The WNK-regulated SPAK/OSR1 kinases directly phosphorylate and inhibit the K+–Cl− co-transporters

Precise homoeostasis of the intracellular concentration of Cl− is achieved via the co-ordinated activities of the Cl− influx and efflux. We demonstrate that the WNK (WNK lysine-deficient protein kinase)-activated SPAK (SPS1-related proline/alanine-rich kinase)/OSR1 (oxidative stress-responsive kinase 1) known to directly phosphorylate and stimulate the N[K]CCs (Na+–K+ ion co-transporters), also promote inhibition of the KCCs (K+–Cl− co-transporters) by directly phosphorylating a recently described C-terminal threonine residue conserved in all KCC isoforms [Site-2 (Thr1048)]. First, we demonstrate that SPAK and OSR1, in the presence of the MO25 regulatory subunit, robustly phosphorylates all KCC isoforms at Site-2 in vitro. Secondly, STOCK1S-50699, a WNK pathway inhibitor, suppresses SPAK/OSR1 activation and KCC3A Site-2 phosphorylation with similar efficiency. Thirdly, in ES (embryonic stem) cells lacking SPAK/OSR1 activity, endogenous phosphorylation of KCC isoforms at Site-2 is abolished and these cells display elevated basal activity of 86Rb+ uptake that was not markedly stimulated further by hypotonic high K+ conditions, consistent with KCC3A activation. Fourthly, a tight correlation exists between SPAK/OSR1 activity and the magnitude of KCC3A Site-2 phosphorylation. Lastly, a Site-2 alanine KCC3A mutant preventing SPAK/OSR1 phosphorylation exhibits increased activity. We also observe that KCCs are directly phosphorylated by SPAK/OSR1, at a novel Site-3 (Thr5 in KCC1/KCC3 and Thr6 in KCC2/KCC4), and a previously recognized KCC3-specific residue, Site-4 (Ser96). These data demonstrate that the WNK-regulated SPAK/OSR1 kinases directly phosphorylate the N[K]CCs and KCCs, promoting their stimulation and inhibition respectively. Given these reciprocal actions with anticipated net effects of increasing Cl− influx, we propose that the targeting of WNK–SPAK/OSR1 with kinase inhibitors might be a novel potent strategy to enhance cellular Cl− extrusion, with potential implications for the therapeutic modulation of epithelial and neuronal ion transport in human disease states.


INTRODUCTION
In eukaryotic cells, the regulation of intracellular Cl − concentration ([Cl − ] i ) is important for multiple essential cellular processes, including cell volume homoeostasis, which counters excessive shrinkage or swelling amidst changes in extracellular tonicity and/or intracellular ion content [1]. The SLC12A (solute carrier family 12) family of electroneutral CCCs (cation-Cl − cotransporters), by transporting Cl − against or with its electrochemical gradient, are critical determinants of [Cl − ] i [2]. Cl − influx is powered by the Na + -driven CCC members, which includes the NCC (Na + -Cl − co-transporter) in the distal convoluted tubule of the kidney [3,4], the NKCC2 (Na + -K + -2Cl − co-transporter 2) in the thick ascending limb of the kidney [5] and the ubiquitously expressed NKCC1 [6][7][8]. Cl − efflux is powered by the K + -driven CCCs, which include four different K + -Cl − co-transporters (KCC1-KCC4) [9], including the neuron-specific KCC2. There are two well-studied splice variants of KCC2, termed KCC2 and KCC2A [10], and of KCC3, termed KCC3 and KCC3A [11]. The physiological importance of the CCCs is illustrated by the human Mendelian diseases or mouse phenotypes that result from their mutation or dysfunction [12], and that two CCCs are the targets of the most commonly used drugs in medicine, the loop-diuretic furosemide (inhibiting NKCC2) and thiazide diuretics (inhibiting NCC) [13]. The activities of the NCC/NKCC1/NKCC2 (i.e. N[K]CCs [Na + -K + ion co-transporters]) and KCCs are reciprocally regulated by protein (de)phosphorylation [9,14,15]. Phosphorylation activates NCC/NKCC1/NKCC2, but inhibits KCCs [9,[15][16][17]. Dephosphorylation has the opposite effect. This reciprocal regulation of Na + -and K + -driven CCCs ensures that cellular Cl − influx and efflux is tightly co-ordinated [9,18]. The importance of this mechanism is exemplified by its evolutionary conservation from worms to humans [19].
In the kidney, the WNK-SPAK/OSR1-mediated activation of NCC and NKCC2, which together mediate ∼ 25 % of renal salt reabsorption, is critical for extracellular volume (influencing blood pressure) and electrolyte homoeostasis. The importance of this pathway in human renal physiology is underscored by the findings that: (i) gain-of-function mutations in WNK1 and WNK4 resulting in increased NCC and NKCC2 activities cause a Mendelian syndrome featuring thiazide-sensitive hypertension and hyperkalaemia (pseudohypoaldosteronism type II, also known as PHAII [32]); (ii) loss-of-function mutations in NCC [33] and NKCC2 [34] cause Gitelman's and Bartter's type 1 syndromes respectively, featuring hypotension and hypokalaemia; and (iii) a mutation of NCC at a residue (T60M) that ablates the key activating WNK-regulated SPAK/OSR1 phosphorylation event causes Gitelman's syndrome in Asian people [35]. Moreover, SPAK-knockout mice [36], or knockin mice expressing a form of SPAK or OSR1 that cannot be activated by WNK kinase isoforms [37], exhibit low blood pressure and are resistant to hypertension when crossed to animals bearing a PHAII-causing knockin mutation that enhances WNK4 expression [38].
In contrast with the N[K]CCs, the direct mediators of KCC phospho-regulation are not known, although early experiments suggested the WNK-SPAK/OSR1 kinases may be involved [39][40][41]. Work to date indicates that two threonine residues that are conserved in all KCC isoforms, termed Site-1 (Thr 991 in KCC3) and Site-2 (Thr 1048 in KCC3), both located within the CTD (C-terminal cytoplasmic domain), play a critical role in controlling the activity of the KCCs [42]. Hypotonic high K + conditions that activate KCC isoforms induce a rapid and robust dephosphorylation of Site-1 and Site-2 [42]. Consistent with these sites representing critical regulatory residues, mutation of Site-1 and Site-2 to alanine in KCC3 results in a constitutively active co-transporter [42]. The identity of the protein kinases that phosphorylate Site-1 and Site-2, however, are currently unknown, although the WNK1 signalling pathway has been implicated, as siRNA-mediated knockdown of WNK1 partially suppressed the phosphorylation of these residues in KCC3 [42]. Consistent with a role for WNK isoforms in controlling KCC activity, overexpression of WNK3 and WNK4 in oocytes and mammalian cells leads to the inhibition of KCC isoforms, and kinase-dead overexpression of WNK3 was found to potently stimulate KCC activity [39,40,[43][44][45].
In the present study, we explore the mechanism by which WNKs might regulate KCC phosphorylation. Our data reveal for the first time that the WNK-controlled SPAK/OSR1 act as direct phosphorylators and major regulators of the KCC isoforms, and explain how activation of the WNK signalling pathway can co-ordinately regulate Cl − influx and efflux by reciprocally controlling the SLC12 family N[K]CC and KCC isoforms. Given these reciprocal actions with anticipated net effects of increasing Cl − influx, we propose that inhibition of the WNK-SPAK/OSR1 signalling pathway might be a potent strategy to enhance cellular Cl − extrusion, with potential therapeutic implications.

