The SPAK (STE20/SPS1-related proline/alanine-rich kinase) and OSR1 (oxidative stress-responsive kinase-1) kinases interact and phosphorylate NKCC1 (Na+–K+–2Cl− co-transporter-1), leading to its activation. Recent studies indicated that SPAK and OSR1 are phosphorylated and activated by the WNK1 [with no K (lysine) protein kinase-1] and WNK4, genes mutated in humans affected by Gordon's hypertension syndrome. In the present study, we have identified three residues in NKCC1 (Thr175/Thr179/Thr184 in shark or Thr203/Thr207/Thr212 in human) that are phosphorylated by SPAK and OSR1, and have developed a peptide substrate, CATCHtide (cation chloride co-transporter peptide substrate), to assess SPAK and OSR1 activity. Exposure of HEK-293 (human embryonic kidney) cells to osmotic stress, which leads to phosphorylation and activation of NKCC1, increased phosphorylation of NKCC1 at the sites targeted by SPAK/OSR1. The residues on NKCC1, phosphorylated by SPAK/OSR1, are conserved in other cation co-transporters, such as the Na+–Cl− co-transporter, the target of thiazide drugs that lower blood pressure in humans with Gordon's syndrome. Furthermore, we characterize the properties of a 92-residue CCT (conserved C-terminal) domain on SPAK and OSR1 that interacts with an RFXV (Arg-Phe-Xaa-Val) motif present in the substrate NKCC1 and its activators WNK1/WNK4. A peptide containing the RFXV motif interacts with nanomolar affinity with the CCT domains of SPAK/OSR1 and can be utilized to affinity-purify SPAK and OSR1 from cell extracts. Mutation of the arginine, phenylalanine or valine residue within this peptide abolishes binding to SPAK/OSR1. We have identified specific residues within the CCT domain that are required for interaction with the RFXV motif and have demonstrated that mutation of these in OSR1 inhibited phosphorylation of NKCC1, but not of CATCHtide which does not possess an RFXV motif. We establish that an intact CCT domain is required for WNK1 to efficiently phosphorylate and activate OSR1. These data establish that the CCT domain functions as a multipurpose docking site, enabling SPAK/OSR1 to interact with substrates (NKCC1) and activators (WNK1/WNK4).
The protein kinases SPAK (STE20/SPS1-related proline/alanine-rich kinase) and OSR1 (oxidative stress-responsive kinase-1) were originally identified through their ability to interact with and stimulate the activity of NKCC1 (Na+–K+–2Cl− co-transporter) [1,2]. SPAK and OSR1 are 68% identical in sequence, possess a highly similar kinase catalytic domain as well as a CCT (conserved C-terminal) domain, comprising 92 amino acids that are 79% identical in sequence. The CCT domain interacts with RFXV (Arg-Phe-Xaa-Val) motifs in NKCC1  as well as other potential binding partners [3,4]. Recent studies indicate that the WNK1 [with no K (lysine) protein kinase-1] and WNK4 enzymes phosphorylate and activate the SPAK and OSR1 protein kinases [5,6]. Mutations in the genes encoding WNK1 and WNK4 enzymes are found in families with an inherited hypertension and hyperkalaemia (elevated plasma K+) disorder termed Gordon's syndrome or pseudohypoaldosteronism type II . The WNK1 and WNK4 protein kinases phosphorylate SPAK and OSR1 at their T-loop residue (Thr233 in SPAK, Thr185 in OSR1) in addition to a non-catalytic serine residue lying in a conserved motif between the catalytic and CCT domain (Ser373 in SPAK, Ser325 in OSR1) . Mutational analysis indicates that phosphorylation of the T-loop in SPAK and OSR1 rather than the phosphorylation of the serine residue mediates activation by WNK1 . Consistent with this, mutation of the T-loop threonine residue, phosphorylated by WNK1/WNK4, to a glutamate residue to mimic phosphorylation increases the basal activity of OSR1  and SPAK (A. C. Vitari, unpublished work) over 10-fold and prevents further activation by WNK1. In the present study, we have mapped the residues on NKCC1 that are phosphorylated by SPAK and OSR1 and we have shown that endogenous NKCC1 phosphorylation at these sites is increased in response to hyperosmotic stress in cells. Furthermore, we have developed a peptide substrate that can be employed to easily assess the activity of these kinases. We have also characterized the binding properties of the CCT domain and establish that it plays an important role in regulating the activation of SPAK/OSR1 by WNK1 as well as the phosphorylation of SPAK and OSR1 substrates such as NKCC1.
