Rice transcription factor bHLH25 confers resistance to multiple diseases by sensing H2O2

Introduction

Hydrogen peroxide (H₂O₂), a main reactive oxygen species (ROS), plays central roles in many biological processes in eukaryotes.^1,2,3 Plants respond to pathogen infection with a burst of oxidants including superoxide anions (O₂^–) in the apoplastic spaces produced by nicotinamide adenine dinucleotide phosphate (NADPH) oxidases to inhibit pathogen growth.^4,5,6 These extracellular oxidants are converted by peroxidases to H₂O₂ as the final product, which is transported into the host cell to trigger immune responses.^3,7,8 The role of H₂O₂ in immunity was first reported in 1974 for animals⁹ and in 1983 for plants.¹⁰ Cells sense H₂O₂ and transduce the signal mainly through H₂O₂-mediated oxidative post-translational modifications of sulfur-containing amino acids cysteine and methionine in proteins at different subcellular locations.^3,11,12

It has taken over four decades to find proteins other than peroxidases that directly sense H₂O₂ to control innate immunity. Master immune regulator non-expressor of pathogenesis-related 1 (NPR1) was recently reported to function as a sensor of redox rhythm in Arabidopsis to gate plant immune response towards the morning and minimize costs on growth at night.¹³ Redox sensor Quiescin sulfhydryl oxidase homolog 1 (QSOX1) regulates ROS levels in Arabidopsis to modulate plant immunity.¹⁴ However, whether NPR1 and QSOX1 directly sense and become oxidized by H₂O₂ remains undetermined. Interestingly, leucine-rich-repeat receptor kinase hydrogen-peroxide-induced Ca²⁺ increases 1 (HPCA1) protein directly recognizes H₂O₂ in Arabidopsis,¹⁵ but whether HPCA1 regulates innate immunity is undetermined. However, our current knowledge about transcription factors which directly sense H₂O₂ in the nucleus and trigger global gene expression remains quite limited. It is worth noticing that H₂O₂ can be transported into the nucleus to regulate immune responses.¹⁶ Thus, a clear gap remains in the identification of transcription factors that directly sense H₂O₂ to regulate immunity and in the understanding of their underlying mechanisms.

Results

H₂O₂ promotes immunity partially through OsLAC7/28/29-mediated lignin biosynthesis

We treated rice plants with exogenous H₂O₂ on roots and detected the accumulation of H₂O₂ content and OsPR1b and OsPR10b mRNAs in rice leaves (Supplementary information, Fig. S1a–d). Plants treated with H₂O₂ showed enhanced resistance to Magnaporthe oryzae, the causal pathogen of the destructive blast disease (Fig. 1a), confirming that H₂O₂ promotes plant immunity. We then performed global transcriptomic analysis via RNA-sequencing and identified 1596 genes up-regulated by H₂O₂ treatment (Supplementary information, Fig. S1e and Table S1). Gene ontology analysis revealed that these up-regulated genes fall into different biological processes (Supplementary information, Fig. S1f and Table S2), with carbohydrate metabolism, response to oxidative stress, and cell wall biogenesis as the most significantly induced processes (Supplementary information, Fig. S1f).

**Fig. 1: H₂O₂ promotes *OsLAC7/28/29*-mediated lignin biosynthesis and disease resistance by repressing *miR397b* expression.**

Cell wall is the first physical barrier that plants use to defend against pathogen penetration at early infection.¹⁷ We therefore screened for most highly up-regulated genes related to cell wall biogenesis and found five genes involved in pectin biosynthesis, five in xyloglucan biosynthesis, one in cellulose biosynthesis, and three in lignin biosynthesis (Supplementary information, Table S2). Because we found lignin accumulation after H₂O₂ treatment (Supplementary information, Fig. S1g) and lignin is a key molecule that strengthens cell wall to enhance disease resistance,^18,19 we focused on the three lignin biosynthesis genes, OsLAC7, OsLAC28 and OsLAC29 (abbreviated as OsLAC7/28/29), which encode laccases that polymerize monolignols to form lignin polymers for cell wall reinforcement.^20,21 We then validated the induction of OsLAC7/28/29 expression by H₂O₂ in rice by reverse transcription quantitative PCR (RT-qPCR) (Fig. 1b).

We then determined the role of OsLAC7/28/29 in plant disease resistance. Single knockout (KO) plants for any of OsLAC7/28/29 and double KO plants of OsLAC7/29 showed no changes in disease resistance or lignin contents (Supplementary information, Figs. S2, S3 and S4a–e). Most importantly, triple KO plants of OsLAC7/28/29 showed decreased disease resistance and lower lignin contents (Fig. 1c, d; Supplementary information, Fig. S4f–k). Plants overexpressing anyone of OsLAC7/28/29 accumulated more lignin contents and stronger disease resistance than Kitaake plants (Supplementary information, Figs. S4c–e, l–n). Besides, overexpression plants displayed thicker sclerenchyma cells and cell walls than Kitaake plants (Supplementary information, Fig. S4o, p). These results suggest that OsLAC7/28/29 positively regulate lignin biosynthesis and disease resistance in a functionally redundant manner. More importantly, H₂O₂-induced disease resistance was significantly weakened in Oslac7/28/29 triple KO plants compared to the wild-type Kitaake plants (Fig. 1e), indicating that H₂O₂ induces disease resistance partially through OsLAC7/28/29.

OsLAC7/28/29 are targeted and repressed by miR397b

We next examined OsLAC7/28/29 promoters for response to H₂O₂. Although OsLAC7/28/29 promoters are enriched with cis-elements for TATA-box-binding proteins (TBPs) and CAAT-box-binding factors (CBFs) (Supplementary information, Fig. S5a and Table S3), neither rice TBP nor CBF genes were induced or repressed by H₂O₂ (Supplementary information, Fig. S5b, c and Table S4). We then cloned 18 TBPs and CBFs and performed transactivation assays to determine their transcriptional regulatory activity on OsLAC7/28/29 promoters. The results showed that none of these 18 TBP or CBF transcription factors simultaneously activated OsLAC7/28/29 promoters (Supplementary information, Table S5), indicating that some factor(s) other than TBPs and CBFs promote OsLAC7/28/29 expression upon H₂O₂ induction. As microRNAs (miRNAs) are well-known for regulating gene expression,²² we screened for potential miRNA-binding sites (MBS) and found that only one miR397b-binding site was commonly shared by OsLAC7/28/29 (Supplementary information, Fig. S5d, e and Table S6). Interestingly, among all genes induced by H₂O₂, only OsLAC7/28/29 mRNAs contain miR397b MBS (Supplementary information, Table S6). These results suggest that miR397b may specifically target OsLAC7/28/29.

In transactivation assays, Nicotiana benthamiana leaves co-expressing 35S:miR397b and 35S:MBS^LACs-YFP (yellow fluorescent protein sequence tagged with miR397b MBS) showed lower YFP RNA levels than those co-expressing 35S:miR397b and 35S:mMBS^LACs-YFP (YFP fused with mutated MBS abolishing miR397b recognition) (Supplementary information, Fig. S5f, g). Nevertheless, 35S:MBS^LACs-YFP alone and 35S:mMBS^LACs-YFP alone expressed similar YFP RNA levels (Supplementary information, Fig. S5g). RT-qPCR assays showed that miR397b-KO plants expressed higher OsLAC7/28/29 levels, whereas miR397b overexpression (miR397b-OE) plants expressed lower OsLAC7/28/29 levels than Kitaake plants (Fig. 1f; Supplementary information, Fig. S5h–j). In addition, we found that H₂O₂ represses miR397b expression (Fig. 1g) and that miR397b-KO blocks the induction of OsLAC7/28/29 expression by H₂O₂ (Supplementary information, Fig. S6a). These results indicate that H₂O₂ represses miR397b expression to increase OsLAC7/28/29 expression.

