Dimerization-activated PP2C-D2 and D5 phosphatases negatively regulate brassinosteroid signaling by antagonizing BRI1

Significance

Reversible protein phosphorylation mediated by kinases and phosphatases is essential for plant growth and development. While plant phospho-signaling research has largely focused on kinases and phosphorylation, phosphatases and dephosphorylation are comparatively understudied. Here, we show that protein phosphatase PP2C-Ds are activated by dimerization. Unlike the monomeric function of most PP2C-type protein phosphatases, PP2C-D2 and D5 form homo- and heterodimers to directly dephosphorylate key residues in the BRI1 kinase activation loop, thereby inhibiting BR signaling. Upon BR perception, BRI1-mediated phosphorylation and SAUR15 binding return PP2C-Ds to a monomeric inactive state. Our findings establish PP2C-Ds as key repressors of BR signaling via a unrecognized regulatory mechanism.

Abstract

Brassinosteroids (BRs) are key hormones that promote plant growth and development. While reversible phosphorylation of the BR receptor BRASSINOSTEROID-INSENSITIVE 1 (BRI1) is critical for BR signaling, much remains to be learned about the dephosphorylation mechanisms of BRI1. Here, we demonstrate that the D-clade type 2C protein phosphatases (PP2C-Ds) negatively regulate BR signaling by dephosphorylating key residues in the kinase activation loop of BRI1. The phosphatase activity of PP2C-D2 and D5 is activated through homo- or heterodimerization, a process antagonized by BR-induced BRI1 phosphorylation or SMALL AUXIN UP RNA15 (SAUR15) binding. BL treatment, BRI1-mediated phosphorylation, or SAUR15 binding disrupts dimerization, returning PPC2C-Ds to a monomeric, inactive state. These data not only reveal another phosphorylation/dephosphorylation cascade regulating BR signaling but also decode unrecognized regulatory mechanisms of PP2C-Ds in plants.

Brassinosteroids (BRs) are steroid plant hormones with versatile roles in growth and development, immunity, and abiotic stress responses. Loss-of-function mutants in BR synthesis and signaling display severe defects, including dwarfism, dark-green leaves, and photomorphogenesis when grown in the dark (1, 2).

BR signaling is triggered when the hormone is perceived at the cell surface by the extracellular domain of BRASSINOSTEROID-INSENSITIVE 1 (BRI1) and its coreceptor, BRI1-ASSOCIATED RECEPTOR KINASE 1/SOMATIC EMBRYOGENESIS RECEPTOR KINASE 3 (BAK1/SERK3), which results in the autophosphorylation and transphosphorylation of BRI1 and BAK1 (3–5). The activated signaling inhibits glycogen synthase kinase 3 (GSK3)-like kinase BRASSINOSTEROID INSENSITIVE 2 (BIN2), leading to the nuclear accumulation of dephosphorylated BRASSINAZOLE-RESISTANT 1 (BZR1) and BRI1-EMS-SUPPRESSOR 1 (BES1) (6–8). Subsequently, activated BES1 and BZR1 regulate the expression of BR-responsive genes in association with other transcription factors (7). In addition to this signal transduction mechanism, the BR-triggered BRI1 modulates cell expansion by activating plasma membrane (PM) H⁺-ATPases (9, 10), and BRI1 activity can be further promoted by SMALL AUXIN UP RNA 15 (SAUR15), which is upregulated by both auxin and BR (10, 11). As a central regulator, the balance between BRI1 activation and deactivation precisely controls the initiation, timing, and amplitude of BR signaling.

The signaling activity of BRI1 is closely linked to the phosphorylation of serine (S), threonine (T), and tyrosine (Y) residues in its cytoplasmic domain (12–14). Phosphorylation of S1044, T1045, and T1049 in the activation loop is critical for BRI1 kinase activity and BR signal transduction. These sites can be phosphorylated by BRI1 itself or by BAK1 as well as other SERKs (5, 12, 14). We previously found that SAUR15 enhances BRI1 phosphorylation and kinase activity, although its exact role is unclear (10). BRI1 deactivation occurs through several mechanisms, including phosphorylation at S891 in the ATP binding domain, its C-terminal autoinhibitory domain, interaction with BRI1 KINASE INHIBITOR1 (BKI1), or endocytosis-resulted decrease in the PM pool of functional BRI1 (15–20). Additionally, BR-activated BRI1 could be attenuated by protein phosphatase 2A (PP2A)-dependent dephosphorylation at the C-terminal autoinhibitory region (21, 22). However, dephosphorylation mechanisms of the key sites in BRI1 kinase activation loop remain largely unresolved.

Protein phosphatase 2Cs (PP2Cs), like PP2As, belong to the type 2 protein phosphatase (PP2) family (23). While PP2As function as a holoenzyme, comprising a catalytic C subunit, a regulatory B subunit, and a scaffolding A subunit, PP2Cs form a distinct subfamily in which the catalytic and regulatory domains are on a single polypeptide (24). The Arabidopsis PP2C subfamily contains twelve clades, PP2C-A to PP2C-L, plus six genes that remain unclustered (24). Several members dephosphorylate key components of signal transduction cascades, as well as ion pumps and transporters. PP2C-As negatively regulate ABA signaling by dephosphorylating SNF1-RELATED PROTEIN KINASE2s (SnRK2s), a process inhibited by ABA-binding PYRABACTIN RESISTANCE1 (PYR1)/PYR1-LIKE/REGULATORY COMPONENTS OF ABA RECEPTORS (PYR/PYL/RCAR) (25). PP2C-Bs affect plant growth and abiotic stress tolerance by dephosphorylating mitogen-activated protein kinases (MAPKs) and CBL-interacting kinases (CIPKs) (26, 27). PP2C-Cs suppress pattern-triggered immunity and CLAVATA3/EMBRYO SURROUNDING REGION-RELATED (CLE) peptide signaling by inhibiting corresponding empty receptors. Upon ligand binding, the cytosolic kinases activated by receptors phosphorylate PP2C-Cs, causing them to dissociate from these receptors (28). PP2C-Ds play other assorted roles in plants. PP2C-D3/PP2C38 dephosphorylates BOTRYTIS-INDUCED KINASE1 (BIK1) to suppress innate immune signaling, a process counteracted by BIK1- and MITOGEN-ACTIVATED PROTEIN KINASE KINASE KINASE KINASE 4 (MAP4K4)-mediated phosphorylation (29, 30). PP2C-D6 and D7 inhibit the Na⁺/H⁺ antiporter activity of SALT OVERLY SENSITIVE 1 (SOS1) to maintain Na⁺ homeostasis under the control of SOS3-LIKE CALCIUM-BINDING PROTEIN8 (SCaBP8) (31). We previously showed that PP2C-D2 and D5 inhibit cell expansion in developing organs by dephosphorylating PM H⁺-ATPases, a process negatively regulated by SAUR15 (11). The mechanism by which SAUR15 inhibits PP2C-Ds remains unclear. Our previous work also showed that BR stimulates SAUR15 interaction with BRI1 leading to the phosphorylation and activation of PM H⁺-ATPases (10). These data suggest that SAUR15, BRI1, and PP2C-Ds may coregulate BR signal transduction.

