A conserved phospho-switch controls receptor kinase SDS2 activation and homeostasis in rice immunity

Abstract

Plant surface immune receptors are tightly controlled to maintain homeostasis and prevent overactivation. However, the mechanisms coordinating the balance remain largely unknown. The monocot-specific receptor-like kinase SDS2 regulates cell death and immunity in rice. Here, we identify Thr671, a conserved phospho-switch in RD-type Ser/Thr kinases across kingdoms, as the central node that orchestrates both immune activation and turnover of SDS2. Autophosphorylation of Thr671 initiates signaling but also primes SDS2 for ubiquitination and degradation by the E3 ligase SPL11. This degradation is antagonized by the phosphatase SIPP1, which stabilizes SDS2 through Thr671 dephosphorylation to promote immunity. Strikingly, a second phosphatase, SIPP2, displaces SIPP1 from SDS2 via competitive binding, thereby enabling SPL11-dependent SDS2 degradation and ensuring timely immune attenuation. Thus, the SIPP1-SPL11-SIPP2 module establishes a dynamic equilibrium essential for SDS2 homeostasis, providing new insights into the sophisticated regulation of surface receptors.

Dual phosphatases SIPP1 and SIPP2 orchestrate SDS2 immune dynamics from activation to attenuation.

INTRODUCTION

Surface immune receptors on the plasma membrane recognize pathogen-associated molecular patterns (PAMPs)—such as bacterial flagellin and LPS, or fungal chitin—to activate pattern-triggered immunity (PTI), and are thus termed pattern recognition receptors (PRRs) (1, 2). Plant PRRs fall into two structural classes: receptor-like kinases (RLKs), which contain an extracellular domain (ECD), a transmembrane helix, and a cytoplasmic kinase domain; and receptor-like proteins (RLPs), which lack the cytoplasmic kinase and require co-receptors for signal transduction. Both RLKs and RLPs depend on cytoplasmic kinases to propagate PTI signaling (2–4).

Plant RLKs can be divided into RD and non-RD kinases based on the conservation of an arginine (Arg) residue preceding the catalytic aspartate (Asp) in the HRD motif of catalytic loop. RD-kinases typically depend on phosphorylation in the activation loop to achieve full activity (5, 6). The majority plant RLK members are RD-type Ser/Thr kinases which phosphorylate on Ser/Thr residues of substrates (2, 4, 7). The RD-type Ser/Thr kinases possess unique features in protein sequence and structure, especially in the catalytic loop and activation loop (8, 9).

PRRs complexes are tightly controlled. Ubiquitination and dephosphorylation are two prevailing protein modifications occur on kinase domain for PRR regulation (10). Ubiquitination mediated PRRs degradation plays important roles in PTI negative regulation. Arabidopsis FLS2 is targeted by ubiquitin ligases (E3) PUB12/13 for ubiquitination and degradation, thus to attenuate immune responses (11). Arabidopsis chitin receptor LYK5 is under PUB13 negative regulation (12). In rice, the SD-1-type RLK SDS2 is negatively regulated by the E3 ligase SPL11 mediated ubiquitination and degradation (13).

Dephosphorylation is mediated by protein phosphatases and plays diverse roles in the regulation of PRR. The most common role by dephosphorylation is negative regulation on PRR activation (10). The LPS receptor LORE, an SD-1 RLK, confers basal xylem immunity in Arabidopsis against Ralstonia solanacearum. To prevent prolonged immunity, the PP2C phosphatases LOPP is recruited to dephosphorylate LORE to attenuate signaling (14). In the case of flagellin sensing, both FLS2 and its coreceptor BAK1 are targeted by protein phosphatases KAPP and PP2A respectively for dephosphorylation, thus to achieve down regulation of flagellin triggered immunity (15, 16). Rice XA21 is an LRR type RLK that sensing bacterial pathogen Xanthomonas oryzae pv. oryzae (Xoo) secreted peptide RaxX21(sulphated) and confers resistance to specific Xoo strains. XB15 directly dephosphorylate XA21 to attenuate PTI (17). Arabidopsis protein phosphatases POLTERGEIST-LIKE 4 and 5 (PLL4 and PLL5), orthologs of rice XB15, act as inhibitors of receptor complexes in both immunity and development which are released from the receptor complexes upon elicitation (18). Arabidopsis chitin receptor CERK1 is under CIPP1 dephosphorylation regulation which controls the switch of CERK1 states from standby to nascent, leading to recycling of the receptor (19).

Although the negative regulation is prevailing, the regulation of PRR homeostasis is largely unknown. Homeostasis is crucial for PRRs to maintain decent immune activation. An excess of PRRs could lead to autoimmunity that manifests as stunted growth and spontaneous cell death, which is often seen in PRR overexpression lines and in lesion mimic mutants (2, 10, 20, 21). On the other end, insufficient PRR levels leaves the plant vulnerable to pathogens which is exampled by PRR mutants (13, 22). SDS2 is a monocot specific RLK which positively modulates cell death and immunity through phosphorylation of OsRLCK118 and OsRLCK176 which in turn regulate ROS production. SDS2 mutants are susceptible to blast infection. In contrast, plus enhanced resistance, SDS2 overexpression lines demonstrates stunted growth and spontaneous lesions on leaves (13).

In this study, we identify Thr671 in the receptor-like kinase SDS2 as a conserved phospho-switch controlling immunity and turnover. Thr671 represents a critical phospho-switch conserved among RD-type Ser/Thr kinases across kingdoms. Immune activation triggers Thr671 autophosphorylation in SDS2, which promotes signaling but also primes SDS2 for SPL11-mediated ubiquitination and degradation. The phosphatase SIPP1 sustains SDS2 homeostasis by dephosphorylating Thr671, thereby antagonizing SPL11-mediated ubiquitination to promote defense. The second protein phosphatase, SIPP2, interacts with SDS2 and negatively regulates disease resistance by modulating the SDS2-SIPP1 interaction. Specifically, SIPP2 competes with SIPP1 for SDS2 binding to enable SPL11-dependent SDS2 degradation. This study elucidates a novel mechanism for immune receptor homeostasis centered on a conserved phospho-switch, which is dynamically regulated by the antagonistic actions of the SIPP1-SPL11-SIPP2 module.

RESULTS

Phosphorylation in the activation loop modulates SDS2 activity and stability

To investigate the regulatory mechanism of phosphorylation and stability of SDS2 in immunity, we conducted in vitro kinase assay (auto-phosphorylation) with bacterially purified SDS2 (cytoplasmic kinase domain) and determined the phosphorylation sites by mass spectrometry (MS). Three residues in the activation loop, Ser664, Thr667, and Thr671, were identified to be phosphorylated in the MS assay (fig. S1). Of which, Thr667 and Thr671 are conserved among multiple RD-type RLKs, as demonstrated by multiple alignment (Fig. 1A). Though not identified in the MS assay, Thr677 demonstrates high degree of conservation among RLKs, suggesting conserved role may share. Phosphorylation in the activation loop is critical for kinase activation (6). To investigate the role of phosphorylation on SDS2 activation, in vitro kinase assay was carried out with Thr-to-Ala (T-to-A) mutants of SDS2. T-to-A substitutions were introduced to mimic unphosphorylation (phospho-dead). GST-tagged cytoplasmic domain of wild type SDS2, SDS2^T667A, SDS2^T671A, and SDS2^T677A were purified and applied to in vitro kinase assay. Compared to GST-SDS2 and GST-SDS2^T667A, GST-SDS2^T671A and GST-SDS2^T677A exhibited reduced phosphorylation levels, suggesting that Thr671 and Thr677 are the primary auto-phosphorylation residues or are critical for kinase activity of SDS2 (Fig. 1B).

Fig. 1. Phosphorylation in the activation loop modulates SDS2 activity and stability.

(A) Schematic of SDS2 domains and sequence alignment of the activation loop. Red asterisks indicate Thr667, Thr671 and Thr677. (B) In vitro kinase assay to detect the autophosphorylation of SDS2 and its variants. All the GST-SDS2 recombinant protein and its variants used in this study is tagged to cytoplasmic domain (kinase domain). The phosphorylation level of SDS2 was detected with anti-pThr antibody. Protein loading is indicated by Coomassie brilliant blue (CBB) staining in the study. (C) Degradation of SDS2-HA and variants in protoplasts treated with 50 μM CHX. GFP is internal control. Numbers indicate relative protein levels. Protein levels (band intensities) were quantified using ImageJ software in this study. (D) Quantification of SDS2-HA levels in (C). Values are mean ± SD (n = 3). Different letters (a-b) indicate significant differences [one-way ANOVA least significant difference (LSD) test, P < 0.05]. (E) In vivo ubiquitination assay of SDS2-HA and variants co-expressing FLAG-Ub with 50 μM MG132. (F) Co-IP to detect the interactions between SPL11 and SDS2 variants in rice protoplasts. (G and H) Trans-phosphorylation assays. MBP-RLCK176^K108E (G) or MBP-RLCK118^K116E (H) were incubated with GST-SDS2, GST-SDS2^T671A, or GST-SDS2^T677A for 1 hour. Phosphorylation detected with anti-pThr/pTyr antibody. (I) Relative lesion area of punch inoculated leaves in ZH11, sds2 mutants and SDS2, SDS2^T667A, SDS2^T671A, and SDS2^T677A complemented lines. The lesion area on leaves was determined with ImageJ in this study. Values are mean ± SD (n = 18, 9, 9, 11, 11, 9, 10, 10, 9, 15, and 14, respectively, one-way ANOVA LSD test, P < 0.05). (J) Relative fungal biomass of the leaves in (I). Relative fungal biomass of leaves in (I) determined by MoPot2 qPCR, normalized to OsUBQ. Values are mean ± SD (n = 3, one-way ANOVA LSD test, P < 0.05).

