana leaves (Fig. 2d), split luciferase complementation (SLC) assays (Fig. 2e), and co-immunoprecipitation (Co-IP) experiments in rice protoplasts (Fig. 2f). Furthermore, we observed that OsbHLH148 exhibits nuclear-cytoplasmic localization, whereas OsBIHD1 is located exclusively in the nucleus (Supplementary Fig. 5b, c). The identification of these two functionally distinct TFs as NLR8 interactors suggested that NLR8 may be associated with separate signaling pathways. We further checked the tissue expression profiles of NLR8 and the two TF genes, which displayed partially overlapping expression patterns (Supplementary Fig. 5d, data from the Collection of Rice Expression Profiles (CREP)33). We also examined their transcriptional responses to drought stress and rice blast inoculation. OsbHLH148 and NLR8 were concurrently upregulated under drought conditions, whereas OsBIHD1 and NLR8 were coordinately induced upon blast infection (Supplementary Fig. 5e). Collectively, these expression patterns further support the functional relevance of NLR8 and the two TF genes in rice plants.Fig. 2NLR8 interacts with transcription factors (TFs) OsbHLH148 and OsBIHD1.a Schematic of the two NLR8 protein variants, showing the coiled-coil (CC), nucleotide-binding (NB-ARC), and leucine-rich repeat (LRR) domains. The NLR8-Hap2 protein is truncated and lacks the LRR domain. b Subcellular localization of YFP-fused NLR8 variants in rice protoplasts, showing distribution in both the nucleus (marked by Ghd7-CFP) and cytoplasm. Scale bar, 10 μm. c Yeast two-hybrid (Y2H) assay and β-galactosidase assay demonstrating that both NLR8 haplotypes interact with OsbHLH148 and OsBIHD1. SD/-2: medium lacking Trp and Leu; SD/-4: medium lacking Trp, Leu, His, and adenine. d Bimolecular fluorescence complementation (BiFC) assay in Nicotiana benthamiana leaves, validating the interactions between NLR8 haplotypes and OsbHLH148 or OsBIHD1. Scale bar, 20 μm. e Split firefly luciferase complementation (SLC) assay in Nicotiana benthamiana leaves further confirms the interactions. GST-cLUC served as a negative control. f Co-immunoprecipitation (Co-IP) assay in rice protoplasts validating the interactions. NLR8 variants were FLAG-tagged, and OsbHLH148 and OsBIHD1 were GFP-tagged. Co-expression of GFP with the FLAG-tagged NLR8 served as a negative control. Source data are provided as a Source Data file. NLR8 interacts with transcription factors (TFs) OsbHLH148 and OsBIHD1. a Schematic of the two NLR8 protein variants, showing the coiled-coil (CC), nucleotide-binding (NB-ARC), and leucine-rich repeat (LRR) domains. The NLR8-Hap2 protein is truncated and lacks the LRR domain. b Subcellular localization of YFP-fused NLR8 variants in rice protoplasts, showing distribution in both the nucleus (marked by Ghd7-CFP) and cytoplasm. Scale bar, 10 μm. c Yeast two-hybrid (Y2H) assay and β-galactosidase assay demonstrating that both NLR8 haplotypes interact with OsbHLH148 and OsBIHD1. SD/-2: medium lacking Trp and Leu; SD/-4: medium lacking Trp, Leu, His, and adenine. d Bimolecular fluorescence complementation (BiFC) assay in Nicotiana benthamiana leaves, validating the interactions between NLR8 haplotypes and OsbHLH148 or OsBIHD1. Scale bar, 20 μm. e Split firefly luciferase complementation (SLC) assay in Nicotiana benthamiana leaves further confirms the interactions. GST-cLUC served as a negative control. f Co-immunoprecipitation (Co-IP) assay in rice protoplasts validating the interactions. NLR8 variants were FLAG-tagged, and OsbHLH148 and OsBIHD1 were GFP-tagged. Co-expression of GFP with the FLAG-tagged NLR8 served as a negative control. Source data are provided as a Source Data file. Two TFs mediate distinct NLR8 stress responses To dissect the functional specificity of the NLR8-TF interactions, we generated CRISPR/Cas9-mediated knockout and overexpression lines for OsbHLH148 and OsBIHD1 (Supplementary Fig. 6). Phenotypic analyses revealed distinct stress-specific functions (Fig. 3; Supplementary Fig. 7). Under drought conditions, OsbHLH148 knockout (osbhlh148) mutants were more sensitive to drought, exhibiting severe leaf rolling and yield loss (Fig. 3a–d), whereas its overexpression (OsbHLH148-OE) conferred enhanced drought tolerance across all measured parameters (Fig. 3e–h). We next examined representative drought-responsive marker genes associated with the OsbHLH148-mediated pathways31. Under drought conditions, OsDREB1A and OsDREB1G showed higher expression in OsbHLH148-OE plants and lower expression in the osbhlh148 mutant compared with WT (Supplementary Fig. 7a), supporting an association between OsbHLH148 and drought-responsive gene expression. OsBIHD1 knockout (osbihd1) mutants were more susceptible to blast disease, indicated by larger lesions post-inoculation with M. oryzae, whereas its overexpression (OsBIHD1-OE) enhanced blast resistance with reduced lesion size (Fig. 3i–k). After M. oryzaeinfection, the representative defense-responsive marker genes OsPR1, OsPR1a, and OsPR1b were upregulated in OsBIHD1-OE plants and reduced in the osbihd1 mutant compared with WT (Supplementary Fig. 7b), supporting an association between OsBIHD1 and defense-responsive gene expression. osbhlh148 mutants and OsbHLH148-overexpression lines showed WT-like blast susceptibility, and osbihd1 mutants and OsBIHD1-overexpression lines exhibited no significant difference in drought resistance compared to WT (Supplementary Fig. 7c–g). These results support a functional specificity in which OsbHLH148 mediates drought tolerance, and OsBIHD1 mediates blast resistance. This functional orthogonality suggests that NLR8 is associated with these genetically separable stress-response branches.Fig. 3OsbHLH148 and OsBIHD1 are specifically associated with NLR8-mediated drought and blast responses.a Non-transgenic segregant control corresponding to the CRISPR/Cas9 lines (CRI-NT) and osbhlh148 mutants before and after drought treatment. Scale bar, 20 cm. LRS (b), seed-setting rate (c), and grain yield per plant (d) for CRI-NT and osbhlh148 mutants under drought (n = 7 plants). e Non-transgenic segregant control corresponding to the OsbHLH148-overexpression lines (OE-NT) and OsbHLH148-OE lines before and after drought treatment. Scale bar, 20 cm. LRS (f), seed-setting rate (g), and grain yield per plant (h) for OE-NT and OsbHLH148-OE lines under drought (n = 6 plants). The knockout and overexpression assays were performed independently under drought intensities optimized for each genotype group. Relative soil water content (RSWC) after drought treatment was approximately 14% for the knockout assay and approximately 10% for the overexpression assay; statistical comparisons were therefore restricted to genotypes within each assay. i Blast symptoms of osbihd1, OsBIHD1-OE, and