Here, we show that PP2C-D2 and D5 interact with and regulate BR signaling by dephosphorylating S/T sites in the activation loop of BRI1. D2 and D5 form homo- and heterodimers, which are disrupted by BRI1-catalyzed phosphorylation and SAUR15 interaction, shifting them to monomeric state and inhibiting phosphatase activity. Our findings identify a another phosphorylation/dephosphorylation cascade governing BR signaling and shed a light on PP2C-D regulation mode.

Results

PP2C-Ds Are Negative Regulators of BR Signaling.

BR is the key hormone promoting hypocotyl cell elongation (1). Ren et al. (32) and our previous studies (11) demonstrate that PP2C-D2 (D2), PP2C-D5 (D5), and PP2C-D6 (D6) are the primary PP2C-D phosphatases that negatively regulate cell expansion during hypocotyl growth. To investigate their functions, we analyzed the phenotype of loss- and gain-of-function D2, D5, and D6 related lines. Since single PP2C-D mutants showed only mild defects (SI Appendix, Fig. S1 A–C; 32), double (d2 d5, d2 d6, d5 d6) and triple (d2 d5 d6) mutants were used. Transgenic lines in which D2 (pD2::D2 L1 and L2), D5 (pD5::D5 L1 and L2), and D6 (pD6::D6 L1 and L2) are expressed under their native promoters in Col-0 were generated (SI Appendix, Fig. S1B). Hypocotyl and primary root growth were promoted in d2 d5, d2 d6, d5 d6, and d2 d5 d6 to a similar degree as pBRI1::BRI1 seedlings (Fig. 1 A–D and SI Appendix, Fig. S1 A and B). Elevating PP2C-Ds expression under their native promoters inhibited both hypocotyl and primary root growth, resembling the weak bri1-301 allele of BRI1 (Fig. 1 A–D; 33).

Fig. 1.

PP2C-D phosphatases inhibit BR signaling. (A) Seedlings phenotypes of PP2C-D-related lines grown on ½ MS medium. Upper panel, 10-d-old seedlings grown in light. Lower panel, 4-d-old seedlings grown in dark. (Scale bar, 1 cm.) (B–D) Primary root length (B) and hypocotyl length of light-grown (C) and dark-grown (D) seedlings shown in (A). Boxplots show first and third quartiles split by median lines, with whiskers showing the minimum and maximum values. Different letters indicate significant differences (n = 35 per group; one-way ANOVA with Tamhane’s T2 test). The P values for each line compared with Col-0 are all less than 0.0001. (E) Primary root length of seedlings grown on ½ MS medium containing different BL concentrations in light. (F) Relative root length of seedlings in (E) calculated as the ratio of BL-treated to untreated (0 nM BL) seedlings. (G) Hypocotyl length of seedlings grown on ½ MS medium containing different BRZ concentrations in dark. (H) Relative hypocotyl length of seedlings in (G) calculated as the ratio of BRZ-treated to untreated (0 nM BRZ) seedlings. For (E and G), values are means ± SD (n = 35 per group). (I) BZR1 phosphorylation in 10-d-old seedlings grown on ½ MS medium with 1 μM BRZ in light. BZR1 was detected by immunoblotting with an α-BZR1 antibody. The integrated density of phosphorylated BZR1 bands (BZR1^+P) and unphosphorylated BZR1 bands (BZR1^−P) was quantified using ImageJ to calculate the BZR1^+P/BZR1^−P ratio. (J–L) Relative transcript levels of CPD (J), DWF4 (K), and SAUR15 (L) in 10-d-old seedlings grown on ½ MS medium with 1 μM BRZ in light. Transcript levels were analyzed by RT-qPCR using ACT2 as the reference gene and normalized to Col-0. Data show mean ± SD of three technical replicates (5 seedlings each). Statistical significance between Col-0 and other lines was determined by the independent samples t test (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). For (J), the P values for bri1-301, pBRI1::BRI1, d2 d5, d2 d5 d6, and pD6::D6 L2 are less than 0.0001, and for other lines compared with Col-0 are as follows: d2 d6, P = 0.002; d5 d6, P = 0.005; pD2::D2 L1, P = 0.004; pD2::D2 L2, P = 0.003; pD5::D5 L1 P = 0.004; pD5::D5 L2 P = 0.003; and pD6::D6 L1 P = 0.025. For (K), the P values for bri1-301, pBRI1::BRI1, d2 d5 d6, pD2::D2 L2, pD5::D5 L1 and L2, and pD6::D6 L2 are less than 0.0001, and for other lines compared with Col-0 are as follows: d2 d5, P = 0.003; d2 d6, P = 0.009; d5 d6, P = 0.016; pD2::D2 L1, P = 0.008; and pD6::D6 L1, P = 0.021. For (L), the P values for bri1-301, pBRI1::BRI1, d2 d5, d2 d6, d2 d5 d6, pD2::D2 L2, pD5::D5 L1 and L2 are less than 0.0001, and for other lines compared with Col-0 are as follows: d5 d6, P = 0.009; pD2::D2 L1, P = 0.002; pD6::D6 L1, P = 0.001; and pD6::D6 L2, P = 0.002. Experiments were repeated three times with similar results.

Here, expression of D2, D5, and D6 from the respective native promoters complemented the d2 d5 or d2 d6 phenotype, confirming these defects result from loss of PP2C-D activity (SI Appendix, Fig. S1 D−I). As previously reported (34), pBRI1::BRI1 rescued the null mutant bri1-701 phenotype (SI Appendix, Fig. S1 D−I). To confirm that these transgenic lines expressed PP2C-D or BRI1 at endogenous levels, we carried out semiquantitative PCR. We found similar transcript levels of D2 in D2-com, D5 in D5-com, D6 in D6-com, and BRI1 in BRI1-com as in wild type plants (SI Appendix, Fig. S1E).

