Ac/Ds-Induced Receptor-like Kinase Genes Deletion Provides Broad-Spectrum Resistance to Bacterial Blight in Rice

Rice bacterial blight caused by Xanthomonas oryzae pv. oryzae (Xoo) seriously affects rice yield production. The discovery and application of broad-spectrum resistance genes are of great advance for disease resistance breeding. Previously, we identified that multiple receptor-like kinase (RLK) family gene deletions induced by the Ac/Ds system resulted in a lesion mimic symptom. In this study, the mutant #29 showed that this lesion mimic symptom was isolated. Further analysis identified that four RLK genes (RLK19-22) were deleted in the #29 mutant. The #29 mutant exhibited broad-spectrum resistance to Xoo and subsequent analyses identified that pathogenesis-related genes PR1a, PBZ1, and cellular H2O2 levels were significantly induced in the mutant compared to wild-type plants. A genetic analysis revealed that reconstruction of RLK20, RLK21, or RLK22 rescued the lesion mimic symptom of the #29 mutant, indicating that these three RLKs are responsible for broad-spectrum resistance in rice. Further yeast two hybrid and bimolecular fluorescence complementation assays demonstrated that RLK20 interacts with RBOHB, which is a ROS producer in plants. Compared to wild-type plants, the #29 mutant was more, while #29/RLK20 ox was less, susceptible to MV (methyl-viologen), an ROS inducer. Co-expression of RLK20 and RBOHB reduced RBOHB-promoted H2O2 accumulation in the cells. Taken together, our research indicated that the RLKs may inhibit RBOHB activity to negatively regulate rice resistance to Xoo. These results provide the theoretical basis and valuable information about the target genes necessary for the successful breeding of rice cultivars resistant to bacterial blight.