Plasmids
Full-length human KCCs isoforms were amplified employing SuperScript III (Invitrogen) from placenta total RNA (Stratagene) using the appropriate oligonucleotides. The RT-PCR (reverse transcription-PCR) products were ligated into the pCR2.1-TOPO vector and sequenced. The sequence-verified KCCs were subcloned into bacterial (pGEX-6P-1) and mammalian (pCMV5) expression vectors using BamHI and NotI. Required amino acid mutations were introduced into the pCR2.1-TOPO clone using site-directed mutagenesis by QuikChange method (Stratagene), but substituting the Taq enzyme with KOD Hot Start DNA polymerase (Novagen). Oligonucleotides were obtained from Invitrogen Life Sciences. DNA sequencing was performed by The Sequencing Service, College of Life Sciences, University of Dundee (http://www.dnaseq.co.uk).

Cell culture, transfections and stimulations
HEK (human embryonic kidney)-293 cells were cultured on 10-cm-diameter dishes in DMEM (Dulbecco's modified Eagle's medium) supplemented with 10 % (v/v) FBS, 2 mM Lglutamine, 100 units/ml penicillin and 0.1 mg/ml streptomycin. For transfection experiments, each dish of adherent HEK-293 cells was transfected with 20 μl of 1 mg/ml polyethylenimine (Polysciences) and 5-10 μg of plasmid DNA as described previously [46]. At 36 h post-transfection cells were stimulated with either control isotonic or hypotonic medium for a period of 30 min. Cells were lysed in 0.3 ml of ice-cold lysis buffer/dish, lysates were clarified by centrifugation at 4 • C for 15 min at 26 000 g and the supernatants were frozen in aliquots in liquid nitrogen and stored at − 20 • C. Protein concentrations were determined using the Bradford method. Where indicated cells were treated with the indicated concentrations of the SPAK/OSR1 CCT domain inhibitor STOCK1S-50699 [30], which was purchased from InterBioScreen.

Expression and purification of OSR1 T185E and D164A, SPAK D212A, and KCCs N-terminal and C-terminal domains in Escherichia coli cells
All pGEX-6P-1 constructs were transformed into BL21 E. coli cells and 1-litre cultures were grown at 37 • C in LB broth containing 100 μg/ml ampicillin until the attenuance at 600 nm was 0.8. IPTG (30 μM) was then added and the cells were cultured for a further 16 h at 26 • C. Cells were isolated by centrifugation (4 • C for 30 min at 3500 g), resuspended in 40 ml of ice-cold lysis buffer and lysed in one round of freeze-thawing, followed by sonication (Branson Digital Sonifier; ten 15-s pulses with a setting of 45 % amplitude) to fragment DNA. Lysates were centrifuged at 4 • C for 15 min at 26 000 g. The GST-fusion proteins were affinity-purified on 0.5 ml of glutathione-Sepharose and eluted in buffer A containing 0.27 M sucrose and 20 mM glutathione.

Immunoblotting and phospho-antibody immunoprecipitation
Cell lysates (15 μg) in SDS sample buffer were subjected to electrophoresis on polyacrylamide gels and transferred on to nitrocellulose membranes. The membranes were incubated for 30 min with TTBS containing 5 % (w/v) non-fat powdered skimmed milk. The membranes were then immunoblotted in 5 % (w/v) non-fat powdered skimmed milk in TTBS with the indicated primary antibodies overnight at 4 • C. Sheep antibodies were used at a concentration of 1-2 μg/ml. The incubation with phosphospecific sheep antibodies was performed with the addition of 10 μg/ml of the dephospho-peptide antigen used to raise the antibody. The blots were then washed six times with TTBS and incubated for 1 hour at room temperature (15-20 • C) with secondary HRP-conjugated antibodies diluted 5000-fold in 5 % (w/v) non-fat powdered skimmed milk in TTBS. After repeating the washing steps, the signal was detected using an ECL reagent. Immunoblots were developed using a film automatic processor (SRX-101; Konica Minolta Medical) and films were scanned with a 600 d.p.i. resolution on a scanner (PowerLook 1000; UMAX). Figures were generated using Photoshop/Illustrator (Adobe). For phospho-antibody immunoprecipitation, KCC isoforms were immunoprecipitated from indicated cell extracts. An aliquot (2 mg) of the indicated clarified cell extract was incubated with 15 μg of the indicated phospho-specific KCC antibody conjugated to 15 μl of Protein G-Sepharose in the presence of 20 μg of the dephosphorylated peptide encompassing the residues on KCC that the antibody was raised against. Incubation was for 2 h at 4 • C with gentle agitation, and the immunoprecipitates were washed three times with 1 ml of lysis buffer containing 0.15 M NaCl and twice with 1 ml of buffer A. Bound proteins were eluted with 1× LDS sample buffer. c.p.m./pmol) in a total reaction volume of 25 μl. The reaction was terminated by the addition of LDS sample buffer. DTT was added to a final concentration of 10 mM, the samples heated at 95 • C for 5 min and cooled for 20 min at room temperature. Iodoacetamide was then added to a final concentration of 50 mM and the samples left in the dark for 30 min at room temperature to alkylate the cysteine residues. The samples were subjected to electrophoresis on a Bis-Tris 10 % polyacrylamide gel, which was stained with Colloidal Blue and then autoradiographed. The corresponding phosphorylated bands were excised, cut into smaller pieces and washed sequentially for 10 min on a vibrating platform with 1 ml of each of the following: water, 1:1 (v/v) mixture of water and acetonitrile, 0.1 M ammonium bicarbonate, 1:1 mixture of 0.1 M ammonium bicarbonate and acetonitrile, and finally acetonitrile. The gel pieces were dried before incubation for 16 h at 30 • C in 25 mM triethylammonium bicarbonate containing 5 μg/ml trypsin as described previously [47]. Following tryptic digestion, >95 % of the 32 P radioactivity incorporated in the gel bands was recovered and the samples were chromatographed on a Vydac 218TP5215 C 18 column (Separations Group) equilibrated in 0.1 % trifluoroacetic acid in water. The column was developed with a linear acetonitrile gradient at a flow rate of 0.2 ml/min and fractions of 0.1 ml were collected. Some phosphopeptide fractions (Tables 1 and 2) were purified further by immobilized metalchelate affinity chromatography on Phos-Select resin (Sigma) as described previously [48].

