A tripartite rheostat controls self-regulated host plant resistance to insects

Plants deploy receptor-like kinases and nucleotide-binding leucine-rich repeat receptors to confer host plant resistance (HPR) to herbivores1. These gene-for-gene interactions between insects and their hosts have been proposed for more than 50 years2. However, the molecular and cellular mechanisms that underlie HPR have been elusive, as the identity and sensing mechanisms of insect avirulence effectors have remained unknown. Here we identify an insect salivary protein perceived by a plant immune receptor. The BPH14-interacting salivary protein (BISP) from the brown planthopper (Nilaparvata lugens Stål) is secreted into rice (Oryza sativa) during feeding. In susceptible plants, BISP targets O. satvia RLCK185 (OsRLCK185; hereafter Os is used to denote O. satvia-related proteins or genes) to suppress basal defences. In resistant plants, the nucleotide-binding leucine-rich repeat receptor BPH14 directly binds BISP to activate HPR. Constitutive activation of Bph14-mediated immunity is detrimental to plant growth and productivity. The fine-tuning of Bph14-mediated HPR is achieved through direct binding of BISP and BPH14 to the selective autophagy cargo receptor OsNBR1, which delivers BISP to OsATG8 for degradation. Autophagy therefore controls BISP levels. In Bph14 plants, autophagy restores cellular homeostasis by downregulating HPR when feeding by brown planthoppers ceases. We identify an insect saliva protein sensed by a plant immune receptor and discover a three-way interaction system that offers opportunities for developing high-yield, insect-resistant crops.

Plants deploy receptor-like kinases and nucleotide-binding leucine-rich repeat receptors to confer host plant resistance (HPR) to herbivores 1 . These gene-for-gene interactions between insects and their hosts have been proposed for more than 50 years 2 . However, the molecular and cellular mechanisms that underlie HPR have been elusive, as the identity and sensing mechanisms of insect avirulence effectors have remained unknown. Here we identify an insect salivary protein perceived by a plant immune receptor. The BPH14-interacting salivary protein (BISP) from the brown planthopper (Nilaparvata lugens Stål) is secreted into rice (Oryza sativa) during feeding. In susceptible plants, BISP targets O. satvia RLCK185 (OsRLCK185; hereafter Os is used to denote O. satvia-related proteins or genes) to suppress basal defences. In resistant plants, the nucleotide-binding leucine-rich repeat receptor BPH14 directly binds BISP to activate HPR. Constitutive activation of Bph14-mediated immunity is detrimental to plant growth and productivity. The fine-tuning of Bph14-mediated HPR is achieved through direct binding of BISP and BPH14 to the selective autophagy cargo receptor OsNBR1, which delivers BISP to OsATG8 for degradation. Autophagy therefore controls BISP levels. In Bph14 plants, autophagy restores cellular homeostasis by downregulating HPR when feeding by brown planthoppers ceases. We identify an insect saliva protein sensed by a plant immune receptor and discover a three-way interaction system that offers opportunities for developing high-yield, insect-resistant crops.
During the millions of years of plant and insect co-evolution, herbivorous insects have developed diverse feeding strategies to retrieve nutrients from host plants. Such insects can cause up to 18% of annual losses of global crop yield 3 . Because plants are sessile and cannot escape continuous attack by pests, plants have evolved various defence strategies to combat herbivory. Unlike microbial pathogens, insects can move swiftly on or among plants as they feed, making it challenging to understand the mechanisms that underlie their interactions.
Herbivorous insects actively select their feeding sites and secrete saliva during feeding to facilitate nutrient acquisition from host plants 1,4,5 . Some salivary proteins induce host defences 1,5 . In the past decade, plant genes that encode nucleotide-binding leucine-rich repeat (NLR) receptors that confer resistance to piercing and sucking insects have been isolated through map-based cloning [6][7][8][9][10][11][12] . NLR receptors function as intracellular immune receptors that directly or indirectly detect cognate effectors 13,14 and activate effector-triggered immunity (ETI) [6][7][8][9][10] , also known as HPR. The identities of the insect avirulence effectors that trigger ETI have not yet been discovered. Hence, how a host plant recognizes the insect and deploys and modulates resistance remains largely unknown.
The brown planthopper (BPH; N. lugens Stål) is the most destructive insect pest that threatens rice production globally 15 . Using its stylet mouthparts to probe and penetrate plant cells, it consumes phloem sap and eventually causes plant death and substantial yield losses [16][17][18] . BPHs migrate long distances, crossing national boundaries in search of better host plants [15][16][17][18] , which makes regional control of BPH challenging. For this reason, HPR has been crucial for BPH control. The first BPH-resistance gene isolated in rice, Bph14, encodes a coiled-coiled nucleotide-binding leucine-rich repeat (CC-NB-LRR) protein that forms a homomeric complex and interacts with WRKY transcription factors 19 . BPH14 enhances WRKY activity to confer resistance to BPH 19 .
Here we show that the BPH salivary protein BISP is directly recognized by BPH14. Notably, the perception of BISP not only triggers Bph14-mediated resistance but also activates NBR1-mediated selective autophagy. This results in the degradation of BISP to restrict hyperactivation of Bph14-regulated defences. We elucidate the molecular and Individual data points are plotted. c-e, BISP interacted with BPH14 in Y2H (c), co-IP (d) and BiFC (e) assays. N14 and BISP(125-241) served as negative controls. DDO, SD/-Leu-Trp; QDO, SD/-Leu-Trp-His-Ade. bZIP63-CFP, nuclear marker. f, Quantification of relative YFP intensities in BiFC assays. Data are the mean ± s.d.
(n = numbers of biologically independent cells). g, Amino acid sequence of BISP. The asterisk indicates the stop codon. Glycine residues are marked in red. The 13-amino acid and 4-amino acid tandem repeats are underlined in blue and black, respectively. h,i, Immunohistochemical localization of BISP in female BPH salivary glands using pre-immune rabbit serum (h) or anti-BISP antibodies (i). Red fluorescence (Cy3) and blue fluorescence show the localization of BISP and DAPI-stained nuclei, respectively. j, BISP was delivered into rice leaf sheaths during BPH infestation. Leaf sheath proteins were analysed using anti-BISP antibodies. Ponceau S staining served as the loading control. k, Immunohistochemical staining showing BISP in BPH-infested rice leaf sheaths. The non-infected (middle) and BPH-infested (bottom) sheaths were detected by anti-BISP antibodies and pre-immune rabbit serum, respectively, served as negative controls. In b and f, P values were derived by one-way analysis of variance (ANOVA). Experiments (a-e,h-k) were repeated at least three times, each giving similar results. The results of the other two repeats are presented in Supplementary Fig. 2. Scale bars, 5 μm (e), 25 μm (k), 100 μm (h,i) or 10 cm (a).
between BISP and BPH14 and their nucleocytoplasmic co-localization in rice protoplasts (Fig. 1e,f and Extended Data Fig. 1e). Bisp (LOC111051577) resides on chromosome 12 of BPH and encodes a 241-amino acid protein with an amino-terminal signal peptide (amino acids  and no transmembrane domain (Fig. 1g), which suggests that BISP is a secretory protein. Rich in glycine residues (32%), BISP has three 13-amino acid (GLADVGGLTNTVS) and 4-amino acid (CGGN/R) tandem repeats in its carboxy-terminal region (Fig. 1g). Quantitative PCR with reverse transcription (RT-qPCR) confirmed that Bisp is highly expressed in salivary glands and in whole female adults (Extended Data Fig. 1f,g). Immunohistochemical analysis using BISP antiserum revealed that BISP accumulated to higher levels in female salivary glands than in male salivary glands and guts (Fig. 1h,i and Extended Data Fig. 1h-m,p), a result that was consistent with Bisp RNA levels. Moreover, BISP signals were higher than the salivary protein NlSP1 (Extended Data Fig. 1n-p).
BPHs were allowed to feed on rice plants to determine whether BISP is secreted into rice leaf sheaths. Immunoblots detected BISP in leaf sheath protein extracts from BPH-infested but not non-infested plants (Fig. 1j). Immunolocalization assays verified the delivery of BISP into rice leaves. BISP was detected along the penetration path of stylets within the leaf sheaths (Fig. 1k). Immunogold electron microscopy further confirmed that BISP was secreted into rice tissues during BPH feeding (Extended Data Fig. 1q-t).

