OPEN STOMATA 1 activates SLAC1 anion channel primarily through CPK15 in ABA-induced stomatal closure in Arabidopsis

Significance

Abscisic acid (ABA) activates Ca²⁺ signaling to promote stomatal closure. The Ca²⁺-independent kinase OST1 and the Ca²⁺-dependent protein kinases CPKs directly phosphorylate and activate SLAC1 and SLAC1 homologues (SLAHs). However, the priming of Ca²⁺ signaling and the relationship between these pathways remain unclear. Here, we demonstrate that the Ca²⁺-dependent protein kinase CPK15 is essential for ABA-induced stomatal closure, with OST1 primarily activating SLAC1 anion channel activity through CPK15. OST1 and Ca²⁺ work synergistically to enhance CPK15 activity. These findings reveal a mechanism where OST1 connects Ca²⁺-independent and Ca²⁺-dependent pathways via CPK15 to regulate ABA-induced stomatal movement.

Abstract

The Ca²⁺-independent OST1 and Ca²⁺-dependent protein kinases CPKs both activate the anion channel SLAC1 during ABA-induced stomatal closure pathway. However, the mechanism by which OST1 regulates SLAC1 activation and its relationship with CPKs remain unclear. Here, we identify that OST1 primarily activates SLAC1 in this process through CPK15. Mutation of CPK15 significantly impairs ABA-induced stomatal closure and increases drought sensitivity. OST1 interacts with CPK15 and phosphorylates it at T103, which is essential for ABA-induced stomatal closure. Moreover, CPK15 can phosphorylate eight sites in the N terminus of SLAC1 to activate its anion currents in oocytes. Expression of SLAC1^8D (a phosphomimetic form) in oocytes constitutively activates anion channel activity and effectively restores the impaired ABA-induced stomatal closure of cpk15-1 but not by SLAC1^8A (a phospho-dead form). Furthermore, activated CPK15 by OST1 and Ca²⁺ enhances its activity toward SLAC1, and mutations of both OST1 and CPK15 have additive effect on ABA-induced stomatal closure, suggesting that CPK15 activates SLAC1 through both direct and indirect mechanisms. These findings demonstrate the key role of CPK15 in ABA-induced stomatal closure, revealing a connection between OST1 and CPKs in both Ca²⁺-independent and Ca²⁺-dependent pathways in ABA-induced stomatal closure.

Stomatal pore, controlled by guard cells, governs gas exchange between the plant and its surroundings for photosynthesis and transpiration. Guard cells respond to environmental signals (e.g., drought, CO₂, ABA, and ethylene) by adjusting stomatal opening or closure via turgor pressure changes (1–4). ABA, a pivotal phytohormone, promotes stomatal closure to limit water loss through a complex signaling network (5, 6).

SLOW ANION CHANNEL-ASSOCIATED 1 (SLAC1) is a key effector in stomatal closure (7, 8). ABA triggers SLAC1 phosphorylation at both N- and C-terminus via Ca²⁺-independent and Ca²⁺-dependent pathway (9–12). In the Ca²⁺-independent pathway, the SnRK2 kinase OPEN STOMATA 1 (OST1) acts as a central player in SLAC1 activation (9). Under basal conditions, OST1 is inhibited by class 2C protein phosphatases (PP2Cs) by direct binding (13, 14). However, when ABA is present, ABA-bound PYR/PYL (PYRABACTIN RESISTANCE 1/PYR1-LIKEs) receptors sequester PP2Cs, enabling OST1 activation via Raf-like kinases and autophosphorylation (6, 15–18). Active OST1 phosphorylates SLAC1, inducing anion efflux, membrane depolarization, and stomatal closure (9, 19). In this signaling pathway, changes in cytosolic Ca²⁺ levels do not affect the activity of OST1.

ABA also elevates cytosolic Ca²⁺ in guard cells (20, 21), activating SLAC1 and its homologous channel SLAHs via Ca²⁺-dependent protein kinases (CPKs) such as CPK3/4/6/11/21 (22–24). The process of ABA-induced stomatal closing is slightly impaired in cpk3/6 and cpk4/11 double mutants (25, 26) and is further suppressed in the cpk3/4/5/6/11 quintuple mutant (27, 28). In vitro assays have shown that CPK21 and CPK23 phosphorylate SLAC1/SLAHs (12, 22, 29), yet their individual knockouts enhance drought tolerance (30, 31), contradicting the general role of CPK21/23 and other CPKs in regulating anion channels. It is noteworthy that CPKs are encoded by a large multigene family consisting of 34 members in Arabidopsis, suggesting the functional redundancy or the presence of yet undiscovered CPKs dominate under physiological conditions. While the activity of CPKs is typically regulated by calcium, it remains unclear whether other protein kinases can modulate the activity of CPKs during ABA-mediated stomatal closure.

SLAC1 activation requires phosphorylation at multiple sites, including major sites such as S59, T62, S65, S86, S113, T114, S116, S120, and T513, along with some additional minor sites (10, 11, 32–35). OST1 has been shown to phosphorylate S120 (19, 36), while CPK6 specifically targets S59 for phosphorylation (10). Notably, expression of SLAC1^S120A can largely rescue the impaired ABA-induced stomatal closure of slac1-3 (37), suggesting that there may be additional phosphorylation sites or roles of OST1 beyond its action on S120. Moreover, it is unclear which kinases are responsible for phosphorylating the other identified sites in vivo, whether additional phosphorylation sites exist within SLAC1, and which specific site plays a predominant role in ABA-induced stomatal closure.

Although Ca²⁺-dependent and -independent pathways in ABA signaling have been long regarded as separate, OST1 dysfunction impairs Ca²⁺ signaling and inhibits Ca²⁺-induced stomatal closure (21, 38). Futhermore, OST1 directly phosphorylates the cyclic nucleotide-gated channels CNGCs (CNGC5/6/9/12) (39, 40), thereby facilitating Ca²⁺ influx during ABA signaling. CPKs such as CPK3/4/6/11/27 can also activate OST1 under osmotic stress (41), hinting an interplay between these two pathways. However, it remains unknown whether OST1 interacts with specific CPKs in the ABA signaling.