Introduction
Rice is an important crop that feeds more than 50% of the world's population. Rice bacterial blast (BB) caused by Xanthomonas oryzae pv. oryzae (Xoo) is a serious disease that severely threatens yield production. In the 1980s, large outbreaks of the disease were frequent [1]. The disease has been effectively controlled with the application of resistance genes such as Xa3/Xa26 and Xa4 during the breeding process [2,3]. More than 40 resistance genes have been identified to date and 11 of these genes have been successfully cloned [4,5]. These resistance genes encode different types of proteins. For example, Xa3/Xa26, Xa4, and Xa21 encode receptor-like kinase; Xa13, Xa25, and Xa41 encode sugar and are eventually exported as a transporter (SWEET); Xa10, Xa23, and Xa27 encode executor proteins; and Xa1 and Xa5 encode other types of proteins [6].
Xa3/Xa26, Xa4, and Xa21 encoding receptor-like kinase are involved in PAMP-Triggered Immunity (PTI) [3,7,8]. Both Xa3/Xa26 and Xa21 confer broad-spectrum resistance to various Xoo races [8,9]. Xa26 was first identified in the rice indica variety Minghui 63 [9]. It is the same gene as Xa3 identified in the japonica variety Wase Aaikolu 3 [7]. XA3/XA26 interacts with somatic embryogenesis receptor kinase 2 (OsSERK2) and triosephosphate isomerase 1.1 (OsTPI1.1) to further regulate rice resistance to Xoo. Suppression of OsTPI1.1 in rice weakens its resistance to Xoo [10]. Similar to the role of OsTPI1.1, OsSERK2 positively regulates the rice resistance via the interaction with XA3/XA26 [11]. XA21 originated from Oryza longistaminata, is a transmembrane immune receptor that responds to sulfated derivatives from Xoo, and induces XA21-mediated immunity X (RaxX) in rice [12]. While Xa21-mediated resistance is not sustained throughout the entire growing season, the rice plants achieve full resistance only at the adult stage [13]. The XA21/RaxX interaction fits the "gene-for-gene relationship" theory, like plants without XA21 that are susceptible to Xoo strains even when RaxX is produced. Furthermore, rice plants that harbor XA21 in the genome also fail to respond to Xoo strains without RaxX production [14][15][16][17]. Xa4 encodes a kinase, which belongs to the subfamily of receptor-like kinases (RLKs), which are localized on the cell wall. Unlike Xa3/Xa26 and Xa21, Xa4 is a race-specific resistance gene to Xoo and strengthens the cell wall during the entire growing season of rice [3,18]. Xa4 is one of the most widely used genes in Xoo resistance breeding in rice, since it does not compromise yield production [3]. Incompatible interactions of rice-Xoo that induce Xa4 expression further increase cellulose synthase (CesA) levels to strengthen the cell wall, leading to Xoo resistance. In addition, Xa4 induces the production of phytoalexins sakuranetin and momilactone A to inhibit Xoo [19][20][21].
Although there have been many resistance-breeding studies over recent years, complete control of the disease remains a challenge. One important reason for the disease outbreaks may be due to the production of new toxic effectors or the loss of avirulence (Avr) function of the effector belonging to Xoo, which can further cause the loss of resistance of a previously resistant cultivar [22,23]. The phenomenon has been reported that, due to the large-scale cultivation of rice varieties with a single resistance background, selection pressure was increased, thus further inducing Xoo mutation and breakthrough variety resistance [24,25]. The polymerization of resistance genes is an effective strategy for disease resistance breeding, however, it is time-consuming. The discovery of broad-spectrum resistance genes to multiple races and elucidation of their resistance mechanisms can provide resources and a theoretical basis for disease control.
In rice, 29% of all predicted 37,544 genes are family genes clustered on the genome (International Rice Genome Sequencing Project, 2005). However, the evolutionary significance and function of these family genes remain largely unknown. Due to the redundant function of these family genes, mutation of a single gene has often failed to generate an identifiable phenotype, making gene function studies challenging. The Ac/Ds system [26] was put forward as an excellent tool for functional studies and germplasm innovation. The maize Ac (Activator) element encodes a transposase, which catalyzes the transposition of Ds (Dissociation) elements. In general, the transposition results in the excision of the element from a donor site and insertion into a target site. However, recognition of the 5 and 3 ends of different Ac/Ds elements by Ac transposase could induce alternative transposition events, including deletions, duplications, inversions, and other sequence rearrangements [27][28][29]. In addition, the transposition of Ac/Ds preferentially occurs in the genic regions, which would shuffle the coding and regulatory sequences, and thereby generate new genes [30]. The frequency of transposon-induced chromosomal rearrangements increases by at least three times than found in natural populations of maize regenerated via tissue culture [26].
In this study, Ac/Ds-induced chromosomal deletions at RLK locus were identified. Among the deletion mutants, RLK (19-22) mutant #29 with four RLKs deletions exhibited broad-spectrum resistance to Xoo races. PR genes and H 2 O 2 were largely induced in the #29 mutant. Genetic analysis indicated that RLK20, RLK21, RLK22 have redundant functions in regulating the lesion mimic phenotype of the #29 mutant. Furthermore, RLK20 was identified to interact with RBOHB. Compared to the wild-type plants, the #29 mutant plants were more, while #29/RLK20 ox plants were less, susceptible to MV, which is an ROS inducer. Furthermore, co-expression of RBOHB and RLK20 reduced RBOHB-promoted H 2 O 2 generation. These results indicated that the RLKs negatively regulate rice broadspectrum resistance to multiple races of Xoo by controlling the H 2 O 2 and PR gene levels.
These results provide useful information for using the Ac/Ds system to study clustered gene families in plants through the identification of the RLK functions in rice defense.

A Pair of Ds Elements Generate Diverse Chromosomal Rearrangements
The amount of various family genes that display redundant function clustering at the chromosomes hampers the study of gene function in rice. Ac/Ds transposable elements generate chromosomal rearrangements including deletions, inversions, and duplications via the alternative transposition mechanism in rice [29]. To study the function of the redundant family genes, many chromosome fragment rearrangement/deletion mutants induced by the Ac/Ds system were developed. The T-DNA provided for transposase was constructed containing the CaMV 35S promoter to drive the Ac cDNA. Another T-DNA was constructed containing a modified Ds element (Figure 1a). A schematic diagram of the mechanism of a pair of closely located Ds-induced deletion/rearrangements and homologous recombination on chromosomes is presented in Figure 1 b-e. The mechanism was analyzed in detail in a previous study [28]. and PR gene levels. These results provide useful information for using the Ac/Ds system to study clustered gene families in plants through the identification of the RLK functions in rice defense.