Table 1 SPAK in the presence of MO25 phosphorylates KCC isoforms at two conserved Site-2 and Site-3 residues
The NTD and CTD of the indicated KCC isoforms were phosphorylated by SPAK (except for 32 P cycle release data indicated with an asterisk that was phosphorylated by OSR1) in the presence of MO25 with either radioactive [γ 32 P]ATP or non-radioactive Mg-ATP. Phosphorylated fragments were trypsin digested and the resulting tryptic peptides were isolated by HPLC for the 32 P-labelled ones. The identity and site of phosphorylation was determined by a combination of MS and solid-phase Edman sequencing. Non-radioactive KCC fragments phosphorylated by SPAK were analysed by orbitrap MS. The deduced amino acid sequence of each peptide is shown and the phosphorylated residue is underlined and in bold font. The resultant data was searched using the Mascot search algorithm (http://www.matrixscience.com) run on a local server. For conservation and location of sites see Figure 2. It should be noted that although experiments were done with the shorter KCC3A splice variant that lacks nine N-terminal residues, in order to be consistent with most previous literature studies, numbering of residues is based on the nine-residue-longer KCC3 variant. 32

Table 2 Summary of other in vitro SPAK phosphorylation sites that were identified, but not investigated further in the present study
Analysis was undertaken as described in the legend of Table 1. For conservation and location of these sites see Figure 2. Peptides with a m/z (observed) marked with 1P correspond to peptide mass with one phosphate and those with 2P correspond to a peptide with two phosphates. Phosphorylated residues are underlined and in bold font. Lower case m indicates oxidized methionine. 32

MS analysis
Isolated phosphopeptides were analysed by MS analysis as follows. LC-MS/MS was performed using a linear ion traporbitrap hybrid mass spectrometer (OrbiTrap Classic; Thermo Fisher Scientific) equipped with a nanoelectrospray ion source (Proxeon Biosystems or Thermo) and coupled to a Proxeon EASY-nLC system. Peptides were typically injected on to a Thermo/Dionex (part number 160321) Acclaim PepMap100 reverse-phase C 18

P-labelled phosphopeptide sequence analysis
The site of phosphorylation of the 32 P-labelled peptides was determined by solid-phase Edman degradation, on an Applied Biosystems 494C sequencer, of the peptide coupled to Sequelon-AA membrane (Applied Biosystems) as described previously [49].

KCC3A phosphorylation sites in vivo identification
Fifteen 10-cm diameter dishes of HEK-293 cells were transfected with 5-10 μg of the pCMV5 construct encoding human FLAG-KCC3A using the polyethylenimine method [46]. The cells were cultured for a further 36 h, treated with either basic or hypotonic medium, and lysed in 0.3 ml of ice-cold lysis buffer. The pooled lysate was clarified by centrifugation at 4 • C for 10 min at 26 000 g. FLAG-KCC3A was affinity-purified by incubating 50 μl of FLAG-agarose beads with 10 mg of the clarified lysate for 2 h at 4 • C. The immunoprecipitates were washed four times with 1 ml of lysis buffer containing 0.15 M NaCl and twice with 1 ml of buffer A. Proteins were eluted from the FLAG beads by the addition of 25 μl of 1× NuPAGE LDS (lithium dodecyl sulfate) sample buffer to the beads. Eluted proteins were reduced by the addition of 10 mM DTT and were prepared as described above. For the identification of phosphorylation sites, the tryptic digests were analysed by LC-MS on an Applied Biosystems 4000 QTRAP system with precursor ion scanning as described previously [50]. The resultant data were searched using the Mascot search algorithm run on a local server. Tryptic phosphopeptides were also analysed as above under MS analysis and XICs (extracted ion chromatograms) of the isolated phosphopeptides were obtained from the MS RAW data files and analysed with Thermo Scientific TM Xcalibur software Version 2.2. Observed variations in charge states, oxidation states and partial cleavages were taken into account when calculating the total XIC for a particular phosphorylation site. The mass range used for the XICs was the observed m/z value + − 10 p.p.m. of the relevant peptide.

