For all experiments, we used Oryza sativa var. japonica, cultivar “Dongjin” to generate transgenic lines and act as a WT control. Seeds were germinated on the ½ Murashige and Skoog (MS) medium, supplemented with 3% sucrose and 0.3% agar, and seedlings were maintained in the medium until 10 days after germination (DAG). The plants were then transferred to growth rooms with controlled conditions (14-h-light at 28°C/10-h-dark at 22°C, with relative humidity maintained at 50% –70%), greenhouses, or the paddy field, depending on the experimental requirements.

We generated oswrky26 mutants by clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated nuclease 9 (Cas9)-mediated gene editing. To do so, we designed two single guide RNA (sgRNA) targets using two web-based tools: http://crispr.dbcls.jp and http://crispor.org (Naito et al., 2015; Concordet and Haeussler, 2018). The structures of the RNA scaffolds of the target sequences were verified using the bioinformatic tool (http://rna.tbi.univie.ac.at/cgi-bin/RNAWebSuite/RNAfold.cgi). The specificity of target sequences was confirmed using CRISPR RGEN Tools (http://www.rgenome.net/cas-offinder/).

For transformation, synthesized and annealed sgRNA sequences were first ligated into the BsaI-digested pRGEB32 vector (Xie et al., 2015). The resulting plasmids were then transformed into Agrobacterium tumefaciens strain LBA4404 using the freeze-thaw method (An, 1987). These were subsequently transformed into rice embryonic calli according to a previously published method (Lee et al., 1999). Transgenic T0 plants were then selected on ½ MS agar medium including 50 mg L⁻¹ hygromycin. After selection, genomic DNA was extracted from transformed plants, and the target sequence was PCR amplified using primers flanking the target site. Finally, the resulting PCR products were subcloned into T-easy vectors, and five colonies from each line were sequenced to validate sequence identities.

For the overexpression of OsWRKY26 in transgenic plants, the full-length cDNA of OsWRKY26 was amplified by PCR using primers ( Supplementary Table 1 ) with leaf cDNA as a template and inserted into the binary vector pGA3438 (Kim et al., 2009) using the 5x In-fusion HD enzyme (STO345, Takara). The resulting construct was then introduced into A. tumefaciens LBA4404 and transformed into rice. As a control, transgenic plants expressing the Myc tag alone were also generated. To confirm the expression of the target gene, total protein was extracted from the leaf blades of six transgenic T1 lines at 50 DAG, and exogenous OsWRKY26 was detected by western blotting using a Myc antibody (9B11 with HRP conjugate; Cell Signaling Technology). Additionally, OsWRKY26 transcript levels were analyzed by qRT-PCR, with OsUbi5 as the reference gene.

The OsWRKY26 promoter:β-glucuronidase (GUS) vector (pWRKY26:GUS) was created by insertion of a fragment 3,000 bp upstream of the start codon into a pGA3519 vector (Yoon et al., 2014) through the KpnI and HpaI restriction enzymes sites. pWRKY26:GUS transgenic plants were then stained using GUS solution (50 mM sodium phosphate pH 7.0, 1 mM potassium ferricyanide, 1 mM potassium ferrocyanide, 0.1% triton X-100, 10 mM EDTA, pH 8.0, 0.1% X-Glu dissolved in 1% DMSO and 5% methanol). Plants showing a strong GUS signal were selected for further analysis.

RNA was isolated, and transcript levels were quantified following the method as previously reported (Tun et al., 2023). Transcript levels of specific genes were examined at different time points; these levels were normalized to those of rice Ubiquitin 5 (OsUbi5, LOC_Os01g22490) (Jain et al., 2006), which was used as an internal control. At least three biological replicates were performed for each time point. Relative expression levels were calculated using the ΔΔCt method; primer specificity was verified by observing a single sharp peak in a melting curve analysis (Schmittgen and Livak, 2008; Cho et al., 2016). The primers used in this study are listed in Supplementary Table 1 .