MATERIALS AND METHODS
Sequencing-grade trypsin and protease-inhibitor cocktail tablets were from Roche. Dialysed foetal bovine serum and other tissue-culture reagents were from Life Technologies. [γ-32P]ATP, streptavidin–Sepharose high-performance and glutathione–Sepharose 4B were from Amersham Biosciences. Colloidal Blue staining kit was from Invitrogen. Tween 20 was from Sigma. Nonidet P40 was from Fluka. P81 phosphocellulose paper was from Whatman. GST (glutathione S-transferase)–PreScission protease was expressed and purified from Escherichia coli using plasmids kindly provided by Professor John Heath (School of Biosciences, University of Birmingham, Birmingham, U.K.). All peptides were synthesized by Dr Graham Bloomberg (Molecular Recognition Centre, University of Bristol School of Medical Sciences, Bristol, U.K.). Streptavidin-coated (SA) sensor chips were from BiaCore AB (Stevenage, Herts., U.K.).
General methods, buffers and DNA constructs
Tissue culture, transfection, immunoblotting, restriction-enzyme digests, DNA ligations and other recombinant DNA procedures were performed using standard protocols. All mutagenesis steps were carried out using the QuikChange® site-directed mutagenesis kit (Stratagene) using KOD polymerase from Thermococcus kodakaraensis (Novagen). DNA constructs used for transfection were purified from E. coli DH5α cells using Qiagen plasmid Mega or Maxi kit according to the manufacturer's protocol. All DNA constructs were verified by DNA sequencing, which was performed by the Sequencing Service, School of Life Sciences, University of Dundee, Dundee, Scotland, U.K., using DYEnamic ET terminator chemistry (Amersham Biosciences) on Applied Biosystems automated DNA sequencers. Lysis buffer was 50 mM Tris/HCl, pH 7.5, 1 mM EDTA, 1 mM EGTA, 1% (w/w) Nonidet P40, 1 mM sodium orthovanadate, 50 mM NaF, 5 mM sodium pyrophosphate, 0.27 M sucrose, 1 mM DTT (dithiothreitol) and Complete™ protease-inhibitor cocktail (one tablet per 50 ml). Buffer A was 50 mM Tris/HCl, pH 7.5, 0.1 mM EGTA and 1 mM DTT. Sample Buffer was 1× NuPAGE® LDS (lithium dodecyl sulphate) sample buffer (Invitrogen) containing 1% (v/v) 2-mercaptoethanol. For expression of proteins in E. coli, pGEX-6P-1 constructs were transformed into E. coli BL21 cells, expressed and purified as described previously . For expression of proteins in HEK-293 (human embryonic kidney) cells, this was undertaken as described previously . The cloning of human WNK1 , SPAK, OSR1 and a fragment corresponding to amino acids 1–260 of dogfish shark NKCC1  were described previously.
Mapping phosphorylation sites on NKCC1
The activation assay mixtures were set up in a total volume of 25 μl of buffer A, containing 0.25 μM GST–WNK1-(1–661) (expressed in E. coli), 5 μM GST–SPAK or 5 μM GST–OSR1 (expressed in E. coli), 10 mM MgCl2 and 0.1 mM non-radioactive ATP. After incubation for 40 min at 30 °C, 5 μl of the activation assay mix was transferred to 20 μl of buffer A containing 6.25 μM NKCC1-(1–260) expressed in E. coli, 10 mM MgCl2 and 0.1 mM [γ-32P]ATP (∼3000 c.p.m./pmol). After incubation for 40 min at 30 °C, incorporation of phosphate was determined following electrophoresis of samples on 10% NuPAGE® Bis-Tris gels and autoradiography of Coomassie Blue-stained gels. The protein bands corresponding to NKCC1 were excised, and phosphate incorporation was quantified on a PerkinElmer liquid-scintillation counter. Following tryptic digestion, >87% of the 32P radioactivity incorporated in the gel bands was recovered and the samples were chromatographed on a reverse-phase HPLC Vydac 218TP5215 C18 column (Separations Group) as described in the legend to Figure 1. Fractions corresponding to the major 32P-containing peaks were analysed using an Applied Biosystems 4700 proteomics analyser [MALDI–TOF/TOF (matrix-assisted laser-desorption ionization–tandem time-of-flight)] and solid-phase Edman degradation on an Applied Biosystems 494C sequenator of the peptide coupled to Sequelon-AA membrane (Milligen) .