Rb + -uptake assay in ES and HEK-293 cells
ES cells were plated in 12-well plates (2.4-cm diameter/well) and the 86 Rb + -uptake assay was performed on cells that were 80 % confluent. HEK-293 cells were plated at a confluence of 50-60 % in 12-well plates (2.4-cm diameter per/well) and transfected with wild-type or various mutant forms of full-length FLAG-tagged human KCCs. Each well of HEK-293 cells was transfected with 2.5 μl of 1 mg/ml polyethylenimine and 1 μg of plasmid DNA. The 86 Rb + -uptake assay was performed on the cells at 36 h post-transfection. In both cases, culture medium was removed from the wells and replaced with either isotonic or hypotonic medium for 15 min at 37 • C. Cell medium was removed by means of aspiration with a vacuum pump and replaced with stimulating medium plus inhibitors including 1 mM ouabain and 0.1 mM bumetanide, to prevent 86 Rb + uptake via NKCC1, for a further 15 min. After this period, the medium was removed and replaced with isotonic medium plus inhibitors containing 2 μCi/ml 86 Rb + for 10 min at 37 • C. After this incubation period, cells were rapidly washed three times with the respective ice-cold non-radioactive medium. The cells were lysed in 300 μl of ice-cold lysis buffer and 86 Rb + -uptake tracer activity was quantified on a PerkinElmer liquid scintillation analyser.

Figure 2 Conserved and unique identified regulatory phospho-sites in KCC isoforms
Sequence alignment of KCC1, KCC2, KCC2A, KCC3, KCC3A and KCC4 isoforms that show the position of the in vitro and in vivo phosphorylation sites identified in the present study (boxed residues). Boxed residues are phosphorylation sites that are listed in Tables 1-3. ᭡, in vitro phosphorylation sites listed in Tables 1 and 2. ᭹, in vivo KCC3A phosphorylation sites that are dephosphorylated following exposure of cells to hypotonic high K + conditions listed in Table 3. ᭺, in vivo KCC3A phosphorylation sites that are not influenced by exposure of cells to hypotonic high K + conditions listed in Table 3. ଙ, in vivo KCC3A phosphorylation sites that were identified by the Lifton group [42] to become dephosphorylated following exposure of cells to hypotonic high K + conditions. ଝ, in vivo KCC3A phosphorylation sites that were previously identified [42] and not influenced by hypotonic high K + conditions. It should be noted that as all experiments were done with the shorter KCC3A splice variant that lacks nine amino acid N-terminal residues, in order to be consistent with most previous literature studies, numbering of residues for KCC3A is based on the nine-residue-longer KCC3 variant.
of KCC fragments being mediated specifically by SPAK, the kinase-dead SPAK [D212A] mutant failed to phosphorylate KCC isoforms in the presence of MO25α ( Figure 1B). OSR1, in the presence, but not absence, of MO25α, also phosphorylated the NTD and CTD of all KCC isoforms ( Figure 1C). In contrast, kinase-dead OSR1 [D164A] failed to phosphorylate all KCC isoforms ( Figure 1D). These data demonstrate that both SPAK and OSR1 are capable of phosphorylating the NTD and CTD of multiple KCC isoforms when complexed with MO25α.

Identification of in vitro SPAK/OSR1 phosphorylation sites
To map the sites that the SPAK-MO25α complex phosphorylates on KCC isoforms, we performed phosphorylation reactions in the presence of radioactive [γ 32 P]ATP, and then digested them with trypsin. The resulting 32 P-labelled tryptic peptides were isolated by HPLC and the identity and site of phosphorylation was determined by employing a combination of MS and solidphase Edman sequencing ( Figure 2, and Tables 1 and 2). We also analysed directly the phosphorylation reactions using orbitrap MS analysis to pinpoint residues phosphorylated by SPAK-MO25α ( Figure 2 and Table 1).
Notably, the SPAK-MO25α complex was found to phosphorylate all KCC isoforms at the critical Site-2 (Thr 983 in KCC1, Thr 1030 in KCC2A, Thr 1048 in KCC3A and Thr 980 in KCC4; Figure 2 and Table 1). SPAK-MO25α also phosphorylated KCC1, KCC2A and KCC4 at a conserved N-terminal threonine residue that we have termed Site-3 (Thr 5 in KCC1 and Thr 6 in KCC2A/KCC4; Figure 2). Site-3 is not conserved in the KCC3A splice variant employed, but is conserved in the KCC3 splice variant (Thr 5 ) (Figure 2). Consistent with previous findings [51], SPAK-MO25α also phosphorylated KCC3A at Ser 96 (Site-4) within a motif not present in other KCCs (Figure 2). Our analysis also revealed a limited number of other phosphorylation sites observed in only single isoforms, which have not been further investigated (Table 2). Importantly, we did not observe phosphorylation of the critical Site-1 residues by SPAK-MO25α in any of the KCC isoforms tested.

Mapping of KCC3A phosphorylation sites in vivo
We used MS orbitrap analysis of KCC3A transiently expressed in HEK-293 cells that were stimulated with either hypotonic low Cl − conditions (where phosphorylation of Site-1 and Table 3 In vivo KCC3A phosphorylation sites identified in the present study N-terminal FLAG-tagged KCC3A was overexpressed in HEK-293 cells. At 36 h post-transfection, cells were treated with either hypotonic low Cl − or hypotonic high K + conditions to induce dephosphorylation and hence activation of the KCC3A isoform for 30 min. KCC3A was immunoprecipitated and after electrophoresis on a polyacrylamide gel and Colloidal Blue staining, the bands corresponding to KCC3A were excised and digested with trypsin. Phosphorylated peptides were identified by combined MS and MS/MS analysis The data was searched using the Mascot search algorithm run on a local server. The deduced amino acid sequence of each peptide is shown and the phosphorylated residue, where identified, is underlined and in bold font. The ratios of the abundance of each phosphopeptide that was detected in hypotonic high K + samples compared with the hypotonic low Cl − conditions was calculated using Xcalibur software. Lower case m indicates oxidized methionine.The conservation and location of each phosphorylated residue is illustrated in Figure 2. It should be noted that although experiments were done with the shorter KCC3A splice variant that lacks nine N-terminal residues, in order to be consistent with most previous literature studies, the numbering of residues is based on the nine-residue-longer KCC3 variant.

Phosphopeptide
Mascot score (hypotonic high K + ) m/z (observed) Phosphorylation site Xcalibur ratio (hypotonic low Cl − /hypotonic high K + ) Site-2 should be maximal) or hypotonic high K + conditions (which should induce dephosphorylation of Site-1 and Site-2), to map further KCC3A phosphorylation sites in vivo [42]. This confirmed that hypotonic high K + conditions did indeed induce a >10-fold reduction in phosphorylation of Site-1 (Thr 991 ) and Site-2 (Thr 1048 ) ( Figure 2 and Table 3). Our analysis identified a number of other phosphorylated residues on KCC3A whose phosphorylation was not inhibited significantly by hypotonic conditions (Figure 2 and Table 3). Several of these sites have also been identified previously [42] (Figure 2).

