The zinc finger protein DHHC09 S-acylates the kinase STRK1 to regulate H₂O₂ homeostasis and promote salt tolerance in rice

Abstract

Soil salinity results in oxidative stress and heavy losses to crop production. The S-acylated protein SALT TOLERANCE RECEPTOR-LIKE CYTOPLASMIC KINASE 1 (STRK1) phosphorylates and activates CATALASE C (CatC) to improve rice (Oryza sativa L.) salt tolerance, but the molecular mechanism underlying its S-acylation involved in salt signal transduction awaits elucidation. Here, we show that the DHHC-type zinc finger protein DHHC09 S-acylates STRK1 at Cys5, Cys10, and Cys14 and promotes salt and oxidative stress tolerance by enhancing rice H₂O₂-scavenging capacity. This modification determines STRK1 targeting to the plasma membrane or lipid nanodomains and is required for its function. DHHC09 promotes salt signaling from STRK1 to CatC via transphosphorylation, and its deficiency impairs salt signal transduction. Our findings demonstrate that DHHC09 S-acylates and anchors STRK1 to the plasma membrane to promote salt signaling from STRK1 to CatC, thereby regulating H₂O₂ homeostasis and improving salt stress tolerance in rice. Moreover, overexpression of DHHC09 in rice mitigates grain yield loss under salt stress. Together, these results shed light on the mechanism underlying the role of S-acylation in RLK/RLCK-mediated salt signal transduction and provide a strategy for breeding highly salt-tolerant rice.

S-acylation of the kinase STRK1 promotes salt signaling from STRK1 to catalase C, in turn regulating hydrogen peroxide homeostasis and improving salt tolerance in rice.

IN A NUTSHELL.

Background: Soil salinity results in oxidative stress and heavy losses to crop production. S-acylation of proteins occurs extensively in plants and plays important roles in many essential cellular functions. Receptor-like kinases (RLKs) often act as receptors to perceive extracellular signals or stimuli, resulting in dimerization followed by autophosphorylation and activation of receptor-like cytoplasmic kinases (RLCKs) by transphosphorylation. Activated RLCKs then phosphorylate downstream target proteins to initiate the stress response. The S-acylated RLCK protein SALT TOLERANCE RECEPTOR-LIKE CYTOPLASMIC KINASE 1 (STRK1) phosphorylates and activates CATALASE C (CatC) to improve rice (Oryza sativa L.) salt tolerance, but the molecular mechanism underlying its S-acylation in salt signal transduction awaits elucidation.

Question: What is the molecular mechanism underlying S-acylation in RLK/RLCK-mediated salt signal transduction?

Findings: We identified the DHHC-type zinc finger protein DHHC09, which S-acylates STRK1 at several cysteine residues and positively regulates salt tolerance in rice. DHHC09-mediated S-acylation mainly determines STRK1 targeting to the plasma membrane and promotes salt signaling from STRK1 to CatC via transphosphorylation, thereby regulating H₂O₂ homeostasis and improving rice salt tolerance. DHHC09 deficiency impairs this signaling cascade and causes hypersensitivity to salt stress. Moreover, DHHC09 overexpression in rice mitigates grain yield loss under salt stress.

Next steps: The plasma membrane is highly compartmentalized into lipid nanodomains, which play a pivotal role in cellular signal transduction. The unknown upstream RLK of STRK1 and the micro-environment of lipid nanodomains in which S-acylated STRK1 resides require further investigation.

Introduction

Soil salinity results in heavy losses to crop production and threatens global food security. Salt stress induces osmotic, ionic, and oxidative stresses in plants (Yang and Guo 2018). To cope with salt stress, plants have evolved a set of mechanisms to perceive and respond to internal and environmental signals to reestablish ionic, osmotic, and redox homeostasis. Hydrogen peroxide (H₂O₂) is a vital signaling molecule in many biological processes, including growth and development, while its excessive accumulation results in cellular damage. Recently, a G_γ protein encoded by Alkaline tolerance 1 (AT1) gene was reported to inhibit the phosphorylation of H₂O₂ exporter plasma membrane (PM) intrinsic protein PIP2; 1 to regulate H₂O₂ homeostasis and improve alkaline tolerance in crops (Zhang et al. 2023). Catalase (CAT) is a vital antioxidant enzyme that degrades the excessive H₂O₂ from salt stress, thus protecting plant cells against salt stress by maintaining H₂O₂ hemostasis. Overexpression and activation of CatC, a member of the CAT family, improve salt tolerance (Zhou et al. 2018; Deng et al. 2021), whereas its deficiency results in the accumulation of H₂O₂ and leaf cell death in rice (Oryza sativa L.) (Lin et al. 2012). Dephosphorylation of CatC at Ser18 improves salt and oxidative stress tolerance by promoting its tetramerization in rice (Wang et al. 2023). The precise regulation of CAT activity is essential for plant growth, development, and stress response (Liu et al. 2023a). For example, a zinc finger protein LESION SIMULATING DISEASE1 (LSD1) interacts with CATs and enhances their activity to regulate light-dependent cell death in Arabidopsis (Arabidopsis thaliana) (Li et al. 2013). PHOSPHATASE OF CATALASE 1 (PC1) was recently reported to specifically dephosphorylate CatC at Ser9, inhibit its tetramerization and enzymatic activity, and act as a molecular switch to balance the growth and salt tolerance in rice (Liu et al. 2023b; Molla 2023).

S-acylation, also known as palmitoylation, is a reversible post-translational modification involving the transfer of a fatty acid group to the sulfhydryl group of cysteine residues of target proteins (Hemsley 2020). The hydrophobic nature of the acyl group can dramatically alter the protein properties. S-acylation regulates membrane localization, protein trafficking, activity, stability, conformation, and protein interactions (Hemsley and Grierson 2008; Baekkeskov and Kanaani 2009; Greaves and Chamberlain 2011). S-acylation modification extensively occurs in plants and plays important roles in many essential cellular functions. A proteomic study has identified approximately 450 putatively S-acylated proteins in poplar (Populus trichocarpa) (Srivastava et al. 2016). Recently, Kumar et al. (2022) identified 1,094 S-acylated proteins representing around 6% of the proteome in Arabidopsis.

The acyl group is added by a family of DHHC-type zinc finger proteins, also known as DHHCs or protein acyl transferases (PATs), characterized by a DHHC motif within their catalytic domain. DHHC proteins have S-acyltransferase activity and their DHHC motifs are essential for S-acyltransferase activity and for their auto-S-acylation (Qi et al. 2013; Tian et al. 2022). There are 7 DHHC proteins in yeast (Saccharomyces cerevisiae), 24 in Arabidopsis, and 30 in rice (Hemsley et al. 2005; Greaves and Chamberlain 2011; Batistic 2012; Li et al. 2016). Among the many S-acylated proteins characterized to date, only a limited number of DHHC-S-acylated protein pairs have been identified in plants (Zhang et al. 2019; Jiang et al. 2021; Zeng et al. 2021; Gao et al. 2022). Recently, we constructed an OsDHHC cDNA library and screened 5 DHHC-S-acylated protein pairs using bimolecular fluorescence complementation (BiFC) assays in rice. Among these pairs, the interaction of OsDHHC30 with the calcineurin-B-like protein OsCBL2 and OsCBL3 confer the tolerance to salt and oxidative stresses (Tian et al. 2022). An ABHD17-like hydrolase screening system has also been established to identify pairs of protein substrates and de-S-acylation enzymes in Arabidopsis (Liu et al. 2021). These protein S-acylation and de-S-acylation enzyme screening systems contribute to the elucidation of the molecular mechanism underlying this reversible modification in plant biological processes.

Receptor-like kinases (RLKs) are one of the major cell-surface receptors, and receptor-like cytoplasmic kinases (RLCKs) are an RLK subfamily in plants. RLKs often act as receptors to perceive extracellular signals or stimuli, resulting in dimerization followed by autophosphorylation and activation of RLCKs by transphosphorylation. Then the activated RLCKs phosphorylate the downstream target proteins to initiate the stress response in plants (Osakabe et al. 2013; Macho et al. 2015). Thus, RLCKs play central roles in this signaling pathway of the phosphorylation cascade. Besides the phosphorylation by RLKs for signal transduction, some RLCKs are S-acylated for membrane anchoring. The RLCK protein CONSTITUTIVE DIFFERENTIAL GROWTH 1 (CDG1) is anchored to the PM by S-acylation and mediates brassinosteroid (BR) signal transduction from the BR receptor BRASSINOSTEROID INSENSITIVE1 (BRI1) to BRI1 SUPPRESSOR1 (BSU1) phosphatase in Arabidopsis (Kim et al. 2011).

We previously found that the SALT TOLERANCE RECEPTOR-LIKE CYTOPLASMIC KINASE 1 (STRK1) is anchored to the PM by S-acylation, and this modification is necessary for its function in salt signal transduction in rice. Upon salt stress, STRK1 phosphorylates CatC at Tyr210 and activates CatC at the PM to regulate H₂O₂ homeostasis and improve salt tolerance (Zhou et al. 2018). However, which DHHC-S-acylates STRK1 and the molecular mechanism underlying S-acylation involved in salt signal transduction are still unclear. In this study, we characterized DHHC09 as an upstream component of STRK1-mediated salt signal transduction pathway to S-acylate STRK1 and regulate the H₂O₂ homeostasis and salt stress response in rice. Moreover, overexpression of DHHC09 in rice not only improved growth at the seedling stage but also markedly limited the grain yield loss under salt stress conditions.

Results

STRK1-mediated salt signal transduction depends on S-acylation

STRK1 is a tyrosine kinase that phosphorylates CatC at Tyr210 and positively regulates salt tolerance in rice, and its 3 cysteine residues (Cys5, 10, 14) are speculated to be S-acylated for targeting the PM (Zhou et al. 2018). To further determine the effect of S-acylation of STRK1 on the salt stress response, the fusion proteins STRK1-yellow fluorescence protein (YFP) and STRK1^C5,10,14A-YFP (substitution of Cys5, 10, 14 with alanine) were overexpressed in rice plants and their salt stress response was investigated. Compared with wild-type (WT) plants, the salt tolerance was significantly enhanced in STRK1-YFP overexpressing lines, whereas such effects were lost in STRK1^C5,10,14A-YFP overexpressing lines (Fig. 1, A to C), indicating that S-acylation is required for STRK1 to function in salt signal transduction. Moreover, STRK1-YFP overexpressing lines showed lower Na⁺ accumulation and Na⁺/K⁺ ratio under salt stress, while STRK1^C5,10,14A-YFP overexpressing lines displayed the similar results to WT plants (Fig. 1, D to F).

Figure 1.

