Kinase-mediated phosphorylation is a pivotal regulatory process in stomatal responses to stresses. Through a redox proteomics study, a sucrose non-fermenting 1-related protein kinase (SnRK2.4) was identified to be redox-regulated in Brassica napus guard cells upon abscisic acid treatment. There are six genes encoding SnRK2.4 paralogs in B. napus. Here, we show that recombinant BnSnRK2.4-1C exhibited autophosphorylation activity and preferentially phosphorylated the N-terminal region of B. napus slow anion channel (BnSLAC1-NT) over generic substrates. The in vitro activity of BnSnRK2.4-1C requires the presence of manganese (Mn2+). Phosphorylation sites of autophosphorylated BnSnRK2.4-1C were mapped, including serine and threonine residues in the activation loop. In vitro BnSnRK2.4-1C autophosphorylation activity was inhibited by oxidants such as H2O2 and recovered by active thioredoxin isoforms, indicating redox regulation of BnSnRK2.4-1C. Thiol-specific isotope tagging followed by mass spectrometry analysis revealed specific cysteine residues responsive to oxidant treatments. The in vivo activity of BnSnRK2.4-1C is inhibited by 15 min of H2O2 treatment. Taken together, these data indicate that BnSnRK2.4-1C, an SnRK preferentially expressed in guard cells, is redox-regulated with potential roles in guard cell signal transduction.
Reversible protein phosphorylation/dephosphorylation is a universal regulatory mechanism for biological processes. This reversible modification can bring about activation or deactivation of target proteins with diverse functions, thus acting as an on/off switch to control physiological processes . Understanding how this regulatory mechanism contributes to plant survival under stress conditions is critically important because plants are sessile and continuously subjected to different environmental stimuli and challenges. Within the genome of the reference plant Arabidopsis thaliana, more than 1000 genes encode protein kinases, indicating broad and fundamental functions of kinases in plants . One group of plant kinases, sucrose non-fermenting 1-related protein kinases (SnRKs), plays crucial roles in plant responses towards a spectrum of environmental factors. Among the three subfamilies of SnRKs , SnRK1s mainly participate in carbon and nitrate metabolism in plants , whereas SnRK2s and SnRK3s are plant-unique kinases mainly functioning in stress responses . The SnRK2 subfamily in A. thaliana comprises 10 genes. Nine SnRK2s are activated by hyperosmotic stress such as mannitol treatment while five (SnRK2.2, SnRK2.3, SnRK2.6, SnRK2.7, and SnRK2.8) are activated by abscisic acid (ABA) in A. thaliana suspension cells .
There have been a few studies of redox regulation of plant protein kinases. It has been reported that deactivation of phosphoenolpyruvate carboxylase kinase by oxidized glutathione (GSSG) was reversed by the addition of dithiothreitol (DTT) and the process could be accelerated by thioredoxin . In contrast, an inhibitory effect of thioredoxin has been observed on an S-locus kinase isolated from Brassica oleracea stigmas . Similarly, the in vitro activities of rice mitogen-activated protein kinase MPK3 and MPK6 were inhibited by a cold-inducible thioredoxin h from rice . A salinity-activated SnRK2 from tobacco (Nicotiana tabacum), NtOSAK (an ortholog of Arabidopsis SnRK2.10) formed a cellular complex with glyceraldehyde 3-phosphate dehydrogenase. Upon salt stress, NtOSAK is activated partially through a nitric oxide (NO)-dependent process, and the glyceraldehyde 3-phosphate dehydrogenase is regulated directly by NO through S-nitrosylation. However, no direct NO modifications on cysteine residues were observed in NtOSAK . The activity of the Arabidopsis brassinosteroid-insensitive 1-associated receptor-like kinase 1 (BAK1) was sensitive to H2O2; an interaction between BAK1 and glutaredoxin AtGRXC2 was observed and glutathionylation on BAK1 was catalyzed by AtGRXC2 in vitro . A mitogen-activated protein kinase from B. napus, BnMAPK4, aggregated upon H2O2 treatment in vitro but retained its activity . However, mutation at cysteine 232 of BnMAPK4 caused loss of aggregation upon H2O2 treatment, indicating thiol-based redox regulation of kinase biochemical features . Reactive oxygen species (ROS) and NO are secondary messengers in both ABA and jasmonic acid signal transduction in guard cells [12,13]. Although several lines of evidence indicate that OPEN STOMATA 1 (OST1; AtSnRK2.6) acts upstream of ROS and NO production, the question of whether OST1 is under redox regulation remained untackled until recently. Wang et al.  reported that OST1 activity is negatively regulated by NO through S-nitrosylation at cysteine 137, a residue close to the catalytic site of OST1. These data together indicate an orchestrated regulatory network interconnecting redox and phosphorylation events in plants.
B. napus is one of the most important oilseed crops and is widely grown worldwide . This species is a close relative of A. thaliana  and its complete genome sequence was published recently . Thus, B. napus has great potential as a crop model for genetic engineering towards enhanced stress tolerance . We previously identified a serine/threonine protein kinase SnRK2 to be redox responsive in ABA-treated B. napus guard cells through a redox proteomics study . Yoo et al.  recently mapped this kinase gene to the B. napus C genome (progenitor B. oleracea) and thus named it BnSnRK2.4-1C. We use BnSnRK2.4-1C here to be consistent with the established nomenclature. We also found that H2O2, NO donor S-nitrosoglutathione (GSNO), and GSSG can inhibit the in vitro BnSnRK2.4-1C autophosphorylation activity and the inhibitory effect could be reversed by DTT .