To determine whether phosphorylation regulates SDS2 stability, we investigated the stability of SDS2 and its T-to-A mutants in rice protoplasts. HA-tagged SDS2, SDS2^T667A, SDS2^T671A, and SDS2^T677A were transiently expressed in rice protoplasts, treated with protein synthesis inhibitor cycloheximide (CHX), and applied to immunoblot. SDS2-HA, SDS2^T667A-HA, and SDS2^T677A-HA were rapidly degraded after 3 h treatment with CHX, whereas SDS2^T671A-HA demonstrated largely reduced degradation (Fig. 1, C and D). We also examined the stability of SDS2 and its T-to-D (phospho-mimic) mutants in rice protoplasts. SDS2^T671D-HA accumulated to much lower levels than SDS2-HA, SDS2^T667D-HA, and SDS2^T677D-HA. Following 3 hours of CHX treatment, SDS2^T671D-HA was barely detectable (fig. S2A). Because SDS2 undergoes ubiquitination-mediated degradation, we wonder whether Thr phosphorylation controls SDS2 ubiquitination. To this end, we examined the ubiquitination levels of SDS2 and its T-to-A mutants in rice protoplasts. The ubiquitination level of SDS2^T671A-HA was much lower than that of SDS2-HA, SDS2^T667A-HA, or SDS2^T677A-HA (Fig. 1E and fig. S2B). Consistent with the reduced ubiquitination level, non-ubiquitinated SDS2^T671A-HA accumulated to a much higher level than that of SDS2-HA or other mutants (Fig. 1E). We also examined the ubiquitination levels of SDS2 and its T-to-D mutants in rice protoplasts. The ubiquitination levels of SDS2^T671D-HA were markedly enhanced compared to SDS2-HA, SDS2^T667D-HA, and SDS2^T677D-HA (fig. S2C). Previous study revealed that SPL11 interacts with and ubiquitinates SDS2 for degradation (13). Thus, the reduced ubiquitination level of SDS2^T671A-HA is probably attributed to impaired interaction with SPL11. To test this possibility, we performed Co-IP assay to assess the interaction between SPL11 and SDS2 or its T-to-A variants. Compared to wild-type or SDS2^T677A, SDS2^T671A demonstrated a markedly weakened interaction with SPL11 (Fig. 1F). Noteworthy, SDS2^T671A-HA accumulated to a much higher level in this assay, which is consistent with the reduced degradation in the stability assay (Fig. 1C) and with reduced ubiquitination level in the ubiquitination assay (Fig. 1E). We also performed in vitro pull-down assays to examine the interaction between SPL11 and SDS2, SDS2^T671A, or SDS2^T671D. Compared to GST-SDS2, GST-SDS2^T671D showed substantially stronger binding to MBP-SPL11 whereas GST-SDS2^T671A exhibited markedly reduced interaction (fig. S2D). Collectively, these results indicate that the auto-phosphorylation of the Thr671 residue controls SDS2 interaction with SPL11 and its stability, whereas the Thr677 residue does not.

Previous study showed that SDS2 modulates rice immunity by phosphorylating OsRLCK118/176 (13). Since Thr671A or Thr677A mutations impair SDS2 auto-phosphorylation or kinase activity, it is not clear whether trans-phosphorylation against OsRLCK118 and OsRLCK176 is affected. To examine this, GST-SDS2 or its T-to-A variants alone or together with MBP-RLCK were applied to in vitro kinase assay. Compared to GST-SDS2, GST-SDS2^T671A demonstrated reduced trans-phosphorylation activity while GST-SDS2^T677A totally lost the activity (Fig. 1, G and H), suggesting that Thr671 and Thr677 are critical for SDS2 kinase activity. Altogether, phosphorylation on Thr671 in the activation loop is critical for the kinase activity and stability of SDS2.

To verify the functional roles of the Thr671 and Thr677 sites in rice disease resistance, we generated sds2 CRISPR-Cas9 knockout mutants (sds2–4 and sds2–28) (fig. S3A). With sds2–4 mutant line, we generated stable transgenic rice lines expressing SDS2-YFP, SDS2^T667A-YFP, SDS2^T671A-YFP, or SDS2^T677A-YFP driven by the native SDS2 promoter. Protein abundance analysis revealed that SDS2^T671A-YFP accumulated to markedly higher levels than SDS2-YFP or other mutants in the complemented lines (fig. S3B), supporting that Thr671 is critical for the SDS2 stability. Following inoculation with Magnaporthe oryzae, the sds2 knockout mutants exhibited enhanced susceptibility to blast fungus. SDS2-YFP and SDS2^T667A-YFP complemented lines partially restored resistance comparable to wild-type plants. However, SDS2^T677A-YFP transgenic lines failed to complement sds2 mutation. Notably, SDS2^T671A-YFP transgenic lines conferred even stronger resistance than wild-type plants (Fig. 1, I and J, and fig. S3C), which could be explained by the combination of basal level of kinase activity and highly accumulation of SDS2^T671A. Although the T671A substitution retains basal level of kinase activity (Fig. 1, G and H), it reduces SPL11-mediated degradation (described later). Taken together, phosphorylation of T671 in the activation loop modulates SDS2 activity and stability.

Protein phosphatase SIPP1 interacts with SDS2 and positively regulates disease resistance

PRR activation is transient and dynamic (4, 23, 24). To determine the dynamics of SDS2 phosphorylation and abundance upon immune activation, SDS2-HA was expressed in rice protoplasts, treated with chitin or flg22 at time points of 15, 30, and 45 min, and subjected to immunoprecipitation and immunoblotting. The result revealed that both SDS2 phosphorylation and abundance peaked at 15 min post-treatment and declined thereafter (fig. S4, A to F), suggesting protein phosphatases may involve in the stability and phosphorylation of SDS2. To further verify that the transient accumulation of SDS2 upon PAMP treatment is due to its protein stability, we performed a time-course degradation assay of SDS2. The results showed that SDS2 degradation was readily detectable at 15 min after CHX treatment, and this degradation was significantly inhibited upon MG132 treatment (fig. S4, G and H), indicating SDS2 undergoes a rapid turnover which is mediated by 26S proteasome.

To investigate how SDS2 stability is controlled by dephosphorylation, we performed immunoprecipitation coupled with mass spectrometry (IP-MS) to identify SDS2 associated protein phosphatase(s). SDS2-HA fusion protein was expressed in rice protoplasts, treated with flg22, and applied to IP-MS. Two protein phosphatases were identified to interact with SDS2 and designated SDS2-Interacting Protein Phosphatase 1 (SIPP1, LOC_Os05g04370) and SDS2-Interacting Protein Phosphatase 2 (SIPP2, LOC_Os02g13110) (table S1). SIPP1 belongs to a putative RNA polymerase II subunit B1 C-terminal domain (CTD) phosphatase, also known as C-terminal domain phosphatase-like (CPL), which involve in RNA polymerase II transcription regulation and in other cellular processes. To confirm the interaction between SIPP1 and SDS2, SDS2-HA is co-expressed with SIPP1-GFP or GFP in rice protoplasts with or without chitin treatment and subjected to Co-IP. As detected by immunoblotting, SDS2-HA co-immunoprecipitated with SIPP1-GFP, but not with GFP control. A weak interaction was detected without chitin treatment, and the interaction between SDS2-HA and SIPP1-GFP was significantly enhanced after chitin treatment for 15 min (Fig. 2A). To further confirm the interaction, bimolecular fluorescence complementation (BiFC) assay was performed in rice protoplasts with or without chitin treatment. SDS2-VC (SDS2 fused with VC) and SIPP1-VN (SIPP1 fused with VN) were co-transfected into rice protoplasts and observed with confocal 20 hours after transfection. Yellow fluorescence was observed in the plasmamembrane of protoplasts co-expressing SDS2-VC and SIPP1-VN rather than in the controls, and was detectable in the absence or presence of chitin treatment (Fig. 2B). In addition, pull-down assay was conducted to confirm SIPP1-SDS2 interaction. Purified GST-SDS2 was incubated with His-SIPP1 or His-PP70 (as a negative control) in the presence of glutathione conjugated agarose beads and applied to immunoblot. The result showed that His-SIPP1 was co-purified with GST-SDS2 while His-PP70 was not (Fig. 2C). To investigate whether Thr671 phosphorylation influence SIPP1-SDS2 interaction, a pull-down assay was conducted. The result demonstrated that the T671A or T677A substitutions does not impair the binding to SIPP1 (fig. S5), suggesting that SIPP1 may constitutively bind SDS2. Collectively, these results indicate that SIPP1 interacts with SDS2.

Fig. 2. SIPP1 interacts with SDS2 and positively regulates rice disease resistance.