STRK1 is S-acylated at Cys5, 10, 14 and the mutation in S-acylation sites affects its protein location. A) Phenotypic comparison of seedlings grown under salt stress at the seedling stage. 15-d-old seedlings of STRK1 and STRK1^C5,10,14A overexpressing plants (STRK1-YFP-2/9 and STRK1^C5,10,14A-YFP-2/6) as well as their corresponding WT were treated with 140 mM NaCl for 10 d and recovered for 6 d. Bar = 2.5 cm. B and C) Survival rates of transgenic and WT plants in (A) after 6 d of recovery. Forty plants in each line were used for survival rate analysis. Data in (B) and (C) are presented as mean ± SD (n = 4, **P ≤ 0.01, one-way ANOVA). D) to F) Ion contents in shoots of 15-d-old plants after 5 d of salt stress (140 mM NaCl). The normal and salt-stressed samples of shoots were used for analysis of ion contents. Normal represents day 0. The values of Na⁺ (D) and K⁺(E) together with the ratio of Na⁺/K⁺ (F) are presented. DW, dry weight. Data are presented as mean ± SD (n = 3, *P ≤ 0.05, **P ≤ 0.01, Two-way ANOVA). (G) Analysis of STRK1 S-acylation in STRK1-YFP and STRK1^C5,10,14A-YFP transgenic plants by an ABE assay. S-acylated STRK1 was detected with an anti-GFP antibody. The lanes in “S-acylation” show the amount of STRK1-YFP and STRK1^C5,10,14A-YFP bound to the neutravidin-agarose beads with (+) or without (−) treatment with NH₂OH. “Loading” controls show equal amounts of proteins loaded. (H) Mass-shift detection of STRK1-YFP in the transgenic plants by an APE assay. Samples were analyzed by immunoblot using an anti-GFP antibody. The number of PEGylation events is indicated by asterisks (*), which corresponded to mono-, di-, and tri-PEGylated proteins, respectively. Apo refers to non-PEGylated protein. mPEG-Mal, methoxy-PEG-maleimide. The experiments were repeated 3 independent times. I) Subcellular fractionation assay in rice protoplasts. The STRK1-YFP, STRK1^C5,10,14A-YFP, and STRK1^K95E-YFP were co-expressed in rice protoplasts with an S-acylated protein GSD1 that requires S-acylation to locate at the PM. GSD1-FLAG was used as a PM maker. S, soluble protein fraction; U, microsomal fraction; EM, microsomal fraction isolated from the endomembrane; PM, microsomal fraction isolated from the PM. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH)/Actin and ATPase 3 (AHA3) were used as the cytoplasm and PM makers, respectively. J) Quantification of the percentage of STRK1-YFP, STRK1^C5,10,14A-YFP, and STRK1^K95E-YFP in S and U in (I) using Image LabTM software version 4.1. These statistics came from 3 independent experiments. Data are mean values ± SD (n = 3). K) Analysis of STRK1-YFP, STRK1^C5,10,14A-YFP, and STRK1^K95E-YFP S-acylation in S and U. The S-acylation levels of STRK1-YFP, STRK1^C5,10,14A-YFP, and STRK1^K95E-YFP in U and S were detected with an anti-GFP antibody. The lanes in “S-acylation” show the amount of STRK1-YFP, STRK1^C5,10,14A-YFP, and STRK1^K95E-YFP bound to the neutravidin-agarose beads with (+) or without (−) NH₂OH treatment. “Loading” controls show equal amounts of proteins loaded.

Then, the S-acylation was determined by an acyl-biotinyl exchange (ABE) assay with or without the hydroxylamine (NH₂OH) thioester-cleavage treatment as described previously (Tian et al. 2022). The signals of STRK1-YFP were detected in the NH₂OH-treated and loading samples, but those of STRK1^C5,10,14A-YFP were only detected in the loading samples (Fig. 1G), indicating that STRK1-YFP but not STRK1^C5,10,14A-YFP was S-acylated in planta. An acyl-PEG exchange (APE) assay further determined the S-acylation site. STRK1-YFP was efficiently labeled by 10 kDa methoxy-PEG-maleimide (mPEG-Mal), which resulted in 3 slow-migrating STRK1-PEGylated polypeptides (Fig. 1H). The same results were also detected in another fusion protein STRK1-FLAG labeled by 5 kDa mPEG-Mal (Supplementary Fig. S1A). These results indicate that the 3 cysteine residues (Cys5, 10, 14) are S-acylation sites of STRK1.

We previously found that S-acylation mediates STRK1 association with the PM (Zhou et al. 2018). Subcellular fractionation assays showed that STRK1-YFP and STRK1^K95E (a kinase-dead form of STRK1 with the mutation of Lys95 to glutamic acid) were only detected in the microsomal fraction and the PM, whereas about 67% and 33% of STRK1^C5,10,14A-YFP was detected in the cytoplasm and the PM or microsomal fraction, respectively (Figs. 1, I and J, Supplementary Fig. S1B). To exclude the influence of background, an S-acylated protein GRAIN SETTING DEFECT 1 (GSD1) and its corresponding DHHC protein DHHC14, which were reported to be localized at the PM (Gui et al. 2015; Tian et al. 2022), were co-expressed in rice protoplasts with STRK1-YFP, STRK1^C5,10,14A-YFP, and STRK1^K95E-YFP as controls for analysis. Just like STRK1-YFP and STRK1^K95E-YFP, both GSD1-FLAG and DHHC14-FLAG were only detected in the samples derived from the microsomal fraction and the PM (Figs. 1, I and J, Supplementary Fig. S1B). Both STRK1-YFP and STRK1^K95E-YFP were S-acylated, but STRK1^C5,10,14A-YFP was not S-acylated, even in the part that remained in the microsomal fraction or PM (Fig. 1K). These results were further validated in stable genetic lines (STRK1-YFP and STRK1^C5,10,14A-YFP) (Supplementary Figs. S1C–S1E). STRK1 was previously considered to form a complex with an unknown RLK with a transmembrane domain (Zhou et al. 2018), thus we speculate that the PM-anchored STRK1^C5,10,14A-YFP might result from its complex with an unknown RLK. Together, these results reveal that Cys5, 10, 14 are indeed the S-acylation sites of STRK1, and their S-acylation is required for STRK1-mediated salt signal transduction in rice.

DHHC09 that interacts with STRK1 was screened from an OsDHHC cDNA library

To screen potential DHHCs that S-acylate STRK1, the full-length cDNA of STRK1 was subcloned into a cCFP BiFC expression vector pE3449 and then BiFC assays were performed on an OsDHHC cDNA library as described previously (Tian et al. 2022). The results showed that STRK1 specifically interacted with DHHC09 (also known as OsDHHC09) at the PM, but not with other DHHCs (Fig. 2A; Supplementary Fig. S2A). DHHC09 (LOC_Os03g24900) has 2 splice variants, LOC_Os03g24900.1 and LOC_Os03g24900.2, which encode DHHC09-I and DHHC09-II, respectively (Supplementary Figs. S3, A and B).

Figure 2.

STRK1 physically interacts with DHHC09. A) Screening the DHHC that interacts with STRK1 using an OsDHHC cDNA library and BiFC assay (only shown the DHHC09–STRK1 interaction). The fusion proteins STRK1-cCFP and DHHC09-nVenus were co-expressed in rice protoplasts. The pairings of STRK1-cCFP and nVenus, cCFP and DHHC09-nVenus were co-expressed in rice protoplasts and used as negative controls. Before imaging, the protoplasts were treated with FM4-64 (2 μM) for 5 min to show the PM. BF, bright field; GFP, green fluorescence protein. Fluorescence in the protoplasts was imaged under a confocal laser scanning microscope. Bar = 10 μm. B) Pull-down assay for interaction between STRK1 and DHHC09. Purified GST-STRK1 from E. coil was used for copurifying DHHC09-FLAG from N. benthamiana leaves. An S-acylated protein GSD1 and its corresponding DHHC protein DHHC14 (also known as OsDHHC14), which are localized at the PM and used as the negative controls, were also incubated with DHHC09-FLAG and GST-STRK1, respectively. Immunoblots were developed with anti-GST antibody to detect STRK1 and GSD1, and with anti-FLAG antibody to detect DHHC09 and DHHC14. C) Co-IP assay for interaction between STRK1 and DHHC09 in N. benthamiana leaves. Protein extracts (Input) were immunoprecipitated with an anti-GFP antibody (IP). The PM-anchored proteins GSD1 and DHHC14 were used as negative controls. Immunoblots were developed with anti-FLAG antibody to detect DHHC09 and DHHC14, and with anti-GFP antibody to detect STRK1 and GSD1. D) Cys190 is a key site for interaction between STRK1 and DHHC09. DHHC09Δ1 contains 2 predicted S-acylation sites, and DHHC09Δ2 contains DHHC motif, DHHC09Δ3 contains the NXTTXE domain. DHHC09^C190A was the mutation of cysteine to alanine in the DHHC motif of DHHC09. BF, bright field; GFP, green fluorescence protein. BiFC assays were performed as described above. Bar = 10 μm.

Then the STRK1–DHHCs interaction was analyzed by pull-down assay. The results showed that only DHHC09-I-FLAG and DHHC09-II-FLAG, not other DHHCs, isolated from rice protoplasts were specifically pulled down by GST-STRK1 purified from Escherichia coli (Supplementary Fig. S2B). DHHC09-I and DHHC09-II have 5 and 4 transmembrane domains, respectively, and their DHHC motifs are located at the cytosolic side of PM (Supplementary Fig. S3C). Like DHHC09-I, DHHC09-II also interacted with STRK1 at the PM (Supplementary Figs. S2A and S3D). Both DHHC09-I and DHHC09-II transcripts were found in all tissues examined with the lowest levels in leaf sheaths and the highest in leaves (Supplementary Fig. S3E). Compared to DHHC09-II, the expression of DHHC09-I is more responsive to salt, alkali, and H₂O₂ stresses but less responsive to abscisic acid (ABA) and high temperature (Supplementary Fig. S3F). Because their interaction partner STRK1 promotes the salt tolerance in rice (Zhou et al. 2018), we focused our analysis on the DHHC09-I–STRK1 interaction, and DHHC09 refers to DHHC09-I unless otherwise noted hereafter.

The STRK1–DHHC09 interaction was further verified by pull-down and coimmunoprecipitation (co-IP) assays, and the PM-anchored proteins GSD1 and DHHC14 (Gui et al. 2015; Tian et al. 2022) were used as negative controls. Pull-down assays showed that DHHC09-FLAG isolated from Nicotiana benthamiana was successfully pulled down by GST-STRK1 but not by GST-GSD1, and GST isolated from E. coli, while DHHC14-FLAG could not be pulled down by GST-STRK1 (Fig. 2B). Co-IP assays in N. benthamiana revealed that DHHC09-FLAG could be coimmunoprecipitated with STRK1-YFP but not with GSD1-YFP, whereas DHHC14-FLAG could not be coimmunoprecipitated with STRK1-YFP (Fig. 2C). The interaction between truncated DHHC09 with STRK1 was further analyzed by BiFC assays. The results showed that only DHHC09Δ2 (110–229 AA) containing the DHHC motif could interact with STRK1 at the PM (Fig. 2D). However, the catalytically inactive form DHHC09^C190A (substitution of Cys190 in the DHHC motif with alanine) failed to interact with STRK1 (Fig. 2D), indicating that the Cys190 in the DHHC motif of DHHC09 is essential for their interaction. These results demonstrate that DHHC09 specifically interacts with STRK1 in vitro and in vivo. Additionally, subcellular localization assays showed that DHHC09-YFP was specifically distributed at the PM (Supplementary Fig. S4A). Expression pattern analysis found that DHHC09 was induced by NaCl and H₂O₂ (Supplementary Fig. S4, B and C), suggesting that it is involved in response to salt and oxidative stresses.