In the present study, BnSnRK2.4-1C was heterologously expressed and characterized using biochemical approaches. The results demonstrate that both autophosphorylation and substrate phosphorylation activities of BnSnRK2.4-1C are redox-regulated in vitro. The in vitro kinase activity inhibited by H2O2, GSNO, and GSSG can be recovered by thioredoxins. Cysteines contributing to the redox regulation were identified. Additionally, the activity of this kinase, which is preferentially expressed in guard cells , was shown to be responsive to external H2O2 treatment in vivo. This work reveals a new regulatory mechanism interconnecting phosphorylation and redox switches in a B. napus guard cell protein kinase.
Materials and methods
RNA extraction, reverse transcription, and PCR
Guard cell protoplasts were prepared as previously described . RNA was extracted from B. napus guard cell protoplasts with the RNeasy® Plant Mini Kit (Qiagen, Inc., U.S.A.) following the manufacturer's protocol. The quality and quantity of extracted RNA were measured using a NanoDrop® 1000 Spectrometer (Thermo Fisher Scientific, Inc., U.S.A.). cDNA was synthesized from 1 µg of total RNA using a SuperScript® II Kit (Invitrogen, U.S.A.) in a 20 µl reaction with oligo(dT) following the manufacturer's protocol.
Recombinant protein expression and purification
BnSnRK2.4-1C was cloned and heterologously expressed as previously described . To generate the cysteine (C) mutants and serine (S)/threonine (T) mutants, the residues were substituted by alanine (A) or aspartic acid (D) residues using a site-directed mutagenesis kit (Stratagene, U.S.A.) with appropriate primers (Supplementary Table S1). Fidelity of the mutated sequences in the BnSnRK2.4-1C mutant constructs was confirmed by DNA sequencing. All the constructs were individually expressed in E. coli strain BL21 (DE3) by growth in LB medium [1% (w/v) tryptone, 0.5% (w/v) yeast extract, and 1% (w/v) NaCl] at 37°C to an absorbance of 0.6, followed by induction with 1 mM isopropyl-β-d-thiogalactopyranoside at 37°C for 4 h. BnSnRK2.4-1C and mutant proteins were purified as His-tagged proteins using the Midi PrepEase® Kit (Affymetrix/USB, U.S.A.). The homogeneity of the purified proteins was determined by SDS–PAGE, and the identities were confirmed by liquid chromatography–tandem mass spectrometry (LC–MS/MS). Other heterologously expressed proteins such as A. thaliana SLAC1 N-terminal region (AtSLAC1-NT), BnSLAC1-NT, and thioredoxin isoforms m, f, and h were prepared similarly as described above. AtSLAC1-NT contains the N-terminal sequence (189 amino acids with a predicted size of 22 kDa) of an Arabidopsis slow-type anion channel (AT1G12480). The N-terminal sequence of the AtSLAC1 ortholog from B. rapa (GenBank Accession: ADQ28082, similarity 91.2%), one of the progenitor species of B. napus, was used to design primers in order to amplify the corresponding B. napus BnSLAC1-NT (encoding 190 amino acids with a predicted size of 22 kDa). For kinase assays, purified protein was dialyzed against 25 mM Tris–HCl (pH 7.5) containing 0.5 mM DTT and a bacterial protease inhibitor cocktail (Sigma–Aldrich Co., U.S.A.) at 4°C overnight. The protein preparations were concentrated by ultraﬁltration using an Amicon® Ultra 4 ml centrifugal filter with a 3 kDa cutoff membrane (EMD Millipore Corp., U.S.A.) at 4°C. Protein concentration was determined by the Bradford protein assay (Bio-Rad Laboratories, Inc., U.S.A.) with bovine serum albumin as a standard.
Plant growth and generation of transgenic line
Arabidopsis and B. napus used in the present study were grown under short day conditions, with 8/16 h day/night and 22/20°C day/night temperature cycles. Light intensity was 125 ± 15 µmol m−2 s−1. The BnSnRK2.4-1C cDNA with a FLAG tag was cloned into pCAMBIA1301 vector (Cambia, Australia) and introduced into A. thaliana wild-type Col-0 using the floral dip transformation method. The transgenic plants were screened on ½ MS plates containing hygromycin (50 mg/l). Protein expression was confirmed by western blot using anti-FLAG antibody (Sigma–Aldrich Co., U.S.A.).
Plant treatment, protein extraction, western blot, and immunoprecipitation
Two-week-old seedlings of the above-mentioned transgenic line were sprayed with either water (as control) or H2O2 (10 mM). Seedlings were collected at each time point (15 and 30 min) during the treatment and frozen with liquid nitrogen for later protein extraction. Proteins from control or H2O2-treated seedlings were extracted in a buffer containing 50 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA, 0.1% Triton X-100, 0.5% β-mercaptoethanol, and protease and phosphatase inhibitor cocktail (product no. 88 669, Thermo Fisher Scientific, Inc., U.S.A.) after grinding tissue into a fine powder under liquid nitrogen. The supernatant was collected after centrifugation at 20 000g, 4°C for 15 min and the protein concentration was quantified using the Bradford assay (Bio-Rad Laboratories, Inc., U.S.A.).
