The CC‐NB‐LRR OsRLR1 mediates rice disease resistance through interaction with OsWRKY19

Summary

Nucleotide‐binding site–leucine‐rich repeat (NB‐LRR) resistance proteins are critical for plant resistance to pathogens; however, their mechanism of activation and signal transduction is still not well understood. We identified a mutation in an as yet uncharacterized rice coiled‐coil (CC)‐NB‐LRR, Oryza sativa RPM1‐like resistance gene 1 (OsRLR1), which leads to hypersensitive response (HR)‐like lesions on the leaf blade and broad‐range resistance to the fungal pathogen Pyricularia oryzae (syn. Magnaporthe oryzae) and the bacterial pathogen Xanthomonas oryzae pv. oryzae, together with strong growth reduction. Consistently, OsRLR1‐overexpression lines showed enhanced resistance to both pathogens. Moreover, we found that OsRLR1 mediates the defence response through direct interaction in the nucleus with the transcription factor OsWRKY19. Down‐regulation of OsWRKY19 in the rlr1 mutant compromised the HR‐like phenotype and resistance response, and largely restored plant growth. OsWRKY19 binds to the promoter of OsPR10 to activate the defence response. Taken together, our data highlight the role of a new residue involved in the NB‐LRR activation mechanism, allowing identification of a new NB‐LRR downstream signalling pathway.

Introduction

Plants have developed sophisticated and inducible defence mechanisms against pathogen attack that involves recognition of and response to pathogen molecules at the plant cell surface or within the cell by host receptors. The plant pattern‐recognition receptors recognize conserved microbe‐associated molecular patterns (MAMPs) at the cell surface to stimulate MAMP‐triggered immunity (MTI), which functions as the basal defence response in plants (Jones and Dangl, 2006; Tena et al., 2011). To interfere with plant MTI, pathogens produce virulence effectors that can be recognized directly or indirectly by a host resistance (R) protein. Successful recognition between pathogen effector and R protein can trigger effector‐triggered immunity (ETI) in plants to prevent the forward colonization of pathogens (Dodds and Rathjen, 2010; Jones and Dangl, 2006). Activation of ETI often causes strong accumulation of reactive oxygen species (ROS) and cell death at infected sites, termed hypersensitive response (HR). The activation of both MTI and ETI can result in activation of systemic acquired resistance (SAR), which enhances the overall degree of resistance to plant diseases (Jones and Dangl, 2006).

The classic ‘gene‐for‐gene’ concept of ETI in plants (HammondKosack and Jones, 1997) proposes a receptor–ligand model with a higher degree of complexity, where a plant R protein recognizes an avirulence (Avr) protein directly or indirectly by detecting effector‐mediated modification of host proteins to initiate the defence response (Mackey et al., 2002; Mestre and Baulcombe, 2006). R proteins that contain a nucleotide‐binding domain (NB) and apoptosis APAF‐1 and CED‐4 conserved domain (ARC) and C‐terminal leucine‐rich repeats (LRR) are known as NB‐LRR receptors, hereafter NLRs. NLRs can be divided into two broad classes on the basis of their N‐terminal domains, which shows homology to the Drosophila Toll protein and mammal interleukin‐1 receptor, and is known as Toll/interleukin‐1 receptor (TIR) or be a coiled‐coil (CC) domain (Monosi et al., 2004). These are termed TIR‐NB‐LRR (TLRs) and CC‐NB‐LRR (CLRs), respectively. Generally, the CC domain of CLRs is required for association with accessory proteins, which is vital for downstream signalling (Lukasik and Takken, 2009; Swiderski et al., 2009). The CC domain can activate a defence response cascade through association with WRKY transcription factors. Over‐ or ectopic‐expression of this domain alone often causes localized cell death (Cheng et al., 2015; Hu et al., 2017; Inoue et al., 2013). The NB‐ARC domain predominantly participates in nucleotide binding, and strong conformational changes in this domain are associated with exchange between ADP and ATP (Takken et al., 2006; Tameling et al., 2002). Many conserved motifs have been identified in this region. The P‐loop (Walker A or kinase 1a motif) and Walker B are two conserved motifs of P‐loop NTPases (Tameling et al., 2006), mutation of which commonly causes loss of function of the NLR protein (Lolle et al., 2017). Mutation of other conserved motifs, such as RNBS‐A, RNBS‐D and MHD, may cause effector‐independent conformational changes that result in autoactivation or loss of function of NLR proteins (Mestre and Baulcombe, 2006; Tameling et al., 2006; Zhang et al., 2003). The LRR domain is relatively diverse and predominantly functions in the recognition of effectors (Shen et al., 2003).

In the absence of effectors, the NLRs are in an ADP‐bound state in which intra‐molecular interactions maintain the proteins inactive; in the presence of effectors, conformational changes associated with the ATP‐bound state lead to disruption of the repressive intra‐molecular interactions and exposition of the N‐terminal domain (CC or TIR), allowing remodelling of signal complexes (Mestre and Baulcombe, 2006; Sukarta et al., 2016). Autoimmunity caused by mutation of NLRs often causes stunted growth and development of plants, accumulation of reactive oxygen species, elevated expression of defence‐related genes and lesion‐like cell death in the leaves (De Oliveira et al., 2016; Grant et al., 2000; Zhang et al., 2003). In‐depth analysis showed that dimerization of invariant CLRs in a CC domain‐dependent manner is a minimal functional unit in immune signalling (Maekawa et al., 2011). Forward studies revealed that the CC or TIR domains of NLRs formed a homo‐ or heteromeric complex that was associated with a chaperone and one or more WRKYs simultaneously to activate the downstream signal (Maekawa et al., 2011; Shirasu, 2009; Tran et al., 2017).