Creation and characterization of phospho-specific KCC antibodies
To develop a tool that is able to monitor the phosphorylation state of these critical KCC phospho-residues in cells and tissues, we raised phospho-specific antibodies that recognize the critical regulatory Site-1 (Thr 991 ), Site-2 (Thr 1048 ) or Site-4 (Ser 96 ) residues in KCC3A. The antibodies recognized by immunoblot or immunoprecipitation ( Figure 3A) wild-type KCC3A overexpressed in HEK-293 cells, but not a mutant version of KCC3A in which the phosphorylated residue was changed to alanine. These antibodies confirm that treatment of HEK-293 cells with hypotonic high K + conditions for 30 min induces almost complete dephosphorylation of KCC3A at Site-1 (Thr 991 ) and Site-2 (Thr 1048 ) ( Figure 3A). We observed that hypotonic high K + conditions induced a significantly more moderate dephosphorylation of Site-4 (Ser 96 ) compared with Site-1 and Site-2 ( Figure 3A). Individual mutation of Site-1 (Thr 991 ), Site-2 (Thr 1048 ) or Site-4 (Ser 96 ) did not significantly affect the phosphorylation of the other sites ( Figure 3A). The residues surrounding Site-1 and Site-2 are almost identical in all KCC isoforms (Figure 2), indicating that the phosphospecific antibodies raised against these residues would probably detect the equivalent residues on other KCC isoforms. Indeed, immunoblot and immunoprecipitation analysis revealed that the KCC3A phospho-specific Site-1 and Site-2 antibodies readily detected KCC1, KCC2 and KCC4 ( Figure 3B) overexpressed in HEK-293 cells grown in basic control isotonic medium, but to a lesser extent in cells treated with hypotonic high K + conditions that induce the dephosphorylation of these residues.

Linking SPAK/OSR1 with KCC3A phosphorylation and transport activity
To determine whether activation of the SPAK/OSR1 pathway promotes phosphorylation of KCC3A at Site-2 and Site-4, the residues that are phosphorylated in vitro ( Figure 2 and Table 1), we exposed HEK-293 cells to hypotonic low Cl − conditions that stimulate SPAK/OSR1 [4]. We monitored WNK1 phosphorylation at its activating T-loop residue (Ser 382 ), SPAK/OSR1 phosphorylation at the residue phosphorylated by WNK1 (Ser 373 /Ser 325 ) and NKCC1 phosphorylation at residues phosphorylated by SPAK/OSR1 (Thr 203 , Thr 207 and Thr 212 ) over a 30 min time period ( Figure 4A). These experiments revealed that within 5 min of exposing HEK-293 cells to hypotonic low Cl − conditions, an increase in the phosphorylation of WNK1, SPAK/OSR1 and NKCC1 was observed, which was maximal within 10 min and sustained for 30 min. Consistent with SPAK/OSR1 mediating the phosphorylation of KCC3A at Site-2 and Site-4, we observed an increased phosphorylation of these residues within 5-10 min of exposing HEK-293 cells to hypotonic low Cl − conditions, and this phosphorylation was sustained for at least 30 min. Exposure of HEK-293 cells to hypotonic low Cl − conditions also promoted the phosphorylation of KCC at Site-1 over a similar time course as for Site-2 and Site-4 ( Figure 4A).
We next analysed KCC3A activity by measuring 86 Rb + uptake in isotonic or hypotonic low Cl − conditions ( Figure 4B). As expected given that KCC3A is maximally phosphorylated following hypotonic low Cl − conditions, very low KCC3A activity was observed ( Figure 4B). Individual mutation of either Site-1 or Site-2 to alanine moderately activated KCC3A, and these mutants were still significantly inhibited by hypotonic low Cl − conditions ( Figure 4B). In contrast, a KCC3A mutant in which both Site-1 and Site-2 were mutated to alanine was highly active and not inhibited further by hypotonic low Cl − conditions, thereby confirming that these are indeed the essential inhibitory sites.
We also treated HEK-293 cells with hypotonic high K + conditions to induce dephosphorylation of KCC3A. Time course analysis revealed that the most pronounced dephosphorylation of Site-1, Site-2 and Site-4 occurred within 15-30 min. Levels of WNK1, SPAK/OSR1 and NKCC1 phosphorylation were also reduced markedly after a 15-30 min exposure to hypotonic high K + conditions ( Figure 4C). We then analysed KCC3A activity by measuring 86 Rb + uptake in isotonic or hypotonic high K + conditions ( Figure 4D). As expected, hypotonic high K + conditions induced a marked activation of wild-type KCC3A ( Figure 4D, upper panel). This was associated with a significant dephosphorylation of KCC3A Site-1 and Site-2 and a marked reduction in WNK1, SPAK/OSR1 and NKCC1 phosphorylation ( Figure 4D, lower panel). Individual alanine mutation of Site-1 and Site-2, preventing phosphorylation at these sites, led to a moderate activation of KCC3A under isotonic conditions, and the activity of these mutants was stimulated further following exposure to hypotonic high K + conditions. Mutation of both Site-1 and Site-2, as reported previously [52], resulted in constitutive activation of KCC3A that was not activated further by hypotonic high K + conditions. Individual mutation of Site-4 was found to have no significant effect on KCC3A activity in isotonic or hypotonic high K + conditions ( Figure 4D, upper panel), as reported previously [51].