A Pair of Ds Elements Generate Diverse Chromosomal Rearrangements
The amount of various family genes that display redundant function clustering at the chromosomes hampers the study of gene function in rice. Ac/Ds transposable elements generate chromosomal rearrangements including deletions, inversions, and duplications via the alternative transposition mechanism in rice [29]. To study the function of the redundant family genes, many chromosome fragment rearrangement/deletion mutants induced by the Ac/Ds system were developed. The T-DNA provided for transposase was constructed containing the CaMV 35S promoter to drive the Ac cDNA. Another T-DNA was constructed containing a modified Ds element (Figure 1a). A schematic diagram of the mechanism of a pair of closely located Ds-induced deletion/rearrangements and homologous recombination on chromosomes is presented in Figure 1 b-e. The mechanism was analyzed in detail in a previous study [28]. The 3′ and 5′ ends from two different Ds elements were re-inserted into the OsRLG1-36 region after being cut by the Ac transposase, which is known to be the alternative transposition [29]. OsRLG1-36 homologous genes encoding receptor-like kinases clustered on the short arm of rice chromosome 1. One line, OsRLG5::DS, was isolated by screening the transformed rice plants. This line possessed a single copy Ds insertion in the promoter region of OsRLG5 (Receptor Like Kinase Gene 5) (Figure 1a), which is related to Leaf rust resistance 10 (Lr10) The 3 and 5 ends from two different Ds elements were re-inserted into the OsRLG1-36 region after being cut by the Ac transposase, which is known to be the alternative transposition [29]. OsRLG1-36 homologous genes encoding receptor-like kinases clustered on the short arm of rice chromosome 1. One line, OsRLG5::DS, was isolated by screening the transformed rice plants. This line possessed a single copy Ds insertion in the promoter region of OsRLG5 (Receptor Like Kinase Gene 5) (Figure 1a), which is related to Leaf rust resistance 10 (Lr10).

Ac/Ds-Induced RLK Deletion Mutants Exhibited Broad-Spectrum Resistance to Xoo
In a previous study, we identified a Ds element at the RLK19/RLG5 locus. We found that this RLK family consists of 36 members and contains a cluster located on chromosome 1 [29]. To test the RLK family function, two Ds elements closely located at the RLK19 locus were first generated and the large fragment deletions with one to 11 RLKs deletions were further identified. Among them, the deletion mutants with a loss of eight and 11 RLKs exhibited lesion mimic symptoms. In this study, the deletion lines were further analyzed and the #29 mutant with the RLK19-22 deletion was identified (Figure 2a). The #29 mutant displayed lesion mimic symptoms (Figure 2b). To analyze whether the #29 mutant had the characteristics of autoimmunity, mutant plants were first inoculated with five different Xoo races (PXO61, PXO71, PXO79, PXO86, PXO99). The lesion lengths were measured two weeks after the inoculation and the results demonstrated that the #29 mutant plants exhibited high and broad-spectrum resistance to Xoo (Figure 2c,d).

Ac/Ds-Induced RLK Deletion Mutants Exhibited Broad-Spectrum Resistance to Xoo
In a previous study, we identified a Ds element at the RLK19/RLG5 locus. We found that this RLK family consists of 36 members and contains a cluster located on chromosome 1 [29]. To test the RLK family function, two Ds elements closely located at the RLK19 locus were first generated and the large fragment deletions with one to 11 RLKs deletions were further identified. Among them, the deletion mutants with a loss of eight and 11 RLKs exhibited lesion mimic symptoms. In this study, the deletion lines were further analyzed and the #29 mutant with the RLK19-22 deletion was identified (Figure 2a). The #29 mutant displayed lesion mimic symptoms (Figure 2b). To analyze whether the #29 mutant had the characteristics of autoimmunity, mutant plants were first inoculated with five different Xoo races (PXO61, PXO71, PXO79, PXO86, PXO99). The lesion lengths were measured two weeks after the inoculation and the results demonstrated that the #29 mutant plants exhibited high and broad-spectrum resistance to Xoo (Figure 2c,d).