NLRs participate in defence responses to multiple pathogens, such as fungi, bacteria, viruses and insects. In rice, although considerable numbers of NLR genes have been identified and studied to elucidate the mechanism of immunity mainly against pathogens such as Pyricularia oryzae (syn. Magnaporthe oryzae), Xanthomonas oryzae pv. oryzae (Xoo) and Rhizoctonia solani, previous studies have typically focused on resistance to a single pathogen, especially the fungal pathogen P. oryzae, and only a limited number of reports have concentrated on two or more pathogens in rice (Chen and Ronald, 2011; Li et al., 2019). Associated with one of the most devastating diseases in cultivated rice, the rice blast, P. oryzae infects epidermal cells by generating specialized appressorium and spreads biotrophically in plant cells (Wilson and Talbot, 2009). X. oryzae are agents of bacterial blight and leaf streak in rice, possessing type III secretion system for self multiplication and promotion of pathogenicity in host plants (Timilsina et al., 2020). The fungal pathogen R. solani, the agent of the rice sheath blight, developed a necrotrophical means of colonizing and spreading, leading to considerable research on the pathogenesis involved (Molla et al., 2020). Among the cloned NLRs in rice, the most intensively studied protein is BPH14, which is associated with two transcription factors, OsWRKY46 and OsWRKY72, which bind to the promoter of RLCK281 and a callose synthase gene to mediate insect resistance (Hu et al., 2017). In the present study, we identified a CLR protein, named OsRLR1 due to its high homology to RPM1 in Arabidopsis thaliana. RPM1, first identified in 1995, conducts recognition of two effectors: avrRpm1 and avrB from the bacteria Pseudomonas syringae (Grant et al., 1995; Mackey et al., 2002). Mutation and overexpression of OsRLR1 conferred resistance to the fungal pathogen P. oryzae and the bacterial pathogen Xoo. OsRLR1 associated with the transcription factor OsWRKY19 to regulate defence in rice.

Results

Point mutation in the NLR OsRLR1 induces spontaneous HR‐like lesions

From a rice mutant library subjected to EMS treatment, we identified a recessive mutant, rlr1 (previously designated lmes2), that developed HR‐like lesions on the leaf blade with an early senescence phenotype in the absence of pathogen infection (Xing et al., 2016). We previously mapped the mutation in a 66‐kb interval on chromosome 10 (Xing et al., 2016). We sequenced all genes in this interval and found a single‐base transversion (A to T) at the 953 bp site of the coding sequence in LOC_Os10g07978, causing an amino acid change from Glu to Val at position 318 in the protein (Figure 1a). The single exon of LOC_Os10g07978 encodes a 923 amino acid protein. A BLASTP search of the NCBI database (https://blast.ncbi.nlm.nih.gov/Blast.cgi) revealed that the most similar homolog of LOC_Os10g07978 is the CLR from Arabidopsis thaliana RPM1 (33% identity and 52% similarity). Thus, LOC_Os10g07978 was a putative homolog of RPM1 and the gene was designated Oryza sativa RPM1‐like resistance gene 1 (OsRLR1). As RPM1, OsRLR1 is a typical CLR protein with N‐terminal CC domains (32–165), NB domain (178–330), ARC domain (331–523) and the C‐terminal LRR domain (551–923). The mutation was located in the NB domain and caused a change from the hydrophilic, polar negatively charged amino acid Glu to the nonpolar hydrophobic amino acid Val (Figure 1b).

Figure 1.

Point mutation in the CLR OsRLR1 leads to broad‐range resistance against pathogens. (a) Genes in previously mapped lmes2 region and the mutation site of OsRLR1; blue box represents the OsRLR1 open reading frame, and black line represents the mutation site. (b) Schematic diagram of OsRLR1 protein structure. (c) Relative expression level of OsRLR1 in WT, rlr1 mutant and complementation lines (COM; rlr1 mutant line transformed with ProUbi:OsRLR1); values represent means ± SD (n = 3). L1 and L2 represent individual COM lines. (d) Images of plants at tillering stage. Bar = 5 cm. (e) Leaves of WT (1); rlr1 mutant (2); and COM (3). (f) Diseased leaves at 10 days after P. oryzae inoculation. (g) Lesion length at 10 days after P. oryzae inoculation (n = 30). (h) Lesion number of the whole leaf blade at 10 days after P. oryzae inoculation (n = 30). (i) Diseased leaves at 15 days after Xoo inoculation. (j) Lesion length at 15 days after Xoo inoculation (n = 15). (k) Xoo population in inoculated leaves at 15 days after inoculation (n = 3×5). (l) Sequence alignment of OsRLR1 and other CLR proteins; the mutated amino acid is highlighted in blue. (m) Co‐localization of OsRLR1‐GFP with a nucleus‐localized RFP in rice protoplast; free GFP was used as control (upper row). (n) Immunodetection of OsRLR1‐GFP after cell fractionation of N. benthamiana leaves. *P < 0.05; **P < 0.01 (Student’s t‐test).

To verify that OsRLR1 is the gene causing the phenotype, we cloned the coding sequence of LOC_Os10g07978 from the wild type (WT) and overexpressed it in rlr1 plants. We sequenced the mutated site in the WT, rlr1 mutant and T₀ transgenic plants. Only the transgenic plants were polymorphic at the mutation site (both A and T), which confirms that the transformation was successful (Figure S1a). The transcript level of OsRLR1 was much higher in the homozygous T₂ transgenic plants than in the WT or rlr1 mutant plants (Figure 1c). The transgenic plants recovered a phenotype similar to that of the WT (Figure 1d‐e). At the same time, overexpressing mutated OsRLR1 (E318V, thereafter called OsRLR1_M) in the WT showed the HR‐like phenotype on the leaf blade like rlr1 (Figure S1b‐c). We also measured the transcript levels of pathogenesis‐related (PR) genes OsPR1a and OsPR10, two marker genes of immunity activation. The expression levels of the two genes were elevated significantly (Figure S1d). These results show that the mutation in OsRLR1 (LOC_Os10g07978) accounted for the rlr1 phenotype. Considering the role of CLRs in immunity, we analysed the disease resistance/susceptibility of the WT, rlr1 mutant and the transgenic lines. At 10 dpi for P. oryzae, the lesion length and lesion number in the rlr1 leaf blades were considerably shorter and lower, respectively, than that seen in the WT and transgenic plants (Figure 1f‐h). Similarly, at 15 dpi for Xoo, the lesion length and bacterial population were considerably reduced in the rlr1 mutant compared with the WT and transgenic plants (Figure 1i‐k). In contrast, no notable difference was observed between the WT and rlr1 mutant plants in the lesions caused by the fungal R. solani (Figure S1e, f).