(A) Co-IP assay to detect the interaction between SDS2-HA and SIPP1-GFP in rice protoplasts with or without chitin treatment for 15 min. Co-IP assay was performed with anti-GFP beads. (B) BiFC assay to detect the interaction between SDS2 and SIPP1 in rice protoplasts with or without chitin treatment for 15 min. Co-transfected protoplasts were observed with confocal 20 hours. after transfection. Scale bars, 10 μm. (C) GST pull-down detecting interaction between GST-SDS2 and His-SIPP1. His-PP70 as negative control. (D) Disease symptoms on the leaves of the ZH11, sipp1–1, and sipp1–2 plants after punch inoculation with the M. oryzae isolate RB22. The images were taken at 14 days after inoculation. (E) Relative lesion area of punch inoculated leaves in (D). Values are mean ± SD (n = 8, 7, and 7, respectively, ***P < 0.001, **P < 0.01, Student’s t test). (F) Relative fungal biomass of the leaves in (D). Values are mean ± SD (n = 3,***P < 0.001, Student’s t test). (G) ROS production in ZH11, sipp1–1, and sipp1–2 plants. ROS burst was induced with the sheath of 10-day-old seedlings by 10 μg/ml chitin and determined with L-012 based chemiluminescence assay. (H) Disease symptoms on the leaves of the ZH11, SIPP1-OE-23, and SIPP1-OE-29 plants after punch inoculation with the M. oryzae isolate RB22. (I) Relative lesion area of punch inoculated leaves in (H). Values are mean ± SD (n = 9, 9, and 7, respectively, **P < 0.01, Student’s t test). (J) Relative fungal biomass of the leaves in (H). Values are mean ± SD (n = 3, ***P < 0.001, Student’s t test). (K) ROS production in ZH11, SIPP1-OE-23 and SIPP1-OE-29 plants with 10 μg/ml chitin treatment.

To investigate the function of SIPP1 in rice immunity, disease resistance and immune responses were analyzed in sipp1 mutants and SIPP1-overexpression lines. Two independent sipp1 knockout mutants (sipp1–1 and sipp1–2) were generated in the ZH11 background with CRISPR/Cas9 technology (fig. S6A). The sipp1 mutant lines were inoculated with Magnaporthe oryzae (isolate RB22) to examine the disease resistance. The sipp1 mutants displayed severer symptoms with larger lesions and more fungal biomass than that with ZH11 (Fig. 2, D and E). To analyze the immune response, PAMP-induced ROS bursts were examined in the sipp1 mutant lines. The sipp1 mutant lines generate less ROS than ZH11 did (Fig. 2G and fig. S6B). To further confirm SIPP1 functions, we constructed SIPP1 overexpressed lines and analyzed the phenotypes of disease resistance and PTI response. HA-tagged SIPP1, driven by the promoter of maize ubiquitin, was introduced into ZH11 cultivar. Two homozygous SIPP1-overexpressing lines, SIPP1-OE-23 and SIPP1-OE-29, were chosen for fungus inoculation and ROS induction (fig. S6, C and D). Both punch- and spray-inoculation were performed in the SIPP1 overexpressing lines with isolate RB22. The results demonstrated that SIPP1-overexpressing lines were more resistant than ZH11 (Fig. 2, H to J, and fig. S6, E and F). Consistently, SIPP1-overexpressing lines generated more ROS than ZH11 did (Fig. 2K and fig. S6G). Altogether, SIPP1 positively regulates disease resistance and PTI responses in rice.

SIPP1 controls SDS2 turn-over by dephosphorylating Thr671

To investigate how SIPP1 regulates SDS2-mediated rice immunity, we analyzed the dephosphorylation of SIPP1 on SDS2. Firstly, we performed in vitro dephosphorylation assay to examine whether SIPP1 can dephosphorylate SDS2. The result showed that SDS2 phosphorylation decreased progressively with increasing His-SIPP1 levels, no reduction was observed with His-PP70 (an irrelevant phosphatase) or His-SIPP2 (a regulatory PP2A β-subunit that lacks intrinsic phosphatase activity in vitro) (Fig. 3A), indicating specific dephosphorylation of SDS2 by SIPP1. A series of Asp-to-Ala (D-to-A) substitutions were introduced to SIPP1 to determine the active residues of SIPP1, because Asp act as catalytic residues in CTD protein phosphatases (25, 26). The D-to-A mutants were examined phosphatase activities and found that the D270A mutant demonstrated abolished phosphatase activity on GST-SDS2 (Fig. 3B and fig. S7A), suggesting Asp270 is required for SIPP1 phosphatase activity. Although Asp 270 is likely involved in catalysis, indirect effects on substrate binding or conformational dynamics cannot be completely excluded.

Fig. 3. SIPP1 controls SDS2 turn-over by dephosphorylating Thr671.

(A) SIPP1 dephosphorylates SDS2 in the in vitro dephosphorylation assay. Purified GST-SDS2 was incubated with His-SIPP1, His-SIPP2, or His-PP70 for 1 hour. SDS2 phosphorylation was detected with anti-pThr antibody. (B) D270A substitution abolished phosphatase activity of SIPP1. His-SIPP1 and His-SIPP1^D270A were incubated with GST-SDS2 in the in vitro dephosphorylation assay. SDS2 phosphorylation was detected with anti-pThr antibody. (C) The abundance and phosphorylation levels of SDS2-myc in the transfected protoplasts of wild-type (ZH11) and of sipp1 mutants. SDS2-myc was transiently expressed in the ZH11, sipp1–1 and sipp1–2 rice protoplasts. Transfected rice protoplasts were treated with 10 μg/ml chitin for 15 min. Total protein from rice protoplasts were immunoprecipitated with anti-myc beads and immunoblotted with anti-pThr and anti-myc antibodies. GFP is internal control. (D) The abundance and phosphorylation levels of SDS2-myc in the transfected protoplasts of wild-type (ZH11) and of SIPP1 overexpression lines treated with 10 μg/ml chitin for 15 min. (E) MBP pull-down assay to detect the interaction between MBP-SIPP1 or its variants and GST-SDS2. (F) SDS2-myc was co-expressed with SIPP1-HA or SIPP1^R205A-HA in rice protoplasts with chitin treatment for 15 min. (G) SIPP1 dephosphorylated SDS2 at Thr671 in vitro. Purified GST-SDS2 or GST-SDS2^T671A was incubated with His-SIPP1, His-SIPP2, or His-PP70 for 1 hour in the in vitro dephosphorylation assay. SDS2 phosphorylation was detected with anti-pThr and anti-pT671 antibodies. (H) SDS2 phosphorylation on Thr671 is increased in the protoplasts of sipp1 mutants. SDS2-myc was transiently expressed in the protoplasts of ZH11, sipp1–1 and sipp1–2 and treated with chitin for 15 min. Total protein from rice protoplasts were immunoprecipitated with anti-myc beads and detected with anti-pT671 and anti-myc antibodies. (I) SDS2 phosphorylation on Thr671 is reduced in the protoplasts of SIPP1-overexpression lines with chitin treatment for 15 min.

To validate dephosphorylation of SDS2 by SIPP1, we examined the phosphorylation levels of SDS2 in rice protoplasts of sipp1 mutants and of SIPP1 over-expressed lines. SDS2-myc was transiently expressed in sipp1 mutant rice protoplasts, treated with chitin or flg22, and was subsequently subjected to immunoprecipitation and immunoblot analysis to determine its phosphorylation levels. The results showed that SDS2 phosphorylation levels were elevated in the mutants compared to that in ZH11 (Fig. 3C and fig. S7, B to F). In contrast, the phosphorylation levels of SDS2 in the SIPP1 over-expressed lines is much lower than that in ZH11 (Fig. 3D and fig. S7, G to K). These results indicate that SIPP1 dephosphorylates SDS2 in vivo. Intriguingly, SDS2 protein abundance are less in the sipp1 mutants while much more in the SIPP1 over-expressed lines compared to that in ZH11 (Fig. 3, C and D, and fig. S7, C, D, F, H, I, and K), suggesting that SIPP1 modulates SDS2 homeostasis. To further confirm whether the dephosphorylation and stabilization of SDS2 by SIPP1 requires their direct interaction, we performed structure-guided mutagenesis to identify critical SIPP1 residues that mediating SDS2 binding. Based on the predicted interaction interface, we generated a series of SIPP1 point mutants, including D196A, E199A, R205A, D543A, S550A, and T552A. MBP pull-down assays revealed that the MBP-SIPP1^R205A mutant completely failed to interact with SDS2 (Fig. 3E). To assess the functional consequences of disrupting this interaction, we co-expressed SDS2 with either SIPP1-HA or SIPP1^R205A-HA in rice protoplasts and examined protein abundance and phosphorylation levels. SIPP1-HA effectively inhibited chitin-induced phosphorylation of SDS2 and promoted SDS2 protein accumulation. In contrast, SIPP1^R205A-HA had no effect on either the phosphorylation status or the protein abundance of SDS2 (Fig. 3F), indicating that direct binding to SDS2 is required for SIPP1-mediated regulation. Taken together, SIPP1 modulates SDS2 homeostasis by dephosphorylating the latter.