RLK20, RLK21, and RLK22 Regulate the Broad-Spectrum Resistance to Xoo in Rice
DAB staining showed that the level of H2O2 was significantly higher in the #29 mutants compared to the wild-type plants (Figure 3a). In addition, the qRT-PCR results showed that relative expression levels of PR genes (PR1a and PBZ1) were significantly induced in #29 compared to those of the wild-type (Figure 3b). To investigate which RLK regulates lesion mimic and defense against Xoo in the #29 mutants, RLK19, RLK20, RLK21, or RLK22 genes were individually expressed using the non-specific promoter 35S in the #29 mutant plants. The reconstruction of RLK20, RLK21, or RLK22 all complemented the mutant phenotype, with RLK19 proving to be the exception (Figure 3c). Leaves of the four kinds of transgenic complementary plants were inoculated with PXO86. The results showed that #29/RLK20 ox, #29/RLK21 ox, and #29/RLK22 ox plants were susceptible to PXO86, similar to the wild-type plants (Figure 3d). The #29/RLK19 ox plants, however, were resistant to PXO86. qRT-PCR was performed to detect the gene levels in the complementation plants. The qRT-PCR results demonstrated that RLK19, RLK20, RLK21, and RLK22 were highly expressed in leaves, while no transcripts were detected in the #29 2.3. RLK20, RLK21, and RLK22 Regulate the Broad-Spectrum Resistance to Xoo in Rice DAB staining showed that the level of H 2 O 2 was significantly higher in the #29 mutants compared to the wild-type plants (Figure 3a). In addition, the qRT-PCR results showed that relative expression levels of PR genes (PR1a and PBZ1) were significantly induced in #29 compared to those of the wild-type (Figure 3b). To investigate which RLK regulates lesion mimic and defense against Xoo in the #29 mutants, RLK19, RLK20, RLK21, or RLK22 genes were individually expressed using the non-specific promoter 35S in the #29 mutant plants. The reconstruction of RLK20, RLK21, or RLK22 all complemented the mutant phenotype, with RLK19 proving to be the exception (Figure 3c). Leaves of the four kinds of transgenic complementary plants were inoculated with PXO86. The results showed that #29/RLK20 ox, #29/RLK21 ox, and #29/RLK22 ox plants were susceptible to PXO86, similar to the wild-type plants (Figure 3d). The #29/RLK19 ox plants, however, were resistant to PXO86. qRT-PCR was performed to detect the gene levels in the complementation plants. The qRT-PCR results demonstrated that RLK19, RLK20, RLK21, and RLK22 were highly expressed in leaves, while no transcripts were detected in the #29 mutant plants (Figure 3e). The expression levels of PR1a and PBZ1 in #29/RLK20 ox, #29/RLK21 ox, and #29/RLK22 ox plants were similar to the wild-type plants (Figure 3f). The #29/RLK19 ox plants, however, demonstrated significantly higher expression compared to the wild-type. These results suggest that RLK19, 20, and 21 are required for the lesion mimic and bacterial resistance in the #29 mutant plants. mutant plants (Figure 3e). The expression levels of PR1a and PBZ1 in #29/RLK20 ox, #29/RLK21 ox, and #29/RLK22 ox plants were similar to the wild-type plants (Figure 3f). The #29/RLK19 ox plants, however, demonstrated significantly higher expression compared to the wild-type. These results suggest that RLK19, 20, and 21 are required for the lesion mimic and bacterial resistance in the #29 mutant plants.