To explore the function of OsRLR1, its basal expression pattern and subcellular localization were analysed. Results of qRT‐PCR show that OsRLR1 is mainly expressed in leaf blades (Figure S1g). A 3484 bp sequence of the OsRLR1 promoter was cloned in transcriptional fusion with the GUS gene. GUS staining of ProOsRLR1:GUS transgenic plants shows that OsRLR1 was mainly expressed in the vascular bundles of the leaf blade (Figure S1h, i). When OsRLR1‐GFP and nuclear marker‐RFP were co‐expressed in rice protoplasts, the green and red fluorescent protein (RFP) signals were superimposed (Figure 1m). Western blot showed that the OsRLR1‐GFP fusion protein was in the nuclear fraction after cell fractionation (Figure 1n). These results indicate that OsRLR1 exerts its biological function in the nucleus.

A multiple sequence alignment of OsRLR1 and reported CLRs from rice, Arabidopsis, tomato, potato, maize and wheat was generated (accession numbers are listed in Table S2). The mutated amino acid was located close to a relatively conserved motif, RNBS‐B/sensor 1, in the NB domain (Figure 1l). Nineteen CLRs from an additional 10 species were included in the alignment for phylogenetic analysis. OsRLR1 was clustered with the rice blast resistance protein PID3, A. thaliana RPM1 and ZmRPM1 (Figure S1j). Taken together, these results suggest that OsRLR1 may function as a typical CLR gene involved in the disease resistance pathway in rice.

OsRLR1 positively regulates defence responses

To examine the role of OsRLR1 in the defence response, we overexpressed OsRLR1 in the WT. Three OsRLR1 overexpression homozygous transgenic T₂ (OsRLR1‐OX) lines were obtained. At the tillering stage, the expression level of OsRLR1 in the three OsRLR1‐OX lines was more than 100 times that of the WT (Figure 2a). At this stage, OsRLR1‐OX plant growth was similar to the WT. In contrast to the rlr1 mutant line, no HR‐like lesions were observed on the leaf blades of the OsRLR1‐OX lines (Figure S2a‐f). Moreover, overexpression of OsRLR1 had no significant influence on the agronomic traits of the OsRLR1‐OX lines (Figure S2g).

Figure 2.

OsRLR1 positively regulates the defence responses. (a) Relative expression level of OsRLR1 in WT, rlr1 mutant and OsRLR1‐overexpressing lines in the WT background (OsRLR1‐OX); values represent means ± SD (n = 3). (b) Diseased leaves at 10 days post‐inoculation (dpi) with P. oryzae. (c) Lesion length of diseased leaves at 10 dpi with P. oryzae (n = 30). (d) Lesion number of the whole leaf blade at 10 dpi with P. oryzae (n = 30). (e) Diseased leaves at 15 dpi with Xoo. (f) Lesion length of diseased leaves at 5, 10 and 15 dpi with Xoo (n = 15). (g) Bacterial population in inoculated leaves at 15 dpi with Xoo (n = 3×5). *P < 0.05; **P < 0.01 (Student’s t‐test).

We inoculated the WT, rlr1 mutant and the three OsRLR1‐OX lines with P. oryzae. The lesion length and lesion number at 10 dpi in the OsRLR1‐OX plants were significantly reduced compared with the WT (Figure 2b–d). The same lines were also inoculated with Xoo. The lesion length was measured at 5, 10 and 15 dpi. The OsRLR1‐OX lines developed a shorter lesion than the WT at all time points (Figure 2e,f). Consistently, the bacterial population in the WT was larger than that in the OsRLR1‐OX lines at 15 dpi (Figure 2g). Therefore, OsRLR1 positively regulates defence resistance in rice.

Leaf blades were sampled at 0, 3, 12, 24, 48, 72 and 96 hours post‐inoculation (hpi) with P. oryzae. At all time points, the expression level of OsRLR1 in the transgenic lines was much higher than that in the WT, which confirmed that the transformation was successful (Figure 3a). We also measured the transcript levels of OsNPR1, an immunity regulator and the PR genes OsPR1a and OsPR10. The expression levels of the three genes were significantly higher in the OsRLR1‐OX and in the rlr1 mutant plants than in the WT at almost all time points (Figure 3a). These results indicate that an enhanced defence response was activated by P. oryzae infection both in the rlr1 mutant lines and in OsRLR1‐OX lines. A similar trend was observed at 5 and 10 dpi with Xoo (Figure 3b). Taken together, these results show that OsRLR1 positively regulates defence response in rice.

Figure 3.

Defence‐related gene expression in OsRLR1‐OX lines after P. oryzae and Xoo inoculation. (a) Relative expression level of OsRLR1, OsNPR1, OsPR1a and OsPR10 in the WT, rlr1 mutant and three OsRLR1‐OX lines at 0, 3, 12, 24, 48, 72 and 96 h post‐inoculation (hpi) with P. oryzae; values represent means ± SD (n = 3). (b) Relative expression level of OsNPR1, OsPR1a and OsPR10 in WT, rlr1 and three OsRLR1‐OX lines at 5 and 10 dpi with Xoo; values represent means ± SD (n = 3). *P < 0.05; **P < 0.01 (Student’s t‐test).