WT plants after inoculation with Magnaporthe oryzae isolate 99-20-2. Scale bar, 1 cm. Lesion length (j) and area (k) after blast inoculation. Sample sizes: osbihd1-#2, n = 22; osbihd1-#3, n = 20; WT, n = 24; OsBIHD1-OE2, n = 20; OsBIHD1-OE4, n = 20. l OsbHLH148 epistasis analysis. Phenotypes of OsbHLH148-OE, WT, nlr8-#3, and nlr8-#3/OsbHLH148-OE plants before and after drought. Scale bar, 20 cm. m Green-projected area ratio (GPAR) and LRS for the lines in (l) (n = 6 plants). n OsBIHD1 epistasis analysis. Blast symptoms of OsBIHD1-OE, WT, nlr8-#3, and nlr8-#3/OsBIHD1-OE plants. Scale bar, 1 cm. Lesion length (o) and area (p) for the lines in (n). Sample sizes: nlr8-#3, n = 65; WT, n = 63; OsBIHD1-OE2, n = 58; nlr8-#3/OsBIHD1-OE2, n = 36. For (j), (k), (o), and (p), n denotes lesions measured on punch-inoculated leaves. Box plots show medians, interquartile ranges, and whiskers extending to the most extreme values within 1.5× the interquartile range; dots show individual values. Statistical significance was determined using two-sided Welch’s t-tests. Source data are provided as a Source Data file. OsbHLH148 and OsBIHD1 are specifically associated with NLR8-mediated drought and blast responses. a Non-transgenic segregant control corresponding to the CRISPR/Cas9 lines (CRI-NT) and osbhlh148 mutants before and after drought treatment. Scale bar, 20 cm. LRS (b), seed-setting rate (c), and grain yield per plant (d) for CRI-NT and osbhlh148 mutants under drought (n = 7 plants). e Non-transgenic segregant control corresponding to the OsbHLH148-overexpression lines (OE-NT) and OsbHLH148-OE lines before and after drought treatment. Scale bar, 20 cm. LRS (f), seed-setting rate (g), and grain yield per plant (h) for OE-NT and OsbHLH148-OE lines under drought (n = 6 plants). The knockout and overexpression assays were performed independently under drought intensities optimized for each genotype group. Relative soil water content (RSWC) after drought treatment was approximately 14% for the knockout assay and approximately 10% for the overexpression assay; statistical comparisons were therefore restricted to genotypes within each assay. i Blast symptoms of osbihd1, OsBIHD1-OE, and WT plants after inoculation with Magnaporthe oryzae isolate 99-20-2. Scale bar, 1 cm. Lesion length (j) and area (k) after blast inoculation. Sample sizes: osbihd1-#2, n = 22; osbihd1-#3, n = 20; WT, n = 24; OsBIHD1-OE2, n = 20; OsBIHD1-OE4, n = 20. l OsbHLH148 epistasis analysis. Phenotypes of OsbHLH148-OE, WT, nlr8-#3, and nlr8-#3/OsbHLH148-OE plants before and after drought. Scale bar, 20 cm. m Green-projected area ratio (GPAR) and LRS for the lines in (l) (n = 6 plants). n OsBIHD1 epistasis analysis. Blast symptoms of OsBIHD1-OE, WT, nlr8-#3, and nlr8-#3/OsBIHD1-OE plants. Scale bar, 1 cm. Lesion length (o) and area (p) for the lines in (n). Sample sizes: nlr8-#3, n = 65; WT, n = 63; OsBIHD1-OE2, n = 58; nlr8-#3/OsBIHD1-OE2, n = 36. For (j), (k), (o), and (p), n denotes lesions measured on punch-inoculated leaves. Box plots show medians, interquartile ranges, and whiskers extending to the most extreme values within 1.5× the interquartile range; dots show individual values. Statistical significance was determined using two-sided Welch’s t-tests. Source data are provided as a Source Data file. To examine the genetic interplay between NLR8 and these TFs in regulating rice drought and blast resistance, we generated OsbHLH148- and OsBIHD1-overexpression lines in the nlr8-#3 mutant background, respectively (nlr8/OsbHLH148-OE and nlr8/OsBIHD1-OE, respectively). Under drought stress, both OsbHLH148-OE and nlr8/OsbHLH148-OE plants showed significantly better drought resistance (indicated by higher green-projected area ratio (GPAR) and lower LRS) than WT plants, while the nlr8-#3 mutant showed the most sensitive phenotype (Fig. 3l, m). Similarly, both OsBIHD1-OE and nlr8/OsBIHD1-OE plants showed significantly enhanced blast resistance compared to WT, while the nlr8-#3 mutant exhibited the most susceptible phenotype (Fig. 3n–p). Furthermore, we generated knockout mutants of OsbHLH148 and OsBIHD1 in the KY131-NIL(Hap2) background via CRISPR/Cas9, designated KY131-NIL(Hap2)/osbhlh148 and KY131-NIL(Hap2)/osbihd1 (Supplementary Fig. 8a). Under drought conditions, KY131-NIL(Hap2)/osbhlh148 lines exhibited significantly reduced drought tolerance, as indicated by increased LRS and decreased biomass, compared with KY131-NIL(Hap2) plants (Supplementary Fig. 8b–d). However, these lines retained blast resistance comparable to that of KY131-NIL(Hap2), with no significant differences in lesion length or lesion area (Supplementary Fig. 8e–g). In contrast, KY131-NIL(Hap2)/osbihd1 lines showed compromised blast resistance, whereas their drought tolerance remained comparable to that of KY131-NIL(Hap2) (Supplementary Fig. 8b–g). Together, these genetic analyses support a model in which OsbHLH148 and OsBIHD1 act as genetically separable components associated with NLR8-Hap2-mediated drought and blast resistance, respectively. NLR8-Hap2 is associated with enhanced TF stability We hypothesized that the superior function of NLR8-Hap2 might stem from an enhanced ability to regulate its downstream partners. To determine whether NLR8-Hap2 affects OsbHLH148 and OsBIHD1 at the transcriptional level, we first examined their transcript abundance in WT, nlr8, NLR8-Hap1-COM, and NLR8-Hap2-COM plants. No significant differences in OsbHLH148 or OsBIHD1 transcript levels were detected among these genotypes (Supplementary Fig. 9a). We next examined the stability of OsbHLH148 and OsBIHD1 in the presence and absence of NLR8. Nuclear-cytoplasmic fractionation revealed reduced OsbHLH148 protein levels in both total and nuclear fractions of nlr8 mutant protoplasts compared to WT (Fig. 4a). In vitro degradation assays demonstrated that OsbHLH148 underwent 26S proteasome-dependent degradation, as evidenced by MG132-mediated stabilization in WT extracts (Fig. 4b). Degradation was accelerated in nlr8 extracts (Fig. 4c), with MG132 treatment rescuing this effect, suggesting that NLR8 is associated with increased OsbHLH148 accumulation/stability under these assay conditions. Interestingly, NLR8-Hap2-COM lines showed greater OsbHLH148 accumulation than NLR8-Hap1-COM lines (Fig. 4d). Consistent with this observation, in vitro degradation assays showed delayed degradation of OsbHLH148 in NLR8-Hap2-COM extracts compared to NLR8-Hap1-COM extracts (Fig. 4e), suggesting that the Hap2 variant is associated with enhanced OsbHLH148 stability under these assay conditions. OsbHLH148 is a known activator of the drought-responsive genes OsDREB1A and OsDREB1G31. Accordingly, the enhanced stabilization of OsbHLH148 by NLR8-Hap2 resulted in significantly higher expression