In this study, we identify CPK15 as an OST1-regulated kinase in ABA-induced stomatal closure. CPK15 phosphorylates eight residues in SLAC1 N terminus. OST1 enhances CPK15 activity on SLAC1 via direct phosphorylation and Ca²⁺ elevation during ABA treatment, linking Ca²⁺-independent and Ca²⁺-dependent pathways through the OST1–CPK15 module.

Results

OST1 Physically Interacts with CPK15.

OST1 is central to both Ca²⁺-independent and -dependent pathways. To investigate the relationship between OST1 and CPKs, we identified several highly expressed CPKs in guard cells that are either phosphorylated during ABA treatment or involved in ABA-induced stomatal movement (28, 42, 43). We conducted split-Luciferase assays in Nicotiana benthamiana to examine interactions between OST1 and CPK4, CPK5, CPK6, CPK11, CPK15, CPK21, and CPK32. All seven interacted with OST1 at comparable expression levels, with the strongest interaction observed between CPK15 and OST1 (Fig. 1 A and B and SI Appendix, Fig. S1 A and B). While it has been reported that CPK4/6/11 phosphorylate OST1 during osmotic stress (41), we focused on further characterizing CPK15. No enhancement was detected for the interaction of OST1 and CPK15 under ABA treatment (SI Appendix, Fig. S1 C and D). To validate their in vivo interaction, we conducted coimmunoprecipitation (co-IP) experiments by coexpressing OST1–YFP and CPK15–Flag in Arabidopsis protoplasts. Results showed that CPK15–Flag was specifically immunoprecipitated by OST1–YFP, but not by YFP alone (Fig. 1C). These results clearly demonstrate that OST1 interacts with CPK15 in vivo, indicating a role of CPK15 in guard cells.

Fig. 1.

OST1 interacts with CPK15. (A) The interaction of OST1 with CPKs by split-LUC assay. N-terminal LUC (nLUC) fused to N terminus of OST1 and C-terminal LUC (cLUC) fused to C-terminus of CPKs were coexpressed in N. benthamiana leaf epidermis. (B) Quantitative analysis of the luminescence intensity in (A). Data shown are mean ± SEM (n = 3 independent biological replicates). Different letters above the error bars indicate the significant differences at P < 0.05 by One-way ANOVA with Tukey’s test. (C) Coimmunoprecipitation of OST1 with CPK15 in Arabidopsis protoplasts coexpressing 35S–OST1–YFP and 35S–CPK15–FLAG, or 35S–YFP and 35S–CPK15–FLAG. Protein extracts were immunoprecipitated with GFP-beads. The input and coimmunoprecipitated proteins were detected with anti-GFP and anti-FLAG antibodies as indicated. IP, immunoprecipitation. Experiments were repeated three times.

CPK15 Is Involved in ABA-Induced Stomatal Movement.

Public database and our experiments show that CPK15 is highly expressed in guard cells (43, 44) (Fig. 2A and SI Appendix, Fig. S2A). ABA treatment induced dynamic CPK15 expression, peaking at 2 h (SI Appendix, Fig. S2B). Promoter-GUS assays revealed CPK15 expression in cotyledons, rosette leaves, floral organs, as well as both the basal and distal ends of siliques (SI Appendix, Fig. S2A). Additionally, we examined the expression level of CPK15 in guard cells by GUS staining on the leaf epidermis of ProCPK15:GUS transgenic plants with or without ABA. Quantitative analysis confirmed ABA-induced upregulation of CPK15 in guard cells (Fig. 2A and SI Appendix, Fig. S2C). Furthermore, subcellular localization in N. benthamiana leaves showed that CPK15–YFP localized to the cytosol and plasma membrane, showing signals along the cell periphery and within intracellular punctate structures that partially colocalized with the plasma membrane marker CBL1 (45) (SI Appendix, Fig. S2 D and E).

Fig. 2.

CPK15 is involved in ABA-induced stomatal movement and drought performance. (A) GUS staining of epidermal peels from rosette leaves of ProCPK15:GUS transgenic plants treated with or without 20 μM ABA for 1 h. (Scale bar, 8 μm.) (B) Time-resolved stomatal conductance of Col-0, cpk15 mutants, and CPK15 complementation lines in response to ABA. (C) Time-resolved relative stomatal conductance in response to ABA in Col-0, cpk15-1, and cpk3/4/5/6/11 quintuple mutant plants. Data in (B and C) are presented as means ± SEM. n = 3 leaves per genotype per experiment. Experiments were repeated three times. (D) Drought performance assessment of Col-0, cpk15-1, and its CPK15 complementation plants (CPK15_COM). (Scale bar, 2.5 cm.) Experiments were repeated 3 times, and the survival rates of different genotypes were depicted as means ± SEM.

To investigate the role of CPK15, we generated two mutants by CRISPR/Cas9 technology: cpk15-1 (with a 1,489-bp deletion from 132th nucleotide) and cpk15-2 (with one bp deletion at 190th nucleotide), both creating premature stop codons at the 55th and 74th AA, respectively (SI Appendix, Fig. S3A). Both mutant lines are Cas9-free homozygous lines after backcrossing with Col-0. They showed significantly reduced stomatal closure to ABA treatment compared to Col-0 plants, as assessed by measuring stomatal aperture in leaf epidermal layers (SI Appendix, Fig. S3B) and time-resolved stomatal conductance in intact detached leaves (Fig. 2B and SI Appendix, Fig. S3D), which better reflect real-time in vivo stomatal responses (46). These impaired phenotypes were fully rescued by expression of native CPK15 (Fig. 2B and SI Appendix, Fig. S3 C and D). CPK15 overexpressing under the 35S promoter did not enhance ABA responses compared to Col-0 with equivalent treatment conditions (SI Appendix, Fig. S3 E–G).