RLK20 Interacts with RBOHB to Modulate ROS Generation
To analyze the function of RLKs, RLK20 was selected as bait for the isolation of the interacting proteins via a yeast-two hybrid (Y2H) screening. Y2H was performed to isolate the interacting protein using the kinase domain of RLK20. Among more than 20 interactors screened, one interactor was RBOHB (Figure 4a). Sequencing of the AD-RBOHB clone identified an RBOHB fragment that contained only the N-terminal cytosolic part of Data indicates average ± standard error (SE) (n > 6). **, p < 0.01 (f) Relative expression levels of PR1a and PBZ1 in WT, #29/RLK19, #29/RLK20, #29/RLK21, and #29/RLK22 OX plants. Data indicates the average ± standard error (SE) (n > 6). The letters a and b denote significant differences. p < 0.01.

RLK20 Interacts with RBOHB to Modulate ROS Generation
To analyze the function of RLKs, RLK20 was selected as bait for the isolation of the interacting proteins via a yeast-two hybrid (Y2H) screening. Y2H was performed to isolate the interacting protein using the kinase domain of RLK20. Among more than 20 interactors screened, one interactor was RBOHB (Figure 4a). Sequencing of the AD-RBOHB clone identified an RBOHB fragment that contained only the N-terminal cytosolic part of the protein. To further examine the RLK20 and RBOHB interaction, the full-length RBOHB and RLK20 were analyzed in the bimolecular fluorescence complementation (BiFC) system. The results showed that RLK20 and RBOHB interacted at the plasma membrane (Figure 4b).
Since RBOHB encodes NADPH oxidase, which catalyzes ROS production, the ROS levels in #29/RLK20 ox, #29 mutant, and wild-type leaves were tested using a 1µM MV treatment. The results indicated that the #29 mutant was more, while #29/RLK20 OX was less, sensitive to MV compared to wild-type plants (Figure 4c). Expression of RBOHB induced H 2 O 2 accumulation, while co-expression of RBOHB and RLK20 reduced RBOHB-promoted H 2 O 2 accumulation (Figure 4d). These results indicated that RLK20 interacts with and inhibits RBOHB to reduce ROS generation. the protein. To further examine the RLK20 and RBOHB interaction, the full-le RBOHB and RLK20 were analyzed in the bimolecular fluorescence complementa (BiFC) system. The results showed that RLK20 and RBOHB interacted at the pla membrane (Figure 4b). Since RBOHB encodes NADPH oxidase, which catalyzes production, the ROS levels in #29/RLK20 ox, #29 mutant, and wild-type leaves were te using a 1µM MV treatment. The results indicated that the #29 mutant was more, w #29/RLK20 OX was less, sensitive to MV compared to wild-type plants (Figure 4c) pression of RBOHB induced H2O2 accumulation, while co-expression of RBOHB RLK20 reduced RBOHB-promoted H2O2 accumulation (Figure 4d). These results cated that RLK20 interacts with and inhibits RBOHB to reduce ROS generation.

Discussion
Rice bacterial blight seriously threatens yield production [1]. Resistance breedi an economically and eco-friendly way to protect crops from disease. A single resist gene can easily induce genetic mutations in the pathogen due to the widespread grow regions of rice varieties, leading to the loss of resistance in rice varieties. The discove resistance genes and functional studies are the basis of durable disease control. The genome sequencing data demonstrate that 29% of the genes are predicted to be organ in clustered gene families (International Rice Genome Sequencing Project 2005), posi challenge in the examination of the functions of the gene families. Furthermore, f tional dissection or annotation of these clustered gene families may be of significanc use in future breeding.
Previously, we identified rice plants with eight receptor-like gene (RLK19-26) d