of these target genes in NLR8-Hap2-COM plants compared to NLR8-Hap1-COM plants (Fig. 4f).Fig. 4NLR8-Hap2 is associated with enhanced stability of OsbHLH148 and OsBIHD1 compared with NLR8-Hap1.a Nuclear-cytoplasmic fractionation showing reduced total and nuclear OsbHLH148 protein levels in nlr8 mutant protoplasts compared to WT. GFP-FLAG co-expression served as a control. Actin and histone H3 served as cytoplasmic and nuclear markers, respectively. b Cell-free degradation assay showing that OsbHLH148 is degraded via the 26S proteasome, and this is inhibited by MG132. c Cell-free assay showing accelerated OsbHLH148 degradation in nlr8 extracts compared to WT, with degradation suppressed by MG132 treatment. d Nuclear-cytoplasmic fractionation showing higher OsbHLH148 accumulation in NLR8-Hap2-COM lines than in NLR8-Hap1-COM lines. GFP-FLAG co-expression served as a control. Actin and histone H3 were used as cytoplasmic and nuclear markers, respectively. e Cell-free assay showing delayed OsbHLH148 degradation in NLR8-Hap2-COM extracts compared to NLR8-Hap1-COM extracts. f Relative expression of OsDREB1A and OsDREB1G in WT, nlr8 mutant, NLR8-Hap1, and NLR8-Hap2 complementation lines. g Western blot showing reduced OsBIHD1 protein levels in nlr8 protoplasts. h Cell-free assay showing OsBIHD1 degradation is inhibited by MG132. i Cell-free assay showing accelerated OsBIHD1 degradation in nlr8 extracts. j Western blot showing higher OsBIHD1 accumulation in NLR8-Hap2-COM lines. k Cell-free assay showing delayed OsBIHD1 degradation in NLR8-Hap2-COM extracts compared to NLR8-Hap1-COM extracts. l Relative expression of OsBIHD1 target genes OsACO2 and OsACO3. The WB bands were quantified and analyzed by ImageJ v2.3.0. The numerical values are labeled below the bands. For (f) and (l), n = 3 biological replicates per genotype. Box plots show medians (center lines), interquartile ranges (boxes), and whiskers extending to the most extreme values within 1.5× the interquartile range; dots show individual values. Statistical significance was determined using two-sided Welch’s t-tests. Source data are provided as a Source Data file. NLR8-Hap2 is associated with enhanced stability of OsbHLH148 and OsBIHD1 compared with NLR8-Hap1. a Nuclear-cytoplasmic fractionation showing reduced total and nuclear OsbHLH148 protein levels in nlr8 mutant protoplasts compared to WT. GFP-FLAG co-expression served as a control. Actin and histone H3 served as cytoplasmic and nuclear markers, respectively. b Cell-free degradation assay showing that OsbHLH148 is degraded via the 26S proteasome, and this is inhibited by MG132. c Cell-free assay showing accelerated OsbHLH148 degradation in nlr8 extracts compared to WT, with degradation suppressed by MG132 treatment. d Nuclear-cytoplasmic fractionation showing higher OsbHLH148 accumulation in NLR8-Hap2-COM lines than in NLR8-Hap1-COM lines. GFP-FLAG co-expression served as a control. Actin and histone H3 were used as cytoplasmic and nuclear markers, respectively. e Cell-free assay showing delayed OsbHLH148 degradation in NLR8-Hap2-COM extracts compared to NLR8-Hap1-COM extracts. f Relative expression of OsDREB1A and OsDREB1G in WT, nlr8 mutant, NLR8-Hap1, and NLR8-Hap2 complementation lines. g Western blot showing reduced OsBIHD1 protein levels in nlr8 protoplasts. h Cell-free assay showing OsBIHD1 degradation is inhibited by MG132. i Cell-free assay showing accelerated OsBIHD1 degradation in nlr8 extracts. j Western blot showing higher OsBIHD1 accumulation in NLR8-Hap2-COM lines. k Cell-free assay showing delayed OsBIHD1 degradation in NLR8-Hap2-COM extracts compared to NLR8-Hap1-COM extracts. l Relative expression of OsBIHD1 target genes OsACO2 and OsACO3. The WB bands were quantified and analyzed by ImageJ v2.3.0. The numerical values are labeled below the bands. For (f) and (l), n = 3 biological replicates per genotype. Box plots show medians (center lines), interquartile ranges (boxes), and whiskers extending to the most extreme values within 1.5× the interquartile range; dots show individual values. Statistical significance was determined using two-sided Welch’s t-tests. Source data are provided as a Source Data file. Parallel experiments with OsBIHD1 revealed similar regulatory mechanisms. OsBIHD1 protein levels were reduced in nlr8 protoplasts (Fig. 4g) and showed MG132-sensitive degradation (Fig. 4h). Degradation was accelerated in nlr8 plant extracts (Fig. 4i), similar to the pattern observed for OsbHLH148. NLR8-Hap2-COM lines exhibited higher OsBIHD1 levels than those in NLR8-Hap1-COM (Fig. 4j), corresponding to delayed degradation in in vitro assays (Fig. 4k). OsBIHD1 is known to regulate OsACO2 and OsACO3, genes involved in ethylene biosynthesis and blast defense34. Consistent with increased OsBIHD1 accumulation, NLR8-Hap2-COM lines showed higher expression of these target genes compared to NLR8-Hap1-COM (Fig. 4l). Taken together, these results suggest that NLR8-Hap2 is associated with increased accumulation or stability of the tested TF partners under these assay conditions, thereby amplifying their downstream signaling. The LRR domain attenuates CC-mediated interactions We examined the protein abundance of NLR8 in the NLR8-Hap1-COM and NLR8-Hap2-COM lines. No significant difference was observed between the two genotypes (Supplementary Fig. 9b), suggesting that the functional differences between haplotypes may arise from their structural differences. The increased accumulation/stability associated with the LRR-truncated Hap2 variant suggested that the LRR domain might attenuate detectable NLR8-mediated TF association. To test this, we first determined the interaction domains. Y2H assays confirmed that both OsbHLH148 and OsBIHD1 interact specifically with the N-terminal CC domain of NLR8, but not with the NB-ARC or LRR domains (Fig. 5a). Specifically, the CC domain of NLR8 interacts with the bHLH domain and C-terminal region (amino acids 114-299) of OsbHLH148 (Supplementary Fig. 9c), and the N-terminal, KN, and POX domains (amino acids 1-438) of OsBIHD1 (Supplementary Fig. 9d). This indicates that the CC domain may be the primary interface for TF interaction. We then investigated potential intramolecular interactions within NLR8. SLC assays revealed that the CC domain interacts with both the NB-ARC domain and with the LRR2 subdomain, but not with the LRR1 subdomain (Fig. 5b; Supplementary Fig. 9e). The NB-ARC domain interacts with the CC domain as well as both LRR1 and LRR2 subdomains.Fig. 5The LRR domain attenuates CC-domain-mediated interactions with downstream TFs.a Y2H assay showing that OsbHLH148 and OsBIHD1 interact specifically with the CC domain of NLR8, but not the NB-ARC or LRR domains. b SLC assay mapping intramolecular interactions within NLR8. The