The Arabidopsis thaliana K+ channel KAT1 has been suggested to have a key role in mediating the aperture of stomata pores on the surface of plant leaves. Although the activity of KAT1 is thought to be regulated by phosphorylation, the endogenous pathway and the primary target site for this modification remained unknown. In the present study, we have demonstrated that the C-terminal region of KAT1 acts as a phosphorylation target for the Arabidopsis calcium-independent ABA (abscisic acid)-activated protein kinase SnRK2.6 (Snf1-related protein kinase 2.6). This was confirmed by LC-MS/MS (liquid chromatography tandem MS) analysis, which showed that Thr306 and Thr308 of KAT1 were modified by phosphorylation. The role of these specific residues was examined by single point mutations and measurement of KAT1 channel activities in Xenopus oocyte and yeast systems. Modification of Thr308 had minimal effect on KAT1 activity. On the other hand, modification of Thr306 reduced the K+ transport uptake activity of KAT1 in both systems, indicating that Thr306 is responsible for the functional regulation of KAT1. These results suggest that negative regulation of KAT1 activity, required for stomatal closure, probably occurs by phosphorylation of KAT1 Thr306 by the stress-activated endogenous SnRK2.6 protein kinase.
The aerial epidermal tissues of higher plants contain stomatal pores that play an important role in enabling gas exchange with the external environment, providing CO2 for photosynthesis and releasing O2. On the other hand, these pores also permit water loss through transpiration, and the stomatal aperture must therefore be tightly controlled according to environmental conditions. Stomatal aperture is determined by the turgidity of two bordering guard cells, and stomatal closing is mediated by the release of ions and organic solutes from the guard cells, together with concomitant inhibition of K+ influx [1–7]. At least six Shaker-type (voltage-dependent) K+ channels are known to be expressed in the guard cells of Arabidopsis thaliana: KAT1, KAT2, AKT1, AtKC1, AKT2, and GORK [8–11]. Among them, KAT1 is the most characterized channel with respect to its structure and function [10,12,13]. KAT1 has six transmembrane domains, containing a voltage sensor at the N-terminus, and a relatively long cytosolic region, consisting of more than 360 amino acids, at the C-terminus . The inward-rectifying K+ channel KAT1 is of great interest as it has been suggested to have a key role in stomatal opening , and the inhibition of KAT1 channel activity is understood to be one of the requirements to enable stomatal closure .
Stomatal closure occurs in response to unfavourable environmental conditions, such as drought stress. The stress-related plant hormone ABA (abscisic acid) is known to act on guard cells early in a signal transduction pathway that leads to reduced turgidity and hence stomatal closure. The regulation of potassium and anion channels in response to ABA is thus essential for the stomatal closing event [6,16–19]. Arabidopsis has ten members of the SnRK2 (Snf1-related protein kinase 2) family . One of these, SRK2E/OST1/SnRK2.6 (hereafter referred to as SnRK2.6), has been identified as a key component of ABA signal transduction in guard cells [21,22]. Inactivation of SnRK2.6 results in loss of the ABA-induced stomatal closure response [21–23]. The ABI1 PP2C (protein phosphatase 2C) regulates SnRK2.6 activation, and has been shown to interact with SnRK2.6 in an ABA-dependent manner in response to low humidity . These results show that the SnRK2.6 protein kinase is a positive regulator of stomatal closure in guard cells.
The amino acid sequence of SnRK2.6 shows high identity (79%) with the Vicia faba AAPK (ABA-activated protein kinase)/ABR (ABA-responsive) protein kinase, which is also activated by ABA [25,26]. AAPK/ABR was specifically activated by pretreatment of V. faba guard cell protoplasts with ABA, but not by pretreatment with the other plant hormones IAA (indoleacetic acid), 2,4-D (2,4-dichlorophenoxy acetic acid), kinetin or GA3 (gibberelic acid). AAPK/ABR was also insensitive to Ca2+ [26–28], and a mutant AAPK/ABR eliminated the ABA activation of slow anion channels in V. faba guard cells . We demonstrated previously that activated AAPK/ABR was able to phosphorylate the Arabidopsis KAT1 C-terminal region in a V. faba protoplast system, although the identity of the specific KAT1 target sites and the significance for channel activity remained to be determined .