M. oryzae inoculation was conducted as a previously described method (Tun et al., 2023). Briefly, M. oryzae isolate RO1-1 was first grown on V8 juice agar plates (80 mL L^−l V8 Juice [Campbell’s Soup Company, Camden, NJ], 15 g L⁻¹ agar, pH 6.8) for 2 weeks under continuous fluorescent light. Conidia were collected, examined, and quantified under a microscope, then suspended in water to reach a final concentration of 5 x 10⁶ mL⁻¹. Leaves from 50 DAG plants were used for spot inoculation (Kanzaki et al., 2002). Particularly, 3 μl of conidia suspension was applied to 2 mm press-injured spots on the leaves. After inoculation, the infected plants were placed in a closed box to maintain 100% relative humidity at 25°C in the dark for 24 h before being transferred to a moist incubator. Infected leaves were harvested at 9 days post inoculation (DPI) for lesion length quantification and photographed.

The Xoo strain PXO99 was used for bacterial infection analyses. Xoo inoculation on the leaves were carried out using a previously described leaf clipping method (Ke et al., 2017). For the calculation of colony-forming units (CFUs), leaf fragments (6 cm in length) from below the Xoo lesion area were collected at 12 DPI. CFUs were calculated according to the previous method (Ke et al., 2017). Lesion length was measured at 14 DPI to assess the degree of disease symptoms.

The second leaves from WT and OsWRKY26 mutant plants at 50 DAG were used for transcriptome analyses. Total RNA was extracted in triplicates from each line and subjected to RNA sequencing. Sequencing data quality was assessed using FastQC v0.11.9 (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). To obtain high-quality clean reads, low-quality bases from both 5′ and 3′ ends of the paired-end reads were trimmed using Trimmomatic v0.39 (Bolger et al., 2014). The rice reference genome was downloaded from the Rice Genome Annotation Project database (http://rice.plantbiology.msu.edu/pub/data/Eukaryotic_Projects/o_sativa/annotation_dbs/pseudomolecules/version_7.0/all.dir/) (Kawahara et al., 2013). Index construction for the reference genome and alignment of the cleaned reads to the reference genome were both performed using Hisat2 v2.2.1 using the default parameters (Kim et al., 2015). Aligned SAM files were then converted to BAM format and sorted using SAMtools v1.10 (Danecek et al., 2021). Raw read counts were generated from the alignment files with featureCounts v2.0.0, which is part of the Subread package v1.6.2 (Liao et al., 2014). Post-normalizations of read counts were calculated as Fragments Per Kilobase of transcript per Million mapped reads (FPKM) to facilitate the estimation of gene expression levels of each sample. Differentially expressed genes were then identified using the DEseq2 package v1.38.3 (Love et al., 2014), using an adjusted p-value threshold of 0.05 and an absolute log₂ fold change ≥1. Functional annotations of genes were obtained by performing a BLASTP search against the NCBI Nr protein database with an E-value cutoff of 1.0e-8. Using the MSU7 functional annotation files, description, Gene Ontology (GO) terms, and Pfam IDs were assigned to each gene. GO enrichment analysis of all upregulated genes in the biological process category was carried out using the BiNGO tool, employing a hypergeometric test and Benjamini & Hochberg False Discovery Rate correction to obtain an adjusted p-value cutoff of < 0.05 (Maere et al., 2005).

The cDNA sequence of the OsWRKY26 coding region was inserted into the BamHI and KpnI-digested pGA3452 vector containing the ZmUbi1 promoter, a green fluorescence protein (GFP) coding region, and the nos terminator (Kim et al., 2009). The resulting fusion construct was then co-transformed with a vector expressing a nuclear localization signal (NLS)-red fluorescence protein (RFP) conjugate under the control of the ZmUbi1 promoter (NLS-RFP) into protoplasts isolated from rice Oc cells (Baba et al., 1986). Fluorescence signals were observed using a confocal microscope (K1-Fluo, Nanoscope, Korea) 14 h after transformation. GFP and RFP were excited at 488 nm and 561 nm, respectively, and their emission signals were collected using 525–550 nm and 561 nm long-pass filters, respectively.