DHHC09 positively regulates salt and oxidative stress tolerance in rice

To determine the effects of DHHC09 on the response to salt stress, DHHC09 overexpressing and knockout lines were generated (Supplementary Figs. S3G and S5A to S5D). DHHC09-FLAG overexpressing and dhhc09 seedlings were treated with salt (140 mM NaCl) for 10 and 8 d, respectively, and recovered for 6 d. Compared with the corresponding WT plants, DHHC09-FLAG overexpressing lines exhibited more tolerance to salt stress with higher survival rates, whereas dhhc09 lines displayed hypersensitivity to salt stress with lower survival rates (Fig. 3, A to C), indicating that DHHC09 positively regulates salt tolerance in rice. Moreover, compared to WT, DHHC09-FLAG overexpressing lines showed lower Na⁺ accumulation, Na⁺/K⁺ ratio, and relative ion leakage as well as higher chlorophyll content under salt stress, but dhhc09 lines displayed the opposite results (Figs. 3, D to H). However, no noticeable difference in these physiological indexes was observed between these transgenic and WT plants under normal conditions. Similarly, DHHC09-II-FLAG overexpressing lines also exhibited higher survival rate and chlorophyll content as well as lower relative ion leakage than WT plants under salt stress (Supplementary Figs. S3H to S3K). These observations suggest that DHHC09, including DHHC09-II, confer tolerance to salt stress in rice at the seedling stage.

Figure 3.

DHHC09 confers the tolerance to salt and oxidative stresses in rice. A) Phenotypic comparison of seedlings grown under salt stress at the seedling stage. DHHC09 overexpressing plants (DHHC09-FLAG-3/5) and knockout mutants (dhhc09-10/18) as well as their corresponding WT were treated with 140 mM NaCl for 10 and 8 d, respectively, and recovered for 6 d. Bar = 2.5 cm. B and C) Survival rates of transgenic and WT plants in (A) after 6 d of recovery. Forty plants in each line were used for survival rate analysis. Data in (B) and (C) are presented as mean ± SD (n = 4, **P ≤ 0.01, One-way ANOVA). D) to F) Ion contents in shoots of 15-d-old plants after 5 d of salt stress. WT, DHHC09-FLAG-3/5, and dhhc09-10/18 plants were exposed to salt stress (140 mM NaCl) for 5 d, and the shoots were harvested for analysis of ion contents. Normal represents day 0. The values of Na⁺(D) and K⁺(E) together with the ratio of Na⁺/K⁺(F) are presented. DW, dry weight. G) Relative ion leakage in leaves of 15-d-old plants after 140 mM NaCl treatment for 3 d. H) Chlorophyll content in leaves of 15-d-old plants treated with 140 mM NaCl for 7 d. FW, fresh weight. I) Phenotypic comparison of rice plants subjected to MV stress. The germinated seeds were transplanted into either 1/2 MS medium or 1/2 MS medium supplemented with 4 μM MV for 6 d. MV, methylviologen. Bar = 1 cm. J and K) Seedling height (J) and chlorophyll content in leaves (K) of DHHC09 transgenic and WT plants under normal and MV stress conditions. Data in (D) to (H) and (J) to (K) are presented as mean ± SD (*P ≤ 0.05, **P ≤ 0.01, Two-way ANOVA). L) Leaf phenotype of DHHC09 transgenic and WT plants at the 3-leaf stage under normal conditions or after 100 mM H₂O₂ stress for 2 d. Bar = 1 mm. M) DAB staining for H₂O₂ in leaves from unstressed and H₂O₂-stressed DHHC09 transgenic and WT plants for 1 d. Bar = 1 mm.

To determine the effects of DHHC09 on the response to oxidative stress, these transgenic and WT seedlings were treated with methylviologen (MV) and H₂O₂. After 6 d MV (4 μM) treatment, DHHC09-FLAG overexpressing seedlings exhibited more green and higher seedling height and chlorophyll content than WT, but dhhc09 seedlings displayed an etiolated and dwarf phenotype with lower chlorophyll content (Fig. 3, I to K). However, no noticeable difference in these phenotypes was observed among the transgenic and WT seedlings under normal conditions. Additionally, severe necrosis was observed in the leaves of dhhc09 mutants but not in those of DHHC09-FLAG overexpressing lines after 2 d H₂O₂ (100 mM) treatment (Fig. 3L). H₂O₂ was accumulated less in the leaves of DHHC09-FLAG overexpressing lines but more in those of dhhc09 mutants by 3,3′-diaminobenzidine (DAB) staining, whereas no noticeable staining was observed in all transgenic and WT plants under normal conditions (Fig. 3M). These results indicate that overexpression of DHHC09 in rice enhanced the H₂O₂-scavenging capacity, thereby improving the oxidative stress tolerance. Our findings demonstrate that DHHC09 positively regulates salt and oxidative stress tolerance in rice.

DHHC09 S-acylates STRK1 in vivo

The S-acylation activity of DHHC09 was first determined by knocking out Akr1, encoding an S-acyltransferase, in yeast as described previously (Tian et al. 2022). DHHC09 and DHHC09^C190A were fused with a GKPIPNPLLGLDST peptide (V5) and expressed in the akr1 yeast mutant, and their response to high temperature was investigated. The results showed that DHHC09 but not DHHC09^C190A successfully complemented the thermosensitive phenotype of akr1 mutant (Fig. 4A), indicating that DHHC09 is an S-acyltransferase and the cysteine residue of DHHC motif is required for its activity. ABE assays revealed that the signals of DHHC09-V5 were detected in the NH₂OH-treated and loading samples derived from DHHC09-V5/akr1 strains, but those of DHHC09^C190A-V5 were only detected in the loading samples (Fig. 4B). The same results were also detected in the DHHC09-YFP and DHHC09^C190A-YFP expressed in transgenic rice plants (Fig. 4C) and in N. benthamiana leaf cells (Supplementary Fig. S6A). These results indicate that DHHC09 but not DHHC09^C190A has auto-S-acylation activity in yeast and in planta, and the mutation from cysteine to alanine in DHHC motif abolishes auto-S-acylation.

Figure 4.

DHHC09 mediates the S-acylation of STRK1. A) Functional complementation assays of DHHC09 and DHHC09^C190A in akr1 yeast mutant. EV/WT (W303-1A) strains transformed with the pYES2 EV exhibited mononuclear cells. EV/akr1 strains transformed with the pYES2 EV exhibited elongated multiple nuclei cells. DHHC09/akr1 strains transformed with pYES2-DHHC09 but not DHHC09^C190A/akr1 strains transformed with pYES2-DHHC09^C190A exhibited the similar phenotype to EV/WT strains at 30 °C (upper panel) and 37 °C (lower panel). Bar = 10 μm. B) Auto-S-acylation analysis of DHHC09 and DHHC09^C190A in akr1 yeast mutant by ABE assay. The lanes in “S-acylation” exhibited the amount of S-acylated DHHC09-V5 bound to neutravidin-agarose beads with (+) or without (−) NH₂OH treatment. “Loading” controls exhibited same amounts of protein. The S-acylation signals were detected by an anti-V5 antibody. C) Auto-S-acylation analysis of DHHC09 and DHHC09^C190A in transgenic rice plants by ABE assay. The S-acylation signals were detected using an anti-GFP antibody. D) to F) DHHC09 but not DHHC09^C190A could complement the salt-hypersensitive phenotype of dhhc09 plants. Phenotypic comparison (D) of the transgenic seedlings overexpressing DHHC09 and DHHC09^C190A in the dhhc09-10 background (DHHC09-YFP/dhhc09-1/3 and DHHC09^C190A-YFP/dhhc09-2/4) under salt stress. Fifteen-d-old transgenic seedlings as well as their corresponding WT were treated with 140 mM NaCl for 10 and 8 d, respectively, and recovered for 6 d. Survival rates (E) and (F) of transgenic and WT plants in (D) were analyzed after 6 d of recovery. Forty plants in each line were used for survival rate analysis. Data are presented as mean ± SD (n = 3, **P ≤ 0.01, One-way ANOVA). Bar = 2.5 cm. G) S-acylation analysis of STRK1, STRK1^C5,10A, STRK1^C5,14A, STRK1^C10,14A, STRK1^C5,10,14A, and STRK1^K95E co-expressed with DHHC09 or DHHC09^C190A in akr1 yeast mutant by ABE assay. All the pairings of fusion proteins were co-expressed in the akr1 yeast mutant as indicated. The pairing of EV + STRK1-V5 and DHHC14-V5 + STRK1-V5 were also expressed in akr1 yeast mutant as the negative controls. DHHC09-V5 is marked by asterisks. The WT and mutant STRK1 are marked with pound signs. DHHC14-V5 is marked by circles. The S-acylated bands were detected by an anti-V5 antibody. The bands marked with anti-V5 antibody indicate the S-acylated STRK1 as well as the auto-S-acylated DHHC09 and DHHC14 under the lanes in “S-acylation”. H) S-acylation analysis of STRK1 in rice by ABE assay. WT, wild-type plants; DHHC09-FLAG, DHHC09-FLAG-3 transgenic rice plants (T1); dhhc09, dhhc09-10 plants (T1). S-acylated STRK1 bands were detected using the polyclonal anti-STRK1 antibody. The experiments were repeated 3 independent times and one replicate is presented.

DHHC09-YFP and DHHC09^C190A-YFP were overexpressed in WT and dhhc09 rice plants (Supplementary Fig. S5, A and D), and their salt stress response was analyzed. We found that DHHC09 but not DHHC09^C190A markedly improved the salt tolerance (Supplementary Figs. S6B to S6D). Similarly, DHHC09 but not DHHC09^C190A could complement the salt-hypersensitive phenotype of dhhc09 plants (Figs. 4D to 4F). These results indicate that DHHC09 is a functional S-acyltransferase, and the Cys190 in the DHHC motif is essential for its S-acyltransferase activity and salt signal transduction in rice.