To determine whether these phenotypes were linked to BR signaling activation or disruption, we conducted BR sensitivity tests. Root measurements showed that d2 d5, d2 d6, d5 d6, and d2 d5 d6 were highly sensitive to 2,4-epibrassinolide (BL, a bioactive BR), whereas PP2C-D transgenic seedlings were less responsive compared to Col-0 (Fig. 1 E and F). Hypocotyl elongation of dark-grown etiolated seedlings treated with brassinazole (BRZ, a specific inhibitor of BR biosynthesis) was assessed. Compared to Col-0, double and triple mutants of PP2C-Ds as well as pBRI1::BRI1 plants showed decreased sensitivity to BRZ, whereas PP2C-D transgenic lines as well as bri1-301 were hypersensitive (Fig. 1 G and H).

BZR1 phosphorylation was assessed by immunoblotting with an α-BZR1 antibody in PP2C-D related lines. Compared to Col-0, the ratio of phosphorylated BZR1 (BZR1^+P, upper bands) versus unphosphorylated BZR1 (BZR1^−P, lower bands) was higher in bri1-301 and PP2C-D transgenic lines whereas it was lower in pBRI1::BRI1 and PP2C-D mutants (Fig. 1I). To further confirm BR signaling alterations, RT-qPCR was used to measure transcript levels of the BR biosynthesis genes, CONSTITUTIVE PHOTOMORPHOGENIC DWARF (CPD) and DWARF4 (DWF4), which are feedback-repressed by BR signaling, and SAUR15, a BR-upregulated gene (10, 35, 36). Compared to Col-0, CPD and DWF4 transcripts were lower in PP2C-D mutants, resembling pBRI1::BRI1, whereas they were higher in bri1-301 and PP2C-D transgenic lines (Fig. 1 J and K). However, SAUR15 expression showed an opposite trend. Compared to Col-0, SAUR15 transcripts were lower in bri1-301 and PP2C-D transgenic lines but higher in pBRI1::BRI1 and PP2C-D mutants (Fig. 1L). In addition, we measured D2, D5, and D6 expression in Col-0, bri1-301, bri1-701, pBRI1::BRI1, and cpd (a strong allele of CPD, 37), with and without BL treatment, and found that transcriptional expression of these PP2C-Ds is independent of BR (SI Appendix, Fig. S1 J−L). Together, these data suggest that D2, D5, and D6 phosphatases negatively regulate BR signaling.

Phylogenetic analysis of PP2C-Ds in Arabidopsis indicated that D5 and D6 form part of the same subclade, whereas the relationship of D2 to other D clade members was unresolved (SI Appendix, Fig. S1M; 24, 29). Moreover, d5 was the only single mutant with a hypocotyl elongation phenotype (SI Appendix, Fig. S1C; 24, 29). Therefore, we chose to focus on D2 and D5 for further functional and biochemical properties analysis.

PP2C-D2, D5 Interact With BRI1 to Regulate BR-Mediated Growth.

Our previous studies showed that SAUR15 promotes auxin- and BR-mediated growth by inhibiting D2, D5, or activating BRI1 (10, 11). Since SAUR15 interacts with D2, D5, and BRI1 on the PM (10, 11), we tested whether BRI1 also interacts with D2 and D5. Yeast two-hybrid (Y2H) assays and in vitro pull-down assays showed that both phosphatases directly interact with the BRI1 cytoplasmic domain (BRI1_CD) (Fig. 2 A–C). Bimolecular Fluorescence Complementation (BiFC) assays further confirmed BRI1 association with D2 and D5 on the PM in Nicotiana benthamiana leaf epidermal cells (Fig. 2D). Coimmunoprecipitation (Co-IP) assays in seedlings coexpressing BRI1-FLAG with D2-HA or D5-HA fusion protein showed that BRI1-FLAG pulled down both fusion proteins, which was unaffected by BL treatment (Fig. 2 E and F). These results indicate that BRI1 interacts with D2 and D5 in vitro and in vivo.

Fig. 2.

PP2C-D2, D5 interact with BRI1 and mediate BRI1’s dephosphorylation under the control of SAUR15. (A) Y2H assay showing D2 and D5 interact with the cytoplasmic domain of BRI1. PP2C-Ds, BAK1, and mTagBFP2 (BFP2) were each fused to the C-terminal half of ubiquitin (Cub). BFP2, Full-length BRI1 (BRI1_FL), its extracellular domain (BRI1_ED), and its cytoplasmic domain (BRI1_CD) were each fused to the N-terminal half of ubiquitin (Nub). BAK1-Cub with BRI1_FL-Nub (BAK1/BRI1_FL) served as a positive control (4). BFP2-Cub with BFP2-Nub (BFP2/BFP2) served as a negative control. (B and C) In vitro MBP pull-down assay of GST-D2-HA and GST-D5-HA with MBP-BRI1_CD. GST-D2-HA (B) and GST-D5-HA (C) fusion proteins were incubated with amylose resin bound to either MBP or MBP-BRI1_CD. Proteins were immunoblotted with α-MBP and α-HA antibodies, separately. IP, immunoprecipitate; IB, immunoblot. (D) BiFC assays showing D2 and D5 interaction with BRI1 in N. benthamiana leaf epidermal cells. BRI1 was fused to the C-terminal half of YFP (YC). HA and PP2C-Ds were fused to the N-terminal half of YFP (YN). Leaves coinfiltrated with GV3101 carrying BRI1-YC and HA-YN constructs served as the negative control. YFP, Bright field (Bright) and YFP/Bright merged (Merged) images are shown. (Scale bar, 10 μm.) (E and F) Co-IP assays of D2-HA and D5-HA with BRI1-FLAG. BRI1-FLAG was coexpressed with BFP2-HA, D2-HA (E) and D5-HA (F). 10-d-old seedlings grown on ½ MS medium were treated with or without 1 μM BL for 120 min. IP and Input proteins were analyzed by immunoblotting with α-FLAG and α-HA antibodies, separately. The integrated density of IP proteins bands was measured using ImageJ to calculate the D2-HA/BRI1-FLAG and D5-HA/BRI1-FLAG ratio. (G) Phenotypic analysis of different PP2C-D and BRI1-related lines grown on ½ MS medium. Upper panel, 10-d-old seedlings grown in light. Lower panel, 4-d-old seedlings grown in dark. (Scale bar, 1 cm.) (H) Primary root length of light-grown seedlings shown in (G). (I) Hypocotyl length of light-grown seedlings shown in (G). (J) Hypocotyl length of dark-grown seedlings shown in (G). Boxplots were presented as described in Fig. 1B. Different letters indicate significant differences (n = 35 per group, one-way ANOVA with Tamhane’s T2 test). For (H), the P value for pBRI1::BRI1 d2 d5 compared with Col-0 is 1.000, and for other lines compared with Col-0 are all less than 0.0001. For (I), the P values for pBRI1::BRI1 pD2::D2 and pBRI1::BRI1 pD5::D5 are 0.312 and 1.000, and for other lines compared with Col-0 are all less than 0.0001. For (J), the P values for each line compared with Col-0 are all less than 0.0001. (K) Phosphorylation state of BRI1-YFP in Col-0, d2 d5, as well as D2 and D5 transgenic lines under 35S promoter (D2-HA, D5-HA). BRZ-pretreated seedlings were treated with or without 100 nM BL for 60 min. BRI1-YFP was immunoprecipitated using α-GFP nanobody agarose beads. BRI1-YFP loading and phosphorylation was assessed by immunoblotting with α-GFP and α-T^+P/Y^+P antibodies, separately. (L) In vitro phosphatase assays showing BRI1_CD dephosphorylation by D2 and D5. Loading and phosphorylation of MBP-BRI1_CD was assessed by immunoblotting with α-MBP and α-T^+P/Y^+P antibodies, separately. CIP, a generic alkaline phosphatase from calf intestine, served as a positive control. PI, phosphatase inhibitor. (M) BRI1 amino acid sequence features and potential dephosphorylation sites identified by mass spectrometry. JM, intracellular-juxtamembrane domain; KD, kinase catalytic domain; CT, C-terminus. (N and O) In vitro phosphatase assays showing dephosphorylation of D2 (N) and D5 (O) on BRI1_CD incubated with or without SAUR15. Protein loading and MBP-BRI1_CD phosphorylation were assessed by immunoblotting with α-MBP, α-HA, α-FLAG, and α-T^+P/Y^+P antibodies, separately.