Despite the known roles for the SnRK2.6 protein kinase and the KAT1 K+ channel in stomatal closure, it was not clear how these two activities are linked. To further understand the interaction between SnRK2.6 and KAT1, we have evaluated whether SnRK2.6 is able to regulate KAT1 activity by phosphorylation of specific residues in the C-terminus using biochemical and electrophysiological approaches. Phosphorylation assays confirmed an interaction between SnRK2.6 and KAT1, and resulted in identification of target phosphorylated residues in KAT1. The functional significance of this phosphorylation was also demonstrated using Xenopus oocyte and yeast systems. Together, these results demonstrate a direct link between SnRK2.6-mediated phosphorylation and KAT1 activity, suggesting this is an important step for ABA-induced signal transduction of stomatal closure.
Construction and purification of recombinant KAT1 peptides
His6-tagged peptides of the C-terminal region of KAT1 were constructed as follows. A plasmid containing full-length KAT1 served as a template for PCR amplification. All KAT1 peptide fragments were amplified by PCR using a NdeI-site-containing sense primer, and a BamHI-site-containing antisense primer. The NdeI–BamHI fragments were ligated into the equivalent sites of the pET16b vector (Novagen). Escherichia coli BL21 harbouring the recombinant plasmids were grown under ampicillin selection in LB (Luria–Bertani) medium. Cultures, grown overnight at 37 °C, were diluted 100-fold and grown at 37 °C to a D600 of 0.6–0.7. IPTG (isopropyl β-D-thiogalactoside) was then added to a final concentration of 0.5 mM and the cultures were grown for a further 12 h at 25 °C. All subsequent purification steps were carried out at room temperature (25 °C). A 200 ml aliquot of the culture was centrifuged for 5 min at 3860 g, and the pellet was resuspended in 15 ml of lysis buffer (20 mM Tris/HCl, pH 8.0, 0.1 M NaCl and 6 M urea). Resuspended cells were then disrupted by sonication and cell extracts centrifuged at 100000 rev./min for 15 min using a Biosafe Optima TLX (Beckman Coulter). The resulting supernatant was transferred to a Ni-NTA (Ni2+-nitrilotriacetate) agarose column (bed volume, ~2 ml). The column was washed with 5 column vols of wash buffer (20 mM Tris/HCl, pH 8.0, 6 M urea, 0.1 M NaCl and 10 mM imidazole) and then the KAT1 peptides were eluted with elution buffer (wash buffer with 100 mM imidazole). Proteins were concentrated using Amicon Ultra (Millipore) or Vivaspin2 ml (10 kDa molecular-mass cut-off; Vivascience) concentrators and stored at −20 °C.
In-gel and in vitro kinase assays
In-gel kinase assays using protein extracts from Arabidopsis T87 cells expressing 35S::SnRK2.6-GFP or 35S::GFP (where GFP is green fluorescent protein) were performed as described previously . Purification of recombinant SnRK2.6 and in vitro kinase assays with KAT1 peptides were performed as described by Belin et al. . The in vitro kinase assays were performed for 45 min at room temperature with 100 ng of the purified SnRK2.6 and 136 ng of the recombinant KAT1 peptides or histone III-S substrate (Sigma) in 12.5 μl of buffer [20 mM Hepes, pH 7.5, 0.5% (v/v) Triton X-100, 2 mM MnCl2, 1× protease inhibitor cocktail (Roche), 10 mM NaF and 5 mM β-glycerophosphate] with 5 μCi of [γ-32P]ATP (3000 Ci/mmol). Reactions were stopped by the addition of 12.5 μl of sample buffer containing 100 mM EDTA and heating at 95 °C for 5 min . Proteins were analysed by SDS/PAGE (16.5% gel) and autoradiography.