Previous studies have shown that CPK3/4/5/6/11 participate in ABA-induced stomatal closure (28). While the cpk3/4/5/6/11 quintuple mutant shows significant insensitivity to ABA, single mutants display only weak or negligible phenotypes (25, 26). Strikingly, cpk15-1 and cpk3/4/5/6/11 quintuple mutant exhibited comparable levels of stomatal ABA insensitivity (Fig. 2C and SI Appendix, Fig. S3 H and I). These findings underscore the crucial role of CPK15 as a key player in ABA-induced stomatal closure.

Drought stress stimulates ABA synthesis, and impairment in ABA-induced stomatal closure may increase plant susceptibility to drought stresses. Under drought stresses, cpk15 mutants wilted earlier than Col-0, with 100% mortality after rewatering (SI Appendix, Fig. S3J). Complementation with CPK15 restored the drought-sensitive phenotype in cpk15-1 mutants (Fig. 2D), linking CPK’s role in stomatal closure to drought adaptation. These findings demonstrate the critical role of CPK15 in ABA-induced stomatal closure and its contribution to drought resilience.

OST1 Phosphorylates CPK15 at Thr103.

CPK15 is proposed to be a Ca²⁺-dependent protein kinase. We assessed its kinase activity by a luminescent adenosine diphosphate (ADP) detection assay (ADP-Glo assay) (47). Our results showed that His-tagged CPK15 phosphorylated the general substrate Histone III-S protein (Fig. 3A), and this phosphorylation was significantly enhanced upon the addition of Ca²⁺ (Fig. 3A), confirming that CPK15 is a functional Ca²⁺-dependent protein kinase. Mutation of the conserved Aspartic acid (Asp) 226 to Ala in the kinase domain (48) (SI Appendix, Fig. S4) almost completely abolished its kinase activity on Histone III-S protein (Fig. 3A), suggesting that CPK15^D226A is a “kinase-dead” form of CPK15. In contrast, OST1 exhibited Ca²⁺-independent activity, with K50R mutation eliminating its function (Fig. 3A), consistent with the previous report (18).

Fig. 3.

CPK15 is phosphorylated by OST1 and acts downstream of OST1 in ABA-induced stomatal movement and drought performance. (A) Kinase activities of His-CPK15, His-CPK15^D226A, GST-OST1, and GST-OST1^K50R by using the ADP-Glo method. Values are means ± SEM. (B) The phosphorylation of GST-OST1 on His-CPK15^D226A and of His-CPK15 on GST-OST1^K50R by in vitro kinase assays with phos-tag. (C) The peptide and site of CPK15 phosphorylated by OST1 identified via LC/MS-MS in two independent replicates. (D) The phosphorylation of GST-OST1 on His-CPK15^D226A and His-CPK15^D226A/T103A by in vitro kinase assays with phos-tag. GST-OST1 and His-CPK15 in (B and D) were detected by immunoblotting with anti-GST and anti-His antibodies, respectively. M, protein marker. Experiments were repeated three times. (E) Phosphorylation levels of CPK15 in CPK15–Flag transiently expressing Col-0 and ost1-3 protoplasts treated with 20 μM ABA for 1 h. CPK15–Flag was immunoprecipitated by Flag magnetic beads. CPK15–Flag protein and CPK15–Flag phosphorylation levels were detected with anti-Flag and anti-pS/T antibodies, respectively. M, protein marker. Experiments were repeated three times. (F) The ratio of phosphorylated CPK15 to total CPK15 protein levels in (E). Values are mean ± SEM (n = 3 independent experiments). (G) Time-resolved stomatal conductance of Col-0, ost1-3, cpk15-1, and ost1cpk15 in response to ABA. (H) The initial rates of stomatal conductance changes in the first 10 min of ABA treatment. Values are means ± SEM (n = 5 independent experiments, each with five leaves per genotype). (I) Time-resolved relative stomatal conductance of Col-0, ost1-3, and ost1-3 plants expressing CPK15, CPK15^T103A, or CPK15^T103D in response to ABA. n = 3 leaves per genotype per experiment. Values are mean ± SEM. Experiments were repeated three times. (J) Drought performance of Col-0, ost1-3, and ost1-3 plants expressing CPK15, CPK15^T103A, or CPK15^T103D. (Scale bar, 2 cm.) Experiments were repeated at least 3 times and the survival rates of different genotypes are depicted as means ± SEM. Different letters above the error bars in (A, F, and H) indicate the significant differences at P < 0.05 by One-way ANOVA with Tukey’s test.

We hypothesized that the interaction between OST1 and CPK15 may affect their phosphorylation status. Both CPK15 and OST1 exhibit autophosphorylation activity; therefore, we conducted in vitro Phos-tag kinase assays using GST-tagged OST1 on His-tagged CPK15^D226A, as well as His-tagged CPK15 on GST-tagged OST1^K50R. Our results showed that OST1 phosphorylated His-CPK15^D226A in the presence of ATP but not in its absence (Fig. 3B), while CPK15 failed to phosphorylate OST1^K50R, even in the presence of 10 μM CaCl₂ (Fig. 3B).

To identify the specific residues in CPK15 targeted by OST1 for phosphorylation, we analyzed the tryptic peptides from the phosphorylated His-CPK15^D226A bands using high-pressure liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS). Thr103 was consistently identified as the sole phosphorylation site in CPK15 by OST1 across two independent replicates; however, no phosphorylation sites were identified with the kinase-inactive OST1^K50R (Fig. 3C and Dataset S2). Further in vitro assays showed that mutating T103 in CPK15^D226A abolished OST1-mediated phosphorylation (Fig. 3D), suggesting that T103 in CPK15 is indeed a critical residue targeted for phosphorylation by OST1.

To validate the phosphorylation of CPK15 by OST1 in vivo, we transiently expressed CPK15–Flag in ost1-3 and Col-0 protoplasts to evaluate its phosphorylation level under ABA treatment. Under normal conditions, both Col-0 and ost1-3 plants showed low levels of CPK15 phosphorylation. However, ABA-induced CPK15 phosphorylation was significantly reduced in the ost1-3 mutant plants compared to Col-0 (Fig. 3 E and F). These results suggest that OST1 primarily phosphorylates CPK15 at Thr103 during ABA signaling.