BISP suppresses plant defence
To determine whether BISP influenced the success of BPH on rice, we disrupted Bisp expression by microinjecting Bisp double-stranded RNAs (dsRNAs) into individual BPHs to induce RNA-mediated interference (RNAi). Bisp transcripts and proteins were significantly reduced in BPHs that received Bisp-RNAi compared with those that received GFP-RNAi (Extended Data Fig. 2a). Bisp-RNAi insects exhibited lower weight gain and honeydew excretion (Extended Data Fig. 2b,c). These insects also experienced significantly higher mortality when fed on the susceptible N14 rice plants than BPHs that received no injection (controls) or GFP-RNAi (Extended Data Fig. 2d). Therefore, Bisp plays a crucial role in the feeding and performance of BPHs on susceptible rice plants.
To determine the function of BISP in rice, we ectopically expressed BISP (without its signal peptide) in BPH-susceptible N14 rice (N14-Bisp) plants (Fig. 2a). The stature of N14-Bisp plants was similar to control N14 plants (Extended Data Fig. 3a,b). However, N14-Bisp plants were hypersensitive to BPHs, displayed more severe symptoms after BPH feeding and died more quickly than the BPH-infested N14 plants (Fig. 2b,c). Moreover, the weight gain and honeydew excretion of BPHs feeding on N14-Bisp plants were significantly higher than those feeding on N14 plants (Extended Data Fig. 3c,d). A two-host choice test showed that BPHs preferentially settled on N14-Bisp plants rather than N14 plants (Extended Data Fig. 3e). These results show that BISP increases the susceptibility of rice to BPHs.
We reasoned that the BISP effector was acting as a virulence factor that suppresses rice defence responses, which makes susceptible plants more vulnerable to BPHs. The endogenous levels of free salicylic acid (SA), a plant hormone vital for BPH resistance in rice 10,11,19 , were lower in N14-Bisp plants than in N14 plants (Fig. 2d). Similarly, the levels of SA biosynthesis and signalling genes (OsICS1 and OsNPR1) and downstream defence-related genes (OsPR1a, OsPR1b, OsPR5 and OsPR10) transcripts were all significantly lower in N14-Bisp plants than in N14 plants (Extended Data Fig. 3f).
Receptor-like kinases (RLKs) play crucial roles in plant defence against pathogens and herbivores 1,20 . We aimed to identify the target of BISP and to elucidate the molecular mechanism that underlies its ability to suppress defence responses in susceptible plants. To that end, we examined the interactions of RLKs, receptor-like cytoplasmic kinases (RLCKs) and MAP kinases with BISP in Y2H assays ( Fig. 2e and Extended Data Fig. 4a-c). Among the eight kinases tested, only OsRLCK185 interacted with BISP. BISP interacted with the OsRLCK185 kinase domain (86-353 amino acids). The association of OsRLCK185 with BISP was confirmed by co-IP in rice protoplasts (Fig. 2f). As OsRLCK185 regulates plant immunity through autophosphorylation 21,22 , an in vitro phosphorylation assay and immunoblotting with phosphoserine/ phosphothreonine-specific antibodies confirmed that OsRLCK185 was an active kinase with autophosphorylation activity (Extended Data Fig. 4d). Furthermore, OsRLCK185 autophosphorylation was attenuated by BISP ( Fig. 2g and Extended Data Fig. 4d).
We used three OsRLCK185 knockout rice lines (Osrlck185-1-Osrlck185-3 plants) to further explore the roles of OsRLCK185 in rice defence against BPHs (Extended Data Fig. 4e). Osrlck185 lines were     and OsRLCK159 served as negative controls. g, OsRLCK185 autophosphorylation activity was reduced by BISP. BISP(26-124)-GST-His proteins served as the negative control. Kinase activity was detected by autoradiography and input proteins shown after Coomassie Brilliant Blue staining (CBB). h,i, Phenotypes (h) and BPH-resistance scores (i) of Osrlck185 and WT plants after 4 days of BPH infestation. n = 45 plants examined over 3 independent experiments. In box plots in c and i, the box limits indicate the 25th and 75th percentiles, the whiskers indicate the full range of the data and the centre line indicates the median. Individual data points are plotted. In c,d and i, P values were derived by one-way ANOVA. The experiments (a,b,d-h) were repeated at least three times, each giving similar results. Scale bars, 10 cm (b,h). Article more susceptible to BPHs, as they were severely damaged compared with wild-type (WT) plants 4 days after BPH infestation (Fig. 2h,i). BPHs that fed on the Osrlck185 lines gained significantly more weight and excreted significantly more honeydew than those fed on WT plants (Extended Data Fig. 4f,g), which indicated that OsRLCK185 positively regulates basal immunity in BPH-susceptible plants.
To further assess the function of BISP in BPH, we analysed the performance of BPHs treated with Bisp-RNAi or GFP-RNAi on Osrlck185-1 and WT plants. Bisp-RNAi insects showed significantly lower weight gain, honeydew excretion and survival rate than GFP-RNAi insects on Osrlck185-1 plants (Extended Data Fig. 2e-g). This result suggests that BISP has a specific function related to OsRLCK185 and may target other defence-related proteins in rice.
BPH14 forms a homomeric complex and interacts with WRKY72 to increase WRKY72 accumulation, which in turn activates downstream defence signalling 19 . To examine the effects of BISP on the formation of BPH14 complexes and on WRKY72 protein levels, we co-expressed BPH14 and BISP in rice protoplasts. The levels of BPH14 self-association ( Fig. 3f) and WRKY72 levels ( Fig. 3g) were increased when BPH14 and BISP were co-expressed in rice protoplasts, which suggested that BISP stimulated the activity of BPH14.
To examine the activation of BPH14 by BISP, we constructed transgenic Bph14-Bisp rice lines in which Bisp was constitutively expressed in the Bph14 background (Extended Data Fig. 5g). WRKY72 protein and transcript levels were higher in three homozygous Bph14-Bisp plants relative to Bph14 plants (Fig. 3h,i). In addition, non-infested Bph14-Bisp plants had higher levels of free SA than Bph14 plants (Fig. 3j). Furthermore, the SA-responsive OsICS1 and OsNPR1 and four rice PR genes were significantly upregulated (Extended Data Fig. 5h). The intensity of Bph14-mediated resistance responses in the Bph14-Bisp plants positively correlated with the levels of BISP protein.
We evaluated BPH resistance in Bph14-Bisp plants. Whereas Bph14 plants showed little damage after 7 days of infestation (Fig. 1a), prolonged BPH feeding (14 days) caused substantial damage. By contrast, prolonged BPH feeding caused little damage to Bph14-Bisp plants (Fig. 3k,l), which suggested that ectopic expression of Bisp in Bph14 plants enhances resistance to BPHs. In addition, BPH adults preferred Bph14 over Bph14-Bisp plants in two-host choice assays and had poor performance on Bph14-Bisp plants based on BPH weight gain, honeydew excretion and survival ( Fig. 3m and Extended Data Fig. 5i-k).
We further tested the performance of Bisp-RNAi and GFP-RNAi insects on Bph14 plants and susceptible control MH63 plants (Extended Data Fig. 2h-j). The GFP-RNAi and non-injected insects exhibited significantly decreased weight gain, honeydew excretion and survival on Bph14 plants compared with those on MH63 plants. By contrast, the Bisp-RNAi insects showed similar weight gain, honeydew excretion and survival rate when fed on Bph14 or MH63 plants. Moreover, the Bisp-RNAi insects exhibited more weight gain, honeydew excretion and a higher survival rate than GFP-RNAi insects on Bph14 plants. These results show that the knockdown of Bisp improved BPH performance on Bph14 plants.
Constitutive activation of immune responses to pathogens usually negatively affects plant growth 23,24 . We observed that ectopic expression of Bisp in the Bph14 background had significant fitness costs. Bph14-Bisp plants were smaller than Bph14 plants at both the seedling and heading stages (Fig. 3n,o and Extended Data Fig. 6a,b). Furthermore, the heading dates of Bph14-Bisp plants were advanced (Extended Data Fig. 6c). Finally, Bph14-Bisp plants exhibited poor agronomic traits, with significantly lower yields than Bph14 plants (Extended Data Fig. 6d-j). The degree of fitness costs in Bph14-Bisp plants positively correlated with the levels of BISP and the intensity of Bph14-mediated resistance responses (Fig. 3h-m and Extended Data Figs. 5g-k and 6). These results indicate that constitutively expressed BISP induces BPH14-mediated resistance in Bph14-Bisp plants, which has a substantial fitness cost in rice.