Sites on NKCC1 phosphorylated by SPAK and OSR1
In vitro NKCC1 phosphorylation reactions
Assays were set up in a total volume of 25 μl of buffer A containing 1 μM GST–[T185E]OSR1 or GST–[T233E]SPAK expressed in E. coli, 5 μM NKCC1-(1–260), GST–NKCC1-(1–260) or GST–[T175A/T179A/T184A]NKCC1-(1–260), 10 mM MgCl2 and 0.1 mM [γ-32P]ATP (∼300 c.p.m./pmol). After incubation for 40 min at 30 °C, incorporation of phosphate was determined as described above.
OSR1 kinase assays employing CATCHtide (cation chloride co-transporter peptide substrate) or NKCC1-(1–260)
Assays were set up in a total volume of either 50 μl (CATCHtide assay) or 25 μl (NKCC1 assay) in buffer A containing 10 mM MgCl2, 0.1 mM [γ-32P]ATP (∼300 c.p.m./pmol), 0.08–0.5 μM GST–[T185E]OSR1, 300 μM CATCHtide (RRHYYYDTHTNTYYLRTFGHNTRR) or 5 μM NKCC1-(1–260). After incubation for 10–60 min at 30 °C, the reaction mixture was applied on to P81 phosphocellulose paper, the papers were washed in phosphoric acid and incorporation of 32P radioactivity into CATCHtide was quantified as described previously . For the NKCC1 assay, the reaction was stopped by the addition of SDS sample buffer and phosphorylation of NKCC1 was assessed as described above following SDS/PAGE (10% Bis-Tris).
For immunoblotting of WNK1, an antibody raised in sheep against WNK1 protein encompassing residues 61–661 expressed in E. coli was used . For the immunoprecipitation of WNK1 (see Figure 8), we employed an antibody raised against the C-terminus of human WNK1 (residues 2360–2382: QNFNISNLQKSISNPPGSNLRTT) . For immunoblotting of phosphorylated NKCC1 (see Figure 2), an antibody raised in sheep against a peptide encompassing residues 198–217 of human NKCC1 phosphorylated at Thr203, Thr207 and Thr212 was used (HYYYDpTHT-NpTYYLRpTFGHNT). For the immunoprecipitation and immunoblotting of NKCC1 from HEK-293 cells in Figure 2(B), an antibody raised in sheep against shark NKCC1-(1–260) expressed and purified from E. coli was used. T4 anti-NKCC1 mouse monoclonal antibody was purchased from Developmental Studies Hybridoma Bank (Iowa University, Iowa City, IA, U.S.A.) and used for immunoblotting of NKCC1 as shown in Figures 4 and 5. Mouse monoclonal antibodies recognizing GST were purchased from Sigma (#G1160). Secondary antibodies coupled to fluorophores were from Molecular Probes and Rockland Technologies.
Phosphorylation of endogenous NKCC1
Samples were heated in sample buffer, subjected to SDS/PAGE (3–8% Tris/acetate) and transferred on to nitrocellulose membranes. Membranes were blocked for 5 min in TBST [Tris-buffered saline with Tween 20; 50 mM Tris/HCl (pH 7.5), 0.15 M NaCl and 0.2% Tween 20], containing 10% (w/v) dried skimmed milk. The membranes were then incubated for 16 h at 4 °C in TBST containing 10% (w/v) dried skimmed milk with 0.5 μg/ml anti-WNK1 antibody, 2 μg/ml anti-phospho-NKCC1 antibody, 5 μg/ml anti-NKCC1-(1–260) antibody, 10000-fold dilution for anti-GST antibody or 100 μg/ml ascites containing monoclonal T4 anti-NKCC1 antibody. Detection was performed using fluorophore-conjugated antibody and a Li-Cor Odyssey® IR imaging system.