The ability of DHHC09 to S-acylate STRK1 was investigated by individually co-expressing DHHC09 and DHHC09^C190A with STRK1 in the akr1 yeast mutant. Both S-acylated STRK1 and auto-S-acylated DHHC09 were detected in the NH₂OH-treated samples from DHHC09 + STRK1/akr1 strains, whereas these S-acylation bands were hardly detected in the samples from DHHC09^C190A + STRK1/akr1 and empty vector (EV)+STRK1/akr1 strains, regardless of NH₂OH treatment (Fig. 4G). In addition, the S-acylation signal was still detected in STRK1^C5,10A, STRK1^C5,14A, and STRK1^C10,14A but not in STRK1^C5,10,14A co-expressed with DHHC09 in the akr1 yeast mutant, indicating that the 3 cysteine residues (Cys5, 10, 14) of STRK1 are the S-acylation sites recognized by DHHC09. By contrast, another PM-anchored DHHC protein DHHC14 could not S-acylate STRK1 in the akr1 yeast mutant, but its auto-S-acylation signal was still detected. Strikingly, similar to the WT STRK1, the kinase-dead form STRK1^K95E could be S-acylated by DHHC09 in the akr1 yeast mutant (Fig. 4G), indicating that DHHC09 S-acylates STRK1 regardless of its kinase activity. These results indicate that STRK1 is a substrate directly and specifically S-acylated by DHHC09 in yeast, and this modification is dependent on the auto-S-acylation activity of DHHC09. Additionally, we constructed STRK1 knockout lines and prepared anti-STRK1 polyclonal antibodies by immunizing mice to determine the S-acylation of STRK1 by DHHC09 in planta (Supplementary Fig. S7, A to C). The S-acylated STRK1 bands were detected in the NH₂OH-treated samples from DHHC09-FLAG and WT lines but not in the samples from dhhc09 mutant (Fig. 4H). These results demonstrate that DHHC09 is a functional S-acyltransferase and specifically S-acylates STRK1 in vivo.

DHHC09-mediated S-acylation mainly determines STRK1 targeting to the PM

To determine the effect of DHHC09-mediated S-acylation on the subcellular localization of STRK1, the fusion protein STRK1-YFP was transiently expressed in WT and dhhc09 rice protoplasts. We found that the PM localization of STRK1-YFP in WT protoplasts was mostly altered to the cytoplasmic localization in dhhc09 protoplasts, but a small proportion of STRK1-YFP was still detected at the PM of dhhc09 protoplasts (Fig. 5A). Subcellular fractionation assays further demonstrated that STRK1-YFP was only detected in the microsomal fraction and PM derived from WT and DHHC09-YFP overexpressing plants. In contrast, most of STRK1-YFP was detected in the cytoplasm, with a small proportion in the microsomal fraction and PM derived from dhhc09 mutant (Fig. 5B), consistent with the subcellular localization of STRK1^C5,10,14A-YFP in transgenic rice plants (Fig. 1I). S-acylated STRK1 was detected in the microsomal fraction treated with NH₂OH from DHHC09-YFP overexpressing and WT plants but not in the microsomal fraction and cytoplasm from dhhc09 mutant (Fig. 5C). These results indicate that STRK1 targeting to the PM mainly depends on the DHHC09-mediated S-acylation.

Figure 5.

The salt signal transduction is dependent on DHHC09-mediated STRK1 PM location. A) Subcellular localization of STRK1-YFP in WT and dhhc09 protoplasts. The protoplasts were treated with FM4-64 (2 μM) for 5 min to show the PM. BF, bright field; YFP, yellow fluorescence protein. Bar = 10 μm. B) Distribution of STRK1 in WT, DHHC09 overexpressing, and dhhc09 plants. STRK1 and DHHC09-YFP were detected by immunoblot with anti-STRK1 and anti-GFP antibodies, respectively. S, soluble protein fraction; U, microsomal fraction; EM, microsomal fraction isolated from the endomembrane; PM, microsomal fraction isolated from the PM. GAPDH/Actin and AHA3 were used as loading controls for the cytoplasm and PM, respectively. C) S-acylation of STRK1 in S and U. S-acylation states of STRK1 in U and S were detected with the anti-STRK1 antibody (left panels). The auto-S-acyaltion states of DHHC09-YFP were detected with an anti-GFP antibody (right panels). The lanes in “S-acylation” show the amount of STRK1 and DHHC09-YFP bound to the neutravidin-agarose beads with (+) or without (−) NH₂OH treatment. “Loading” controls show equal amounts of proteins loaded. D) to (F) The function of STRK1 in salt signal transduction depends on the DHHC09-mediated S-acylation. Phenotypes (D) of the transgenic seedlings overexpressing STRK1 (S-COM-4/5) and STRK1 + DHHC09 (SD-COM-4/6) in the dhhc09-10 background under salt stress. S-COM-4/5 and SD-COM-4/6 seedlings as well as their corresponding wild-type (WT) were treated with 140 mM NaCl for 8 and 10 d, respectively. Survival rates (E) and (F) of transgenic and WT plants in (D) were analyzed after 6 d of recovery. Data are presented as mean ± SD (n = 3, **P ≤ 0.01, One-way ANOVA). Bar = 2.5 cm.

A complementation test was performed to investigate the effect of STRK1 and DHHC09 on the salt stress response of dhhc09 mutant. When only the fusion protein STRK1-green fluorescence protein (GFP) was overexpressed in dhhc09 mutant (STRK1-GFP/dhhc09, S-COM) (Supplementary Fig. S5E), no mitigating effect was observed under salt stress conditions (Fig. 5, D and E). Conversely, when both STRK1-GFP and DHHC09-FLAG were co-overexpressed in dhhc09 mutant (STRK1-GFP + DHHC09-FLAG/dhhc09, SD-COM) (Supplementary Fig. S5E), the salt tolerance was significantly improved compared to WT (Fig. 5, D and F). It was notable that the survival rates (63.3% to 68.3%) of SD-COM lines (Fig. 5F) were markedly higher than those (51.6% to 52.5%) of DHHC09-YFP/dhhc09 lines under salt stress conditions (Fig. 4F). These results indicate that the function of STRK1 in salt signal transduction depends on the DHHC09-mediated S-acylation.

DHHC09 promotes the salt signaling from STRK1 to CatC via transphosphorylation

We previously identified that a tyrosine phosphorylation cascade might exist in the proposed salt signaling pathway RLK–STRK1–CatC and STRK1 has autophosphorylation in rice (Zhou et al. 2018). The endogenous STRK1 and CATs were immunoprecipitated using the anti-STRK1 and anti-CAT antibodies, respectively, from WT and transgenic plants to determine the effects of DHHC09 on their tyrosine phosphorylation levels. The tyrosine phosphorylation levels of STRK1 and CATs in DHHC09 overexpressing plants were significantly higher than those in WT plants, whereas that of CATs was markedly lower in dhhc09 mutant (Fig. 6, A to C). Notably, the tyrosine phosphorylation level of STRK1 in dhhc09 mutant was markedly higher than that in DHHC09 overexpressing plants (Fig. 6, A and B). These results indicate that DHHC09 promotes the intracellular signaling from STRK1 to CATs via transphosphorylation, and its deficiency leads to a blockage in this signaling pathway and the accumulation of tyrosine phosphorylation in STRK1.

Figure 6.

DHHC09 promotes the salt signaling from STRK1 to CatC via tyrosine phosphorylation. A) Tyrosine phosphorylation levels of STRK1 and CATs in WT and DHHC09 transgenic plants. B and C) Statistics on tyrosine phosphorylation levels of STRK1 (B) and CATs (C) in (A). Data are presented as mean ± SD (n = 6, **P ≤ 0.01, One-way ANOVA). D) Tyrosine phosphorylation levels of STRK1 and CATs in DHHC09 transgenic plants treated with 140 mM NaCl for the indicated time. E and F) Statistics on tyrosine phosphorylation levels of STRK1 (E) and CATs (F) in (D). Endogenous STRK1 and CATs in (A) and (D) were immunoprecipitated from WT, DHHC09-FLAG, and dhhc09 plants with anti-STRK1 and anti-CAT mouse antibodies (IP), respectively. The tyrosine phosphorylation levels of STRK1 and CATs were analyzed by immunoblot (WB) using an anti-phospho-tyrosine (anti-pTyr) rabbit antibody. The experiments were repeated 3 independent times and all the data were presented. G and H) CAT activities (G) and H₂O₂ contents (H) in DHHC09 transgenic seedlings treated with 140 mM NaCl for the indicated time. Data are presented as mean ± SD (n = 3). I) Analysis of STRK1-YFP and STRK1^C5,10,14A-YFP in DRMs by a sucrose density gradient. The assays were performed in stable genetic lines (STRK1-YFP and STRK1^C5,10,14A-YFP). STRK1-YFP and STRK1^C5,10,14A-YFP were detected by immunoblot using an anti-GFP antibody. J) Analysis of endogenous STRK1 in DRMs. WT and dhhc09 plants were used for analysis. The endogenous STRK1 was detected by immunoblot using the anti-STRK1 antibody. PMs in (I) and (J) were treated or not with methyl β-cyclodextrin (mβCD) and submitted to the DRM isolation procedure. Twelve fractions collected from the top to the bottom of the gradient were precipitated, and then were analyzed by SDS-PAGE and immunoblot. The fractions 5–7 contained lipid nanodomains. DRMs, detergent-resistant membranes (the low-density fractions 5–7). DSMs, detergent-soluble membranes (the high-density fractions 9–12). The experiments were repeated 3 independent times and one replicate is presented.

Salt stress was previously reported to stimulate the tyrosine phosphorylation of CatC by STRK1, which enhances the enzymatic activity of CatC in rice (Zhou et al. 2018). The effects of DHHC09 on the tyrosine phosphorylation levels of STRK1 and CATs in response to salt stress were further investigated. After 10 and 20 min of NaCl treatment, the tyrosine phosphorylation levels of STRK1 were moderately increased in WT plants but sharply increased with the extension of NaCl treatment time in DHHC09 overexpressing and dhhc09 plants (Fig. 6, D and E). Especially, compared with no salt-treated controls, the tyrosine phosphorylation levels of STRK1 treated with NaCl for 20 min were increased to about four-fold in dhhc09 plants but only two-fold in DHHC09 overexpressing plants (Fig. 6, D and E), further indicating that DHHC09 deficiency results in the accumulation of tyrosine phosphorylation in STRK1. Conversely, compared with the WT plants, the tyrosine phosphorylation levels of CATs were substantially increased in DHHC09 overexpressing plants but markedly decreased in dhhc09 plants after salt treatment (Fig. 6, D and F). These results further demonstrate that DHHC09 promotes the salt signaling from STRK1 to CATs via transphosphorylation.

Moreover, the S-acylation levels of DHHC09 and STRK1 were enhanced in WT and DHHC09 overexpressing plants after salt treatments (Supplementary Figs. S7D to S7H), indicating that, besides the tyrosine phosphorylation, salt stress also stimulates the S-acylation of DHHC09 and STRK1. The CAT activity and H₂O₂ concentration were measured in the shoots of these transgenic and WT plants. DHHC09 overexpressing plants exhibited higher CAT activities and lower H₂O₂ accumulation than WT plants, whereas dhhc09 plants displayed the opposite results under salt stress (Figs. 6, G and H). These results further support the hypothesis that DHHC09 improves salt tolerance by promoting the salt signaling from STRK1 to CATs via transphosphorylation to regulate H₂O₂ homeostasis. Additionally, STRK1 and CatC still interact in dhhc09 protoplasts. However, their subcellular site of interaction was altered from only the PM in WT to the cytoplasm with a small proportion at the PM in dhhc09 protoplasts (Supplementary Fig. S8), indicating that DHHC09-mediated S-acylation does not affect the STRK1–CatC interaction.