CC domain interacts with the NB-ARC and LRR2 subdomain, but not with the LRR1 subdomain. The NB-ARC domain interacts with the CC domain, as well as the LRR1 and LRR2 subdomains. c, d Competition SLC assays in Nicotiana benthamiana. Co-expression of the LRR2 subdomain reduces the detectable interaction signals between NLR8-Hap1/2 and OsbHLH148 (c) or OsBIHD1 (d). Co-expression with GFP serves as a control. e, f Competition Co-IP assays. Co-expression of LRR2-HA reduces co-immunoprecipitation of OsbHLH148-GFP (e) or OsBIHD1-GFP (f) with NLR8-Hap2-FLAG. g, h Co-expression assays in rice protoplasts. Co-expression of LRR2 reduces NLR8-associated stabilization of OsbHLH148 (g) and OsBIHD1 (h). Source data are provided as a Source Data file. The LRR domain attenuates CC-domain-mediated interactions with downstream TFs. a Y2H assay showing that OsbHLH148 and OsBIHD1 interact specifically with the CC domain of NLR8, but not the NB-ARC or LRR domains. b SLC assay mapping intramolecular interactions within NLR8. The CC domain interacts with the NB-ARC and LRR2 subdomain, but not with the LRR1 subdomain. The NB-ARC domain interacts with the CC domain, as well as the LRR1 and LRR2 subdomains. c, d Competition SLC assays in Nicotiana benthamiana. Co-expression of the LRR2 subdomain reduces the detectable interaction signals between NLR8-Hap1/2 and OsbHLH148 (c) or OsBIHD1 (d). Co-expression with GFP serves as a control. e, f Competition Co-IP assays. Co-expression of LRR2-HA reduces co-immunoprecipitation of OsbHLH148-GFP (e) or OsBIHD1-GFP (f) with NLR8-Hap2-FLAG. g, h Co-expression assays in rice protoplasts. Co-expression of LRR2 reduces NLR8-associated stabilization of OsbHLH148 (g) and OsBIHD1 (h). Source data are provided as a Source Data file. These findings raised the possibility that the LRR region may interfere with CC-mediated association with the tested TFs. We tested this possibility using competition assays. Co-expression with the LRR2 subdomain in N. benthamiana reduced the detectable interaction signal between NLR8-Hap1/Hap2 and OsbHLH148 or OsBIHD1 (Fig. 5c, d; Supplementary Fig. 9f, g). Furthermore, competitive Co-IP assays showed that LRR2 co-expression reduced the binding of NLR8-Hap2 to OsbHLH148 and OsBIHD1 (Fig. 5e, f). Co-expression with LRR2 in rice protoplasts reduced the ability of NLR8-Hap2 to stabilize both OsbHLH148 and OsBIHD1 (Fig. 5g, h). These results are consistent with an LRR-associated inhibitory effect on CC-mediated TF association, providing a possible explanation for why the LRR-truncated NLR8-Hap2 variant showed enhanced association with and stabilization of the tested TF partners. To examine whether a similar effect could be observed in another rice NLR context, we examined whether the LRR region of BPH1412 could affect the CC-NB-mediated interaction with OsWRKY72. Co-expression of the BPH14-LRR region reduced the BPH14 CC-NB—OsWRKY72 interaction and attenuated BPH14 CC-NB-mediated stabilization of OsWRKY72 (Supplementary Fig. 10). These results suggest that a similar LRR-associated attenuation effect exists in the BPH14 context, yet broader analyses are needed to assess whether this effect occurs more widely in the NLR family. Breeding potential of NLR8-Hap2 To assess the breeding potential of the NLR8-Hap2 variant for developing rice varieties with dual resistance to drought and blast, we analyzed its haplotypic distribution using the RiceVarMap v2 database35. This analysis showed that the frequency of NLR8-Hap2 is markedly higher in the indica subgroup (31.17%; 860/2,759 varieties) than in the japonica (1.65%; 25/1,512 varieties) or aus (0.74%; 2/269 varieties) (Fig. 6a). This highlights its potential utility, particularly for japonica breeding. To validate this, we developed near-isogenic lines (NILs) by introgressing NLR8-Hap2 into two NLR8-Hap1 backgrounds: the indica restorer line Minghui 63 (MH63) and the northern japonica variety Kongyu131 (KY131) (Supplementary Fig. 11). In pot-based drought testing, the near-isogenic line MH63-NIL(Hap2) displayed improved drought tolerance compared with the corresponding control (Supplementary Fig. 12a–d). To substantiate these results under field conditions, we conducted small-plot (0.5 m2) field trials in Wuhan (2024). Consistent with the pot experiments, the near-isogenic line MH63-NIL(Hap2) exhibited stronger drought tolerance at the panicle development stage compared with the control, as reflected by lower leaf rolling, higher seed-setting rate, and yield per plant (Fig. 6b–e). To further validate the breeding value of NLR8-Hap2, we conducted medium-sized plot field trials in Wuhan (2025), with each plot covering 2.25 m². Under drought conditions, MH63-NIL(Hap2) lines exhibited a significantly higher grain yield per plot compared with the control (Fig. 6f, g). Concurrently, MH63-NIL(Hap2) showed enhanced resistance to rice blast under natural field disease pressure (Fig. 6h, i). A similar trend was observed in the northern japonica variety KY131 under the tested conditions. The KY131-NIL(Hap2) line also showed improved drought tolerance in the field and pot (Fig. 6j–m; Supplementary Fig. 12e–h), as well as enhanced blast resistance under the tested inoculation conditions (Fig. 6n, o; Supplementary Fig. 12i–k). Under the tested normal field conditions, introgression of NLR8-Hap2 did not cause detectable yield penalties, as both NILs maintained normal agronomic performance and yield-related traits comparable to their parental lines under normal irrigation conditions (Supplementary Fig. 13a–c).Fig. 6Breeding application of NLR8-Hap2 for dual-stress-resilient rice.a Frequency distribution of the causal variant (InDel 7708) in NLR8 across 4726 rice varieties, showing higher prevalence of NLR8-Hap2 in indica. b Representative phenotypes of the near-isogenic line MH63-NIL(Hap2) and its recurrent indica parent MH63 (Hap1) during and after field drought (Wuhan, 2024). LRS (c), seed-setting rate (d), and grain yield per plant (e) in 0.5-m2 plots under drought (MH63(Hap1), n = 24 plants; MH63-NIL(Hap2), n = 25 plants). Representative phenotypes before drought, after drought, and after 45 d of recovery (f), and LRS and grain yield per plot (g), in a 2.25-m2 field-plot trial (Wuhan, 2025; n = 4 independent field plots per genotype). h Representative blast symptoms of MH63(Hap1) and MH63-NIL(Hap2) under natural field disease pressure. i Lesion length and lesion area measured on infected leaves collected from field-grown plants (MH63(Hap1), n = 38; MH63-NIL(Hap2), n = 39). Here, n denotes the number of lesions measured on infected leaves collected from field-grown plants. Scale bar, 1 cm. j Representative phenotypes of KY131-NIL(Hap2) and its recurrent japonica parent KY131 (NLR8-Hap1) before and after field