Affinity-purification of SPAK and OSR1 employing the biotin–RFQV peptide
A 1 mg portion of clarified HEK-293 cell lysate expressing forms of GST–OSR1 (see Figure 6B) or 10 mg of untransfected HEK-293 cell lysate (see Figure 6D) was incubated with 3 μg of the indicated biotinylated peptide for 15 min at 4 °C. A 10 μl aliquot of streptavidin beads equilibrated in lysis buffer was added. After 15 min of incubation at 4 °C under gentle agitation, the beads were washed twice with lysis buffer containing 0.5 M NaCl, followed by two washes with buffer A. The beads were boiled in 1× NuPAGE® LDS sample buffer, and the samples were subjected to SDS/PAGE (10% Bis-Tris). The gels were either transferred on to nitrocellulose membranes for immunoblotting or stained with Colloidal Blue staining kit according to the manufacturer's protocol and were processed for MS fingerprinting.
Immunoprecipitation of endogenous WNK1 and NKCC1
HEK-293 cells grown on 10-cm-diameter dishes were lysed in 0.5 ml of lysis buffer and were clarified by centrifugation at 14000 g for 5 min. For immunoprecipitations, 0.5 mg (WNK1) or 2 mg (NKCC1) of lysate was incubated at 4 °C for 1 h with 5 μl of Protein G–Sepharose beads conjugated to 5 μg of anti-(WNK1 C-terminus) antibody, 1.5 μg of anti-NKCC1-(1–260) antibody or the equivalent amount of pre-immune antibody. The immunoprecipitates were washed twice with lysis buffer containing 0.5 M NaCl, and twice with buffer A.
Binding was analysed in a BiaCore 3000 system. The indicated biotinylated peptides were bound to an SA sensor chip (400 response units). The indicated concentrations of bacterially expressed wild-type and mutant forms of GST–OSR1 fusion protein in HBS-EP [Hepes-buffered saline with EDTA and polysorbate 20; 10 mM Hepes, pH 7.4, 0.15 M NaCl, 3 mM EDTA and 0.005% (v/v) polysorbate 20], were injected over the immobilized peptides at a flow rate of 30 μl/min. Interactions between each peptide and the GST–OSR1 forms were analysed and steady-state binding was determined at each concentration. Dissociation of GST–OSR1 forms from each peptide was monitored over 90 s. Regeneration of the sensor chip surface between each injection was performed with three consecutive 5 μl injections of a solution containing 10 mM NaOH and 1 M NaCl.
Mapping residues in NKCC1 that are phosphorylated by SPAK and OSR1
To define the residues in the N-terminal regulatory domain of NKCC1 that are phosphorylated by SPAK and OSR1, we employed a fragment of shark NKCC1 encompassing residues 1–260 that SPAK and OSR1 are known to phosphorylate efficiently [2,5,6]. Under the conditions employed, NKCC1-(1–260) was phosphorylated by active SPAK or OSR1 to ∼0.15 mol of phosphate/mol, digested with trypsin and chromatographed on a C18 column to isolate 32P-labelled phosphopeptides (Figures 1A and 1B). Both profiles were similar, revealing the presence of two major phosphopeptides termed P1/P3 and P2/P4. MS and Edman sequencing established the identity of P1/P3 as a peptide doubly phosphorylated at Thr175 and Thr179 (corresponding to Thr203 and Thr207 in human NKCC1; see Figure 3A), and P2/P4 as a peptide phosphorylated at Thr184 (corresponding to Thr212 in human NKCC1; see Figure 3A). We next assessed how mutation of these residues affected the phosphorylation of NKCC1-(1–260). Individual mutation of Thr175, Thr179 or Thr184 to alanine moderately reduced phosphorylation of NKCC1 by SPAK (Figure 1C) and OSR1 (Figure 1D), whereas a mutant of NKCC1 in which all three residues were replaced by alanine was no longer phosphorylated by either SPAK (Figure 1C) or OSR1 (Figure 1D).