The Phosphorylation of CPK15 at Thr103 Is Crucial for OST1 Regulation of ABA-Induced Stomatal Closure and Drought Performance.

To investigate the role of OST1 and CPK15 at the genetic level, we generated the ost1cpk15 double mutant by crossing ost1-3 and cpk15-1. Stomatal ABA responses in these mutants were assessed by measuring stomatal aperture and conductance. The ost1cpk15 double mutant exhibited greater ABA insensitivity than either single mutant (Fig. 3 G and H and SI Appendix, Fig. S5 A and B), indicating that OST1 and CPK15 function through overlapping yet independent mechanisms in ABA-induced stomatal closure.

To explore OST1-mediated phosphorylation of CPK15 in ABA-induced stomatal closure, we introduced wild-type CPK15, phospho-dead CPK15^T103A, or phosphomimetic CPK15^T103D driven by the CaMV 35S promoter in ost1-3. Two independent transgenic lines per variant with comparable expression levels were chosen for analyses (SI Appendix, Fig. S5C). Both stomatal aperture and conductance measurements showed that CPK15^T103D expression largely rescued the impaired stomatal response to ABA in ost1-3, whereas neither CPK15 nor CPK15^T103A expression could rescue this response (Fig. 3I and SI Appendix, Fig. S5 D–F).

Consistent with these findings, CPK15^T103D expression also significantly alleviated the drought hypersensitivity of ost1-3, while neither CPK15 nor CPK15^T103A exhibited such effect (Fig. 3J). These results suggest that phosphorylation at Thr103 is essential for the functional role of CPK15 downstream of OST1 in ABA-induced stomatal closure and drought tolerance.

CPK15 Interacts with SLAC1 and Phosphorylates the N Terminus of SLAC1.

Previous studies suggest that OST1 regulates ABA-induced stomatal closure pathway primarily by phosphorylating SLAC1 (9, 10, 19). Given that CPK15 acts downstream of OST1 in this pathway (Fig. 3), we hypothesized that OST1 may activate SLAC1 mainly through CPK15. Co-IP experiments in Arabidopsis protoplasts confirmed that SLAC1 interacts with CPK15 in vivo (Fig. 4A). BiFC analyses in N. benthamiana leaves further revealed that CPK15 binds to the N terminus of SLAC1 (SLAC1-N, 1-180 AA) but not to a truncated version lacking this region (SLAC1-ΔN) (Fig. 4B), indicating that their interaction depends on the N terminus of SLAC1.

Fig. 4.

CPK15 phosphorylates eight residues at the N terminus in SLAC1. (A) Coimmunoprecipitation of SLAC1 and CPK15 in Arabidopsis protoplasts expressing 35S–SLAC1–YFP with 35S–YFP or 35S–CPK15–FLAG. Protein extracts were immunoprecipitated with GFP-beads and detected with anti-GFP and anti-FLAG antibodies. IP, immunoprecipitation. (B) The interaction of CPK15 with SLAC1 and its truncated variants by BiFC assays. CPK15-nYFPand cYFP fused SLAC1, SLAC1-N, or SLAC1-ΔN were coexpressed in N. benthamiana leaf epidermis. The combination of PIP2;1 (PLASMA MEMBRANE INTRINSIC PROTEIN 2;1) and CPK15 was used as a negative control. BF, bright field. Merged, the merged images of YFP and BF. (Scale bar, 10 μm.) (C and D) The phosphorylation of GST-CPK15 on His-SLAC1-N without Ca²⁺ (C) and under various concentrations of Ca²⁺ (D) by in vitro kinase assays. SLAC1 phosphorylation was detected by anti-His antibody with phos-tag. M, protein marker. (E) The phosphorylation of GST-CPK15 and GST-CPK15ΔEF on His-SLAC1-N with and without 1 μM CaCl₂ by in vitro kinase assays. SLAC1 phosphorylation was detected by anti-His antibody with phos-tag, and GST-CPK15 and GST-CPK15ΔEF were detected with anti-GST antibody. M, protein marker. (F) Eight sites in SLAC1-N are phosphorylated by CPK15 via LC/MS-MS analysis in two independent replicates. (G and H) Phosphorylation of CPK15 on SLAC1-N and its variants (SLAC1-N^4A, SLAC1-N^5A, SLAC1-N^6A, SLAC1-N^7A-1, SLAC1-N^7A-2, and SLAC1-N^8A) by in vitro kinase assays. SLAC1 phosphorylation was detected by anti-His antibody with phos-tag. M, protein marker. Experiments in (A, C, D, E, G, and H) were repeated three times.

Using purified GST-CPK15 and His-SLAC1-N from Escherichia coli, we conducted in vitro Phos-tag kinase assays, which confirmed that CPK15 phosphorylated SLAC1-N (Fig. 4C). As CPK15 is a Ca²⁺-dependent protein kinase (Fig. 3A), reducing the CPK15 concentration (1/50 dilution) or omitting Ca²⁺ markedly decreased SLAC1-N phosphorylation, whereas increasing the concentration of Ca²⁺ enhanced it (Fig. 4D). It has been reported that Ca²⁺ typically activates CPKs via EF-hand domains (49). We purified EF-hand-deleted CPK15 (CPK15ΔEF, 1-360 AA) from E. coli and found that it phosphorylated SLAC1-N more strongly. However, additional Ca²⁺ did not further enhance this activity (Fig. 4E), suggesting that intracellular Ca²⁺ levels play a crucial role in this process.