Thr671 is critical for SDS2 stability, as demonstrated by the degradation assay and ubiquitination assay with SDS2^T671A (Fig. 1, C to E). It is not clear whether SIPP1 targets phosphorylation on Thr671 thus to modulate SDS2 homeostasis. To determine whether SIPP1 dephosphorylates SDS2 at Thr671, we raised a specific antibody against synthesized SDS2 peptide with phosphorylation on Thr671 (CETQANpTNRVV). The specificity of anti-pThr671 was confirmed by immunoblot. Purified GST-SDS2 (auto-phosphorylated) was treated with Lambda protein phosphatase (λPP) and immunoblotted with anti-pThr671 and anti-pThr. Like with anti-pThr, the immunoblot signals with anti-pThr671 disappeared after λPP treatment (fig. S8A). To further validate the specificity of the pT671 antibody, we performed in vitro phosphorylation assays using recombinant GST-SDS2, GST-SDS2^T671A, and GST-SDS2^T667A proteins. Following in vitro kinase reactions, the phosphorylation status of these proteins was examined using the pT671 antibody. The pT671 antibody detected a strong signal in GST-SDS2 and GST-SDS2^T667A but not in the GST-SDS2^T671A mutant (fig. S8B), demonstrating the specificity of the phospho-Thr671 antibody. To determine whether SIPP1 dephosphorylates SDS2 at Thr671, in vitro dephosphorylation assay was conducted. Immunoblot analysis with anti-pThr and anti-pThr671 antibodies showed that His-SIPP1, but not His-SIPP2 or His-PP70, markedly reduced both total threonine phosphorylation and specific Thr671 phosphorylation, and SIPP1 exhibited no dephosphorylation effect on T671A, suggesting that T671 is the critical dephosphorylation site targeted by SIPP1 (Fig. 3G). To confirm that SIPP1 dephosphorylates SDS2 at Thr671, we examined the Thr671 phosphorylation of SDS2 in rice protoplasts of SIPP1 mutants and over-expressed lines. After chitin or flg22 treatment and immunoprecipitation, Thr671 phosphorylation levels were dramatically reduced in SIPP1 over-expressed lines while elevated in sipp1 mutants, compared to that in ZH11 (Fig. 3, H and I, and fig. S8, C to H). Noteworthy, SDS2-myc accumulated to a much higher levels in SIPP1 over-expressed lines but to decreased levels in sipp1 mutants (Fig. 3, H and I), which is consistent with the findings in the above assays. Taken together, SIPP1 negatively regulates SDS2 turn-over by dephosphorylating SDS2 at Thr671.

SIPP1 antagonizes SPL11-mediated SDS2 ubiquitination and degradation

As revealed above, SIPP1 over-expression enhanced the resistance to blast fungus (Fig. 2). Does the enhanced resistance rely on SDS2? To answer this question, the sds2 knockout mutants were crossed with the SIPP1 over-expressed line SIPP1-OE-29 to generate SIPP1-OE-29/sds2 hybrids, which were subjected to inoculation (fig. S9A). Punch inoculation was performed with M. oryzae isolate RB22. Unlike SIPP1 over-expressed line SIPP1-OE-29, the line SIPP1-OE-29/sds2 demonstrated reduced resistance to blast fungus with larger lesions and elevated fungal biomass, which is comparable to that of sds2 mutant (Fig. 4, A to C), indicating the enhanced resistance conferred by SIPP1 over-expression relies on SDS2.

Fig. 4. SIPP1 modulates SPL11-mediated SDS2 ubiquitination and degradation.

(A) Disease symptoms on the leaves of the ZH11, SIPP1-OE-29, sds2–4, and SIPP1-OE-29/sds2–4 plants after punch inoculation with the M. oryzae isolate RB22. (B and C) Relative lesion area (B) and relative fungal biomass (C) of inoculated leaves in (A). Values are mean ± SD [(B) n = 12, 13, 10, and 10, respectively; (C) n = 3], Different letters indicate statistically significant differences according to one-way ANOVA LSD test (P < 0.05). (D) SIPP1 enhances SDS2 accumulation which is dependent on SPL11. SDS2-myc or SDS2^T671A-myc were co-expressed with SIPP1-HA, SIPP2-HA, or SIPP1^D270A-HA in rice protoplasts of wild type IR64 and spl11 mutant with or without chitin treatment for 15 min. (E) SIPP1 inhibits the interaction between SDS2 and SPL11 in vivo. Total proteins from transfected protoplasts were subject to Co-IP assays with anti-myc beads and immunoblotted with anti-HA, anti-myc, and anti-GFP antibodies. (F) In vivo ubiquitination assay to show the ubiquitination levels of SDS2-myc or SDS2^T671A-myc co-expressed with SIPP1, SIPP1^D270A, or SIPP2 in rice protoplasts. FLAG-Ub was co-expressed in rice protoplasts in the presence of 50 μM MG132. The poly-ubiquitination of SDS2-myc was detected by immunoblotting with anti-myc and anti-FLAG antibodies.

As previously reported, SDS2 ubiquitination and degradation depends on SPL11 (13). To investigate whether SIPP1-mediated SDS2 turn-over requires SPL11, SDS2-myc was expressed alone or together with SIPP1-HA, SIPP1^D270A-HA, or SIPP2-HA in the rice protoplasts of spl11 mutant or wild-type IR64, respectively. In the IR64 protoplasts, SDS2-myc accumulated to a higher level co-expressed with SIPP1-HA than that with SIPP2-HA or SIPP1^D270A-HA. Noteworthy, SDS2^T671A-myc accumulated to a much higher level and SIPP1-HA failed to enhance the accumulation of SDS2^T671A-myc. In contrast, SDS2-myc levels in the spl11 protoplasts did not differ from each other (Fig. 4D), indicating SIPP1 mediated SDS2 accumulation depends on SPL11 and the regulation is determined by Thr671 in SDS2.

To investigate whether SIPP1 affects the SDS2-SPL11 interaction and SDS2 ubiquitination, SDS2-HA or SDS2^T671A-HA and SPL11-myc were co-expressed with increasing amounts of SIPP1-GFP in rice protoplasts. Protein samples from the protoplasts were subjected to Co-IP with anti-myc beads and followed by immunoblot. With the amounts of SIPP1-GFP increasing, the SPL11-myc co-purified SDS2-HA decreased accordingly, even though the total amounts of SDS2-HA accumulated higher. Notably, compared to SDS2-HA, SDS2^T671A-HA exhibited markedly reduced interaction with SPL11-myc and this interaction was not further weakened by SIPP1-GFP co-expression (Fig. 4E). The result suggests that SIPP1 suppresses SDS2-SPL11 interaction which could be explained by SIPP1 dephosphorylation of SDS2 or by competition between SIPP1 and SPL11. To rule out the possibility of competition, we performed a competitive binding assay to investigate whether SIPP1 competitively bind SDS2 against SPL11. GST-SDS2 was incubated with MBP-SPL11, with addition of gradient amounts of His-SIPP1. The mixtures were subjected to glutathione beads binding and followed by immunoblot. The result demonstrated that SDS2-SPL11 binding was not interrupted by the addition of His-SIPP1 (fig. S9B), suggesting absence of competition between SPL11 and SIPP1 in binding SDS2. Furthermore, the binding interfaces on SDS2 are probably different from each other. To test whether SIPP1 interact with SPL11, pulldown assay was conducted. The result showed that SIPP1 did not interact with SPL11 (fig. S9C), which supports the absence of competition between SIPP1 and SPL11. To determine whether SIPP1 inhibits SPL11-mediated ubiquitination of SDS2, we performed an in vivo ubiquitination assay. SDS2-myc and FLAG-Ub were co-expressed with SIPP1-HA, SIPP2-HA, or SIPP1^D270A-HA in rice protoplasts and subjected to Co-IP and immunoblot. SDS2-myc demonstrated dramatically reduced ubiquitination level when co-expressed with SIPP1-HA, but no difference was observed when co-expressed with SIPP1^D270A-HA or with SIPP2-HA. Interestingly, compared to wild type SDS2-myc, SDS2^T671A-myc demonstrated markedly reduced ubiquitination and showed no further reduction when co-expressed with SIPP1-HA (Fig. 4F), suggesting that SIPP1 suppresses SDS2 ubiquitination by dephosphorylation SDS2 Thr671. The ubiquitination levels between SDS2 variants are consistent with the interactions between SDS2 variants and SPL11 (Fig. 4, E and F), which supporting the finding that Thr671 phosphorylated SDS2 is the bona fide substrate of SPL11. Overall, SIPP1 modulates SPL11-mediated SDS2 ubiquitination and degradation thus to control the homeostasis of SDS2.

Thr671 represents a conserved phospho-switch modulating RD-type Ser/Thr kinases activity

Protein kinases are evolutionarily ancient components that widespread across all kingdoms (9, 27). Based on the presence or absence of a conserved HRD motif in the catalytic loop, the protein kinases can be divided into RD type or non-RD type. The majority Ser/Thr kinases, which phosphorylate at Ser or Thr on substrates, belong to RD-type kinase (28). Most protein kinases are activated by phosphorylation on Ser/Thr residues within the activation-loop (6, 29). Activated kinases adopt to a unique conformation that is conserved among all the structurally determined kinases (5, 9, 30). SDS2 belongs to a typical RD-type Ser/Thr kinase. To examine the conservation of SDS2 Thr671 and Thr677, we made multiple sequence alignment with RD-type Ser/Thr kinases from plants and animals in the regions of catalytic-loop and activation loop. The alignment was visualized on sequence logo. RD motif, DFG motif, APE motif, and a couple of residues, including Lys^RD+2, Asn^RD+5, Arg/Lys^DFG+3, Thr^APE-5, and Thr^APE-10/−11, are highly conserved (Fig. 5A), suggesting indispensable role in kinase activity. Optimal phosphotransfer, or kinase activity, requires the precise spatial arrangement of several catalytic residues including those highly conserved ones (5, 30). To demonstrate the spatial arrangement, the structure of active PKA (PDB: 1atp) was showed in catalytic-loop and activation loop with conserved residues highlighted in green, which forms two contact clusters (Fig. 5B).

