Salicylic acid attenuates brassinosteroid signaling via protein de‐S‐acylation

Abstract

Brassinosteroids (BRs) are important plant hormones involved in many aspects of development. Here, we show that BRASSINOSTEROID SIGNALING KINASEs (BSKs), key components of the BR pathway, are precisely controlled via de‐S‐acylation mediated by the defense hormone salicylic acid (SA). Most Arabidopsis BSK members are substrates of S‐acylation, a reversible protein lipidation that is essential for their membrane localization and physiological function. We establish that SA interferes with the plasma membrane localization and function of BSKs by decreasing their S‐acylation levels, identifying ABAPT11 (ALPHA/BETA HYDROLASE DOMAIN‐CONTAINING PROTEIN 17‐LIKE ACYL PROTEIN THIOESTERASE 11) as an enzyme whose expression is quickly induced by SA. ABAPT11 de‐S‐acylates most BSK family members, thus integrating BR and SA signaling for the control of plant development. In summary, we show that BSK‐mediated BR signaling is regulated by SA‐induced protein de‐S‐acylation, which improves our understanding of the function of protein modifications in plant hormone cross talk.

Salicylic acid moderates brassinosteroid signaling in Arabidopsis by initiating the release of a key kinase family from the plasma membrane.

Introduction

Brassinosteroids (BRs) are predominant steroid hormones in plants that contribute to plant development and crop yield (Vert et al, 2005). BRs are critical for the enhancement of plant growth, but their endogenous concentrations are typically low and signaling transduction is precisely controlled (Kim & Russinova, 2020). BRs bind to the extracellular domain of their cognate receptor BRI1 (BRASSINOSTEROID INSENSITIVE 1) residing at the plasma membrane, which results in transphosphorylation between BRI1 and its co‐receptors (Jiang et al, 2013). Activated BRI1 phosphorylates RECEPTOR‐LIKE CYTOPLASMIC KINASEs (RLCKs), including BSKs (BRASSINOSTEROID SIGNALING KINASEs) and CDG1 (CONSTITUTIVE DIFFERENTIAL GROWTH 1), which mediate the phosphorylation of the phosphatase BSU1 (BRI1 SUPPRESSOR 1). BSU1 then inactivates the kinase activity of BIN2 (BRASSINOSTEROID INSENSITIVE 2) via dephosphorylation, leading to the accumulation of activated transcription factors BZR1 (BRASSINAZOLE RESISTANCE 1) and BES1 (BRI1 EMS SUPPRESSOR 1) in the nucleus to control transcription to evoke BR responses (Kim & Wang, 2010; Mao & Li, 2020).

The BSK family is critical for linking BR signaling from the plasma membrane to the cytoplasm. BSKs belong to an RLCK family consisting of 12 members in Arabidopsis (Arabidopsis thaliana) that are redundant positive regulators in the BR pathway (Sreeramulu et al, 2013). In Arabidopsis, BRI1 interacts with and phosphorylates BSKs to activate downstream BR signaling, and overexpression of BSKs suppresses the dwarf phenotypes of BR‐insensitive and BR‐deficient mutant plants (Tang et al, 2008). In rice (Oryza sativa), OsBSK3 also functionally interacts with OsBRI1 (Zhang et al, 2016), supporting a conserved function for this kinase family in BR signaling transduction across plant species. BSKs also participate in other pathways: For instance, several BSK members are associated with embryo development (Neu et al, 2019). Furthermore, BSK1 interacts with the receptor kinase FLS2 (FLAGELLIN SENSING 2) at the plasma membrane (Shi et al, 2013; Su et al, 2021) and directly activates downstream kinase cascades for immunity responses (Yan et al, 2018; Shi et al, 2022). However, how BSKs mediate the cross talk between development and stress responses remains unknown.

BRASSINOSTEROID SIGNALING KINASEs work at the plasma membrane but they lack a transmembrane domain. Myristoylation, a type of protein lipidation, occurs on several BSK members and mediates their localization to the membrane (Shi et al, 2013; Ren et al, 2019). Myristoylation always provides target proteins with a transient membrane association; thus, a second lipid modification may afford a more durable membrane association for intracellular trafficking (Shahinian & Silvius, 1995). Furthermore, given that myristoylation is an irreversible modification (Jiang et al, 2018), dynamic regulation of BSK localization may rely on other strategies. Here, we reveal that most Arabidopsis BSK members are substrates of S‐acylation, a reversible post‐translational modification. Protein S‐acylation, also named S‐palmitoylation, is a type of protein lipidation that transfers long‐chain fatty acids such as palmitate to cysteine residues of protein substrates (Linder & Deschenes, 2007). The S‐acylation status of proteins is regulated via protein S‐acyltransferases (PATs) and de‐S‐acylation enzymes (Li & Qi, 2017).

Although bioinformatic and proteomic approaches have been used to identify potential S‐acylation substrates, the molecular function and catalytic mechanism of this modification remain to be explored in plant cells (Hemsley, 2020). Protein S‐acylation plays a role in plant development and stress responses (Zheng et al, 2019). For instance, this modification is necessary for the plasma membrane localization of cellulose synthase complexes (Kumar et al, 2016). S‐acylation is catalyzed by a conserved family of PATs with a DHHC (Asp‐His‐His‐Cys) motif. There are 24 PATs in Arabidopsis, and several members have been functionally characterized (Batistic, 2012); for example, PAT10 contributes to the localization of calcineurin B‐like proteins to the tonoplast for salt tolerance (Zhou et al, 2013). However, the mechanism of protein de‐S‐acylation has rarely been investigated in plant cells. We recently identified a group of de‐S‐acylation enzymes in Arabidopsis, named ABAPTs (ABHD17‐LIKE ACYL PROTEIN THIOESTERASEs; Liu et al, 2021), but the role of protein de‐S‐acylation in plant hormone signaling transduction is completely unknown.

In the current study, we discovered that salicylic acid (SA), a critical plant hormone contributing to plant immunity (Ding & Ding, 2020), attenuates BR signaling via de‐S‐acylation of BSKs. Phytohormone cross talk is essential for the connection of different biological responses in plant cells (Depuydt & Hardtke, 2011; Ohri et al, 2015). The cross talk between BRs and other hormones is essential for the precise regulation of plant development and stress responses (Nolan et al, 2020). Nevertheless, the functional association between BRs and SA is unclear. BRs stimulate the expression of SA‐activated immunity genes by suppressing BIN2 phosphorylation of TGA (TGACG SEQUENCE‐SPECIFIC BINDING PROTEIN)‐type transcription factors (Han et al, 2022; Kim et al, 2022). However, the mechanism by which SA modulates BR signaling remains unknown. Therefore, our work on suppression of the BSK pathway by SA‐inducible protein de‐S‐acylation will improve our understanding of the cross talk between plant hormones mediated via dynamic protein modification.