Protein extracts (20 µg each sample) were separated on a 12% SDS–polyacrylamide gel and transferred to a nitrocellulose membrane under 100 V for 1 h at 4°C using a Mini-PROTEAN system (Bio-Rad Laboratories, Inc., U.S.A.). The membrane was blocked at room temperature with 5% (w/v) non-fat dry milk in phosphate-buffered saline (PBS) buffer for 1 h and then washed three times for 5 min each with PBS buffer containing 0.5% Tween 20 (PBST). The membrane was incubated with anti-FLAG antibody at 4°C overnight. After removing unbound antibodies by washing with PBST, the blots were incubated with goat anti-mouse IgG secondary antibody (horseradish peroxidase conjugates; Thermo Fisher Scientific, Inc., U.S.A.) in the PBST buffer at a dilution of 1 : 10 000 and visualized using a SuperSignal™ West Femto chemiluminescent kit (Thermo Fisher Scientific, Inc., U.S.A.). The blot was imaged using an Amersham Imager 600 (GE Healthcare, U.S.A.) under the automatic mode.
For immunoprecipitation of BnSnRK2.4-1C, 1 mg of plant protein extract was incubated with 30 µl of anti-FLAG M2 affinity gel (Sigma–Aldrich Co., U.S.A.) with gentle shaking on an end-over-end shaker (40 rpm) at 4°C for 2 h. After the incubation, the gel slurry was washed three times with TBS [50 mM Tris–HCl and 150 mM NaCl (pH 7.4)]. The FLAG-tagged BnSnRK2.4-1C was eluted using 3 × FLAG peptides (150 ng/µl in TBS; Sigma–Aldrich Co., U.S.A.) and then used for in-gel kinase activity assays. Each experiment was repeated three times.
In-gel kinase assay
Recombinant BnSLAC1-NT was purified as described above and then mixed with SDS–polyacrylamide gel mixture at a concentration of 20 µg/ml before gel casting. Immunoprecipitated BnSnRK2.4-1C from different samples (as described above) were separated on a 12% SDS–polyacrylamide gel with embedded BnSLAC1-NT as the substrate. SDS was removed by washing the gel three times at room temperature using a buffer containing 25 mM Tris (pH 7.5), 0.5 mM DTT, 0.1 mM Na3VO4, 5 mM NaF, 0.5 mg/ml bovine serum albumin, and 0.1% Triton X-100. The gel was then incubated with renaturing buffer [25 mM Tris (pH 7.5), 1 mM DTT, 0.1 mM Na3VO4, and 5 mM NaF] overnight at 4°C with gentle shaking. Before the phosphorylation reaction, the gel was incubated with reaction buffer [25 mM Tris (pH 7.5), 2 mM EGTA, 5 mM MgCl2, 5 mM MnCl2, 1 mM DTT, and 0.1 mM Na3VO4] for 30 min at room temperature. ATP (final concentration 0.2 µM) and 32P-ATP (50 µCi) were then added to the reaction buffer and the reaction was conducted for 90 min. The gel was washed at least five times using 5% trichloroacetic acid (TCA) and 1% sodium pyrophosphate, for 1 h for each wash. The gel was dried and then autoradiographed to monitor kinase activity.
In vitro in-solution kinase assay
Substrates used in the kinase assays included myelin basic protein (MBP), β-casein, histone type III (all purchased from Sigma–Aldrich Co., U.S.A.), and recombinant AtSLAC1-NT and BnSLAC1-NT (expressed and purified as described above). The reaction buffer for phosphorylation of BnSnRK2.4-1C with or without substrates contained 50 mM Tris–HCl (pH 7.5), 10 mM MnCl2 (or other divalent chloride salts), 2 µM cold ATP, and 2 µCi [γ-32P] ATP. The total volume of each reaction was 20 µl. One microgram of recombinant BnSnRK2.4-1C was added to initiate the reaction unless otherwise stated. For the kinase–substrate reaction to monitor kinase trans-phosphorylation activity, low kinase/substrate ratios were chosen to reduce the risk of non-specific reaction. After incubation at 30°C for 30 min, the reaction was stopped by adding SDS–PAGE sample buffer and denaturing at 100°C for 5 min. Proteins were then separated on 12% SDS–polyacrylamide gels. Phosphorylated proteins were visualized by autoradiography after the gel was washed with a buffer containing 5% TCA and 1% sodium pyrophosphate and dried. For kinase activity quantification in both in-gel and in-solution kinase assays, signal intensities of autoradiograph and Coomassie Blue staining were measured using ImageJ (National Institutes of Health, U.S.A.). Normalized kinase activity was calculated by dividing autoradiographic signal with Coomassie Blue signal and then normalized to that of the BnSnRK2.4-1C wild type.