PM is known to be highly compartmentalized into lipid nanodomains, also known as lipid rafts, and that lateral segregation of proteins and lipids plays a pivotal role in cellular processes such as signal transduction and membrane protein trafficking (Jaillais and Ott 2020). The fraction of detergent-resistant membranes (DRMs; also referred to as detergent-insoluble membranes) is thought to consist of aggregates of lipid nanodomains, and S-acylation is a hallmark of receptor targeting into nanodomains (Borner et al. 2005; Chen et al. 2019). To determine the effect of DHHC09-mediated S-acylation on STRK1 targeting the lipid nanodomains, DRMs were isolated from the PM using a Triton X-100-to-protein ratio of 15 as described previously (Raffaele et al. 2009), and twelve fractions of equal volume were collected from the top to the bottom of the gradient and detected by immunoblot.

STRK1-YFP was detected in both DRMs (the low-density fractions 5–7) and detergent-soluble membranes (DSMs; the high-density fractions 9–12) derived from STRK1-YFP overexpressing plants, whereas it was only observed in DSMs when treated with methyl b-cyclodextrin (mβCD), a chelator of free sterols used as the nanodomain destroyer (Figs. 6I). By contrast, STRK1^C5,10,14A-YFP was only detected in DSMs with or without mβCD treatment. These results imply that STRK1 might be partially located in lipid nanodomains and this location is determined by S-acylation modification. Similarly, the endogenous STRK1 was detected in both DRMs and DSMs derived from WT plants but only in DSMs derived from the dhhc09 mutant, and the STRK1 originally located in DRMs of WT plants were altered to DSMs when treated with mβCD (Figs. 6J). Importantly, mβCD treatment does not substantially affect the amount of PM-localized STRK1, suggesting a reorganization of this protein within the PM (Fig. 6, I and J). These results further suggest that DHHC09-mediated S-acylation might determine STRK1 targeting to the lipid nanodomains, and the STRK1 localized in lipid nanodomains might play a pivotal role in the salt signaling from STRK1 to CatC via transphosphorylation.

DHHC09 improves the grain yield of rice under salt stress

The effects of DHHC09 on the agronomic traits, especially grain yield, were further investigated in rice at the reproductive stage under salt stress. The transgenic and WT plants were planted in plastic pots under normal growth conditions and then exposed to 1% NaCl at the panicle development stage. After salt treatments, DHHC09 overexpressing lines exhibited more green leaves, more straw weight, more vigorous root system, and longer root length than WT plants, whereas dhhc09 lines displayed the opposite results (Fig. 7A; Supplementary Figs. S9, C and F to S9O). By contrast, no noticeable difference was observed in these traits between WT and transgenic plants under normal conditions (Supplementary Figs. S9, A and C, S9F to S9O). These results indicate that DHHC09 significantly improves the growth at the reproductive stage under salt stress.

Figure 7.

DHHC09 improves the grain yield of rice under salt stress at the reproductive stage. A and B) Phenotypic comparison of rice plants under salt stress. Salt stress of plants was initiated at the panicle development stage by exposure to 1% (about 170 mM) NaCl for the indicated time, and then the plants were recovered with normal irrigation for 10 d (A) and harvested (B). C) to L) Effective panicle number (C) and (D), weight per panicle (E) and (F), seed setting rate (G) and (H), thousand-seed weight (I) and (J), and grain yield per plant (K) and (L) of DHHC09-FLAG overexpressing (OE-3/5), dhhc09 mutant (dhhc09-10/18), and WT plants in (A) after 10 d of recovery. Data are presented as mean ± SD (**P ≤ 0.01, Two-way ANOVA).

Moreover, the effective panicle number, panicle weight, thousand-seed weight, and grain yield of DHHC09 overexpressing lines were significantly higher than those of WT plants, whereas those of dhhc09 lines were markedly lower after salt treatments (Fig. 7, B to L). Notably, the grain yield per plant of DHHC09 overexpressing lines was increased by about 120% compared with WT plants under salt stress. In contrast, no noticeable difference was observed in these traits between WT and transgenic plants under normal conditions (Fig. 7, C to L; Supplementary Fig. S9B). No noticeable difference was detected in plant height and seed setting rate between WT and transgenic plants under salt stress and normal conditions (Fig. 7, G and H; Supplementary Figs. S9, D and E). These results demonstrate that DHHC09 improves salt tolerance at the reproductive stage and limits the grain yield loss of rice under salt stress.

STRK1 was previously found to improve the grain yield of rice under salt stress (Zhou et al. 2018). To investigate whether the improvement of grain yield by STRK1 is dependent on DHHC09, the agronomic traits were also determined in STRK1-GFP/dhhc09 (S-COM) and STRK1-GFP + DHHC09-FLAG/dhhc09 (SD-COM) plants at the reproductive stage under salt stress. Except for the plant height and seed setting rate, all the straw weight, root system, root length, effective panicle number, panicle weight, thousand-seed weight, and grain yield were significantly improved in SD-COM lines but not in S-COM lines compared with WT plants after salt treatments (Supplementary Figs. S10 and S11). Notably, the grain yield per plant of SD-COM lines was increased by about 110% compared to WT plants under salt stress (Supplementary Figs. S10, B and L). These results further demonstrate that STRK1 improves rice grain yield depending on the DHHC09-mediated S-acylation under salt stress.

DISCUSSION

DHHC09 positively regulates salt and oxidative stress tolerance in rice

Salt stress not only inhibits crop growth but also reduces crop yields during the reproductive period. Sustainable growth and crop yield under salt stress become the main criteria for breeding salt-tolerant crops (Zeng and Shannon 2000; Zhu 2001). RLKs perceive the extracellular signals or stimuli through their extracellular domains and propagate the signals to RLCKs via their intracellular kinase domains by transphosphorylation (Shiu et al. 2004). Thus, protein phosphorylation via RLKs/RLCKs plays an essential role in plant responses to biotic and abiotic stresses. OsRLCK176/118 interact with the RLK protein SPL11 cell death suppressor 2 (SDS2) and phosphorylate the NADPH oxidase OsRbohB to stimulate ROS production and plant immunity in rice (Fan et al. 2018). Recently, the RLK protein FERONIA was reported to phosphorylate phytochrome B to regulate plant growth and salt tolerance in Arabidopsis (Liu et al. 2023).

Some RLCKs are subjected to S-acylation modification for membrane anchoring and signal transduction in plants. The RLCKs Lost In Pollen tube guidance 1 (LIP1) and LIP2 are anchored to pollen tube membranes by S-acylation and are involved in the perception of the ovule-secreted peptide signal AtLURE1 (Liu et al. 2013). The RLCK protein PBS1-LIKE 19 (PBL19) is S-acylated to anchor to the PM and participates in the ENHANCED DISEASE SUSCEPTIBILITY (EDS1)-dependent plant immunity in Arabidopsis (Li et al. 2022). Previously, we demonstrated that the RLCK protein STRK1 is S-acylated to anchor to the PM and phosphorylates CatC at Tyr210 to promote salt tolerance in rice (Zhou et al. 2018). However, the molecular mechanism underlying S-acylation modification involved in STRK1-mediated salt signal transduction remains unknown. In this study, we further verified that STRK1 is indeed an S-acylated protein with S-acylation sites at Cys5, 10, and 14 by ABE and APE assays, and this S-acylation is required for STRK1 to function in salt signal transduction (Fig. 1; Supplementary Fig. S1). These results shed light on the mechanism underlying S-acylation modification participating in RLK/RLCK-mediated salt signal transduction in plants.

DHHC proteins catalyze the S-acylation of substrates, and the DHHC motif is required for S-acyltransferase activity (Qi et al. 2013; Tian et al. 2022). DHHC09 which interacts with STRK1 was then successfully screened from an OsDHHC cDNA library, and the Cys190 in its DHHC motif is required for their interaction (Fig. 2; Supplementary Fig. S2). DHHC09 encodes 2 splice variants, DHHC09-I and DHHC09-II, both of which contain the conserved and characteristic DHHC motif and specifically interact with STRK1 at the PM (Supplementary Figs. S2 and S3). DHHC09 was further demonstrated to have auto-S-acylation and to specifically S-acylate STRK1 in yeast and rice (Fig. 4; Supplementary Fig. S6). Especially, the catalytically inactive form DHHC09^C190A did not complement the thermosensitive phenotype of akr1 yeast mutant, nor did it S-acylate STRK1 in vivo. STRK1 was previously found to positively regulate the salt and oxidative stress tolerance in rice (Zhou et al. 2018). Similarly, DHHC09 was also identified to positively regulate the salt and oxidative stress tolerance in rice (Fig. 3). It was worth noting that H₂O₂ was observed to accumulate less in the leaves of DHHC09 overexpressing plants but more in those of dhhc09 mutants by DAB staining (Fig. 3M), which was further confirmed by the following results that DHHC09 overexpressing plants exhibited higher CAT activities and lower H₂O₂ accumulation than WT plants, whereas dhhc09 plants displayed the opposite results under salt stress (Fig. 6, G and H).

Notably, DHHC09-II overexpressing lines also exhibited higher tolerance to salt stress than WT plants (Supplementary Figs. S3H to S3K). These results indicate the importance of DHHC09-mediated S-acylation of STRK1 in salt stress signal transduction in rice and the improved tolerance of DHHC09 overexpressing plants to salt stress resulting from the enhanced reactive oxygen species (ROS)-scavenging capacity. Pot experiments at the reproductive stage showed that the DHHC09 overexpressing plants exhibited more robust growth and higher grain yield than WT plants under salt stress, whereas no noticeable difference was observed under normal conditions (Fig. 7; Supplementary Fig. S9), indicating that DHHC09 enhances salt tolerance without noticeable growth and yield penalties in rice. Together, these results suggest that DHHC09 is a promising candidate gene for yield improvement in rice under salt stress conditions.

DHHC09-mediated S-acylation affects STRK1 Pm localization but not STRK1–CatC interaction

S-acylation is an essential post-translational modification and regulates protein membrane localization and functions (Greaves and Chamberlain 2011). For example, the calcium-dependent protein kinase ZmCPK9 is S-acylated by Tip Growth Defective 1 (ZmTIP1), a DHHC protein, to anchor to the PM, which contributes to drought tolerance in maize (Zea mays) (Zhang et al. 2019). The localization of calcineurin-B-like proteins OsCBL2 and OsCBL3 in the endomembrane (EM) depends on S-acylation mediated by OsDHHC30 and confers to salt and oxidative stress tolerance in rice (Tian et al. 2022). STRK1 was previously identified to anchor and interact with CatC at the PM via S-acylation (Zhou et al. 2018). As expected, subcellular fractionation assays found that STRK1-YFP was distributed only in the PM. By contrast, two-thirds of STRK1^C5,10,14A-YFP, the non-S-acylated form, were transferred to the cytoplasm, while a third remained in the PM (Fig. 1, I to K; Supplementary Figs. S1C to S1E). Accordingly, a small proportion of non-S-acylated STRK1 was still observed at the PM and PM fraction derived from the dhhc09 mutant (Fig. 5, A to C). These results suggest that STRK1 targeting to PM is mainly determined by S-acylation, but is likely also influenced by other factors.