Fig. 5. Thr671 represents a conserved phospho-switch for RD-type Ser/Thr kinases.

(A) Sequence logo of multiple sequence alignment of catalytic loop and activation loop RD-type Ser/Thr kinases from plants (upper panel, n = 4934) and animals (lower panel, n = 312). The plots are the frequency of amino acids across the loops. The conserved residues are denoted with triangles and names above. +, − indicate the residue(s) downstream or upstream of specific motif. The gap indicates variable region. (B) Conserved residues highlighted in green form two contact clusters in active PKA (PDB: 1atp). (C) Superimposition of predicted SDS2 (with phosphorylation on Thr671) and crystallographic structures of 5 active kinases active kinases (PDB: 1atp, 1ctp, 1jst, 2erk, 1o6l) in the region of catalytic loop and activation loop. The RMSD values between SDS2 and the others are labeled. (D) Schema of the interactions between highly conserved residues in the determined kinase structures. The lines indicate interaction between residues. Red, salt bridge; blue, hydrogen bond. (E to G) Kinase assay on selected kinases with T-to-A or T-to-D substitutions. MoPMK1 (F) and ERK2 (G) are RD-type Ser/Thr kinases from M. oryzae and rat respectively. Phosphorylation is immunoblotted with anti-pThr/pTyr and anti-pSer.

Thr671 and Thr677 of SDS2 are corresponding to Thr^APE-10/−11 and Thr^APE-5, respectively. To understand the roles of the two Thr in kinases activity, we aligned the structures of five RD-type Ser/Thr kinases (PDB: 1atp, 1ctp, 1jst, 2erk, 1o6l). The five structures are chosen because they are confirmed to be in the active form and the structures of their inactive form are available. In addition, they are belonging to different kinase categories with varying functions. The structures are well aligned with each other in the regions of catalytic loop and activation loop, especially for the highly conserved residues (Fig. 5C). The overall structure and orientation are almost the same, suggesting conserved roles in kinase activity. Most importantly, interactions occur among the conserved residues (Fig. 5D). Thr^APE-10/−11, corresponding to Thr671 in SDS2, are all phosphorylated and interacts with Arg^RD motif and Arg/Lys^DFG+3 via salt bridge. Thr^APE-5, corresponding to Thr677 in SDS2, interacts with Asp^RD and Lys^RD+2 via hydrogen bond. In addition, Asn^RD+2 interacts via hydrogen bond with two conserved Asp, Asp in DFG motif and Asp in RD motif. The interacting residues form a highly conserved network in the structures (Fig. 5, C and D), which suggests a similar activation mechanism are shared among the RD-type Ser/Thr kinases across kingdoms. To verify whether SDS2 activates in the same way, the structure of SDS2 kinase domain with phosphorylation on Thr671 was predicted with AlphaFold3 and showed high pLDDT values across the sequence (the pLDDT confidence score for the phosphorylated Thr671 residue is 83.20; fig. S10A). The predicted structure showed a similar overall arrangement and orientation for those conserved residues and was well aligned to the crystallographic structures of active kinases, with RSMD values vary from 0.610 to 1.685 Å (Fig. 5C and fig. S10B), suggesting SDS2 shares high similarity in structure and the same mechanism in activation as those determined ones.

Since the interactions between Thr^APE-10/−11 and Arg^RD or Arg/Lys^DFG+3 are mediated by salt bridge, the Thr^APE-10/−11 must be phosphorylated and serve as a switch. To confirm the role of Thr^APE-10/−11 in kinase activity, T-to-A and T-to-D substitutions were made for 8 randomly selected kinases and for SDS2. Kinase activity was tested for all the mutants and wildtype. All the tested T-to-A mutations at Thr^APE-10/−11 position (mimicking unphosphorylation) largely abolished kinase activity. In contrast, most T-to-D substitutions (phospho-mimic) demonstrated elevated kinase activity compared to T-to-A mutations (Fig. 5, E to G, and fig. S10, C to H), suggesting activation of the kinases. In comparison to wildtype kinases, the T-to-D substitutions show enhanced, comparable, or weakened kinase activity, suggesting the strength of phospho-mimics varies among kinases. The results largely support the notion that phosphorylation at Thr^APE-10/−11 position activate the kinases, same as those confirmed by crystallography. Actually, the role of phosphorylation at Thr^APE-11/−10 on kinase activity was extensively reported and is consistent with the findings here (31–33). Taken together, SDS2 Thr671 serves as a phospho-switch which is conserved among RD-type Ser/Thr kinases.

Protein phosphatase SIPP2 interacts with SDS2 and negatively regulates disease resistance

The phosphatase SIPP2, a regulatory β-subunit of PP2A also identified in the SDS2 IP-MS analysis, lacks intrinsic phosphatase activity in vitro. To confirm the SIPP2-SDS2 interaction, we performed co-immunoprecipitation (Co-IP) in rice protoplasts with or without chitin treatment. SDS2-HA co-precipitated with SIPP2-GFP only after chitin treatment. No interaction was detected without chitin treatment or with the GFP control (Fig. 6A). Furthermore, a pull-down assay demonstrated a direct interaction between purified GST-SDS2 and His-SIPP2 (Fig. 6B). Finally, BiFC assays revealed that the fluorescence of SDS2-VC and SIPP2-VN reconstituted at the plasma membrane only after chitin treatment. No fluorescence was detected at the plasma membrane without chitin treatment or in the controls (Fig. 6C). Collectively, these results indicate that SIPP2 interacts with SDS2.

Fig. 6. SIPP2 interacts with SDS2 and negatively regulates rice disease resistance.

(A) Co-IP assay to detect the interaction between SDS2-HA and SIPP2-GFP in rice protoplasts with or without chitin treatment for 15 min. Co-IP assay was performed with anti-GFP beads, and the immuno-precipitated complex was detected with anti-GFP and anti-HA antibodies. (B) GST pull-down to detect the interaction between SDS2 and SIPP2. His-SIPP2 was incubated with GST-SDS2 or GST and detected with anti-His and anti-GST antibodies. (C) BiFC assay to detect the interaction between SDS2 and SIPP2 in rice protoplasts with or without chitin treatment for 15 min. Co-transfected protoplasts were observed with confocal 20 hours after transfection. Scale bars, 10 μm. (D) Disease symptoms on the leaves of the ZH11, sipp2–7, and sipp2–12 plants after punch inoculation with the M. oryzae isolate RB22. The images were taken at 14 days after inoculation. (E) Relative lesion area of punch inoculated leaves in (D). The lesion area on leaves was determined with ImageJ. Values are mean ± SD (n = 9, 8, and 6, respectively, ***P < 0.001, Student’s t test). (F) Relative fungal biomass of the leaves in (D). The fungal biomass was determined by measuring the content of M. oryzae transposable element MoPot2 via quantitative PCR. OsUBQ was used as the reference gene. Values are mean ± SD (n = 3, ***P < 0.001, Student’s t test). (G and H) ROS production in ZH11, sipp2–7, and sipp2–12 plants. ROS burst was induced with the sheath of 10-day-old seedlings by 10 μg/ml chitin (G) or 100 μM flg22 (H) and determined with L-012 based chemiluminescence assay.

To investigate SIPP2 function in rice immunity, we analyzed disease resistance and immune responses in CRISPR/Cas9-generated sipp2 mutants (sipp2–7, sipp2–12; fig. S11A). Both punch- and spray-inoculation were performed in the sipp2 mutants with M. oryzae isolate RB22. Upon inoculation, sipp2 mutants developed significantly smaller lesions and contained less fungal biomass than the ZH11 control, indicating enhanced resistance (Fig. 6, D to F, and fig. S11, B and C). Accordingly, the sipp2 mutants also exhibited a stronger chitin or flg22-induced ROS burst compared to ZH11 (Fig. 6, G and H). These results demonstrate that SIPP2 acts as a negative regulator of both disease resistance and PTI in rice.