BPH14 mediates BISP turnover by autophagy
The fitness costs imposed by ectopic expression of BISP in Bph14 plants suggested that the activation of Bph14-mediated resistance should be tightly controlled in the natural habitats of rice. Several lines of evidence support this hypothesis. First, when BISP is expressed in rice protoplasts, BISP levels were lower when BISP was co-expressed with BPH14 ( Fig. 4a), and BISP levels negatively correlated with BPH14 levels (Extended Data Fig. 7a). Indicative of the specificity of the BISP-BPH14 interaction, BISP levels were not reduced when BISP was co-expressed with N14 (Fig. 4a). These results indicate that BISP is degraded in a BPH14-dependent manner.
The ubiquitin-proteasome system and autophagy are the main protein-degradation pathways in plant cells 25 . The ubiquitin-proteasome system is a key post-translational mechanism for controlling plant immune responses to pathogens 26 . To test whether the ubiquitinproteasome system is involved in the modulation of BISP-BPH14 immune activation, the 26S proteasome inhibitor MG132 was added to protoplasts expressing BISP and BPH14. MG132 blocked the proteasome pathway degradation of WRKY72 (Extended Data Fig. 7b), but did not inhibit the degradation of BISP (Fig. 4b). By contrast, treatment with four autophagy inhibitors blocked the degradation of BISP ( Fig. 4c and Extended Data Fig. 7c), which suggested that the autophagy pathway is involved in the degradation of BISP.
To further assess the role of autophagy in BISP turnover, we used cyan fluorescent protein-tagged ATG8F (CFP-ATG8F) as a marker 27,28 to monitor autophagy (Extended Data Fig. 7d,e). When CFP-ATG8F was expressed with BISP or BPH14 individually in Nicotiana benthamiana leaves (hereafter Nb is used to denote N. benthamiana-related proteins or genes), few punctate CFP structures, corresponding to pre-autophagosomes and autophagosomes, were detected. However, when CFP-ATG8F was expressed with BISP and BPH14, the number of punctate CFP structures significantly increased, which indicated that BISP and BPH14 co-expression induced autophagy. This interaction was specific, as expressing CFP-ATG8F with BISP and N14 did not increase the number of autophagosomes. Concanamycin A (ConA) blocks autophagic flux 28 , and the numbers of autophagic bodies were substantially increased in cells that expressed both BISP and BPH14 and treated with ConA. When we silenced the core autophagy genes NbATG6, NbPI3K and NbATG7, the numbers of autophagic bodies were markedly decreased in the silenced leaves that expressed CFP-ATG8F with BISP-yellow fluorescent protein (YFP) and BPH14 (Extended Data Fig. 7f-h). Furthermore, using transmission electron microscopy, we observed more double-membraned autophagosome structures in the phloem companion cells of non-infested Bph14-Bisp plants than in Bph14 and MH63 plants (Fig. 4d,e). Accordingly, OsATG8 protein levels and OsATG8a, OsATG8b and OsATG8c transcript levels were significantly higher in non-infested Bph14-Bisp plants than in Bph14 and MH63 plants (Extended Data Fig. 7i,j). These observations indicate that BISP is degraded through autophagy in a BPH14-dependent manner.
As BPHs secrete saliva containing BISP into rice tissues when feeding (Fig. 1j,k and Extended Data Fig. 1q-t), we monitored BISP levels in infested BPH-susceptible (MH63) or BPH-resistant (Bph14) plants. In MH63 plants, BISP was first detected at 6 h and increased continuously for 72 h of BPH feeding (Fig. 4f). By contrast, BISP levels were lower in Bph14 plants, with levels sustained after 24 h of feeding (Fig. 4f). OsATG8 levels did not significantly change in BPH-infested susceptible plants. By contrast, in Bph14 plants, OsATG8 levels increased by 6 h after feeding and were maintained for 72 h (Fig. 4f). These results suggest that autophagy is activated in Bph14 plants but not in BPH-susceptible rice. These data correlated well with the upregulation of three OsATG8 genes in Bph14 plants compared with MH63 plants following BPH infestation (Extended Data Fig. 7k). ConA treatment substantially increased levels of ATG8 proteins in BPH-infested Bph14 plants (Extended Data Fig. 7l). These results indicate that BPH feeding activates autophagy and that BISP levels are tightly controlled in Bph14 plants.

OsNBR1-mediated autophagy controls BISP levels
Mechanisms of autophagy are conserved in yeast, plants and metazoans 29 . In rice and other plants, multiple ATG8 proteins mediate autophagy [30][31][32] . To clarify the molecular mechanisms that underlie the BPH14-dependent degradation of BISP, we examined the interactions of four OsATG8 proteins with BISP or BPH14 using Y2H assays. Neither BISP nor BPH14 interacted with the OsATG8 proteins ( Fig. 4g and Extended Data Fig. 8a).
Relative Wrky72 RNA level

Article
In addition to canonical autophagy, selective autophagy targets specific cargoes, such as intracellular pathogens and defective proteins, and plays pivotal roles in eukaryotic organisms [33][34][35] . NBR1 is a selective autophagy cargo receptor and interacts with ATG8 proteins to facilitate the autophagy-dependent degradation of target proteins [33][34][35][36][37] . We therefore examined the interaction of BISP or BPH14 with OsNBR1 in Y2H assays. Both BISP and BPH14 interacted with OsNBR1 ( Fig. 4g and Extended Data Fig. 8a). Co-IP assays in rice protoplasts confirmed these interactions (Fig. 4h). Furthermore, OsNBR1 interacted with all four OsATG8 proteins in Y2H assays ( Fig. 4g and Extended Data Fig. 8a) and in rice protoplasts ( Fig. 4i), which suggested that OsNBR1 might mediate the degradation of BISP.
The roles of OsNBR1 in the autophagic degradation of BISP were examined by immunoblotting. The expression of BISP with OsNBR1 in rice protoplasts did not influence BISP levels. By contrast, the expression of BISP, OsNBR1 and BPH14 reduced the level of BISP (Fig. 4j), which suggested that OsNBR1 facilitated BPH14-dependent degradation of BISP. We used CFP-ATG8F to determine the effects of OsNBR1 on the numbers of autophagic vesicles. The expression of OsNBR1 with BISP and BPH14 induced more punctuate CFP-ATG8F structures    in N. benthamiana leaves than expression of only BISP and BPH14 (Extended Data Fig. 8b,c). Additionally, OsNBR1 protein and transcript levels were significantly higher in non-infested Bph14-Bisp plants than in Bph14 and MH63 plants (Extended Data Fig. 8d,e). Following BPH infestation, OsNBR1 protein and RNA levels were increased in Bph14 plants, but not in BPH-susceptible MH63 plants ( Fig. 4f and Extended Data Fig. 8f). Moreover, ConA treatment significantly increased OsNBR1 levels in Bph14 plants following BPH infestation (Extended Data Fig. 7l). Y2H and co-IP assays showed that the CC and LRR domains of BPH14 interacted with OsNBR1 (Extended Data Fig. 9a,b). Co-IP assays also showed that BPH14 enhanced the interaction of OsNBR1 with BISP (Fig. 4k). NBR1 homopolymerizes through its PB1 domain, binds to ubiquitin through its UBA domain and interacts with ATG8 through a conserved LIR motif to promote selective autophagy 37 . OsNBR1 contains these interaction domains, as well as conserved ZZ-type zinc finger and BRCA1 domains. To determine which OsNBR1 domain (or domains) participates in the degradation of BISP, we generated four truncated forms (N1 to N4) and three site-directed mutants (N5 to N7) of OsNBR1 (Extended Data Fig. 9c). Co-expression assays showed that none of mutants were able to degrade BISP (Extended Data Fig. 9d). Whereas the Osnbr1 N5 mutant lost the ability to form homopolymers (Extended Data Fig. 9h), the other Osnbr1 mutants retained the ability to form homopolymers. However, these mutants failed to interact with BPH14, BISP or OsATG8C (Extended Data Fig. 9e-i). These results suggest that OsNBR1 is involved in BISP degradation through autophagy and that domains outside the UBA in OsNBR1 are important for interacting with its target BISP.
We generated OsNBR1 knockout mutants in the Bph14 background (Bph14-Osnbr1) to further analyse the roles of OsNBR1 in the Bph14 Hours after BPH removal Wrky72 transcript levels in Bph14 and MH63 plants after cessation of BPH feeding. c, Co-IP detection of levels of BPH14 homomeric complex formation. N1 served as a negative control. Numbers above the lanes indicate band intensity relative to co-precipitated BPH14-MYC, quantified using ImageJ. d, Immunoblot detection of BISP and WRKY72 in Bph14 and Bph14-Osnbr1 plants after cessation of BPH feeding. e, Relative Wrky72 transcript levels in Bph14 and Bph14-Osnbr1 plants after cessation of BPH feeding. f,g, Photographs (f) and resistance scores (g) of Bph14 and Bph14-Osnbr1 plants after 14 days of BPH infestation. Scale bar, 10 cm. The box limits indicate the 25th and 75th percentiles, the whiskers indicate the full range of the data and the centre line indicates the median. Individual data points are plotted (n = 42 plants examined over 3 independent experiments). h, A working model for BISP-induced immunity regulation. In BPH-susceptible rice, BISP interacts with OsRLCK185 and suppresses its phosphorylation activity, thereby inhibiting basal defence and promoting BPH feeding. In rice carrying the BPH resistance gene Bph14, BISP binds directly to the LRR domain of BPH14 and activates HPR. Continuous activation of HPR is detrimental to plant growth and reproduction. Therefore, following activation of BPH14-mediated immunity, BISP-BPH14 binds to OsNBR1, which mediates the autophagic degradation of BISP. The degradation of BISP leads to the termination of BPH14-mediated immunity, which restores cellular homeostasis and prevents the fitness costs associated with prolonged activation of BPH14-mediated immunity. The timing of dissociation of the BISP-BPH14 complex after association with OsNBR1 and the fate of BPH14 are yet to be determined. Numbers under lanes (a,d) indicate protein abundance relative to that of actin (loading control, a,c,d), quantified using ImageJ. In b, c, e and g, P values were derived by one-way ANOVA. In b and e, data are the mean ± s.d. (n = 3, biologically independent experiments). Experiments (a,c,d,f) were repeated three times, each giving similar results. PRRs, pattern recognition receptors. Article degradation of BISP (Extended Data Fig. 9j). OsNBR1 was undetectable in the Bph14-Osnbr1 mutants (Extended Data Fig. 9k). We then expressed BISP and BPH14 in Bph14-Osnbr1, Bph14 and MH63 protoplasts. Consistent with a role for OsNBR1 in BISP turnover, BISP levels were higher in Bph14-Osnbr1 protoplasts than in Bph14 and MH63 protoplasts (Fig. 4l). BISP was more abundant in infested Bph14-Osnbr1 plants than in Bph14 plants, but less abundant than in MH63 plants (Fig. 4m). These results indicate that OsNBR1-mediated autophagic degradation controls BISP levels.