Generation of a peptide substrate for OSR1
Phosphorylation of endogenous NKCC1
In order to study whether endogenous NKCC1 was phosphorylated on the residues that are phosphorylated by SPAK/OSR1, we raised a phosphospecific antibody against a peptide encompassing residues 198–217 of human NKCC1 phosphorylated on Thr203, Thr207 and Thr212. This antibody recognized NKCC1-(1–260) only after it was phosphorylated by active OSR1 in vitro (Figure 2A). A mutant form of NKCC1-(1–260) in which Thr203, Thr207 and Thr212 were replaced by alanine was not recognized by the anti-phospho-NKCC1 antibody after incubation with active OSR1 and MgATP (Figure 2A).We next assessed whether endogenous NKCC1 was phosphorylated at the residues targeted by SPAK and OSR1. Previous work demonstrated that hyperosmotic stress such as that induced by sorbitol results in hyperphosphorylation and activation of NKCC1 . To verify whether sorbitol increased phosphorylation of endogenous NKCC1 at the sites phosphorylated by SPAK and OSR1, we treated HEK-293 cells in the presence or absence of 0.5 M sorbitol and immunoblotted cell lysates or NKCC1 immunoprecipitates with the anti-phospho-NKCC1 antibody. These data demonstrate that sorbitol increased phosphorylation of NKCC1 2-fold (Figure 2B).
Generation of a peptide substrate for SPAK and OSR1
A peptide encompassing residues 198–217 of human NKCC1 (almost identical with residues 170–189 of shark NKCC1; Figure 3A) in which two arginine residues were added to the N- and C-termini to increase solubility and ability to interact with P81 phosphocellulose paper was generated. This peptide was phosphorylated by the active [T185E]OSR1 mutant (Figure 3B), with a Km of 330 μM and a Vmax of 5.7 units/mg. Similar results were obtained using the active [T233E]SPAK mutant (not shown). This peptide was termed CATCHtide, and can be employed to assess SPAK/OSR1 activity in a straightforward manner, using the assay described in the Materials and methods section.
Residues of the CCT domain of SPAK and OSR1 required for binding to NKCC1 and WNK1
Previous studies using yeast two-hybrid analysis indicated that the CCT domain of SPAK and OSR1 bound to NKCC1 , as well as WNK4 [3,4], through an RFXV motif. Consistent with this, we found that full-length SPAK (547 amino acids long) or OSR1 (527 amino acids long) or the isolated CCT domain of these enzymes, when expressed in HEK-293 cells, interacted with endogenously expressed NKCC1 (Figure 4A). We were unable to detect endogenously expressed WNK4 in HEK-293 cells, but found that endogenously expressed WNK1 that possesses five RFXV motifs, interacted with the CCT domain-containing fragments of SPAK and OSR1 (Figure 4A). Mutant forms of SPAK and OSR1 lacking the CCT domain failed to bind NKCC1 or WNK1 (Figure 4A). To determine which residues on the CCT domain were required for interaction with NKCC1 and WNK1, we undertook an alanine scan in which we investigated the effect of mutating 46 residues in CCT domain on its ability to bind to endogenously expressed NKCC1 and WNK1. Most mutations did not affect the binding; however, a few, including those of Asp479 and Leu493, disrupted binding of the SPAK CCT domain to both WNK1 and NKCC1 (Figure 4B). We also observed that mutations of other residues, namely Thr480, Val484, Glu487, Val494, Asp498 and Val502 affected binding to NKCC1 to a greater extent than to WNK1 (Figure 4B).