Determination of in vitro phosphorylation sites by LC-MS/MS (liquid chromatography tandem MS)
In vitro kinase reactions were performed for 45 min at room temperature with 200 ng of the purified SnRK2.6 and 400 ng of the recombinant KAT1 peptides in 20 μl of buffer as described above, except that 1 mM ATP was used instead of [γ-32P]ATP. The reaction was stopped by the addition of 20 μl of sample buffer and an incubation at 95 °C for 5 min, then proteins were separated by SDS/PAGE (16.5% gel). Gel fragments containing protein bands corresponding to recombinant KAT1 peptides were excised from the CBB (Coomassie Brilliant Blue)-stained gels and washed three times with HPLC-grade water containing 30% acetonitrile (Kanto Chemical) and 50 mM ammonium bicarbonate. After alkylation with dithiothreitol and iodoacetamide, samples were digested with 0.1 μg of trypsin (Promega) in 10 μl of 50 mM ammonium bicarbonate for 16 h at 37 °C. The digested peptides in the gel pieces were then extracted twice with 20 μl of 5% (v/v) formic acid/50% (v/v) acetonitrile. Phosphopeptides were enriched using Titansphere Phos-Tio Kit (GL Sciences) and loaded on to a Bridge BEH130 C18 column (3.5 μm, 75 μm internal diameter, 150 mm length; Waters) using the CapLC system (Waters). LC-MS/MS analysis was performed according to Fujiwara et al. . The buffers ued were 0.1% acetic acid in 5% (v/v) acetonitrile (buffer A) and 0.1% acetic acid in 95% (v/v) acetonitrile (buffer B). A linear gradient from 5% to 45% of buffer B for 25 min was applied, and the peptides eluted from the column were introduced directly into a Q-TOF Ultima mass spectrometer (Waters) with a flow rate of 200 nl/min. Ionization was performed with a potential of 2200 V applied to the PicoTip nanospray source (New Objective). For survey scans, MS spectra were acquired for the m/z range of 300–1700 and MS/MS spectra were acquired for the two most intense ions from the precursor ion scan. For CID (collision-induced dissociation), the collision energy was set automatically according to the mass and charge state of the precursor peptides. MS/MS spectra were subjected to the MASCOT server (version 2.1, Matrix Science) against a protein database from the NCBI (National Center for Biotechnology Information). For database searches, the following parameters were used: peptide tolerance at ±0.8 Da, and MS/MS tolerance at ±0.6 Da; peptide charge of 2+ or 3+; trypsin as the enzyme, allowing up to four missed cleavages; carbamidomethylation of cysteine residues as a fixed modification and oxidation of methionine residues as a variable modification.
Channel expression in oocytes and current recording
DNA fragments encoding full-length wild-type and mutant KAT1 peptides were amplified by PCR using a HindIII-site-containing sense primer, and a BamHI-site-containing antisense primer. The HindIII–BamHI DNA fragments were ligated into the equivalent sites of modified pYES2 vector (Invitrogen) for expression in oocytes and yeast . Whole-cell currents were measured using the two-electrode voltage-clamp technique at room temperature in Xenopus laevis oocytes as described previously [34,35]. Two-electrode voltage-clamp experiments were performed using a voltage-clamp amplifier (AxoClamp 2B; Axon Instruments). The microelectrodes contained 3 M KCl with a resistance of 0.3–1.0 MΩ. The bath solution was 120 mM KCl, 1 mM MgCl2, 1 mM CaCl2 and 10 mM Hepes (pH 7.3). Step voltage pulses (−30 to −170 mV with a 20 mV decrement) were applied from a holding potential of −40 mV, the duration of each pulse being 500 ms. Data acquisition and analysis were performed using pCLAMP 9.2 (Molecular Devices) and Origin 5.0 software (Axon Instruments).