LC-MS/MS analysis identified eight sites in SLAC1-N, S6, S55, S59, T62, S65, S86, S113, and S120, which were phosphorylated by CPK15 across two independent replicates. In contrast, no phosphorylation sites were detected with CPK15^D226A (Fig. 4F and Dataset S3). We then generated His-tagged SLAC1-N mutants with progressive alanine substitutions: SLAC1-N^S6A, SLAC1-N^4A (S6A/S55A/T62A/S86A), SLAC1-N^5A (S6A/S55A/T62A/S86A/S120A), SLAC1-N^6A (S6A/S55A/T62A/S86A/S113A/S120A), SLAC1-N^7A-1 (S6A/S55A/S59A/T62A/S86A/S113A/S120A), and SLAC1-N^8A (S6A/S55A/S59A /T62A/S65A/S86A/S113A/S120A). In vitro Phos-tag kinase assays revealed a stepwise reduction in CPK15 activity as the number of mutation sites increased (Fig. 4G). Among these sites, only S55 was newly identified, the other seven sites were previously reported (32–35). To assess the role of residue S55, we compared the phosphorylation status of SLAC1-N^S55A and SLAC1-N^7A-2 (all other seven known sites mutated) against wild-type SLAC1-N by in vitro Phos-tag kinase assays. No significant difference was observed for SLAC1-N^S55A, however, a substantial reduction in phosphorylation occurred for SLAC1-N^7A-2 compared to wild-type SLAC1-N (Fig. 4H). Further mutating residue S55 resulted in nearly undetectable phosphorylation in the mutant SLAC1-N^8A (Fig. 4H). These results suggest that all these eight residues within SLAC1-N are indeed targets of CPK15.

The Phosphorylation of SLAC1 at Eight Sites Is Essential for the Activation of the Channel Activity and ABA-Induced Stomatal Closure.

To explore this in vivo, we transiently expressed SLAC1–YFP in the protoplasts of both Col-0 and cpk15 mutants. ABA treatment induced the phosphorylation of SLAC1, however, this induction was significantly reduced in cpk15 mutants compared to Col-0 (Fig. 5 A and B). To further examine CPK15-mediated eight-site phosphorylation of SLAC1 in vivo, we coexpressed YFP-tagged SLAC1 (wild-type, individual site mutations, and an eight-site mutation) alongside CPK15–Flag in cpk15-1 protoplasts. ABA slightly increased SLAC1 phosphorylation in cpk15-1, which was dramatically enhanced by CPK15 coexpression. Variants with single phospho-site mutations displayed similar phosphorylation levels to wild-type SLAC1 upon ABA activation. In contrast, mutating all eight phosphorylation sites in SLAC1 (SLAC1^8A) completely blocked ABA-induced phosphorylation by CPK15 (Fig. 5 C and D). These findings confirm that CPK15 primarily mediates the phosphorylation of these eight N-terminal sites on SLAC1 during ABA treatment.

Fig. 5.

CPK15 activates SLAC1 channel activity by phosphorylation on eight sites in response to ABA. (A) SLAC1 phosphorylation in Col-0 and cpk15 protoplasts transiently expressing SLAC1–YFP treated with ABA (20 μM) or not. SLAC1–YFP was immunoprecipitated by GFP magnetic beads, and its protein and phosphorylation levels were detected by immunoblotting with anti-GFP and anti-pS/T antibodies, respectively. M, protein marker. (B) The ratio of phosphorylated SLAC1 to total SLAC1 protein levels in (A) from three independent replicates. (C) ABA-induced phosphorylation levels of SLAC1 and its site-mutation variants by CPK15 in cpk15-1. YFP-tagged SLAC1 variants were coexpressed with CPK15–Flag in cpk15-1 protoplasts, treated with 20 μM ABA for 1 h, then immunoprecipitated by GFP magnetic beads. Protein and phosphorylation levels were detected by immunoblotting with anti-GFP and anti-pS/T antibodies, respectively. M, protein marker. (D) The ratio of phosphorylated SLAC1 or SLAC1 variants to their corresponding protein levels in (C) from three independent replicates. Values are mean ± SEM. (E) Whole-cell voltage-clamp recordings of oocytes expressing SLAC1 or SLAC1^8A with or without CPK15 or CPK15ΔEF. (F) CPK15- and CPK15ΔEF-mediated SLAC1 or SLAC1^8A anion currents in oocytes are shown as in (E). Data are presented as means ± SEM. n = 10 oocytes recorded for each condition in the same batch. Different letters above the error bars in (B, D, and F) indicate the significant differences at P < 0.05 by One-way ANOVA with Tukey’s test.

We next assessed how these phosphorylation events affect SLAC1 channel activity using Xenopus laevis oocytes. SLAC1-mCherry localization at plasma membrane and its expression levels were unaffected by the eight-site mutations (SI Appendix, Fig. S6). Two-electrode voltage clamp (TEVC) experiments showed that infection of SLAC1-mCherry cRNA alone generated negligible currents. However, coinjection with CPK15-eGFP elicited typical SLAC1-like anion currents for SLAC1-mCherry but not for SLAC1^8A-mCherry (Fig. 5 E and F). Notably, CPK15ΔEF (lacking Ca²⁺-binding EF hands) induced stronger currents than wild-type CPK15 (Fig. 5 E and F), aligning with its enhanced phosphorylation activity on SLAC1 (Fig. 4E). These results suggest that CPK15 activates SLAC1 channel function by phosphorylating the eight sites in a Ca²⁺-dependent manner.

To dissect the contributions of individual site in SLAC1 to ABA-induced stomatal closure, we generated the transgenic plants expressing phosphomimetic variants of the eight individual phosphorylation sites (mutated within SLAC1^8A) in the slac1-3 mutant. Two independent transgenic lines for each variant, with comparable expression levels and growth phenotypes, were chosen for further study (SI Appendix, Fig. S7A). Measurements of both stomatal aperture and time-resolved stomatal conductance revealed that SLAC1^8A–YFP was unable to rescue the ABA insensitivity of slac1-3 (Fig. 6A and SI Appendix, Figs. S7B and S8A); in contrast, SLAC1^7A-T62D–YFP (7A-T62D), which carries mutations S6A/S55A/S59A/S65A/S86A/S113A/S120A/T62D to mimic T62 phosphorylation, significantly restored ABA-induced stomatal closure. The remaining seven phosphomimetic variants partially alleviated the slac1-3 phenotype (Fig. 6 B–D and SI Appendix, Figs. S7 C–E and S8 B and C). Further drought assessments showed that SLAC1^7A-T62D–YFP significantly restored drought tolerance of slac1-3 mutants, and SLAC1^7A-T55D–YFP conferred a measurable but less pronounced rescue effect relative to the phospho-dead variant SLAC1^8A–YFP (SI Appendix, Fig. S7F). These findings suggest that all eight CPK15-targeted sites contribute to ABA-induced stomatal closure, with T62 playing a predominant role.