Discussion
Rice bacterial blight seriously threatens yield production [1]. Resistance breeding is an economically and eco-friendly way to protect crops from disease. A single resistance gene can easily induce genetic mutations in the pathogen due to the widespread growing regions of rice varieties, leading to the loss of resistance in rice varieties. The discovery of resistance genes and functional studies are the basis of durable disease control. The rice genome sequencing data demonstrate that 29% of the genes are predicted to be organized in clustered gene families (International Rice Genome Sequencing Project 2005), posing a challenge in the examination of the functions of the gene families. Furthermore, functional dissection or annotation of these clustered gene families may be of significance for use in future breeding.
Previously, we identified rice plants with eight receptor-like gene (RLK19-26) deletions (from a clustered gene family consisting of 36 RLKs), by a pair of closely located Ds transposable elements, exhibiting lesion mimic symptoms [26]. To further investigate which RLK was responsible for the lesion mimic symptoms, more deletion lines were isolated to identify the chromosomal regions. Eventually, the #29 deletion mutant showed the lesion mimic symptom was isolated. Similar to other lesion mimic mutants reported [31][32][33][34][35], the growth of #29 was seriously inhibited. Further analysis identified that four RLKs (RLK19-22) were deleted in the #29 mutant. Furthermore, RLK19 shared a 39% sequence similarity with barley Lr10 [29], a leaf rust-resistant gene, implying its potential function in plant defense. The inoculation of five different races of Xoo strains demonstrated that the #29 plants were broad-spectrum resistant mutants. These results are similar to a previous report that identifies the broad-spectrum resistance symptoms of lesion mimic mutants [34,36]. These mutants have characteristics of chlorophyll degradation, H 2 O 2 accumulation and apoptosis, which affect growth and development [33,34,36,37]. Furthermore, the individual RLK functions in the #29 mutants were analyzed by the overexpression of each RLK in the #29 background to ensure that the lesion mimic phenotype was caused by the loss of RLKs. Reconstruction of RLKs rescued the lesion mimic phenotype of the #29 mutant, while the RLK19 did not rescue the mutant phenotype, indicating that three RLKs (RLK20-22) negatively regulate rice broad-spectrum resistance.
A previous study reported that PR gene expression levels and H 2 O 2 content are significantly higher in the lesion mimic mutant compared to wild-type plants [34,38,39]. Similar to other lesion mimic mutants, PR genes (PBZ1 and PR1b) expression levels and H 2 O 2 content was significantly higher in the #29 mutants compared to wild-type plants. Interestingly, further yeast-two hybrid screening using RLK20 kinase domain as bait, identified that RBOHB (a ROS biogenesis enzyme) interacts with RLK20. MV (ROS inducer) treatment and subsequent DAB staining results showed that the #29 mutant contained more, while #29/RLK20 ox contained less H 2 O 2 compared to wild-type plants. These results suggest that RLK20 may interact with RBOHB to inhibit its function. To further confirm this hypothesis, RLK20 and RBOHB were expressed in tobacco leaves, and the H 2 O 2 level was monitored. The results showed that RLK20 expression inhibited RBOHB-mediated ROS production, implying that RLK20 may phosphorylate RBOHB to inhibit ROS generation. Previous studies demonstrated that the calcium-dependent protein kinase (CDPK) and Rac/ROP small GTPase Rac1 interact with RBOHB to activate ROS production [40,41]. RLKs may inhibit CDPK or Rac1 binding to RBOHB to reduce ROS production. However, further studies are required to explore the function of RLKs in ROS production. Alternatively, the loss of RLK20-22 may activate RBOHB-mediated ROS production to highly accumulate the ROS, by which cell death and subsequent lesion mimic symptom were produced. Further studies are required to elucidate this issue.
Isolation and utilization of resistance-related genes is an efficient way to control the disease. However, technical limitations and the complexity of the plant genome made the resistant gene isolation task even more difficult. Diverse genetic approaches have been developed, which significantly accelerate the speed of functional genomic analysis. Due to the functional redundancy, elucidation of the clustered family gene functions is still challenging. The current study analyzed the clustered RLK family functions by Ds-induced chromosomal deletions and proposed the potential for using Ac/Ds-induced deletions as a tool for future investigations of clustered gene family functions. Taken together, we identified the RLK functions in rice broad-spectrum resistance providing target genes for future resistant cultivar breeding.