When inoculated with M. oryzae, miR397b-KO plants developed smaller lesions and harbored less M. oryzae than Kitaake plants, whereas miR397b-OE plants responded oppositely (Fig. 1h; Supplementary information, Fig. S6b, c). miR397b-KO plants displayed thicker sclerenchyma cell walls with more lignin; miR397b-OE plants showed the opposite (Supplementary information, Fig. S6d–f). These results suggest that the miR397b-laccases module regulates disease resistance by controlling lignin accumulation and cell wall thickness. Additionally, Oslac7/28/29-KO/miR397b-KO plants and Oslac7/28/29-KO plants displayed the same levels of blast resistance, both lower than the wild-type Kitaake and miR397b-KO plants (Fig. 1i), indicating that the effects of miR397b in disease resistance completely rely on the suppression of OsLAC7/28/29.

bHLH25 represses miR397b to enhance lignin biosynthesis and disease resistance

To search for transcription factors that regulate miR397b expression in response to H₂O₂, we identified various cis-elements in miR397b promoter (pmiR397b) (Supplementary information, Table S3) which makes it difficult to identify relevant transcription factors. Therefore, we used yeast one-hybrid to screen a rice cDNA library with pmiR397b as the bait and identified a basic/helix-loop-helix transcription factor, bHLH25, which is localized in rice nucleus and able to bind to pmiR397b in yeast (Fig. 2a, b). DNA affinity purification coupled with quantitative PCR (DAP-qPCR) and chromatin immunoprecipitation coupled with qPCR (ChIP-qPCR) using bhlh25-KO plants overexpressing bHLH25 showed that bHLH25 is enriched in rice chromatin containing pmiR397b with a G-box-like-2 motif (Fig. 2c–e; Supplementary information, Fig. S7a). Electrophoretic mobility shift assay (EMSA) showed that bHLH25 directly bound to biotin-labeled G-box-like-2 motif in pmiR397b (bio-pmiR397b-G-box), and un-labeled competitors carrying wild-type G-box-like-2 motif (com-pmiR397b-G-box) markedly reduced the binding of bHLH25 to G-box-like-2 motif, whereas competitors with mutant G-box-like-2 motif (mut-pmiR397b-G-box) carried significantly reduced competing ability (Fig. 2f), indicating that bHLH25 binds to pmiR397b mainly through the G-box-like-2 motif, but also partially through nucleotides flanking the G-box-like-2 motif. These results are also consistent with previous reports showing that nucleotides flanking the core cis-elements on promoters are also important for the binding specificity and affinity of bHLH-type transcription factors.^23,24

**Fig. 2: bHLH25 represses *miR397b* expression to enhance lignin biosynthesis and disease resistance.**

We further performed a transactivation assay and found that bHLH25 overexpression (bHLH25-OE) inhibited the RNA level of YFP driven by pmiR397b in N. benthamiana (Supplementary information, Fig. S7b). RT-qPCR assays showed that miR397b expression decreased in bHLH25-OE rice transgenic plants, but increased in bhlh25-KO plants compared with that in Kitaake (Fig. 2g; Supplementary information, Fig. S7c–h). In contrast, OsLAC7/28/29 mRNA levels increased in bHLH25-OE plants but decreased in bhlh25-KO plants (Supplementary information, Fig. S7i, j). Moreover, we found that bHLH25 KO abolished the H₂O₂-mediated repression of miR397b expression (Supplementary information, Fig. S7k). These results suggest that the repression of miR397b expression by H₂O₂ is dependent on bHLH25.

We next determined whether bHLH25 regulates disease resistance. bhlh25-KO plants exhibited larger lesions and harbored more fungus than Kitaake, whereas bHLH25-OE plants responded oppositely (Fig. 2h; Supplementary information, Fig. S7l). In agreement, bHLH25-OE plants showed more lignin and thicker sclerenchyma cell walls than Kitaake plants, whereas bhlh25-KO plants showed the opposite (Fig. 2i, j; Supplementary information, Fig. S7m). Taken together, these results indicate that bHLH25 promotes lignin biosynthesis, cell wall reinforcement and disease resistance by repressing miR397b expression. In addition, the disease resistance level of miR397b-OE/bHLH25-OE plants was much higher than that of miR397b-OE plants, but significantly lower than that of bHLH25-OE (Fig. 2k), suggesting that the bHLH25-conferred disease resistance is partially mediated by miR397b.

bHLH25 promotes CPS2-mediated phytoalexin biosynthesis to enhance disease resistance

Interestingly, bHLH25-OE plants also exhibited a cell death phenotype on leaves in the absence of pathogen attack (Fig. 3a), which was not observed in either miR397b-KO or OsLAC7-OE plants. We speculated that bHLH25 likely regulates other defense pathway(s) independent of the miR397b-laccases pathway. A previous study showed that bHLH25 putatively bound to the N-box-like motif in the CPS2 promoter (pCPS2).²⁵ CPS2 produces phytocassanes,²⁶ a type of antimicrobial phytoalexins, which attracted us because of their ability to directly inhibit M. oryzae spore germination and fungal hypha growth and to trigger cell death.^27,28,29 Thus, we further examined whether bHLH25 regulates CPS2 expression.

Fig. 3: bHLH25 promotes *CPS2*-mediated phytoalexin biosynthesis and disease resistance in rice in a manner independent of *miR397b.*

EMSA validated binding of bHLH25 to the N-box-like motif in pCPS2 (Fig. 3b). Transactivation assays revealed that bHLH25 enhanced the RNA level of YFP driven by pCPS2 in N. benthamiana (Supplementary information, Fig. S8a). Consistently, CPS2 expression was elevated in bHLH25-OE plants but deeply repressed in bhlh25-KO plants (Fig. 3c; Supplementary information, Fig. S8b). Besides, the contents of phytocassane C were higher in CPS2 overexpression (CPS2-OE) and bHLH25-OE plants than in Kitaake plants, but undetectable in cps2-KO and bhlh25-KO plants (Fig. 3d, e; Supplementary information, Fig. S8c–f). Importantly, CPS2-OE plants exhibited a cell death phenotype similar to that observed in bHLH25-OE plants (Fig. 3f). When challenged with M. oryzae, CPS2-OE plants developed smaller disease lesions and harbored less fungal biomass than Kitaake plants, whereas cps2-KO plants responded oppositely (Fig. 3g; Supplementary information, Fig. S8g). Moreover, the disease resistance level of cps2-KO/bHLH25-OE plants was lower than that of bHLH25-OE plants (Fig. 3h). Therefore, bHLH25 promotes CPS2-mediated phytoalexin biosynthesis to enhance disease resistance.