To assess genetic interactions between D2 and D5 with BRI1, we characterized plant growth phenotypes in reciprocal crosses of gain- and loss-of-function lines of BRI1, D2, and D5. Supplemented expression of D2 and D5 under their native promoters enhanced root and hypocotyl elongation as well as rosette growth defects in bri1-301, whereas d2 and d5 loss-of-function partially rescued these traits (Fig. 2 G–J and SI Appendix, Fig. S2 A–I). Similarly, elevated D2 and D5 expression partially suppressed the cell elongation phenotype of pBRI1::BRI1, whereas this phenotype was enhanced in d2 d5 (Fig. 2 G–J and SI Appendix, Fig. S2 A–I). Additionally, pBRI1::BRI1 d2 d5 grown in the light or dark exhibited a twisted root or hypocotyl phenotype (Fig. 2G), resembling BL-treated seedlings (7, 37). Petiole elongation in pBRI1::BRI1 was also further enhanced by D2 and D5 mutations (SI Appendix, Fig. S2 G and I).

To elucidate the function of D2 and D5 in BRI1-mediated BR signaling, we tested sensitivity of these lines to BL and BRZ. In bri1-301, supplemented D2 and D5 expression further reduced BL sensitivity but enhanced BRZ sensitivity, whereas d2 d5 loss-of-function partially restored BL sensitivity and reduced BRZ sensitivity (SI Appendix, Fig. S2 J–M). In the pBRI1::BRI1 background, BL sensitivity was suppressed by gain of D2- and D5- function, but further strengthened in d2 d5 (SI Appendix, Fig. S2 J and K). Moreover, compared to pBRI1::BRI1, BRZ sensitivity was increased in pBRI1::BRI1 pD2::D2 and pBRI1::BRI1 pD5::D5 but decreased in pBRI1::BRI1 d2 d5 (SI Appendix, Fig. S2 L and M). These results suggest that D2 and D5 regulate BR-mediated growth through their interaction with BRI1.

PP2C-D2, D5 Mediate the Dephosphorylation of BRI1, Which Is Restrained by SAUR15.

To confirm whether D2 and D5 modulate BRI1 activity, we first analyzed BRI1 phosphorylation in D2-, D5-altered seedlings treated with or without BL. As expected, under BL treatment, phosphorylation of BRI1 was reduced in D2 and D5 enhanced lines but promoted in the d2 d5 mutant compared to Col-0 (Fig. 2K and SI Appendix, Fig. S2N). In vitro phosphatase assays further confirmed that D2 and D5 directly dephosphorylated the BRI1_CD (Fig. 2L). To identify BRI1 site(s) targeted by PP2C-Ds, we used mass spectroscopy. Five sites in the kinase domain of BRI1 (Y898, Y945, S1026, S1044/T1045, and T1049) were identified as targets of D2 dephosphorylation in vitro (Fig. 2M and SI Appendix, Fig. S3). S1044/T1045 and T1049, located in the activation loop, are essential for BRI1 kinase activity and BR signaling (12). Since PP2Cs are generally thought to act as S/T phosphatases (24), we tested whether D2 and D5 are also able to dephosphorylate Y sites. Malachite green dephosphorylation assays showed that D2 and D5 directly dephosphorylated rSAPK3^+P peptide (RQADSEMTGY^+PVVTR), which is phosphorylated at Y (Y^+P) (SI Appendix, Fig. S2 O and P). Dephosphorylation of BRI1 Y sites by D2 and D5 was further confirmed by immunoblotting with a specific α-Y^+P antibody (SI Appendix, Fig. S2Q). Thus, D2 and D5 are dual-specificity phosphatases, and directly dephosphorylate BRI1 to regulate its activity.

Since SAUR15 suppresses D2, D5 and activates BRI1 (10, 11), we investigated its role in PP2C-D-mediated BRI1 regulation. We found that incubation with SAUR15 repressed D2, D5-catalyzed BRI1 dephosphorylation (Fig. 2 N and O). These results indicated that SAUR15 restrains D2 and D5-mediated inhibition of BRI1.

PP2C-D2, D5 Phosphatase Activity Is Directly Suppressed by BRI1 Through Phosphorylation.