Phosphopeptide enrichment and phosphorylation site identification using LC–MS/MS
For phosphosite identification, an in vitro kinase reaction was conducted with non-radioactive ATP. After the reaction, ∼5 µg of BnSnRK2.4-1C was run on a 12% SDS–polyacrylamide gel and stained with Coomassie Blue. Gel bands were excised and chopped into small pieces (∼1 × 1 × 1 mm cubes), washed with water once, and destained with 50% acetonitrile (ACN) in 50 mM ammonium bicarbonate. After reduction with DTT (10 mM), alkylation with iodoacetamide (55 mM), and digestion with trypsin (∼100 ng) in 50 mM ammonium bicarbonate overnight, peptides were extracted from the gel pieces twice with 70% ACN and 0.1% of trifluoroacetic acid (TFA) and then lyophilized. Peptides were first dissolved in 10 µl of binding solution (80% ACN, 15% H2O, and 5% TFA, pH <3) and then mixed with 10 µl of sampling solution (3% ACN, 0.1% acetic acid, and 0.01% TFA). Phosphopeptides were enriched with TiO2 NuTip microcolumns (GlygenSci, Columbia, MD, U.S.A.) as previously described . Enriched phosphopeptides were collected and lyophilized. The lyophilized peptides were solubilized in 10 µl of loading buffer (3% ACN, 0.1% acetic acid, and 0.01% TFA) and loaded onto a C18 capillary trap cartridge (LC Packings, U.S.A.) followed by separation on a 15 cm nano-flow analytical C18 column (PepMap 75 µm i.d., 3 µm, 100 Å) at a flow rate of 300 nl/min on a nano-LC ultra 1D plus system (AB Sciex, U.S.A.) using solvents A and B. Solvent A contained 3% ACN (v/v) and 0.1% acetic acid (v/v), whereas solvent B contained 97% ACN (v/v) and 0.1% acetic acid (v/v). Peptide separation was performed with a linear gradient from 3 to 40% of solvent B for 20 min, followed by an increase to 90% of solvent B for 5 min and maintained for 5 min. The eluted peptides were directly electrosprayed into an LTQ Orbitrap-XL mass spectrometer (Thermo Fisher Scientific, Inc., U.S.A.). MS/MS spectra were acquired in a data-dependent mode. An Orbitrap full MS scan (resolution: 3 × 104; mass range 400–1800 Da) was followed by MS/MS scans in the ion trap, which was performed on the top 10 most abundant ions via collision-induced dissociation. The ion isolation window was 3 Da. The normalized collision energy was set at 28%. The dynamic exclusion time was 20 s . Additionally, if a phosphate neutral loss of 98, 49, 32.66 and 24.5 m/z below the precursor ion mass was detected, multistage activation was conducted for the top five ions in a data-dependent manner if the precursor exceeded a threshold of 500 ion counts .
In vitro redox-responsive cysteine identification by isotope labeling and LC–MS/MS
Thiol-reactive isotope-coded afﬁnity tagging (ICAT) was employed to identify cysteines responsive to different redox treatments (i.e. H2O2, GSNO, or GSSG, all at 1 mM) using a reverse labeling strategy [18,24,25] (Supplementary Figure S1A). A 100 µg aliquot of purified BnSnRK2.4-1C was incubated with or without oxidant for 15 min at room temperature. Iodoacetamide was added to the reaction mixture at a final concentration of 100 mM to block free thiols. The alkylation reaction was performed at 37°C for 1 h. TCA was then added to a final concentration of 10% (v/v) and proteins were precipitated on ice for 2 h. After centrifugation at 20 000g for 20 min at 4°C, the pellets were washed once with 80% acetone, once with 100% acetone, and were briefly dried in a lyophilizer. Each pellet was dissolved in 80 µl of ICAT-denaturing buffer from the ICAT kit (AB Sciex, U.S.A.). Protein samples were reduced and then labeled. The control sample was labeled with the ICAT light tag and the oxidant-treated sample was labeled with the ICAT heavy tag; in other words, oxidant-modified cysteines would be captured with the heavy tag on LC–MS/MS analysis. Trypsin digestion and strong cation exchange fractionation were conducted according to the ICAT kit manual. The peptides in each fraction were puriﬁed using an avidin afﬁnity cartridge, dried, and suspended in TFA at 37°C for 2 h to release the peptides. LC–MS/MS and data analysis were performed as previously described .