The full-length coding sequence of OsWRKY26 was inserted between the ZmUbi1 promoter and the HA tag through the HpaI and KpnI restriction enzymes into pGA3698 (Kim et al., 2009) using In-fusion HD enzymes. The resulting vector was used as an effector construct. For luciferase reporter assays, different lengths of the OsXa39 promoter sequence were subcloned into a pGL4.23 vector that consisted of a minimal promoter and the Luciferase coding sequence (E8411, Promega). To normalize luciferase activity, the ZmUbi : GUS vector (Cho et al., 2008) was used as an internal control.

Chromatin immunoprecipitation (ChIP) experiments were performed as a previously described protocol (Haring et al., 2007). Briefly, fresh leaves from OsWRKY26-OX-9 and Myc tag alone transgenic plants were harvested at 10 DAG. Proteins and chromatin complexes were cross-linked with 1% formaldehyde in solution A (10 mM Tris pH 8.0, 400 mM sucrose, 0.1 mM PMSF, and 5 mM β-mercaptoethanol). After 10 min of vacuum infiltration, 125 mM glycine was added to quench the reaction. Following nuclei isolation, extracted DNA was sonicated to fragments of approximately 200-1,000 bp in length. Immunoprecipitation was performed using an anti-Myc antibody (9B11, cell signaling) and agarose beads conjugated with protein G and A (EMD Millipore). Washing and reverse cross-linking steps were then carried out according to the protocol of Haring et al. (2007). The enriched chromatin was analyzed by PCR using the primer sets listed in Supplementary Table 1 .

We previously reported that the expression of OsWRKY26 in rice Oc cells was induced by supplementation with sucrose within 2 h, reached its highest level after 4 h, and then rapidly declined (Tun et al., 2023). To further characterize the gene, its expression pattern was studied during plant development. The transcript level of OsWRKY26 gradually increased during the vegetative growth stage and reached the maximum level at 50 DAG ( Figure 1A ). The gene remained highly expressed until 70 DAG, after which transcript levels dropped rapidly at the heading stage (80-90 DAG).

To investigate the tissue-specific expression pattern of OsWRKY26, a 3,000 bp sequence upstream of the start codon of OsWRKY26 was cloned to a GUS reporter vector, and the resulting pOsWRKY26:GUS vector was transformed into rice plants. GUS staining of tissues from transgenic plants at various developmental stages showed that this signal was detected mainly in leaf vascular bundles ( Figures 1B–E ). In addition, weak GUS staining was found in spikelets ( Supplementary Figure 1A ). However, the staining was not detected in roots ( Supplementary Figures 1B, C ).

To investigate the functional role of OsWRKY26, null mutants of the gene were generated by the CRISPR/Cas9 methods. Among several transgenic plants, two mutant lines were selected for further study ( Figure 2A ). In the first mutant line, oswrky26-1 (wr26-1), 4 nucleotides were deleted along with a 12-nucleotide substitution at the sgRNA binding site. The second mutant line, oswrky26-2 (wr26-2), carried biallelic mutations: a five base pair deletion in the target site on one chromosome and a one base pair deletion on another chromosome. Both mutant lines were found to grow and show normal head morphology ( Figure 2B ). However, plant height and grain yield were slightly reduced in mutant plants compared to the WT control ( Supplementary Figure 2 and Supplementary Table 2 ).