BISP degradation attenuates resistance
On susceptible plants, BPHs consume phloem sap for prolonged times 17 . By contrast, on BPH-resistant plants, various resistance factors prevent sustained phloem ingestion 11,12,16,17,38 . BPHs have a higher frequency of stylet probing, penetrating numerous sites in an attempt to find an acceptable feeding site 17,39 .
To provide insights into the duration of Bph14-mediated immunity, we assessed the longevity of BISP and the immune status of rice plants after BPH feeding had ceased.  (Fig. 5a). Correlating well with the decreased BISP levels, the levels of OsATG8 and OsNBR1 proteins and their transcripts were more abundant in Bph14 plants than in MH63 plants ( Fig. 5a and Extended Data Fig. 10a). These results suggest that after BPH feeding and secretion of saliva have ceased, BISP is stable for prolonged times in susceptible rice but is rapidly eliminated through autophagy in Bph14 plants.
Moreover, the degradation of BISP was accompanied by a mitigation of the signal outputs of BPH14-activated resistance (Fig. 5a,b and Extended Data Fig. 10b). For example, WRKY72 protein and transcript levels rapidly reduced (within 12 h) in Bph14 plants and were slower in susceptible plants after BPH feeding had ceased (Fig. 5a,b). In addition, OsICS1 and OsNPR1 transcripts showed a similar regulation pattern (Extended Data Fig. 10b). The rapid dissipation of BPH14-regulated immunity based on BISP depletion was supported by the dynamics of formation of the BPH14 homomeric complex, which was reduced when BPH14 was expressed with BISP and OsNBR1 to induce autophagy (Fig. 5c). These results suggest that autophagy contributes to the elimination of BISP and the attenuation of resistance in Bph14 plants after BPH feeding is terminated.
Further support for this theory was garnered from examining the fate of BISP and defence-signalling components in Bph14-Osnbr1 plants and WT Bph14 plants after BPH feeding was terminated. BISP levels were significantly higher in Bph14-Osnbr1 plants than in Bph14 plants at 0, 24 and 48 h after removing the insects (Fig. 5d), which confirmed that OsNBR1 deficiency slowed the degradation of BISP. WRKY72 protein and transcript levels, as well as Osics1 and Osnpr1 transcripts, were significantly higher at 0-48 h in the Bph14-Osnbr1 plants than the Bph14 plants (Fig. 5d,e and Extended Data Fig. 10c). This result suggests that the time span of BPH14-activated signalling is extended in Bph14-Osnbr1 plants.
Finally, we examined the impact of OsNBR1 on rice resistance to BPHs. Whereas Bph14 plants were damaged by prolonged BPH infestation (14 days), Bph14-Osnbr1 plants had limited symptoms and displayed a higher level of BPH resistance (Fig. 5f,g). BPH adults preferred Bph14 plants to Bph14-Osnbr1 plants in a two-host choice test. On the basis of honeydew excretion, weight gain and survival, BPHs had poorer performance when feeding on Bph14-Osnbr1 plants (Extended Data Fig. 10d-g). These results indicate that OsNBR1 deficiency enhances BPH resistance in Bph14 plants.

Discussion
HPR conferred by NLR proteins is widely used for controlling insect pests 40 . Here we provide molecular and cellular insights into the long-standing question of how plant NLR receptors perceive insect-feeding signals to deploy and modulate ETI resistance mediated by these proteins. Notably, we discovered that the tripartite interactions of BISP, its cognate NLR receptor BPH14 and the selective-autophagy receptor NBR1 enable a self-regulated resistance mechanism to BPHs (Fig. 5h).
In this tripartite interaction system, the BISP effector is secreted into rice leaf sheaths to bind to one of its targets, OsRLCK185. By interfering with OsRLCK185 autophosphorylation, BISP suppresses rice immune responses to promote BPH feeding and success. However, in resistant rice, the NLR receptor BPH14 directly binds to BISP. Following recognition, resistance is activated, which stops BPHs from ingesting phloem sap 8,17,19 . Constitutive activation of BISP-triggered resistance in Bph14 plants is detrimental to plant fitness, reducing plant stature and yields. Therefore, mechanisms to temper the resistance response must be deployed to promote plant vitality. The importance of the ubiquitin-proteasome system in effector and NLR protein turnover in pathogen-plant ETI is known 26,41 . Here we revealed that the governor of Bph14-mediated resistance is OsNBR1-regulated autophagy, which controls BISP turnover. Both BISP and BPH14 physically interact with OsNBR1. In turn, OsNBR1 interacts with OsATG8 to promote the degradation of BISP and, thereby, fine-tune the level of Bph14-mediated resistance. Whether the BISP-BPH14 complex dissociates before BISP turnover by autophagy or whether BPH14 is also degraded by autophagy is an essential topic of research in the future. Finally, although autophagy can be induced by transient reprogramming in response to plant hormones 42 , the role of SA, which is induced in the BISP-BPH14 interaction, remains to be elucidated.
The BISP-BPH14-OsNBR1 resistance response adapts to changes in BPH feeding pressure. BPHs terminate feeding when a host becomes undesirable, which can induce individual insects or entire BPH populations to migrate long distances in search of better host plants [15][16][17][18] . When BPHs feed on Bph14 plants, BISP triggers resistance, which makes Bph14 plants unpalatable and forces the insects to stop feeding. With the cessation of feeding and the departure of insects from these sites, the activated autophagy pathway rapidly degrades BISP to restore cellular homeostasis. Immune systems resume their 'off' status, thereby allowing rice to reallocate resources to growth and reproduction. The newly described BISP-BPH14-OsNBR1 tripartite interaction system enables rice to resist BPHs without compromising yield performance in the natural habitats 43 . By leveraging the autoregulatory mechanisms used in Bph14 gene-for-gene resistance in new ways, we have the ability to create a new generation of resistant rice that can feed the increasing world population in a sustainable manner and, significantly, to control insect pests without relying on insecticide sprays.

Online content
Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41586-023-06197-z.

Insect materials and growth conditions
The BPH insects were maintained on susceptible cultivar Taichung Native 1 at 28 ± 2 °C with 60-80% relative humidity and a photoperiod of 16 h of light-8 h of dark. RNAi of BPH by dsRNA injection was developed as part of this study: GFP-RNAi (microinjected with dsGFP) and Bisp-RNAi (microinjected with dsBisp).

BPH bioassays on rice plants
To evaluate BPH resistance of transgenic and WT rice plants, sets of around 15 or 20 seeds from the same plant were sown in a plastic cup (10 cm in diameter, 20 cm in height). At the three-leaf stage, all plants were placed under a large gauze cover, and each seedling was infested with ten second-or third-instar BPH nymphs. When all of the susceptible plants had died (scored as 9), each seedling of the other cultivars or lines was given a score of 0, 1, 3, 5, 7, or 9 according to the degree of damage, as previously described 8,46,47 , and the BPH resistance score was calculated. At least three replicates were used for each cultivar or line.
To measure the weight gain and honeydew excretion of BPHs, a newly emerged short-winged female adult and a Parafilm sachet were weighed using a 1/100,000 Shimadzu AUW120D electronic balance 11,12 . The insect was placed in the pre-weighed sachet, which was then attached to the leaf sheath near the lower part of a rice plant. After 48 h, the insect and Parafilm sachet were weighed again. The difference between the first and second measurements of the weight of the insect was recorded as the BPH weight gain; the difference between the first and second measurements of the weight of the sachet was recorded as amount of honeydew excreted. At least ten replicates were used for each cultivar or line.
For the two-host choice tests, a transgenic plant and a WT plant were grown in a plastic cup as described elsewhere [10][11][12] . At the four-leaf stage, the cup was covered with gauze, and 20 or 30 second-or third-instar nymphs were released into the cup (the exact number depended on the BPH resistance of the WT plant). The number of nymphs that settled on each plant was recorded at 6, 12, 24, 48, 72, 96, and 120 h after release (the moment of release was considered time 0). Ten cups were analysed for each cultivar or line examined.
To study BPH nymph survival on rice, second-or third-stage BPH nymphs were released (ten insects per plant), and the cups were covered with a light-transmitting mesh. The nymphs on each plant were counted 8 or 9 days after release. Survival rates were calculated as the number of surviving nymphs divided by the total number of nymphs released at the start of the experiment. Ten cups were analysed for each cultivar or line.