Mass spectroscopy also identified several phosphorylated residues in D2 in the presence of BRI1_CD and ATP in vitro (SI Appendix, Fig. S4), suggesting that BRI1 may regulate D2 phosphorylation. In vitro kinase assays confirmed that BRI1_CD phosphorylates D2 (Fig. 3 A–C and SI Appendix, Fig. S5). Phos-PAGE and Phos-tag^™ Biotin assays were performed to assess in vivo phosphorylation. Compared to untreated seedlings, the phosphorylation of D2 was promoted by BL treatment (Fig. 3 D and E). Two in vivo phosphorylation sites of D2 were identified by mass spectroscopy (SI Appendix, Fig. S6 A–E). Combined with in vitro mass spectrometry results, three phosphorylation sites in D2, including S193, Y241, and S253, were identified (Fig. 3F and SI Appendix, Figs. S4 and S6 A–E). In vitro and in vivo phosphorylation assays proved that D5 was also phosphorylated by BRI1 (SI Appendix, Fig. S5 and S7 A–E). Since S193 and Y241 are highly conserved in PP2C-Ds (SI Appendix, Fig. S6 F and G), we carried out site-directed mutation of these sites to determine whether they are targets of phosphorylation. Our results showed that S-to-A and Y-to-F substitutions weaken BRI1-mediated phosphorylation of D2 and D5 (Fig. 3G and SI Appendix, Fig. S7F), confirming that BRI1 regulates phosphorylation at S193 (S190 in D5) and Y241 (Y238 in D5). Given that BRI1 also interacts with D1, D2, D3, D4, D5, D6, D7, and D9 (SI Appendix, Fig. S6H), it may likewise phosphorylate other PP2C-Ds.

Fig. 3.

BRI1 directly phosphorylates and represses PP2C-D2. (A and B) BRI1_CD phosphorylation of D2. Protein loading and phosphorylated D2 were assessed by immunoblotting with α-HA, α-FLAG, α-S^+P (A) and α-T^+P/Y^+P antibodies (B), separately. (C) BRI1_CD-mediated phosphorylation of D2 determined by in vitro phosphorylation assays followed by p-nitrobenzyl mesylate (PNBM) alkylation. Protein loading and phosphorylated D2 were assessed by immunoblotting with α-HA, α-FLAG, and α-thiophosphate ester antibodies, separately. (D) D2 phosphorylation in seedlings treated with or without BL. BRZ-pretreated seedlings were collected and then treated with or without 1 μM BL for 120 min. Total protein was extracted. D2-HA were immunoprecipitated using α-HA Nanobody Agarose Beads and separated in a Phos-tag gel. D2-HA was assessed by immunoblotting with an α-HA antibody. (E) D2 phosphorylation in plants analyzed by Phos-tag^™ Biotin. Seedlings were treated as described in (D). D2-HA were immunoprecipitated using α-HA Nanobody Agarose Beads. Immunoprecipitated protein was detected with an α-HA antibody and the phosphorylated D2 was assessed with Phos-tag™ Biotin. D2^+P/D2 ratio were calculated as described in Fig. 1I. (F) Diagram showing amino acid sequence features and potential phosphorylation sites on D2. Phosphorylation sites on D2 identified by in vitro mass spectrometry were marked with asterisk, and by in vivo mass spectrometry were underlined. (G) Phosphorylation of wild-type as well as S-to-A and Y-to-F mutated D2 by BRI1_CD. Protein loading and phosphorylated protein were assessed as described in (C). D2^+P/D2 ratio was calculated as described in Fig. 1I. For (A to C, G), BRI1_CD-mediated phosphorylation of MBP-FLAG was analyzed in (SI Appendix, Fig. S5). (H) BRI1 suppressed D2 phosphatase activity. GST-tagged protein elution buffer (no protein), GST-HA, and GST-D2-HA were pretreated with or without MBP or MBP-BRI1_CD. (I) Mutations at S193 and Y241 affect the catalytic activity of D2. GST-HA, wild-type, and mutant D2 were pretreated with MBP-BRI1_CD and ATP. For (H and I), pNPP was used as the dephosphorylation substrate, and data shown are means ± SD (n = 4 per group). Statistical significance analysis for the absorbance of 405 nm at 5, 10, 15, and 20 min is shown in (SI Appendix, Fig. S8 A and C). (J and K) Effect of site-directed mutagenesis of Y241 (J) and S193 (K) on BRI1 dephosphorylation. Protein loading and BRI1_CD phosphorylation was assessed by immunoblotting with α-MBP, α-HA, and α-T^+P/Y^+P antibodies, separately.

Phosphatase assays using a chemical substrate p-nitrophenylphosphate (pNPP) showed that BRI1 inhibits D2 and D5, as pretreatment with MBP-BRI1_CD and ATP, but not MBP and ATP, significantly suppressed their phosphatase activity (Fig. 3H and SI Appendix, Figs. S7G and S8 A and B; S8A and B). To test whether this inhibition was phosphorylation-dependent, we compared phosphatase activity of wild-type D2 and D5 with phospho-mimetic and phospho-deficient variants. Compared to the wild-type, phospho-mimetic variants Y241D/E and Y238D/E showed reduced activity, while phospho-deficient variants Y241F and Y238F showed increased activity (Fig. 3I and SI Appendix, Figs. S7H and S8 C and D). Immunoblotting confirmed that Y241D/E and Y238D/E substitutions reduced phosphatase activity toward BRI1_CD, whereas Y241F and Y238F substitutions increased activity (Fig. 3J and SI Appendix, Fig. S7I). Additionally, mutations S193F/A/D/E in D2 or S190F/A/D/E in D5 significantly reduced activity (Fig. 3 I and K and SI Appendix, Fig. S7 H and J and S8 C and D), indicating that modifications at this critical residue disrupt function. These findings show that phosphorylation is a key regulator of D2 and D5 activity.

PP2C-D2, D5 Are Activated by Dimerization, a Process Inhibited by Phosphorylation.

We showed that BRI1 inhibits D2 and D5 through phosphorylation, but the mechanism underlying this control was unclear. Protein structure analysis has indicated that Oryza sativa photosystem II core phosphatase (OsPBCP, from K clade of PP2C) form homo-oligomers (38). Y2H and pull-down assays indicate a direct interaction between D2 with itself as well as D5 (Fig. 4A and SI Appendix, Fig. S9 A and B). BiFC and Co-IP assays confirmed that D2 interacts with itself as well as D5 in planta (Fig. 4 B–D). Likewise, D5 formed homodimers both in vitro and in planta (SI Appendix, Fig. S9 C–E). These results indicate that D2 and D5 form homo- and heterodimers both in vitro and in vivo.