Interestingly, lignin content remained unchanged in cps2-KO and CPS2-OE plants (Fig. 3i), in which phytoalexin contents were altered. Similarly, phytoalexin content remained unchanged in miR397b-KO and miR397b-OE plants, in which lignin contents were altered (Fig. 3j). Thus, bHLH25 promotes disease resistance through two independent defense pathways: miR397b-mediated lignin accumulation and CPS2-mediated phytoalexin accumulation.

bHLH25 is directly oxidized at methionine residues by H₂O₂

Our previous study showed that post-translational modification in a transcription factor serves as a functional switch to control two independent biological processes.³⁰ We therefore performed mass spectrometry analysis on potential post-translational modifications of bHLH25. The results showed that the oxidation level of bHLH25 incubated with extracts of M. oryzae-inoculated leaves increased when compared to bHLH25 oxidation level incubated with extracts of mock-treated leaves (Supplementary information, Fig. S9a and Table S7). This oxidation occurred at methionine residues 213, 256, 392, and 418, but not on cysteines, of bHLH25 (Fig. 4a, b).

**Fig. 4: bHLH25 directly senses H₂O₂ to regulate *miR397b*– and *CPS2*-mediated defense pathways in rice.**

In addition, we also directly treated bHLH25 protein with H₂O₂ in vitro and examined by mass spectrometry analysis the amino acids that became oxidized by H₂O₂. The results showed that the oxidation levels at M213/M256/M392/M418 were significantly higher upon direct exposure to H₂O₂ than mock (Supplementary information, Fig. S9b and Table S7). These results suggest that bHLH25 is directly oxidized by H₂O₂ and upon M. oryzae infection.

Oxidized bHLH25 represses miR397b expression and non-oxidized bHLH25 activates CPS2 expression

DAP-qPCR and EMSA analyses showed that H₂O₂ enhanced the binding ability of bHLH25 to the G-box-like-2 motif of pmiR397b, but reduced its binding ability to the N-box-like motif of pCPS2 (Fig. 4c; Supplementary information, Fig. S9c). Transactivation analysis in N. benthamiana showed that H₂O₂ treatment increased the ability of bHLH25 to repress pmiR397b expression, while decreasing the ability of bHLH25 to promote pCPS2 expression (Fig. 4d). These results indicate that oxidized bHLH25 prefers binding to pmiR397b to inhibit miR397b expression, while non-oxidized bHLH25 prefers binding to pCPS2 to promote CPS2 expression.

During M. oryzae infection, H₂O₂ was induced at 6–12 hpi (Fig. 4e) and miR397b repression and OsLAC7/28/29 induction occurred at 12–60 hpi in rice plants (Fig. 4f, g; Supplementary information, Fig. S9d, e), which boosted lignin accumulation simultaneously (Fig. 4h). H₂O₂ content gradually decreased after 24 hpi due to the activation of lignin biosynthesis, which is supported by our results that H₂O₂ levels dropped in miR397b-KO and OsLAC7-OE plants (Supplementary information, Fig. S9f, g) and by a previous report showing that laccases-mediated lignin polymerization consumes H₂O₂.³¹ Consistently, CPS2 expression increased mostly at 60–96 hpi with M. oryzae (Fig. 4i), which boosted phytoalexin accumulation (Fig. 4j).

bHLH25 and miR397b are required for H₂O₂-induced disease resistance

Further analyses showed that lignin and phytoalexin induction by M. oryzae infection in bhlh25-KO plants was obviously delayed and diminished compared to that in Kitaake plants (Supplementary information, Fig. S10a, b). While resistance to M. oryzae infection in Kitaake was clearly enhanced by H₂O₂ treatment, this enhancement was weakened in bhlh25-KO plants, indicating that bHLH25 is required for H₂O₂-induced disease resistance; bHLH25-OE plants showed a resistance level much higher than Kitaake after H₂O₂ treatment (Fig. 4k), supporting the positive role of bHLH25 in H₂O₂-induced disease resistance.

miR397b-KO plants showed a disease resistance level so high that H₂O₂ treatment had no clear effects on miR397b-KO plants in disease resistance (Supplementary information, Fig. S10c), indicating that H₂O₂ effects on disease resistance are mostly mediated by repression of miR397b. Consistently, H₂O₂-induced disease resistance was weakened in miR397b-OE plants (Supplementary information, Fig. S10c). On the contrary, H₂O₂-induced disease resistance was not affected by CPS2 KO or CPS2 overexpression in rice plants (Supplementary information, Fig. S10d). These results suggest that the disease resistance induced by H₂O₂ treatment requires both bHLH25 and miR397b, but not CPS2. Taken together, the oxidation of bHLH25 by H₂O₂ is essential for the disease resistance mediated by the bHLH25–miR397b-OsLAC7/28/29 module.

M256 is essential for bHLH25 function

In silico analysis showed that M256 is in proximity to residues H255, E259 and R263 (Fig. 5a) that are key to DNA-binding specificity for bHLH-type transcription factors.³² We therefore replaced M256 with comparable hydrophobic residue valine to obtain bHLH25^M256V and analyzed the importance of M256. This replacement neither changed the nuclear localization nor the protein stability of bHLH25 in rice (Supplementary information, Fig. S11a, b). Moreover, M256V replacement did not significantly affect the alpha helix structure of bHLH25 basic domain when analyzed by the Self-Optimized Prediction³³ and Missense3D methods³⁴ (Supplementary information, Fig. S11c, d).

**Fig. 5: M256 is essential for bHLH25 to sense H₂O₂ and promote disease resistance.**

EMSA analysis showed that upon H₂O₂ treatment, oxidation of bHLH25 enhanced its binding to pmiR397b, but weakened its binding to pCPS2, whereas bHLH25^M256V lost binding to pmiR397b and pCPS2 and was not affected by H₂O₂ treatment (Supplementary information, Fig. S11e, f). Unlike bHLH25 plants, bHLH25^M256V plants failed to repress miR397b expression, promoted CPS2 expression, and enhanced disease resistance to M. oryzae (Fig. 5b; Supplementary information, Fig. S11g, i). In addition, bHLH25^M256V plants showed no cell death phenotype (Supplementary information, Fig. S12a, b). Thus, M256 is essential for bHLH25 to regulate miR397b and CPS2 expression and disease resistance.

The oxidation level of M256 is induced in vivo by H₂O₂ treatment and pathogen infection

Immunoblot analysis using a customized antibody (anti-oM256) specifically recognizing oxidized M256 in bHLH25 showed that anti-oM256 antibody detected an increase in oxidized M256 in bHLH25 protein, but no signals in bHLH25^M256V protein, after exposure to increased levels of H₂O₂ (Fig. 5c). To test the effects of H₂O₂ on bHLH25 in vivo, we treated Kitaake roots with H₂O₂ and assess bHLH25 RNA levels, protein levels, and M256 oxidation levels. We found no significant changes in bHLH25 RNA levels (Supplementary information, Fig. S12c). bHLH25 protein levels remained stable when protein extracts from purified nuclei isolated from rice plants treated with H₂O₂ were probed in immunoblot analysis; whereas the bHLH25 oxidation level at M256 clearly increased upon H₂O₂ treatment in the same in vivo experiment (Fig. 5d). We also performed a similar experiment by directly treating rice protoplasts with H₂O₂ and obtained similar results (Fig. 5e). Moreover, direct treatment of rice leaves with H₂O₂ significantly increased the oxidation level of bHLH25 and the RNA level of OsPR1b in rice leaves (Fig. 5f; Supplementary information, Fig. S12d).

We further tested the H₂O₂-sensing ability of bHLH25 in vivo in rice protoplasts by increasing H₂O₂ levels with OsRBOHA, which encodes an NADPH oxidase responsible for H₂O₂ production upon pathogen infection.³⁵ Rice protoplasts co-overexpressing bHLH25 and OsRBOHA displayed higher H₂O₂ levels and higher M256 oxidization levels than control protoplasts co-overexpressing bHLH25 and HA proteins (Supplementary information, Fig. S12e). Furthermore, protoplasts of Osrboha-KO plants showed lower H₂O₂ levels and lower M256 oxidization levels than those of wild-type Kitaake plants (Fig. 5g; Supplementary information, Fig. S12f, h). These results suggest that H₂O₂ directly targets bHLH25 at M256 in rice nucleus.