Fig. 6.

Phosphorylation of SLAC1 by CPK15 plays a critical role in ABA-induced stomatal movement and drought performance. (A) Time-resolved relative stomatal conductance of Col-0, slac1-3, and slac1-3 expressing SLAC1 or SLAC1^8A transgenic plants in response to ABA. (B–D) Time-resolved relative stomatal conductance in response to ABA in Col-0, slac1-3, and slac1-3 transgenic plants expressing YFP-tagged SLAC1^7A-S6D (S55A/S59A/T62A/S65A/S86A/S113A/S120A/S6D), SLAC1^7A-S55D (S6A/S59A/T62A/S65A/S86A/S113A/S120A/S55D), SLAC1^7A-S59D (S6A/S55A/T62A/S65A/S86A/S113A/S120A/S59D), SLAC1^7A-T62D (S6A/S55A/S59A/S65A/S86A/S113A/S120A/T62D), SLAC1^7A-S65D (S6A/S55A/S59A/T62A/S86A/S113A/S120A/S65D), SLAC1^7A-S86D (S6A/S55A/S59A/T62A/S65A/S113A/S120A/S86D), SLAC1^7A-S113D (S6A/S55A/S59A/T62A/S65A/S86A/S120A/S113D), or SLAC1^7A-S120D (S6A/S55A/S59A/T62A/S65A/S86A/S113A/S120D). (E) Time-resolved relative stomatal conductance in response to ABA in Col-0, cpk15-1, and cpk15-1 expressing SLAC1, SLAC1^8A, or SLAC1^8D transgenic plants. Data in (A–E) are presented as means ± SEM. Experiments were repeated three times, each with 5 leaves per genotype for (A) and 3 for (B–E). (F) Drought performance of Col-0, cpk15-1, and cpk15-1 expressing SLAC1, SLAC1^8A, or SLAC1^8D transgenic plants. (Scale bar, 2 cm.) Experiments were repeated 3 times, and the survival rates of different genotypes are depicted as means ± SEM.

CPK15 Acts Upstream of SLAC1 in ABA-Induced Stomatal Closure and Drought Performance through Phosphorylation.

To investigate whether CPK15 regulates ABA-induced stomatal closure and drought performance through SLAC1, transgenic plants expressing YFP-fused SLAC1^8D, SLAC1^8A, and wild-type SLAC1 in the cpk15-1 mutant were generated (SI Appendix, Fig. S7G). Time-resolved stomatal dynamic response to ABA in the detached intact leaves (Fig. 6E and SI Appendix, Fig. S7H), changes in stomatal aperture in epidermal layers (SI Appendix, Fig. S8 D and E), and drought performance assessment (Fig. 6F) revealed that SLAC1^8D–YFP significantly restored the insensitive stomatal response to ABA and drought hypersensitivity of cpk15-1. In contrast, SLAC1–YFP only slightly alleviated these phenotypes compared to SLAC1^8D–YFP, while SLAC1^8A–YFP had no effect on these phenotypes (Fig. 6 E and F and SI Appendix, Figs. S7H and S8D). These results demonstrate that Ca²⁺-dependent protein kinase CPK15 primarily regulates ABA-induced stomatal closure and drought performance primarily by activating SLAC1 through phosphorylation at eight N-terminal sites.

Phosphorylation of CPK15 by OST1 Enhances Its Ability to Phosphorylate SLAC1.

Our results establish a phosphorylation cascade where Ca²⁺-independent protein kinase OST1 phosphorylates Ca²⁺-dependent protein kinase CPK15 at T103 (Fig. 3D), and CPK15 subsequently phosphorylates SLAC1 at eight sites to regulate ABA-induced stomatal movement (Figs. 4G and 6E). We compared the activity of recombinant CPK15 and CPK15^T103D proteins. ADP-Glo assays (47) revealed that CPK15^T103D exhibited higher kinase activity toward Histone III-S protein than wild-type CPK15, particularly at elevated Ca²⁺ concentrations. In contrast, the kinase-dead CPK15^D226A exhibited no kinase activity (Fig. 7A). Subsequently, we evaluated the phosphorylation activity of CPK15, CPK15^T103D, and CPK15^T103A toward SLAC1-N. Phos-tag kinase assays further revealed that when equal amounts were used, CPK15^T103D exhibited markedly enhanced kinase activity toward SLAC1-N compared to both CPK15 and CPK15^T103A, while CPK15^T103A showed only a slight reduction in kinase activity compared to CPK15 (Fig. 7B).

Fig. 7.

