A natural variation in rice susceptibility factor Rsf1 confers resistance to root-knot nematode Authors: Jingwen Yu, Chao Xiang, Deliang Peng, Huan Peng, Shiming Liu, Lingan Kong, Ling Qing, Mingjun Li, Xiyue Yu, Huanhuan Hao, Yaqin Liu, Wencui Wu, Xue Han, Yanxue Chen, Houxiang Kang, Wenkun Huang Published: 2026 ABSTRACT AbstractRice root-knot nematode (RRKN, Meloidogyne graminicola) is a destructive nematodes threatening global rice production. Few RRKN-resistant resources have been reported and their natural variations remain largely unknown. Here, we identify RRKN susceptibility factor 1 (Rsf1), encoding a cysteine-rich protein, via Genome-Wide Association Study (GWAS), multi-omics analysis, and functional validation. A natural variation in the TATA-box of OsRsf1 is observed, wherein RRKN-susceptible rice accessions contain haplotype (Hap) A while resistant accessions contain HapG. OsRsf1 with HapA is highly expressed and induced by RRKN infection, while HapG OsRsf1 shows low expression, conferring enhanced RRKN resistance. We further demonstrate that transcription factor Myb44 promotes OsRsf1 expression. Knockouts of OsRsf1 in rice and its ortholog AtRsf1 in Arabidopsis significantly enhanced nematode resistance, indicating that functional conservation. Mutations in OsRsf1 trigger a general defence response during RRKN infection. Our finding reveals a conserved Rsf1-mediated susceptibility mechanism underlying RRKN infection and provides a promising target for breeding nematode-resistant crops.Subject terms: Molecular engineering in plants, Transgenic plants, Biotic, Agricultural genetics Introduction Plant parasitic nematodes (PPNs) have a devastating impact on global crop yields, causing annual economic losses of 80−173 billion USD1–3. Root-knot nematodes (RKNs) are the most severe PPNs, infecting almost all economically important crops4,5. More than 100 different PPNs infect rice (Oryza sativa L.), the staple food for at least two-thirds of the global population6, causing 10–25% yield loss annually in infected rice fields worldwide7,8. Rice root-knot nematode (RRKN, Meloidogyne graminicola) is one of the most damaging PPN. RRKN occurs in upland, lowland, and deepwater rice, as well as in rice grown in nurseries9. Between 2013 and 2022, RRKN affect an average of 30,282 hectares of land in China each year, causing 20–80% yield losses10,11. The rising incidence of RRKN is closely associated with increasing water scarcity and changes in rice cultivation practices12. Classical breeding for nematode-resistant crops typically relies on the deployment of conventional resistance (R) genes. Resistance to RKNs is primarily governed by multiple quantitative trait loci (QTLs). To date, more than 30 resistance QTLs associated with RKN resistance have been identified13,14. For example, tomato Mi-1.2 gene15 encodes a canonical NB-LRR resistance protein that has been widely utilized to confer resistance against RKN. Similarly, rice MG1 gene12 encodes an NB-LRR protein that interacts with a protease inhibitor and confer resistance to RRKN. R gene-mediated resistance is quick and strong due to the specific recognition between R protein and pathogen effector. However, only a few nematode-resistant R genes were identified, and nematodes often readily overcome the resistance mediated by these genes upon large-scale deployment. For example, tomato Mi-1-mediated resistance was broken by RKNs16. Nearly 88% of soybean cyst nematode populations in Indiana have overcome the soybean resistance gene Rhg1-b17. In recent years, genome-wide association studies (GWAS) have proven to be a powerful tool for dissecting the genetic architecture of host resistance to RRKN in rice18,19. For instance, an investigation utilizing a global rice diversity panel identified 11 genomic regions associated with RRKN resistance, located on chromosomes 1, 3, 4, 5, 11, and 1218. Similarly, a GWAS study of Indian wild rice accessions uncovered 17 single-nucleotide polymorphisms (SNPs) significantly linked to RRKN resistance19. With the advancement of technologies such as gene editing, mutants of susceptibility (S) genes have emerged as an alternative approach in nematode-resistance breeding20. Several nematode-associated S genes have been reported. For instance, a loss-of-susceptibility mutant of the soybean α-SNAP gene confers resistance to soybean cyst nematode Heterodera glycine21. Similarly, mutant of Arabidopsis ERF6 gene improves resistance to RKN M. incognita22. Additionally, disruption of the OsHPP04 S gene enhances resistance to RRKN23,24. Pathogens often utilize host S genes to facilitate infection by altering cell structure, manipulating nutrient transport, or suppressing immune signaling. For instance, powdery mildew fungi exploit barley S gene Mlo25, which regulates cell wall protrusions, to facilitate host invasion. Stripe rust fungi utilize wheat S gene TaPsIPK126, which modulates phosphoinositide-mediated membrane signaling, to suppress host immune responses. Similarly, the rice bacterial leaf streak pathogen manipulates sugar transporter S genes SWEET27 to draw nutrients and metabolic resources from the host. In another case, rice blast fungus exploits S genes Pi2128 and Bsr-d129 to suppress host defense responses. To facilitate parasitism, RRKN secretes effectors, such as MgGPP, to suppress host defense responses30. In addition, many nematode effectors also manipulate host developmental reprogramming, cellular processes, and transport nutrients towards the feeding site31. For instance, cyst nematode CLE effectors mimic plant peptide hormones to reprogram host developmental pathways and induce syncytium formation32,33. RKN secretes effector MiPFN3 targets actin cytoskeleton to induce depolymerization, facilitating structural reorganization for parasitism34. Mi8D05 of M. incognita physically interacts with the plant aquaporin TIP-2 and is predicted to facilitate giant cell formation by modulating turgor pressure through water and solute transport35. In this work, we conduct a large-scale field evaluation for nematode resistance using a diversified rice population comprising 345 accessions. By integrating GWAS with multi-omics analyses and functional validation, we identify OsRsf1, a gene encoding a cysteine-rich protein, as a key susceptibility factor for RRKN. We demonstrate that the disruption of OsRsf1 results in robust resistance to RRKN in rice. We also find that T-DNA insertional mutants of AtRsf1, the Arabidopsis ortholog of OsRsf1, also significantly enhance RRKN resistance in Arabidopsis, indicating that Rsf1 functions as a conserved S gene in both rice and Arabidopsis. Results Rsf1 is tightly associated with RRKN resistance To identify potential rice resistance resources, we conducted a large-scale field screen for RRKN resistance using 345 representative rice accessions from the Core Rice Diversity Panel II (C-RDP-II)36, including 56 aromatic (ARO), 57 aus (AUS), 91 indica (IND), 61 temperate japonica (TEJ), and 75 tropical japonica (TRJ) accessions (Fig. 1a). Various responses to RRKN infection were