Pharmacological evidence that SPAK/OSR1 directly phosphorylate KCC3A Site-2 and Site-4
A small molecule compound termed STOCK1S-50699 has recently been described that binds the CCT domain of SPAK/OSR1, thereby preventing WNK-mediated activation of the SPAK/OSR1, and consequently functioning as a WNK-SPAK/OSR1 inhibitor [30]. To obtain further evidence that SPAK/OSR1 regulate KCC3A phosphorylation, we exposed HEK-293 cells for 30 min to either basic control isotonic conditions or hypotonic low Cl − (to activate SPAK/OSR1), and then treated the cells in the same conditions with increasing doses of STOCK1S-50699 for an additional 30 min. These experiments confirmed that 1-10 μM STOCK1S-50699 inhibited markedly, in a dose-dependent manner, SPAK/OSR1 phosphorylation at Ser 373 /Ser 325 (i.e. the activating site phosphorylated by WNK1) and NKCC1 phosphorylation at Thr 203 /Thr 207 /Thr 212 (SPAK/OSR1 sites whose phosphorylation is required for maximal transporter activity) in both hypotonic low Cl − conditions, as well as basic control isotonic conditions ( Figure 5). These effects were paralleled by a similar suppression of KCC3A Site-2 (Thr 1048 ) and Site-4 (Ser 96 ) phosphorylation, consistent with the phosphorylation of these residues being mediated by SPAK/OSR1. Phosphorylation of Site-1 was also inhibited by STOCK1S-50699, but to a lower extent than Site-2 ( Figure 5).

Evidence that SPAK/OSR1 regulate the endogenous phosphorylation and activity of KCC2A, KCC3A and KCC4
To investigate how the lack of SPAK and OSR1 activity affects endogenous KCC isoform phosphorylation, we utilized previously described wild-type and double OSR1 185A/185A /SPAK 243A/243A knockin ES cells, in which the T-loop phosphorylation site on SPAK and OSR1, that is phosphorylated by WNK isoforms, is mutated to alanine to prevent their activating phosphorylation [29]. Wild-type and knockin cells were treated with basic control isotonic or hypotonic high K + conditions for 30 min to induce dephosphorylation of KCC isoforms, and then the harvested lysates were subjected to immunoprecipitation with Site-1 and Site-2 phospho-specific antibodies that immunoprecipitate all KCC isoforms when phosphorylated at these residues ( Figure 3). The immunoprecipitates were then immunoblotted with antibodies that specifically recognize different individual KCC isoforms. In wild-type ES cells, endogenous KCC2A, KCC3A and KCC4 were phosphorylated at Site-1 and Site-2, as these cotransporters could readily be immunoprecipitated with the Site-1 and Site-2 phospho-specific antibodies ( Figure 6, lower panel). As expected, immunoprecipitation of KCC2A, KCC3A and KCC4 isoforms was inhibited when wild-type ES cells were treated with hypotonic high K + conditions to induce dephosphorylation of these residues. In the OSR1 185A/185A /SPAK 243A/243A knockin ES cells, we still observed that endogenous KCC2A, KCC3A and KCC4 were phosphorylated at Site-1 to a similar extent as wildtype cells; however, consistent with SPAK/OSR1 predominately phosphorylating Site-2, Site-2 phosphorylation in endogenous KCC2A, KCC3A and KCC4 was abolished in knockin ES cells lacking SPAK/OSR1 activity ( Figure 6, lower panel).
We then analysed endogenous KCCs activity in the OSR1 185A/185A /SPAK 243A/243A knockin ES cells by measuring 86 Rb + uptake under basic control isotonic or hypotonic high K + conditions, and compared these results with wild-type ES cells. As expected, exposure of wild-type ES cells to hypotonic high K + conditions induced a significant ∼ 2-fold activation of 86 Rb + uptake ( Figure 6, upper panel), consistent with KCC activation. In contrast, in the double OSR1 185A/185A /SPAK 243A/243A knockin cells that lack SPAK/OSR1 activity, we observed an elevated basal activity of 86 Rb + uptake that was not stimulated markedly further by hypotonic high K + conditions, suggesting an increase in KCC activity secondary to decreased phosphorylation at Site-2 ( Figure 6, upper panel). Together, these data provide evidence that endogenous silencing SPAK/OSR1 kinase activity is able to stimulate KCC activity by preventing Site-2 inhibitory phosphorylation.
We also investigated whether the SPAK/OSR1 inhibitor STOCK1S-50699 would affect the phosphorylation of endogenous KCC isoforms in either ES cells ( Figure 7A) or HEK-293 cells ( Figure 7B). These studies revealed that in both ES cells and HEK-293 cells treated under either basic isotonic or hypotonic low chloride conditions, STOCK1S-50699 inhibited the phosphorylation of KCC3a (ES and HEK-293) and KCC2a (ES) at Site-2 and Site-4, but not Site-1 (Figure 7).