It has been previously reported that OsWRKY26 expression can be induced by Xoo and M. oryzae infection (Choi et al., 2017; Li et al., 2021). We therefore examined whether these mutant lines showed altered responses to Xoo infection. Bacterial inoculation of leaves during the juvenile stage (30 DAG) resulted in slightly reduced lesion lengths in mutant lines compared to WT ( Supplementary Figures 3A, B ), with no significant differences observed in heading stage plants (90 DAG) ( Supplementary Figures 3C, D ). However, significant levels of resistance were found in mutant plants at the adult vegetative stage (50 DAG), as evidenced by shorter lesion lengths ( Figures 2C, D ). The bacterial populations were also reduced in the lesions of wr26 lines compared to the WT ( Figure 2E ). Taken together, these results suggest that enhanced resistance to Xoo in the mutants was observed when OsWRKY26 expression was highest. Notably, no changes in resistance were observed at the heading stage, when OsWRKY26 expression was low, suggesting that its impact on resistance likely depends on expression levels.

To determine whether OsWRKY26 also responded to M. oryzae infection, we inoculated fungal spores on the second leaf blades of WT, wr26-1, and wr26-2 plants at 50 DAG. Lesion measurements nine days after infection showed similar lesion lengths between WT and mutant plants, suggesting that OsWRKY26 does not play a significant role in defense against M. oryzae ( Supplementary Figures 4A, B ).

OsWRKY26 belongs to the WRKY family of TFs, which are known to regulate responses to biotic and abiotic stress, senescence, and various developmental processes (Li et al., 2021; Wang et al., 2023; Javed and Gao, 2023). To confirm the subcellular localization of OsWRKY26, GFP signals from OsWRKY26-GFP and RFP from a nucleus localization signal (NLS-RFP) marker were tracked in protoplasts derived from Oc cells. As a control, the NLS was connected to RFP under the ZmUbi1 promoter (NLS-RFP). Both constructs were co-introduced into rice protoplasts and transiently expressed for 14 h. Visualization of the expressed proteins under a fluorescence microscope showed that the OsWRKY26-GFP protein co-localized with NLS-RFP within the nucleus, indicating that OsWRKY26 is a nuclear-localized protein ( Supplementary Figure 5 ).

To investigate the potential targets of OsWRKY26, we performed transcriptome analyses of wr26-1 and WT leaves at the adult vegetative stage. Analysis of three biological replicates revealed 1,298 differentially expressed transcripts between mutant and WT plants. Among them, 777 transcripts were upregulated and 521 were downregulated in the mutant plants ( Figure 3A , Supplementary Table 3 ). GO analysis revealed that stress- and stimulus-related transcripts were among the most abundantly induced categories in the mutants ( Figure 3B , Supplementary Table 4 ). Based on their annotated functions, we focused on five groups: Group 1 - NLR genes, Group 2 - PR genes, Group 3 - WRKY genes, Group 4 - hormone-related genes, and Group 5 - sugar allocation genes ( Figure 3C ).

Since wr26 mutant plants showed enhanced resistance to bacterial pathogen infection, we chose eight genes from Group 1 and Group 2 that were related to disease resistance for further analysis. The expression levels of several NLR genes, OsXa39, OsXa47, OsBBR1, and OsRSR1, and PR genes, OsPR1a, OsPR1b, OsPR2, and OsPR4c, were significantly higher in wr26 mutant leaves compared to the WT ( Figures 4A–H ). In contrast, RNA-seq data revealed that transcript levels of genes involved in sugar allocation were significantly lower in mutant plants compared to the WT ( Figures 4I–L ). These results indicate that OsWRKY26 plays a significant role in regulating defense-related genes, thereby enhancing resistance to Xoo infection, while also affecting sugar metabolism pathways.