Fig. 4.

PP2C-D2 and D5 form homo- and heterodimers, which are inhibited by phosphorylation. (A) Y2H assay showing interactions between D2 and D5. BFP2, D2, and D5 proteins were fused to the Cub and Nub. SAUR15 protein was fused to Nub. D2-Cub with SAUR15-Nub (D2/SAUR15) served as a positive control (11). BFP2-Cub with BFP2-Nub (BFP2/BFP2) served as a negative control. (B) BiFC analysis showing D2 and D5 interactions in N. benthamiana leaf epidermal cells. D2 and D5 were fused to the YC or YN. Leaves coinfiltrated with GV3101 carrying D2-YC and HA-YN constructs served as the negative control. YFP, Bright and Merged images are shown. (Scale bar, 10 μm.) (C and D) Co-IP assays of D2-HA with D2-GFP-FLAG and D5-GFP-FLAG. D2-GFP-FLAG (C) and D5-GFP-FLAG (D) were coexpressed with D2-HA or BFP2-HA. 10-d-old seedlings grown on ½ MS medium were collected. IP and Input proteins were analyzed by immunoblotting with α-FLAG and α-HA antibodies, separately. (E) Y2H assays showing the effect of site-directed mutagenesis on the interaction between D2 and D5. Yeast was diluted (10⁻¹, 10⁻², and 10⁻³) of saturated cultures. Autoactivation and control assays are shown in (SI Appendix, Fig. S11A). (F and G) rBiFC analysis of interaction between site-mutant D2 (F), D2 and D5 (G). YFP, RFP, and YFP/RFP merged (Merged) images are shown. (Scale bar, 10 μm.) Fluorescence intensity of YFP and RFP was measured using ImageJ and the quantified YFP/RFP ratio for different combinations are shown at the Bottom. Boxplots were presented as described in Fig. 1B. Different letters represent significant differences (n = 50 per group, one-way ANOVA with Tamhane’s T2 test). For (F), the P values for 241F compared with D2 are 0.998, and for other combination compared with D2 are all less than 0.0001. For (G), the P values for Y238F/Y241F compared with D5/D2 are 0.993, and for other combination compared with D5/D2 are all less than 0.0001.

Since D2 and D5 activity is repressed by phosphorylation, we tested whether phospho-mimetic and phospho-deficient mutations at S193/S190 and Y241/Y238 affect D2 and D5 interactions. Y2H, pull-down, and ratiometric bimolecular fluorescence complementation (rBiFC) assays showed that while Y241F or Y238F had no effect, Y241D/E, Y238D/E, S193F/A/E/D, and S190F/A/E/D substitutions repressed dimerization (Fig. 4 E–G and SI Appendix, Fig. S10).

Together, these findings suggest that D2 and D5 function as homo- and heterodimers, and that their dimerization is inhibited by phosphorylation.

BR-triggered BRI1 and SAUR15 Inhibit PP2C-D2 and D5 Activity by Disrupting Dimerization.

Given that BL treatment promotes phosphorylation of D2 and D5, we decided to investigate whether BL impacts D2 and D5 dimerization. Co-IP assays using seedlings coexpressing D2-HA with D2-GFP-FLAG or D5-GFP-FLAG fusion protein treated with or without BL showed that D2 and D5 dimerization was repressed by BL treatment (Fig. 5 A and B). To assess the effect of BRI1 on D2 and D5 dimerization, we performed pull-down assays. Results showed that homo- and heterodimerization was weaker in the presence of BRI1_CD, which was further confirmed by rBiFC assays (Fig. 5 C–E and SI Appendix, Fig. S11 A and B).

Fig. 5.

BRI1 and SAUR15 inhibit PP2C-D2 and D5 dimerization. (A and B) BL treatment represses interaction between D2 and D5. D2-GFP-FLAG (A) and D5-GFP-FLAG (B) were coexpressed with D2-HA or BFP2-HA. BRZ-pretreated seedlings were collected and treated with or without 1 μM BL for 120 min. IP and Input proteins were analyzed by immunoblotting with α-FLAG and α-HA antibodies, separately. The HA/FLAG ratio was calculated as described in Fig. 2E. (C and D) Pull-down assays showing the effect of BRI1 on D2 and D5 dimerization. GST-D2 fusion proteins with MBP-MBP-HA or MBP-MBP-BRI1_CD-HA fusion proteins were incubated with amylose resin bound with MBP-D2-FLAG (C) or MBP-D5-FLAG (D). IP and Input proteins were analyzed by immunoblotting with α-FLAG, α-GST, and α-HA antibodies, separately. (E) rBiFC analysis of PP2C-Ds dimerization in the presence and absence of BRI1 in N. benthamiana leaf epidermal cells. BFP, YFP, RFP, and YFP/RFP merged (Merged) images are shown. The x-axis denotes leaves coinfiltrated with GV3101 carrying constructs for rBiFC, with “-BRI1” indicating BFP2 alone and “+BRI1” indicating BRI1 fused to BFP2. (Scale bar, 10 μm.) (F and G) Pull-down assays showing the effect of SAUR15 on D2 and D5 dimerization. GST-D2-HA fusion proteins with GST-FLAG or GST-SAUR15-FLAG fusion proteins were incubated with amylose resin bound with MBP-D2 (F) or MBP-D5 (G). IP and Input proteins were analyzed by immunoblotting with α-MBP, α-HA, and α-FLAG antibodies, separately. (H) rBiFC analysis of PP2C-Ds dimerization in the presence or absence of SAUR15 in N. benthamiana leaf epidermal cells. BFP, YFP, RFP, and YFP/RFP merged (Merged) images are shown. The x-axis denotes leaves coinfiltrated with GV3101 carrying constructs for rBiFC, with “-SAUR15” indicating BFP2 alone and “+SAUR15” indicating SAUR15 fused to BFP2. (Scale bar, 10 μm.) For (E and H), fluorescence intensity of YFP and RFP was measured and calculated as described in Fig. 4F and YFP/RFP ratio is shown at the Right. Boxplots were presented as described in Fig. 1B. Statistical significance between “–BRI1” and “+BRI1” as well as “–SAUR15” and “+SAUR15” was determined by the independent samples t test (n = 30 per group, ****P < 0.0001). (I) Model illustrating how D2 and D5 regulate BR signaling. There is a phosphorylation and dephosphorylation balance between BRI1 and D2, D5. When BRs are absent (Left), the dephosphorylation by D2 and D5 on BRI1 overtakes the phosphorylation by BRI1 on D2 and D5. Dimerized D2 and D5 actively dephosphorylate BRI1, keeping BR signaling off. When BRs such as BL are present (Right), the phosphorylation by BRI1 on D2 and D5 overtakes the dephosphorylation by D2 and D5 on BRI1. Ligand perception promotes autophosphorylation and transphosphorylation of BRI1 and BAK1. The activated BRI1 phosphorylates D2 and D5, which suppresses interaction between D2 and D5, and thus inhibits their phosphatase activity. BR-induced SAUR15 further reduces D2 and D5 catalytic activity by promoting monomerization. The highlighted portion indicates the central findings of this study.