In vivo redox-responsive cysteine identification
To capture the in vivo redox modifications on cysteine residues of BnSnRK2.4-1C (Supplementary Figure S1B), 20 mM N-ethylmaleimide (NEM) was included in the protein extraction buffer to irreversibly label-free thiol groups. Subsequently, 2.5 mg of the total protein extract from the 35S::BnSnRK2.4-1C-FLAG transgenic plants was incubated with 30 µl of anti-FLAG affinity gels. Samples were eluted and then reduced with 10 mM DTT. Excess of DTT was removed by buffer exchange using a Zeba™ desalting column (Thermo Fisher Scientific, Inc., U.S.A.). The resultant samples were labeled with a thiol-specific reagent monobromobimane (mBBr) and then subjected to SDS–PAGE. The mBBr-labeled BnSnRK2.4-1C was visualized under UV light, and the protein band was excised for in-gel trypsin digestion. LC–MS/MS was performed on an EASY-nLC™ 1200 System coupled online with an Orbitrap Fusion™ Tribrid™ mass spectrometer (Thermo Fisher Scientific, Inc., U.S.A.). Briefly, the samples were loaded onto an Acclaim PepMap® 100 trapping column in solvent A (0.1% formic acid in water) and then separated on a 25 cm PepMap® C18 analytical column with a 30 min gradient. The gradient program was 2–5% of solvent B (0.1% formic acid and 99.9% ACN) in 2 min, 5–35% of B over 2–22 min, 35–98% of B over 22–24 min, and isocratic at 98% of B over 24–30 min. Data-dependent acquisition was performed using a cycle of 2 s, with the full MS and the MS/MS detected in the Orbitrap and the ion trap, respectively. The full MS was scanned over 350–1800 m/z at a resolution of 120 000 (m/z 400) with an automatic gain control of 4E5, and the selected precursor ions were isolated in the quadrupole with a 1.3 Da isolation window. For MS/MS detection, the maximum inject time was 50 ms and the automatic gain control target was 1E4. The raw MS spectra were searched against the Arabidopsis TAIR10 database (https://www.arabidopsis.org/) supplemented with the BnSnRK2.4-1C sequence using Proteome Discoverer 2.1 (Thermo Fisher Scientific, Inc., U.S.A.). NEM (+125.048 Da) and mBBr (+190.074 Da) on cysteine residues and phosphorylation (+79.966 Da) on serine, threonine, and tyrosine residues were selected as dynamic modifications. Quantification of redox-responsive cysteines was based on the measurement of peak intensity of NEM-modified cysteine-containing peptides.
The serine/threonine protein kinase BnSnRK2.4 is an SnRK2b subfamily member
BnSnRK2.4-1C encodes a protein kinase with 354 amino acids and a predicted molecular mass of 40.2 kDa (GenBank Accession: ADP24128). Sequence alignment suggests that BnSnRK2.4-1C belongs to the SnRK2 subfamily (Figure 1A) and contains all 11 conserved kinase subdomains characteristic of serine/threonine kinases . The stretch of glutamic acid residues present at the C-terminus classifies the kinase to the SnRK2b subfamily (Figure 1A); SnRK2a has a dominant aspartic acid repeat at the C-terminus. Phylogenetic analysis showed that closely related kinases of BnSnRK2.4-1C from other plant species included SnRK2.4 and SnRK2.10 from Arabidopsis and an osmotic stress-activated protein kinase from tobacco (NtOSAK) (Figure 1B). Arabidopsis SnRK2.4 (AtSnRK2.4) and AtSnRK2.10 have been found to be activated by both ionic (salt) and non-ionic (mannitol) osmotic stresses ; particularly, upon ABA treatment an increase in phosphorylation was observed in AtSnRK2.4 at the activation loops . NtOSAK is rapidly activated within 1 min after osmotic stress and the activity is maintained for ∼2 h . The known functions of AtSnRK2.4, AtSnRK2.10, and NtOSAK imply that BnSnRK2.4-1C may also be stress-responsive and involved in stress signal transduction.
Sequence alignment and phylogenetic analysis of BnSnRK2.4-1C with several related kinases from other plant species.
BnSnRK2.4-1C requires Mn2+ for
in vitro autophosphorylation activity with multiple phosphorylation sites
Autophosphorylation is one of the most important features of kinases , but some recombinant kinases do not exhibit autophosphorylation activities under in vitro conditions [30,31]. Here three different divalent cations, Mg2+, Mn2+, and Ca2+, were tested alone and in combination as cofactors for in vitro autophosphorylation of purified recombinant BnSnRK2.4-1C (Supplementary Figure S2C). Autophosphorylation activity was observed in the presence of Mn2+ only (Supplementary Figure S2A). Other divalent cations, alone or in combination with Mn2+, had little or no effect on autophosphorylation (Supplementary Figure S2A). Inclusion of calf intestinal phosphatase led to undetectable activity. The maximum BnSnRK2.4-1C activity was achieved with 5 or 10 mM MnCl2 (Supplementary Figure S2B). To ensure optimal assay conditions, 10 mM MnCl2 was used for all subsequent kinase assays.
Phosphorylation site mapping revealed multiple autophosphorylated residues of BnSnKR2.4-1C (Supplementary Figure S3A), including three residues in the activation loop, i.e. serine 154 (Ser154), serine 158 (Ser158), and threonine 159 (Thr159) (Supplementary Figure S3). The counterparts of Ser158 and Thr159 in OST1/SnRK2.6, i.e. Ser175 and Thr176, have been reported to be critical for the OST1 in vitro activity . We mutated Ser158 and Thr159 of BnSnRK2.4-1C to alanine (A) or aspartic acid (D) to test whether they were essential for kinase activity. Figure 2 shows that BnSnRK2.4-1C autophosphorylation activity was greatly reduced when the Ser158 was mutated to alanine, which indicates that Ser158 might be the major site of autophosphorylation or that Ser158 phosphorylation may be a prerequisite for BnSnRK2.4-1C autophosphorylation at the other sites. However, the phosphorylation activity of the S158A mutant with the generic substrate β-casein was only slightly reduced. This implies that although Ser158 is a site of autophosphorylation, phosphorylation of this residue is not required for trans-phosphorylation activity. However, the ‘phosphomimic’ mutant S158D had greater activity, suggesting that phosphorylation at this residue does enhance substrate phosphorylation (Figure 2). In contrast, the mutation of Thr159 to alanine resulted in increased phosphorylation activity against β-casein, while no activity of the phosphomimic mutant T159D was detected, as would be expected for a negative regulatory role of phosphorylation at this site.