As NLR genes are preferentially respond to Xoo strains (Lu et al., 2022), we focused on the Group I from Figure 1C . Because OsXa39 confers broad-spectrum resistance to various Xoo strains compared with other NLR genes (Shi et al., 2022; Zhang et al., 2015a, 2015b; Zhang et al., 2015c), we investigated whether this gene is a direct target of OsWRKY26 by employing a transient expression assay. The OsXa39 promoter region (−2,572 to −63 bp upstream of the start ATG codon) was fused to the Luciferase gene ( Figure 5A ). As a reference, the GUS reporter was driven by the ZmUbi1 promoter. Both constructs were co-transformed into rice protoplasts along with an OsWRKY26-HA fusion protein-expressing plasmid ( Figure 5A ). As a control, a plasmid expressing only the HA tag was also introduced. After 15 h of incubation, the luciferase and GUS activities were measured. The results showed that luciferase activity driven by the OsXa39 promoter was significantly reduced when co-expressed with OsWRKY26 compared to the HA alone ( Figure 5B ). As a further control, we introduced OsWRKY7, a gene involved in resistance to both M. oryzae and Xoo, and found that this did not affect OsXa39 expression ( Figure 5B ). These results indicated that OsXa39 promoter activity was specifically repressed by OsWRKY26.

To locate the region responsible for the repression, two constructs were generated: one containing an upstream region (−2,572 to −1,022 bp) and the other containing a downstream region (−818 to −63 bp) ( Figure 5C ). These regions were connected to the Luciferase gene and introduced into rice protoplasts along with OsWRKY26-HA or with only the HA tag. Measurement of luciferase activity showed that expression of Luciferase driven by the upstream region was not inhibited by OsWRKY26 ( Figure 5D ). However, the marker expression linked to the downstream region was significantly suppressed by OsWRKY26 ( Figure 5D ). To further narrow down the repressor region, two additional fragments (−531 to −63 bp and −183 to −63 bp) were tested ( Figure 5C ). Transient assay results revealed that the longer fragment retained the repressor activity, whereas the shorter fragment did not ( Figure 5D ). Therefore, the OsWRKY26 repressor region is likely located between −531 and −183 bp.

OsWRKY26 may bind to and interfere with an unspecified positive regulatory transcription factor that induces expression of OsXa39. To find potential TFs interacting with OsWRKY26, we conducted a yeast two-hybrid screen, which identified 19 interacting proteins, including E3 ligases, enzymes, hormone-related genes, and other proteins of unknown function. However, no TFs known to be involved in Xoo resistance were identified ( Supplementary Table 5 ). These results suggest that OsWRKY26 may directly bind to the promoter region of OsXa39 and repress its expression.

To confirm whether OsWRKY26 directly binds to the OsXa39 promoter, we performed ChIP-PCR analysis. Overexpression lines of OsWRKY26 were generated using a construct containing the OsWRKY26 coding region between ZmUbi1 promoter and a Myc tag ( Supplementary Figure 6A ). Leaf samples from six transgenic plants were harvested to measure OsWRKY26-Myc protein levels by western blot analysis ( Supplementary Figure 6B ). Three transgenic plants expressed the fusion protein at high levels, and the transcript levels of lines #9 and #10 were further confirmed ( Supplementary Figure 6C ). These transgenic plants did not show altered responses to Xoo ( Supplementary Figures 6D, E ).

Transgenic plants expressing OsWRKY26 at high levels (OsWRKY26-OX-9) and those expressing only the Myc tag were used for ChIP-PCR. Chromatins from transgenic nuclei were sonicated and immunoprecipitated with Myc antibody. Precipitated DNA was amplified with fifteen overlapping primer sets spanning 96-397 bp of the promoter region of OsXa39 ( Figure 6A ). In these tests, OsUbi5 was used as a negative control to test for non-specific binding of the antibody. The analysis showed that three regions, P12 (−680 to −531 bp), P13 (−554 to −324 bp), and P14 (−348 to −191 bp), were enriched with OsWRKY26, indicating that OsWRKY26 binds directly to the OsXa39 promoter region between −680 and −191 bp upstream of the start codon ( Figure 6B ). We also noted that this region overlapped with the OsWRKY26 repressor region (−531 and −183 bp) identified by the transient expression assay using the Luciferase marker. These results highlight the repressive activity of OsWRKY26 on the OsXa39 promoter.