OsRLR1 interacts with OsWRKY19

Given that CLRs often function in the defence response in association with WRKY transcription factors, five defence‐related WRKY proteins (OsWRKY13, OsWRKY19, OsWRKY47, OsWRKY68 and OsWRKY76) were screened for ability to interact with the CC domain of OsRLR1 (OsRLR1_CC) using a yeast two‐hybrid system. The full‐length WRKY proteins often auto‐activate the Y2H system (Inoue et al., 2013); hence, truncated proteins were used in this study. All co‐transformed yeasts grew on DDO medium (Figure 4a left), whereas only OsWRKY19 interacted with OsRLR1_CC to allow growth on QDO/AbA medium (Figure 4a right). For further characterization of the interaction, truncated and full‐length OsRLR1 and OsRLR1_M proteins were each expressed with OsWRKY19 into yeast. All co‐transformed yeasts grew on DDO medium (Figure 4b left). All truncated OsRLR1 and OsRLR1_M lacking the LRR domain, but not the full‐length OsRLR1 nor OsRLR1_M, interacted with OsWRKY19 (Figure 4b right). In addition, in the Y2H system, the OsRLR1 CC domain, but not the NB, ARC, or LRR domains, formed self‐association (Figure S3).

Figure 4.

OsRLR1 interacts with OsWRKY19. (a) Yeast two‐hybrid (Y2H) assay between OsRLR1 CC domain and several rice WRKY proteins. (b) Y2H assay between truncated OsRLR1 and OsWRKY19 proteins. (c) Bimolecular fluorescence complementation (BiFC) between OsRLR1‐cYFP or OsRLR1_M‐cYFP and OsWRKY19‐nYFP in Nicotiana benthamiana leaves; nYFP and cYFP alone were used as negative controls. (d) Co‐IP of OsRLR1‐Myc with OsWRKY19‐GFP (left) or OsRLR1_M‐Myc with OsWRKY19‐GFP (right). (e‐f) Split luciferase assays. OsRLR1‐Cluc (e) or OsRLR1_M‐Cluc (f) was co‐expressed with OsWRKY19‐Nluc in N. benthamiana leaves. 1: OsRLR1‐Cluc or OsRLR1_M‐Cluc with OsWRKY19‐Nluc; 2: OsWRKY19‐Nluc with Cluc; 3: OsRLR1‐Cluc or OsRLR1_M‐Cluc with Nluc; and 4: Cluc with Nluc. Cluc and Nluc alone were used as negative controls.

In rice protoplasts, the GFP signal of the OsWRKY19‐GFP fusion protein co‐localized with the RFP signal of the nuclear marker protein, while the free GFP was observed both in cytoplasm and nucleus (Figure S5a). The OsWRKY19‐GFP fusion protein was detected by Western blot in the nuclear fraction after cell fractionation (Figure S5b). In Nicotiana benthamiana leaves, the GFP fluorescence of OsRLR1‐GFP, OsRLR1_M‐GFP and OsWRKY19‐GFP was also localized in the nucleus as labelled by DAPI (Figure S5c). We next verified the interaction between OsRLR1 and OsWRKY19 in planta using a bimolecular fluorescence complementation (BiFC) assay. The YFP fluorescence was observed in the nucleus of N. benthamiana leaves that co‐expressed OsRLR1‐cYFP and OsWRKY19‐nYFP (Figure 4c). Similar YFP fluorescence in the nucleus was observed when OsRLR1_M‐cYFP and OsWRKY19‐nYFP were co‐expressed (Figure 4c). A co‐immunoprecipitation assay (Figure 4d) and a split LUC assay (Figure 4e,f) confirmed that OsRLR1 and OsRLR1_M interacted with OsWRKY19 in N. benthamiana leaves. Thus, both OsRLR1 and OsRLR1_M interacted with OsWRKY19 in the nucleus.

The results described above suggest that OsWRKY19 participates in the OsRLR1‐mediated defence response. Consistent with this conclusion, we found that OsWRKY19 was up‐regulated in OsRLR1‐OX and in rlr1 plants at almost all sampling time points after inoculation with P. oryzae (Figure S6).

OsWRKY19 mediates OsRLR1‐induced defence responses

To explore further the role of OsWRKY19 in the OsRLR1‐mediated resistance response, we used RNAi to silence OsWRKY19 in rlr1. The OsWRKY19 expression level in all T₀ transgenic lines was less than that in rlr1 (Figure 5a). The phenotype of the rlr1 OsWRKY19‐silenced plants (rlr1 OsWRKY19_RNAi) was assessed in T₂ plants. At the tillering stage, compared with rlr1 plants, the growth and development of rlr1 OsWRKY19_RNAi plants were better, although not comparable to WT plants (Figure 5b–d). The first, second and third youngest leaves of rlr1 OsWRKY19_RNAi plants developed fewer HR‐like lesions than rlr1 plants (Figure 5e–g). The DAB staining of the second and third leaves indicates that H₂O₂ accumulation was reduced in rlr1 OsWRKY19_RNAi plants compared with rlr1 plants, while almost no H₂O₂ was detected in the WT (Figure 5h). At the maturity stage, rlr1 OsWRKY19_RNAi plants were considerably taller and developed a greater number of tillers (Figure 5i), longer ears and more grains (Figure 5j) than rlr1 plants. Agronomic trait analysis of the rlr1 OsWRKY19_RNAi line that showed the best development reveals that all traits were significantly increased compared with the rlr1 mutant although not restored to the level of the WT (Figure S7a‐h). Thus, the growth and development of rlr1 OsWRKY19_RNAi plants were partially restored.

Figure 5.

Spontaneous HR‐like lesion phenotype of the rlr1 mutant is compromised by silencing OsWRKY19. (a) Relative expression of OsWRKY19 in rlr1 WRKY19_RNAi T₀ plants; values represent means ± SD (n = 3); L1‐L4 are individual transgenic lines. (b–d) WT, rlr1 mutant and homozygous T₂ rlr1 WRKY19_RNAi plants at the tillering stage. (e‐g) Leaves of WT, rlr1 and rlr1 OsWRKY19_RNAi plants; 1, 2 and 3 represent the first, second and third youngest leaves. (h) DAB staining of leaves of WT, rlr1 and rlr1 OsWRKY19_RNAi plants; 2 and 3 represent the second and third leaves from the shoot apex. (i) WT, rlr1 and rlr1 WRKY19_RNAi plants at the maturity stage. (j) Panicles of WT, rlr1 and rlr1 WRKY19_RNAi plants at the maturity stage.