Analysis of the interaction of SPAK and OSR1 with WNK1 and NKCC1
CCT domain of OSR1 is required for efficient phosphorylation of NKCC1
To verify whether removal of the CCT domain affected the intrinsic catalytic activity of OSR1, we first assayed this enzyme with CATCHtide, which does not possess an RFXV motif, and would therefore not be expected to interact directly with the CCT domain. Both full-length [T185E]OSR1 and a mutant of OSR1 lacking the CCT domain ([T185E]OSR1-ΔCCT; residues 1–435) phosphorylated CATCHtide with similar efficiency (Figure 5A), indicating that removal of the CCT domain does not affect the intrinsic OSR1 catalytic activity. We next tested the efficiency with which the [T185E]OSR1-ΔCCT mutant phosphorylated NKCC1, and observed that it was only able to phosphorylate NKCC1-(1–260) at low background level, similar to that observed with an inactive [T185A]OSR1 mutant. We also investigated how mutations in the CCT domain affected the ability of full-length OSR1 expressed in HEK-293 cells to interact with NKCC1 and WNK1 (Figure 5B). Consistent with the results of the isolated CCT domain, the full-length [T185E/D459A]OSR1, [T185E/V464A]-OSR1 and [T185E/L473A]OSR1 mutants failed to interact with endogenously expressed WNK1 and NKCC1 under conditions in which wild-type OSR1 and [T185E]OSR1 bound strongly (Figure 5B). Moreover, the [T185E/D459A]OSR1, [T185E/V464A]-OSR1 and [T185E/L473A]OSR1 mutants that failed to interact with NKCC1 only phosphorylated NKCC1-(1–260) poorly, while they retained normal catalytic activity towards CATCHtide (Figure 5B).
Role of the CCT domain in enabling OSR1 to phosphorylate NKCC1 and CATCHtide
Characterization of binding properties of the CCT domain
In order to analyse the properties of the CCT domain in more detail, we synthesized a biotinylated peptide that encompasses the RFXV motif on WNK4, proposed to mediate the interaction with SPAK . We investigated the ability of this peptide to interact with wild-type and mutant forms of OSR1, deploying a quantitative BiaCore surface plasmon resonance binding assay. The isolated OSR1 CCT domain interacted with the RFQV (Arg-Phe-Gln-Val) motif-containing peptide with high affinity (apparent Kd of 8 nM; Figure 6A). Several of the CCT domain mutants tested, including [D459A]OSR1, [V464A]OSR1 and [L473A]OSR1, as well as OSR1-ΔCCT, failed to bind the RFQV motif-containing peptide. Other mutants ([V474A]OSR1, [T460A]OSR1) interacted with the RFQV motif-containing peptide with lower affinity than the wild-type CCT domain (Figure 6A). Similar results were obtained when binding of the RFQV motif-containing peptide to these mutants was assessed in a HEK-293 cell lysate peptide pull-down assay (Figure 6B). We next studied how mutation of the RFQV motif affected binding of the peptide to the CCT domain. The AFQV (Ala-Phe-Gln-Val), RAQV (Arg-Ala-Gln-Val) and RFQA (Arg-Phe-Gln-Ala) motif-containing peptides failed to bind the OSR1 CCT domain, whereas the RFAV (Arg-Phe-Ala-Val) motif-containing peptide bound to this domain with an affinity similar to that of the parental RFQV motif-containing peptide (Figure 6C). Employing the biotinylated RFQV or RFAV peptides coupled to streptavidin–Sepharose, we were able to affinity-purify sufficient amounts of endogenously expressed SPAK and OSR1 from 10 mg of HEK-293 cell lysate to visualize them on a Coomassie Blue-stained gel and identify them by MS (Figure 6D). In contrast, we were unable to affinity-purify SPAK and OSR1 using the non-SPAK/OSR1-binding peptides (Figure 6D).
Analysis of the binding properties of the CCT domain
The RFQV peptide did not affect the ability of [T185E]OSR1 to phosphorylate CATCHtide (Figure 7A), indicating that interaction of the CCT domain with an interacting partner does not stimulate kinase activity directly. In contrast, the RFQV peptide suppressed, in a dose-dependent manner, phosphorylation of NKCC1-(1–260) by [T185E]OSR1, establishing further the necessity of the CCT domain in enabling OSR1 to phosphorylate NKCC1 (Figure 7A). Consistent with this, peptides that are unable to bind the CCT domain did not suppress phosphorylation of NKCC1-(1–260) by [T185E]OSR1 (Figure 7B).