Tissue Culture Regeneration and Transgenic Plant Generation
The Ac and Ds gene trap cassettes were developed according to previously published methodology [42]. Ac and Ds elements were cloned into a T-DNA vector pSB11 (Figure 1a) and transformed into LBA4404 cells [29]. The tissue culture regeneration was conducted according to previously described methods [31,43]. Briefly, seeds were hulled and sterilized with 0.6% H 2 O 2 . Tissue culture media were used to produce plantlets. The regenerated plants were transplanted into bottles with solid 0.5× MS medium. The plants were then transferred to the greenhouse at 28 • C.
For the functional validation of each of the RLK19, 20, 21, and 22 genes, the entire ORFs were cloned and connected to the pCAMBIA1381-Ubi vector. The recombination vectors were then transformed into the rice cultivar Dongjin via Agrobacterium tumefaciens-mediated methods [44]. Primers designed by Primer Premier 5 for genes cloning were as follows in Table 1:

Rice Cultivation and Xoo Inoculation
Top second leaves of two-month-old rice plant were inoculated with five Xoo strains (PXO61, PXO71, PXO79, PXO86, PXO99, from Zhejiang Academy of Agricultural Sciences) using the leaf-cutting method according to a previous report [45]. The inoculated plants were stored on a plastic-covered shelf to keep moist for three days. To activate the Xoo strains before inoculation, the strains were inoculated on a potato semisynthetic agar (PSA) plate at 28 • C. The mature colony was further inoculated in 20 mL fluid PSA medium in a 50 mL centrifuge tube, shaken at 220 rpm until the bacterial suspension was ready for use when the OD 600 = 0.5-1.0. The length of the disease spots from the top second leaves from six different plants was measured two weeks after the inoculation.

Determination of H 2 O 2 Content
The entire ORF of RLK20 and RBOHB were separately cloned into the pCAMBIA1302. RBOHB and RLK20 + RBOHB were next injected into tobacco leaf via A. tumefaciensmediated transformation. The tobacco leaves from different treatment groups were treated with 1 µM MV (Sigma, St Louis, USA) or sterile water (negative control) for 4 h. The leaves treated with MV were rinsed with sterile water, then the excess moisture on the leaves was absorbed by filter paper. The hydrogen peroxide levels of the leaves were determined according to the previously described methods [46]. Leaf tissues (1 g) were homogenized in an ice bath with 10 mL 0.1% (w/v) trichloroacetic acid (TCA). The homogenate was first centrifuged at 12,000× g for 15 min and then 1 mL of the supernatant was removed to 1 mL of 10 mM potassium phosphate buffer (pH 7.0) and 2 mL of 1 M (w/v) KI. The blank control consisted of 0.1% TCA without leaf extract. After the reaction was developed for one hour in darkness, the absorbance of the supernatant was measured at 390 nm by spectrophotometer. The content of H 2 O 2 was calculated using a standard curve prepared with known concentrations of H 2 O 2 .

qRT-PCR Analysis
Rice leaves were collected for total RNA extraction using TRIZol reagent (Takara, Dalian, China). For qRT-PCR, RNA was reverse-transcribed to cDNA using the PrimeScript RT reagent Kit (Takara, Dalian, China) and the relative expression levels of different genes were detected using Ssofast EvaGreen Supermix (BIO-RAD, Hercules, CA, USA) with Mx3005P (Agilent, Palo Alto, CA, USA) [47]. Three technical replicates for each sample in the experiment were performed. Ubiquitin was used as an internal reference gene. All qRT-PCR primers in Table 2.

DAB Staining Assay
To detect H 2 O 2 levels, rice leaves were treated with 1 µM MV or sterile water (negative control) for 24 h and stained with diaminobenzene (DAB) according to a previously published method [48]. The leaves were cut into 2 cm pieces and placed in DAB solution and incubated in a growth chamber with 25 • C for 8 h. The leaf sections were examined by light microscopy. The areas where H 2 O 2 production occurred were reddish-brown.

Yeast Two-Hybrid Screening
The yeast two-hybrid screening was conducted according to the Library Construction and Screening Kits (Clontech, Dalian, China) instructions. The RLK20 kinase domain sequences were cloned into pGBKT7 as a bait vector. RLK20-pGBKT7 and cDNA-pGADT7-Rec or lam-pGADT7 (negative control) were co-transformed into Y2H gold yeast strain for library screening. Yeast transformants were grown on synthetic dropout-Leu-Trp-His-Ade plates.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.