To determine whether silencing of OsWRKY19 compromised the OsRLR1_M induced resistance to pathogens, WT, rlr1 and rlr1 OsWRKY19_RNAi plants were inoculated with P. oryzae at the four‐leaf stage. The rlr1 OsWRKY19_RNAi plants developed longer lesion length and larger lesion number compared with the rlr1 mutant (Figure 6a,b,c). Similar results were observed after Xoo inoculation at the tillering stage (Figure 6d,e,f). Thus, down‐regulation of OsWRKY19 in the rlr1 mutant weakened the OsRLR1_M ‐induced resistance to pathogens.

Figure 6.

OsRLR1_M‐induced resistance to P. oryzae and X. oryzae is compromised in silencing OsWRKY19. (a) Diseased leaves at 10 days post‐inoculation (dpi) with P. oryzae. (b) Lesion length of diseased leaves at 10 dpi with P. oryzae (n = 30). (c) Lesion number of the whole leaf blade at 10 days after P. oryzae inoculation (n = 30). (d) Diseased leaves at 15 dpi with Xoo. (e) Lesion length of diseased leaves at 5, 10 and 15 dpi with Xoo (n = 15). (f) Bacterial population in inoculated leaves at 15 dpi with Xoo (n = 3×5). (g) Relative expression level of OsPR1a and OsPR10 in WT, rlr1 mutant and rlr1 WRKY19_RNAi plants at 0, 3, 12, 24, 48, 72 and 96 hpi with P. oryzae; values represent means ± SD (n = 3). (h) Relative expression level of OsPR1a and OsPR10 in WT, rlr1 and rlr1 WRKY19_RNAi plants at 5 dpi with Xoo. L1 and L2 represent individual lines of WRKY19_RNAi in rlr1. *P < 0.05; **P < 0.01 (Student’s t‐test).

OsWRKY19 expression level in P. oryzae‐inoculated leaves of two rlr1 OsWRKY19_RNAi lines was significantly reduced compared with rlr1 plants at all sampling time points, which shows that OsWRKY19 silencing is maintained during P. oryzae infection (Figure S8a). In general, OsRLR1 expression level in rlr1 OsWRKY19_RNAi lines was reduced, but not at all time points (Figure S8b). OsRLR1 expression is induced during the infection by P. oryzae (Figure 3a). It is also induced in the rlr1 mutant, likely because a positive feedback regulation. It is thus coherent that its expression is decreased in the rlr1 OsWRKY19_RNAi lines, in which signalling downstream OsRLR1 is decreased. Similarly, OsNPR1 expression level was lower only at some time points after inoculation with P. oryzae and Xoo (Figure S8c,d). In contrast, the expression levels of OsPR1a and OsPR10 in rlr1 OsWRKY19_RNAi plants were significantly decreased compared with those in rlr1 plants at all time points (Figure 6g). Similar results were observed at 5 dpi with Xoo (Figure 6h). Thus, OsWRKY19 contributes to OsRLR1‐mediated defence responses in rice.

OsWRKY19 activates expression of OsPR10

To test the transcription activity of OsWRKY19, we used the luciferase (LUC) reporter system in rice protoplasts and found that expressing OsWRKY19‐GAL4 BD enhanced LUC activity by a factor of more than 5 compared with the empty vector pGreenII800, which indicates that OsWRKY19 functions as an active transcription factor (Figure 7a). Rice protoplasts that expressed OsWRKY19 with ProOsPR10:LUC demonstrated enhanced LUC activity by a factor of 6 compared with the control transformed with ProOsPR10:LUC and the empty vector (Figure 7b), which indicates that OsWRKY19 can activate the expression of OsPR10. To verify whether OsWRKY19 can directly activate the expression of OsPR10, we analysed the promoter region of OsPR10 using the PlantCARE database and detected four potential elements: three W‐boxes and one WLE 1 (Figure 7c). CHIP assay was conducted in vivo using the WT and rlr1 leaf blades at the seedling stage (Figure 7d,e). The results showed that for the sample IP with OsWRKY19 antibodies from WT there was slight enrichment of the PCR products using primers of sites P2 and P4, and significant enrichment of the PCR product using primers of site P3 (Figure 7d). In the rlr1 mutant, sample IP with OsWRKY19 antibodies showed significant enrichment of the PCR products at sites P2, P3 and P4 (Figure 7e). No PCR product in sample IP with OsWRKY19 antibodies was detected using primers of site P1 in the WT or rlr1 mutant (Figure 7d,e). Taken together, these results show that OsWRKY19 may directly activate expression of OsPR10. Transactivation of ProOsPR10:LUC by OsWRKY19 was also observed in N. benthamiana leaves. Interestingly, LUC activity was induced by co‐expression of OsRLR1 and even more by co‐expression of OsRLR1_M (Figure 7f). These results suggest that interaction with OsRLR1 facilitates the function of OsWRKY19 in the activation of the OsPR10 gene. These results allow to propose a model of the defence response signal conduction from OsRLR1 to OsWRKY19 to OsPR10 (Figure 7g).

Figure 7.

OsWRKY19 controls OsPR10 expression. (a) Transcription factor activity assay. The empty vector Pro35S:GAL, and the positive control Pro35S:GAL‐VP16 or Pro35S:OsWRKY19‐GAL construct were introduced in rice protoplasts. LUC activity was measured 12 h after protoplast transformation. (b) Transactivation assay in rice protoplasts. LUC activity was measured 12 h after transformation with the Pro35S:OsWRKY19 and ProOsPR10:LUC constructs. (c) Distribution of potential binding sites in the promoter of OsPR10. Blue bars indicate the DNA fragments used for the CHIP assays; red lines indicate the W‐box element and violet line indicates the WLE1 element. (d‐e) CHIP‐qPCR for P1‐P4 sites of OsPR10 promoter in wild type (d) and the rlr1 mutant (e) was conducted using samples before IP (input) and after IP with anti‐OsWRKY19 polyclonal antibodies or pre‐immune IgG (control); values represent means ± SD (n = 3). The presented qPCR results are ratios relative to the value of the IP sample with IgG. f: Transactivation assays in N. benthamiana leaves. LUC activity was measured 2 days after transformation with the ProOsPR10:LUC, Pro35S:OsWRKY19, Pro35S:OsRLR1 and/or Pro35S:OsRLR1_M . (g): Proposed model for OsRLR1 and OsWRKY19 mediated defence response in rice. Minus (‐) and plus (+) denote the presence or absence of the indicated constructs. *P < 0.05; **P < 0.01 (Student’s t‐test).