Effects of RFQV motif-containing peptide on OSR1 kinase activity
CCT domain facilitates activation of OSR1 by WNK1
We next studied how removal of the CCT domain of OSR1 affected its phosphorylation (Figure 8A) and activation (Figure 8B) by full-length WNK1 in vitro. We compared the ability of endogenous WNK1, immunoprecipitated from HEK-293 cell lysates, to phosphorylate full-length OSR1 or OSR1-ΔCCT. WNK1 phosphorylated full-length OSR1 more efficiently than the OSR1-ΔCCT mutant lacking the CCT domain (Figure 8A). In addition, we found that endogenously expressed WNK1 immunoprecipitated from HEK-293 cells activated full-length OSR1 30-fold over a 40 min period (Figure 8B). In contrast, the OSR1-ΔCCT mutant was only activated ∼10-fold in a parallel experiment (Figure 8B).
Role of the CCT domain in regulating OSR1 activation by WNK1
It is well established that phosphorylation of the N-terminal intracellular regulatory domain of NKCC1 plays a crucial role in regulating the activity of this cation co-transporter [11,12]. In vivo32P-labelling experiments of shark rectal gland tubules identified three phosphorylation sites on NKCC1 (Thr184, Thr189 and Thr202) that were stimulated with forskolin and calyculin-A . Mutation of Thr189, but not Thr184 and Thr202, was found to inhibit activation of NKCC1 . Another more recent study indicated that the individual mutation of the equivalent sites on rabbit NKCC2 did not significantly affect co-transporter activity, but mutation of all three residues markedly inhibited activity . SPAK and OSR1 were originally identified as kinases that interacted with and activated NKCC1 [1,2,4]. Further evidence that SPAK regulates the activity of NKCC1 came from the finding that overexpression of dominant-negative SPAK inhibited NKCC1 phosphorylation and activation . Direct mapping of the residues on NKCC1 that are phosphorylated by SPAK and OSR1 has not previously been undertaken. In the present study, we have demonstrated that SPAK and OSR1 phosphorylate a cluster of three residues on shark NKCC1 (Thr175, Thr179 and Thr184) and that mutation of all three sites is required to prevent phosphorylation of NKCC1 by SPAK and OSR1 (Figure 1). Importantly, we also established that sorbitol stimulation of cells, which is known to lead to phosphorylation and activation of NKCC1 , increases the phosphorylation of NKCC1 at the residues targeted by SPAK and OSR1 in vitro (Figure 2B). Further work will be required to assess the effect that mutation of Thr175, Thr179 and Thr184 has on NKCC1 activity in unstimulated and osmotically stressed cells. In addition, it would be important to establish how inhibition of SPAK/OSR1 and WNK isoforms affects activation and phosphorylation of NKCC1 at the Thr175, Thr179 and Thr184 residues. As mentioned above, only the effect of mutating Thr184 of NKCC1 has been investigated previously, and its mutation did not substantially affect NKCC1 activity [13,14]. It is therefore possible that phosphorylation of Thr175 and Thr179, in addition to Thr184, contribute to activation of NKCC1 by SPAK and OSR1, and this will need to be verified by further experimentation. Our peptide-mapping studies indicate that Thr189 and Thr202 on NKCC1 are not phosphorylated by SPAK and OSR1 in vitro, suggesting that other protein kinase(s) are likely to be targeting those sites in vivo.
The residues in NKCC1 phosphorylated by SPAK and OSR1 as well as the residues surrounding them, are conserved in NKCC2 and NCC (Na+–Cl− co-transporter) (Figure 3A). NCC is the target of the thiazide drugs which effectively lower blood pressure in patients with Gordon's syndrome [7,15]. WNK1 is reportedly overexpressed in humans with mutations in the intronic non-coding region of the WNK1 gene, indicating that overexpression of WNK1 results in hypertension . Consistent with this notion, heterozygous WNK1−/+ mice have lower blood pressure . It would be interesting to investigate the activity of the SPAK/OSR1 kinases, as well as the activity and phosphorylation of NKCC1/2 and NCC, in patients with Gordon's syndrome and WNK1−/+ mice. The SPAK and OSR1 phosphorylation sites are not conserved on any of the isoforms of the KCC (K+–Cl− co-transporter) family (Figure 3A), suggesting that they may not be controlled in the same manner by SPAK/OSR1. However, KCC3, but not KCC1, KCC2 and KCC4, possesses a putative CCT-binding RFMV (Arg-Phe-Met-Val) motif in a similar position to that found in NKCC1/2 and NCC and this may therefore explain why KCC3 was found to interact with SPAK/OSR1 in a yeast two-hybrid screen . It would be interesting to investigate whether SPAK/OSR1 are capable of phosphorylating and regulating the activity of KCC3.