Plasmid construction
Routine molecular cloning techniques were used to prepare the constructs. The plasmids and primers used in this work are listed in Supplementary Table 3. All of the resulting recombinant vectors were sequenced.
To construct the plasmid for Y2H screening, the full-length Bph14 sequence was amplified from Bph14 leaf sheath cDNA and cloned into the Xma I site of the pGBKT7 vector, which produced the BPH14-BD vector.
To construct the plasmids for the Y2H assays, the coding sequences of Bisp (ORF sequence without its 1-25 amino acid residue signal peptide; amino acids 26-241), Bisp 26-124 (amino acids  and Bisp 125-241 (amino acids 125-241) were amplified and cloned into the Xma I site of the pGADT7 vector to produce BISP-AD, BISP(26-124)-AD and BISP(125-241)-AD, respectively. The coding sequences of the CC, NB and LRR domains and full-length Bph14 were amplified from the BPH14-BD vector and cloned into the Xma I sites of the pGADT7 and pGBKT7 vector to produce CC-AD, NB-AD, LRR-AD, BPH14-AD, CC-BD, NB-BD, and LRR-BD, respectively. The coding sequence of N14 (Bph14 allele in the rice variety Nipponbare) was amplified from N14 leaf sheath cDNA and cloned into the Xma I site of the pGBKT7 vector to produce N14-BD.
Constructs used for rice protoplast transfection were generated using pCXUN-4×HA and pCXUN-4×Myc 19  To analyse the subcellular localization of BISP, the coding sequence of Bisp (ORF sequence without its signal peptide) was cloned downstream of the maize (Zea mays) Ubiquitin 1 promoter, in-frame with GFP in the binary vector pCAMBIA1300, which produced the BISP-GFP construct.
The Escherichia coli recombinant protein vectors used for the expression and purification of proteins in the phosphorylation activity assays, designated pET-MBP-His and pET-GST-His, were created by adding a C-terminal MBP or GST tag, respectively, to the pET-28a expression vector (EMD Biosciences, Novagen). The coding sequences of Bisp (ORF sequence without its signal peptide), Bisp   To generate the CRISPR-Cas9 construct, the target sites for CRISPR-Cas9 were designed using the web-based tool CRISPR-P (v.2.0; http:// crispr.hzau.edu.cn/CRISPR2/). The CRISPR-Cas9 binary construct was prepared as previously described 49 , which produced the construct OsNBR1-Cas9.

Y2H screening
The Matchmaker Gold Yeast Two-Hybrid System (Clontech, 630489) was used to screen for BPH14-interacting proteins. Y2H Gold cells carrying BPH14-BD were mixed with 1 ml of a BPH salivary gland cDNA library and incubated overnight before plating on TDO (SD/-Leu-Trp-His)-selective medium. Candidate clones growing on TDO medium were confirmed following the manufacturer's protocol (Clontech, 630489). The BPH cDNA library was constructed using the Make Your Own "Mate & Plate" Library System (Clontech, 630490) with a simple and highly efficient protocol. Total RNA was isolated from BPH salivary glands using TRIzol reagent (Takara, 9109), and cDNA was synthesized from the isolated mRNA using SMART cDNA synthesis technology (Clontech, 634926).
For the Y2H assay of protein interactions, the yeast strain AH109 was transformed with the indicated bait and prey plasmids according to the manufacturer's instructions. Co-transformants were simultaneously plated on selection medium containing DDO (SD/-Leu-Trp), TDO (SD/-Leu-Trp-His) with the appropriate concentration of 3-amino-1,2,4-triazole (3-AT) or QDO (SD/-Leu-Trp-His-Ade) and incubated at 30 °C.

Cloning and sequencing of Bisp cDNA
The cDNA sequences of Bisp were obtained from BPH salivary gland transcriptomes 5 . To obtain the full-length counterparts of the truncated sequences in the transcriptomes, 5′-and 3′-RACE amplification was performed using a SMARTer RACE cDNA Amplification kit (Clontech, 634923) following the manufacturer's instructions. Bisp-specific primers were obtained based on sequencing data. The RACE products were amplified, and the purified products were ligated into the pMD18-T vector (Takara, 6011) and sequenced.

Protein extraction, immunoblot analysis and in vivo co-IP assays
Proteins were extracted from the leaf sheaths of rice seedlings at the four-leaf stage. The leaf sheaths were wiped with a cotton ball soaked in alcohol to remove honeydew and ground in liquid nitrogen. Total protein was extracted from the leaf tissue in rice protein extraction buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 10% glycerol, 0.1% NP-40, plant protease inhibitor cocktail (Roche)). Equal amounts of total protein were analysed by SDS-PAGE and detected by immunoblotting using anti-actin (Abbkine, ABL1050; 1:3,000), anti-BISP (1:500), anti-OsWRKY72 (Beijing Protein Innovation, AbP80456-A-SE; 1:500) and anti-AtNBR1 (Agrisera, AS194281; 1:1,000) antibodies. For OsATG8 analysis, equal amounts of proteins were subjected to SDS-PAGE with 6 M urea 50 and detected by immunoblotting using anti-AtATG8A antibodies (Abcam, ab77003; 1:1,000). The anti-BISP and anti-NISP1 antibodies were prepared by expressing BISP and NISP1 (cloned into pET28a) in E. coli strain BL21 (DE3), respectively. The expressed recombinant proteins were collected and injected into two rabbits (DIA·AA Biotech). Ponceau S solution (Sigma-Aldrich, P7170) was used to stain the PVDF membrane and was used as a loading control. The intensities of protein signals were quantified using ImageJ.
For the co-IP assays, rice protoplasts were incubated at 28 °C for 14-20 h after transfection. Total protein extracts were collected from the protoplasts in the rice protein extraction buffer as described above. The supernatants were incubated with 10 μl Protein G Agarose beads (Millipore,   For the yeast protein immunoblots, total yeast protein was extracted as previously described 51 . In brief, single yeast colonies were resuspended in 4 ml YPDA liquid medium and grown to OD 600 = 0.6 at 30 °C. The pelleted yeast cells were resuspended in 1 ml distilled water, combined with 100 μl 0.2 M NaOH, incubated for 5 min at room temperature and pelleted and resuspended in 50 μl loading buffer. The proteins were boiled for 10 min at 95 °C and detected by immunoblotting with anti-HA or anti-MYC antibodies. Ponceau S solution was used to stain the PVDF membrane for loading control.

Rice protoplast isolation and transfection
Rice protoplasts were prepared from etiolated seedlings and transfected as previously described 11,19 . About 3-10 μg of plasmid DNA was used to transfect around 2 × 10 5 protoplasts using the PEG-mediated method. The fluorescent or epitope-tagged proteins were detected at 14-20 h after transfection.

Subcellular localization and BiFC analysis
For subcellular localization, the BISP-GFP plasmid and the nuclear marker bZIP63-RFP 11,19 were transformed into rice protoplasts. Following incubation at 28 °C for 14-20 h, the protoplasts were imaged using a confocal microscope (Leica, DMi8).
For BiFC analysis, protoplasts were transfected with the indicated expression vectors. Following incubation at 28 °C for 14-20 h, the protoplasts were imaged by confocal microscopy. BiFC fluorescent signals were quantified using previously described methods 52 . The intensity of YFP and CFP signals in individual protoplasts was determined using ImageJ software. The average YFP/CFP intensity ratio (per cent) of the whole cell was measured, and the means were calculated from five to ten independent cells. The fluorescent or epitope-tagged proteins were detected with anti-GFP (Roche, 11814460001; 1:1,000), anti-HA or anti-MYC antibodies.

RNA isolation, real-time and semi-quantitative RT-PCR analysis
Total RNA was isolated from rice leaf sheaths, BPHs or N. benthamiana leaves using TRIzol reagent (Takara, 9109). First-strand cDNA was synthesized using a PrimeScript RT Reagent kit with gDNA Eraser (Takara, RR047A) following the manufacturer's instructions.
Real-time RT-qPCR was carried out using a CFX96 Real-Time system (Bio-Rad) with SYBR Green Supermix (Bio-Rad, 172-5274) following the manufacturer's instructions. For RT-qPCR in rice, three housekeeping genes, Osubiquitin, Ostbp and Oshsp 53 , were used as reference genes for calibration of real-time RT-PCR data. For RT-qPCR in BPH, NlRPS18, Nlubiquitin and NlACTB (which encodes β-actin) were used as reference genes. The sequences of the primers used are listed in Supplementary Table 3.
For the semi-quantitative RT-PCR assays, first-strand cDNA (5 μl) was diluted to 50 μl final volume with TE buffer, and 1 μl of the dilution was used as template for PCR amplification using gene-specific primers. The internal standard NbACTB was used. The number of cycles used for amplification with each primer pair was adjusted to ensure that the amplification was in the linear range. The PCR products were sequenced to ensure that they were derived from the targeted genes. The sequences of the primers used are listed in Supplementary Table 3.