SIPP2 competes with SIPP1 for SDS2 binding in the late phase of PTI

To determine whether SIPP2 dephosphorylates SDS2, we evaluated SDS2 phosphorylation levels in sipp2 mutant protoplasts treated with chitin or flg22 for 15 min. SDS2-myc was transiently expressed and immunoprecipitated, revealing higher phosphorylation levels in the mutants compared to ZH11. In contrast, Thr671 phosphorylation levels showed no difference between sipp2 mutants and ZH11 (Fig. 7A and fig. S12, A to H). Notably, SDS2 protein accumulation was unaltered in the sipp2 mutants after a 15-min chitin or flg22 treatment (Fig. 7A and fig. S12, B and F). To clarify this, we co-expressed SDS2 with SIPP1, SIPP2, or the phosphatase-dead SIPP1^D270A in protoplasts and treated them with chitin or flg22 for 15 min. Both SIPP1 and SIPP2 suppressed SDS2 phosphorylation. Interestingly, SIPP1 co-expression promoted SDS2 protein accumulation while SIPP2 had no effect (Fig. 7, B and C). To investigate how SIPP1 and SIPP2 co-regulate SDS2 phosphorylation, we co-expressed SDS2-myc with SIPP1-GFP and SIPP2-HA in rice protoplasts and performed a time-course Co-IP with chitin or flg22 treatment. The binding between SIPP1 and SDS2 peaked at 15 min and subsequently weakened gradually. In contrast, the interaction between SIPP2 and SDS2 was progressively strengthened with prolonged chitin or flg22 treatment (Fig. 7, D and E). And the SDS2 protein level peaked at 15 min before declining (Fig. 7, D and E), a pattern that aligned with its accumulation dynamics upon PAMP treatment (fig. S4, A to F). These results prompted us to examine SDS2 protein accumulation in sipp2 mutants at 45 min after PAMP treatment. The results showed that when treated with chitin for 45 min, both SDS2 phosphorylation levels and protein accumulation were higher in the sipp2 knockout mutants than in ZH11 (Fig. 7F and fig. S12, I and J). The opposite interaction dynamics between the SIPP2-SDS2 and SIPP1-SDS2 following PAMP treatment leads us to propose a competitive binding mechanism and SIPP2 may displace SIPP1 on SDS2. To confirm, we conducted a competitive binding with Co-IP. The result showed that SIPP2 binding to SDS2 was enhanced as its amount increased, while the binding between SIPP1 and SDS2 was progressively reduced (Fig. 7G). In addition, we performed an in vitro competitive binding assay to investigate whether SIPP2 competitively binds SDS2 against SIPP1. GST-SDS2 was incubated with His-SIPP1, with addition of gradient amounts of MBP-SIPP2, and followed by immunoblot. The results demonstrated that the SDS2-SIPP1 interaction was interrupted by the addition of MBP-SIPP2 (Fig. 7H). These results collectively suggest that SIPP2 competes with SIPP1 for binding to SDS2, and dephosphorylates SDS2 via sites distinct from Thr671, thereby inhibiting immune signaling during the late phase of PTI.

Fig. 7. SIPP2 competes with SIPP1 for SDS2 binding in the late phase of PTI.

(A) The abundance and phosphorylation levels of SDS2-myc in the transfected protoplasts of wild-type (ZH11) and of sipp2 mutants. Transfected rice protoplasts were treated with 10 μg/ml chitin for 15 min. Total protein from rice protoplasts were immunoprecipitated with anti-myc beads and immunoblotted with anti-pT671, anti-pThr/Tyr and anti-myc antibodies. (B and C) The abundance and phosphorylation of SDS2-myc expressed with SIPP1-HA, SIPP2-HA and SIPP1^D270A-HA. SDS2-myc was transiently expressed alone or together with SIPP1-HA, SIPP2-HA or SIPP1^D270A-HA in rice protoplasts treated with 10 μg/ml chitin (B) or 5 μM flg22 (C) for 15 min. Total protein extracts from rice protoplasts were immunoprecipitated with anti-myc beads. GFP was transiently co-expressed as an internal control. (D and E) SDS2-SIPP1 and SDS2-SIPP2 interactions were monitored over time following 10 μg/ml chitin or 5 μM flg22 treatment. SDS2-myc, SIPP1-GFP, and SIPP2-HA were transiently co-expressed in rice protoplasts, followed by chitin (D) or flg22 (E) treatment for the indicated time periods. (F) The abundance and phosphorylation levels of SDS2-myc in the transfected protoplasts of wild-type (ZH11) and of sipp2 mutants with 10 μg/ml chitin treatment for 45 min. (G) SIPP2 competitively binds to SDS2 against SIPP1. SDS2-myc and SIPP1-GFP were co-expressed with SIPP2-HA or PP70-HA (as a negative control) in rice protoplasts with flg22 treatment. (H) In vitro competitive binding assay. His-SIPP1 was incubated with GST-SDS2 in the presence of increasing concentrations MBP-SIPP2 with glutathione agarose beads.

Furthermore, we found that the kinase domain of SDS2, SIPP1 and SIPP2 are highly conserved among the monocot species analyzed (Fig. 8, A to C, and figs. S13 and S14). We speculate that the regulatory module of SDS2-SIPP1-SIPP2 may represent a conserved mechanism in monocots.

Fig. 8. SDS2, SIPP1 and SIPP2 are highly conserved in monocots.

(A) Sequence alignment and phylogenetic analysis of the SDS2. Alignment of the catalytic loop and activation loop, and maximum-likelihood phylogeny of SDS2 in Sorghum bicolor, Miscanthus lutarioriparius, Zea mays, Panicum hallii, Urochloa decumbens, Hordeum vulgare, Triticum aestivum, Triticum turgidum, Panicum virgatum, Oryza sativa, and Zizania latifolia. Conserved phosphorylation site Thr671 was marked with asterisks. (B) Sequence alignment of a functionally conserved motif (highlighting D270, marked with asterisks) and phylogenetic reconstruction of SIPP1 in S. bicolor, Miscanthus lutarioriparius, Z. mays, P. hallii, U. decumbens, H. vulgare, T. aestivum, T. turgidum, P. virgatum, O. sativa, Z. latifolia, and Panicum miliaceum. (C) Sequence alignment of a conserved motif and phylogenetic reconstruction of SIPP2 in S. bicolor, Miscanthus lutarioriparius, Z. mays, P. hallii, U. decumbens, H. vulgare, T. aestivum, T. turgidum, P. virgatum, O. sativa, Z. latifolia, and P. miliaceum.

DISCUSSION

Surface immune receptors, or PRRs, must be tightly regulated to avoid inappropriate activation of defense responses (10). However, the mechanisms maintaining their homeostasis remains largely unknown. SDS2 is a monocot specific surface immune receptor which controls cell death and immunity in rice (13). In this study, we revealed that protein phosphatase SIPP1 controls the activation and stability of SDS2 by modulating the phosphorylation on Thr671 to promote plant immunity. And another protein phosphatase SIPP2 competitively binds to SDS2 against SIPP1, thereby enabling SPL11-dependent SDS2 degradation and leading to the attenuation of SDS2-mediated immune signaling. The working model is summarized in Fig. 9.

Fig. 9. Working model of how SIPP1, SIPP2, and SPL11 coordinately regulate SDS2 activation and homeostasis.

Thr671 on activation loop serves as a phospho-switch which controls SDS2 activation and homeostasis. The phosphorylation of Thr671 is monitored by SIPP1 and SPL11. Thr671-phosphorylated SDS2 is either ubiquitinated by SPL11 for degradation or dephosphorylated by SIPP1 to return to unphosphorylated. Unphosphorylated SDS2 is resistant to degradation and remains stable. Therefore, SIPP1 antagonizes SPL11-mediated ubiquitination by dephosphorylating the Thr671, thereby sustaining SDS2 homeostasis. Upon immune activation, Thr671 undergoes autophosphorylation to transduce immune signals to downstream RLCKs which trigger immune responses. To prevent excessive immune activation, Thr671 phosphorylated SDS2 is ubiquitinated and degraded by the E3 ligase SPL11 which leads to immune attenuation. SIPP2 displaces SIPP1 from SDS2 via competitive binding to favor SPL11-dependent SDS2 degradation. Created in BioRender, S. Wenfeng (2026) https://BioRender.com/skcoytx).

Protein phosphatases play diverse roles in the regulation of PRRs. PRR complexes are hot spots for phosphatase mediated immune regulation. The negative regulation of PRR complexes is much common (10). LOPP dephosphorylates LORE at Ser761 to attenuate LPS triggered responses in xylem immunity (14). KAPP and PP2A negatively regulate the flagellin receptor FLS2 and its coreceptor BAK1respectively by dephosphorylation (15, 16). Rice XA21 is under the negative regulation of XB15 which directly dephosphorylate XA21 (17). The XB15 orthologs in Arabidopsis, PLL4 and PLL5, serve as inhibitors of receptor complexes which dampen signaling by dephosphorylating the receptor kinases in the resting stage (18). In addition, three PP2C phosphatases PP2CH1, EGR1, and EGR2 negatively regulate the versatile receptor FERONIA (FER) in the pollen-stigma interaction in Arabidopsis (34). This regulatory module may also function in plant immunity, since FER participates in immune signaling widely (35–40). Similar to the negative regulation, Arabidopsis CIPP1 controls the recycling of CERK1 by dephosphorylating at Tyr428 which converts the CERK1 from activation to nascent (19). In contrast to the aforementioned phosphatases that negatively regulate immunity, SIPP1 plays a positive role in rice defense by dephosphorylating SDS2 and controlling its activation and homeostasis, working in concert with another protein phosphatase SIPP2. Specifically, SIPP2 competes with SIPP1 for SDS2 binding to enable SPL11-dependent SDS2 degradation at the late stage of PTI (Fig. 6). Alternatively, SIPP2 may target other phosphorylation site(s) of SDS2 for dephosphorylation since SDS2 phosphorylation level is affected in sipp2 mutants (Fig. 6). The synergistic action between SIPP1 and SIPP2 enables the precise control of immune activation and attenuation. This regulatory module is likely widely distributed in monocots, since the components are conserved among monocot crops such as maize, wheat, sorghum, and barley (Fig. 8).