Yeast transformation and growth conditions
Yeast transformations were carried out as described previously [35,36]. Transformed yeast cells were selected on a minimal solid medium [YNB (yeast nitrogen base) medium containing 100 mM KCl, 0.7% YNB without amino acids, 2% sucrose, 2% galactose, the required amino acids except uracil and 2% agarose]. YNB medium and uracil-free arginine:phosphate synthetic medium, similar to those described previously by Rodriguez-Navarro and Ramos , were used for complementation tests. Uracil-free arginine:phosphate synthetic medium contained 2% sucrose, 2% galactose, the required amino acids except uracil, 10 mM L-arginine, 2 mM MgSO4, 0.2 mM CaCl2, oligoelements (8 μM H3BO3, 0.25 μM CuSO4, 0.6 μM KI, 2.7 μM MnSO4, 1 μM Na2MoO4, 1 μM ZnSO4, 0.5 μM CoCl2 and 0.5 μM NiCl2), vitamins (2 μg/l biotin, 400 μg/l calcium panthenate, 2 μg/l folic acid, 2 mg/l inositol, 400 μg/l niacin, 400 μg/l pyridoxine hydrochloride, 200 μg/l riboflavin and 400 μg/l thiamine hydrochloride) and 2% agarose, and the pH was adjusted to 5.9 with orthophosphoric acid.
Measurement of K+ uptake in yeast
K+ uptake in Saccharomyces cerevisiae CY162 containing the KAT1 wild-type and mutant constructs were measured as described previously . Cells were grown to a D600 of 1–2 in YNB medium containing 2% sucrose and 2% galactose supplemented with 100 mM KCl and the required amino acids, but lacking uracil. Cells were harvested by centrifugation and washed twice in starvation buffer (50 mM Tris/succinate, pH 5.9). K+ starvation was achieved by incubating the cells in the starvation buffer for 6 h at 30 °C. The cells were washed once with lowsalt buffer and the concentration adjusted to a D600 of 1 with the same buffer containing 2% glucose. KCl (15 mM) was added to the cell suspension at 0 min. The net uptake of K+ was measured by the silicone filtration technique [38,39], and the K+ content of the cell pellets was determined by flame photometry .
KAT1 peptide is phosphorylated by ABA-activated SnRK2.6
On the basis of the fact that KAT1 and the SnRK2.6 protein kinase are both endogenous to Arabidopsis and localize to the same cell type to regulate stomata aperture, we first examined whether SnRK2.6 was able to directly phosphorylate KAT1. In-gel protein kinase assays were performed using extracts of Arabidopsis T87 cultured cells, which overexpressed the SnRK2.6–GFP fusion proteins and were treated with ABA or NaCl to activate SnRK2.6 [21,24]. In these assays, gels were treated with either the entire KAT1 C-terminal region from the His301 residue immediately after the end of the S6 transmembrane segment (His301–Asn677, His-tagged peptide), or with a shorter peptide from the Tyr376 residue, including the putative cyclic-nucleotide-binding domain (Tyr376–Asn677) (Figure 1A). Furihata et al.  had confirmed previously that autophosphorylation signals for the SnRK2-type kinases were not present at a level to interfere with this analysis in parallel in-gel experiments. In contrast, clear phosphorylation signals were observed for ABA- and NaCl-activated SnRK2.6–GFP when exposed to the histone III-S substrate as a positive control, whereas no background signal was observed for non-activated SnRK2.6–GFP (Figure 1B). A strong phosphorylation signal was also observed when ABA-activated SnRK2.6–GFP was treated with the KAT1(His301–Asn677) substrate. In contrast, only very weak phosphorylation was observed for the KAT1(Tyr376–Asn677) substrate. These results suggested that activated SnRK2.6 was able to phosphorylate at least the C-terminal region of KAT1, most probably between His301 and Tyr376.