Discussion

OsRLR1 is a typical CLR protein in rice

Resistance proteins are widely used by plants and mammals to aid recognition and response to pathogen attack by monitoring the presence of corresponding Avr proteins (Roberts et al., 2013). The activation of R proteins upon recognition of an intruding pathogen commonly induces cell death at the intrusion site to disable pathogen growth and halt infection. Thus, in the absence of pathogens, R genes must be tightly and precisely regulated to avoid inappropriate activation. In Arabidopsis, RPM1 is located on the peripheral plasma membrane and is repressed by a plasma membrane‐localized protein, RIN4 (Boyes et al., 1998; Mackey et al., 2002). Phosphorylation of RIN4 caused by effectors from the bacterial pathogen Pseudomonas syringae can release RPM1 to trigger the ETI response (Mackey et al., 2002). In the present study, we identified a rice CLR gene, OsRLR1, homologous to RPM1, which is expressed predominantly in vascular bundles of leaves. Unlike RPM1, OsRLR1 is localized in the nucleus. Similarly to most CLR, for example MLA10 in barley (Maekawa et al., 2011) and BPH14 in rice (Hu et al., 2017), OsRLR1 likely self‐associates through its CC domain. Although OsRLR1 cannot perceive the presence of the available Avr proteins directly (Figure S4), it cannot be excluded that OsRLR1 may recognize effector in an indirect way. Point mutation and overexpression of OsRLR1 increase resistance to the bacterial pathogen Xoo and fungal hemibiotroph pathogen P. oryzae. Interestingly, it does not increase resistance or susceptibility to the fungal necrotrophic pathogen R. solani, suggesting that OsRLR1 controls specific defence responses leading to resistance to biotrophic/ hemibiotrophic pathogens. Alternatively, we cannot exclude that OsRLR1‐mediated immunity is active only in leaf blades and not in leaf sheaths.

Mutation of E318 in the NB domain leads to autoactivation of OsRLR1

Mutation in the RNBS‐D domain in AtLOV1 (P437L) and RPM1 (P442L) causes loss of function of the proteins (Sweat et al., 2008), whereas mutation in StRx (E442L, close to RNBS‐D) enables autoactivation of the protein (Bendahmane et al., 2002). The rlr1 mutant phenotype, including the leaf‐tip necrosis, the strong ROS production and expression of PR genes in absence of pathogens as well as the ability of OsRLR1_M to induce OsWRKY19 transcriptional activity in heterologous systems suggest that OsRLR1_M codes for an autoactive form of CLR. Mutation in the NB domain of two CLRs of tomato (Solanum lycopersicum), I‐2 and Mi‐1, impaired ATP hydrolysis and caused autoactivation (Tameling et al., 2002; Tameling et al., 2006). Modelling of the three‐dimensional OsRLR1 structure revealed that the mutated site E318 is on the other face of the NB domain compared with the nucleotide‐binding site (Figure S9c,d,e). Interestingly, the corresponding residue in AtZAR1, the only full‐length plant NLR that has been crystallized to date (Wang et al., 2019), is located outside the protein in the ADP‐bound state (Figure S9a) and becomes extremely close to an α‐helix of the ARC domain (Figure S9b) due to protein rearrangement, including a rotation of the NB domain in the ATP‐bound state. Thus, the E318 or neighbouring residues of OsRLR1 may be involved in the stability of the protein in the ATP‐bound state. The mutation OsRLR1_M might lock the tertiary structure corresponding to the activated state by altering the interaction between the NB and the ARC domains. Therefore, we speculate that the mutation in OsRLR1 might influence ATP hydrolysis either directly or indirectly. Many mutations in residues of the NB domain were found to lead to autoactivation of NLR (Figure S9, Table S3) (Bendahmane et al., 2002; Tameling et al., 2006; Tang et al., 2011; Van Ooijen et al., 2008); however, no previous study has reported that a mutation close to the RNBS‐B motif may cause autoactivation of an NLR. Thus, we have identified a new site that may be critical in NLR functioning.

OsRLR1 positively regulates the defence response in rice

NPR1 is vital in plant resistance. Mutation of NPR1 in Arabidopsis abolishes expression of PR genes and completely compromises activation of SAR (Cao et al., 1997), while overexpression of which enhanced pathogen resistance in a dose‐dependent manner (Cao et al., 1998). Overexpression of OsNPR1 or NH1 spontaneously activates the defence response and enables enhanced resistance to Xoo (Chern et al., 2005). The higher expression level of OsNPR1 and PR genes in both the rlr1 and in the OsRLR1‐OX plants, compared with the WT plants, suggests that OsRLR1 controls OsNPR1 expression and consequently the OsNPR1‐mediated defence response in rice. Expression of PR genes induces cell death to prevent pathogen growth at the invasion site (Sels et al., 2008; Loon et al., 2006). In the rlr1 mutant, OsPR1a and OsPR10 expressions were highly induced. This might account for massive cell death observed in the rlr1 mutant. Moreover, lower level of OsPR1a and OsPR10 expression at 0 hpi in OsRLR1‐OX lines than in the rlr1 mutant line (Figure 3a,b) might explain the absence of HR‐like lesions in OsRLR1‐OX lines compared with the rlr1 mutant line. Previous studies have shown that some NLR‐mediated resistance is dosage‐dependent (Howles et al., 2005) and overexpression of NLR genes may result in autoactivity (Tao et al., 2000). In the present study, OsRLR1 overexpression in rice enhanced the defence response and led to broad‐range resistance but caused neither an HR‐like phenotype nor any effect on agronomic traits, which may encourage its potential use in enhancing disease resistance in rice breeding.