observed among rice accessions in C-RDP-II (Suppl. Data 1). The result showed that only nine accessions were resistant, 142 accessions exhibited moderate resistance, 139 accessions were moderately susceptible, 54 accessions were susceptible, and one accession was highly susceptible (Fig. 1b and Suppl. Fig. 1a–c). No significant differences in RRKN susceptibility were observed among the five subpopulations (Suppl. Fig. 1d). Distribution histograms and P–P plots (Suppl. Fig. 1e, f) indicated that the resistance trait among these rice populations followed a normal distribution. Two accessions, KPM18 (Khao Pahk Maw, a TEJ accession from Thailand) and PTB30 (IRGC6304-2, an ARO accession from India), along with the known resistant japonica cultivar ZH1137, exhibited strong resistance to RRKN under controlled conditions (Fig. 1c–f). Evaluation of RRKN resistance under field and greenhouse conditions. a Phylogenetic tree for the 345 accessions in the C-RDP-II population. b Distribution of gall index of C-RDP-II rice accessions. R resistant, MR moderately resistant, MS moderately susceptible, S susceptible, HS highly susceptible. c Gall index of RRKN resistant accessions (KPM, PTB30, ZH11) and susceptible control NPB at 14 dpi under greenhouse conditions. d RRKN disease phenotypes in roots of resistant accessions (KPM, PTB30, ZH11) and susceptible control NPB at 14 dpi under greenhouse conditions. Arrowheads indicate galls. e Gall numbers in resistant (KPM, PTB30, ZH11) and susceptible (NPB) accessions at 14 dpi with RRKN. Each fourteen-day-old rice seedling was inoculated with 300 J2s RRKN. Sample sizes (n = number of plants) and the error bars indicate statistical significance groups at P < 0.05 (One-way ANOVA analysis followed by Tukey’s multiple comparisons test). All experiments were performed three times with similar results. Data are means ± SD. f Acid fuchsin staining of rice roots at 14 dpi. The top panel shows stained rice roots with arrowheads indicating galls. The bottom panel shows images of a typical gall. Scale bars, 4 mm (Top), 200 μm (Bottom). Source data are provided as a Source Data file. To elucidate the genetic architecture underlying RRKN resistance, we conducted a genome-wide association study using 673,494 single-nucleotide polymorphism (SNP) markers in the C-RDP-II population. We identified 358 SNPs that were significantly associated with RRKN resistance (Fig. 2a and Suppl. Data 2). These SNPs delineated 18 distinct loci associated with RRKN resistance (LARRKNs) (Suppl. Data 3). Of the 18 identified LARRKNs, three were co-localized with previously reported quantitative trait loci (QTLs) for RRKN resistance. Specifically, LARRKN7 overlapped with qGN6.1/qGN6.214, LARRKN13 with the OsbZIP8019 locus, and LARRKN17 with OsGF14d19. The remaining 15 loci represent RRKN-associated regions identified in this study. Genome-wide association study (GWAS) for rice RRKN resistance, coupled with genetic and allelic variation analyses for candidate loci. a Manhattan plot of RRKN resistance-associated loci. Red dots indicate SNPs with −log10(P) > 4. Two-sided P-values were determined using canonical correlation analysis. Arrows indicate previously identified QTLs, including qGN6.1/qGN6.2 (in LARRKN7), OsbZIP80 (in LARRKN13), and OsGF14d (in LARRKN17), and the Rsf1 gene (in LARRKN8). b Heatmap of top 10 significantly differentially expressed genes in LARRKN8 during RRKN infestation. NPB: Nipponbare. c Relative RNA expression levels of Rsf1 in the roots of resistant (KPM, PTB30) and susceptible (NPB) rice accessions under non-infected conditions. Data represent means ± SD from three independent biological replicates. d Relative RNA expression levels of Rsf1 in the roots of resistant (KPM, PTB30) and susceptible (NPB) rice accessions with or without (mock) RRKN infection. e Haplotype analysis for Rsf1 in 12 RRKN-susceptible and 12 RRKN-resistant rice accessions. f Linkage disequilibrium (LD) heat map of the Rsf1 gene. The LD value indicates the linkage between two SNPs. g Dual-luciferase activity assay for Rsf1 promoter variants. Left: reporter constructs carrying a HapA- or HapG-type TATA-box. Right: LUC/REN activities of promoter fragments of HapA (200, 500, 1000 bp) and HapG (200, 500, 1000 bp). LUC/REN ratios were measured 16 h after transfection, with empty vector (EV) as a control. Different letters above the bars in (c, d, g) indicate statistical significance groups at P < 0.05 (one-way ANOVA followed by Tukey’s multiple comparisons test). Experiments were independently performed three times. Data represent means ± SD. Source data are provided as a Source Data file. LARRKN8, the largest peak located on chromosome 7 identified in this study, did not overlap with any previously characterized RRKN-resistance region (Fig. 2a). We therefore focused on LARRKN8. To further investigate the genetic basis of LARRKN8, we analyzed a 4.5 Mbp genomic region surrounding the top-associated SNPs using the Nipponbare (NPB) reference genome (MSU v7.0). This region contains 175 annotated protein-coding genes. In NPB (RRKN susceptible rice), RNA-seq analysis revealed that four Thionin-like genes, Thil30 (LOC_Os07g24830, subsequently designated as Rsf1), Thil60 (LOC_Os07g25060), Thil20 (LOC_Os07g24820) and Thil50 (LOC_Os07g25050), in LARRKN8 were significantly upregulated (-log2 (fold change) ≥ 1, p < 0.05) at 3 days post-infestation (dpi) with RRKN (Fig. 2b and Supplementary Data 4), showing FPKM values of 1235.6 (Rsf1), 184.3 (Thil60), 2.4 (Thil20) and 816.8 (Thil50), respectively (Suppl. Fig. 2a, b). These four genes encode cysteine-rich proteins, sharing 56.12%−56.55% sequence identity with the previously reported Thionin1/2/6/7 proteins38 (Supplementary Fig. 2c, d). To assess allelic variations, we cloned these Thionin-like genes and sequenced from both resistant and susceptible accessions. In particular, Rsf1 contained three haplotypes: Hap1/Hap2 (encoding the same protein) and Hap3 (Suppl. Fig. 3a–d). Hap3 was predominant in the C-RDP-II population, with a frequency of 70.3% (Suppl. Table 1). To characterize their functions in rice response to nematode infestation, we generated overexpression (OE) lines for each of the five distinct candidate genes in the NPB background (Suppl. Fig. 4a, b). Specifically, these lines overexpressed either Rsf1_Hap1, Rsf1_Hap3, Thil60, Thil20, or Thil50. Preliminary assessment revealed that all transgenic lines displayed normal growth and development, with phenotypes comparable to those of wild-type NPB plants (Suppl. Fig. 4c–f). Our RRKN inoculation results showed that both Rsf1_Hap3-OE (Rsf1-OE) and Thil50-OE lines exhibited significantly increased susceptibility to RRKN compared to wild-type NPB (Suppl. Fig. 4g, h). In contrast, no significant differences in RRKN resistance were observed for Thil60-OE, Thil20-OE, or Rsf1_Hap1-OE lines. To further investigate the roles of Rsf1 and Thil50 in nematode infection, we conducted proteomic analysis for the nematode infection process in NPB39. The results demonstrated that only Rsf1 protein level was significantly