A canonical RLK includes an extracellular domain, a transmembrane domain, and an intracellular protein kinase domain (Torii 2000). RLKs act as transmembrane proteins to perceive the signals through their extracellular domains and propagate them to the downstream RLCKs via their intracellular kinase domains (Shiu et al. 2004). Although RLCKs lack the transmembrane domains, they are potentially anchored to the PM through the formation of a complex with RLK or S-acylation and/or myristoylation motifs (Veronese et al. 2006; Kim et al. 2011; Tanaka et al. 2012; Li et al. 2016; Li et al. 2022). Since a small proportion of non-S-acylated STRK1 and STRK1^C5,10,14A remained in the PM (Figs. 1, I to K and 5, A to C), we thus speculate that the PM localization of non-S-acylated STRK1 including STRK1^C5,10,14A might result from their interaction with an unknown RLK. Interestingly, some RLKs are also S-acylated by DHHC proteins, which does not alter their PM anchoring but affects their functions, especially signal transduction in plants. For instance, the immune receptor P2K1, an RLK protein, is S-acylated by AtPAT5/9, which ultimately dampens the immune response in Arabidopsis (Chen et al. 2021).

The S-acylation modification of some RLCKs not only determines their subcellular localization but also affects their interaction with other proteins. For example, the RLCK protein PBL19 is S-acylated to anchor to the PM and activated by the LYSIN MOTIF-CONTAINING RECEPTOR-LIKE KINASE 5 (LYK5)−CHITIN ELICITOR RECEPTOR KINASE 1 (CERK1) receptor complex under chitin elicitation; a fraction of non-S-acylated PBL19 proteins are relocated to the nucleus and induce transcriptional self-amplification mainly through the transcription factor WRKY8, while a proportion of them are cleaved to interact with and phosphorylate the substrate protein EDS1 in the cytoplasm (Li et al. 2022). Strikingly, DHHC09-mediated S-acylation did not affect the STRK1–CatC interaction, and STRK1 could normally interact with CatC in dhhc09 protoplasts (Supplementary Fig. S8). These results indicate that S-acylation modification affects STRK1 PM localization but not STRK1–CatC interaction. Although a proportion of non-S-acylated STRK1 and STRK1^C5,10,14A were still localized at the PM, all the STRK1^C5,10,14A overexpressing plants, dhhc09 mutants, and S-COM plants completely lost the tolerance to salt stress (Figs. 1, 3, and 5). These results suggest that S-acylation of STRK1 by DHHC09 determines its function in salt stress response.

S-acylation of STRK1 by DHHC09 is required for the salt signaling from STRK1 to CatC via tyrosine transphosphorylation

RLCKs often functionally and physically associate with RLKs as a complex that relays intracellular signaling via transphosphorylation events (Kim et al. 2011; Lin et al. 2013). Particularly, tyrosine phosphorylation is an essential regulatory mechanism in initiating and transducing RLK/RLCK-mediated BR and innate immune signaling (Macho et al. 2015). CAT is a phosphoprotein whose function is tightly regulated by phosphorylation (Cao et al. 2003; Zou et al. 2015; Zhou et al. 2018). Tyr231 and Tyr386 of CAT are phosphorylated by c-Abl/Arg tyrosine kinase in mice (Mus musculus), promoting CAT's ubiquitination and degradation (Cao et al. 2003). We previously proposed that a tyrosine phosphorylation cascade exists in the salt signaling pathway RLK–STRK1–CatC, where STRK1 acts as a key component to transduce the signal from an unknown RLK to the effector CatC via tyrosine phosphorylation to maintain H₂O₂ homeostasis upon salt stress (Zhou et al. 2018).

DHHC09 overexpressing plants exhibited higher tyrosine phosphorylation levels of STRK1 and CATs than WT, whereas dhhc09 mutants displayed lower tyrosine phosphorylation levels of CATs but higher tyrosine phosphorylation levels of STRK1, especially under salt stress (Fig. 6, A to F). These results indicate that DHHC09 promotes the salt signaling from STRK1 to CATs via tyrosine phosphorylation, and its deficiency results in a blockage of this signaling transduction and an accumulation of tyrosine phosphorylation in STRK1, but does not affect the phosphorylation of STRK1 by an unknown RLK and/or its autophosphorylation. The Na⁺ sensor glycosyl inositol phosphorylceramide (GIPC) sphingolipids and the H₂O₂ sensor HYDROGEN PEROXIDE-INDUCED Ca²⁺ INCREASE 1 (HPCA1) have been identified to anchor on the cell surface and trigger Ca²⁺ influx in Arabidopsis (Jiang et al. 2019; Wu et al. 2020). We propose that an unknown upstream RLK might perceive the extracellular Na⁺ and H₂O₂, phosphorylating and activating the S-acylated STRK1 by DHHC09. Then the activated STRK1 phosphorylates CatC at Tyr210 and activates CatC at the PM to eliminate the excessive H₂O₂, thereby improving rice salt and oxidative stress tolerance.

CATs have been demonstrated to be activated at the PM and then transported to the peroxisome or even the nucleus to perform their function (Zou et al. 2015; Zhou et al. 2018; Zhang et al. 2020; Gao et al. 2021; Al-Hajaya et al. 2022). The PM anchoring of STRK1 mediated by S-acylation plays a critical role in the activation of CatC upon salt stress in rice. Surprisingly, a third of non-S-acylated STRK1 remained in the PM, but did not phosphorylate CatC in dhhc09 mutants (Figs. 5, A and B, and 6, A to E). Lipid nanodomains on the cell surface participate in signal transduction. In general, upon intra- or extracellular stimuli, lipid nanodomains recruit proteins to a new micro-environment, where the phosphorylation state can be modified by local kinases and phosphatases, resulting in downstream signaling cascades (Simons and Toomre 2000; Furt et al. 2010). We hypothesize that DHHC09-mediated S-acylation determines STRK1 targeting to a lipid nanodomain of PM, where STRK1 can properly function in the salt signal transduction in rice.

As expected, STRK1 was detected in both DRMs and DSMs, but only in DSMs when treated with mβCD; STRK1^C5,10,14A in transgenic plants and the endogenous STRK1 in dhhc09 mutant were only detected in DSMs regardless of mβCD treatment (Fig. 6, I and J). These preliminarily results imply that STRK1 might be partially located in lipid nanodomains, while its non-S-acylated forms, the endogenous STRK1 in dhhc09 mutant and STRK1^C5,10,14A, cannot be properly located in lipid nanodomains. Given the blockage of salt signaling transduction and the accumulation of tyrosine phosphorylation in STRK1 resulting from the deficiency of DHHC09 (Fig. 6, A to F), the S-acylated STRK1 in lipid nanodomains might be considered to play a pivotal role in the salt signaling from STRK1 to CatC via transphosphorylation. Besides Tyr210 phosphorylated by STRK1, Tyr360 is another tyrosine phosphorylation site of CatC, which might be phosphorylated by an unknown kinase (Zhou et al. 2018). We speculated that the tyrosine phosphorylation signal of CATs in dhhc09 mutant might result from Tyr360 phosphorylation. The tyrosine phosphorylation level of CATs in dhhc09 mutant was markedly lower than that of WT plants under normal conditions (Fig. 6, A and C), indicating that a small proportion of CATs is still phosphorylated by STRK1 in WT plants under normal conditions.

Based on these data, we propose a working model of DHHC09 participating in the STRK1-mediated salt or oxidative stress signaling (Fig. 8). Under normal conditions, only a small proportion of STRK1 is S-acylated at Cys5, 10, 14 by DHHC09, located in the lipid nanodomains and phosphorylated by an unknown RLK. A small proportion of CatC is phosphorylated at Tyr210 by the S-acylated and phosphorylated STRK1 at the lipid nanodomains and then transported to the peroxisome to keep lower CAT activity, thus keeping an appropriate H₂O₂ level as the molecular signal to sustain rice growth and development. Upon salt stress, the unknown RLK might perceive the extracellular Na⁺ and H₂O₂, and DHHC09 S-acylates most of STRK1 at Cys5, 10, 14. Then the S-acylated STRK1 is located in the lipid nanodomains and phosphorylated by the unknown RLK. The phosphorylated STRK1 forms homodimers, phosphorylates CatC at Tyr210 and activates CatC at the lipid nanodomains. Ultimately, the activated CatC is transported to the peroxisome to eliminate the excessive H₂O₂ resulting from salt stress, thereby improving rice salt and oxidative stress tolerance. The unknown upstream RLK and the micro-environment of related lipid nanodomains merit further investigation.

Figure 8.

A proposed model for the role of DHHC09 in regulating salt stress tolerance. Under normal conditions, only a small proportion of STRK1 is S-acylated at Cys5, 10, 14 by DHHC09, located in the lipid nanodomains and phosphorylated by an unknown RLK. A small proportion of CatC is phosphorylated at Tyr210 by the S-acylated STRK1 at the lipid nanodomains and then transported to the peroxisome to keep lower CAT activity, thus keeping an appropriate H₂O₂ level to sustain rice growth and development. Upon salt stress, the unknown RLK might perceive the extracellular Na⁺ and H₂O₂, and most of STRK1 are S-acylated by DHHC09 and phosphorylated by the unknown RLK, which in turn phosphorylates and activates CatC at the lipid nanodomains. Finally, the activated CatC is transported to the peroxisome to eliminate the excessive H₂O₂ resulting from salt stress, thereby relieving the oxidative damage and improving rice salt tolerance. The question mark denotes the unknown upstream RLK. The non-S-acylated STRK1 is also shown to interact with the unknown RLK at the PM but not in the lipid nanodomain. The pink wavy lines denote the palmitoyl groups. The red lightning symbols denote the salt or oxidative stress stimulation. The unexplored step between the unknown RLK and STRK1 is marked with a blue dashed line with an arrow. P represents the phosphorylation modification site. The size of arrows represents the amount of signal transduction traffic.

In conclusion, these results demonstrate that DHHC09 S-acylates and anchors STRK1 to the PM, promoting the salt signaling from STRK1 to CatC via transphosphorylation and thereby improving salt tolerance in rice. The significantly improved growth and grain yield under salt stress of DHHC09 overexpressing rice plants indicates that DHHC09 is a promising candidate gene for maintaining yield in crop plants exposed to salt stress.

Materials and methods

Plant materials and stress treatments

Rice (Oryza sativa L.) cv. Kitaake plants were used in this study. The plants were grown in a plant growth chamber (Conviron E7/2, Canada) under a light/dark period of 16/8 h with 50% to 70% humidity and 30/24 ± 1 °C day/night temperature except for the rice at the reproductive stage. Nicotiana benthamiana plants were grown in a greenhouse under controlled conditions (approximately 70% humidity, 16-h-light/8-h-dark cycles, 22/20 °C day/night cycles).