Results

Most Arabidopsis BSKs are S‐acylated for their plasma membrane localization in plant cells

To explore the dynamic regulation of BSK localization, we used the bioinformatics software CSS‐PALM (Ren et al, 2008) to analyze protein sequences of all 12 Arabidopsis BSKs; we identified potential S‐acylation sites in all members (Fig 1A). We employed a biotin‐switch assay (Hemsley et al, 2008) to detect S‐acylation of all BSK members in plant cells. We expressed GFP (green fluorescent protein) fused BSKs in Arabidopsis protoplasts and chemically blocked all free cysteine residues in proteins 24 h after transfection. We then specifically removed S‐acyl groups via hydroxylamine treatment to expose cysteine residues previously protected by an S‐acyl group. This treatment was followed by cysteine conjugation with biotin, capture on streptavidin resin, and elution by reducing reagents. We determined that 11 of the 12 BSKs are S‐acylated in plant cells under the conditions tested, BSK4 being the exception (Fig 1B).

Figure 1. Most members of the Arabidopsis BSK family are S‐acylated in plant cells.

1. Sequence alignment of N‐terminal motifs of 12 Arabidopsis BSKs. Full‐length BSK proteins were used for analysis via ClustalW, and their N‐terminal regions in the alignment are shown. Conserved amino acid residues are shown using the ClustalW color code (each residue is assigned a color if the amino acid profile of the alignment at that position meets some minimum criteria specific for the residue type).
2. Detection of S‐acylation of BSKs in plant cells. Twelve BSK proteins were fused with GFP at their C terminus and expressed in Arabidopsis protoplasts for 24 h, and the cells were collected for Biotin‐switch assay. Cell lysates are shown as input, and the S‐acylated proteins enriched on the resin which is dependent on hydroxylamine (NH₂OH) are indicated as pulldown samples. All images of anti‐GFP immunoblots are representative of three biological independent experiments.
3. Effect of 2‐BP on subcellular localization of BSKs in protoplasts. BSK‐GFP was transiently expressed in protoplasts, and 40 μM (final concentration) of 2‐BP, an S‐acylation inhibitor, was added to the medium 8 h before confocal microscopy. The same volume of solvent DMSO was supplemented in the control samples. For the plasma membrane (surface) localization, the fluorescence was maintained as a circle in different scanning layers; for the cytoplasm (nonsurface) localization, the fluorescence was detected inside the cells, but in some scan layers, cortical fluorescence might also be detectable. If the signals were detectable inside the cells in any scan layer, we determined them as cytoplasm localization. The representative GFP (green) signals from three biologically independent experiments were shown. Bars, 5 μm.
4. Effect of 2‐BP on subcellular localization of BSK1 in root cells of intact transgenic plants. Six‐day‐old seedlings of 35S:BSK1‐GFP were treated with 80 μM 2‐BP for 2 h; DMSO was supplemented in the control samples. Representative GFP (green) and merged (with bright field) signals from three biologically independent experiments are shown. Bars, 10 μm.
5. Effect of 2‐BP on the localization of BSK1‐GFP in transgenic plants using a cell fractionation assay. Six‐day‐old seedlings were treated with 80 μM 2‐BP (DMSO was used in the control sample) for 3 h before protein extraction. Total proteins (T) were divided into soluble (S) and pellet (P) fractions using ultra‐centrifugation. The representative immunoblot using an anti‐GFP antibody is shown.

Source data are available online for this figure.

Given that attachment of fatty acids improves the affinity of proteins to lipid bilayers, S‐acylation may contribute to the membrane localization of BSKs. To test this possibility, we treated protoplasts expressing each GFP‐fused BSK protein with 2‐BP (2‐bromopalmitate), a protein S‐acylation inhibitor (Jennings et al, 2009). Under control conditions in the absence of 2‐BP, BSK4, which is not S‐acylated, localized to the nonsurface regions of plant cells, while the remaining 11 BSK members localized to the plasma membrane, precisely on the surface of protoplasts. 2‐BP treatment resulted in a similar distribution throughout the nonsurface regions in the protoplasts for all BSKs (Fig 1C). Because BSKs are also modified by myristoylation, 2‐BP may not completely dissociate them from intracellular membrane systems, but these data supported the function of S‐acylation in the maintenance of the plasma membrane localization of BSKs. To confirm the effect of S‐acylation on BSK localization, we generated stable Arabidopsis transgenic lines for BSK1 and BSK3, which harbor different S‐acylation motifs (two cysteine residues for BSK1 and three cysteine residues for BSK3), as GFP fusions. Confocal microscopy analysis revealed a localization at the plasma membrane for BSK1‐GFP and BSK3‐GFP in intact root and leaf cells under control conditions that was disrupted by treatment with 2‐BP (Figs 1D and EV1A–C). We independently performed a cell fractionation assay, whose results also showed that 2‐BP treatment decreased the membrane association of BSK1 and BSK3 (Figs 1E and EV1D).

Figure EV1. Effect of 2‐BP on the membrane localization of BSK1 and BSK3.

• A, B
  The effect of 2‐BP on the localization of BSK1 and BSK3 in intact leaf cells of the transgenic plants. The 6‐day‐old seedlings of 35S:BSK1‐GFP (A) and 35S:BSK3‐GFP (B) were treated with 80 μM 2‐BP (DMSO was used in the control sample) for 3 h. The representative GFP (green) and merged (with bright field) signals from three biologically independent experiments are shown. Bars, 10 μm.
• C
  The effect of 2‐BP on subcellular localization of BSK3 in root cells of intact transgenic plants. Six‐day‐old seedlings of 35S:BSK3‐GFP were treated with 80 μM 2‐BP for 2 h; DMSO was supplemented in the control samples. The representative GFP (green) and merged (with bright field) signals from three biologically independent experiments are shown. Bars, 10 μm.
• D
  The effect of 2‐BP on the localization BSK3‐GFP in transgenic plants using a cell fractionation assay. The 6‐day‐old seedlings were treated with 80 μM 2‐BP (DMSO was used in the control sample) for 3 h before protein extraction. The total proteins (T) were divided into soluble (S) and pellet (P) fractions using ultra‐centrifugation. The immunoblotting using an anti‐GFP antibody is shown.

Source data are available online for this figure.

We hypothesized that mutations of S‐acylation sites may also disrupt the plasma membrane association of BSKs. Accordingly, we mutated each cysteine residue predicted to be a potential S‐acylation site in BSK1 and BSK3 and repeated the biotin‐switch assay to identify those sites modified in planta. We observed that the C3 residue was the predominant S‐acylation site of BSK1 over the C4 residue (Fig 2A, Appendix Fig S1A), while the C5, C10, and C11 residues of BSK3 all contributed to its S‐acylation (Fig 2B, Appendix Fig S1B). We assessed the subcellular localization of BSK1‐GFP and BSK3‐GFP variants by confocal microscopy in transfected protoplasts and intact root cells, which revealed that mutations of the S‐acylation sites of BSK1‐GFP and BSK3‐GFP disrupted their distribution to the plasma membrane and resulted in their accumulation in the cytoplasm (Fig 2C–F, Appendix Fig S2A–D). Consistently, cellular fractionation assays indicated that mutating the S‐acylation sites dramatically increased the accumulation of both BSK1 and BSK3 variants in the soluble fraction, in contrast to the membrane concentration of the wild‐type BSK1 and BSK3 (Fig 2G and H). The contribution of S‐acylation to the membrane association of BSK1 was smaller than the irreversible myristoylation at its glycine residue at position 2 (G2), as evidenced by the relative distribution of the BSK1 C3S and G2A single mutant variants between the soluble and pellet fractions (Fig EV2A). We conclude that the residual fraction of S‐acylation‐defective BSK1 associated with the plasma membrane may be mediated by myristoylation. Moreover, mutating the myristoylation site (G2A variant) reduced the S‐acylation level of BSK1 (Fig EV2B), suggesting that myristoylation may increase the membrane affinity of BSK1 for subsequent S‐acylation, consistent with the interaction model of dual lipidations (Fhu & Ali, 2021). Taken together, these data support the notion that S‐acylation is necessary for the plasma membrane localization of BSK members.