drought. LRS (k), seed-setting rate (l), and grain yield per plant (m) (n = 22 plants). Representative blast symptoms after inoculation with Magnaporthe oryzae isolate 99-20-2 (n) and lesion length and area (o) (n = 24 lesions per genotype, measured on punch-inoculated leaves). Scale bar, 1 cm. p A proposed model for NLR8-Hap2-mediated dual-stress resistance. In full-length NLR8-Hap1, the LRR domain attenuates CC domain interactions with the tested TFs. In naturally truncated NLR8-Hap2, the absence of the LRR domain is associated with reduced attenuation, enhanced binding and stabilization of OsbHLH148 and OsBIHD1, with enhanced outputs of the corresponding drought- and blast-response branches, respectively. These effects confer dual resilience without detectable penalties under the tested normal field conditions. Box plots show medians, interquartile ranges, and whiskers extending to the most extreme values within 1.5× the interquartile range; dots show individual values. Statistical significance was determined using two-sided Welch’s t-tests. Source data are provided as a Source Data file. Breeding application of NLR8-Hap2 for dual-stress-resilient rice. a Frequency distribution of the causal variant (InDel 7708) in NLR8 across 4726 rice varieties, showing higher prevalence of NLR8-Hap2 in indica. b Representative phenotypes of the near-isogenic line MH63-NIL(Hap2) and its recurrent indica parent MH63 (Hap1) during and after field drought (Wuhan, 2024). LRS (c), seed-setting rate (d), and grain yield per plant (e) in 0.5-m2 plots under drought (MH63(Hap1), n = 24 plants; MH63-NIL(Hap2), n = 25 plants). Representative phenotypes before drought, after drought, and after 45 d of recovery (f), and LRS and grain yield per plot (g), in a 2.25-m2 field-plot trial (Wuhan, 2025; n = 4 independent field plots per genotype). h Representative blast symptoms of MH63(Hap1) and MH63-NIL(Hap2) under natural field disease pressure. i Lesion length and lesion area measured on infected leaves collected from field-grown plants (MH63(Hap1), n = 38; MH63-NIL(Hap2), n = 39). Here, n denotes the number of lesions measured on infected leaves collected from field-grown plants. Scale bar, 1 cm. j Representative phenotypes of KY131-NIL(Hap2) and its recurrent japonica parent KY131 (NLR8-Hap1) before and after field drought. LRS (k), seed-setting rate (l), and grain yield per plant (m) (n = 22 plants). Representative blast symptoms after inoculation with Magnaporthe oryzae isolate 99-20-2 (n) and lesion length and area (o) (n = 24 lesions per genotype, measured on punch-inoculated leaves). Scale bar, 1 cm. p A proposed model for NLR8-Hap2-mediated dual-stress resistance. In full-length NLR8-Hap1, the LRR domain attenuates CC domain interactions with the tested TFs. In naturally truncated NLR8-Hap2, the absence of the LRR domain is associated with reduced attenuation, enhanced binding and stabilization of OsbHLH148 and OsBIHD1, with enhanced outputs of the corresponding drought- and blast-response branches, respectively. These effects confer dual resilience without detectable penalties under the tested normal field conditions. Box plots show medians, interquartile ranges, and whiskers extending to the most extreme values within 1.5× the interquartile range; dots show individual values. Statistical significance was determined using two-sided Welch’s t-tests. Source data are provided as a Source Data file. Discussion In this study, we identified NLR8-Hap2, a natural rice NLR8 allele encoding an LRR-truncated protein that contributes to enhanced drought and blast resistance and does not cause detectable yield penalties under the tested normal field conditions. Our data support a model in which the LRR truncation reduces attenuation of CC-domain association with and stabilization of the two tested TFs, OsbHLH148 and OsBIHD1, and is associated with enhanced outputs of the corresponding stress-response pathways (Fig. 6p). This dual-pathway regulation distinguishes NLR8-Hap2 from other NLRs, such as PibH8, which enhances drought and blast resistance via a single metabolic pathway24. While interactions between full-length NLRs and related TF pathways have been previously suggested32,36, our findings suggest that a natural LRR truncation in NLR8-Hap2 is associated with enhanced stabilization of two TF partners and improved resistance to the tested drought and blast stresses. This provides a case in which natural variation in an NLR gene contributes to combined abiotic and biotic stress resistance in rice. Previous studies of RBA1, TN2, the RPS4/RRS1 receptor complex, NRG1C/NRG1.3 and ADR1-L family members have shown that truncation, derepression or structural diversification of NLR-related proteins can lead to different immune regulatory outputs26–30. Structural studies of selected CNLs show that ligand-induced ADP release can prime activation and that subsequent activation involves interdomain reorganization and resistosome assembly; in the helper NLR SlNRC2, inactive oligomeric assemblies reinforce autoinhibition13,15,37. Together, these observations highlight the importance of the specific NLR architecture and interaction context in shaping the outcome of NLR truncation. In NLR8-Hap2, the interaction and competition results support a model in which the LRR2 region of NLR8 attenuates CC-mediated association with OsbHLH148 and OsBIHD1. The BPH14 assays provide an additional observation in another rice NLR context, but broader genetic, biochemical, and structural analyses will be required to determine whether similar inhibitory effects occur in other NLR proteins. A key aspect of the NLR-mediated dual-resistance mechanism is how NLR8-Hap2 contributes to two genetically separate stress-response branches. Our data are consistent with NLR8-Hap2 acting as an interaction platform that promotes the accumulation of the tested TF partners, OsbHLH148 and OsBIHD1. For drought tolerance, NLR8 stabilizes OsbHLH148, a key regulator of jasmonate-mediated dehydration responses31,38, thereby amplifying the expression of downstream genes, such as OsDREB1A. For blast resistance, it stabilizes OsBIHD1, a BELL-type homeodomain protein involved in ethylene biosynthesis and immunity32,34. These observations are consistent with reduced proteasome-associated turnover of the tested TFs under our assay conditions. By slowing their degradation, NLR8-Hap2 is associated with higher steady-state levels of these TFs, thereby allowing enhanced transcriptional responses. Crucially, our genetic analyses indicate that these two signaling pathways are functionally independent of each other. OsbHLH148 does not contribute to blast resistance (Supplementary Fig. 7c–e), nor does OsBIHD1 affect drought tolerance (Supplementary Fig. 7f, g). This apparent