The salt stress, MV treatment, and H₂O₂ treatment were performed as described previously (Zhou et al. 2018; Liu et al. 2023). For salt stress at the seedling stage, 15-d-old seedlings (40 plants each genotype) were transferred to a hydroponic culture solution (0.3 mM KH₂PO₄, 0.35 mM K₂SO₄, 1.0 mM MgSO₄·7H₂O, 0.5 mM Na₂SiO₃·9H₂O, 1.0 mM CaCl₂·2H₂O, 9.0 μM MnCl₂·4H₂O, 20.0 μM H₃BO₃, 0.77 μM ZnSO₄·7H₂O, 0.32 μM CuSO₄·5H₂O, 20.0 μM NaFeEDTA, and 0.39 μM Na₂MoO₄·2H₂O, pH5.5; Liu et al. 2023) containing 140 mM NaCl. To obtain more apparent phenotypes between transgenic and WT plants, 8- and 10-d NaCl treatments were applied for the putative salt-sensitive (STRK1^C5,10,14A-YFP-2/6, dhhc09-10/18, DHHC09^C190A-YFP/dhhc09-2/4, S-COM-4/5) and salt-tolerant seedlings (STRK1-YFP-2/9, DHHC09-FLAG-3/5, DHHC09-YFP/dhhc09-1/3, SD-COM-4/6) as well as their corresponding WT seedlings, respectively. Then, the seedlings were transferred to a normal hydroponic culture solution to recover for 6 d and the survival rates were measured. For salt stress treatment at the reproductive stage, 2 positive transgenic lines with WT plants (30 plants each genotype) were planted in each plastic plot, and were exposed to 1% (w/v, about 170 mM) NaCl for 20 or 26 d at the panicle development stage (about 50-d-old). Then, the NaCl solution was completely emptied from the plastic pots, and the plants were irrigated with normal water to recover for 10 d. For the MV treatment, 4 μM MV was used to treat newly germinated seeds (30 seeds each genotype). After 6 d of MV treatment, the seedling height and chlorophyll content were analyzed. The H₂O₂ treatment was 100 mM H₂O₂ at the 3-leaf stage, and then the leaves treated for 1 d were stained with DAB solution (50 mM Tris-HCl, pH3.8, 1 mg/mL DAB, 0.01% (V/V) Triton X-100). Reverse transcription quantitative PCR (RT-qPCR) analyses were performed using the rice seedlings at the 3-leaf stage treated with 140 mM NaCl or 1% (V/V) H₂O₂. For relative expression levels of the 2 splice variants of DHHC09 (DHHC09-I and DHHC09-II), 5 types of stress [ABA treatment (6 μM); alkaline treatment (75 mM mixed alkali: 62.5 mM NaHCO₃ and 12.5 mM Na₂CO₃, pH9.2–9.4); H₂O₂ treatment (1% (V/V)); high-temperature treatment (45 °C); NaCl treatment (140 mM)] were conducted. The primers are shown in Supplemental Data Set S1.

To study the significance of S-acylated Cys190 of DHHC09 in planta, the coding sequence (CDS) of mutated DHHC09 (DHHC09^C190A) was generated by site-directed mutagenesis and subcloned into the pCAMBIA1300 vector (Cambia) (primers shown in Supplemental Data Set S1). DHHC09^C190A sequences were overexpressed in dhhc09-10, an DHHC09 knockout line, via Agrobacterium (Agrobacterium tumefaciens)-mediated transformation as described previously (Lin et al. 2009). To study the significance of DHHC09-mediated STRK1 S-acylation, STRK1 alone, as well as STRK1 and DHHC09, were overexpressed in dhhc09-10 mutant. The T2 generation of transgenic rice plants was used to investigate the stress response as above. The details of the above plant lines are presented in Supplemental Data Set S2.

Physiological measurements

Following Schmidt et al. (2013) and Ouyang et al. (2007), relative ion accumulation of Na⁺ and K⁺ and relative ion leakage were measured, respectively.

Total chlorophyll content and CAT activity were measured as described previously (Ouyang et al. 2010) with slight modifications. The chlorophyll was extracted using 80% acetone from 150 mg leaf sample powder. The chlorophyll concentration was calculated by spectrophotometrically measuring absorption at 663 and 645 nm. Catalase activity was measured by following the decomposition rate of H₂O₂. 10 μL of enzyme homogenate (10% w/v) was added to 2 ml reaction mixture [50 mM phosphate buffer (pH 7.0), 10 mM H₂O₂] in a final assay volume of 2.0 mL. The disappearance of H₂O₂ is monitored at 240 nm wavelength. Hydrogen Peroxide Assay Kit (Beyotime, China) was used to determine H₂O₂ content in rice leaves. 0.1 g samples were rapidly ground into powder in liquid nitrogen. Then, the samples were homogenized at 2 ml hydrogen peroxide lysis buffer and centrifuged at 12,000 g at 4 °C for 5 min. The 50 μL supernatant was added to the inspection hole and incubated with 100 μL hydrogen peroxide test reagent at 28 °C for 30 min. The absorbance was immediately determined by BioTek Instruments Synergy 2 Multi-Mode Microplate Reader. The concentration of H₂O₂ in the sample was calculated according to the standard curve prepared with known concentrations of H₂O₂. Three independent biological replicates were carried out for each sample. Statistical analysis was determined by Two-way ANOVA.

Analysis of agronomic traits

Various agronomic traits were measured as described by Wei et al. (2022) with minor modification. Effective panicle number, seed setting rate, grain yield per plant, plant height, straw weight, root weight, and root length were measured on a single-plant basis. Filled and unfilled grains of the main panicle were separated manually for measurement of seed setting rate [filled grains/(filled grains + unfilled grains) × 100%]. All filled grains from a single plant were collected and dried for measurements of grain yield per plant. The filled grains were harvested and oven-dried at 105 °C for 30 min and then at 80 °C to constant weight. One thousand seeds were randomly selected and weighed as a replicate, and 10 replicates were performed for each line. Plant height was determined as the height of the main tiller. The straw (including leaves, shoots and panicles) and root parts of these plants were oven-dried at 105 °C for 30 min and then oven-dried at 80 °C until constant weight and weighed to acquire dry weights. Statistical analysis was conducted on 30 individuals from each line. Statistical analysis was determined by Two-way ANOVA.

Acyl-biotinyl exchange (ABE) assays

ABE assay was performed as described previously (Tian et al. 2022). Briefly, each plant sample supernatant was added to saponin and N-ethylmaleimide (NEM) to block free cysteines. The incubated mixture was precipitated with methanol/chloroform. The precipitated protein was then divided into 2 equal portions: one sample served as a control portion, while the other treated with hydroxylamine (NH₂OH), a Cys-palmitoyl thioester linkages cleavage reagent. Following biotinylation a sample of each reaction is removed to act as a loading control. Both remaining liquids were then incubated with high-capacity neutravidin-agarose beads (Thermo Scientific, USA), and then the S-acylated proteins were eluted from beads and detected by immunoblot using the Anti-V5 (ab309485, Abcam, UK, dilutions: 1:5,000), anti-GFP (M20008, Abmart, China, dilutions: 1:5,000) and anti-STRK1 antibodies.

Acyl-PEG exchange (APE) assays

APE assay was performed as described previously (Tian et al. 2022). Briefly, the sample supernatant was incubated with SDS and Tris (2-carboxyethyl) phosphine (TCEP). Free sulfhydryls were alkylated with NEM. The incubated samples were precipitated with methanol/chloroform and then NH₂OH cleavage was performed. Proteins were precipitated, resuspended, and treated with mPEG-Mal for 2 h. The resulting proteins were precipitated and analyzed by immunoblot.

Isolation of PM fractions and DRMs, and mβCD treatment of PMs

PM was purified by two-phase partitioning, adapting the protocol described in Tian et al. (2022). Briefly, plant lysates were supplemented with phenylmethylsulfonyl fluoride and centrifuged at 50,000 g to collect the microsomes. PM and EM were separated from the microsomes over a two-phase system with PEG/Dextran (20% Dextran T-500, 40% PEG, 0.923 g sucrose, 0.2 M potassium phosphate buffer, pH7.8, 2 M KCl, adding sterile water to 9 g). The upper and bottom phases were individually collected and purified again over the two-phase system. Finally, the upper and bottom phases were diluted 3 and 10 times, respectively, and centrifuged at 100,000 g for 1 h. For subcellular fractionation assays, PM and EM were resuspended in 1×SDS loading buffer and then analyzed by immunoblot. For isolation of DRM from PM, PM was resuspended in buffer (5 mM potassium phosphate buffer, pH7.8, 330 mM sucrose, 0.1 mM EDTA, 1 mM DTT, 3 mM KCl). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH)/Actin and ATPase 3 (AHA3) were used as the cytoplasm and PM makers, respectively. Immunoblot analysis was then conducted using anti-FLAG (M20004, Abmart, China, dilutions: 1:5,000), anti-GFP, Anti-AHA3 (bs-2247R, Bioss, China, dilutions: 1:5,000), Anti-Actin (CW0096M, CWBIO, China, dilutions: 1:5,000), Anti-GAPDH (P30008S, Abmart, China, dilutions: 1:5,000) antibodies.

The isolation of DRMs from the PM of rice leaves or protoplasts was performed as described by Raffaele et al. (2009) with minor modification. Briefly, the resuspended PM was divided into 2 equal parts, one as control and the other incubated with 20 mM methyl b-cyclodextrin (mβCD) for 30 min with continuous shaking. Both PM fractions were treated with 1% (V/V) Triton X-100 with shaking on ice for 30 min. Treated PM extracts were brought to a final concentration of 52% sucrose (W/W), overlaid with a sucrose step gradient (concentration decreasing from 60% to 10% with a step size of 5%) in buffer (25 mM Tris, pH7.8, 150 mM NaCl, and 5 mM EDTA) and centrifuged at 175,000 g for 22 h in an SW32Ti rotor (BECKMAN COULTER, USA). DRMs (an opaque band) could be collected below the 30% to 35% layers. Fractions were collected from top to bottom, and proteins were precipitated successively by 20% and 15% trichloroacetic acid (W/V), 1 h on ice, and cold acetone. Pellets were resuspended in 1×SDS loading buffer and then analyzed by immunoblot.

Bimolecular fluorescence complementation (BiFC) assays

Screening DHHCs of STRK1 using an OsDHHC cDNA library and BiFC assays was performed as described previously (Tian et al. 2022). The full-length cDNA of STRK1 was subcloned into a cCFP BiFC expression vector pE3449 (Lee et al. 2008) and then BiFC assays were performed on the OsDHHC cDNA library. Additionally, the sequences of truncated DHHC09 (DHHC09Δ1, DHHC09Δ2, DHHC09Δ3), mutated DHHC09 (DHHC09^C190A), and CatC were individually subcloned into an nVenus BiFC expression vector pE3308 (Lee et al. 2008). The sequence of STRK1 was subcloned into a cCFP BiFC expression vector pE3449 (primers shown in Supplemental Data Set S1). DHHC09Δ1/Δ2/Δ3/CatC and STRK1-cCFP pairing were individually co-expressed in WT or dhhc09 mutant rice protoplasts. Rice protoplasts were prepared according to the method described by Ni et al. (2019). Transient expression was performed using the PEG-mediated transformation method (Zhou et al. 2018). Incubation for 12–16 h after transformation, the rice protoplasts were observed under a confocal laser scanning microscope (Nikon Ti-E + A1 SI, Japan). GFP was excited by 488 nm argon-ion laser, with detection in the fluorescein range (roughly 505 to 550 nm). FM4-64 was excited by 561 nm laser, with detection in the fluorescein range (roughly 610 to 640 nm). The bright field, fluorescent signals (FM4-64 or GFP), and merged signals were recorded.