Our results are consistent with previous studies [1,3,4] indicating that the CCT domain of SPAK and OSR1 interacts with RFXV motifs on NKCC1 as well as WNK1 and WNK4 with high affinity. However, our findings do not exclude the possibility that variants of the RFXV motif might also interact with the CCT domain. It would be of interest to undertake binding analysis employing peptide arrays, in which each of the residues within the RFXV motif was mutated to all of the other amino acids. Our work also defines for the first time a number of mutations in the CCT domain that specifically abolish interaction with NKCC and WNK1. It would be informative to crystallize the CCT domain to define the molecular mechanism by which this domain binds the RFXV motif with high affinity. Our findings in the present study demonstrate for the first time that a functional CCT domain is essential for efficient binding and phosphorylation of the NKCC1 substrate by SPAK and OSR1, but not CATCHtide, which does not possess an RPXV motif. These results indicate that the CCT domain functions as a specific substrate-recognition domain, which enables docking of RFXV motif-containing substrates and hence their efficient phosphorylation by SPAK and OSR1. Our results also support the notion that the CCT domain is required for efficient phosphorylation and activation of OSR1 by WNK1, as we observe that the OSR1 mutant lacking the CCT domain is markedly less phosphorylated and activated by full-length WNK1 than the wild-type OSR1 (Figure 8).
Many kinases, such as MAPKs (mitogen-activated protein kinases), GSK3 (glycogen synthase kinase 3) and PDK1 (phosphoinositide-dependent kinase 1), rely on substrate-recognition sites termed docking sites that recognize motifs on their substrate distant from the sites that are phosphorylated (reviewed in ). In the MAPKs GSK3 and PDK1, the substrate docking regions are located within a specific pocket of the catalytic domain itself. In contrast, SPAK and OSR1 have evolved a specific relatively large 92-residue domain separate from the kinase domain to interact with substrates as well as upstream regulators. Further work is required to explore how the CCT domain interaction with WNK1 or NKCC1 is co-ordinated in a manner that permits SPAK/OSR1 to become first activated by WNK1/WNK4 and subsequently interact with their substrates. It would also be interesting to determine whether SPAK and OSR1 interact with substrates other than NKCC or other co-transporters through their CCT domain.
We thank Steven L. Roberds (Pfizer, St. Louis, MO, U.S.A.) for support, Gursant Kular for help with the Biacore analysis, David Campbell for MS analysis, Agnieszka Kieloch and Gail Fraser for technical assistance, the Sequencing Service (School of Life Sciences, University of Dundee) for DNA sequencing, the Post Genomics and Molecular Interactions Centre for Mass Spectrometry facilities (School of Life Sciences, University of Dundee), and the protein production and antibody purification teams [Division of Signal Transduction Therapy (DSTT), University of Dundee] co-ordinated by Hilary McLauchlan and James Hastie for expression and purification of antibodies. A.C.V. is the recipient of a Pfizer-sponsored studentship. H.K.R.K. is supported by a long-term FEBS (Federation of European Biochemical Societies) fellowship. D.R.A. is supported by the Association for International Cancer Research, Diabetes UK, the Medical Research Council and the Moffat Charitable Trust, as well as the pharmaceutical companies that support the DSTT (AstraZeneca, Boehringer-Ingelheim, GlaxoSmithKline, Merck & Co. Inc., Merck KGaA and Pfizer).
cation chloride co-transporter peptide substrate
domain, conserved C-terminal domain
glycogen synthase kinase 3
human embryonic kidney
lithium dodecyl sulphate
matrix-assisted laser-desorption ionization–tandem time-of-flight
mitogen-activated protein kinase
oxidative stress-responsive kinase-1
phosphoinositide-dependent kinase 1
STE20/SPS1-related proline/alanine-rich kinase
Tris-buffered saline with Tween 20
with no K (lysine) protein kinase