Immunohistochemistry
For immunolocalization of BISP in rice leaf-sheath sections, N14 plants at the two-leaf stage were infested with ten fourth-or fifth-instar nymphs for 2 days or served as non-infested controls. The outer leaf sheaths were collected, fixed, dehydrated, and embedded in paraffin (Paraplast Plus, Sigma-Aldrich, P3683) as previously described 11,17 . The sections were incubated in dimethylbenzene and an alcohol gradient to remove the paraffin and to gradually rehydrate. The sections were incubated with anti-BISP antibodies or pre-immune rabbit serum (prepared as described above, 1:100) at 4 °C in the dark overnight. The next day, the sections were washed 3 times (15 min each time) with PBST (PBS with 0.1% (v/v) Triton X-100), blocked with 5% bovine serum albumin in PBST at room temperature for 1 h, washed 3 times at 15-min intervals with PBST, and incubated with the secondary Cy3-conjugated goat anti-rabbit antibodies ( Jackson ImmunoResearch, 111-165-003, 1:500) at 4 °C overnight. The sections were washed extensively with PBST at 15-min intervals and imaged under a confocal microscope (Leica, DMi8).
For immunolocalization in BPH salivary glands and guts, female or male insects were anaesthetized with carbon dioxide (CO 2 ). The salivary glands and guts were dissected in 0.65% NaCl, washed 3 times in PBST and fixed in 4% (v/v) formaldehyde in PBST at 4 °C overnight. The next day, the organs were extensively washed with PBST, incubated with anti-BISP antibodies, pre-immune serum or anti-NlSP1 antibodies 54 (1:100) at 4 °C in the dark overnight, and treated as described above. Nuclei were stained with 100 nM 4,6-diamidino-2-phenylindole (DAPI) at room temperature in the dark for 5 min. The salivary glands and guts were extensively washed with PBST and placed on a glass slide. Photographs were taken under a confocal microscope (Leica, DMi8). For the quantification of Cy3, the Cy3 and DAPI fluorescence signals in individual cells were quantified using ImageJ. The average Cy3/DAPI intensity ratio of each cell was measured, and mean ratios were calculated from five to ten independent cells.

RNAi and bioassay of BPHs after dsRNA injection
A 506-bp fragment of Bisp was amplified with the primers listed in Supplementary Table 3, including the T7 promoter sequence in the forward primer. After cloning the Bisp fragment into the pMD18-T vector (Takara, 6011), dsRNA was synthesized from PCR-generated DNA templates using a MEGAscript T7 Transcription kit (Ambion, AM1354). BPH nymphs in their third, fourth or fifth stage were injected with Bisp-RNAi or GFP-RNAi using a Nanoliter 2010 injector (World Precision Instruments) as previously described 5 . The efficiency of gene silencing was determined by RT-qPCR at 1-10 days after dsRNA treatment.
After the injected BPHs were placed and reared on rice plants for 12 h, the survival rate, weight gain and honeydew excretion of GFP-RNAi BPHs, Bisp-RNAi BPHs and non-injected control insects were measured as described above. Five replicates were set up for each treatment to measure BPH survival rates, and 20 replicates were performed for each treatment group to measure weight gain and honeydew excretion.

Plant transformation
To constitutively express Bisp in rice plants, the BISP-MYC plasmids were transformed into N14 or Bph14 plants through Agrobacterium-mediated transformation 8 . To knockout OsNBR1 by genome editing, the CRISPR-Cas9 plasmids OsNBR1-Cas9 were introduced into Agrobacterium tumefaciens strain EHA105 and transformed into Bph14 as described elsewhere 8 . Genomic DNA was extracted from these transformants, and PCR amplification was performed using primer pairs flanking the designed target site. The PCR products were sequenced to determine whether the gene editing was successful.

Quantification of free SA
Endogenous free SA levels were determined in rice leaf sheaths. Fifteen leaf sheaths from each plant or line were ground to a fine powder in liquid nitrogen as a replicate. Three biological replicates of each frozen sample were prepared. Each sample (100 mg) was added to 750 μl precooled extraction buffer (methanol:water:acetic acid, 80:19:1, v/v/v) supplemented with 3 μg naphthaleneacetic acid (Sigma-Aldrich, 35745) as an internal standard and shaken at 4 °C overnight. After centrifugation, the supernatant was filtered through a 0.22-μm nylon membrane and dried under a nitrogen flow at room temperature for at least 4 h. Finally, 200 μl methanol was added to dissolve the extracts. Endogenous free SA measurements were performed as described elsewhere 55 .
For the BISP antibody specificity analysis, the purified BISP-GST-His protein and total protein from whole insect and a twofold dilution series were separated on SDS-PAGE gels. BISP-GST-His was detected in immunoblots using anti-BISP and anti-His (GenScript, A00186; 1:2,000) antibodies, respectively. Ponceau S solution was used to stain the PVDF membrane and used as the total BPH protein loading control.

Recombinant protein expression and purification in insect cells and BLI analyses
The binding kinetics and affinities of LRR and N14-LRR with BISP, BISP  or BISP(125-241) was determined using an Octet RED96 system (FortéBio) as previously described 56,57 . Purified LRR and N14-LRR were biotinylated with NHS-biotin (Sangon Biotech, C100212) according to the manufacturer's instructions. All streptavidin-coated biosensors (FortéBio) were hydrated in PBS buffer (2 mM KH 2 PO 4 , 8 mM Na 2 HPO 4 , 136 mM NaCl, 2.6 mM KCl, 0.05% Tween 20, and 0.5% bovine serum albumin, pH 7.4) for 10 min. Biotinylated LRR and N14-LRR were diluted in PBS buffer to a final concentration of 20 μg ml −1 and immobilized onto a streptavidin-coated biosensor. Biosensors with immobilized LRR and N14-LRR were diluted in binding buffer containing different concentrations of BISP, BISP  or BISP(125-241) for the binding kinetics analysis (association step). Subsequently, the streptavidin-coated biosensor was returned to PBS buffer for 180 s (dissociate step). PBS buffer was used in all BLI experiments to reduce nonspecific binding to the biosensors except for experiments in the presence of BISP, BISP  or BISP(125-241). The experiments included the following steps at 25 °C: (1) loading (60 s); (2) baseline (120 s); (3) immobilization of proteins onto sensors (60 s); (4) association (180 s); and (5) disassociation (180 s). The data were analysed, and sensor-grams were step-corrected and reference-corrected. Global fitting of the kinetic rates was performed, then the equilibrium dissociation constant (K d ), association (K on ) and dissociation (K off ) rate constants and their standard errors were analysed using FortéBio data analysis software (v.1.1.0.16, FortéBio). The assays were repeated two times for each affinity measurement.
For the competition assay, an Octet RED96 system (FortéBio) was used as previously described 58

MST assay
The MST assay was performed as previously described 59,60 . The affinity of purified LRR or N14-LRR with BISP, BISP  or BISP(125-241) was measured using a Monolith NT.115 (Nanotemper Technologies). Proteins were fluorescently labelled using a Monolith Protein Labelling kit (RED-NHS 2nd Generation; Nanotemper Technologies, MO-L011) according to the manufacturer's protocol, and the labelled protein used for each assay was about 20 nM. A solution of unlabelled BISP, BISP  or BISP(125-241) protein was diluted for an appropriate serial concentration gradient. The samples were loaded into capillaries (Nanotemper Technologies, MO-K022) after incubation at room temperature for 5 min. Measurements were performed in 20 μl of PBS buffer containing 0.05% Tween 20. The assays were repeated three times for each affinity measurement. Data analyses were performed using Nanotemper Analysis software provided by the manufacturer.
For the competition assays between LRR with BISP and BISP(26-124) by MST, 50 nM LRR proteins and 200 nM BISP  were combined in PBS buffer with 0.05% Tween 20 and incubated for 10 min to enable protein interactions. A series of different concentrations of BISP was then added to the LRR-BISP(26-124) mixture before measurement. The assays were repeated three times for each measurement. The dissociation constant (K i ) from the MST assays was calculated using the K i Finder software from the Nanotemper website (https://nanotemper.my.site.com/ explore/s/article/Can-competition-assays-be-performed-with-MST).