OsRLR1 functions with OsWRKY19 to regulate the defence response

WRKY transcription factors are specific to plants and algae, and play a variable role in developmental processes and biotic and abiotic stress responses (Cheng et al., 2015; Devaiah et al., 2007; Eulgem et al., 2000). In plants, WRKY transcription factors often activate expression of PR genes by recognizing and binding to W‐box or WLE 1 cis‐regulatory elements (Choi et al., 2015; Hu et al., 2017; Inoue et al., 2013). In rice, many WRKY proteins involved in pathogen resistance have been identified. The group II OsWRKY62 is a negative regulator of rice resistance to Xoo, and its overexpression suppresses resistance by inhibiting the expression of defence‐related genes (Peng et al., 2008). In contrast, the group III OsWRKY45 positively regulates defence responses to P. oryzae and Xoo but not to R. solani (Shimono et al., 2012). The NLR protein Pb1 mediates blast resistance in rice through interaction with OsWRKY45 by the CC domain (Inoue et al., 2013). In the present study, we found that OsRLR1 interacts with OsWRKY19 through the CC domain. As OsWRKY45, OsWRKY19 belongs to group III of the WRKY family (Wang et al., 2015; Xie et al., 2005), which are predominantly responsive to pathogens and/or salicylic acid (Kalde et al., 2003). OsWRKY19, activating the transcription of the PR gene, is a positive regulator of the defence response induced by OsRLR1. Down‐regulation of OsWRKY19 partially compromised the OsRLR1_M‐mediated resistance to P. oryzae and Xoo and partially recovered the OsRLR1_M‐mediated defective development. The partial recovery may be attributable to residual OsWRKY19 expression in the OsWRKY19‐silenced lines or to interaction of OsRLR1 with other WRKY transcription factors that would have a redundant function in the defence response.

In conclusion, we have characterized a new rice CLR, OsRLR1 and at least part of its downstream signalling pathway leading to defence activation. Moreover, we have identified a new residue E318 in the NB domain that might be critical for NLR function. Finally, we have shown that overexpression of OsRLR1 leads to broad‐range resistance with almost no penalty in terms of yield in rice.

Methods

Plant materials and growth conditions

The rlr1 mutant was identified from a mutant library of the rice indica cultivar ‘Jinhui 10’ (JH10) generated by ethyl methanesulphonate (EMS) treatment. For agronomic trait measurement, rice plants were grown in experimental fields in Chongqing, China, under natural conditions. For pathogen inoculation assays, rice plants were grown in a growth chamber (26 °C and under high‐pressure sodium lamp) at Southwest University. All N. benthamiana plants were grown in a greenhouse at the Rice Research Institute of Southwest University, Chongqing, China.

Pathogen inoculation and symptom assessment

At the four‐leaf stage, plants in the growth chamber were spray‐inoculated with a suspension of P. oryzae spores from three races (ZB7, ZB11 and ZB31) in 0.01% Tween‐20 (1 × 10⁵ spores/mL). Spores were obtained from P. oryzae races cultured on autoclaved sorghum seeds. At different time points after inoculation, leaves were excised for total RNA extraction. The length of diseased leaves and number of lesions on the leaf blade with the most serious disease symptoms of each plant were measured at 10 days post‐inoculation (dpi).

At about 30 days old, plants grown in chamber were inoculated with the Xoo strain zhe173. Preparation of suspension and inoculation was performed as described previously (Peng et al., 2008). After 5, 10 and 15 dpi, the lesion length was measured and bacterial growth was determined by counting colony‐forming units (CFU) as described previously (Peng et al., 2008). At booting stage, plants grown in field were inoculated with Rhizoctonia solani (RH‐9) as described previously (Jia et al., 2013).

Histological analysis

Staining with 3,3‐diaminobenzidine (DAB) (Suolaibao) was conducted to detect H₂O₂ accumulation in leaves as described previously with slight modifications (Choi et al., 2015). Briefly, leaves were detached and placed in DAB solution (10 mm MES, 1 mg/mL DAB) for 16 h, then cleared in alcohol. The leaves were examined with a stereomicroscope (SZX16/SZX10; Olympus, Japan). At the seedling stage, leaves of T₀ plants transformed with ProOsRLR1:GUS were stained for about 12 h at 37 °C in β‐glucuronidase (GUS) solution (Sigma). The stained leaf samples were used for paraffin sectioning. Sections were imaged using a Nikon Eclipse E600 microscope (Nikon, Japan).

Rice transformation

To generate the complementation and overexpression construct, the coding sequence of OsRLR1 was amplified from JH10 and cloned into the digested pCA1300 vector containing a maize Ubiquitin promoter. Likewise, the coding sequence of OsRLR1_M was amplified from rlr1 and cloned into pCA1300 vector with the same method. A 3484 bp (before start codon) promoter sequence of OsRLR1 was amplified from JH10 and ligated to the digested pCA1300 vector to drive GUS expression. All primers used are listed in Table S1.

Transient expression assays in rice protoplasts

The vector pAN580 harbouring GFP under the control of 2 × CaMV35S promoter was used for analyses of OsRLR1 and OsWRKY19 subcellular localizations in the rice protoplasts. The rice protoplasts were isolated following a method described previously (Ma et al., 2017). The GFP and RFP fluorescence signals were observed using a confocal microscope (LSM800; Zeiss, Germany). Transient expression of the LUC reporter was achieved via the same method. The Pro35S:Renilla‐LUC activity was measured as a control to normalize LUC activity among samples using a luminometer (GloMax; Promega, USA).

Yeast two‐hybrid assay

Yeast two‐hybrid assays were conducted with the Y2H Gold yeast strain using the GAL4 two‐hybrid system (Clontech) following the manufacturer’s instructions. The pGADT7 and pGBKT7 vectors were used for plasmid constructions. Double dropout plates lacking Trp and Leu were used to select co‐transformed colonies. Protein interactions were detected by measuring the growth of colonies on QDO plates lacking Trp, Leu, Ade and His but containing AbA (100 ng/mL).