The majority Ser/Thr kinases belong to RD-type which possess unique structure and conserved residues in the catalytic loop and the activation loop (8, 9, 30). Phosphorylation on the activation loop is crucial for kinase activity (6, 29, 41). Activated kinases adopt to a unique conformation in which conserved residues form a rigid interaction network (Fig. 5). Full activation of kinase activity requires the phosphorylation on Thr^APE-10/−11 residues of activation loop which contributes to the formation of the interaction network (5). Therefore, Thr residues at APE-10/−11 position, Thr^APE-10/−11, serve as an activation switch. To facilitate description, we term this switch as APE-10 switch, based on the position of Thr residues in the activation loop which distributed from APE-9 to −11 (centered at −10). Briefly, the APE-10 switch could be defined as the phospho-switch at Thr^APE-10/−11 residues in the activation loop of RD-type Ser/Thr kinases. Phosphorylation of the APE-10 switch turn on kinase activity, as seen in the structurally determined active kinases. T-to-A substitutions (phospho-dead) of Thr^APE-10/−11 residues largely block the activation. In contrast, T-to-D substitutions (phospho-mimic) enhance kinases activity (Fig. 5, E to G, and fig. S7, B to G). The APE-10 switch is conserved among RD-type Ser/Thr kinases across kingdoms (Fig. 5A). SDS2 is a typical RD-type Ser/Thr kinase. Thr671 resides at the position of APE-10/−11 in the activation loop, representing the APE-10 switch of SDS2. Our results demonstrated that Thr671 phosphorylation leads to the activation of SDS2 which triggers immune responses in rice. On the other side, Thr671 phosphorylation recruits SPL11 to ubiquitinate SDS2 and modulate the stability. Dephosphorylation on Thr671 or phospho-dead mutation of the residue (SDS2^T671A) results in impaired kinase activity and compromised ubiquitination of SDS2. The findings reveal that Thr671 serves as a phospho-switch.

The regulation on the APE-10 switch is robust. SIPP1 modulates SDS2 kinase activity and stability by dephosphorylating Thr671. Phosphorylation on Thr671 associates with kinase activation and susceptibility to degradation. Action of SIPP1 directly reverse the effects, shutting down SDS2 activity and keep it stable. In the case of FER negative regulation, three PP2C phosphatases also target the phosphorylation on residues (Ser695 and Thr696) on the activation loop, which is the same position as Thr671 on SDS2. The phosphatases dephosphorylate FER at Ser695 and Thr696 and inhibit its kinase activity thus to reduce ROS production on stigma and to promote pollen hydration (34). Due to the indispensable role in kinase activity, the APE-10 switch is also targeted by pathogen effector. The effector AvrAC from the bacterial pathogen Xanthomonas campestris pv. campestris (Xcc) harbors uridylyl transferase activity which adds uridine 5′-monophosphate modification (uridylylation) to Ser/Thr residues corresponding to the APE-10 switch in BIK1 and RIPK. Uridylylation on the conserved residues prevents the phosphorylation dependent activation of these kinases thereby suppresses immune responses to facilitate bacterial infection (42). The uridylylation functionally equals to non-phosphorylation or dephosphorylation in the prevention of kinase activation. Considering the conservation of the APE-10 switch in RD-type Ser/Thr kinases, the regulation of kinase activity on this switch is probably widely distributed.

Plant genomes encode a big group of protein phosphatases which belong to different families (25, 26, 43–45). The above-mentioned protein phosphatases CIPP1, LOPP, KAPP, XB15, PLL4 and PLL5, PP2CH1, EGR1 and EGR2C are all belong to PP2C family, suggesting important roles of PP2C in immune regulation (14–19, 34). SIPP2 belongs to the PP2A subfamily of the PPP family and functions as a PP2A regulatory subunit. PP2A enzymes form heterotrimeric complexes consisting of a scaffolding A subunit, a regulatory B subunit (such as SIPP2), and a catalytic C subunit. The C and A subunits together constitute the core enzyme, which binds to the B subunit that confers substrate specificity, cellular localization, and enzymatic activity to the ABC trimer (25–26, 44). In contrast, SIPP1 belongs to the C-terminal domain phosphatase-like (CPL) subfamily, which is cation-dependent and exhibits Tyr/Ser phosphatase activity (25). CPL phosphatases have emerged as master regulators of plant immunity as well as development. For example, Arabidopsis CPL1 targets mitogen-activated protein kinase cascades to suppress plant immunity (46). Arabidopsis CPL3 negatively regulates plant immunity by dephosphorylating the RNA Polymerase II C-terminal domain, thereby suppressing immunity-related gene expression and preventing hyperactivation of defense responses (47, 48). In addition to the role in plant defense, Arabidopsis CPL1 is involved in floral transition and flowering (49, 50). OsCPL3 is involved in brassinosteroid signaling in rice (51). Noteworthy, SIPP1 functions in cytosol and proximal to plasmmembrane to modulate SDS2 activation and homeostasis while Arabidopsis CPL1 and CPL3 exert dephosphorylation in the nucleus to suppress MAPK cascades and immunity-related gene expression respectively (46, 47, 50).

The activity and homeostasis of SDS2 are precisely regulated by SIPP1 and SIPP2. However, it remains unclear how SIPP1 and SIPP2 are temporally regulated during immune responses. Whether and how SIPP2 targets other phosphorylation site(s) in SDS2 to modulate rice immunity deserve further investigation. In addition, the precise temporal control mechanisms governing SIPP1 and SIPP2 during immune activation and attenuation await further elucidation.

MATERIALS AND METHODS

Plant materials and growth conditions

The rice cultivars Zhonghua 11 (ZH11), IR64, and Nipponbare (NIP) were used in this study. SIPP1 overexpression lines, sipp1, sipp2 and sds2 mutants used in this study are in the ZH11 background. SIPP1 cDNA was cloned into pRHVcHA-pUbi to generate SIPP1-overexpressing constructs, which were transformed into Agrobacterium tumefaciens EHA105 via freeze-thaw and introduced into rice using Agrobacterium-mediated transformation. The sipp1, sipp2, and sds2 mutants were generated by CRISPR/Cas9. The SIPP1-OE-29/sds2–4 plants were obtained by crossing SIPP1 overexpression line SIPP1-OE-29 with mutant sds2–4. Stable transgenic rice lines in the sds2 CRISPR-Cas9 knockout background complementing SDS2-YFP, SDS2^T667A-YFP, SDS2^T671A-YFP, or SDS2^T677A-YFP were driven by the native SDS2 promoter. The spl11 mutant in IR64 background has been described previously. Rice was grown in a greenhouse under the conditions of 12-hour day/12-hour night at 30°C/28°C, 80% relative humidity (RH)/ 60% RH, respectively. Rice seedlings for protoplast isolation or ROS burst assay were cultured on half-strength Murashige and Skoog (½ MS) medium in a growth chamber (28°C, 200 μmol m⁻² s⁻¹ light).

Blast Fungus (Magnaporthe oryzae) strains and culture conditions

The M. oryzae isolate RB22 was cultured on oat meal plate [3% (w/v) oat and 1.5% (w/v) Agar] under light at room temperature for 10 days for sporulation. The spores were collected by scratching the plates and suspended in inoculation solution (pure water with 0.03% v/v Tween-20). The spore suspension was filtered with Miracloth and adjusted to the concentrations of 1 × 10⁵ for spray inoculation or 5 × 10⁵ for punch inoculation.

In vitro kinase assay

In vitro phosphorylation assays were performed as previously described with a minor modification (13, 52). Autophosphorylation of SDS2 was assayed by incubating purified GST-tagged cytoplasmic domain of SDS2 (GST-SDS2) (2 μg) or its mutants (GST-SDS2^T667A, GST-SDS2^T671A, GST-SDS2^T677A) in kinase buffer (20 mM Tris-HCl, pH 7.5, 10 mM MgCl₂, 1 mM CaCl₂, 10 μM ATP, 1 mM DTT) at 30°C for 1 hour. Reactions were terminated by adding 1 × SDS-PAGE loading buffer. Phosphorylation was analyzed by immunoblotting using anti-pThr (9386S, Cell Signaling Technology) or anti-pSer (ab9332, Abcam) antibodies. Protein loading was confirmed by Coomassie Brilliant Blue (CBB) staining. For transphosphorylation assays, GST-SDS2 (2 μg) was incubated with substrate RLCK176^K108E or MBP-RLCK118^K116E (2–4 μg) in kinase buffer supplemented with 1 mM MnCl₂ at 30°C for 1 hour. Phosphorylation was detected by immunoblotting using anti-pThr/Tyr (9381S, Cell Signaling Technology) antibodies.

Co-IP assay

Protoplasts were isolated from Nipponbare (NIP) seedlings as previously described (53). Protoplasts (2.5 × 10⁶ cells/ml) were transfected with 50 μg of the indicated constructs and cultured for 16 hours to allow protein expression. Before harvest, the protoplasts were treated with 5 μM flg22 (Sangon Biotech) for 20 min. Total proteins were extracted using lysis buffer [150 mM NaCl, 50 mM Tris-HCl (pH 7.5), 5 mM EDTA (pH 8.0), 10% Glycerol, 0.5% Triton X-100, 1 mM DTT, 1 mM PMSF and protease inhibitor cocktail (P1006, Beyotime)]. The protein extracts were incubated with anti-GFP (KTSM1334, AlpaLifeBio), anti-Myc (KTSM1336, AlpaLifeBio) or anti-HA (KTSM1335, AlpaLifeBio) Nanobody Magarose Beads at 4°C for 2 hours with gentle rotation. After incubation, the beads were washed five times with ice-cold 1× PBS. The bound proteins were eluted with 1× SDS-PAGE loading buffer and applied to immunoblotting using specific antibodies.