In bHLH25-OE plants, the M256 oxidation level was highly induced upon M. oryzae infection at 12–24 hpi, and started to decrease at ~36 hpi, reaching the lowest level at ~72 hpi. In contrast, mock-treated rice plants did not show obvious oxidation of bHLH25 (Fig. 5h). There is a delay in bHLH25 oxidation compared to H₂O₂ level, which may be due to the facts that transport of H₂O₂ from other parts of the tissue/cell into the nucleus takes time and that oxidation of bHLH25 also takes time to occur.^36,37 We also noticed a sharp increase in the total H₂O₂ content in rice leaves at 72 hpi with M. oryzae (Fig. 4e), while the M256 oxidation level was not increased at this time (Fig. 5h). It is likely that at 72 hpi, H₂O₂ mostly accumulates in mitochondria, chloroplasts, and cytoplasm due to cell death induced by phytoalexins³⁸ that may not oxidize nuclear bHLH25. Taken together, these results demonstrate that M256 is required for bHLH25 to directly sense H₂O₂ upon pathogen infection.

M256-like methionine residues are highly conserved among bHLH25 orthologues from other plant species

To assess the broadness and importance of potential M256-like residues in bHLH25 orthologues, we surveyed all bHLH25 orthologues from 110 plant species that have genome sequences available. Interestingly, M256-like methionine residues are highly conserved in the basic domains of the bHLH25 orthologues from all these 110 plant species (Fig. 6a; Supplementary information, Table S8).

**Fig. 6: bHLH25 confers resistance to multiple diseases and its M256 is highly conserved among bHLH25 homologs in plants.**

Moreover, immunoblot analysis on A. thaliana bHLH25 orthologue (AtbHLH25) showed that its conserved methionine (M291) became oxidized by H₂O₂ in wild-type AtbHLH25, but not in AtbHLH25^M291V (Fig. 6b), similar to the results of rice bHLH25. Thus, the mechanism by which bHLH25 senses H₂O₂ may broadly exist in the plant kingdom.

bHLH25 confers resistance to multiple diseases caused by fungi or bacteria

Next, we determined the role of bHLH25 in plant biotic and abiotic stress responses. We found that unlike the accumulation of H₂O₂ in rice leaves triggered by pathogen infection, which was high and concentrated at and around infection sites, the accumulation of H₂O₂ induced by salt and high-temperature stresses was relatively weak and dispersed (Supplementary information, Fig. S13a). Consistently, the oxidation level of bHLH25 was significantly increased in plants upon pathogen infection (Fig. 5h). In contrast, the oxidation level of bHLH25 showed no obvious changes in plants under salt stress or high-temperature stress (Supplementary information, Fig. S13b). More importantly, bHLH25 showed no significant effects on rice resistance to salt stress or high-temperature stress (Supplementary information, Fig. S13c, d). These results suggest that bHLH25 is specifically oxidized by H₂O₂ in response to pathogen attack to activate defense responses.

Biotic stresses caused by various pathogens induce H₂O₂ production in plants.³ We therefore further examined whether bHLH25 confers plant disease resistance to other pathogens, including R. solani (a major necrotrophic fungal pathogen to which no effective resistance genes are available), and Xanthomonas oryzae pv. oryzae (which causes devastating rice bacterial leaf blight).^39,40 bHLH25-OE plants showed smaller lesions, while bHLH25^M256V plants showed larger lesions, than Kitaake plants (Fig. 6c, d). In agreement, miR397b-KO plants and CPS2-OE plants developed significantly smaller lesions, whereas miR397b-OE plants and cps2-KO plants developed larger lesions (Supplementary information, Fig. S14). These results indicate that bHLH25 confers resistance to multiple diseases in rice and M256 of bHLH25 is required for this function.

Discussion

Our study discovers that rice bHLH25 directly senses H₂O₂ in the nucleus through M256 to regulate plant disease resistance. In contrast, PRXIIB and QSOX1 may sense cytosolic H₂O₂,^14,36 and HPCA1 senses extracellular H₂O₂ through their cysteine residues.¹⁵ We also uncover that bHLH25 senses H₂O₂ to activate two independent defense pathways. Mechanistically, pathogen infection promotes production of H₂O₂, which oxidizes bHLH25 at M256 to repress miR397b expression, elevating laccase expression and lignin biosynthesis to reinforce cell walls to prevent pathogens from penetrating plant cells. Lignin biosynthesis in turn consumes H₂O₂ and causes accumulation of non-oxidized bHLH25, allowing bHLH25 to preferentially promote CPS2 expression to increase phytoalexin biosynthesis, which inhibits growth and expansion of those pathogens that escape into the plant (Fig. 6e). This oxidation/non-oxidation status change of bHLH25 endows plants with an efficient defense system by reinforcing a physical barrier and accumulating antimicrobial phytoalexins. Moreover, this mechanism reveals how plants employ a single protein to coordinate two independent but essential defense pathways.

Excess levels of H₂O₂, lignin or phytoalexin cause severe side-effects on plant growth: excess H₂O₂ causing oxidation on DNA, RNA, and lipids is toxic to cells^11,41; elevated lignin levels inhibit plant growth^42,43,44; phytoalexin accumulation causes cell death as shown here (Fig. 3a, f) and in previous reports.^45,46 By using this oxidization/non-oxidization status change of bHLH25, plants homeostatically maintain H₂O₂, lignin, and phytoalexin at optimized levels by fine-tuning bHLH25 activity. This precise fine-tuning not only enables plants to effectively defend against pathogen attacks but also prevents the accumulation of these molecules that harm plants.

Our study suggests that high and concentrated accumulation of H₂O₂ at and around the pathogen infection sites in plant (Supplementary information, Fig. S13a) leads to clear oxidation of bHLH25 at M256 (Fig. 5h). However, the relatively weak and dispersed H₂O₂ accumulation induced by abiotic stresses, like salt and high temperature (Supplementary information, Fig. S13a), might not be able to oxidize bHLH25 or oxidizes bHLH25 at very low level which is not enough for bHLH25 function on lignin biosynthesis (Supplementary information, Fig. S13b). This indicates that plants utilize bHLH25 to sense accumulated high-intensity H₂O₂ under pathogen infection and then bHLH25 is oxidized to promote lignin biosynthesis for defending against pathogen attack, while in plants producing relatively dispersed and low-intensity H₂O₂ under abiotic stress, bHLH25 cannot be oxidized at an enough level for effectively activating related downstream response. Thus, we speculate that plants may employ bHLH25 to differentiate stimuli from biotic and abiotic stresses through the level of accumulated H₂O₂.