Transient expression in Nicotiana benthamiana leaves

Agrobacterium tumefaciens strain GV3101 containing the target plasmid was grown in YEB medium with antibiotic selection to OD₆₀₀ = 0.6–0.8. Cells were suspended to an OD₆₀₀ = 0.4 in MES buffer (10 mm MgCl₂, 10 mm MES; pH 5.6) and kept in the dark for 2–4 h before inoculation. Leaves were analysed at 2–3 days after transformation.

Bimolecular fluorescence complementation assays

OsRLR1 and OsRLR1_M (the mutated form of OsRLR1) were fused with the sequence coding the C‐terminal fragment of the yellow fluorescent protein (cYFP) in the pxy104 vector. OsWRKY19 was fused with the sequence coding the N‐terminal portion of the yellow fluorescent protein (nYFP) in the pxy106 vector. A. tumefaciens strains containing the different plasmids were mixed prior to inoculation in N. benthamiana leaves. Leaves were stained with DAPI (Sigma) 1 h before examination with a confocal microscope.

Cell fractionation

A plant nuclear and cytoplasmic protein extraction kit (BestBio) was used to extract nuclear and cytoplasmic fractions from N. benthamiana leaves expressing OsWRKY19‐GFP or OsRLR1‐GFP, following the protocol provided by the manufacturer. The rabbit anti‐GFP antibodies (Abcam) and anti‐Rabbit IgG light‐chain antibodies (Abcam) were used to detect the fusion proteins by Western blot.

Split luciferase assay

The 2 × CaMV35S promoter was used for expressing OsRLR1‐Cluc or OsRLR1_M‐Cluc with OsWRKY19‐Nluc in N. benthamiana, respectively. The split luciferase assay was conducted as described before (Chen et al., 2008). 1 mm luciferin was sprayed to the inoculated leaves before LUC activity was captured by CCD imaging camera (Lumina Series III; IVIS, USA).

Chromatin immunoprecipitation analysis

A ChIP assay was conducted using the EpiQuik Plant ChIP kit (EpiGentek) following the protocol provided by manufacturer with slight improvement. Leaf blades of the wild type and rlr1 mutant were used for the ChIP assay. The rabbit anti‐OsWRKY19 polyclonal antibodies (the specific C‐terminal 175–277 amino acid was purified from E.coli and injected into rabbit to get the antibodies) and pre‐immune IgG (control) were used in protein/DNA immunoprecipitation. The total genomic DNA extract (input) and purified DNA eluent after IP were used in a qPCR assay.

Co‐immunoprecipitation assay

Pro2 × 35S:OsRLR1‐Myc or Pro2 × 35S:OsRLR1_M ‐Myc was co‐expressed with Pro2 × 35S:OsWRKY19‐GFP and P19 (a silencing inhibitor) in N. benthamiana leaves. Proteins were extracted from the agroinfiltrated N. benthamiana leaves with the extraction buffer (140 mm NaCl, 2.7 mm KCl, 25 mm Na₂HPO₄, 1.5 mm KH₂PO₄, 0.01 mm EDTA and 0.05% NP‐40) supplemented with a protease inhibitor cocktail, PMSF (0.5 mm) and DTT (5 mm). 100 μL of extract after centrifugation at 13 000 rpm for 30 min was used as input and the remainder was incubated with anti‐mouse Myc antibodies (Abcam: Cambridgeshire, England, UK) or anti‐mouse IgG antibodies (Abcam) at 4 °C overnight. The protein and antibody mixture was then incubated with agarose protein A beads at 4 °C for 4 h. The suspension was centrifuged at 1000g for 5 min, and the coimmunoprecipitate mixture was washed four times. The protein in precipitate and input were separated by SDS‐PAGE and analysed by Western blot using GFP and Myc antibodies.

Conflicts of interest

Authors declare no conflict of interest.

Author contributions

Guanghua He and Dan Du designed the research. Dan Du and Xin Lu performed the subcellular location, and BiFC and Y2H analysis. Yadi Xing performed the map‐based cloning and phenotype analysis. Yadi Xing performed the expression analysis of OsRLR1. Changwei Zhang designed the pathogen inoculation analysis. Linjun Cai, Han Yun and Dan Du performed the pathogen inoculation analysis. Qiuli Zhang, Yingying Zhang and Mingming Liu performed agronomy analysis. Dan Du, Yadi Xing and Xinlong Chen performed the vector construction. Guanghua He and Xianchun Sang performed the EMS study and identified the mutant. Yinghua Ling and Zhenglin Yang performed the planting of the materials in the field. Yunfeng Li performed protein alignment analysis. Benoit Lefebvre performed 3D structure construction of OsRLR1 and revision of the manuscript. Dan Du and Changwei Zhang wrote the article.

Supporting information

Figure S1 Expression analysis of OsRLR1.

Figure S2 Phenotype and agronomic trait analysis of OsRLR1‐OX plants.

Figure S3 Yeast two‐hybrid assay between the same domains of OsRLR1.

Figure S4 Yeast two‐hybrid assay between the AVR protein in P. oryzae and OsRLR1.

Figure S5 Localization of OsWRKY19 and OsRLR1 in the nucleus.

Figure S6 Relative expression level of OsWRKY19.

Figure S7 Agronomic traits in the wild type (WT), rlr1 mutant, and WRKY19_RNAi plants.

Figure S8 Relative expression level of defence related genes after P. oryzae and Xoo inoculation.

Figure S9 Modelling of OsRLR1 3D protein structure.

Table S1 Primers used in this study.

Table S2 CLR proteins used in multiple sequence alignment.

Table S3 Mutated sites in NBS domain of other NLRs.

Du, D. , Zhang, C. , Xing, Y. , Lu, X. , Cai, L. , Yun, H. , Zhang, Q. , Zhang, Y. , Chen, X. , Liu, M. , Sang, X. , Ling, Y. , Yang, Z. , Li, Y. , Lefebvre, B. and He, G. (2021) The CC‐NB‐LRR OsRLR1 mediates rice disease resistance through interaction with OsWRKY19. Plant Biotechnol J, 10.1111/pbi.13530