induced by RRKN infection (P < 0.0001) (Suppl. Fig. 5a). The Rsf1 protein contains 11.6% (11 of 95) cysteine and is localized subcellularly to the plasma membrane (Suppl. Fig. 5b–d). Subsequently, we utilized CRISPR/Cas9 technology to generate Rsf1 knockout lines in the NPB genetic background. Three independent homozygous Rsf1-KO lines, each confirmed to harbor a frame-shift mutation, were characterized (Supplementary Fig. 6a). These mutants unequivocally showed significantly enhanced resistance to RRKN compared to wild-type NPB, evidenced by a drastic reduction in both gall and nematode counts (Suppl. Fig. 6b–i). These results collectively support a role for Thil30 as a susceptibility factor associated with RRKN infection in rice. Accordingly, we named it Rsf1 (RRKN susceptibility factor1). Natural variation in Rsf1 TATA-box is associated with RRKN resistance We examined Rsf1 gene expression among resistant and susceptible rice accessions before and after nematode infestation. The results revealed that the Rsf1 RNA level was markedly and significantly higher in the susceptible accession (NPB) than in the resistant accessions (KPM and PTB30) without nematode infection (Fig. 2c). Upon RRKN infection, Rsf1 remained high and was further induced in NPB at 3 dpi, but remained low and not induced in KPM and PTB30 (Fig. 2d). We quantified Rsf1 RNA levels in 12 susceptible and 17 resistant rice accessions (9 resistant accessions and 8 moderately resistant accessions; see Suppl. Data 5) before and after inoculation with RRKN. Consistently, Rsf1 RNA levels were significantly higher in susceptible accessions than in resistant ones without RRKN inoculation. Upon inoculation, Rsf1 was further induced in susceptible accessions, yielding a 6.1-fold (P = 0.0006) greater RNA level compared to those in resistant accessions at 3 dpi (Suppl. Fig. 7a, b). This phenomenon was consistent with previous observation that Rsf1 expression was significantly induced by RRKN in the susceptible cultivar NPB (O. sativa), but not induced in resistant African rice species TOG5681 (O. glaberrima)40. To assess the cause of differential Rsf1 expression levels between RRKN-resistant and RRKN-susceptible rice accessions, we analysed 1,903-bp genomic sequences of Rsf1 from 12 resistant and 12 susceptible rice accessions. The analyzed region covering promoter region (1 kb), transcription start site (TSS), 5’ untranslated region (UTR, 73 bp), two exons (288 bp in total) separated by a single intron (242 bp), and the 3’ UTR (300 bp). This analysis uncovered a total of 31 SNPs in Rsf1 (Fig. 2e, f and Suppl. Data 6). At SNP-14118000 (TATATAA/G), 59% of accessions carrying Haplotype_G (HapG) were resistant (R and MR), whereas 71.4% of accessions carrying Haplotype_A (HapA) were susceptible (HS, S, and MS) in the C-RDP-II population (Table 1 and Suppl. Data 7). Linkage disequilibrium (LD) analysis indicated that SNP-14118000 and SNP-14117493 were tightly linked within Rsf1 gene (Fig. 2f). We furtherly employed transcriptional activation assays to test the effects of SNP-14118000 and SNP-14117493. The results showed that the HapA (200 bp) allele of SNP-14118000 conferred significantly higher promoter activity than the HapG allele, while SNP-14117493 had no observable differences (Fig. 2g and Suppl. Fig. 7c). SNP-14118000 is located within the TATA-box region. Table 1Frequencies of Rsf1 TATA-box haplotypes in the C-RDP-II populationHaplotypeResistant accessionsSusceptible accessionsRMRMSSHSHapA3.6% (1 of 28)25.0% (7 of 28)50.0% (14 of 28)17.8% (5 of 28)3.6% (1 of 28)HapG8.0% (8 of 100)51.0% (51 of 100)40.0% (40 of 100)1.0% (1 of 100)0% (0 of 100)R resistant, MR moderately resistant, MS moderately susceptible, S susceptible, HS highly susceptible. Source data are provided as a Source Data file. Frequencies of Rsf1 TATA-box haplotypes in the C-RDP-II population R resistant, MR moderately resistant, MS moderately susceptible, S susceptible, HS highly susceptible. Source data are provided as a Source Data file. We analyzed the geographic distribution of SNP-14118000 HapA and HapG in diverse rice germplasms. Compared with the distribution regions of RRKN (https://gd.eppo.int/taxon/MELGGC/distribution, reference time: 2025-10-30), we found that the distribution of HapG accessions is associated with regions of severe RRKN infestation (R2 = 0.767, P < 0.0001) (Fig. 3a, b). HapG was predominantly found in O. sativa rufipogon and O. sativa indica (Fig. 3c, d). Infrequent gene flow of Rsf1 was detected between O. sativa rufipogon and O. sativa indica (Fig. 3e, f, and Suppl. Data 8). Neutral tests suggest that the Rsf1 gene underwent a purifying selection during the domestication of cultivated rice from wild rice (Fig. 3g and Suppl. Data 9). Geographic distribution and evolutionary patterns of Rsf1. a Geographic distribution of HapA- and HapG-type TATA-box of Rsf1 promoter among cultivated and wild rice accessions. Green and blue represent HapA and HapG, respectively. The map was downloaded from the EPPO Global Database (https://gd.eppo.int/taxon/MELGGC/distribution, reference time: 2025-10-30). b Gall index of HapA and HapG rice accessions in the C-RDP-II. The bars within violin plots represent the 25th percentile, median, and 75th percentile; statistical significance (P-value) was determined using a two-sided Student’s t-test. c, d Haplotype network of Rsf1 constructed using the coding region and 1 kb upstream sequences of cultivated and wild rice accessions. Black dots along the branches represent mutational steps between haplotypes. Ind: O. sativa indica; Tej: O. sativa temperate japonica; Trj: O. sativa tropical japonica; Ruf: wild rice O. rufipogon. e Fst values of Rsf1 reflects the level of genetic differentiation between cultivated and wild rice. Jap: O. sativa japonica (Tej and Trj). f Nm values of Rsf1 reflects the number of migrants between cultivated and wild rice. |Fst| > 0.33 or |Nm|< 1 suggests infrequent gene flow; |Fst|< 0.33 or |Nm|> 1 suggests frequent gene flow. g Neutrality tests of cultivated and wild rice accessions. Source data are provided as a Source Data file. Rsf1 expression is regulated by transcription factor Myb44 To further investigate how Rsf1 expression is regulated, we used the core promoter region (1000 bp) of Rsf1 to pull down nucleoproteins of nematode-infected rice roots. Protein complexes (Fig. 4a, red arrows) bound to the promoter were isolated after protein-DNA pull-down assays followed by mass spectrometry (LC-MS). LC-MS identified 24 annotated proteins that were specifically pulled down by the Rsf1 promoter at 0/1/3 dpi. Among these proteins, only Myb44 (LOC_Os06g14010) was found to be a transcription factor (Fig. 4b, Suppl. Data 10 and Suppl. Fig. 8a–f), which was therefore a priority for further study. Electrophoretic mobility shift assays (EMSA) demonstrated that Myb44 directly binds to the Rsf1 promoter, and this binding was unaffected by natural variation (HapA and HapG) within the Rsf1 TATA-box region (Fig. 4c, Suppl. Fig. 9a, b). Myb44 is induced upon nematode infection in the susceptible accession (NPB, carry HapA) and the resistant accessions (KPM and PTB30, carry HapG), with higher expression in NPB at 3 dpi (P < 0.05) (Suppl. Fig. 9c). In rice protoplast transactivation assays, Myb44 significantly activated the Rsf1 promoter, but activation level was 3.7-fold lower for HapG than HapA (P < 0.0001), indicating that natural variation within the Rsf1 TATA-box region affects Myb44-mediated transcriptional activation (Fig. 4d). Additionally, confocal microscopy revealed that Myb44 is localized in the rice nucleus (Fig. 4e). Myb44 directly binds to Rsf1 promoter to regulate its expression. a Protein-DNA pull-down assay using biotin-labeled or unlabeled Rsf1 promoter as probe to pull down total nuclear proteins extracted from rice roots at 0, 1, and 3 dpi. Biotin-labeled DNA-protein complexes were pulled down with streptavidin-labeled beads. Proteins pulled down were separated in an SDS–PAGE. b Venn diagram showing numbers of proteins that specifically bind to the Rsf1 promoter. c EMSA experiments for binding of Myb44 to the TAACCA motif in the Rsf1 promoter. Wildtype competitors were added at 20-, 100-, or 200-fold excess over the labeled probe, while mutant competitors were added at 200-fold excess. d Dual-luciferase assay for transcriptional activation of Rsf1 promoter (HapA and HapG) by Myb44 in rice protoplasts. LUC/REN ratios were measured 16 h post-transfection. Data are mean ± SD from three independent experiments. Statistical significance (P-value) was determined by a two-sided Student’s t-test. e Subcellular localization of Myb44 in rice protoplasts. Scale bars, 4 μm. f Rsf1 RNA levels in wild-type (NPB, WT) and Myb44-knockout (Myb44-KO-1 and Myb44-KO-15, NPB background) mutants at 1, 3, 7, and 14 dpi by RRKN measured by RT-qPCR. Data are presented as mean ± SD of three independent biological replicates. P-values were determined by one-way ANOVA followed by Tukey’s multiple comparisons test. g, h Gall numbers (g) and nematode numbers (h) in Myb44 knockout and WT plants at 14 dpi. Each fourteen-day-old rice seedling was inoculated with 300 J2s RRKN. Three biological replicates were performed. Sample sizes are indicated by n = number of plants; in each boxplot, the center line indicates the median, the box edges represent the first and third quartiles, and the whiskers extend to the minima and maxima. P-values were determined by one-way ANOVA followed by Tukey’s multiple comparisons test. i Acid fuchsin staining of roots at 14 dpi. Top: overview; bottom: representative gall. Scale bars: 1 cm (top) and 200 μm (bottom). Source data are provided as a Source Data file. To further investigate the role of Myb44 in Rsf1-mediated nematode infection progress, we generated Myb44 knockout lines (Myb44-KO-1 and Myb44-KO-15) in the NPB background using CRISPR/Cas9 (Supplementary Fig. 9d). All mutant lines developed normally without any detectable growth abnormalities (Supplementary Fig. 9e–h). RT-qPCR assays showed a substantial reduction in Rsf1 RNA levels at 3 dpi with nematodes in the roots of Myb44-KO lines compared to that in WT roots, further confirming the positive role of Myb44 in Rsf1 expression (Fig. 4f). The inoculation results revealed that the Myb44-KO lines exhibited higher RRKN resistance, showing significantly reduced numbers of root galls and nematodes compared to WT (P < 0.0001, Fig. 4g–i and Supplementary Fig. 9i, j). These findings collectively suggest that transcription factor Myb44 binds to the Rsf1 promoter to promote its expression, contributing to RRKN infection. Rsf1 facilitates RRKN parasitism in rice To further assess the biological function of Rsf1, we tested purified recombinant Rsf1-GST fusion protein and GST control protein (expressed from pGEX-4T-1) for in vitro activity assays to see whether Rsf1 affects RRKN activity. The results showed that the Rsf1-GST protein has no significant effects on nematode activity compared to water, elution buffer, and GST protein controls (Supplementary Fig. 10a). To evaluate how Rsf1 indirectly influences RRKN infection, we conducted nematode attraction assays using Rsf1 knockout (Rsf1-KO) and overexpression (Rsf1-OE) lines. The results showed that Rsf1-KO rice seedlings most significantly reduced attraction to RRKN at the 0–1.5 cm zone 7dpi, showing a 61.37% decrease in the number of nematodes migrating toward root surfaces (Fig. 5a, b). In contrast, Rsf1-OE seedlings did not show a significant difference in nematode distribution compared to wild-type (NPB) control across all measured distance intervals (Suppl. Fig. 10b). These findings suggest that knockout of Rsf1 substantially reduces the attractiveness of rice roots to RRKN, thereby contributing to enhanced resistance. Rsf1 facilitates RRKN infestation in rice. a Schematic of the experimental design for testing attraction of wild-type (NPB, WT) and Rsf1-KO plants to RRKN. Created in BioRender. JW, Y. (2026) https://BioRender.com/dl2sllb and https://BioRender.com/b0lvjti. b Attraction of WT and Rsf1-KO plants to RRKN at different distances. Sample sizes are indicated by n = number of plants; in each boxplot, the center line indicates the median, the box edges represent the first and third quartiles, and the whiskers extend to the minima and maxima. Statistical significance (P value) was determined using two-sided Student’s t-test. c, d Percentage of J3 (c) and J4/female (d) nematodes in WT, Rsf1-KO and Rsf1-OE plants. Each fourteen-day-old rice seedling was inoculated with 300 J2 RRKN. P-values were determined by one-way ANOVA followed by Tukey’s multiple comparisons test. e Acidic magenta staining of rice roots at 14 dpi. The white fonts in the figure indicate the stages of RRKN in the root galls of WT, Rsf1-KO and Rsf1-OE plants. J3 represents third-stage juveniles; J4 represents fourth-stage juveniles; female represents mature adults. Scale bar = 400 μm. In (c, d), the center line indicates the median, the box edges represent the first and third quartiles, and the whiskers extend to the minima and maxima. Source data are provided as a Source Data file. We next conducted a greenhouse experiment to further evaluate the effects of Rsf1 mutants on RRKN infection. Microscopic observations after staining at 14 dpi showed that more RRKN juveniles remained at the third stage (J3) in Rsf1-KO plants, while fewer developed to the fourth stage (J4) and adult females (34.98%) than in WT (68.60%) and Rsf1-OE (63.07%) roots (Fig. 5c–e). These findings indicated that knockout of Rsf1 in rice delays nematode development within rice roots. Field experiments in a heavy RRKN field (Hanshou, Hunan province) further showed that Rsf1-KO plants displayed a 63.68% reduction in root gall formation (Fig. 6a and Supplementary Fig. 10c, d) and Rsf1-KO rescued over 16.12% (Rsf1-KO plants) of the 25.01% grain yield loss in WT plants ((30.15 g in Fig. 6p - 22.61 g in Fig. 6g)/30.15 g in Fig. 6p) (Fig. 6b–i). Furthermore, in the absence of RRKN stress, no significant differences were detected in major agronomic traits including yield-related parameters, among Rsf1-OE, Rsf1-KO, and WT plants (Fig. 6j–r). These results collectively indicated that manipulating Rsf1 does not exert deleterious effects on plant development. Rsf1-KO plants show significantly lower grain yield losses caused by RRKN in field. a Number of galls in wild-type (NPB, WT), Rsf1-KO and Rsf1-OE plants grown in RRKN-infested fields. The bars within violin plots represent the 25th percentile, median, and 75th percentile. b Root images with galls of Rsf1-KO and Rsf1-OE plants grown in RRKN-infested fields. c–r Plant height (c, j), number of grains (e), grain weight (g, p), number of tillers (l), and representative images for each measurement (d, f, h, i, k, m, o, q, r) of WT, Rsf1-KO and Rsf1-OE plants under RRKN-infested (c–i) or non-infested (j–r) condition in the field. In (c, e, g, j, l, n, p), sample sizes are indicated by n = number of plants; the center line indicates the median, the box edges represent the first and third quartiles, and the whiskers extend to the minima and maxima. Different letters above the bars in (a, c, e, g, j, l, n, p) indicate statistical significance groups at P < 0.05 (one-way ANOVA followed by Tukey’s multiple comparisons test). Scale bars, 1 cm (b, d, f, h, i, o, q, r), 2 cm (k, m). Source data are provided as a Source Data file. Loss-of-susceptibility mutant of Rsf1 enhances defense responses during RRKN infection To investigate the potential resistance mechanism mediated by Rsf1 mutants, we conducted a comprehensive comparative transcriptomic analysis on WT, Rsf1-KO, and Rsf1-OE plants at key stages of nematode infection (3 dpi and 7 dpi). Analysis of differentially expressed genes (DEGs) and Gene Ontology (GO) enrichment revealed 828 DEGs in the Rsf1-KO plants, including 57 genes associated with defense responses (P-adjusted < 0.05) (Fig. 7a, b, Supplementary Fig. 11a–d and Suppl. Data 11 and Suppl. Data 12). Furthermore, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis revealed that 42 genes in plant-pathogen interactions and 43 genes in MAPK signaling pathways were significantly enriched in Rsf1-KO plants (Fig. 7c and Suppl. Data 13). Further RT-qPCR analysis revealed significantly enhanced expression of defense-related genes, including OsMAPK5 (LOC_Os03g17700), OsPR1 (LOC_Os07g03730), OsPR1b (LOC_Os01g28450), and OsPR10a (LOC_Os12g36880), in Rsf1-KO and Myb44-KO plants compared to wild-type (WT) and Rsf1-OE plants at 3 and/or 7 dpi (Fig. 7d). Concurrently, combined transcriptomic and RT-qPCR analyses also demonstrated significant upregulation of callose-deposition-related gene OsUGE3 (LOC_Os09g35800) in Rsf1-KO and Myb44-KO plants relative to WT and Rsf1-OE plants (Suppl. Fig. 11e, f). Consistently, Rsf1-KO plants exhibited a markedly higher level of callose accumulation (18.64%) compared to wild-type (NPB) plants (P < 0.0001), with an increased average callose-deposition area (Fig. 7e, f). We further investigated the involvement of GO enrichment and plant hormone signal transduction pathways in Rsf1-mediated nematode resistance. Salicylic acid (SA)-related gene OsNH1 was induced at 3 dpi, and ethylene (ET)-related genes (OsERF3 and OsACS1) were upregulated at 3 or 7 dpi in Rsf1-KO plants, suggesting activation of general defence network (Fig. 7g). These results suggest that loss-of-susceptibility mutant of Rsf1 enhances general defense responses in rice. Rsf1-KO plants show significantly higher defense responses upon RRKN infection. a Number of DEGs in wild-type (NPB, WT), Rsf1-KO, and Rsf1-OE plants at 3 and 7 dpi with RRKN. b Venn diagram displaying overlapping DEGs in WT and Rsf1-KO plants. c KEGG enrichment analysis for 828 common DEGs in Rsf1-KO plants, with P-values calculated using a one-sided hypergeometric test. d RT-qPCR analysis for defense-related genes OsMAPK5, OsPR1, OsPR1b, and OsPR10a at 3 and 7 dpi. Data are mean ± SD from three independent biological replicates. e Callose deposition area in WT, Rsf1-KO, and Rsf1-OE plants at 7dpi. Each fourteen-day-old rice seedling was inoculated with 300 J2 RRKN. Sample sizes are indicated by n = number of plants; the center line indicates the median, the box edges represent the first and third quartiles, and the whiskers extend to the minima and maxima. f Callose deposition in WT, Rsf1-KO, and Rsf1-OE plants at 7dpi. Scale bar = 200 μm. g Heatmap for expression levels of phytohormone-related genes in WT, Rsf1-KO and Rsf1-OE plants at 3 dpi and 7 dpi. The TPM values were normalized in the row direction. Red and blue colors indicate relatively higher and lower RNA levels, respectively. Different letters above the bars in (d, e) indicate statistical significance groups at P < 0.05 (one-way ANOVA followed by Tukey’s multiple comparisons test). h Rsf1 working model. Left side: In susceptible rice, the HapA-type TATA-box of Rsf1 promoter enhances Rsf1 transcription, increasing Rsf1 mRNA and protein levels to promote RRKN infection. Right side: Resistant rice accessions carry HapG-type Rsf1 promoter, which reduces Rsf1 transcription, conferring RRKN resistance. Myb44 binds to the TAACCA cis-element upstream of the TATA-box to elevate Rsf1 expression. Rsf1 protein, a cysteine-rich protein, accumulates during RRKN infection that facilitates parasitism by suppressing the general plant defence responses. Source data are provided as a Source Data file. Created in BioRender. JW, Y. (2026) https://BioRender.com/gmljrmo. Rsf1 gene function is conserved between rice and Arabidopsis thaliana Because RRKN can infect many plant species, including many monocotyledonous and dicotyledonous plants41, we investigated whether Rsf1 also serves as a functionally conserved susceptibility factor for nematode infection in dicotyledonous plant species (Suppl. Fig. 12a). We identified T-DNA insertional mutants of AtRsf1 (AT1G66100) (AtRsf1-2 and AtRsf1-7), the Arabidopsis ortholog of OsRsf1, and used them to assess the effects of AtRsf1 on RRKN infection. All Arabidopsis mutant lines showed no significant differences in growth or development compared to WT plants without RRKN inoculation (Suppl. Fig. 12b–f). RRKN inoculation results showed that AtRsf1 mutants exhibited a 53.46% reduction in root galls and a 51.48% reduction in nematode number compared to wild-type plants (P < 0.0001) (Suppl. Fig. 12g–i). These findings suggest that Rsf1 is a conserved susceptibility factor for RRKN infection in plants, and its disruption confers enhanced resistance to RRKN in both rice and Arabidopsis. Discussion Rice root-knot nematode poses a significant threat to global rice production. Changes in rice cultivation systems and global warming have accelerated nematode dissemination and expanded their suitable habitats10. Utilizing resistant crop varieties to control nematode infection is an effective and economical solution to address this problem12,42. Unfortunately, most rice cultivars are susceptible to RRKN, and few typical R genes against RRKN have been identified in rice12,43,44. Recently, modification of host S genes has emerged as an effective strategy for breeding crops with broad-spectrum disease resistance26,28,29,45,46. However, few RRKN