Methionine oxidation is an important post-translational modification for protein functions.⁴⁷ Our results showed that M256 oxidation promotes bHLH25 binding to pmiR397b but weakens bHLH binding to pCPS2. Importantly, the M256V replacement blocks binding of bHLH25 to both pmiR397b and pCPS2 (Supplementary information, Fig. S11e, f). These results are consistent with a previous report showing that the V replacements of ROS-sensing M281/282 in CaMKII abolish the oxidation-dependent kinase activity of CaMKII.⁴⁸ We further performed EMSA to determine the effect of M256 replacement by other amino acids on the DNA binding ability of bHLH25. The results showed that the replacements of M256 in bHLH25 by hydrophobic amino acids (A, P), hydrophilic amino acids (C, T, G, S, N), acidic amino acid (E) and basic amino acid (H) significantly weakened the ability of bHLH25 to bind to pmiR397b and pCPS2 (Supplementary information, Fig. S15). Thus, the replacements of M256 by any other amino acids mainly weaken the DNA binding ability of bHLH25, instead of changing the oxidation state of bHLH25.

Few genes conferring resistance against a wide range of diseases have been identified. As the mechanism used by bHLH25 to sense H₂O₂ very likely broadly exists in the plant kingdom (Supplementary information, Table S8) and bHLH25 confers broad-spectrum resistance to diverse diseases caused by fungi or bacteria (Figs. 5b, 6c, d), our study may directly lead to the discovery of a whole class of bHLH25-like transcription factors that confer broad-spectrum disease resistance in other plant species. In Arabdopsis, the transcription factor CCA1 HIKING EXPEDITION (CHE) is recently reported to undergo sulfenylation in its conserved cysteine residue in an H₂O₂ concentration-dependent manner to promote the expression of salicylic acid-biosynthesis gene to enhance systemic acquired resistance.⁴⁹ Our study, together with this finding, implicate that the Met- and Cys-oxidation mechanism may broadly be employed by eukaryotic transcription factors to sense H₂O₂ to change their roles. Thus, our findings may have great impacts on the research field of H₂O₂ signaling in both plants and animals.

Materials and methods

Plant materials and growth conditions

The full-length cDNA sequences of OsLAC7 (Os01g0850550), OsLAC28 (Os12g0257600), OsLAC29 (Os12g0258700), CPS2 (Os02g0571100), bHLH25 (Os01g0196300) and OsRBOHA (Os01g0734200) were cloned from mRNAs of Nipponbare rice. For generating CRISPR/Cas9 vectors to knock out single genes, 20 bp target sequences were selected by online tool (http://crispr.hzau.edu.cn/CRISPR2/), and were cloned into the BGK032-Cas9 vector. For generation of double KO of OsLAC7/29 and triple KO of OsLAC7/28/29, we cloned the target sequences into the pTCRISPR-Cas9 vector. For generating the construct for pbHLH25:YFP, the 1500 bp native promoter of bHLH25 was amplified from Nipponbare DNA to replace the 35S promoter in pCAMBIA1300–35S:YFP. To study M256 for bHLH25 function, we generated mutant bHLH25^M256V in which M256 of bHLH25 was replaced by a valine, which is a hydrophobic amino acid commonly used for amino acid substitution to minimize structural changes as previously reported.⁴⁸ For generating vectors overexpressing genes other than bHLH25, including miR397b, CPS2 and OsLAC7/28/29, we cloned the precursor of miR397b (pre-miR397b) DNA with extended 300 bp toward both 3’ and 5’, and the cDNA sequences of CPS2, OsLAC7, OsLAC28 and OsLAC29 into the 35S promoter-driven overexpression vector individually.

For generating transgenic plants, vectors were transferred into Agrobacterium tumefaciens EHA105 as previously reported.⁵⁰ EHA105 strains containing vectors were used to transform Kitaake background (Oryza sativa japonica). bHLH25 KO plants were transformed with constructs carrying bHLH25 or bHLH25^M256V, driven by a native bHLH25 promoter or driven by 35S promoter to determine the importance of M256 for bHLH25. For genes other than bHLH25, including miR397b, CPS2 and OsLAC7/28/29, we made their single, double, or triple KO plants to verify their roles in biosynthesis of lignin or phytoalexin and disease resistance, and introduced the corresponding overexpression constructs of these genes into wild-type Kitaake plants to make their overexpression plants. The Oslac7/28/29-KO/miR397b-KO plants were created by crossing Oslac7/28/29-KO#1 and miR397b-KO#1. The miR397b-OE/bHLH25-OE plants were created by crossing miR397b-OE#1 and bHLH25-OE#4. The cps2-KO/bHLH25-OE plants were created by crossing cps2-KO#1 and bHLH25-OE#4.

Rice plants were grown in the experimental field of the Sichuan Agriculture University (Chengdu, Sichuan, China). We used PCR to amplify the Cas9-edited target sequences for KO plants and sequenced the PCR products to ensure that the plants we used for experiments were progeny of sequenced stable homozygous plants. For H₂O₂ treatment and M. oryzae spray inoculation assays, plants were grown with Yoshida’s nutrient solution in an incubator or grown in a greenhouse under a short-day photoperiod (11 h light/13 h dark) with light strength of approximately 20,000 lux at 28 °C/22 °C (day/night) cycles.⁵¹

H₂O₂ treatment and H₂O₂ content measurement

For treating roots of rice plants with H₂O₂, we transferred three-week-old seedlings of Kitaake from the basal nutrition solution to nutrient solution containing 1 mM H₂O₂ or water mock for 0–72 h, then used treated leaves from the same part of plants for further detection as previously described.⁵² For treating rice leaves with H₂O₂, we clipped rice leaves to 1 cm² and treated them with 1 mM H₂O₂ or mock solution. For treating rice protoplasts with H₂O₂, we added water or H₂O₂ to protoplasts of two-week-old rice seedlings to 0 or 1 mM H₂O₂ and the treatment lasted for 0–2 h. For treating protein with H₂O₂ in vitro, we incubated 10 μg purified GST-tagged protein with different concentrations of H₂O₂ in 300 μL solution at 28 °C for 15 min.⁵³

For measuring the total H₂O₂ contents of rice plants, we finely ground ~20 mg of each rice leaf sample or used 0.5 mL rice protoplasts for further measurement using a hydrogen peroxide assay kit (Beyotime, #S0038) following the manufacturer’s instructions and previous studies.^52,54 This assay is based on the formation of ferric ion–xylenol orange complexes from xylenol orange and ferrous ion upon oxidation by H₂O₂. As H₂O₂ accumulation, expression levels of PR genes and OsLAC7/28/29 genes were most profound at 60 and 72 hpt with H₂O₂, we therefore focused on these two time points for examining rice gene expression in rice leaves and disease resistance after treatment with H₂O₂. These experiments were done with three biologically independent replications.

Plant disease resistance assay

For inoculation with M. oryzae, different strains of Zhong10-8-14 and ZB25 were used to verify disease resistance of rice plants. For punch inoculation assays, rice leaf strips of three-week-old seedlings were punched, inoculated with 5 μL spore suspension (1 × 10⁵ spore/mL) of M. oryzae, and incubated at 28 °C as previously described.⁵⁵ Lesion length was measured 7 days after inoculation. For relative fungal growth measurement, DNA was extracted from infected leaves to examine the relative DNA level of the Pot2 gene (AF314096) of M. oryzae, relative to rice ubiquitin (Os03g0234200), by qPCR analysis.⁵⁶ For spray inoculation assays, leaves of three-week-old seedlings were sprayed with spore suspension of M. oryzae.⁵⁷ Then the leaves from the same part of the plants inoculated with M. oryzae were taken for further measurement of H₂O₂ content, lignin, phytocassane C, and gene expression at time points from 0–96 h.^55,58 For determination of plant disease resistance by lesion measurement in spray inoculation, the number of lesions were counted for each infected leaf for 7 days after M. oryzae inoculation.