ABA- and NaCl-stress-activated SnRK2.6 phosphorylated the C-terminal region of KAT1
Mutational analysis suggested that Thr306 of KAT1 is a target site for recombinant SnRK2.6
To further examine the phosphorylation of the KAT1 C-terminal region by SnRK2.6, we adopted an in vitro kinase assay approach using the His-tagged recombinant SnRK2.6 protein that had been purified in an activated state from E. coli . As the kinase reaction is subsequently size-fractionated, this approach enables increased specificity and autophosphorylation signals to be distinguished from substrate phosphorylation signals. Two additional KAT1 C-terminal peptides were included in this assay in an attempt to narrow down the target region (Figure 2A). SnRK2.6 activity in these reactions was confirmed by the presence of an SnRK2.6 autophosphorylation signal at approx. 50 kDa (Figure 2B), which was inactivated by the G33R mutation . Consistent with that shown in Figure 1, the KAT1 (His301–Tyr376) peptide showed distinct phosphorylation, whereas very weak phosphorylation of the KAT1 (Tyr376–Asn677) peptide was observed (Figure 2B). Further truncation of the KAT1 (His301–Tyr376) peptide to His301–Arg360 resulted in an increase in phosphorylation signal, whereas removal of the first 17 residues (Glu318–Tyr376) resulted in loss of the phosphorylation signal. The above data suggested that SnRK2.6 targeted the His301–Ser317 region of KAT1. Due to the fact that SnRK2.6 belongs to the serine/threonine protein kinase family , we next focused on the serine or threonine residues within this region (Figure 3A). A series of mutant peptides based on His301–Arg360 were constructed containing single- or multiple-site replacement of serine or threonine residues with an alanine residue in the His301–Ser317 region. As shown in Figure 3(B), phosphorylation signals were still observed in peptides containing up to five mutations (Thr303, Ser304, Thr308, Ser312 and Ser317). The only mutant peptide to show clear reduction in phosphorylation signal was that containing the T306A mutation, strongly suggesting that Thr306 was the primary target amino acid for SnRK2.6 activity.
Phosphorylation of the C-terminal region of KAT1 peptide by purified recombinant SnRK2.6 kinase
Thr306 of KAT1 is a target site of recombinant SnRK2.6
Identification of Thr306 and Thr308 phosphorylation by SnRK2.6 in LC-MS/MS analysis
Although Thr306 of KAT1 was suggested to be the primary target for SnRK2.6, the introduction of amino acid substitutions can change peptide conformation and accessibility of the kinase. We therefore applied LC-MS/MS analysis to confirm the Thr306 phosphorylation and to identify other phosphorylated residues in the original His301–Tyr376 peptide (Figure 4). The His301–Tyr376 KAT1 peptide was phosphorylated in an in vitro kinase reaction with SnRK2.6, digested in-gel with trypsin and subjected to LC-MS/MS analysis. Two phosphorylated sites, Thr306 and Thr308, were found by a combination of exact mass measurement of precursor ions and observation of neutral loss (H3PO4) from these ions (Figure 4). Taken together with the results shown in Figure 3(B), this confirmed that SnRK2.6 targets Thr306 of KAT1 and can also phosphorylate Thr308.
Identification of Thr306 and Thr308 in the C-terminal region of KAT1 as phosphorylation sites for recombinant SnRK2.6
Functional significance of Thr306 modification for KAT1 function
On the basis of the results of the kinase assays, the functional significance of the threonine residues at positions 306 and 308 in the C-terminal region of KAT1 was investigated. Wild-type and mutant full-length KAT1 peptides were expressed in Xenopus oocytes and two-electrode voltage-clamp measurements were carried out to determine K+ ion conduction properties. Wild-type KAT1 channel showed the expected K+ conductance, which was slightly reduced by mutation of Thr308 (Figure 5). In contrast, modification of Thr306 to Ala, Asp, Glu, Asn or Gln resulted in loss of channel activity and no detectable K+ current amplitude. Interestingly, the T306S mutation also abolished the activity. The significant loss of K+ current by the substitution of Thr306 strongly suggested that Thr306 is essential for activity of the KAT1 K+ channel in this system.
Mutation of Thr306 impaired the function of KAT1 in Xenopus oocytes
The above results showed that any modification of the side chain of Thr306 conferred lack of the channel activity. To investigate the functional significance of Thr306 with greater sensitivity, we took advantage of a yeast complementation system that occurs over a longer time scale and enables measurement even when channel activities are greatly reduced . Six mutant forms of the full-length KAT1 channel protein, T306A, T306D, T306E, T306N, T306Q and T306S, were tested using the yeast mutant strain S. cerevisiae CY162 that lacks the K+ transporters TRK1 and TRK2 . In this system, the T306A, T306N and T306S mutant KAT1 proteins were able to complement the K+-uptake mutation phenotype of yeast strain CY162 on solid medium containing either 0.2 mM or 7 mM KCl, suggesting these did retain at least some K+-uptake activity (Figure 6A). In contrast, yeast cells expressing the T306D, T306E and T306Q mutants were not able to grow on medium containing 0.2 mM KCl (Figure 6A). It should be noted that mutation of threonine to asparagine can be expected to mimic the phosphorylation state , and thus lack of growth by the T306D mutant is consistent with an impairment of KAT1 channel activity by Thr306 phosphorylation.