Multiple studies have shown that SAURs directly inhibit PP2C-Ds phosphatase activity (11, 32, 39, 40), but the underlying mechanisms remain unclear. Dimerization is essential for D2 and D5 activity, which is suppressed by BL treatment (Figs. 3 I–K, 4 E–G, and 5 A and B and SI Appendix, Figs. S7 H–J, S8 C and D, and S11). Given that BR-upregulated SAUR15 inhibits D2 and D5 (10, 11), we tested whether SAUR15 modulates D2 and D5 activity by disrupting their homo- and heterodimerization. Pull-down and rBiFC assays both showed that SAUR15 inhibits their dimerization (Fig. 5 F–H and SI Appendix, Fig. S11 C and D).

Together, these results suggest that BL, BRI1, and SAUR15 inhibit D2 and D5 by repressing their dimerization.

Discussion

The activation and deactivation of BRI1 is essential for regulating the initiation and duration of BR signaling (15). BR perception triggers phosphorylation between BRI1 and SERK coreceptors, activating signaling, while PP2A dephosphorylates the C-terminal autoinhibitory domain, leading to BRI1 signaling inactivation (22). However, the mechanisms controlling BRI1 activity through dephosphorylation of its kinase activation loop are not fully understood.

This study identifies PP2C-D phosphatases as key negative regulators of BR signaling. We show that D2 and D5 dephosphorylate multiple BRI1 sites, including S1044/T1045 and T1049 in the activation loop, to suppress BR signaling. D2 and D5 are activated through dimerization, which is in turn disrupted by BRI1-mediated phosphorylation (Fig. 5I). Likewise, SAUR15, upregulated by BR signaling (10), inhibits D2 and D5 activity by disrupting their dimerization (Fig. 5I).

PP2C-D2 and D5 Inhibit BR Signaling by Dephosphorylating Residues in BRI1’s Activation Loop.

Dozens of S, T, and Y residues in the cytoplasmic domain of BRI1 undergo reversible phosphorylation to regulate BR signaling (12–14). Unlike PP2A, which dephosphorylates S and T residues in BRI1’s C-terminal autoinhibitory domain to inhibit BR signaling (22), our biochemical data show that D2 and D5 dephosphorylate multiple S and T residues, including S1026, S1044/T1045, and T1049 (Fig. 2 and SI Appendix, Fig. S3). Since mutations at different BRI1 phosphorylation sites variably affect plant growth and BR signaling (17, 41–43), we propose that PP2C-Ds and PP2As regulate distinct BR-mediated growth and development processes. Among the S/T residues dephosphorylated by PP2C-Ds, S1044/T1045 and T1049 in BRI1’s activation loop are critical for kinase activity (Fig. 2 and SI Appendix, Fig. S3; 12, 14). These findings suggest that PP2C-Ds deactivate BRI1 signaling, consistent with enhanced cell expansion and increased BR sensitivity in D2, D5, D6 mutants, as well as reduced cell expansion and decreased BR sensitivity when expression of PP2C-Ds was enhanced using their native promoters (Figs. 1 and 2 and SI Appendix, Fig. S1 and S2).

Y phosphorylation and dephosphorylation also regulate BR signaling (13). Aside from BRI1 SUPPRESSOR 1 (BSU1), which dephosphorylates Y200 of BIN2 (44), the phosphatases acting on other phosphorylated Y residues in BR signaling are unknown. We identified D2 and D5 as dual-specificity phosphatases, acting on both S/T and Y residues, including Y898 and Y945 of BRI1 (Fig. 2 and SI Appendix, Figs. S2 and S3). Y898, part of a Y₈₉₈KAI motif, is critical for adaptor protein complex-2-dependent clathrin-mediated endocytosis of BRI1 (19). Y898 dephosphorylation suggests a role of D2 and D5 in reducing the PM pool of BRI1, further suppressing BR signaling. PP2C-Ds are involved in various cellular processes (11, 29, 31, 39, 45), including dephosphorylation of PM H⁺-ATPase downstream of the BR signaling pathway, which curtails BR-mediated growth (45). Our findings indicate that PP2C-Ds also function upstream in the BR signaling pathway, directly regulating BRI1 kinase activity through dephosphorylation.

PP2C-D2 and D5 Are Regulated by Phosphorylation and Dimerization.

Protein phosphatases, once thought to be constitutively active housekeeping enzymes, are now recognized as specific molecular switches, whose activity is controlled by expression, subcellular localization, posttranslational modification, and dimerization (46, 47). Here, we show that PP2C-D2 and D5 are tightly regulated by phosphorylation and dimerization and uncover the connection between these two regulatory mechanisms.

Phosphorylation negatively regulates several PP2C family members but, in most instances, this has been shown to involve S residues. For example, BIK1 and MAP4K4-mediated phosphorylation of S77 in PP2C38/PP2C-D3 removes its repression of innate immune signaling (29, 30). Similarly, receptor-like cytoplasmic kinases RLCK-VII/PBS1-LIKEs (PBLs) phosphorylate the S-X-X-L motif of PP2Cs POLTERGEIST-LIKE (PLLs), inhibiting their activity in CLAVATA/BARELY ANY MERISTEM (CLV/BAM) signaling (28). Our data show that D2 and D5 are phosphorylated by BRI1 at S193/S190 and Y241/Y238 (Fig. 3 and SI Appendix, Figs. S4–S7). This phosphorylation is required for PP2C-Ds deactivation, as phospho-mimetic variants Y241D/E and Y238D/E exhibit reduced catalytic activity, while phospho-deficient variants Y241F and Y238F show enhanced activity in biochemical assays using BRI1 or a chemical substrate (Fig. 3 and SI Appendix, Figs. S7 and S8). Notably, both phospho-mimetic and phospho-deficient substitutions at S sites (S193F/A/E/D and S190F/A/E/D) abolished activity, suggesting that this S residue is critical for D2 and D5 function (Fig. 3 and SI Appendix, Figs. S7 and S8).