For inoculation with R. solani, fully expanded leaves of tillering-stage rice plants were inoculated with mycelial clumps of the AG-1-IA strain following a previously reported method.⁵⁹ For inoculation with Xanthomonas oryzae pv. oryzae, fully expanded leaves of tillering-stage rice plants were cut at about 1 cm from the leaf tip using a pair of scissors pre-dipped in the bacterial suspension of strain PXO99A (OD600 = 0.6) following a previously reported method.¹⁹ The spontaneous cell death phenotype of bHLH25-OE plants is completely different in appearance from the disease lesions in rice leaves caused by pathogen infection, therefore the cell death phenotype does not affect the determination of disease resistance in plants. Three biologically independent replications were performed for all plant disease assays.

RNA extraction, RT-qPCR, semi-qPCR and RNA-seq analysis

For RNA extraction to perform RT-qPCR and RT-semi-quantitative PCR (RT-semi-qPCR), samples were frozen and finely ground in liquid nitrogen and total RNA was extracted using TRIzol (Invitrogen, #15596018). Gene transcript levels were determined by RT-qPCR (QIAGEN, #208054) or RT-semi-qPCR (Vazyme, #P505-d1). The first strand cDNA was synthesized using a reverse transcription kit (Takara, #RR047B). At least three biologically independent replications were performed. To specifically detect the expression of miR397b, we designed primers to examine pre-miR397b levels. Amplification cycles were normalized against rice ubiquitin, or NbEF1α in N. benthamiana,⁶⁰ by calculating differences between the threshold cycle (C_T) of the target gene and the C_T of UBQ5 or NbEF1α. Relevant primer sequences are given in Supplementary information, Table S9.

For RNA-seq analysis, total RNA extracted from leaves of three-week-old rice plants at 60 and 72 hpt with or without H₂O₂ were used to construct Illumina sequencing libraries (NEB, #102715-922). Illumina sequencing libraries were sequenced using the Illumina system Hiseq4000 with 250–300 bp read lengths (Novogene, Bejing, China). The paired-end clean reads of RNA-seq were aligned to the Oryza sativa IRGSP 1.0 genome DNA using the HISAT2 software. We performed quantitative analysis of gene expression level for each sample, then combined them to obtain expression matrix for all samples. Analysis of differentially expressed genes (DEGs) was conducted using the edgeR software,⁶¹ and statistical analysis was performed by Hypergeometric test with Bonferroni correction. The enriched functions of DEGs in RNA-seq data were annotated with the gene ontology function in the clusterProfile and agriGOv2⁶² software.

Promoter analysis

Conserved cis-elements in the promoter regions of OsLAC7/28/29 were identified in the 1000 bp flanking sequences upstream of the transcription start sites of OsLAC7/28/29. The promoter sequences were analyzed by the PlantCARE⁶³ online tool (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) to predict cis-elements.

MBS analysis

To decipher miRNAs regulating the expression of OsLAC7/28/29, we used the psRNATarget online tool (https://www.zhaolab.org/psRNATarget/) to predict potential MBSs⁶⁴ in mRNAs of OsLAC7/28/29. MBSs were selected with a miRNA binding expectation value less than 0.5.

Transactivation assay in N. benthamiana

Transactivation assays were performed as previously reported.⁶⁵ The Agrobacterium strain GV3101 was used for transformation of N. benthamiana. For generating reporter plasmids for regulation by miR397b, we fused the miR397b MBS or mutated MBS (mMBS) of OsLAC7/28/29 with YFP in the pCAMBIA1300-35S:YFP vector. 35S:miR397b was used as the effector. For generating reporters for regulation by bHLH25, we amplified the 1000 bp regions of pmiR397b and pCPS2, then inserted them into the pCAMBIA1300-35S:YFP vector to replace 35S promoter individually. 35S:bHLH25-HA acted as an effector, and 35S:Renilla luciferase (RLUC)-HA acted as a negative control. For transformation, GV3101 strains containing constructs in MMA buffer (10 mM MES, 10 mM MgCl₂, 100 µM AS) at OD600 = 1 were infiltrated into leaves of N. benthamiana and incubated for 36 h, followed by injection of different concentrations of H₂O₂ and incubation for another 12 h. 0.5 g leaves were harvested, followed by extraction of RNAs for RT-qPCR analysis or proteins for immunoblot analysis. Each set of data were generated with three biologically independent replications and derived from the total protein extracts obtained from the same N. benthamiana leaf.

To determine the transcriptional regulation of OsLAC7/28/29 by TBPs and CBFs, we only successfully obtained the CDS of 18 of 25 TBPs and CBFs, and then cloned them into the pGreenII-62SK vector driven by 35S promoter to generate effector constructs. The 1000 bp promoter regions of OsLAC7, OsLAC28 and OsLAC29 were separately cloned into the pGreenII-0800 to drive firefly luciferase (LUC) expression. Renilla luciferase (RLUC) driven by 35S promoter was used as a reference. The empty pGreenII-62SK vector was used as a negative control. Rice protoplasts were transfected with different combinations of plasmids and incubated overnight, then collected and lysed for determining relative RNA levels of LUC and RLUC by RT-qPCR analysis. Relevant primer sequences are given in Supplementary information, Table S9.

Yeast one-hybrid assay

For generating the bait vector for yeast one-hybrid library screening, the 1000 bp pmiR397b was fused to the HIS3 reporter in pHIS2 vector. The prey cDNA library, constructed from rice seedlings of Digu inoculated with M. oryzae, was prepared in the pGADT7-Rec2 vector (Clontech) and transformed into yeast strain AH109 (Clontech) as described previously.⁵⁵ Yeast diploids were selected by incubating at 30 °C for 4 days on minimal medium SD lacking Trp, Leu (-T-L) or lacking Trp, Leu, His (-T-L-H) while supplemented with 30 mM 3-amino-1,2,4-triazole (3-AT, Sigma-Aldrich, #A8056). Positive clones were sequenced and confirmed. For confirming bHLH25 DNA binding activity, the CDS of bHLH25 was amplified and cloned into the pGADT7-Rec2 vector fusing to the activation domain. Constructs were co-transformed into AH109. The empty vector pGADT7-Rec2 was co-transformed as a negative control. DNA–protein interactions were determined by the growth of transformants on the nutrient-deficient medium with 30 mM 3-AT. Relevant primer sequences are given in Supplementary information, Table S9. Three biologically independent replications were performed.

Rice protoplast transient expression analysis

Protoplasts of two-week-old etiolated Kitaake seedlings were isolated and transformed with constructs following the method as described previously.⁶⁶ For subcellular localization analysis, we generated constructs expressing bHLH25-GFP from vector pRTVcGFP (Ubi:GFP).⁶⁷ Constructs for Ubi:GFP or Ubi:bHLH25-GFP were co-transformed with nuclear marker pSAT6:RFP-VirD2NLS respectively into protoplasts. Fluorescence was examined under a confocal microscope at 16 h post transformation. For testing whether bHLH25 senses H₂O₂ produced by NADPH oxidase in vivo, we generated a construct expressing immunity-related NADPH oxidase gene OsRBOHA by cloning it into vector pRTVcHA (Ubi:4×HA). About 1 μg plasmid was used to transform 100 μL protoplasts (2 × 10⁶ protoplast/mL) by incubating at 28 °C for 16 h. Relevant primer sequences are given in Supplementary information, Table S9. Three biologically independent replications were performed.