OST1 enhances the activity of CPK15 to activate SLAC1 anion channel. (A) The effect of Ca²⁺ on the kinase activity of CPK15, CPK15^T103D, and CPK15^D226A by using the ADP-Glo method. Values are means ± SEM (n = 3 independent replicates). Differences were performed by Student’s t test (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001). (B) The phosphorylation of GST-CPK15, GST-CPK15^T103A, and GST-CPK15^T103D on His-SLAC1-N by in vitro kinase assays. SLAC1-N phosphorylation was detected by anti-His antibody with phos-tag. M, protein marker. Experiments were repeated three times. (C) Whole-cell voltage-clamp recordings of oocytes expressing the specified cRNAs. (D) The anion currents of SLAC1 variants with or without mediation by CPK15 or CPK15^T103D are shown in (C). Values are means ± SEM. The number of oocytes recorded for each condition is indicated in the panel. (E) ABA-induced phosphorylation of SLAC1 in Col-0, cpk15-1, ost1-3, and ost1cpk15. SLAC1–YFP was transiently expressed in the protoplasts of these genotypes and treated with 20 μM ABA for 1 h. SLAC1–YFP was immunoprecipitated by GFP magnetic beads and detected by immunoblotting with anti-GFP and anti-pS/T antibodies, respectively. (F) The ratio of phosphorylated SLAC1 to total SLAC protein levels in (E). Values are mean ± SEM (n = 3 independent experiments). Different letters above the error bars in (D and F) indicate the significant differences at P < 0.05 by One-way ANOVA, Tukey’s test. (G) The module of OST1–CPK15 in ABA-induced stomatal closure in Arabidopsis. Upon ABA treatment, OST1 is released from inhibition by PP2Cs and activated through phosphorylation by B2/B3 Raf-like kinases. The activated OST1 orchestrates a dual regulatory mechanism: 1) phosphorylating CNGCs to increase cytosolic Ca²⁺ levels, and 2) phosphorylating CPK15 at Thr103 for activating. The elevated Ca²⁺ further amplifies the kinase activity of phosphorylated CPK15, which in turn phosphorylate eight sites on SLAC1 to activate the channel. Simultaneously, OST1 also directly phosphorylates S120 for additional regulation. This coordinated phosphorylation cascade drives anion efflux from guard cells, resulting in rapid stomatal closure. This OST1–CPK15 module acts as the predominant pathway for ABA-induced stomatal closure, effectively bridging Ca²⁺-dependent and -independent signaling mechanisms.

We also utilized oocytes to investigate the kinase activity of CPK15 phosphorylated by OST1 toward SLAC1. Voltage-clamp recordings showed that CPK15^T103D elicited stronger SLAC1 anion currents than wild-type CPK15. Notably, SLAC1^8D exhibited constitutive activation of anion current activity akin to that observed in SLAC1 activated by CPK15^T103D (Fig. 7 C and D and SI Appendix, Fig. S6). S120, which is phosphorylated by OST1 and crucial for SLAC1 activation (19), was examined. The phosphomimetic variant SLAC1^S120D generated small anion currents, exhibiting weaker activity than SLAC1 activated by CPK15, supporting the notion that S120 represents merely one of eight phosphorylation sites targeted by CPK15. Similarly, SLAC1^T62D showed intermediate current levels between CPK15-activated SLAC1 and SLAC1^S120D (Fig. 7 C and D and SI Appendix, Fig. S6). These findings provide further evidence that phosphorylation at eight sites by CPK15 collectively contributes to the activation of SLAC1.

To investigate the interaction of OST1 and CPK15 in activating SLAC1, we transiently expressed SLAC1–YFP in the protoplasts of Col-0, cpk15-1, ost1-3, and ost1cpk15. ABA-induced SLAC1 phosphorylation was diminished in ost1-3 and nearly abolished in cpk15-1 and ost1cpk15 double mutant (Fig. 7 E and F), mirroring their defects in stomatal ABA responses (Fig. 3 G and H). These findings suggest that OST1 enhances CPK15 kinase activity via Thr103 phosphorylation, enabling CPK15 to fully activate SLAC1 channel activity through multisite phosphorylation.

Discussion

ABA-induced stomatal closure involves coordinated action of both Ca²⁺-independent (OST1/SnRK2s) and Ca²⁺-dependent protein kinases (CPKs) to activate anion channels SLAC1/SLAHs. While OST1 initiates Ca²⁺ signaling through phosphorylation of CNGCs (39, 40), and specific CPKs (CPK3/4/6/11/27) regulate SnRK2s during osmotic stress responses (41), our study reveals Ca²⁺-dependent protein kinase CPK15 as a central node connecting these pathways. We demonstrate that OST1 phosphorylates CPK15 at Thr103, significantly enhancing its capacity to activate SLAC1 through multisite phosphorylation, thereby establishing the OST1–CPK15 module as an essential regulatory hub in ABA signaling.

Among the 34 CPKs in Arabidopsis, CPK15 displays unique functional properties. Unlike single or quintuple mutants of other CPKs (CPK3/4/5/6/11) that show only mild ABA response defects (25, 26, 28), or cpk21/23 single mutants that show enhanced drought resistance (30, 31), cpk15 single mutants exhibit severely impaired ABA-induced stomatal closure and drought hypersensitivity comparable to cpk3/4/5/6/11 quintuple mutants (Fig. 2 B–D). Notably, expression of CPK15 fully rescued these phenotypes (Fig. 2 B and D). Therefore, it can be concluded that CPK15 plays a pivotal role as a key regulator within the CPK family during ABA-induced stomatal closure.

SLAC1 activation requires coordinated phosphorylation at multiple N-terminal sites. Although S59, T62, S65, S86, S113, and S120 were previously identified as important residues (10, 33, 34, 36), very few sites have been functionally validated in planta. Our results demonstrate that CPK15 regulates ABA-induced stomatal closure by phosphorylating eight residues in the SLAC1 N terminus, including seven previously identified sites (S6, S59, T62, S65, S86, S113, and S120) (10, 19, 36, 50), and newly identified site S55. In vitro and in vivo assays confirmed that CPK15 targets S55; this phosphorylation contributes to ABA activation of the SLAC1 channel and subsequent stomatal closure (Figs. 4H and 6B). The phospho-dead variant SLAC1^8A lost ABA responsiveness in planta and channel activity in oocytes (Figs. 5 E and F and 6A). Although individual phosphomimetic variants partially restored the impaired stomatal ABA response and rescued the increased initial stomatal conductance of slac1-3 (Fig. 6 B–D and SI Appendix, Fig. S7 C–E), T62 emerged as the most critical site, as phosphomimetic SLAC1^7A-T62D showed strong functional complementation (Figs. 6 B–D and 7D) and SLAC1^T62D exhibited constitutive channel activation (Fig. 7D). These findings demonstrate that multisite phosphorylation, with T62 playing a central role, orchestrates SLAC1 channel activation and stomatal closure. Furthermore, CPK15 failed to phosphorylate SLAC1^8A or activate its anion currents during ABA treatment (Fig. 5 C, E, and F), indicating that CPK15 primarily regulates ABA-induced stomatal closure via SLAC1 by phosphorylating eight specific sites. However, SLAC1^8D could not fully rescue the stomatal ABA response in cpk15-1 (Fig. 6E). These findings imply that CPK15 may also phosphorylate SLAHs, which redundantly mediate ABA-induced stomatal closure alongside SLAC1.