Figure 2. Mutation of the S‐acylation sites on BSKs results in localization changes.

• A, B
  Identification of S‐acylation sites on BSK1 and BSK3. The wild‐type (WT) and indicated mutant versions of BSK1‐GFP (A) and BSK3‐GFP (B) were expressed in protoplasts for biotin‐switch assay. Images of anti‐GFP immunoblots are representative of three biological independent experiments. The detailed screening data of potential S‐acylation sites are included in Appendix Fig S1.
• C, D
  Subcellular localization of BSK1 and BSK3 with S‐acylation site mutations in protoplasts. WT or S‐acylation site mutation versions of BSK1‐GFP (C) and BSK3‐GFP (D) were expressed in protoplasts, and representative GFP and merged images from three biologically independent experiments are shown. Bars, 5 μm.
• E, F
  Effect of S‐acylation site mutation on the localization of BSK1 and BSK3 in intact root cells. Six‐day‐old seedlings of the WT or S‐acylation site mutant versions of 35S:BSK1‐GFP (E) and 35S:BSK3‐GFP (F) were subjected to localization analysis using confocal microscopy. Representative GFP (green) and merged (with bright field) signals in root cells from three biologically independent experiments are shown. Bars, 10 μm.
• G, H
  The effect of S‐acylation site mutations on the localization of BSK1 and BSK3 in a cell fractionation assay. The WT and mutant versions of BSK1‐GFP (G) and BSK3‐GFP (H) were expressed in protoplasts, and the total proteins (T) were divided into soluble (S) and pellet (P) fractions using ultra‐centrifugation. Representative immunoblotting data using an anti‐GFP antibody are shown.

Source data are available online for this figure.

Figure EV2. Effect of myristoylation site mutation on the membrane association and S‐acylation of BSK1.

1. The effect of myristoylation site mutation (G2A) on the localization of BSK1 in a cell fractionation assay. The wild‐type (WT) and mutant versions of BSK1‐GFP were expressed in protoplasts, respectively. The total proteins (input) were divided into soluble (S) and pellet (P) fractions via ultra‐centrifugation. The representative anti‐GFP immunoblots from three biologically independent experiments are shown.
2. The effect of myristoylation site mutation (G2A) on the S‐acylation of BSK1. The WT and G2A mutant versions of BSK1‐GFP were expressed, respectively, in protoplasts for biotin‐switch assay. The anti‐GFP immunoblotting image is representative of three biologically independent experiments.

Source data are available online for this figure.

S‐acylation is essential for BSK function but SA suppresses their function by decreasing their S‐acylation

BRASSINOSTEROID SIGNALING KINASEs interact with BRI1 at the plasma membrane (Tang et al, 2008), and S‐acylation may contribute to this association. Both bimolecular fluorescence complementation (BiFC) and co‐immunoprecipitation (co‐IP) assays indicated that the wild‐type version but not the C3S variant of BSK1 interacted with BRI1 at the plasma membrane (Fig 3A and B, Appendix Fig S3). As BSKs transduce BR signaling from BRI1 to downstream components (Tang et al, 2008), the loss of S‐acylation may disturb the roles of BSKs in BR‐mediated plant development, a possibility that we explored.

Figure 3. S‐acylation is essential for the physiological function of BSKs.

• A
  BiFC data for detecting the effect of BSK1 S‐acylation on its interaction with BRI1. The indicated plasmid pairs were transformed into protoplasts for BiFC; BAK1 which colocalizes with BRI1 was used in the negative controls; the combinations with empty vectors are included in Appendix Fig S3. YN or YC, the N or C terminal fragment of YFP. Representative YFP (yellow) and merged (YFP in yellow, chloroplast auto‐fluorescence in red, and bright field in gray) signals are shown. Bars, 5 μm.
• B
  Co‐IP result for measuring the function of BSK1 S‐acylation in its interaction with BRI1. Free GFP (negative control) or the WT/C3S version of BSK1‐GFP was co‐expressed with BRI1‐MYC in plant cells for co‐IP using an anti‐GFP resin. The proteins in the input and IP samples were detected via immunoblots with anti‐MYC or anti‐GFP antibodies.
• C, D
  Contribution of S‐acylation to the function of BSK1 and BSK3 in the bri1‐5 mutant. The wild‐type (WT) or S‐acylation site mutant version of BSK1‐GFP (C) or BSK3‐GFP (D) was overexpressed under a 35S promoter in the bri1‐5 background. The developmental phenotypes of the 3‐week‐old (top) and 8‐week‐old (bottom) transgenic plants were recorded. 3M: the C5, 10, 11S version of BSK3. The bri1‐5 or Ws (wild‐type) plants were used as controls. Representative images from three biologically independent experiments are shown.
• E, F
  The effect of S‐acylation on the function of BSK3 in BR responses. The WT or 3M (C5, 10, 11S) version of BSK3 driven by a native promoter was introduced into the bsk3‐1 mutant to generate complementation lines. The phenotype of 8‐day‐old seedlings with and without 100 nM of BL treatment is shown in (E). Bars, 1 cm. Quantitative data of hypocotyl and root length under the control or BL treatment condition are shown in (F). Data are means ± SD from 20 seedlings in four independent experiments. Significance analysis using one‐way ANOVA followed by Tukey's multiple comparison test is shown in different lower‐case letters (P < 0.05).

Source data are available online for this figure.

Overexpression of BSK1 or BSK3 suppresses the dwarf phenotype of bri1‐5 mutant plants (Tang et al, 2008), providing a suitable system to evaluate the effect of S‐acylation on BSK functions. We separately overexpressed the intact and S‐acylation site mutant variants of BSK1 and BSK3 in the bri1‐5 background and selected transgenic lines with similar expression levels for phenotypic analysis (Appendix Fig S4A and B). Overexpression of intact BSK1 or BSK3 partially suppressed the dwarf phenotype of bri1‐5 during development, whereas the variants had no effect (Fig 3C and D, Appendix Fig S4C and D). We also performed a complementation assay of the bsk3‐1 mutant, which exhibits BR insensitivity (Tang et al, 2008) by introducing a genomic construct for the wild‐type or S‐acylation site mutant variants of BSK3 under the control of the native promoter (Appendix Fig S5). The wild‐type transgene, but not that producing S‐acylation‐defective BSK3, rescued the hypocotyl and root defects of the bsk3‐1 mutant in response to brassinolide (BL, the most active BR; Fig 3E and F). This phenotypic characterization supports the notion that S‐acylation is critical for the function of BSKs during plant development.