Effect of amino acid substitution of Thr306 on K+-uptake function in yeast
To further characterize the K+-uptake activity of the mutant KAT1 proteins, the K+-uptake rate of the yeast strain CY162 expressing KAT1 mutant forms was measured following growth under K+-depleted conditions. The yeast strain CY162 cells expressing the KAT1 mutants were cultured in yeast starvation buffer (20 mM Tris/succinate, pH 5.9) and then suspended in low-salt buffer in the presence of glucose. Addition of 15 mM KCl led to rapid and extensive net K+ uptake in cells expressing the wild-type KAT1, but not by the cells containing empty vector (Figure 6B). The measurement of the K+-uptake rate by yeast expressing the KAT1 variants showed that all of the variants exhibited a decreased K+-uptake rate compared with wild-type. This was consistent with the Xenopus voltage-clamp results shown in Figure 5, and as observed on solid medium, the T306D, T306E and T306Q mutants showed the greatest reduction in K+ uptake.
In order to determine the contribution of serine and threonine side chains around Thr306 on the K+ transport activity of KAT1, mutants were constructed containing Thr303, Ser304, Thr308, Ser312 or Ser317 replaced with an alanine or asparagine residues. The S304A mutant did not grow on the selective medium containing 0.2 mM and 7 mM KCl, which was similar to the T306D mutant (Figure 7A), whereas the S304D mutant and other mutants were able to grow on the 0.2 mM KCl medium (Figure 7A). It should be noted that the T308A and T308D mutants did not show any clear effect on the K+-uptake-dependent growth. Consistent with this spot assay, measurement of the K+-uptake rate by yeast cells expressing the S304A mutant revealed negligible net K+-uptake activity, comparable with cells containing the empty vector (Figure 7B). These results suggested modification of the Thr306 side chain regulates K+ permeability of KAT1 and that a polar side chain of the residue at position 304 is required for this activity.
Complementation test of K+-uptake-deficient yeast strain CY162 by KAT1 variants
The closure of stomatal pores on plant leaves during drought stress involves a signal transduction pathway induced by the stress-responsive hormone ABA. In the present study we show that the KAT1 inward-rectifying K+ channel, whose inhibition is required for stomatal closure, is phosphorylated by the ABA-activated SnRK2.6 protein kinase. Moreover, we identified the specific residue targeted by SnRK2.6 as Thr306 in the KAT1 C-terminus and found that modification of this residue can explain the inhibition of KAT1 channel activity. To our knowledge, this is the first report identifying a specific KAT1 residue targeted for phosphorylation by a specific kinase, SnRK2.6, whose modification also led to a down-regulation of the channel activity.
Although there is accumulating evidence that inward-rectifying K+ channels are closely involved in stomatal opening in Arabidopsis guard cells [5,9,15,42], there is little information on the involvement of inward-rectifying K+ channels in the closure of stomata. The shrinkage of guard cells by reduced turgidity requires a concerted regulation of ion-efflux channel activation and inhibition of ion-influx systems , such as inward-rectifying K+ channels. ABA-induced closing of stomata are abolished by kinase inhibitors and enhanced by phosphatase inhibitors, indicating the importance of phosphorylation/dephosphorylation networks to modulate channel activity in this process [16,17,43–47]. At least five Shaker-type K+ channels, KAT1, KAT2, AKT1, AKT2 and AtKC1, contribute to the inward K+ channel activity in Arabidopsis guard cells . Heteromeric assembly of these proteins is also suggested to increase the functional and regulatory diversity of guard cell K+ channel activity [49–53]. KAT2 has been identified as an important component of the stomatal opening mechanism [50,54]. Although inactivation of KAT1 did not induce stomata opening , overexpression of a dominant-negative KAT1 mutant did inhibit light-induced stomatal opening . These phenomena may reflect the assembly of KAT1 and KAT2 as heterotetramers in guard cells . For stomata closure, the regulatory mechanism remains elusive, but probably involves suppression of the inward current activity.