S genes have been identified23,24. Here, we conducted a large-scale evaluation for RRKN resistance under field conditions using 345 diverse rice germplasm resources, most of which were found susceptible to RRKN. By integrating GWAS, multi-omics, and functional analyses, we identified Rsf1 as a S gene mediating rice response to RRKN. Field trials demonstrated that Rsf1 knockout markedly enhanced rice resistance to RRKN without yield penalty. Bacteria and fungi often secrete small cysteine-rich proteins (CRPs) to evade host recognition, inhibit cysteine proteases, suppress reactive oxygen species (ROS) bursts, and disable defense-related proteins, thereby attenuating plant immunity47–51. It is well-documented that nematode-secreted CRPs can suppress host immune responses52. Thionin/Thionin-like family proteins belonging to CRPs play a recognized role in plant-nematode interactions. For instance, rice THI7 confers resistance to RRKN38 and Arabidopsis Thil contributes to defense against the beet cyst nematode H. schachtii53. Despite these insights, the possibility of nematodes actively inducing host plant CRPs to promote their own parasitism remains undetermined. Interestingly, we found that RRKN infestation induces expression of the plant CRP gene Rsf1, resulting in suppression of host immunity and enhancement of nematode parasitism in both rice and Arabidopsis plants. During the long-term co-evolution between nematodes and plants, plants have evolved various ways, including transcriptional regulation54, natural genetic variations55, enhanced immune responses56, and epigenetic regulation57, to counter nematode infection. The TATA-box is a conserved sequence located at the core of eukaryotic gene promoters, and crucial for initiating transcription58. Expression levels of eukaryotic genes show a strong correlation with TATA-box sequence conservation59–61. For instance, TATA-box variations in Arabidopsis, maize, and sorghum significantly reduced transcriptional activity of the gene62. We found a natural variation (TATATAA/G, HapA/G) at the TATA-box of Rsf1 that is associated with Rsf1 expression level. HapG Rsf1 has lower expression than HapA and is associated with regions of severe RRKN infestation. HapG is predominantly found in O. sativa indica, a subspecies adapted to tropical and subtropical environments where severe RRKN infestation is prevalent63,64. However, the underlying mechanism of this association remains unclear and warrants further investigation. Loss-of susceptibility mutant of Rsf1 induced higher expression of defense-related genes, activation of the general plant defence network, and enhanced callose deposition in root galls during RRKN infection. In plant-nematode interactions, phytohormone-mediated responses are highly dynamic, with distinct hormonal pathways differentially contributing to defense or susceptibility depending on infection stage9,56. Thus, the specific contribution of plant hormones to Rsf1-mediated nematode susceptibility remains to be further elucidated. Furthermore, homologs of Rsf1 are widely present across plant species. We demonstrated that Rsf1 functions as a conserved susceptibility factor to RRKN in both rice and Arabidopsis. However, whether its role is functionally conserved in other plant species remains to be further investigated. In conclusion, RRKN induce expression of the plant CRP gene Rsf1, which facilitates their parasitism. A natural variation in TATA box in the promoter of Rsf1 enhances rice resistance to nematodes. Targeted Rsf1 gene editing markedly improves nematode resistance without yield penalties. These findings uncover a conserved mechanism underlying RRKN infection mediated by Rsf1 and highlight Rsf1 as a promising candidate for breeding nematode-resistant crop varieties. Methods Nematode collection and greenhouse assays For greenhouse assays, a pure culture of RRKN (Meloidogyne graminicola) obtained from Hanshou County (28° 54’ N, 111° 58’ E), Hunan Province, China, was maintained on the susceptible cultivar Nipponbare (Oryza sativa) in a sand-soil mixture (3:1) at 25 ~ 28 °C. RRKN second-stage juveniles (J2s) were extracted from galls, and 14-day-old seedlings were inoculated with 300 J2s per plant. Root galls and nematode numbers were assessed at 14 days post-inoculation (dpi). Three independent biological replicates were conducted. Field assays for RRKN resistance The 345 rice accessions from the C-RDP-II panel used in this study were obtained from the Guangdong Academy of Agricultural Sciences (Suppl. Data 1). Field trials were conducted at Hanshou County (28° 54’ N, 111° 58’ E), Hunan Province, China, where fields had experienced severe RRKN infestations for over five consecutive years. Thirty days after sowing, root galls were assessed in ten plants per accession, and the gall index was recorded43. GWAS for rice RRKN resistance GWAS was performed for 345 rice accessions using 673,494 high-quality SNPs (MAF ≥ 2%, missing rate < 25%) (Supplementary Data 2). Associations between SNPs and RRKN resistance were analyzed using the Genome Association and Prediction Integrated Tool (GAPIT v 4)65, implementing both the compressed mixed linear model (CMLM) and Bayesian-information and Linkage-disequilibrium Iteratively Nested Keyway (BLINK) models. The genome-wide significance threshold of GWAS (P-value = 3.2 × 10−5) was determined by the minimum Bayes factor (mBF) = -e*P*ln(P)66. Manhattan and P-P plots were generated with R package CMplot v 4.5.167. Linkage disequilibrium (LD) blocks were defined using the Solid Spine method and visualized as heatmaps with “LDBlockShow v 1.40“68. RNA isolation and RT-qPCR analysis Total RNA was extracted from ground rice galls (0.3 g) using TRIzol reagent (Invitrogen) and reverse-transcribed with a PrimeScript RT reagent kit (TaKaRa). Quantitative PCR was performed with SYBR Green Realtime PCR Master Mix (TaKaRa) on a CFX real-time PCR system (Bio-Rad). Transcript abundance was calculated using the 2−ΔCt method69. The expression level of each tested gene was normalized against two rice reference genes, OsUBQ (LOC_Os01g22490) and OsEXPNAR (LOC_Os07g02340). Each reaction included three biological replicates, with at least three technical replicates per sample. Primer sequences are listed in Suppl. Data 14. Vector construction and plant transformation To generate rice overexpression lines, the coding sequences (CDS) of Rsf1_Hap3, Thil20, Thil50, and Thil60 were amplified from Nipponbare (NPB) cDNA, and Rsf1_Hap1 from KPM cDNA70. Each CDS was cloned into the pRHV overexpression vector under the control of the maize ubiquitin promoter. Constructs were introduced into NPB background via Agrobacterium-mediated transformation, and transgenic plants were selected on thiamphenicol-containing medium. CRISPR/Cas9-mediated knockout lines of Rsf1 (Rsf1-KO) and Myb44 (Myb44-KO) in NPB were generated following the Wang method71. Target sites were designed using an online tool72; corresponding oligonucleotides were annealed and cloned into the pRHCas9 vector. Recombinant plasmids were verified by sequencing