specificity provides a possible explanation for how NLR8-Hap2 improves the tested stress responses without detectable yield penalties under the tested normal field conditions. The pronounced difference in the frequency of the Hap2 variant between indica (high frequency) and japonica (low frequency) subspecies suggests that this haplotype has likely undergone selection, which may be driven by a complex interplay of environmental factors or linked selection. Given the high incidence of blast disease in the humid southern regions where indica rice is predominantly cultivated, resistance to M. oryzae may have contributed to the enrichment of the NLR8-Hap2 allele in certain genetic or environmental contexts. In this evolutionary context, the concurrent improvement in drought tolerance associated with NLR8-Hap2 may be interpreted as a valuable ancillary benefit or a form of pre-adaptation. However, linked selection, demographic history, or breeding history cannot be excluded based on the current population data. The lack of detectable yield penalties under the tested normal field conditions, as our data suggest, is also important for the potential breeding value of NLR8-Hap2. This suggests that the enhanced resilience against sporadic or unpredictable drought, a threat that can occur even in rain-fed areas, may have been favored in some environments without an obvious yield cost under the tested conditions, rather than being unequivocally advantageous for selection. From a breeding perspective, NLR8-Hap2 joins other rice drought-resistance genes and alleles with demonstrated agronomic relevance, including DRO1, DROT1, and ROAD1, which improve drought performance through deeper rooting, regulation of cell-wall properties, and potentiation of ABA signaling, respectively39–41. Unlike these drought-focused genes, NLR8-Hap2 also enhances blast resistance through a genetically separable TF-associated branch. It therefore represents a useful, yet largely untapped, genetic resource for crop improvement. The scarcity of this allele in the japonica gene pool indicates that many elite cultivars currently lack this dual-resistance-associated allele, which may help explain why this superior variation has not been fully exploited. Future evaluation using a broad panel of genetically diverse Magnaporthe oryzae isolates will be necessary to define the resistance spectrum and broader breeding applicability of NLR8-Hap2. Collectively, the introgression of the NLR8-Hap2 variant into elite japonica and other susceptible varieties via marker-assisted selection or precision gene editing presents a promising approach. This approach could simultaneously strengthen crops against a major fungal disease and a key abiotic stress, highlighting the potential value of NLR8-Hap2 for improving stress resilience in rice. In conclusion, our research identifies a natural rice NLR8 allele encoding an LRR-truncated protein that confers dual resistance to drought and blast. Consistent with reduced LRR-associated attenuation, the LRR truncation encoded by NLR8-Hap2 is associated with stronger CC-domain interaction with and stabilization of two distinct TFs, thereby enhancing separate drought and blast-response pathways. These findings provide a distinct mechanism by which a natural structural variation in an NLR gene can alter interactions with regulatory partners and contribute to both biotic and abiotic stress resistance in rice. This work provides both a potential genetic resource and a promising strategy for engineering climate-resilient crops, which are urgently needed for future global food security. Methods Field experiment and GWAS A natural population consisting of 240 Oryza sativa accessions was selected from a core panel of rice germplasm35, comprising 135 indica, 57 japonica, 28 aus, 14 admix, and 6 VI accessions. The drought phenotyping experiment was carried out in 2018 at the Drought Resistance Testing Station (30°28'15.84” N, 114°21'6.42” E) of Huazhong Agricultural University in Wuhan, China. The drought experiment was conducted under a movable rain-off shelter (55 m × 15 m) to exclude rainfall interference and impose progressive field drought stress. Staggered sowing was performed based on differences in flowering time within the population, ensuring all test materials were subjected to drought stress of equal intensity during the panicle development stage and minimizing confounding interference caused by differences in flowering time. The drought treatment continued for 30 days until most accessions exhibited significant leaf rolling, after which recovery irrigation was applied to restore water availability. The experiment was conducted using a randomized complete block design with two biological replicates, comprising 480 plots. Each plot (1 m × 1 m) consisted of 20 plants arranged in a 4 × 5 grid, with adjacent plots separated by purple rice border plants to minimize edge effects. Drought response was evaluated through manual leaf rolling score (LRS) assessments during the stress period. Soil moisture at a depth of 25 cm was monitored using frequency domain reflectometry (FDR), and the air temperature (24–38 °C) and relative humidity (50–100%) were recorded throughout the drought treatment. GWAS was performed using the high-density SNP dataset from RiceVarMap v2.0, which included 10,338,860 high-quality variants with a minor allele frequency (MAF) > 0.05 that were present in at least six accessions. Population structure was accounted for using a linear mixed model implemented in FaST-LMM v2.07, with the kinship matrix serving as a random effect. The kinship matrix was constructed by PLINK 2.042, using 222,391 evenly distributed variants (three variants per 5-kb window, MAF > 0.01) to control for population stratification. The effective number of independent markers was calculated as 1,981,633 using GEC v0.243. Significance thresholds were established using the effective number of independent tests method, with genome-wide suggestive (P = 5.04 × 10−7) and significant (P = 2.54 × 10−8) P-value cutoffs. Visualization of candidate gene-associated mapping based on GWAS results and pairwise LD analysis was performed using LDBlockShow v1.4144. Generation of CRISPR/Cas9-edited, complementation, overexpression, and near-isogenic lines To generate nlr8, osbhlh148, and osbihd1 knockout lines, gene-specific guide RNAs were cloned into the TKC CRISPR/Cas9 vector under the control of the OsU3/OsU6 promoters45,46. The resulting constructs were introduced into Zhonghua 11 (ZH11) via Agrobacterium-mediated transformation. In addition, constructs targeting OsbHLH148 or OsBIHD1 were introduced into the KY131-NIL(Hap2) background to generate KY131-NIL(Hap2)/osbhlh148 and KY131-NIL(Hap2)/osbihd1 lines, respectively. Homozygous