IP-MS assay

To identify SDS2-interacting proteins, the SDS2-HA fusion protein was expressed in rice protoplasts which were treated with flg22 for 15 min, and the SDS2 complex was isolated by immunoprecipitation as described above. The beads were submitted to Instrumental Analysis Center of Shanghai Jiao Tong University for LC-MS/MS analysis.

BIFC assay in protoplasts

The coding regions of SIPP1 were fused to the N terminus of the mVenusN tag in pRTVcVN, and SDS2 was fused to the N terminus of the mVenusC tag in pRTVcVC. The constructed vectors were co-transfected into rice protoplasts (2.5 × 10⁶ cells/ml) isolated from the wild-type seedlings. Fluorescence images were obtained using a confocal microscope (Nikon AX) after 5 μM flg22 treatment for 20 min.

Recombinant protein expression and purification

The pGEX-4 T-1, pMAL-C2X, and pET30a vectors were used to express GST-, MBP-, and His-tagged proteins, respectively. Protein expression was induced in E. coli BL21(DE3) cells with 0.1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG, ST1416, Beyotime) at 25°C for 6 hours. The His-, GST-, and MBP-tagged proteins were purified using Ni-NTA resin (P2210, Beyotime), Glutathione Sepharose (P2253, Beyotime), and amylose resin (SA077025, Smart-Lifesciences), respectively, by following the manufacturers’ protocols.

Pull-down assay

For GST pull-down assays, purified His-SIPP1 and His-PP70 were individually incubated with glutathione agarose beads in the presence of GST-SDS2 or GST (control) at 4°C for 3 hours. Similarly, for MBP pull-down assays, purified His-SIPP1 was incubated with amylose resin in the presence of MBP-SPL11 or MBP (control) under same conditions. Following incubation, the beads were washed five times with ice-cold 1 × PBS buffer. Bound proteins were eluted using 1 × SDS loading buffer and examined by immunoblotting with anti-His (M20001S, Abmart), anti-GST (M20007S, Abmart), and anti-MBP (31201ES60, Yeasen) antibodies.

Protein stability assay in protoplasts

The protein stability assay was performed as described previously (54). Briefly, the pRHVcHA-SDS2-HA, pRHVcHA-SDS2^T667A-HA, pRHVcHA-SDS2^T671A-HA, and pRHVcHA-SDS2^T677A-HA plasmids were transfected into rice protoplasts (2.5 × 10⁶ cells/ml) isolated from NIP. pRHVcGFP was co-expressed as an internal reference. Following 16-hour culture, the transfected protoplasts were treated with 50 μM CHX (HY-12320, MCE) to block synthesis of new proteins. SDS2 abundance was monitored at 0 and 3 hours after treatment. Protein levels were determined by immunoblotting with anti-HA (30704ES60, Yeasen) and anti-GFP (11814460001, Roche) antibodies.

SDS2 phosphorylation antibodies production

Anti-phospho-Thr671 antibody was produced by the service agent Abmart. The phospho-peptide CETQAN(pT)NRVV was synthesized as immunogen and applied to immunization of rabbits. The antiserum was purified with affinity chromatography by using the phospho-peptide. The eluate was purified with column coupled with control peptides CETQANTNRVV to remove nonspecific antibodies. The resulting phospho-Thr671 antibody was validated with immunoblotting.

In vivo phosphorylation assay

In vivo phosphorylation assays were performed as previously described with a slight modification (55). Protoplasts (2.5 × 10⁶ cells/ml) from wild-type or mutant lines were transfected with pRHVcMyc-SDS2-myc or pRHVcHA-SDS2-HA, using pRHVcGFP as an internal reference. After transfection, the protoplasts were cultured at room temperature in the dark for 16 hours. Before harvest, the cells were treated with 5 μM flg22 (Sangon Biotech) or 10 μg/ml chitin (C9752, Sigma-Aldrich) for 15 min or indicated time to induce SDS2 phosphorylation. Total proteins were extracted with lysis buffer [150 mM NaCl, 50 mM Tris-HCl (pH 7.5), 5 mM EDTA (pH 8.0), 10% Glycerol, 0.5% Triton X-100, 1 mM DTT, 1 mM PMSF, plus protease inhibitor cocktail and phosphatase inhibitor cocktail (P1049, Beyotime)]. Lysates were incubated with magnetic affinity beads (anti-Myc or anti-HA, as appropriate) at 4°C for 2 hours with gentle rotation. Beads were washed five times with ice-cold 1 × PBS. Beads bound proteins were eluted with 1 × SDS-PAGE loading buffer and subjected to immunoblotting with specific antibodies. The phosphorylated SDS2 was detected by anti-pThr antibodies or anti-phospho-T671 antibody.

In vitro dephosphorylation assay

Purified His-SIPP1, His-SIPP2, and His-PP70 (2–5 μg) were individually incubated with GST-SDS2 (1 μg) in a reaction buffer (20 mM Tris-HCl, pH 7.5, 10 mM MgCl₂, 1 mM CaCl₂, 1 mM MnCl₂,10 μM ATP, 1 mM DTT) at 30°C for 1 hour. Reactions were terminated by adding 1 × SDS-PAGE loading buffer. Phosphorylation was detected by immunoblotting using anti-pThr antibodies. Total protein loadings were verified by CBB staining.

M. oryzae inoculation

M. oryzae inoculation assays were performed as previously described with a slight modification (13). For punch inoculation, M. oryzae (isolate RB22) was cultured on oatmeal agar at room temperature for 10 days to induce sporulation. Fresh spores were harvested using distilled water containing 0.5% (v/v) Tween-20 and adjusted to a concentration of 5 × 10⁵ spores/ml. Six-week-old seedlings were inoculated by punching the middle section of the second leaf (counting from the top). Phenotypic scoring was performed two weeks post-inoculation. Relative lesion area was measured by ImageJ. Relative fungal biomass was quantified using qPCR to measure the M. oryzae transposable element MoPot2, as previously described. For spray inoculation, a conidial suspension (1 × 10⁶ spores/ml) containing 0.025% (v/v) Tween-20 was prepared for the inoculation. Three-week-old seedlings were spray-inoculated with the conidial suspension. Disease symptoms on inoculated leaves were photographed at 7-day post-inoculation.

ROS assay

The ROS assay was performed as previously described (56). Briefly, rice sheaths were cut to 2 mm segments and pretreated with pure water overnight. ROS production was induced with 100 μM flg22 in 50 mM Tris-HCl buffer (pH 7.5) supplemented with 10 μM L-012 (CDFT-120-04891, wako) and 10 μg/ml horseradish peroxidase (P815746, Macklin). Luminescence was monitored immediately after treatment and recorded at 1-min intervals for 45 min using a Varioskan Flash multireader (BioTek).

In vivo ubiquitination assay

In vivo ubiquitination assay was performed as described previously with a minor modification (57). Protoplasts transfected with FLAG-Ub, SDS2-HA (SDS2^T667A-HA, SDS2^T671A-HA, SDS2^T677A-HA) or SDS2-myc, SIPP1-HA (SIPP1^D270A-HA) or SIPP2-HA, and GFP were incubated overnight in the presence of 50 μM MG132 (S42096, Yuanye Bio-Technology). Total proteins were extracted with lysis buffer [150 mM NaCl, 50 mM Tris-HCl (pH 7.5), 5 mM EDTA (pH 8.0), 10% Glycerol, 0.5% Triton X-100, 1 mM DTT, 1 mM PMSF and protease inhibitor cocktail] and enriched with anti-HA-beads or anti-myc-beads in the presence of MG132 (50 μM). After incubated for 3 hours, the beads were washed five times with ice-cold 1× PBS buffer. Bound proteins were eluted using 1× SDS loading buffer and analyzed by immunoblotting using anti-HA, anti-myc (30601ES60, Yeasen), and anti-FLAG (SLAB0100, Smart-Lifesciences) antibodies.

Multiple alignment of kinases

RLCK and RLK kinases of 10 plant species, including rice, Arabidopsis, alfalfa (Medicago truncatula), soybean (Glycine max), wheat (Triticum aestivum), barley (Hordeum vulgare), liverwort (Marchantia polymorpha), cotton (Gossypium hirsutum), oilseeds (Brassica rapa), and sorghum (Sorghum bicolor), are downloaded from RLKdb (58). Animal kinases are from kinomes of human, mouse, fruit fly, and nematode (http://kinase.com/). The kinase domains are trimmed and applied to multiple alignment by MUSCLE with MEGA12. The sequence logo is generated from the aligned sequences with TBtools (59).

Structural modeling and analysis

Protein sequence of SDS2 cytoplasmic domain was submitted to Alphafold3 server with default settings. Crystallography structurals were download from the Protein Data Bank (PDB) under accession codes 1atp, 1ctp, 1jst, 2erk, and 1o6l, respectively. All structures visualized in this article were prepared with PyMOL.

Statistical analysis

Statistical significance was assessed by two-tailed Student t tests for pairwise comparisons (*P < 0.05, **P < 0.01, ***P < 0.001) in Microsoft Excel or GraphPad Prism 8, and by one-way ANOVA with least significant difference (LSD) post-hoc tests (α = 0.05) for multi-group comparisons in SPSS.

Supplementary Materials

The PDF file includes:

Figs. S1 to S14

Legend for table S1

Table S2

Other Supplementary Material for this manuscript includes the following:

Table S1