DISCUSSION
The results of the present study reveal a major direct role for the WNK-regulated SPAK/OSR1 protein kinases in the regulation of KCC phosphorylation and function. Our data provide the first biochemical ( Figure 1 and Table 1), pharmacological ( Figure 5) and genetic ( Figure 6) evidence that SPAK/OSR1 phosphorylates directly all KCC isoforms at the critical regulatory Site-2 residue (Thr 1048 in KCC3A) [42]. We also confirmed that SPAK/OSR1 phosphorylates directly a previously defined KCC3A-specific residue termed Site-4 (Ser 96 ) [51]. Although the dephosphorylation of Site-1 and Site-2 plays the most important role in activating KCC isoforms [42], recent work has shown that dephosphorylation of Site-4 is also required for maximal KCC3A activation [51]. In addition to the finding that SPAK/OSR1 phosphorylates directly residues known to critically inhibit KCC function (Site-2 and Site-4), we have identified a novel N-terminal phosphorylation site (Site-3) present in all KCC isoforms except KCC3A (Thr 5 in KCC1/KCC3 and Thr 6 in KCC2A/KCC4) that is also phosphorylated directly by SPAK/OSR1 ( Figure 2 and Table 1). Together, these data unambiguously define WNKregulated SPAK/OSR1 as key regulators of KCC activity via direct inhibitory phosphorylation at multiple critical residues.
Our findings indicate that the MO25α subunit is required for efficient phosphorylation of KCC isoforms by SPAK/OSR1, at least in vitro (Figure 1). This is consistent with MO25α playing an essential role in stabilizing WNK-phosphorylated SPAK/OSR1 in an active conformation that can then efficiently phosphorylate its substrates [25]. These data emphasize the importance of including the MO25α subunit in future studies for identification  86 Rb + uptake was measured in the presence of 1 mM ouabain (inhibitor of the plasma membrane Na + /K + -ATPase sodium pump) and 0.1 mM bumetanide (inhibitor of NKCC1) in the uptake medium (for details, see the Materials and methods section). 86 Rb + -uptake assays were then carried out over a 10 min period and was then quantified by scintillation counting. 86 Rb + uptake c.p.m. were normalized per mg of protein for each condition and plotted for both isotonic and hypotonic conditions. The results are presented as the means + − S.D. 86 Rb + uptake for triplicate samples. Lysates were also subjected to immunoblot analysis with the indicated total and phospho-specific antibodies. Similar results were obtained in three separate experiments. HEK-293 cells were transfected with constructs encoding wild-type N-terminal FLAG epitope-tagged KCC3A. At 36 h post-transfection cells were exposed for 30 min to either basic control isotonic conditions or hypotonic low Cl − conditions (to activate the SPAK/OSR1 pathway), and then treated in the same conditions with the indicated doses of the SPAK/OSR1 CCT domain inhibitor STOCK1S-50699 [30] for an additional 30 min in the presence of 1 mM ouabain (an inhibitor of the plasma membrane Na + /K + -ATPase sodium pump) and 0.1 mM bumetanide (inhibitor of NKCC1). 86 Rb + uptake proceeded for 10 min and was then quantified by scintillation counting. 86 Rb + uptake c.p.m. values were normalized per mg of protein for each condition and plotted for both isotonic and hypotonic conditions. Cells were lysed and extracts were subjected to immunoblot analysis with the indicated total and phospho-specific antibodies (lower panel). Similar results were obtained in three separate experiments.
of novel SPAK/OSR1 substrates. Our conclusion that SPAK/OSR1 operate as key direct regulators of KCC isoforms is consistent with previous observations that found that knockdown of WNK1 expression, which would be expected to suppress SPAK/OSR1 activity, decreased phosphorylation of KCC3A at Site-2 (Thr 1048 ) [42]. Similarly, the observation that overexpression of wild-type WNK3 (and WNK4), predicted to activate SPAK/OSR1, inhibits KCC isoforms in oocytes [39,40,[43][44][45], but overexpression of a kinase-dead WNK3, predicted to inhibit SPAK/OSR1 in a dominant-negative mechanism, potently activates KCC isoforms in oocytes [39], is congruent with our conclusions that SPAK/OSR1, regulated by WNK kinase activity, play an important role in controlling KCC phosphorylation and activity.
Interestingly, our data suggest that SPAK/OSR1 may not phosphorylate directly the important Site-1 (Thr 991 in KCC3A) regulatory residue on KCC isoforms. The strongest and most direct evidence to support this conclusion is the finding that three endogenous KCC isoforms (KCC2, KCC3 and KCC4) are still phosphorylated on Site-1, but not on Site-2, in knockin ES cells that lack SPAK/OSR1 activity ( Figure 6). Therefore, at least in ES cells that lack SPAK and OSR1 activity, Previously described matched wild-type and double OSR1 185A/185A /SPAK 243A/243A knockin ES cells [29], in which SPAK/OSR1 are inactive as the T-loop phosphorylation site that is phosphorylated by WNK isoforms is ablated, were incubated with either basic control isotonic medium (open bars) or hypotonic high K + medium (closed bars) for 30 min in the presence of 1 mM ouabain (an inhibitor of the plasma membrane Na + /K + -ATPase sodium pump) and 0.1 mM bumetanide (inhibitor of NKCC1). 86 Rb + uptake proceeded for 10 min and was then quantified by scintillation counting. 86 Rb + uptake c.p.m. values were normalized per mg of protein for each condition and plotted for both isotonic and hypotonic conditions. The lysates were also subjected to immunoprecipitation (IP) with the indicated KCC3A Site-1 and Site-2 phospho-antibodies that recognize all KCC isoforms (see Figure 3). The immunoprecipitates were then immunoblotted employing the indicated specific KCC isoform antibodies. Molecular masses are indicated in kDa on the left-hand side of the Western blots (lower panel). Similar results were obtained for three separate experiments.
another currently unknown kinase must exist that can efficiently phosphorylate Site-1. We also found that SPAK/OSR1 did not phosphorylate KCC isoforms at Site-1 in vitro ( Figure 1 and Table 1). However, the finding that knockdown of WNK1 partially reduced phosphorylation of KCC3A at Site-1 [42], does suggest WNK isoforms regulate KCC phosphorylation at Site-1, either directly upstream of another kinase or perhaps via the regulation of a phosphatase. It would also appear that WNK1 itself does not directly phosphorylate Site-1, as we have also tested whether endogenously immunoprecipitated WNK1 is capable of phosphorylating Site-1 directly on isolated KCC fragments or full-length immunoprecipitated KCC3A. We found that under conditions that WNK1 efficiently phosphorylates catalytically inactive OSR1, no marked phosphorylation of Site-1 or Site-2 was observed when employing the phospho-specific antibodies as readouts for this activity (P. de los Heros, unpublished work). Further work is required to understand how WNK isoforms might control Site-1 phosphorylation and to what extent SPAK/OSR1 is involved. It may be that other WNK isoforms, such as WNK3, are directly involved in KCC Site-1 regulation, given the activating effects of kinase-dead WNK3 on KCCs in vitro [39]; this is one obvious candidate requiring further detailed biochemical examination. Also consistent with a role for the WNK-SPAK/OSR1 cascade regulating Site-1 phosphorylation, we observed that the CCT domain inhibitor STOCK1S-50699 partially suppresses KCC3A Site-1 phosphorylation, albeit to a lesser extent than its inhibition of SPAK/OSR1 activity and KCC3A Site-2 phosphorylation ( Figure 5). Notably, to check the selectivity of STOCK1S-50699, we assessed the effect of a high concentration of this inhibitor (10 μM) on the activity of 150 protein kinases in vitro and found that it does not significantly inhibit the activity of any other kinase, confirming that this compound is not a general protein kinase inhibitor (http://www.kinase-screen.mrc.ac.uk/kinase-inhibitors).
Our data emphasize that the phosphorylation of Site-1, Site-2 and Site-4 are co-ordinately regulated. For example, following exposure of cells to conditions that activate the WNK signalling pathway (hypotonic low Cl − ), we observed a stimulation of phosphorylation of Site-1 and Site-2 of the KCCs, closely accompanied with increased phosphorylation of WNK1, SPAK/OSR1 and NKCC1 ( Figure 4A). Conversely, exposure of cells to conditions that inhibit the WNK signalling network (hypotonic high K + ) resulted in dephosphorylation of Site-1 and Site-2 in the KCCs, as well as WNK1, SPAK/OSR1 and NKCC1 ( Figure 4C). One important unresolved question concerns the mechanism by which extracellular conditions, such as hypotonic high K + , are detected, and how this signal is transduced to trigger the dephosphorylation of WNK pathway components and KCC