and introduced into A. tumefaciens strain EHA105 for rice transformation. T-DNA insertional mutants of AtRsf1 were obtained from the Arabidopsis T-DNA insertion mutant collection (https://www.arashare.cn/index/Product/index.html). All primers used to construct vectors are listed in Suppl. Data 14. All constructs were verified by DNA sequencing. Nematode activity assay Approximately 300 second-stage juveniles (J2s) of RRKN were suspended in each well of a 12-well culture plate. Recombinant GST-tagged Rsf1 protein lacking the signal peptide (GST-Rsf1−ΔSP) was added at defined concentrations and gently mixed. Controls included an equal amount of purified GST, glutathione elution buffer, or sterile water. Plates were sealed with parafilm and incubated at 28 °C in the dark for 48 h. Nematode activity was assessed under a stereomicroscope (Olympus SZX16, Japan), and mortality was calculated as the percentage of dead individuals relative to the total number counted. Each treatment was conducted with three independent biological replicates. Nematode attraction assay The effect of Rsf1 on RRKN migration was evaluated in a plastic container (10 × 5 × 5 cm), partitioned by a 200-mesh screen into six 1.5 cm zones (0–1.5, 1.5–3.0, and 3.0–4.5 cm on each side). Each container was filled with 150 g of a sand and synthetic absorbent resin (SPA) mixture. One-week-old seedlings of Rsf1-KO plants and WT plants were transplanted to opposite ends of the container, then 600 J2s of RRKN were applied to the center. In parallel, assays comparing Rsf1-OE versus WT seedlings were performed under identical conditions. Seven days after inoculation, nematodes were recovered from each zone at 1.5 cm intervals using the wet-sieve method73. Each treatment was conducted with at least three independent replicates. Nematode development assay The development of RRKN in rice roots was assessed in Rsf1-KO plants, Rsf1-OE plants, and WT plants. Fourteen-day-old seedlings grown in PVC tubes containing SPA substrate were inoculated with 300 J2s of RRKN. Plants were maintained at 26–28 °C under a 16 h-light/8 h-dark cycle with 70–75% relative humidity. At 14 days post-inoculation (dpi), root gall numbers, nematode numbers, and nematode developmental stages were recorded. Plant growth parameters, including seedling height and root length, were also measured. Each treatment was performed with three independent biological replicates. Callose deposition assay Fourteen-day-old seedlings of Rsf1-KO plants, Rsf1-OE plants, and WT plants were inoculated with 300 J2s of RRKN. Root galls were collected at 7 days post-inoculation (dpi) and fixed in 3:1 ethanol: acetic acid solution. After ethanol rehydration, samples were cleared with 10% sodium hydroxide following the Millet method74 with minor modifications. Galls were stained with 1% aniline blue, and callose deposition was visualized under an Olympus IX83 inverted fluorescence microscope (Olympus Optics, Tokyo, Japan) equipped with ultraviolet illumination. Callose amount was quantified using Image J 1.53c software. Each treatment was performed with three independent biological replicates. Genetic diversity and evolutionary analyses The genomic sequences of 11 wild rice and 343 cultivated rice accessions were obtained from OryzaGenome (http://viewer.shigen.info/oryzagenome/)75 and the Rice Functional Genomics and Breeding Database (http://www.rmbreeding.cn/Snp3k)76, respectively. Haplotypic diversity (h) and nucleotide diversity (π) of Rsf1 were calculated for rice subgroups using DnaSP v5.10.0177. Neutrality was tested using Tajima’s D78. Genetic differentiation among subgroups was evaluated with Ks*, Z*, and Snn79, with significance threshold of P < 0.05. Gene flow was assessed in DnaSP v5.10.01 based on the standardized variance of Fst (a measure of subpopulation differentiation) and the number of migrants (Nm). Low gene flow was inferred when |Fst| > 0.33 or |Nm|< 1, and high gene flow when |Fst|<0.33 or |Nm| > 1. Haplotype networks were constructed using PopART v1.780. Field trials for rice yield and quality Field experiments were conducted to assess the performance of WT plants, Rsf1-OE plants, and Rsf1-KO plants under RRKN infestation. Field trials were performed in Hanshou County (28° 54’ N, 111° 58’ E) under both RRKN-infested and non-infested conditions. Gall index was recorded at 30 days post-sowing with 20 plants per plot. At maturity, number of grains per plant, grain weight per plant, number of tillers per plant, and plant height were recorded. Parallel measurements were performed in non-RRKN-stress fields to assess growth and yield parameters. DNA pull-down assay DNA pull-down assays were performed using a commercial kit (BersinBio, Guangzhou, China). The −1000 bp to −1 bp promoter region of Rsf1 was amplified by PCR and biotin-labeled. Biotinylated DNA was incubated with Bioeast Mag-SA streptavidin beads (Boyue Biology) for 1 h at 4 °C, with non-biotinylated DNA serving as a negative control. Nuclear proteins were extracted from root galls or root tissues of susceptible cultivar Nipponbare (NPB) at 0, 1, or 3 days post-inoculation using the NE-PER Nuclear and Cytoplasmic Extraction kit (Thermo Fisher, 78833). Bead-bound DNA was incubated with 100 µg of nuclear protein overnight at 4 °C with gentle agitation. Proteins were separated by SDS-PAGE, and visualized by silver staining. Mass spectrometry (MS) and bioinformatics analysis Proteins pulled down by biotinylated and non-biotinylated Rsf1 promoter fragments (1000 bp) were subjected to MS analysis. A total of six samples representing three RRKN infection stages (0, 1, and 3 dpi) and two probe treatments were analysed. Pulled-down proteins were reduced, alkylated, and digested overnight with sequencing-grade trypsin (Promega), followed by analysis on an EASY-nLC 1200 system coupled to a Q Exactive mass spectrometer (Thermo Fisher Scientific). Full MS scans were acquired over an m/z range of 350–1800 at a resolution of 70,000, and the top 20 precursor ions were selected for HCD fragmentation and MS/MS analysis at a resolution of 17,500. Raw data were searched against the UniProt Oryza sativa database using MASCOT v2.3.02 with trypsin specificity allowing one missed cleavage, carbamidomethylation of cysteine as a fixed modification, oxidation of methionine as a variable modification, a precursor mass tolerance of 20 ppm, and a fragment mass tolerance of 0.6 Da. Peptide and protein identifications were filtered at a false discovery rate (FDR) of 1%. Proteins specifically enriched by the biotinylated probe relative to the control were retained for downstream GO, KEGG, and STRING analyses (P < 0.05). MS analyses were performed by Zoonbio Biotechnology (Nanjing, China). Subcellular localization assays Full-length Myb44 cDNA was cloned into the pYBA1132-eGFP vector to generate a CaMV35S:Myb44-eGFP fusion. Recombinant constructs were sequence-verified prior to use. The CaMV35S:p2300-35S-H2B-mCherry plasmid was used as a nuclear marker. Rice protoplasts were prepared following the Baba method81 and transfected with Myb44-eGFP