The mechanism by which OST1 regulates SLAC1 has been controversial. Although it is established that OST1 phosphorylates SLAC1 at S120 in vitro (19), SLAC1^S120A fully complements the impaired ABA stomatal response of slac1-3 (37), suggesting the existence of additional regulatory mechanisms. Our findings reveal that OST1 activates SLAC1 primarily through phosphorylating CPK15 at Thr103, which enhances CPK15’s Ca²⁺-dependent kinase activity (Figs. 3 D and E and 7A). Phosphomimetic variant CPK15 (CPK15^T103D) significantly rescued the initial stomatal conductance, restored impaired ABA stomatal response, and alleviated drought sensitivity of ost1-3; however, neither wild-type nor phospho-dead CPK15^T103A produced such effects (Fig. 3 I and J). Furthermore, CPK15^T103D induced stronger SLAC1 currents than those elicited by OST1-phosphorylated SLAC1^S120D in oocytes (Fig. 7 C and D). In the ost1-3 mutant, ABA still activated CPK15 to a limited extent, suggesting that additional kinases may regulate CPK15 (Fig. 3 E and F). Potential regulators include OST1 homologs (e.g., SnRK2.2/2.3) or other kinase types such as CBL–CIPK complexes, which are known to participate in ABA and Ca²⁺ signaling pathways (51). Genetic analyses reveal that double mutations in both OST1 and CPK15 have additive effects on ABA-induced stomatal closure and SLAC1 phosphorylation (Figs. 3 G and H and 7 E and F). These results support the dual role of OST1 in SLAC1 activation, involving both direct phosphorylation as well as indirect phosphorylation mediated by CPK15.

We propose an updated model (Fig. 7G) in which ABA-activated OST1 performs dual functions in Ca²⁺-dependent pathway: 1) phosphorylating CNGCs to elevate cytosolic Ca²⁺ and 2) phosphorylating CPK15 at T103. The elevated Ca²⁺ further enhances the kinase activity of phosphorylated CPK15, which then activates SLAC1 through multisite phosphorylation, with T62 being particularly critical. Additionally, OST1’s direct phosphorylation of SLAC1 (e.g., at S120) through a Ca²⁺-independent pathway may provide additional fine-tuning. These regulatory mechanisms ensure rapid and robust stomatal closure under stress conditions. In conclusion, our findings identify CPK15 as a critical bridge between Ca²⁺-dependent and -independent pathways, functioning as both an OST1 substrate and a SLAC1 activator. The OST1–CPK15–SLAC1 regulatory axis illustrates how sophisticated kinase cascades integrate multiple signals to optimize plant stress responses.

Materials and Methods

Statistical Analysis.

Two-tailed unpaired Student’s t tests were used to evaluate the differences between two groups, and One-Way ANOVAs or Two-Way ANOVAs with Tukey’s tests in Prism v.9.5.1 (GraphPad) were used to evaluate the differences between multiple groups. P < 0.05 was assumed to be a statistically significant difference.

In Vitro and In Vivo Phosphorylation Assays.

For in vitro kinase assays, 5 μg of kinases (GST-OST1, GST-CPK15, His-CPK15, His-CPK15^T103D) and 20 μg of substrates (GST-OST1^K50R, His-CPK15^D226A, His-CPK15^D226A/T103A, or His-SLAC1-N) were incubated in a 50 μL reaction buffer containing 25 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, 1 mM DTT, 0.4 mM ATP, and 5 mM MnCl₂ at 28 °C for 2.5 h. The kinase products were desalted through chloroform/methanol precipitation and dissolved in SDS sample buffer for Phos-tag SDS-PAGE using an 8% resolving gel with 50 μM Phos-tag (Wako Chemicals) and 60 μM MnCl₂.

For in vivo kinase assays, the CPK15–Flag plasmid was transformed into Col-0, cpk15-1, and ost1-3 protoplasts, SLAC1–YFP and its variants were transformed into Col-0, cpk15-1, cpk15-2, ost1-3, and ost1cpk15 protoplasts. After overnight incubation at 22 °C with or without 20 µM ABA for 1 h, the protoplasts were collected and lysed in a buffer [50 mM Tris-HCl, pH 7.5, 2% SDS (w/v), 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10% glycerol (v/v), 2 mM NaF, 2 mM Na₂MoO₄, 0.5 mM EDTA, and 1 mM PMSF]. Flag-beads or GFP-beads were added to the supernatants to assess the phosphorylation levels of CPK15 or SLAC1 by western blotting using phosphoserine/threonine antibody (ECM, PP2551,1:1,000).

Further detailed experimental methods can be found in SI Appendix.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Dataset S02 (XLSX)

Dataset S03 (XLSX)

Data, Materials, and Software Availability

The sequence data mentioned in this paper can be accessed in the Arabidopsis Genome Project or the GenBank database using the provided accession numbers: OST1 (AT4G33950) (52), SLAC1 (AT1G12480) (53), CPK4 (AT4G09570) (54), CPK5 (AT4G35310) (55), CPK6 (AT2G17290) (56), CPK11 (AT1G35670) (57), CPK15 (AT4G21940) (58), CPK21 (AT4G04720) (59), CPK23 (AT4G04740) (60), and CPK32 (AT3G57530) (61). All other data are included in the manuscript and/or supporting information.