We investigated the effects of treatment with different plant hormones or stress conditions on BSK localization, as reversible S‐acylation may dynamically regulate the S‐acylation level of BSKs to modulate BR signaling under specific conditions. Most treatments did not alter BSK localization, but SA induced a fast change in the localization of BSK1‐GFP in transfected protoplasts and intact root cells from stable transgenic lines (Fig 4A and B). A biotin‐switch assay confirmed that S‐acylation of BSK1 reduced quickly during SA treatment (Fig 4C). Overexpression of wild‐type BSK1 in bri1‐5 enhanced leaf expansion and resulted in higher fresh weight compared with either bri1‐5 seedlings or even the wild type which displayed better leaf expansion (Fig 4D and E). When we treated the same genotypes with SA, we observed a marked suppression of the effect of BSK1 overexpression on bri1‐5 development, resulting in similar phenotypes for bri1‐5 seedlings regardless of BSK1 overexpression. Interestingly, although the wild‐type seedlings displayed larger leaf expansion and root elongation in SA treatment, their fresh weights were not significantly different from those of the seedlings in bri1‐5 background (Fig 4D and E). In agreement, we determined that SA diminished the interaction between BSK1 and BRI1, based on a co‐IP assay (Fig EV3A). Therefore, SA may attenuate the function of BSKs by decreasing their S‐acylation.

Figure 4. SA reduces the S‐acylation of BSK1 and disrupts its localization and function.

• A, B
  Effect of SA on subcellular localization of BSK1‐GFP. (A) BSK1‐GFP was expressed in protoplasts and 100 μM of SA was applied in the treatment. GFP (green) and merged (with bright field) signals were recorded at 0, 20, 40, and 60 min during the SA treatment. Bars, 5 μm. (B) Six‐day‐old 35S:BSK1‐GFP seedlings were treated with or without 100 μM of SA for 15 min. Representative GFP and merged (with bright field) signals are shown. Bars, 10 μm.
• C
  Detection of the regulation of BSK1 S‐acylation mediated by SA. BSK1‐GFP was expressed in protoplasts with 0‐, 20‐, 40‐, or 60‐min treatment with 100 μM of SA. Cells were collected for S‐acylation detection in a biotin‐switch assay. Representative immunoblots are shown in the top graph. Immunoblot signals were quantified by ImageJ and the S‐acylation levels were calculated from relative signals ([pulldown+/input+] − [pulldown−/input–]). The relative S‐acylation level of the 0 min sample was set to 1. The data in the bottom graph are means ± SD from three biologically independent experiments. Significance analysis using one‐way ANOVA followed by Tukey's multiple comparison test is shown in different lower‐case letters (P < 0.05).
• D, E
  Effect of SA on the physiological function of BSK1. The indicated seeds were germinated on the MS medium with or without 40 μM of SA. The photographs were taken 8 days after germination. Representative images are shown in (D). Bars, 1 cm. Quantitative data of fresh weight (per 10 seedlings) of different genotypes are shown in (E). Data are means ± SD from three biologically independent experiments. Significance analysis using one‐way ANOVA followed by Tukey's multiple comparison test is shown in different lower‐case letters (P < 0.05).

Source data are available online for this figure.

Figure EV3. Co‐IP for the ABAPT11‐BSK1 interaction and the effect of SA on the BRI1‐BSK1 interaction.

1. The co‐IP data for detecting the effect of SA on the BRI1‐BSK1 interaction. GFP (negative control) or BSK1‐GFP was co‐expressed with BRI1‐MYC ₃ in protoplasts, the cells were treated with or without 100 μM SA for 60 min before collection for co‐IP using an anti‐GFP resin. The proteins in the input and IP samples were detected using immunoblotting with anti‐MYC or anti‐GFP antibodies.
2. The co‐IP result for detecting the interaction between ABAPT11 and BSK1. GFP (negative control) or BSK1‐GFP was co‐expressed with ABAPT11‐MYC ₆ in protoplasts for co‐IP using an anti‐GFP resin. The proteins in the input and IP samples were measured in immunoblots with ant‐MYC or anti‐GFP antibodies.

Source data are available online for this figure.

ABAPT11 is the de‐S‐acylation enzyme of BSKs that dynamically regulates their subcellular localization

The removal of S‐acyl groups from protein substrates is catalyzed by de‐S‐acylation enzymes (Won et al, 2018), which may contribute to the fast decrease of BSK S‐acylation induced by SA. We recently identified a group of protein de‐S‐acylation enzymes (ABAPTs) comprising 11 members in Arabidopsis (Liu et al, 2021). We systemically screened all ABAPTs for their ability to alter BSK1 localization, used as a proxy for its S‐acylation status. We co‐transfected Arabidopsis protoplasts with BSK1‐RFP (encoding a fusion between BSK1 and the red fluorescent protein) with GFP or constructs encoding a GFP fusion for each of the 11 ABAPT members. As a result, only ABAPT11 had an apparent effect on the plasma membrane localization of BSK1‐RFP (Fig 5A; the effect of ABAPT11‐RFP on BSK1‐GFP localization is shown in Appendix Fig S6A and B). We also co‐transfected ABAPT11‐RFP or RFP with BSK‐GFP constructs for each of the 12 family members in protoplasts. BSK4‐GFP was localized in the membrane periphery and the cytosol in most of the cells with RFP or ABAPT11‐RFP; compared with the RFP control, ABAPT11‐RFP overexpression significantly increased the mislocalization of the GFP‐fused BSK1, 2, 3, 5, 6, 8, 9, 11, and 12 (Fig 5B), providing evidence for the localization alterations of most BSKs upon overexpression of ABAPT11. These data support the idea that ABAPT11 de‐S‐acylates BSK members.

Figure 5. ABAPT11 was identified as the de‐S‐acylation enzyme of BSKs in a systemic screening.

1. Identification of the de‐S‐acylation enzyme of BSK1. BSK1‐RFP was co‐overexpressed with 11 GFP‐fused ABAPT members in protoplasts, respectively (free GFP was used as a control). Cells with both RFP and GFP signals were used for analyzing the localization of BSK1‐RFP. The predominant localization patterns of BSK1‐RFP are shown. The RFP (magenta), GFP (green), and the merged (with bright field in gray) signals are included in the top graph. Bars, 5 μm. The quantitative data of the percentages of cells with mislocalization of BSK1‐GFP (100 cells for each sample) are shown in the bottom graph. Data are means ± SD from three biologically independent experiments. Significance analysis using one‐way ANOVA followed by Tukey's multiple comparison test is shown in different lower‐case letters (P < 0.05).
2. Detection of the effect of ABAPT11 on the localization of BSK members. Twelve BSK‐GFP fusion proteins were co‐overexpressed with RFP (control) or ABAPT11‐RFP in protoplasts. BSK‐GFP localization in cells with both RFP and GFP signals was recorded. Representative localization patterns of BSK‐GFP with RFP or ABAPT11‐RFP are shown in the top graph. Bars, 5 μm. Quantitative data of the percentages of cells with localization changes of BSK‐GFP (100 cells for each sample) are shown in the bottom graph. Data are means ± SD from three biologically independent experiments. ****P < 0.0001, ***P < 0.001, *P < 0.05, n.s. (no significance) P > 0.05, Student's t‐test (unpaired, two‐tailed).

Source data are available online for this figure.