and H2B-mCherry plasmids using PEG-mediated transformation. Transfected protoplasts were incubated for 14 h at 28 °C in W5 solution. For Rsf1 subcellular localization assay, full-length Rsf1 cDNA was fused with eGFP to generate an Rsf1-eGFP fusion construct and transiently expressed in rice protoplasts. GFP fluorescence was observed by confocal microscopy to determine subcellular localization. Phylogenetic analysis Thionin-related protein sequences, including Rsf1 from Oryza sativa japonica (NPB), were retrieved for phylogenetic analysis. A neighbor-joining tree was constructed using MEGA1282 with 1000 bootstrap replicates to assess branch support. Transient transcriptional activity assays The Myb44 coding sequence was cloned into the pYBA1143 effector vector, and the 1000 bp promoter region of Rsf1 (HapA and HapG) was cloned into the pGreenII0800-LUC reporter vector. The reporter construct carries firefly luciferase (LUC) under control of the Rsf1 promoter; Renilla luciferase (REN) driven by the CaMV35S promoter was used as an internal reference. Luciferase activity was measured using the Dual-Luciferase reporter assay system (Vazyme, China). Ratios of LUC to REN activity were calculated to determine relative promoter activity. Each experiment was performed with three independent biological replicates. For Rsf1 promoter activity assays, different shorter versions of Rsf1 promoter were inserted into the pGreenII0800-LUC vector to drive expression of LUC; Renilla luciferase was used as an internal control83. Rsf1 promoter activity was then evaluated using the same procedure as described above. Each experiment was performed with three independent biological replicates. In vitro electrophoretic mobility shift assays Electrophoretic mobility shift assays (EMSA) were performed in ref. 84. Complementary oligonucleotides with 5’-FAM labels were annealed to generate double-stranded probes. Different amounts of unlabeled probes (20-, 100-, and 200-fold over labeled probes) were incubated with FAM-labeled probes at 28 °C in binding buffer for 20 min using an EMSA kit (Thermo Fisher Scientific). Samples were resolved on 6% native polyacrylamide gels in 0.5 × TBE at 4 °C. Fluorescent signals were detected using a Typhoon 9410 imager (Amersham, USA). RNA sequencing (RNA-seq) and data analysis RNA-seq was used to compare transcriptomic profiles between WT plants and Rsf1-KO plants or between WT and Rsf1-OE plants with or without inoculation with RRKN. RNA-seq was performed with three independent biological replicates. The criteria for selecting DEGs were set as follows: P < 0.05 and an absolute log2 (fold change) ≥ 1. GO enrichment analysis and KEGG pathway analyses for DEGs were conducted using the Phyper program (http://www.geneontology.org/). Heatmaps for DEG expression patterns were constructed using TBtools v2.225 software85. Statistical analysis Data were analyzed using GraphPad Prism 10.5.0 software. Statistical significance of the results was evaluated using one-way ANOVA with Tukey’s multiple comparisons test or a two-sided Student’s t-test at a 5% probability level. 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 Supplementary Data 5 Supplementary Data 6 Supplementary Data 7 Supplementary Data 8 Supplementary Data 9 Supplementary Data 10 Supplementary Data 11 Supplementary Data 12 Supplementary Data 13 Supplementary Data 14 Reporting Summary Source data Source Data Acknowledgements We thank Dr. Mawsheng Chern and Dr. Muhammad Ayaz for extensive editing. Author contributions W.K.H. and H.X.K. supervised the work. W.K.H., H.X.K. and J.W.Y. designed the experiments and wrote the manuscript with input from all authors. J.W.Y. performed the majority of the experiments and data analysis. J.W.Y. and H.X.K. helped with the GWAS analysis. J.W.Y., C.X., H.P., L.A.K. and S.M.L. contributed to bioinformatics analysis. J.W.Y., L.Q., D.L.P. and M.J.L. contributed to disease assays. J.W.Y., X.Y.Y., W.C.W., H.H.H., Y.Q.L., Y.X.C. and X.H. were involved in the analysis of the germplasms used for screening and developed some of the inoculations used in this study. Peer review Peer review information Nature Communications thanks Tina Kyndt and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Funding This work was supported by grants from The Agricultural Science and Technology Innovation Program (ASTIP) (CAAS-ZDRW202505, W.K.H.), National Natural Science Foundation of China (32172382, W.K.H.; U24A20405, H.X.K., and 32261143468, H.X.K.), and National Key Research and Development Program of China (2024YFC2607600, W.K.H.). Data availability The data supporting the findings of this study are available in the article and its Supplementary Information files. GWAS statistics are detailed in Suppl. Data 2. Genomic DNA sequences for Rsf1 (Thionin-like gene) from the cultivars NPB and KPM, can be found in the National Genomics Data Center (NGDC) GenBank under accessions C_AA121485.1 [https://ngdc.cncb.ac.cn/genbase/search/gb/C_AA121485.1] and C_AA121486.1 [https://ngdc.cncb.ac.cn/genbase/search/gb/C_AA121486.1], respectively. The RNA-seq data have been deposited in the European Bioinformatics Institute (EBI) under accession E-MTAB-15844. The mass spectrometry proteomics data generated in this study have been deposited in the PRIDE database under accession PXD071646. The proteomics data generated in this study have been deposited in the PRIDE database under accession code PXD071707. Source data are provided with this paper. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Houxiang Kang, Email: kanghouxiang@caas.cn. Wenkun Huang, Email: wkhuang2002@163.com. Supplementary information The online version contains supplementary material available at 10.1038/s41467-026-74677-7. References Associated Data Supplementary Materials Supplementary Information Peer Review file Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 Supplementary Data 3 Supplementary Data 4 Supplementary Data 5 Supplementary Data 6 Supplementary Data 7 Supplementary Data 8 Supplementary Data 9 Supplementary Data 10 Supplementary Data 11 Supplementary Data 12 Supplementary Data 13 Supplementary Data 14 Reporting Summary Source Data Data Availability Statement The data supporting the findings of this study are available in the article and its Supplementary Information files. GWAS statistics are detailed in Suppl. Data 2. Genomic DNA sequences for Rsf1 (Thionin-like gene) from the cultivars NPB and KPM, can be found in the National Genomics Data Center (NGDC) GenBank under accessions C_AA121485.1 [https://ngdc.cncb.ac.cn/genbase/search/gb/C_AA121485.1] and C_AA121486.1 [https://ngdc.cncb.ac.cn/genbase/search/gb/C_AA121486.1], respectively. The RNA-seq data have been deposited in the European Bioinformatics Institute (EBI) under accession E-MTAB-15844. The mass spectrometry proteomics data generated in this study have been deposited in the PRIDE database under accession PXD071646. The proteomics data generated in this study have been deposited in the PRIDE database under accession code PXD071707. Source data are provided with this paper.