mutations were identified by PCR amplification and sequencing of the target region. For phenotypic analyses, T2 homozygous mutant plants lacking detectable Cas9/T-DNA sequences were identified by genomic DNA-based qPCR using SPS as the internal reference and used as transgene-free mutant lines. For NLR8 complementation, genomic fragments containing NLR8-Hap1 or NLR8-Hap2, together with their native regulatory sequences, were cloned into pCAMBIA2301U and introduced into the nlr8-#3 background. For overexpression, the full-length coding sequences of OsbHLH148 and OsBIHD1 were cloned downstream of the maize Ubiquitin promoter in pCAMBIA1301U. The resulting constructs were introduced into ZH11 to generate OsbHLH148-OE and OsBIHD1-OE lines, which were then crossed into the nlr8-#3 background to generate nlr8-#3/OsbHLH148-OE and nlr8-#3/OsBIHD1-OE lines. Independent homozygous lines were used for phenotypic analyses. NILs carrying NLR8-Hap2 were developed in two genetic backgrounds: MH63 (an indica restorer line) × ZS97 (Hap2 donor), KY131 (a japonica variety) × Zhong413 (Hap2 donor). NILs were generated through six generations of backcrossing followed by two generations of self-pollination (BC6F3). Introgression of NLR8-Hap2 was verified by sequencing. Primer and guide-RNA sequences used in this study are provided in Supplementary Data 3. Newly generated plant materials and plasmids are available from the corresponding authors upon request. Drought-resistance evaluation To assess drought resistance at the panicle development stage, rice plants were grown in pots. Irrigation was discontinued at the onset of the panicle development stage to initiate drought stress. Soil water content was monitored by weighing the pots. When the assay-specific target relative soil water content (RSWC) was reached, plants were watered daily to maintain that level for approximately 7 days. The target RSWC for each independently performed assay is specified in the corresponding figure legend and was approximately 10–14% for the assays reported here. Once differences in leaf rolling became evident, LRS were recorded. Following phenotypic assessment, plants were rewatered, and yield-related traits were measured. For knockout and overexpression analyses, drought assays were performed independently when different genotype groups required different drought intensities for phenotypic discrimination. WT controls were grown and treated in parallel with the tested genotypes within each assay, and statistical comparisons were made only within the same assay. Absolute phenotypic values were not directly compared across independently performed drought experiments. For the 2024 Wuhan trial, 0.5 m² plots (0.5 m × 1 m) were used for manually controlled drought treatment. Each plot contained 10 rice plants arranged in a 2 × 5 grid. Irrigation was terminated at 40 days after transplanting (DAT), and a one-month drought period was maintained during the critical panicle development stage (drought-sensitive period). During this period, a movable rain-off shelter was used to block all rainfall to ensure well-controlled stress; the relative soil water content (RSWC) decreased to approximately 8% and was maintained for 5 days, after which manual irrigation was conducted to support normal rice growth. For the 2025 Wuhan trial, 2.25 m² plots (1.5 m × 1.5 m) were adopted for manually controlled drought. Each plot contained 64 rice plants arranged in an 8 × 8 grid. The manually controlled drought treatment was consistent with that in the 2024 Wuhan trial. Promoter activity assay To compare promoter activity, the native NLR8-Hap1 promoter and its two mutant derivatives carrying either the 25-bp deletion (InDel 9656) or the A-to-G substitution at the lead SNP were inserted into the pGreenII-0800 dual-luciferase vector, with firefly luciferase (F-Luc) as the reporter and Renilla luciferase (R-Luc) under the 35S promoter as the internal reference. The constructs were transformed into rice protoplasts and cultured for 16 h. Dual-luciferase activity was measured using the Dual-Luciferase® Reporter Assay System (Promega). Promoter activity was quantified as the F-Luc/R-Luc ratio. Yeast two-hybrid (Y2H) screening and interaction assay For Y2H screening, the full-length coding sequences (CDS) of NLR8-Hap1 and NLR8-Hap2 were individually fused to the GAL4 DNA-binding domain in pGBKT7, and the resulting constructs were transformed into the yeast strain Y2HGold. After confirming that there was no autoactivation or toxicity, these bait strains were mated with a rice whole-plant cDNA library cloned into pGADT7 (Clontech), following the manufacturer’s protocols. Positive interactions were selected on quadruple dropout medium (SD/-Trp/-Leu/-His/-Ade) containing X-α-Gal and validated by sequencing. For targeted Y2H assays, full-length CDSs of NLR8 variants and their identified interactors were cloned into pGBKT7 and pGADT7, respectively. Empty vectors served as negative controls, with interactions assessed by growth on selective media and β-galactosidase activity. Bimolecular fluorescence complementation (BiFC) assay For BiFC assays in tobacco leaves, the full-length CDSs of NLR8 variants and their interaction partners were fused to split yellow fluorescent protein (YFP) fragments in the pEarleygate201-YN and pEarleygate202-YC vectors. These constructs were transformed into Agrobacterium tumefaciens strain GV3101 (containing the p19 silencing suppressor) and infiltrated into leaves of three-week-old Nicotiana benthamiana plants. YFP fluorescence was visualized 48 h post-infiltration using a Zeiss LSM980 confocal microscope. Images were acquired using Zeiss ZEN v3.10. Split luciferase complementation (SLC) assay For SLC assays in tobacco leaves, the full-length CDS of NLR8 variants and their interacting proteins were fused to the N-terminal (nLuc) and C-terminal (cLuc) fragments of luciferase in the respective nLuc and cLuc vectors. Following Agrobacterium-mediated expression, luciferase activity was measured 48 h post-infiltration using a chemiluminescence imaging system (Tanon 5200) after application of 1 mM D-luciferin substrate. Each interaction assay included appropriate negative controls and was performed in at least three independent replicates. Rice protoplast isolation and transformation To isolate protoplasts, two-week-old rice seedlings were cut into 0.5-mm strips and digested in solution buffer (0.6 M mannitol, 10 mM MES-KOH [pH 5.7], 1.5% Cellulase RS, 0.75% Macerozyme R10, 0.1% BSA, and 1 mM CaCl2). After a 4-h digestion at room temperature in darkness with gentle shaking (60 rpm), protoplasts were filtered through a 300-mesh nylon sieve and collected by centrifugation (200 × g, 5 min). Purified protoplasts were sequentially