Figure 7 STOCK1S-50699 inhibits phosphorylation of endogenous KCC isoforms at Site-2 and Site-4 in ES and HEK-293 cells
ES (A) or HEK-293 (B) cells were exposed for 30 min to either basic control isotonic conditions or hypotonic low Cl − conditions (to activate the SPAK/OSR1 pathway), and then treated under the same conditions with the indicated doses of the SPAK/OSR1 CCT domain inhibitor STOCK1S-50699 [30] for an additional 30 min. The lysates were subjected to immunoprecipitation (IP) with the indicated KCC3A Site-1, Site-2 and Site-4 phospho-antibodies that recognize all KCC isoforms (see Figure 3). The immunoprecipitates were then immunoblotted (IB) employing the indicated specific KCC isoform antibodies. Cells lysates were also subjected to immunoblot analysis with the indicated total and phospho-specific antibodies. Similar results were obtained in three separate experiments. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.
isoforms. Protein phosphatase(s) and their associated regulatory subunits that tether these enzymes to their targets are likely to play a critical role in this pathway, illustrated by the fact that all KCC isoforms, as well as core WNK pathway components, are robustly dephosphorylated within 15 min of exposure of cells to hypotonic high K + conditions ( Figure 4C). Several previous studies have highlighted that several general protein phosphatase inhibitors [40,53] block activation of KCC isoforms. Previous work has also implicated protein phosphatase-4 in the dephosphorylation of NCC at the critical Thr 60 regulated by the WNK-SPAK/OSR1 pathway [54]. In future studies, it would be important to determine the identity of the protein phosphatase catalytic activity, as well as the targeting regulatory subunit, that acts on the KCC isoforms, and to determine further whether the activity of this phosphatase complex is regulated by osmotic conditions. It would also be important to define whether the same or different protein phosphatase complexes act on different KCC isoforms and components of the WNK signalling system.
The importance of co-ordinating cellular Cl − influx and efflux has long been appreciated [9,17]. The finding that SPAK/OSR1 kinases phosphorylate and thereby trigger activation of the Na + -driven Cl − influx co-transporters (NKCC1, NKCC2 and NCC) and also phosphorylate and inhibit K + -driven Cl − efflux co-transporters (KCC1, KCC2, KCC3 and KCC4), provides a simple mechanism to explain how these critical Na + and K + co-transporters are reciprocally and co-ordinately controlled ( Figure 8). Stimuli such as hypotonic low Cl − conditions that activate the WNK-SPAK/OSR1 pathway will induce net cellular Cl − absorption by promoting phosphorylation, thereby activating the N[K]CCs and concurrently inhibiting the KCCs. In contrast, stimuli that inactivate the WNK-SPAK/OSR1 pathway, such as hypotonic high K + conditions, will suppress cellular Cl − absorption by promoting dephosphorylation, thereby inhibiting the N[K]CCs and concurrently stimulating the KCCs. This coordinated and potent mechanism, with opposite effects on the main Cl − influx and Cl − efflux mediators involved in cellular Cl − homoeostasis, is of obvious interest for drug development. Our data suggest that this mechanism is druggable by the targeting of the CCT domain within SPAK/OSR1, which interferes with WNK activation. Given the anticipated net effects of WNK-SPAK/OSR1 signalling on increasing cellular Cl − influx, we propose that the inhibition of this pathway with kinase inhibitors resembling STOCK1S-50699 might be a novel potent strategy to enhance cellular Cl − extrusion. The phenotypes of human and mice with alterations in the WNK-SPAK/OSR1 pathway suggest that such a drug might be a novel K + -sparing diuretic with special potency as an anti-hypertensive, since it would suppress renal salt re-absorption (and via effects on intravascular volume, blood pressure) in a more co-ordinated and balanced manner than thiazide and loop diuretics, which only suppress the activity of NCC (thiazide) or NKCC2 (loop diuretics) individually. Moreover, given the recent interest and proof of concept of small molecule enhancers of neuronal Cl − extrusion via KCC2 activation to treat diseases featuring GABAergic (GABA is γ -aminobutyric acid) disinhibition secondary to impaired KCC2 activity [55,56], we propose that an especially potent way to enhance neuronal Cl − extrusion would be to concurrently inhibit Cl − influx via NKCC1 and activate Cl − efflux via KCC2 [57]; our results suggest this could be done by inhibiting WNK-SPAK/OSR1 signalling. These hypotheses are testable, given the existence of a SPAK/OSR1 are activated by WNK isoforms and then phosphorylate the NTD of the Na + -driven CCCs (NKCC1, NKCC2 and NCC), thereby stimulating influx of Cl − (and Na + ). Work undertaken in the present study establishes that SPAK/OSR1 also phosphorylates K + -driven CCCs (KCC1, KCC2, KCC3 and KCC4) at residues located in both their NTD and CTD, thereby stimulating efflux of Cl − (and K + ). Phosphorylation of Site-2 by SPAK/OSR1 greatly contributes to the regulation of the KCC activity. Site-1 phosphorylation also plays a major role in regulating KCC activity; the present study indicates that SPAK/OSR1 may not itself phosphorylate this site directly, but WNKs and/or SPAK/OSR1 may influence the activity of the kinase that phosphorylates Site-1. Further work is required to establish the roles that Site-3 phosphorylation within the NTD plays. Site-4 is only present in the KCC3 and KCC3A splice variants, and previous work suggests that phosphorylation at this site plays a minor role in regulating KCC activity [51].
WNK-SPAK/OSR1 inhibitor and phospho-specific antibody readouts of pathway activity, with potential implications for the therapeutic modulation of epithelial and neuronal ion transport.