To confirm the effect of ABAPT11 on the subcellular localization of BSKs in intact plant cells, we overexpressed ABAPT11 in BSK1‐GFP transgenic lines. We detected the partial mislocalization of BSK1‐GFP in root cells upon overexpression of ABAPT11 (Fig 6A). Consistently, cell fractionation assays indicated that ABAPT11 enhanced the accumulation of BSK1 in the cytoplasm (Fig 6B). We determined that ABAPT11 weakly interacted with BSK1 in plant cells by a co‐IP assay (Fig EV3B), possibly reflecting the transient association of the enzyme with its substrate. Furthermore, a biotin‐switch assay confirmed that overexpression of ABAPT11 significantly lowered the S‐acylation levels of BSK1 and BSK3 (Fig 6C and D), providing direct biochemical evidence that ABAPT11 is a de‐S‐acylation enzyme for BSKs.

Figure 6. ABAPT11 reduces the S‐acylation and membrane localization of BSKs.

• A
  Subcellular localization of BSK1‐GFP in intact root cells of ABAPT11 overexpression lines. Six‐day‐old transgenic seedlings were used for localization analysis of BSK1‐GFP in the root cells via confocal microscopy. Representative GFP (green) and merged (with bright field in gray) signals from three biologically independent experiments are shown. Bars, 10 μm.
• B
  Effect of ABAPT11 overexpression on the BSK1 membrane association in a cell fractionation assay. BSK1‐GFP was expressed in protoplasts with or without ABAPT11 overexpression. Total proteins (T) were separated into soluble (S) and pellet (P) fractions using ultra‐centrifugation. A representative immunoblot using an anti‐GFP antibody is shown. The percentages of BSK1‐GFP in the soluble fraction were calculated from relative immunoblotting signals (S/[S + P]). Data are means ± SD from three biologically independent experiments. **P < 0.01, Student's t‐test (unpaired, two‐tailed).
• C, D
  Effect of ABAPT11 overexpression on S‐acylation of BSK1 and BSK3. The S‐acylation level of BSK1‐GFP (C) and BSK3‐GFP (D) with or without ABAPT11 overexpression in protoplasts was detected in a biotin‐switch assay. Representative immunoblotting image is shown in the top graphs. Quantification of relative S‐acylation levels was included in the bottom graphs. Immunoblotting signals were quantified by ImageJ, and the S‐acylation levels were calculated from relative signals ([pulldown+/input+] − [pulldown−/ input−]). The relative S‐acylation level of the control sample was set to 1. Data are means ± SD from three biologically independent experiments. **P < 0.01, Student's t‐test (unpaired, two‐tailed).

Source data are available online for this figure.

SA induces ABAPT11 expression to suppress BR‐mediated growth via de‐S‐acylation of BSKs

As the S‐acylation of BSKs was suppressed by both SA and ABAPT11, we investigated the association between the de‐S‐acylase and the plant defense hormone. We conducted quantitative RT–PCR to measure ABAPT11 transcript levels in seedlings treated with SA. During SA treatment, ABAPT11 transcript levels increased twofold as early as 10 min into treatment, before gradually returning to normal after 40 min (Fig 7A). Similarly, treating seedlings with the SA analog benzothiadiazole (BTH) also increased ABAPT11 transcript levels (Fig EV4A). In a mutant for NPR1 (NONEXPRESSER OF PR GENES 1), encoding a key component of SA signaling, ABAPT11 expression failed to be induced by SA (Fig 7B). Consistently, ABAPT11 transcript levels were lower in the SA biosynthesis mutant sid2 (salicylic acid induction deficient 2) compared with the wild type (Fig EV4B). To assess whether ABAPT11 protein abundance changed upon SA treatment, we generated transgenic plants expressing ABAPT11‐YFP (encoding a fusion between ABAPT11 and the yellow fluorescent protein) under its native promoter. An immunoblot analysis with an anti‐GFP antibody indicated that SA led to the rapid accumulation of ABAPT11‐YFP (Fig 7C). The further confocal microscopy analysis supported the increase of ABAPT11‐YFP protein levels in root and leaf cells under SA treatment (Fig EV4C and D). We conclude that SA rapidly induces the expression of ABAPT11 and the accumulation of ABAPT11 to erase the S‐acylation of BSKs.

Figure 7. SA induces the expression of ABAPT11 to decrease the membrane association of BSK1.

• A, B
  Expression of ABAPT11 during SA treatment. (A) Seven‐day‐old wild‐type (WT) seedlings were treated with 100 μM SA for 0, 10, 20, 30, 40, 50, and 60 min before collection for RNA preparation and RT–qPCR to detect the transcript level of ABAPT11. (B) Transcript levels of ABAPT11 in wild‐type (WT) and npr1 mutant 7‐day‐old seedlings with or without 100 μM SA for 20 min. ACTIN2 was used as an internal control. Data are means ± SD from triplicated replications in an experiment; the pattern was consistent in three biologically independent experiments. Significance analysis using one‐way ANOVA followed by Tukey's multiple comparison test is shown in different lower‐case letters (P < 0.05).
• C
  Accumulation of ABAPT11 proteins in the SA response. Seven‐day‐old seedlings expressing ABAPT11‐YFP under its own promoter were treated with 100 μM SA for 0, 20, 40, and 60 min before collection for immunoblotting with an anti‐GFP antibody. Coomassie blue staining of total proteins was used as the loading control. Quantitative relative protein levels (YFP/loading) from three biologically independent experiments are shown in the bottom graph. Significance analysis using one‐way ANOVA followed by Tukey's multiple comparison test is shown in different lower‐case letters (P < 0.05).
• D
  The S‐acylation level of BSK1 in the ABAPT11 mutant during SA treatment. BSK1‐GFP was expressed in WT or abapt11‐1 mutant cells (with or without 100 μM SA treatment for 40 min) for S‐acylation measurement in a biotin‐switch assay. Representative immunoblots are shown in the top graph. The immunoblotting signals were quantified by ImageJ, and the S‐acylation levels were calculated from relative signals ([pulldown+/input+] – [pulldown−/ input–]). The relative S‐acylation level of the wild‐type sample without SA treatment was set to 1. Data are means ± SD from three biologically independent experiments. Significance analysis using one‐way ANOVA followed by Tukey's multiple comparison test is shown in different lower‐case letters (P < 0.05).
• E, F
  The function of ABAPT11 in the regulation of BSK1 localization mediated by SA. (E) The GFP signals were measured in intact root cells of 6‐day‐old 35S:BSK1‐GFP transgenic seedlings in the WT or abapt11‐1 mutant background, with or without 100 μM SA treatment for 15 min. The GFP (green) signals are representative of three biological independent experiments. Bars, 10 μm. (F) BSK1‐GFP was expressed in WT or abapt11‐1 mutant protoplasts, with or without 100 μM SA treatment for 40 min. The representative GFP (green) and merged (with the bright field in gray) signals from three biologically independent experiments are shown. Bars, 5 μm.
• G, H
  Confirmation of the SA‐induced BSK1 translocation which is mediated by ABAPT11. BSK1‐GFP was expressed in WT or abapt11‐1 mutant cells, with or without 100 μM SA treatment for 40 min. Total proteins (T) were separated into soluble (S) and pellet (P) fractions using ultra‐centrifugation. Representative immunoblots using an anti‐GFP antibody are shown in (G). The percentages of BSK1‐GFP in the soluble fraction in (H) were calculated from relative immunoblotting signals (S/[S + P]). Data are means ± SD from three biologically independent experiments. Significance analysis using one‐way ANOVA followed by Tukey's multiple comparison test is shown in different lower‐case letters (P < 0.05).