washed with W5 solution (154 mM NaCl, 125 mM CaCl2, 5 mM KCl, 2 mM MES-KOH [pH 5.7]) and resuspended in MMG buffer (0.6 M mannitol, 15 mM MgCl2, and 4 mM MES-KOH [pH 5.7]) at a concentration of 1.0 × 107mL–1. For each transformation, 100 μL of protoplasts was gently mixed with 10 μg of plasmid DNA and an equal volume of PEG solution (40% PEG4000, 0.6 M mannitol, 100 mM CaCl2). After a 15-min incubation at room temperature, the reaction was stopped by adding two volumes of W5 solution. Transformed protoplasts were collected by centrifugation (200 × g, 5 min), resuspended in WI solution (0.6 M mannitol, 4 mM KCl, 4 mM MES-KOH [pH 5.7]), and incubated in darkness at room temperature for 16-20 h before analysis. Transformation efficiency was assessed using a GFP-expressing control plasmid. Co-immunoprecipitation (Co-IP) For Co-IP experiments in rice protoplasts, NLR8 variants and their interacting partners were tagged with 3× FLAG and GFP, respectively. These fusion constructs were co-expressed in rice protoplasts for 16 h. Total protein was extracted in Co-IP buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1 mM EDTA, 0.5% Triton X-100, 1 mM DTT, 1 mM PMSF, and 1× protease Inhibitor Cocktail). Cleared lysates (14,000 × g, 10 min, 4 °C) were immunoprecipitated with anti-FLAG M2 magnetic beads for 4–8 h at 4 °C. Beads were washed five times with Co-IP buffer, and bound proteins were eluted by boiling in SDS sample buffer. Both input and immunoprecipitated fractions were analyzed by immunoblotting using anti-GFP and anti-FLAG antibodies. Subcellular localization Subcellular distribution patterns of NLR8 and TFs were examined using fluorescent protein fusions. Full-length CDS were cloned into PM999 vectors to generate CFP-, GFP-, or YFP-tagged fusion proteins. These constructs were co-expressed in rice protoplasts alongside the nuclear marker Ghd7-CFP or Ghd7-mCherry47. After 16 h of expression, fluorescence signals were acquired using a Zeiss LSM980 confocal microscope and Zeiss ZEN v3.10, with fluorophore-specific excitation and emission settings. Blast-resistance evaluation Blast resistance was assessed using a punch-inoculation method on rice seedlings at the 3–4-leaf stage21. Specifically, blast fungus isolates (99-20-2 and ZB25) were cultured on complete agar medium at 25 °C under a 12-h light/12-h dark cycle for 10 days. Spores were harvested in sterile water and adjusted to a concentration of 1 × 105 spores/mL. Inoculation was performed by applying spore suspensions to punched leaf areas. Disease symptoms were evaluated 7 days post-inoculation by measuring lesion length and area using ImageJ v2.3.0. For the evaluation of natural blast resistance in the field, MH63 and MH63-NIL(Hap2) plants were grown in a naturally blast-prone nursery in Enshi Prefecture. Disease severity was assessed at the heading stage. Infected leaves from each genotype were harvested, and the corresponding lesion length and lesion area were quantitatively measured to evaluate field blast resistance. Cell-free protein degradation assay For recombinant protein preparation, the full-length CDSs of TFs were cloned into the pMAL-c2x vector to generate maltose-binding protein (MBP) fusion proteins, which were expressed in E. coli strain BL21 (DE3). Total protein was extracted from 0.1 g rice sample using 500 μL degradation buffer (25 mM Tris-HCl [pH 7.5], 10 mM NaCl, 10 mM MgCl2). Purified MBP fusion proteins were incubated with rice protein extracts supplemented with 10 mM ATP and 5 mM PMSF. Reactions, with or without 50 μM MG132, were conducted at 30 °C for 0, 1, 2, and 4 h and terminated by adding SDS loading buffer. Protein degradation was analyzed by western blotting with an anti-MBP antibody. Rubisco was used as a loading control, and band intensities were quantified using ImageJ v2.3.0 after normalization to the corresponding Rubisco signal. Nuclear fractionation To isolate cytoplasmic and nuclear proteins from rice protoplasts, collected protoplasts were resuspended in cell lysis buffer (20 mM Tris-HCl [pH 7.5], 10 mM NaCl, 3 mM MgCl2, 0.5% NP-40, 1 mM PMSF) and incubated on ice for 15 min to lyse the cells. The lysate was centrifuged at 1000 × g for 10 min at 4 °C, and the supernatant was collected as the cytoplasmic protein fraction. The resulting nuclear pellet was washed twice with cell lysis buffer to remove residual cytoplasmic proteins. Nuclear proteins were subsequently extracted by resuspending the nuclear pellet in SDS loading buffer and heating at 95 °C for 5 min. Immunoblotting For immunoblotting, protein samples were separated by SDS–PAGE and transferred to PVDF membranes. Primary antibodies used were anti-FLAG (Sigma-Aldrich, F3165; 1:10,000), anti-HA (ABclonal, AE105; 1:3000), anti-MBP (ABclonal, AE016; 1:10,000), anti-GFP (ABclonal, AE011; 1:5000), anti-Actin (ABclonal, AC009; 1:10,000), anti-Histone H3 (ABMART, P30266S; 1:50,000), and anti-Luciferase (Sigma-Aldrich, L0159; 1:10,000). RNA extraction and quantitative real-time PCR (RT-qPCR) Total RNA was extracted using the TRIzol reagent (TransGen Biotech, ET101-01) according to the manufacturer’s protocol. Reverse transcription was performed using the RT Master Mix with gDNA Remover kit (Cookgen, KM405) to synthesize cDNA. Quantitative real-time PCR (RT-qPCR) was conducted using 2× SYBR qPCR Mix (Cookgen, KKR012) on a QuantStudio 6 Flex Real-Time PCR system (Thermo Fisher Scientific). Data were collected and initially processed using QuantStudio Real-Time PCR Software v1.1. Gene expression levels were normalized to Ubiquitin and OsActin, as specified for each experiment, and relative expression was calculated using the ΔΔCt method. Primer amplification efficiency (ranging from 90% to 110%) was validated by the standard curve method. Statistics and reproducibility Unless otherwise specified, R (v4.5.2; https://cran.r-project.org/) was used for statistical analysis and plot generation. Two-sided Welch’s t-tests and Wilcoxon tests were performed using the “t.test” and “wilcox.test” functions, respectively. One-way ANOVA followed by Tukey’s honest significant difference test was performed using the “aov” and “TukeyHSD” functions. Unless otherwise specified, default parameters were used. The experiments shown in Fig. 2b, d, f, Fig. 4a–e and g–k, Fig. 5e–h, and Supplementary Figs. 5a–c, 9b, 9e, and 10b were repeated independently at least twice with similar results. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Peer Review file Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 Supplementary Data 3 Supplementary Data 4 Reporting Summary Supplementary Information Peer Review file Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 Supplementary Data 3 Supplementary Data 4 Reporting Summary Source data Source Data Source Data