Inhibition of protein degradation in rice protoplasts
After PEG-mediated transformation, the rice protoplasts were placed in 1 ml W5 buffer containing 10 μM 3-MA (3-methyladenine, Selleck, S2767), 100 μM chloroquine (MCE, HY-17589A), 100 μM leupeptin (MCE, HY-18234A), or 20 μM E-64d (Aloxistatin, Sigma-Aldrich, E8640) to inhibit autophagy or 50 μM MG132 (Selleck, S2619) to inhibit the 26S proteasome. The treated protoplasts were incubated at 28 °C for 14-20 h before protein extraction. Anti-actin (Abbkine, ABL1050) was used as a reference to quantify total protein levels. Rice WRKY72 is degraded by the 26S proteasome pathway 19 and was used as a positive control to ensure that the proteasome inhibitor MG132 was active. The rice HD1 protein is degraded in the dark by the autophagy pathway 61 , and OsNBR1 was used as a positive control to assure autophagy inhibitors were active. The intensity of protein signals in immunoblots was quantified by ImageJ.

Live-cell imaging of autophagic structures and virus-induced gene silencing in N. benthamiana leaves
To observe autophagic structures in N. benthamiana leaves by confocal microscopy, A. tumefaciens strain GV3101 cells harbouring CFP-NbATG8F (designed as CFP-ATG8F) 27  For virus-induced gene silencing (VIGS) assays, pTRV1 and pTRV2 or its derivatives were introduced into Agrobacterium strain GV3101. VIGS assays were performed as previously described 62 . In brief, the Agrobacterium cultures with pTRV1 and pTRV2 or its derivatives were mixed at a 1:1 ratio and were infiltrated into the leaves of six-leaf stage N. benthamiana plants. Silenced phenotypes appeared in the upper leaves about 10 days after infiltration. A. tumefaciens strain GV3101 cells harbouring CFP-NbATG8F and the constructed expression vectors BISP-101 (BISP-YFP) and BPH14-203 were infiltrated into the N. benthamiana upper leaves.

Transmission electron microscopy
For immunogold electron microscopy, N14 leaf sheaths infested with BPH for 24 h were cut into 0.1-mm pieces, fixed in precooled 3% paraformaldehyde with 0.1% glutaraldehyde (in 100 mM PBS, pH 7.4) and immediately placed under a vacuum until the samples were completely submerged in the liquid. The samples were transferred to fresh 0.1% glutaraldehyde and incubated at 4 °C overnight. The next day, the samples were washed 3 times with PBS (pH 7.4) for 30 min. The samples were cleared by successive washes with 30% and 50% ethanol for 60 min each at 4 °C, then 70%, 80%, 85%, 90%, and 95% ethanol for 30 min each at −20 °C, and finally 100% ethanol 3 times for 60 min each at −20 °C. After ethanol washes, the samples were embedded in Lowicry K4M resin stepwise, according to the manufacturer's instructions. Ultrathin sections (70 nm) of the embedded tissues were cut with a diamond knife using an Ultracut E Ultramicrotome (Reichart-Jung) and collected on 3-mm copper (mesh) grids. Immunolabelling of leaf sections was done using standard procedures as previously described 59 with anti-BISP antibodies or pre-immune serum at 100 μg ml -1 and anti-rabbit IgG gold-coupled secondary antibodies (Sigma-Aldrich, G7277-4ML) at 1:50 dilution. The sections were then stained with 2% uranyl acetate (in 70% ethanol) for 15 min, washed in distilled water 3 times, stained with lead citrate for 12 min, washed in Article NaOH buffer, washed in distilled water 3 times, and examined under a JEM-1230 electron microscope.
For transmission electron microscopy observations, 2-week-old Bph14-Bisp transgenic lines, Bph14 and MH63 leaf sheaths were cut into 0.1-mm pieces, fixed in precooled 4% paraformaldehyde with 2% glutaraldehyde (in 100 mM PBS, pH 7.4) and immediately placed under a vacuum until the samples were completely submerged in the liquid. The samples were transferred to fresh 2% glutaraldehyde and incubated at 4 °C overnight. The next day, the samples were washed 5 times with PBS (pH 7.4) for 20 min and submerged in a 1% solution of osmium tetroxide at room temperature for 1-2 h until they turned completely black because of oxidation. The samples were washed 5 times with PBS at 20-min intervals. The samples were cleared by successive washes with 15%, 30%, 50%, and 70% ethanol for 30 min each, then 80%, 85%, 90%, and 95% ethanol for 20 min each, and finally 100% ethanol twice for 45 min each. After ethanol washing, the samples were embedded in Spurr resin stepwise as instructed by the manufacturer. Ultrathin sections (70 nm) of the embedded tissues were cut with a diamond knife using an Ultracut E Ultramicrotome (Reichart-Jung), collected on 3-mm copper (mesh) grids, stained with 2% uranyl acetate (in 70% ethanol) for 10 min, washed in distilled water 3 times, stained with lead citrate for 15 min, washed in NaOH buffer, and examined under a JEM-1230 electron microscope. To quantify autophagosomes, the double-membrane autophagosomes in every eight transmission electron microscope images were counted and served as one biological replicate; each genotype had five replicates.

BPH infestation
Seeds were sown in 10-cm diameter plastic cups and grown in a plant growth chamber (CONVIRON PGC2000) as described elsewhere 11 . To analyse the relationship between BISP levels and autophagy, 4-leaf rice plants were infested with second-or third-instar BPH nymphs (10 per plant) at the selected time points (6,12,24,48, and 72 h before the end of the experiment). Control groups were maintained in parallel but without BPH infestation. At the end of the treatments, leaf sheaths were wiped with cotton wool balls soaked in alcohol to remove honeydew, leaf sheaths were excised, frozen, and ground to a fine powder in liquid nitrogen, and proteins isolated. Immunoblotting was performed following the procedures described above in the section 'Protein extraction, immunoblotting and in vivo co-IP assays'. All treatments were repeated three times.
To measure autophagy and BISP levels following BPH infestation, 4-leaf rice plants were infested with second-or third-instar BPH nymphs (10 per plant) for 24 h. The insects were then removed, and the plants were allowed to grow under standard conditions (described above). At the selected time points after BPH removal, honeydew was removed from leaf sheaths, leaves were excised, frozen, and ground to a fine powder in liquid nitrogen, and proteins isolated. Immunoblotting was performed following the procedures described above. All treatments were repeated three times.
For the autophagy inhibitor ConA treatment, the BPH-infested and non-infested rice plants were treated with or without ConA (10 μM) for 12 h before samples were collected. Leaf sheath total protein extraction and immunoblotting were performed according to the procedures described above. All treatments were repeated three times.

Agronomic performance test
To investigate the agronomic performance of Bph14 and Bph14-Bisp plants, plants were grown in a field in Wuhan, China, under routine management. At harvest, the middle plants in the central row of each plot were assessed for the following nine agronomic traits: plant height, sword leaf length, sword leaf width, number of panicles per plant, number of filled grains per panicle, number of filled grains per plant, percentage of filled grain, 1,000-grain weight, and grain yield per plant 63 .

Statistical analysis
No statistical methods were used to predetermine the sample sizes. The experiments were not randomized, and investigators were not blinded to allocation during the experiments and outcome assessment. Quantification analyses of protein abundance were conducted using ImageJ software. All values are presented as the mean ± s.d. or mean ± s.e.m. as indicated. Data points are plotted on the graphs, and the number of samples is indicated in the corresponding figure legends. In all cases, sampling for replicates and between time points was performed by collecting material from new plants. All statistical analyses were performed by one-way analysis of variance with MS Excel and GraphPad Prism (v.8.0.1, GraphPad Software). Detailed information about statistical analysis values for all experiments is provided in Supplementary Table 4.

Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability
All data are available within this article and its Supplementary Information. Original gel blots are shown in Supplementary Fig. 1. Original data points in graphs are shown in the source data files. Statistical analyses of this study are provided in Supplementary Table 4. The sequences of Bisp have been deposited and made publicly available in GenBank with accession number MH885414. Source data are provided with this paper. Fig. 1 | Host-plant resistance evaluation, BISP subcellular localization in rice protoplasts, Bisp gene and BISP protein expression profile in BPH, and BISP delivery into rice leaf sheath. a, BPH-resistance scores of Bph14 and N14 plants after BPH infestation. Bph14 and N14 plants express the resistant NLR (BPH14) and the susceptible allele variant (N14), respectively. Lower scores indicate higher resistance to the insects. Data were collected 7 days after BPH infestation (n = 36 plants examined over 3 independent experiments). P values were derived by one-way ANOVA. b, Honeydew excretion of BPHs fed on Bph14 and N14 plants for 2 days. P values were derived by one-way ANOVA (n = 20, biologically independent plants). c, BISP-GFP had a nucleocytoplasmic localization in rice protoplasts. bZIP63-RFP was used as a nuclear marker. The confocal images were taken 14-20 h after transformation. Scale bar, 5 μm. d, Immunoblot detection of full-length BISP-GFP expression in rice protoplasts. BISP-GFP was detected by anti-GFP antibody. e, Immunoblot detection of the expressed proteins in BiFC assays. BISP-YN, BISP 125-241 -YN(BISP(125-241)-YN), and BPH14-YC or N14-YC were co-expressed in rice protoplasts and the total proteins were detected using anti-Myc, anti-HA or anti-ACTIN antibody. f, Relative levels of Bisp RNA in different BPH tissues. SG, salivary glands; FB, fat body. Data are mean ± s.d. (n = 3, biologically independent experiments). g, Relative level of Bisp RNAs during different developmental stages. 1st to 5th, are first to fifth instars; Female and Male are adults. Data are mean ± s.d. (n = 3, biologically independent experiments). h, (Upper) Determination of the anti-BISP specificity using purified BISP-GST-His. The purified BISP-GST-His protein was serially diluted and detected by anti-BISP or anti-His antisera. (Bottom) Correlation between the anti-BISP labeled signal and anti-His labeled signals. The densities of anti-BISP and anti-His labeled signals were quantified using ImageJ with the undiluted sample set as 1. i, (Upper) Determination of the anti-BISP specificity using a dilution series of total mixed adults BPH proteins. BISP was detected using the anti-BISP antiserum. Ponceau staining (Ponceau S) is a loading control. (Bottom) Correlation between the anti-BISP labeled and Ponceau S signals. The density of anti-BISP labeled and Ponceau S signals were quantified using ImageJ with the undiluted sample set as 1. j,k, Immunohistochemical localization of BISP in male BPH salivary glands. Anti-BISP ( j) or pre-immune rabbit serum (k) was used as the primary antiserum. Red fluorescence (Cy3) shows the localization of BISP and blue fluorescence shows the location of DAPI-stained nuclei. Scale bars, 100 μm. l,m, Immunohistochemical localization of BISP in mixed adults BPH guts. Anti-BISP (l) or pre-immune rabbit serum (m) was used as the primary antiserum. Red fluorescence (Cy3) shows the localization of BISP and blue fluorescence shows the location of DAPI-stained nuclei. Scale bars, 100 μm. n,o, Immunohistochemical localization of NlSP1 in female BPH salivary glands. Either anti-NlSP1 (n) or pre-immune rabbit serum (o) was used as the primary antiserum. Red fluorescence (Cy3) shows the localization of NlSP1 and blue fluorescence shows the location of DAPI-stained nuclei. Scale bars, 100 μm. p, Quantification of Cy3 intensities. The Cy3 fluorescence in samples using anti-BISP, anti-NlSP1 and pre-immune serum was normalized relative to DAPI fluorescence (blue) intensity. P values were derived by one-way ANOVA. Data are mean ± s.d. (n = 5, biologically independent experiments). q-t, Immunogold electron microscopy analysis of the distribution of BISP in the rice leaf sheaths during BPH feeding. q,s, TEM image of the BPH salivary sheaths in the rice phloem tissue. Either anti-BISP (q) or pre-immune rabbit serum (s) was used as the primary antiserum. Scale bars, 20 μm. r,t, TEM image of the boxed area in (q) and (s). Red arrows indicate the anti-BISP-labeled immunogold particles (dark spots). Scale bars, 200 nm. SH, salivary sheaths; P, phloem cells. In box plots in a,b, the box limits indicate the 25th and 75th percentiles, the whiskers indicate the full range of the data, and the center line indicates the median. Individual data points are plotted. Experiments were independently repeated two (q-t) or three (b-e, h-o) times, each giving similar results. Fig. 2 | BISP played crucial roles in the feeding and performance of BPH on rice plants. a, Effects of dsRNA injection on Bisp transcript levels in BPHs. Bisp-Ri (Bisp-RNAi) insects were assayed daily for 10 days. GFP-Ri (GFP-RNAi) and CK insects were assayed on day 3. Data are mean ± s.d. (n = 3, biologically independent experiments). The inset shows a representative immunoblot demonstrating the relative abundance of BISP in GFP-Ri BPHs and Bisp-Ri BPHs at 3 days after injection. Total BPH proteins were extracted from injected or control insects and immunoblotted. BISP was detected with anti-BISP antibodies and Ponceau S staining was used as the loading control. b,c,d, Weight gain (b), honeydew excretion (c) and survival rates (d) of dsRNA-injected and non-injected BPHs fed on N14 plants. e,f,g, Weight gain (e), honeydew excretion (f) and survival rates (g) of dsRNAinjected and non-injected BPHs fed on the Osrlck185 − -1 (Osrlck185-1) and WT plants. h,i,j, Weight gain (h), honeydew excretion (i) and survival rates ( j) of dsRNA-injected BPHs fed on the Bph14 and MH63 plants. In a-j, BPHs were microinjected with Bisp dsRNA (Bisp-Ri) or GFP dsRNA (GFP-Ri) and noninjected BPHs (CK) served as a control. For b-j, P values were derived by one-way ANOVA. In b,c,e,f,h,i, data are mean ± s.d. (n = 20, biologically independent plants); in d,g,j, data are mean ± s.e.m. (n = 5, biologically independent experiments). In box plots in b,c,e,f,h,i, the box limits indicate the 25th and 75th percentiles, the whiskers indicate the full range of the data, and the center line indicates the median. Individual data points are plotted. In e-j, different letters indicate significant differences (P < 0.05, one-way ANOVA); exact P values are provided in Supplementary Table 4. All the experiments were repeated three times, each giving similar results. Fig. 4 | Identification of OsRLCK185 and its role in plant defense. a, Y2H screen for BISP-interacting kinases. Yeast strain AH109 was co-transformed with the indicated constructs and grown on DDO (SD/-Leu-Trp) and TDO (SD/-Leu-Trp-His) selective medium containing 1.5 mM 3-amino-1,2,4-triazole (3-AT). b, Immunoblot detection of proteins expressed in yeast using anti-HA and anti-Myc (anti-MYC) antibodies, respectively. Ponceau S staining was used as the loading control. Asterisks indicate specific bands detected by immunoblotting analysis. c, Structure of OsRLCK185. The orange box indicates the kinase domain. d, BISP but not BISP  (BISP(26-124)) reduced OsRLCK185 autophosphorylation activity at Ser and Thr residues in vitro. OsRLCK185 proteins were immunoblotted and phosphorylated OsRLCK185 was detected with anti-phosphoserine/ phosphothreonine antibodies. The input proteins were immunoblotted and detected with anti-His antibodies. e, Verification of three independent OsRLCK185 knockout lines (Osrlck185, Osrlck185 − -1, −2, and −3) by PCR-based sequencing. f,g, Weight gain (f) and honeydew excretion (g) of BPHs on three Osrlck185 − lines and WT plants for 2 d. P values were derived by one-way ANOVA (n = 30, biologically independent samples). In box plots, the box limits indicate the 25th and 75th percentiles, the whiskers indicate the full range of the data, and the centre line indicates the median. Individual data points are plotted. Experiments were repeated three times (a,b,d,f,g) or two times (e), each giving similar results. Fig. 8 | OsNBR1 played a role in the autophagic degradation of BISP by BPH14. a, Immunoblot detection of proteins expressed in yeast using anti-HA and anti-Myc antibodies, respectively. Ponceau S staining was used as the loading control. Asterisks indicate specific bands detected by immunoblotting analysis. b, Confocal images of CFP-NbATG8f-labeled autophagic puncta structures in the absence (left) or presence (right) of ConA. The photos showing cells co-expressing CFP-NbATG8f and BISP with N14 and OsNBR1, BPH14, BPH14 and OsNBR1, or BPH14 and OsNBR1 mutant N1 after agroinfiltration of N. benthamiana leaves. Red arrows indicate CFP-NbATG8flabeled autophagic puncta. (Bottom Row) Insets show enlarged autophagosome image at higher magnification. EV, Empty Vector. Scale bars, 25 μm. c, Average number of CFP-NbATG8-labeled autophagic puncta in the absence or presence of ConA. Data are mean ± s.d. (n = 10, biologically independent samples).

Extended Data
Different letters indicate significant differences (P < 0.05, one-way ANOVA); exact P values are provided in Supplementary Table 4. d, Immunoblot detection of OsNBR1 levels in the non-infested Bph14-Bisp transgenic lines, Bph14 and MH63 plants. Total proteins were immunoblotted and detected with anti-AtNBR1 and anti-ACTIN antibodies, respectively. Numbers under the lanes indicate protein abundance relative to that of ACTIN, quantified by ImageJ. e, Relative OsNBR1 transcript levels in the non-infested Bph14-Bisp transgenic lines, Bph14 and MH63 plants. P values were derived by one-way ANOVA. Data are means ± s.d. (n = 3, biologically independent experiments). f, Relative OsNBR1 transcript levels in Bph14 and MH63 plants after BPH infestation. P values were derived by one-way ANOVA. Data are mean ± s.d. (n = 3, biologically independent experiments). Experiments (a,b,d) were repeated at least three times, each giving similar results.