Protein purification in vitro and extraction in vivo

For generating vectors expressing GST-tagged wild-type and mutant bHLH25 proteins in vitro, CDSs of bHLH25 and and its mutant forms were individually cloned into the pGEX-6p-1-small ubiquitin-like modifier (SUMO) vector. To investigate function of the A. thaliana bHLH25 orthologous protein, CDSs of AtbHLH25 (AT5G56960) and AtbHLH25^M291V were individually cloned into the pGEX-6p-1 vector. The constructs were transformed into the E. coli strain Transetta (DE3) (TransGen Biotech, #CD801-02). Relevant primer sequences are given in Supplementary information, Table S9. Transetta strains were grown in LB medium containing 100 mg/mL ampicillin at 37 °C to OD600 = 0.6. The expression of the fusion proteins was induced by adding IPTG (Sigma, #I6758) to 1 mM and incubating at 28 °C for 14 h. The fusion proteins were purified with glutathione sepharose 4B beads (GE, #17075601).

For extracting protein in vivo, total protein was extracted with native protein extraction buffer (50 mM Tris-MES at pH 8.0, 0.5 M sucrose, 1 mM MgCl₂, 10 mM EDTA, 5 mM DTT and protease inhibitor cocktail (TransGen Biotech, #DI111-01)). For bHLH25-YFP immunoprecipitation (IP), we used the same amount of GFP-trap agarose beads (Chromotek, #gta-100) to enrich bHLH25-YFP protein in leaves of the same weight at each time point. The result of immunoblot analysis showed that the level of bHLH25-YFP protein extracted from rice was indeed different at different time points, which may be due to the different protein turnover rates at different time points that regulate bHLH25 protein level. Plasma membrane protein OsRBOHA-4×HA and nuclear protein bHLH25 were extracted by differential centrifugation according to the manufacturer’s instruction of Plasma Membrane Protein and Nuclei Isolation Kit (Invent, #PF-045).

DAP-qPCR analysis

DAP assay was performed with three biologically independent replications as previously described.⁵⁵ In brief, total DNA was extracted from three-week-old Kitaake seedlings and sonicated into 100–500 bp fragments by using ultrasonic crusher. DNA fragments, GST-tagged protein, and glutathione-agarose were co-incubated in incubation buffer (50 mM Tris-HCl, pH 7.0, 1 mM EDTA, 100 mM KCl, 5% Glycerol, 0.1% Triton X-100, 1 mM DTT) at 4 °C for 2 h. After co-incubation, glutathione-agarose beads were washed three times with incubation buffer. Then 4 mL 5 M NaCl was added into the sample for each 100 mL volume and incubated at 65 °C for 4 h to break down cross-linked GST-tagged protein and DNA fragments. Then extracted DNA was applied for qPCR using respective primer pairs. The ubiquitin promoter was used as the internal control. Relevant primer sequences are given in Supplementary information, Table S9.

ChIP-qPCR analysis

Four-week-old 35S:bHLH25#1 (in bhlh25-KO#6 background) plants were used for ChIP analysis according to methods described previously.³⁰ Protein A magnetic beads (Sangon Biotech, #D110560) with no antibody (No Abs) were used as a negative control. Relevant primer sequences are given in Supplementary information, Table S9. Three biologically independent replications were performed.

EMSA

The detailed procedure of EMSA follows instructions of Light Shift Chemiluminescent EMSA Kit (Thermo Scientific, #20148). Biotin was labeled at the 5’ end of oligonucleotides containing cis-elements. For pre-treating bHLH25 proteins with H₂O₂, we used glutathione sepharose 4B beads to pull out bHLH25 protein pre-treated with H₂O₂. We then incubated the pulled-out bHLH25 protein with probes in reaction solution free of H₂O₂ to perform EMSA. The components of 10 μL binding reaction include 10 mM Tris-HCl (pH 7.5), 50 mM KCl, 50 ng/mL Poly (dI·dC), 1 mM DTT, about 0.5 μg purified GST-tagged proteins, 0.5 μM pmiR397b-G-box or 0.1 μM pCPS2-N-box biotin-labeled DNA. Each dataset consisted of three biologically independent replications.

Lignin measurement

Lignin measurement was performed as reported previously with some adjustments with three independent biological replications.⁶⁸ Briefly, fresh samples (0.5 g) were ground in liquid nitrogen and re-suspended in 7 mL ethanol. The sample was pelleted by centrifugation (1400× g, 5 min) and the pellet was extracted with the following solvents for 15 min each: two times with 7 mL ethanol, three times with 7 mL distilled water, two times with 7 mL ethanol/n-hexane (1:2 [v/v]), and once with 5 mL ethanol. The pellet was dried at 60 °C for 72 h. Subsequently, 20 mg of dry pellet was added to 0.5 mL of 25% acetyl bromide (v/v in glacial acetic acid) and incubated at 70 °C for 30 min. After quick cooling on ice, the sample was dissolved in 6 mL solution containing 0.9 mL 2 M NaOH, 0.1 mL 7.5 M hydroxylamine-HCl and 5 mL of glacial acetic acid. The absorbance of the supernatant was measured at 280 nm after centrifugation (1400× g, 5 min). Lignin content was calculated using the Beer-Lambert law (A = Kbc).

Histochemical staining assay

For histochemical staining of lignin, we used phloroglucinol-HCl staining as described before.⁶⁹ In brief, leaves of four-week-old rice plants were sectioned with a razor blade. Sections were stained with 1% phloroglucinol (w/v) in 12% HCl for 5 min and immediately observed under a stereo microscope (SMZ1000, Nikon, Japan). The thickness of sclerenchyma cells was quantitated as described previously.¹⁸ These experiments were done with three biologically independent replications.

Phytoalexin measurement

Phytoalexin content was measured with three biologically independent replications as described previously.²⁹ In brief, rice leaf samples (10 mg fresh weight) were soaked in 1 mL of extraction solvent (MeOH/H₂O, 80:20 [v/v]) in glass tubes and incubated at room temperature overnight. The extract was centrifuged (4 °C, 15 min, 16,000× g). The supernatants were then collected and dried up. Finally, dried samples were subjected to phytoalexin measurement by LC-ESI-MS/MS as described before.⁷⁰

TEM analysis

Leaf segments from four-week-old rice plants were transferred into plastic tubes with fresh TEM fixative (Servicebio, #G1102) for fixation under vacuum extraction until the samples sink to the bottom. Samples were treated according to a previous study.⁷¹ Samples were observed in TEM Hitachi HT7800. Sclerenchyma cell wall thickness was calculated with ImageJ. These experiments were done with three biologically independent replications.

Mass spectrometry analysis of bHLH25 oxidation

For identifying post-translational modifications on bHLH25 protein after M. oryzae infection, we used extraction buffer (50 mM Tris-MES at pH 8.0, 0.5 M sucrose, 1 mM MgCl₂, 10 mM EDTA, 5 mM DTT and protease inhibitor cocktail) to obtain cell extracts from Kitaake rice leaves treated with M. oryzae or mock control for 24 h. Then we incubated purified GST-bHLH25 protein with cell extracts at 4 °C for 4 h. Finally, the GST-bHLH25 protein was purified using glutathione sepharose 4B beads (GE, #17075601) and used for mass spectrometry analysis.