Dimerization is a well-known regulatory mechanism for receptor protein tyrosine phosphatases (RPTPs) in animals (47, 48), but its role in regulating plant phosphatases is relatively unexplored. While Arabidopsis PROTEIN TYROSINE PHOSPHATASE1 (PTP1) and rice OsPBCP form homo-oligomers (38, 49), the significance of their dimerization is unclear. We demonstrated D2 and D5 homo- and heterodimerization through in vitro and in vivo assays (Fig. 4 and SI Appendix, Fig. S9). Notably, S193/S190 substitutions disrupted both dimer formation and catalytic activity, emphasizing dimerization as essential for D2 and D5 activation (Figs. 3 and 4 and SI Appendix, Figs. S7, S8, and S10). We also show that phosphorylation inhibits D2 and D5 dimerization, as BL treatment, BRI1, and phospho-mimetic Y241/Y238 substitutions weakened monomer interactions (Figs. 4 and 5 and SI Appendix, Figs. S10 and S11). Conversely, phospho-deficient Y241/Y238 substitutions enhanced D2 and D5 activity without altering monomer interactions (Figs. 3 and 4 and SI Appendix, Figs. S7, S8, and S10). In animal RPTPs, posttranslational modifications like oxidation and phosphorylation regulate function by altering dimer conformation rather than preventing dimerization (50–52). Similarly, the increased activity of Y241F/Y238F variants may result from conformational changes in D2 and D5 dimers, suggesting a potential autoregulatory mechanism.

SAURs generally associate with PP2C-Ds to modulate their phosphatase activity in plant growth and development, with most acting as inhibitors (11, 39, 40, 53, 54). However, their precise mechanism is unclear. Wong et al. (54) identified a conserved PP2C-D motif critical for SAUR binding, where missense mutations confer resistance to SAUR and constitutive activity. Our data show that SAUR15, which inhibits D2 and D5 (11), disrupts monomer interactions (Fig. 5; SI Appendix, Fig. S11). Since dimerization promotes PP2C-Ds activation (Figs. 3 and 4 and SI Appendix, Figs. S7, S8, and S10), these results suggest that SAURs inhibit PP2C-Ds activity by preventing dimer formation.

As a key mechanism for protein function modification as well as signal transduction, reversible protein phosphorylation is mediated by protein kinases and protein phosphatases. Plants have far more kinases than phosphatases (24). This imbalance, combined with the high specificity of kinases, has contributed to the perception that phosphatases play a less desirable role in signaling (24, 55–57). However, this disparity raises a key question that how does a relatively small set of phosphatases counteract the extensive phosphorylation driven by numerous kinases, and highlights the need for tight regulation of phosphatases activity. Our findings identify dimerization as a critical activation mechanism for PP2C-Ds and prove that BRI1-mediated phosphorylation and SAUR15 binding are important regulatory mechanisms that maintain PP2C-Ds in an inactive monomeric state (Fig. 5I). These observations further suggest that there must be phosphatase(s) that regulate D2 and D5 dephosphorylation in plants. An autoactivation mechanism may exist whereby the D2 and D5 dimers mediate the dephosphorylation of their phosphorylated monomeric proteins.

Our findings establish PP2C-Ds as key negative regulators of BR signaling and uncover mechanisms controlling their activity. Future studies may explore their role in BRI1 endocytosis and the impact of other posttranslational modifications on dimerization.

Materials and Methods

Plant Materials and Growth Conditions.

All plants were Arabidopsis (Arabidopsis thaliana) ecotype Columbia-0 (Col-0) background. Single T-DNA mutants d2, d5, d6, bri1-701, and cpd were obtained from the Arabidopsis Biological Resource Center (10, 11). Double and triple mutants were generated by crossing. T-DNA mutants were genotyped as recommended (http://signal.salk.edu). The point mutant bri1-301 was genotyped using a dCAPS method (33).

To generate native promoter-driven transgenic lines, genomic fragments spanning −2,564 bp to +2,341 bp for D2, −2,559 bp to +1,513 bp for D5, and -2,533 bp to 1,919 bp for D6 were amplified from Arabidopsis genomic DNA. HA followed by a stop codon at the 3’-terminus was fused to the 3’ end of each gene by PCR. For 35S promoter-driven overexpression lines, GFP coding sequence was fused to D2 and D5 by overlap PCR to generate D2-GFP and D5-GFP, and coding sequences for BRI1, D2, D5, mTagBFP2, D2-GFP, and D5-GFP were amplified. Sequences were cloned into pDONR/ZEO using BP clonase. All inserts were verified by sequencing and then transferred into destination vectors pBIB-BASTA-GWR, pBIB-35S-GWR-FLAG (35S promoter with FLAG tag), or pBIB-35S-GWR-3HA (35S promoter with 3HA tag) by LR clonase for plant expression. Agrobacterium tumefaciens strain GV3101 containing each target construct was used for transformation. Arabidopsis plants were transformed by floral dipping as previously described (11).

Plants were grown under long-day conditions (16 h white light per day, ~100 μmol m⁻² s⁻¹ light intensity; 22 ± 2 °C) except for dark treatment. For BR sensitivity assays, seeds were placed on ½ Murashige and Skoog (MS) medium containing 1% (w/v) sucrose with different concentrations of BL under long-day conditions. For BRZ sensitivity assays, seeds were grown in darkness on ½ MS medium containing 1% (w/v) sucrose with different concentrations of BRZ. For liquid culture, seedlings were grown in flasks containing 50 mL of ½ MS medium with 1% (w/v) sucrose. Photos were taken with a digital camera. Root and hypocotyl lengths were measured using ImageJ software.

Primer sequences, T-DNA mutant stock information, details for mutant and transgenic lines generation, plasmid, enzyme, and chemical sources are provided in SI Appendix, Tables S1–S6.

Methods for RNA extraction and RT-qPCR analysis, protein expression and purification, in vitro pull-down assays, Y2H assays, BiFC and rBiFC, Co-IP, phosphorylation and phosphatase assays, in vitro kinase assays, mass spectrometry, statistical analysis, and accession numbers are described in the SI Appendix, Materials and Methods.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Data, Materials, and Software Availability

All study data are included in the article and/or supporting information.