Source data are available online for this figure.

Figure EV4. Detection of ABAPT11 expression under SA treatment.

• A
  The transcript level of ABAPT11 during the BTH treatment. Seven‐day‐old wild‐type seedlings were treated with or without 100 μM BTH for 20 min. RNA was prepared for quantitative RT–PCR.
• B
  The transcript level of ABAPT11 in the sid2 mutant. RNA was extracted from 3‐week‐old plants for quantitative RT–PCR. ACTIN2 was used as an internal control. The data are means ± SD from triplicated replications in an experiment; the pattern was consistent in three biologically independent experiments. **P < 0.01, *P < 0.05, Student's t‐test (unpaired, two‐tailed).
• C, D
  The protein levels of ABAPT11‐YFP expressed under its native promoter in the SA response. The 7‐day‐old transgenic plants expressing ABAPT11‐YFP under its own promoter were treated with or without 50 μM SA (for roots) or 100 μM SA (for leaves) for 1 h before observation via confocal microscopy. The representative images of GFP signals in the root (C) and leaf (D) tissues from three biologically independent experiments are shown; bars, 20 μm. The same parameters of confocal microscopy were used between the control and SA samples in each experiment.

Source data are available online for this figure.

If de‐S‐acylation of BSKs induced by SA is mediated by ABAPT11, the SA‐induced de‐S‐acylation of BSKs may be blocked in the absence of ABAPT11. This possibility prompted us to measure BSK1 S‐acylation in wild‐type and ABAPT11 T‐DNA knockout mutants (Appendix Fig S7A–C). A biotin‐switch assay indicated that SA suppressed BSK1 S‐acylation in the wild type but not in the abapt11‐1 mutant (Fig 7D). To assess the effect of the loss of ABAPT11 on BSK1 localization, we introduced the 35S:BSK1‐GFP transgene into the abapt11‐1 mutant. We observed that SA induced a change in the localization of BSK1‐GFP in root cells, leaf cells, and protoplasts from the wild type but not in the abapt11‐1 mutant (Fig 7E and F, Appendix Fig S8A–C). Consistently, cell fractionation assays indicated that the SA‐induced localization change in BSK1‐GFP required ABAPT11 (Fig 7G and H). These data provide evidence that suppression of BSK S‐acylation by SA is dependent on ABAPT11.

Because SA enhanced the ABAPT11‐mediated de‐S‐acylation of BSKs, loss of ABAPT11 function may affect the cross talk between BR and SA. We noticed that a high concentration of SA (100 μM) severely suppressed the growth of wild‐type seedlings after germination, but this suppression was attenuated in the abapt11‐1 mutant (Fig EV5A and B). This difference suggested a potential mechanism by which SA induces ABAPT11 expression to inhibit BSK‐mediated development, although we cannot exclude the possibility that SA interacts with pathways other than BR signaling. Therefore, two independent ABAPT11 T‐DNA knockout mutants (Appendix Fig S7A–C) were used for phenotype analysis under combined treatments with different concentrations of BL and SA. Because a high concentration of BRs enhances hypocotyl elongation, we germinated wild‐type and abapt11 mutant seeds on a medium containing 80 nM BL. Compared with the control medium with no BL added, the addition of 80 nM BL induced hypocotyl elongation of all seedlings but to a greater extent for the abapt11 mutants than the wild type. Although 40 μM SA alone had no effect on hypocotyl length, 40 μM SA dramatically suppressed the hypocotyl elongation induced by 80 nM BL in the wild‐type but not in the abapt11 mutants (Fig 8A and B), providing direct evidence for the role of ABAPT11 in SA‐BR cross talk. Because a high concentration of BL or SA severely inhibits root growth, we used a low concentration of each phytohormone to test the function of ABAPT11 in the BR‐SA cross talk in root development, choosing phytohormone concentrations that do not elicit a distinct phenotype in abapt11 mutants to SA or BL treatment. We transferred 4‐day‐old seedlings to different media to measure root elongation; 5 days after transfer, root elongation of wild‐type seedlings was largely suppressed by 20 μM SA and slightly inhibited by 15 nM BL. Both abapt11 mutants displayed a similar root elongation as the wild‐type seedlings under control, BL‐containing, or SA‐containing media. However, the combination of 15 nM BL and 20 μM SA resulted in significantly reduced root elongation in the abapt11 mutants compared with the wild type (Fig 8C and D), suggesting that the attenuation of BR sensitivity by SA is blocked in the absence of ABAPT11. Taken together, these data for hypocotyl length and root elongation indicate that ABAPT11 play an important physiological role in the SA‐imposed regulation of BR‐mediated development.

Figure EV5. Developmental effect of a high concentration of SA on the ABAPT11 mutant.

• A, B
  The wild‐type (WT) and abapt11‐1 seeds were germinated on the MS medium with or without 100 μM SA. The photograph taken 5 days after germination is shown in (A). Bars, 1 cm. The quantitative data of fresh weight (per 10 seedlings) are shown in (B). The data are means ± SD from three biologically independent experiments. Significance analysis using one‐way ANOVA followed by Tukey's multiple comparison test is shown in different lower‐case letters (P < 0.05).

Source data are available online for this figure.

Figure 8. ABAPT11 plays a physiological function in the cross talk of SA and BRs.

• A, B
  Effect of BL and SA on hypocotyl development in ABAPT11 mutants. Seeds were germinated on regular medium (Control), medium containing 80 nM BL, medium containing 40 μM SA, or medium containing both 80 nM BL and 40 μM SA. Hypocotyl length was measured 8 days after germination. Representative images from four independent experiments are shown in (A). Quantitative data of hypocotyl lengths are shown in (B). Data are means ± SD from 20 seedlings in four independent experiments. Significance analysis using one‐way ANOVA followed by Tukey's multiple comparison test is shown in different lower‐case letters (P < 0.05).
• C, D
  Effect of BL and SA on root development of ABAPT11 mutants. Four days after germination, seedlings of WT, abapt11‐1, and abapt11‐2 were transferred onto a regular medium (Control), medium containing 15 nM BL, medium containing 20 μM SA, or medium containing both 15 nM BL and 20 μM SA. The phenotype of root development was recorded 5 days after transfer. Representative images from four independent experiments are shown in (C). White arrowheads indicate the positions of root tips when the seedlings were transferred and 5 days after transfer. Bars, 1 cm. Quantitative data of root elongation 5 days after transferring are shown in (D). Data are means ± SD from 20 seedlings in four independent experiments. Significance analysis using one‐way ANOVA followed by Tukey's multiple comparison test is shown in different lower‐case letters (P < 0.05).
• E
  The model that the SA‐induced ABAPT11 mediates de‐S‐acylation of BSKs to release them from the plasma membrane, resulting in the dissociation of BSKs and BRI1 and the attenuation of BR signaling transduction.

Source data are available online for this figure.

Discussion

Brassinosteroids are crucial plant hormones that regulate development and stress responses, with precise modulation of the signal transduction cascade at different nodes of the network (Nolan et al, 2020). BSKs are key components acting at the plasma membrane that transfer BR signals from the receptor BRI1 to downstream factors (Tang et al, 2008). Given that BR signaling is dynamically controlled by endogenous and exogenous factors (Lv & Li, 2020), a reversible post‐translational modification may be required for regulating BSKs. Here, we demonstrated how S‐acylation and de‐S‐acylation contribute to the accurate modulation of BSKs in response to SA.