Pull-down assays

The in vivo pull-down assay was performed as described previously (Wang et al. 2013). GST and GST-STRK1 proteins were individually affinity purified using GST agarose beads (Invitrogen, USA). Total plant proteins (1 mg) of N. benthamiana transiently expressing DHHC09-FLAG were incubated with 10 μg of purified GST-STRK1 bound to GST agarose overnight at 4 °C on a rotary shaker. Immunoblot analysis was then conducted using anti-FLAG and anti-GST (M20007, Abmart, China, dilutions: 1:5,000) antibodies.

Co-immunoprecipitation assays

Co-immunoprecipitation (Co-IP) assay was performed as described previously (Feng et al. 2008). The pair of plasmids Pro35S:DHHC09-FLAG/Pro35S:STRK1-YFP was transiently co-expressed in the N. benthamiana leaves. Anti-GFP beads (Sigma-Aldrich, USA) were used to immunoprecipitate protein complexes. The immunoprecipitated proteins and their partners were then eluted by heating at 95 °C in 2×SDS sample buffer and subjected to 10% SDS-PAGE. Immunoblot analysis was then conducted using anti-FLAG and anti-GFP antibodies (Abmart, China).

Subcellular localization analysis

The CDS of DHHC09 and STRK1 were individually inserted into the pCAMBIA1300-YFP vector (primers shown in Supplemental Data Set S1). The plasmid constructs were transformed into WT and dhhc09 mutant rice protoplasts following Lv et al. (2014). The fluorescence was then observed with a confocal laser scanning microscope (Olympus FV1000, Japan). GFP was excited by 488 nm laser, with detection in the fluorescein range (roughly 505 to 550 nm). FM4-64 was excited by 561 nm laser, with detection in the fluorescein range (roughly 610 to 640 nm).

Complementation of akr1 defects with DHHC09

The CDS of DHHC09 and DHHC09^C190A (mutation from cysteine to alanine in DHHC motif) were individually subcloned into pYES2 vector (ThermoFisher SCIENTIFIC). Two recombinant and empty vectors were individually transformed into akr1 yeast (Saccharomyces cerevisiae) mutant, whose Akr1 encoding an S-acyltransferase was knocked out (Tian et al. 2022). The positive transformants were cultured on selection medium (20 g/L glucose, 6.7 g/L yeast nitrogen base without amino acids, 0.77 g/L -Ura DO sumpplement, 20 g/L agar) at 30 °C and 37 °C. Cell morphology was observed using a confocal laser scanning microscope (Nikon Ti-E + A1 SI, Japan).

Antibody preparation and immunoblot analysis

The STRK1 antibody was generated as described previously (Zhou et al. 2018). Briefly, recombinant GST-STRK1 protein purified from Escherichia coli was used as the antigen. Antigens (preferably in saline) were mixed with an equal volume of the Complete Freund's Adjuvant (for the first immunization, Sigma, USA) or Incomplete Freund's Adjuvant (for the 3 subsequent immunizations, Sigma, USA) to form an emulsion. The BALB/c mice were firstly immunized at 8 weeks of age in the amount of 50 µg/mouse and then immunized every 3 weeks. Two weeks after the fourth immunization, mice were anesthetized by an intraperitoneal injection of tribromoethanol, and blood was collected via cardiac puncture. After overnight clotting at 4 °C, serum was collected by centrifugation at 1,000 g for 10 min and used for further analysis. The WT and T2 generation of STRK1 overexpressing and strk1 knockout plants were used to verify the specificity of anti-STRK1 antibodies. The immunoblots were probed with anti-STRK1 antibodies (1:5,000). The chemiluminescence signal was detected using ECL/fluorescence immunoblot methods. A polyclonal anti-CAT antibody was also prepared from mice using CatC as an antigen (Liu et al. 2023). As it can recognize at least CatB and CatC, it is thus renamed anti-CAT antibody.

Immunoprecipitation and in vivo phosphorylation assays

The DHHC09 transgenic and WT seedlings were treated with 140 mM NaCl for 3 time points (0, 10, 20 min). The total proteins of plant powder were extracted in lysis buffer [50 mM Tris-HCl, pH 7.4, 50 mM NaCl, 1 mM DTT, 0.1% (v/v) Triton X-100, 1 mM PMSF, 10 mM EDTA, 1× protease inhibitor, and 1× phosphatase inhibitor]. The endogenous STRK1 and CATs were immunoprecipitated with protein A/G magic beads (Bimake, USA) with anti-STRK1 and anti-CAT mouse antibodies, respectively, as previously described (Zhou et al. 2018; Liu et al. 2023). The tyrosine phosphorylation levels of STRK1 and CATs were analyzed by immunoblot using anti-phospho-tyrosine rabbit antibody (Abcam, catalog no. ab17302, dilutions: 1:1,000). In addition, the endogenous STRK1 and CATs were immunoblotted by anti-STRK1 mouse antibody and anti-CAT rabbit antibody, respectively.

Reverse transcription quantitative PCR (RT-qPCR)

Total RNA was extracted from the corresponding tissues (root, stem, leaf sheath, leaf, young spikelet), as well as seedling rice of the control (CK) and different stress treatments (ABA, alkali, H₂O₂, high temperature, and salt) using Trizol reagent (Invitrogen, USA) according to the manufacturer's protocols. cDNA was generated using DNase I (Takara, Japan) and SuperScript II Reverse Transcriptase (Invitrogen, USA). For the 2 splice variants of DHHC09 (DHHC09-I and DHHC09-II), the position of the primers is shown in Supplementary Fig. S3A. The sequences of relevant quantitative primers are shown in Supplemental Data Set S1. RT-qPCR was performed with the SYBR Green Mix (Takara, Japan) in an optical 96-well plate with an Applied Biosystems 7900HT Fast Real-Time PCR System (Applied Biosystems, USA). The relative expression of DHHC09-I and DHHC09-II were calculated by the relative quantification method. The data were normalized to the amplification of the OsActin1 gene.

DAB staining assays

DAB staining assays were performed as described previously (Thordal Christensen et al. 1997; Orozco-Cardenas and Ryan 1999; Khokon et al. 2011). Treated (H₂O₂ stress) or untreated seedlings leaves were excised and soaked in DAB solution (50 mM Tris-HCl, pH3.8, 1 mg/mL DAB, 0.01% Triton X-100). The samples were then vacuumed and incubated overnight in the dark. The incubated samples were treated with ethanol until decolorized.

Statistical analysis

Statistical tests were performed as described in the text and figure legends. Statistical data are provided in Supplemental Data Set S3.

Accession Numbers

Sequence data from this article can be found in the Michigan State University Rice Genome Annotation Project database (http://rice.plantbiology.msu.edu) (Ouyang et al. 2007). The accession numbers are presented in Supplemental Data Set S4.

Author contributions

J.-Z.L. and X.-M.L. conceived and supervised the project. Y.T., J.-Z.L., and X.-M.L. designed the experiments. Y.T. performed most of the experiments. H.Z., J.-C.W., C.L., Y.W., Z.-K.Z., and D.-Y.T. performed some experiments. H.-P.Z. performed the protein structure analysis. G.-X.D., G.-F.D., and W.-B.T. analyzed and evaluated the agronomic traits. Y.T. and J.-Z.L. analyzed the data and wrote the manuscript.

Supplementary data

The following materials are available in the online version of this article.

Supplementary Figure S1. S-acylation affects STRK1 Pm localization.

Supplementary Figure S2. Screening the DHHC of S-acylated protein STRK1 using the OsDHHC cDNA library and BiFC assays.

Supplementary Figure S3. Characterization of DHHC09-I and DHHC09-II encoded by 2 splice variants of DHHC09.

Supplementary Figure S4. Subcellular localization and expression pattern of STRK1.

Supplementary Figure S5. Construction and identification of DHHC09 transgenic plants.

Supplementary Figure S6. DHHC09 is a functional S-acyltransferase and improves rice salt tolerance.

Supplementary Figure S7. Effects of DHHC09 on the S-acylation levels of STRK1 and DHHC09 under salt stress.

Supplementary Figure S8. Analysis of STRK1–CatC interaction in WT and dhhc09 mutant protoplasts by BiFC.

Supplementary Figure S9. Growth and yield penalties could not be detected in DHHC09 transgenic rice at the reproductive stage under normal conditions.

Supplementary Figure S10. DHHC09 is required for the improvement of rice grain yield conferred by STRK1 under salt stress.

Supplementary Figure S11. Phenotypic analysis of S-COM and SD-COM transgenic rice plants at the reproductive stage under normal and salt stress conditions.

Supplementary Data Set 1. Primers used in this study.

Supplementary Data Set 2. The details of plant lines used in this study.

Supplementary Data Set 3. Summary of statistical analyses.

Supplementary Data Set 4. Accession numbers of genes mentioned in this study.

Funding

This work was supported by grants from the National Science Foundation of China (32372057 and 31170172 to J.-Z.L., 31871595 to X.-M.L., and 32272139 to Y.W.), China Postdoctoral Innovative Talent Support Program (BX20230110 to C.L.), Hunan Provincial Important Science and Technology Specific Projects (2018NK1010 and 20210897 to X.-M.L.), Natural Science Foundation of Hunan Province, China (2021JJ30097 to D.-Y.T., 2020JJ4004 to J.-Z.L., 2021JJ30101 to H.-P.Z., and 2021JJ40057 to Y.W.), China Postdoctoral Science Foundation (2020M682561 to Y.W.), 2022 National Center of Technology Innovation for Saline-Alkali Tolerant Rice Functional Improvement Project (2022PT1005 to J.-Z.L.), Open Competition Subject of Hainan Yazhou Bay Seed Lab (B21HJ0108 to J.-Z.L.), Public Subject of State Key Laboratory of Hybrid Rice (Hunan Hybrid Rice Research Center) (2019KF02 to J.-Z.L.), and the Public Subject of Guangxi Key Laboratory of Rice Genetics and Breeding (2018-05-Z06-KF05 to J.-Z.L.).

Dive Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

• CDG1 Gramene: At3g26940

• CDG1 Araport: At3g26940

• LIP2 Gramene: AT1G04640

• LIP2 Araport: AT1G04640

• WRKY8 Gramene: AT5G46350

• WRKY8 Araport: AT5G46350

• salt CHEBI: CHEBI:24866

• RLK Gramene: Solyc06g060690

• RLK Araport: Solyc06g060690

• CERK1 Gramene: AT3G21630

• CERK1 Araport: AT3G21630

• LIP1 Gramene: AT2G20860

• LIP1 Araport: AT2G20860

• AHA3 Gramene: AT5G57350

• AHA3 Araport: AT5G57350