Phosphorylation of Ser158 and Thr159 have distinct roles to regulate BnSnRK2.4-1C kinase activity.
BnSnRK2.4-1C phosphorylates AtSLAC1-NT and its ortholog in
B. napus (BnSLAC1-NT)
Three generic substrates, MBP, histone type III, and β-casein, were used to test BnSnRK2.4-1C activity. A 50-fold excess of substrate to kinase (mol/mol) was used to lessen non-specific interaction (Supplementary Figure S4). After 12 h of autoradiography, both MBP and β-casein showed strong phosphorylation signals, but histone type III did not (Supplementary Figure S4), indicating that BnSnRK2.4-1C is not a promiscuous kinase. A slow anion channel in A. thaliana (AtSLAC1) is known to be phosphorylated at amino acid residues within the NT (189 amino acids) and thus activated by OST1/SnRK2.6 . Here, the kinase–substrate interaction between AtSLAC1-NT and BnSnRK2.4-1C was tested in vitro. AtSLAC1-NT (predicted to be 22 kDa) appeared to run slightly larger on the gel (Figure 3). Recombinant BnSnRK2.4-1C phosphorylated AtSLAC1-NT at a much higher level than the generic substrate β-casein after 2 h of autoradiography, which was an exposure time too short to detect β-casein phosphorylation by BnSnRK2.4-1C (Figure 3). The activity of BnSnRK2.4-1C with BnSLAC1-NT as the substrate was comparable to that with AtSLAC1-NT (Figure 3B). The BnSnRK2.4-1C transcript level is high in guard cells and is inducible by ABA treatment . These data suggest that BnSnRK2.4-1C can potentially regulate the activity of SLAC1 in guard cell signal transduction. The phosphorylation of SLAC1 by BnSnRK2.4-1C may play an important role in stomatal closure under stress conditions.
BnSnRK2.4-1C phosphorylates the NT of
A. thaliana slow anion channel AtSLAC1 and its B. napus ortholog BnSLAC1.
in vitro kinase activity is redox-regulated
The autophosphorylation activity of recombinant BnSnRK2.4-1C was previously shown  and further demonstrated here (Figure 4) to be inhibited by oxidants, including H2O2, GSNO, and GSSG, and the inhibitory effect could be reversed by subsequent treatment with DTT. Such inhibitory effects were dose-dependent, as they were not as prominent in BnSnRK2.4-1C treated with lower concentrations (10–100 µM) of oxidants (Supplementary Figure S5).
Autophosphorylation activity of BnSnRK2.4-1C is regulated by thioredoxins
To further assess the relevance of redox regulation of BnSnRK2.4-1C, we tested whether the kinase activity is altered by the presence of thioredoxins. Active recombinant thioredoxin isoforms f, h, and m from B. napus var. Global were added to kinase samples pretreated with different oxidants. As shown in Figure 4, the inhibition caused by H2O2 could be reversed by thioredoxin h, which is a cytosolic isoform, while the isoforms from the chloroplast, f and m, were not as efficient. All three isoforms could partially reverse the inhibitory effect of GSNO at 1 mM (Figure 4). This is consistent with the observation that thioredoxins are effective in catalyzing either transnitrosylation or denitrosylation of some mammalian proteins . However, none of the thioredoxin isoforms was able to recover the activity of GSSG-treated kinase (Figure 4). This is consistent with the idea that thioredoxins are relatively inefficient in deglutathionylation in plant and mammalian species .
We also tested whether the trans-phosphorylation activity of BnSnRK2.4-1C is redox-regulated. Using BnSLAC1-NT as a substrate, decreased phosphorylation of BnSLAC1-NT was observed when BnSnRK2.4-1C was pretreated with H2O2, GSNO, or GSSG (Figure 5). After incubation with DTT, the phosphorylation activity of oxidant-pretreated BnSnRK2.4-1C was recovered or enhanced (Figure 5). Therefore, both the autophosphorylation and phosphorylation activities of BnSnRK2.4-1C are redox-regulated: the oxidants (H2O2, GSNO, and GSSG) inhibit BnSnRK2.4-1C activity, whereas reductants (DTT and thioredoxins) enhance activity.
Trans-phosphorylation of BnSLAC1-NT by BnSnRK2.4-1C is redox-regulated.
Cysteines of BnSnRK2-1C contribute to the redox regulation
Thiol-containing cysteines play an important role in protein redox regulation . BnSnRK2.4-1C has six cysteine residues: Cys90, Cys114, Cys120, Cys142, Cys186, and Cys233. The cysteine residues responsive to each redox treatment in vitro were identified using ICAT (Supplementary Figure S1). Specifically, Cys142 was oxidized by H2O2 treatment, Cys90 was responsive to GSNO, and Cys90, Cys142, and Cys233 reacted with GSSG (Supplementary Figure S6). These results suggest specificity of thiol-based redox regulation of BnSnRK2.4-1C by different oxidants. Among these redox-responsive residues, Cys142 is located just before the activation loop of the kinase (Figure 1A). The crystal structure of a human 5′-AMP-activated kinase, which shares 40% identity with BnSnRK2.4-1C, suggested that Cys142 is spatially close to the ATP-binding motif (PDB ID: 2H6D) . The involvement of these cysteines and the modifications of the redox-responsive cysteines, e.g. disulfide bond, sulfonic acid, and S-nitrosylation, deserve further investigation.