Our data indicated that most Arabidopsis BSKs were S‐acylated, which contributed to their localization to the plasma membrane. We also established that ABAPT11 was a common de‐S‐acylation enzyme that controlled the subcellular localization of most S‐acylated BSKs. These data provide evidence for a global modulation mechanism of BSKs mediated by reversible S‐acylation. However, it remains unclear how BSKs are S‐acylated, which will make the identification of the responsible PATs that catalyze the S‐acylation of BSKs an important objective in the future. Given that BSK homologs are present in other plants (Zhang et al, 2016), it would be valuable to determine whether this regulation is conserved in other species. Although both BSK1 and BSK3 were S‐acylated in plant cells, we noticed that their motifs of S‐acylation were different. In addition, why BSK4 was not S‐acylated in plant cells may be dependent on its structure; these differences suggest that the precise regulation of this modification may vary among BSK members. Previous studies showed that myristoylation mediates the transient membrane affinity of protein substrates, and the additional S‐acylation provides a more durable membrane association for potential intracellular trafficking (Shahinian & Silvius, 1995). For instance, in mammalian cells, some dually myristoylated and S‐acylated proteins undergo trafficking between the plasma membrane and the Golgi apparatus via S‐acylation and de‐S‐acylation cycles (Rocks et al, 2005). Our data showed that the loss of myristoylation reduced the S‐acylation of BSK1; therefore, it will also be interesting to investigate the cross talk between these two types of lipid modification in the regulation of BSK localization. Overexpression of the wild‐type but not the S‐acylation site mutated variants of BSKs partially rescued the developmental defects of bri1‐5 plants, suggesting that S‐acylation is necessary for the function of BSKs in the BR signaling pathway. BiFC and Co‐IP assays further supported the notion that loss of BSK1 S‐acylation reduced its association with BRI1 at the plasma membrane. The release of BSKs without S‐acylation from the plasma membrane may not disrupt their interaction with downstream targets in the cytosol, but the activation of BSKs requires phosphorylation by BRI1 (Tang et al, 2008); thus, BSK de‐S‐acylation blocks their interaction with BRI1 and would interrupt BR signaling. Therefore, reversible BSK de‐S‐acylation constitutes a fast and powerful means to dynamically control BR signaling in response to plant hormone treatments and environmental stresses.

The current study uncovered how the S‐acylation status of BSKs was affected by SA for the modulation of BR signaling in plant development. The interplay of BRs with other plant hormones, such as auxins, gibberellins, and cytokinins, has been investigated in detail in plant cells (Peres et al, 2019). Two recent works showed that BRs increase SA‐mediated immunity gene expression via suppressing BIN2 phosphorylation of TGA transcription factors (Han et al, 2022; Kim et al, 2022). The data from these works supported the functional cross talks between the BRs and SA pathways. These studies focused on the effect of BRs on the SA signaling transduction, but how SA modulates BR signaling remained to be investigated. In our current study, we elucidated the mechanism by which SA dampens the BR pathway via the de‐S‐acylation of BSKs. Our data showed that SA treatment resulted in quick localization changes of BSKs via inducing the expression of ABAPT11, but the detailed mechanism behind this SA‐mediated increase of ABAPT11 expression needs to be investigated in further study. It will also be worth studying the tissue and cell type specificity of SA‐induced ABAPT11 expression and dissecting the dynamic S‐acylation patterns at a single‐cell level in the future. When treated with both BRs and SA, the abapt11 mutants showed attenuation of root elongation and greater hypocotyl elongation relative to the wild type, supporting a function for SA‐induced de‐S‐acylation mediated by ABAPT11 in the suppression of BR responses. Because SA is an important plant hormone associated with plant pathogenesis (Ding & Ding, 2020), its regulation of BSKs may connect these pathways in the trade‐off between plant development and immunity.

In summary, S‐acylation contributes to the maintenance of BSKs at the plasma membrane; SA induces the expression of ABAPT11 to reduce the S‐acylation level of BSKs, resulting in their release from the plasma membrane and decreasing their interaction with BRI1 to impair BR signal transduction (Fig 8E). Our current work on the attenuation of BR signaling via SA‐mediated de‐S‐acylation of BSKs will improve our understanding of the roles of post‐translational modifications in plant hormone cross talk.

Materials and Methods

Plant material and growth conditions

Arabidopsis thaliana seeds were surface sterilized and stratified on Murashige and Skoog (MS) medium with 1.5% (w/v) sucrose and 0.7% (w/v) agar for 2 days in the dark and grown under a long‐day condition with 16 h of light and 8 h of darkness at 22°C. Seedlings were moved to the soil about 1 week after germination. The npr1 and sid2 mutants were described previously (Spoel et al, 2007). The bri1‐5 (in the Ws background; Tang et al, 2008) mutant was obtained from Prof. Jia Li (Guangzhou University), and seeds for bsk3‐1 (SALK_096500, in the Col‐0 background), abapt11‐1 (SALK_058588, in the Col‐0 background) and abapt11‐2 (SALK_087022, in the Col‐0 background) were obtained from The Nottingham Arabidopsis Stock Centre.

Generation of constructs and transgenic plants

For transient expression of GFP‐fused proteins of BSK1‐12 members, the BSK genes without their stop codons were cloned into a pCAMBIA‐35S:GFP vector. The mutation sites of BSK1 and BSK3 were introduced on the primers. The wild‐type and mutant versions of 35S:BSK1‐GFP and 35S:BSK3‐GFP were introduced into Col‐0 plants for localization observation and into bri1‐5 for phenotypic analysis. For BSK3 complementation, the genomic region of BSK3 with a native promoter (2,000 bp) and a stop codon was cloned into pCAMBIA1300‐221. The mutation sites were introduced using overlap PCR. The wild‐type and mutant versions of ProBSK3:BSK3 were introduced into bsk3‐1 plants for phenotypic analysis. For native ABAPT11‐GFP expression, the genomic region of ABAPT11 with its own promoter (2,000 bp) without the stop codon was cloned into the pCAMBIA1300‐221‐YFP vector.

For systemic screening of the BSK1 de‐S‐acylation enzymes, BSK1 was cloned into the pBI221‐35S:RFP vector for expression of BSK1‐RFP. The generation of 11 ABAPT‐GFP constructs was described previously (Liu et al, 2021). For detecting the effect of ABAPT11 on BSKs, ABAPT11 was constructed into pCAMBIA1300‐UBQ:RFP. To detect the effect of ABAPT11 on the localization of BSK1‐GFP, ABAPT11 was cloned into pCanG‐MYC ₆ and introduced into pCAMBIA‐35S:BSK1‐GFP plants; pCAMBIA‐35S:BSK1‐GFP was transformed into abapt11‐1 plants. For measuring the effect of ABAPT11 on BSK1 S‐acylation, BSK1 and ABAPT11 were cloned into two cloning sites of pCAMBIA1302, in which BSK1 was fused to GFP and ABAPT11‐MYC was used to replace the Hygromycin (R) gene.

Arabidopsis transformation was performed using the Agrobacterium‐mediated floral dipping method (Clough & Bent, 1998). Multiple independent lines of transgenic plants were obtained, and the T3 homozygous offspring of at least two independent lines were used for detailed investigation.