There are reports suggesting that KAT1 activity can be modulated by phosphorylation in a Xenopus oocyte expression system. Co-injection of KAT1 cRNA into oocytes with transcripts extracted from V. faba guard cells decreased KAT1 channel activity, unlike that with transcripts from mesophyll cells . In the same heterologous system, KAT1 current amplitudes decreased in the presence of soya-bean CDPK (calcium-dependent protein kinase) . The calcium-independent ABA-activated AAPK/ABR kinase from V. faba [25,26,29] was also shown to phosphorylate the C-terminal region of KAT1, although this analysis did not include the Thr306 residue and only the region from Arg322 to Asn677 of KAT1 was tested . In Arabidopsis, from which KAT1 is derived, SnRK2.6 is an ABA-activated protein kinase that has been revealed to play a key role in signalling stomatal closure [21–23]. The above findings that SnRK2.6 directly phosphorylates Thr306 of KAT1, and that modification of Thr306 inhibits KAT1 function, provides a new mechanistic link between ABA-induced kinase activation and physiological closing of stomata in response to environmental conditions.
The C-terminal region, after the last transmembrane region in prokaryotic and eukaryotic K+ channels, is responsible for the channel gating [57–60]. Consistent with this, mutation of Asn297 and Val299 were shown to drastically alter KAT1 channel properties , suggesting that the KAT1 C-terminal region immediately after the S6 transmembrane domain is also highly sensitive to conformational change for channel gating. The observed hypersensitivity to modification of Thr306 and loss of activity upon mutation of Ser304 to a non-polar residue therefore probably reflects manipulation of KAT1 channel gating. Similar results have been reported for the weakly inward-rectifying K+ channel, AKT2, which interacts with AtPP2CA . Phosphorylation/dephosphorylation of Ser329 in AKT2, which might be recognized by AtPP2CA (corresponding to position Ser312 in KAT1, see Supplementary Figure S1 at http://www.BiochemJ.org/bj/424/bj4240439add.htm), was proposed to regulate the switching of AKT2 from an instantaneous to a time-dependent mode .
Following a systematic approach to evaluate target sites for plant kinases, SnRK2 protein kinases were found to recognize and phosphorylate serine/threonine residues in the consensus sequence R-X-X-S/T [40,64]. Moreover, SnRK2 kinases displayed a preference for hydrophobic amino acids in the target site, particularly a leucine residue in the sequence L-X-R-X-X-S/T, and this amino acid was consistently present in peptides from Arabidopsis and rice b-ZIP transcription factors that were phosphorylated efficiently by SnRK2 [40,65]. KAT1 contains nine R-X-X-S/T consensus sequences. Five of these are located in the transmembrane segments and the remaining four are present in the C-terminal cytosolic regions. In the C-terminal region of KAT1, Thr308, Ser317 and Ser578 conform to the R-X-X-S/T consensus, whereas Ser641 matches the L-X-R-X-X-S/T sequence. The identification of Thr308 as a phosphorylation target of SnRK2.6 is in agreement with this rule. A previous report showing that residues between Arg322 and Asn677 of KAT1 were phosphorylated by a V. faba kinase similar to SnRK2.6 can also be explained by this rule . However, our results suggested that neither of the above four residues were the primary phosphorylation target site for SnRK2.6. Rather, SnRK2.6 was found to primarily target Thr306 of KAT1, even though it does not match the consensus sequence. The role of Thr306 as the primary target is supported by the fact that it was also confirmed to be functionally relevant to KAT1 channel activity. These findings raise the possibility that other members of the SnRK2 family of protein kinases are also able to recognize Ser/Thr phosphorylation target sites outside of the R-X-X-S/T consensus sequence.