in vivo activity is responsive to H2O2 treatment
To investigate whether BnSnRK2.4-1C activity is redox-regulated in vivo, BnSnRK2.4-1C was immunoprecipitated from transgenic Arabidopsis seedlings treated with H2O2 (10 mM) and the kinase activity was determined via in-gel assays. H2O2 treatment led to reduced kinase activity against BnSLAC1-NT at 15 min, but the activity recovered at 30 min of H2O2 treatment (Figure 6A). Quantitative analysis of the autoradiographic signal intensities from four replicate experiments revealed a statistically significant decrease in in vivo BnSnRK2.4-1C kinase activity upon H2O2 treatment at 15 min, but such a change was not significant at 30 min of treatment (Figure 6B). This indicates that BnSnRK2.4-1C activity in vivo is also redox-responsive and this kinase is functioning dynamically in plant stress responses.
External H2O2 application leads to regulation of BnSnRK2.4-1C
in vivo kinase activity and redox change in the BnSnRK2.4-1C cysteine status.
NEM labeling was employed to identify H2O2-responsive cysteine(s) in the immunoprecipitated BnSnRK2.4-1C. Peptides containing Cys142 and Cys233 were detected. Based on the analysis of peak intensities of NEM-labeled peptides, Cys233 was oxidized in the immunoprecipitated BnSnRK2.4-1C upon H2O2 treatment (Figure 6C and Supplementary Figure S7). Such modification of Cys233, also a residue found to be redox-regulated by in vitro GSSG treatment, is potentially important to mediate the regulation of the in vivo BnSnRK2.4-1C activity under stress conditions.
Elevation of H2O2 and NO levels are well-documented downstream signaling events in guard cell ABA signal transduction [38–41]. Besides their roles as secondary messengers in signal transduction, ROS (e.g. H2O2) and reactive nitrogen species (e.g. NO) can potentially change the cellular redox microenvironment and consequently cause oxidative damage to biomolecules [42,43]. Owing to such dual functions, an orchestrated control is essential to balance the oxidative stress and signaling activities of ROS, including reversible redox regulation of proteins, regulation of phosphoproteins, and activation of ROS-responsive genes . There are a few key examples of redox regulation of proteins participating in guard cell biology. The activities of protein phosphatases type 2Cs ABA INSENSITIVE 1, ABA INSENSITIVE 2, and HYPERSENSITIVE TO ABA 1 (HAB1) in guard cells are sensitive to the redox state, with their activities inhibited by H2O2 [44–46]. Specifically, Cys186 and Cys274 of HAB1 are two key residues that are responsive to H2O2 and mediate oxidative inactivation of this phosphatase . Wang et al.  reported that OST1 activity is negatively regulated by NO through S-nitrosylation at Cys137, a residue close to the catalytic site of OST1. These data suggest a role of thiol-based redox regulation in the ABA signaling complex and this might serve as a mechanism to amplify or desensitize ABA signaling [14,46].
Besides OST1 (AtSnRK2.6), AtSnRK2.2 and AtSnRK2.10 were found to be sulfenylated in an investigation on the Arabidopsis H2O2-dependent sulfenome , one of which (AtSnRK2.10) shares 90% sequence identity with BnSnRK2.4–1C (Figure 1). Here, we provide biochemical evidence of reversible redox regulation of BnSnRK2.4-1C kinase activity. The in vitro autophosphorylation activity of BnSnRK2.4-1C has been shown to be inhibited by H2O2, GSNO, and GSSG, and such inhibitory effects can be reversed by DTT . Here, we show that the phosphorylation activity of BnSnRK2.4-1C is inhibited by oxidants and activated by certain thioredoxin isoforms in vitro (Figures 4 and 5). Interaction between kinases and small redox-regulating enzymes, i.e. thioredoxin and glutaredoxin, has been previously observed; however, the effects on kinase activity are case-dependent [8,10,48]. Specific cysteine residues of BnSnRK2.4-1C respond to different redox treatments (Supplementary Figure S5). Interestingly, Cys120 in BnSnRK2.4-1C is the counterpart of the redox-responsive residue Cys137 in OST1 . Although conserved, Cys120 of BnSnRK2.4-1C was not found to be redox-responsive in our assays, suggesting distinct regulatory mechanisms for the activities of BnSnRK2.4-1C and OST1.
The activation loop (T-loop) of a protein kinase is a region between the catalytic subdomains VII and VIII . Phosphorylation in the activation loop has been identified as a regulatory mechanism in many classes of protein kinases . Mutation of phosphorylation sites such as serine (S) or threonine (T) to aspartic acid (D) or glutamic acid (E), i.e. phosphomimetics, has been extensively employed to characterize the roles of individual phosphosites in kinase activity regulation. In the activation loop of BnSnRK2.4-1C, substitution of Ser158 by D (S158D) greatly enhanced the kinase autophosphorylation and phosphotransferase activities (Figure 2), indicating that phosphorylation of Ser158 is essential to stimulate kinase activity. On the other hand, the BnSnRK2.4-1C mutant T159A showed auto- and trans-phosphorylation activities comparable to the wild type (Figure 2), suggesting the phosphorylation of T159 is not essential for BnSnRK2.4-1C activation. However, the phosphomimic mutant T159D showed non-detectable kinase activity (Figure 2). Similar phenomena have previously been observed, with two proposed explanations [50–52]. One explanation is that the negative charge from the glutamic acid failed to mimic the phosphate group, i.e. a phosphomimic failure, possibly as a result of conformational changes due to the residue substitution [50,51]. Another hypothesis is that hyperphosphorylation within the activation loop may serve as a mechanism to down-regulate the kinase activity after initial activation [52,53]. Overall, our data and previous reports suggest an inhibitory role of the T159 phosphorylation to decrease the kinase activity. Such inhibition, in general, might contribute to the temporal regulation of the kinase activity at different stages of signal transduction.