Biotin‐switch assay

The biotin‐switch assay was performed following our previous description (Liu et al, 2021). The indicated plasmids were transformed into Arabidopsis protoplasts for the expression of GFP‐fused BSK proteins. Twenty‐four hours after transfection, the protoplasts were collected for protein extraction. Cells were lysed in lysis buffer (100 mM HEPES pH 7.5, 2 mM TCEP, 1 mM EDTA, 0.1% [w/v] SDS, and 1× protease inhibitor cocktail) and incubated at 50°C for 5 min, and an equal volume of blocking buffer (100 mM HEPES pH 7.5, 1 mM EDTA, 5% [w/v] SDS, and 0.4% [v/v] MMTS) was added and the mixture was incubated at 40°C for 10 min. After being precipitated by mixing with three volumes of acetone overnight at −20°C, the samples were centrifuged at 5,000 g for 10 min. The pellets were rinsed with 70% (v/v) acetone and resuspended in 200 μl resuspension buffer (1 × PBS pH 7.4, 8 M urea, 2% [w/v] SDS). The suspension was divided into two tubes (100 μl each); in each tube, the sample was incubated for 1 h with 2 μl of 100 mM EDTA, 1 μl of 100× protease inhibitor cocktail, and 50 μl of 4 mM biotin‐HPDP, supplemented with 50 μl of 2 M NH₂OH (pH 7.4) or Tris–HCl (pH 7.4). Then, proteins in each tube were collected via methanol‐chloroform precipitation and dissolved in 100 μl resuspension buffer; 20 μl was saved as the input control, and the rest sample was mixed with 720 μl of PBS containing 0.2% (v/v) Triton X‐100 and incubated with Streptavidin‐Agarose beads for 1.5 h. After binding, the beads were rinsed sequentially with wash buffer (500 mM NaCl, 0.1% [w/v] SDS, 1× PBS pH 7.4) and 1× PBS (pH 7.4). Finally, 70 μl of wash buffer containing 5% (v/v) β‐mercaptoethanol for 20 min to elute S‐acylated proteins. The eluates were subjected to regular SDS–PAGE, and an anti‐GFP antibody (TransGen Biotech, HT801‐01) was used in immunoblotting.

Confocal microscopy

For detection of subcellular localization in protoplasts, the indicated plasmid or plasmid pair was transformed into Arabidopsis protoplasts generated from 3‐week‐old rosette leaves (Yoo et al, 2007) and signals were measured 24 h after transfection. To detect the subcellular localization in root cells of transgenic lines, the primary roots of 6‐day‐old seedlings were used. Analyses were performed under a Zeiss LSM 800 laser‐scanning confocal microscope.

Cell fractionation assay

The cell fractionation assays were performed as described previously (Liu et al, 2021). Proteins were extracted in homogenization buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 13% sucrose, with a protease inhibitor cocktail) and incubated at 4°C for 30 min. After centrifugation twice at 6,000 g for 10 min at 4°C, the supernatants were centrifugated at 80,000 g for 1 h at 4°C to separate soluble and pellet fractions. The pellet was resuspended in the same volume of homogenization buffer. The soluble and pellet fractions were used for SDS–PAGE and immunoblotting.

BiFC

The wild‐type or C3S version of BSK1 was cloned into the pSPYNE‐35S vector, while BRI1 or BAK1 was cloned into the pSPYCE‐35S vector. The constructs were co‐transformed into protoplasts; 48 h after transfection, YFP fluorescence was measured under a Zeiss LSM 800 laser‐scanning confocal microscope.

Co‐immunoprecipitation

BRI1 was cloned into a pBluescript‐based 35S:MYC ₃ vector for transient overexpression of BRI1‐MYC ₃ under the control of a 35S promoter. Then, 35S:BRI1‐MYC ₃ was co‐transformed with 35S:BSK1(WT)‐GFP, 35S:BSK1(C3S)‐GFP, or 35S:GFP in Arabidopsis protoplasts. The cells were collected 24 h after transfection, and proteins were extracted in 200 μl of lysis buffer (20 mM Tris–HCl pH 7.4, 100 mM NaCl, 10% [v/v] glycerol, 1% [v/v] NP40, and 1× protease inhibitor cocktail) and incubated at 4°C for 30 min. After centrifugation at 13,000 g for 15 min, the supernatant was mixed with 800 μl of dilution buffer (20 mM Tris–HCl pH 7.4, 100 mM NaCl, 10% [v/v] glycerol, and 0.5× protease inhibitor cocktail). Ninety microliter of this mixture was saved as an input control, and the rest was incubated with Anti‐GFP Nanobody Agarose Beads (AlpaLife, KTSM1301) at 4°C for 2 h, and then the beads were rinsed using wash buffer (20 mM Tris–HCl pH 7.4, 100 mM NaCl, 10% [v/v] glycerol, and 0.1% [v/v] NP40). Finally, 80 μl of 2× protein sample buffer was mixed with the beads and boiled at 95°C for 5 min, and proteins were subjected to SDS–PAGE and immunoblotting using anti‐GFP (TransGen Biotech, HT801‐01) and anti‐MYC (TransGen Biotech, HT101‐01) antibodies.

Expression analysis

Total RNA was extracted using a HiPure Plant RNA Mini Kit (Magen). Reverse transcription was performed using HiScript® III 1^st Strand cDNA Synthesis Kit (Vazyme). Quantitative PCR was conducted in a Bio‐Rad CFX 96 system (C1000 Thermal Cycler).

Accession numbers

The accession numbers of genes in this study are as follows: BSK1 (At4g35230); BSK2 (At5g46570); BSK3 (At4g00710); BSK4 (At1g01740); BSK5 (At5g59010); BSK6 (At3g54030); BSK7 (At1g63500); BSK8 (At5g41260); BSK9 (At3g09240); BSK10 (At5g01060); BSK11 (At1g50990); BSK12 (At2g17090); BRI1 (At4g39400); ABAPT1 (At4g24760); ABAPT2 (At5g38220); ABAPT3 (At5g14390); ABAPT4 (At3g01690); ABAPT5 (At3g30380); ABAPT6 (At1g13610); ABAPT7 (At1g66900); ABAPT8 (At4g31020); ABAPT9 (At2g24320); ABAPT10 (At1g32190); ABAPT11 (At5g20520); and BAK1 (At4g33430).

Author contributions

Xiaoshi Liu: Investigation; writing – original draft. Zian Chen: Investigation. Liting Huang: Investigation. Youwei Ouyang: Investigation. Zhiying Wang: Investigation. Shuang Wu: Investigation. Weixian Ye: Investigation. Boya Yu: Investigation. Yihang Zhang: Investigation. Chengwei Yang: Supervision. Jianbin Lai: Supervision; writing – original draft; writing – review and editing.

Disclosure and competing interests statement

The authors declare that they have no conflict of interest.

Supporting information

Appendix

Expanded View Figures PDF

Source Data for Expanded View and Appendix

Review Process File

PDF+

Source Data for Figure 1

Source Data for Figure 2

Source Data for Figure 3

Source Data for Figure 4

Source Data for Figure 5

Source Data for Figure 6

Source Data for Figure 7

Source Data for Figure 8

The EMBO Journal (2023) 42: e112998

Data availability

This study includes no data deposited in external repositories.