The plant inward-rectifying K+ channel AKT1 functions in plant roots to take up K+ from the external space. AKT1 is positively regulated by phosphorylation, which is mediated by the CBL1/9 (calcineurin B-like protein 1/9)–CIPK23 (CBL-interacting protein kinase 23) complex [66,67]. The SOS cation transport complex, comprising Na+/H+ antiporter (SOS1), CBL4 (SOS2) and CIPK24 (SOS3) in Arabidopsis, is also positively regulated by phosphorylation . Unlike that for AKT1 and the SOS complex, our data provides evidence of negative regulation of a K+ channel in a phosphorylation transduction pathway. Thr306 is highly conserved amongst Shaker-type K+ channels in Arabidopsis, whereas Thr308 is not (Supplementary Figure S1). Both the SnRK2 family and Shaker-type K+ channels are expressed in diverse cell types in plants [12,22,69,70]. Similar functional examination of phosphorylation at residues corresponding to Thr306 in other Shaker-type K+ channels may provide new insights into regulatory mechanisms of K+ channels by SnRK2 family protein kinases. On the other hand, ABA signal transduction in guard cells entails highly complicated mechanisms and many components that participate in the pathway are still missing [16,17,43–47,71]. At present, it thus remains difficult to identify the specific timing and conditions required for in vivo physiological analysis of the SnRK2.6 and KAT1 interaction and Thr306 phosphorylation.
It is important to note that SnRK2.6-mediated phosphorylation would not be exclusive, and additional regulation of KAT1 probably exists during stomatal closure. Both calcium-dependent and -independent signalling pathways are present in the ABA-mediated regulation of guard cell turgidity. A calcium-dependent kinase from the V. faba bean was found to phosphorylate the KAT1 protein translated in vitro , and a calcium-dependent kinase from soya bean reduced KAT1-mediated K+ current amplitudes in Xenopus oocytes . Moreover, it has been reported that the exogenously supplemented animal protein kinase A avoided the rundown of KAT1 expressed in oocytes [72,73], although Thr306 is not predicted to be a protein kinase A target residue (http://www.cbs.dtu.dk/services/NetPhosK/). Our results provide an important insight into specific phosphorylation of KAT1 and functional relevance to channel activity. Further studies that include additional kinases and phosphatases can be expected to provide a complete understanding of KAT1 channel regulation during stomatal opening and closure.
ABA-activated protein kinase
Coomassie Brilliant Blue
calcineurin B-like protein
CBL-interacting protein kinase
green fluorescent protein
liquid chromatography tandem MS
protein phosphatase 2C
Snf1-related protein kinase 2
yeast nitrogen base
Aiko Sato generated all recombinant peptides and performed voltage-clamp experiments. Yuki Sato, Taishi Umezawa, Aiko Sato and Derek Goto performed the in-gel and in vitro phosphorylation assays. Yoichiro Fukao and Masayuki Fujiwara carried out the LC-MS/MS analysis. Kazuo Shinozaki, Takao Hibi, Mitsutaka Taniguchi, Hiroshi Miyake, Derek Goto and Nobuyuki Uozumi were research group leaders who contributed to experiment design and data interpretation. Aiko Sato, Derek Goto and Nobuyuki Uozumi designed the research and wrote the manuscript.
We thank Christophe Belin and Sebastien Thomine (Institut des Sciences du Végétal, Gif-sur-Yvette, France) for supplying the pET16b-SbRK2.6 and pET16b-SnRK2.6(G33R) vectors containing His-tagged recombinant SnRK2.6. We also thank Izumi Mori (Research Institute for Bioresources, Okayama University, Okayama, Japan) and Satoshi Naito (Graduate School of Life Science, Hokkaido University, Sapporo, Japan) for support with phosphorylation assays. We used the Radioisotope Laboratory of the Graduate School of Agriculture, Hokkaido University.
This work was supported in part by a Grant-in-Aid for Scientific Research on Priority Areas [grant number 20053002 (to D. B. G)] and by Grants-in-Aid for Scientific Research [grant numbers 17078005, 20246044, 20-08103 (to N. U.)] from the MEXT (Ministry of Education, Culture, Sports, Science and Technology) Plant Graduate Students programme from the Nara Institute of Science and Technology (to A. S.); and by the JSPS (Japan Society for the Promotion of Science).