The SnRK2 kinase subgroup is known to mediate plant stress responses . In A. thaliana, for example, there are 10 SnRK2s, with the majority (except for SnRK2.9) activated in vivo by stresses such as ABA, mannitol, and salinity . Particularly, AtSnRK2.2, 2.3, and 2.6 (OST1) exhibit functional overlap in ABA signal transduction [54,55], while AtSnRK2.7 and 2.8 are weakly activated by ABA [5,56]. AtSnRK2.4 is osmotically activated  and maintains primary root length and lateral root number under saline conditions . AtSnRK2.4 is also involved in cadmium-inhibited root growth through elevation of stress-triggered ROS accumulation .
Although no genetic or physiological evidence to date has shown AtSnRK2.4 to be ABA-responsive, an increase in phosphorylation within the activation loop of AtSnRK2.4 upon ABA treatment of seedlings was observed in a phosphoproteomic analysis . The simultaneous observation of phosphorylation increase in AtSnRK2.4 together with AtSnRK2.2, 2.3, and 2.6 in the same study suggested ABA activation of AtSnRK2.4, a kinase which had not previously been shown to mediate ABA responses. BnSnRK2.4-1C shares 94% amino acid sequence identity with AtSnRK2.4 and is potentially involved in stress responses in B. napus. BnSnRK2.4-1C is preferentially expressed in guard cells . Here, we show that BnSnRK2.4-1C preferentially phosphorylates the N-terminal region of the slow anion channel SLAC1 over generic substrates (Figure 3). The expression specificity and the in vitro interaction between BnSnRK2.4-1C and SLAC1-NT, which is involved in stomatal turgor change and ABA response , indicates a potential role of BnSnRK2.4-1C in signal perception and transduction within guard cells. Consistent with this conclusion, up-regulation of BnSnRK2.4-1C by ABA was detected at the transcriptional level . We also observed that the in vivo BnSnRK2.4-1C kinase activity is regulated by external H2O2 application (Figure 6). All these data suggest the potential involvement of BnSnRK2.4-1C in stress responses in guard cells. However, whether BnSnRK2.4-1C and BnSLAC1 interact in vivo deserves further investigation, and identification of phosphorylation targets of BnSnRK2.4-1C at the proteome level will improve understanding of its kinase functions.
Our data show that BnSnRK2.4-1C in vitro and in vivo kinase activities are redox-regulated. This enzyme was originally identified in B. napus guard cells through proteomic analysis, but the responsive cysteine residues were not identified at that time due to technical limitations . Here, ICAT was employed to map the redox-regulated cysteines of BnSnRK2.4-1C under a variety of oxidative stress conditions. Autophosphorylation sites of BnSnRK2.4-1C were mapped via metal dioxide enrichment followed by LC–MS/MS analysis. We also provide biochemical and physiological data, showing that BnSnRK2.4-1C participates in stress responses in guard cells. In particular, the observed in vitro phosphorylation of the N-terminal region of SLAC1, a crucial anion channel conveying ABA signaling in guard cells, by BnSnRK2.4-1C indicates a potential role of BnSnRK2.4-1C in regulating stomatal responses to abiotic stresses. This work suggests a potential link between redox regulation and phosphorylation events in guard cell signal transduction.
A. thaliana SLAC1 N-terminal region
brassinosteroid-insensitive 1-associated receptor-like kinase 1
B. napus SLAC1 N-terminal region
thiol-reactive isotope-coded afﬁnity tagging
liquid chromatography–tandem mass spectrometry
myelin basic protein
OPEN STOMATA 1
PBS buffer containing Tween 20
reactive oxygen species
sucrose non-fermenting 1
sucrose non-fermenting 1-related protein kinase
Tris–HCl and NaCl
M.Z., T.Z., W.J., and C.S-S. performed the experiments. A.C.H., S.C., S.M.A., and W.S. designed the experiments and contributed to data analysis. All authors contributed to writing and editing the manuscript, and S.C. and S.M.A. finalized the manuscript.
This work was supported by grants MCB 0818051, MCB 1412547, and MCB 1412644 from the National Science Foundation.
We thank Dr Qiang Chen and Ning Zhu for technical assistance, and Dr Jin Koh in the Proteomics and Mass Spectrometry Core, Interdisciplinary Center for Biotechnology Research, University of Florida for help in LC–MS/MS analysis.
The Authors declare that there are no competing interests associated with the manuscript.
Present address: Department of Biology, Pennsylvania State University, University Park, PA 16802, U.S.A.
These authors contributed equally to the work.