The stained protein band was cut, and the volume of the cut gel was 0.5–1 mm³. The cut gel was de-stained with 50% acetonitrile (ACN) in 50 mM triethylammonium bicarbonate (TEAB) and dehydrated upon washing with 100% ACN till the gel turned white. Proteins in gel were treated with 1 mL of 10 mM DTT for 40 min at 56 °C and subsequently alkylated with 1 mL of 50 mM IAM for 30 min in the dark. The gel was washed with de-staining buffer and treated with ACN as above. Protein digestion: add 10–20 µL 10 ng/µL pancreatic enzyme, rest on ice for 30 min, and wait for the enzyme solution to be completely absorbed by the colloidal particles, add 100 mM TEAB buffer to the total volume of 100 µL, and allow enzyme digestion at 37 °C overnight. The supernatant was collected by centrifugation at low speed, 100 µL 0.1% formic acid was added into the glue, and the supernatant was collected by centrifugation at low speed after oscillating at room temperature for 5 min. Sample cleanup: centrifuge at room temperature 12,000× g for 5 min, take the supernatant slowly through the C18 desalting column, and then wash it with 1 mL cleaning solution (0.1% formic acid, 4% acetonitrile) for 3 consecutive times, then add 0.4 mL eluent (0.1% formic acid, 75% acetonitrile) for 2 consecutive times, and then freeze dry the elution sample after combination.

Mobile phase A (100% water, 0.1% formic acid) and B solution (80% acetonitrile, 0.1% formic acid) were prepared. The lyophilized powder was dissolved in 10 μL of solution A, and centrifuged at 15,000 rpm for 20 min at 4 °C. 1 μg of sample was injected into a C18 Nano-Trap column. Peptides were separated in an analytical column, using a linear gradient elution. The separated peptides were analyzed using a Q Exactive HF-X mass spectrometer (Thermo Fisher), with ion source of Nanospray Flex (ESI), spray voltage of 2.3 kV, and ion transport capillary temperature of 320 °C. Full scan range from m/z 350 to 1500 with resolution of 60,000 (at m/z 200), an automatic gain control (AGC) target value was 3 × 10⁶ and a maximum ion injection time was 20 ms. The top 40 precursors of the highest abundance in the full scan were selected and fragmented by higher energy collisional dissociation and analyzed in MS/MS, where resolution was 15,000 (at m/z 200), the AGC target value was 1 × 10⁵, the maximum ion injection time was 45 ms, a normalized collision energy was set as 27%, an intensity threshold was 2.2 × 10⁴, and the dynamic exclusion parameter was 20 s. The resulting spectra from each fraction were searched against Oryza_sativa_subsp_japonica_uniprot_2021_3_9.fasta database separately by Proteome Discoverer 2.2. The identified protein contains at least 1 unique peptide with FDR no more than 1.0%. Proteins containing similar peptides that could not be distinguished by MS/MS analysis were identified as a same protein group. Oxidation of residues was specified in Proteome Discoverer 2.2 as variable modifications. These experiments were done with three biologically independent replications.

Immunoblot analysis

Proteins were electrophoresed on a 10% SDS-PAGE gel before being transferred to a PVDF membrane (Merck Millipore, Billerica, MA, USA) using a Trans-Blot® SD Semi-Dry Transfer Cell (Bio-Rad, Hercules, CA, USA) according to standard procedures. For detecting YFP and GFP proteins, we used anti-GFP as previously reported.⁷² For detecting biotin, we used streptavidin-HRP (Invitrogen, #434323). The monoclonal antibody against oxidized M256 of bHLH25 (anti-oM256) was generated by ABclonal Technology (Wuhan, China). We used an oxidized bHLH25 short peptide (QLQH-(oxidized M)-ISER) to immunize mice and screen for positive hybridoma clones, and a non-oxidized bHLH25 short peptide (QLQH-(M)-ISER) to exclude any hybridoma clones generating antibodies that reacted with the non-oxidized peptide, in enzyme-linked immunosorbent assay (ELISA) and immunoblot analysis. Relative levels of band gray-scale values were measured by Image Lab 3.0 software (Bio-Rad). Relative oxidation value was calculated by dividing the oxidation value of bHLH25 treated with M. oryzae by that of bHLH25 treated with mock. The antibodies used in this study: anti-GFP (Invitrogen, #PA1-980A, 1:10,000), anti-HA (Invitrogen, #26183, 1:5000), anti-Actin (Sangon Biotech, #D110007, 1:2000), anti-GST (Invitrogen, #A-5800, 1:5000), anti-Rabbit IgG HRP (Invitrogen, #31460, 1:10,000), anti-Mouse IgG HRP (Invitrogen, #62-6520, 1:10,000), and anti-oM256 (1:1000). All experiments were conducted with three biologically independent replications.

In silico analysis

For predicting the structure of bHLH25 protein, we used the SWISS-MODEL (https://swissmodel.expasy.org/)⁷³ or Alphafold (https://www.alphafold.ebi.ac.uk/).^74,75 For the prediction by SWISS-MODEL, bHLH transcription factor MYC2 was selected as the most suitable template for bHLH25 to calculate and simulate the structure of bHLH25. For evaluating the structural effect of bHLH25^M256V on alpha-helix structure in the basic domain of bHLH25, we performed the Missense3D analysis³⁴ based on the structures of bHLH25 and bHLH25^M256V proteins predicted by Alphafold2. For evaluating the structural effect of bHLH25^M256V on the protein secondary structure of bHLH25, we conducted self-optimized prediction method (SOPMA)³³ on Network Protein Sequence Analysis⁷⁶ to analyze the alpha-helix structure in the basic domain of bHLH25.

Protein stability determination

Protein stability was determined by immunoblot analysis according to a previous study.⁷⁷ Plasmids harboring bHLH25-GFP or bHLH25^M256V-GFP were used to transform rice protoplasts. After incubation for 12 h, protoplasts were treated with 100 mM protein synthesis inhibitor cycloheximide for different time periods (time points). Three biologically independent replications were performed.

Orthologue analysis

The protein accessions and detailed information of bHLH25 orthologues in the plant kingdom were obtained from the Basic Local Alignment Search Tool of the National Center for Biotechnology Information (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The alignment was visualized using JalView and colored by ClustalX. All protein accessions used for analysis and other M256 orthologues of bHLH25 in the conserved basic domains were shown in Supplementary information, Table S8.

High-temperature treatment

High-temperature treatment was performed with three biologically independent replications according to a previous report.⁷⁸ In brief, three-week-old seedlings were treated at 42 °C in growth chambers for three days, then they were returned to normal conditions (26 °C) for 5-day recovery.

Salt stress treatment

Salt stress treatment was performed with three biologically independent replications according to a previous report.⁷⁹ Salt stress was initiated 3 h after the start of the light period by incubating 3-week-old seedlings in nutrition solution with 150 mM NaCl for three days, then they were returned to normal nutrition solution for 5-day recovery.

3,3-diaminobenzidine (DAB) staining

DAB staining was conducted with three biologically independent replications as described previously.⁷⁷ Rice leaves were submerged in staining buffer (10 mM MES, pH 6.5 and 1 mg/mL DAB) for 18 h in darkness at room temperature and transferred to 90% (v/v) ethanol at 65 °C until all the chlorophyll had been removed. Then the cleared leaves were photographed.

Quantification and statistical analysis

For statistical analysis, the two-tailed Student’s t-test was performed by using the Office Excel software; the one-way ANOVA with LSD and Dunnett’s multiple comparison test at P < 0.05 was performed by using the IBM SPSS software (v.21); the two-way ANOVA with Tukey’s multiple comparison test at P < 0.05 was performed by using R (v.4.4.0). Data were presented as means ± SD as indicated. P values over 0.05 were considered not significant. Sample sizes, statistical tests used and P values are stated in the figures and figure legends.