WNK3 kinase is a positive regulator of NKCC2 and NCC, renal cation-Cl^- cotransporters required for normal blood pressure homeostasis

Abstract

WNK1 and WNK4 [WNK, with no lysine (K)] are serine-threonine kinases that function as molecular switches, eliciting coordinated effects on diverse ion transport pathways to maintain homeostasis during physiological perturbation. Gain-of-function mutations in either of these genes cause an inherited syndrome featuring hypertension and hyperkalemia due to increased renal NaCl reabsorption and decreased K⁺ secretion. Here, we reveal unique biochemical and functional properties of WNK3, a related member of the WNK kinase family. Unlike WNK1 and WNK4, WNK3 is expressed throughout the nephron, predominantly at intercellular junctions. Because WNK4 is a potent inhibitor of members of the cation-cotransporter SLC12A family, we used coexpression studies in Xenopus oocytes to investigate the effect of WNK3 on NCC and NKCC2, related kidney-specific transporters that mediate apical NaCl reabsorption in the thick ascending limb and distal convoluted tubule, respectively. In contrast to WNK4's inhibitory activity, kinase-active WNK3 is a potent activator of both NKCC2 and NCC-mediated transport. Conversely, in its kinase-inactive state, WNK3 is a potent inhibitor of NKCC2 and NCC activity. WNK3 regulates the activity of these transporters by altering their expression at the plasma membrane. Wild-type WNK3 increases and kinase-inactive WNK3 decreases NKCC2 phosphorylation at Thr-184 and Thr-189, sites required for the vasopressin-mediated plasmalemmal translocation and activation of NKCC2 in vivo. The effects of WNK3 on these transporters and their coexpression in renal epithelia implicate WNK3 in NaCl, water, and blood pressure homeostasis, perhaps via signaling downstream of vasopressin.

In the kidney, the regulation of net renal NaCl reabsorption is a major determinant of blood pressure and is the product of the coordinated function of diverse transcellular and paracellular electrolyte transport pathways distributed along the nephron (1). A key step in this process is the apical entry of Na⁺ with or without Cl^-. In the thick ascending limb of Henle (TAL) and the distal convoluted tubule (DCT), this is mediated by related kidney-specific electroneutral cation/Cl^- cotransporters, the Na-K-2Cl cotransporter NKCC2 (encoded by SLC12A1) in the TAL and the Na-Cl cotransporter NCC (encoded by SLC12A3) in the DCT. The unrelated epithelial Na⁺ channel (ENaC; encoded by SCNN1A, SCNN1B, and SCNN1G) mediates electrogenic Na⁺ reabsorption in the connecting tubule and collecting duct (CD); ENaC activity is accompanied by paracellular Cl^- reabsorption and secretion of K⁺ and H⁺ in these nephron segments. Inherited variation in the activities of these flux mediators or their regulators alters blood pressure in humans, with increased or decreased net NaCl reabsorption raising or lowering blood pressure (1). For example, loss-of-function mutations in the genes encoding NKCC2 and NCC cause Bartter's and Gitelmann's syndromes, respectively, inherited disorders featuring low blood pressure due to renal NaCl wasting (2, 3). Although a number of hormones, such as vasopressin and aldosterone, regulate these transport proteins to maintain NaCl and water and blood pressure homeostasis, the transducers that link hormonal signaling to downstream targets and the mechanisms coordinating the activities of multiple transporters and/or channels are poorly understood (4).

Genetic analysis can provide fundamental insight into the function of complex networks by identifying genes and pathways that, when mutated, disrupt the integration of normally coordinated systems (1). Recent studies have identified WNK1 and WNK4 [WNK, with no lysine (K)] (encoded by PRKWNK1 and PRKWNK4, respectively) as serine-threonine protein kinases that have the biochemical properties and physiologic effects of such integrative regulators (5). Missense mutations in WNK4 cause pseudohypoaldosteronism type II (PHAII), a disease featuring hypertension and hyperkalemia (high serum K⁺ levels) due to a coupled increase in renal NaCl reabsorption and deficiency in renal K⁺ secretion (5). Subsequently, WNK4 has been shown to be a multifunctional regulator of diverse Na⁺, K⁺, and Cl^- flux pathways. Wild-type WNK4 is an inhibitor of NCC, NKCC1, the K⁺ channel ROMK1 (or Kir1.1; encoded by KCNJ1), and an activator of paracellular Cl^- flux; some of these effects are kinase-dependent, whereas others are independent of WNK4's catalytic activity (6-12). Importantly, disease-causing missense mutations in WNK4 cluster within a highly conserved acidic domain (5) and have sharply divergent effects on these downstream targets; PHAII mutations eliminate inhibition of NCC and increase paracellular Cl^- permeability, whereas simultaneously increasing inhibition of ROMK1 (6-12). Together, these effects can account for the observed increase in NaCl reabsorption and the decrease in K⁺ secretion seen in affected subjects. These findings demonstrate that WNK4 is a multifunctional molecular switch capable of having opposing effects on multiple ion flux pathways via independent mechanisms, precisely the sort of properties one would expect for an integrator of systems (13). Mutations that increase expression of WNK1 cause a similar phenotype (5). Recent evidence suggests WNK1 is an upstream regulator of WNK4 at NCC (7) and may also regulate ENaC through SGK1 (14, 15).

WNK kinases are characterized by the substitution of cysteine for lysine at a highly conserved residue in the catalytic domain; they are found in both animal and plant species (16, 17). A total of four such kinases exist in the human genome (5, 18), each sharing significant homology in the kinase domain, an autoinhibitory domain, two putative coiled-coil domains, and a short acidic domain. To date, little is known about the two other members of the WNK kinase gene family, WNK2 and WNK3. Thus, it is unknown whether these other WNK family members act in the same pathways as WNK1 and WNK4, or whether they have broadly different functions. If operating in the same pathways, their functions could either be redundant to or different from those of WNK1 and WNK4. WNK3 transcripts are most abundant in brain and kidney (18, 19), but the protein has not been localized, and its function has not been characterized. Herein, we show WNK3's renal localization, unique among WNK kinases, and explore its biochemical and physiologic functions. Our findings indicate that WNK3 has properties distinct from those of WNK1 and WNK4 and suggest WNK3 participates in the coordinated regulation of NKCC2 and NCC, kidney-specific cation/Cl^- cotransporters necessary for electrolyte and blood pressure homeostasis.

Methods

cDNA Constructs. 5′ EcoRI and 3′ XbaI sites were engineered into PCR primers that were used to amplify WNK3 from a full-length human WNK3 cDNA clone (OriGene, Rockville, MD); this amplicon was subcloned into pGH19 (9). 5′ NotI and 3′ EcoRI sites and a C-terminal hemagglutinin (HA) tag were engineered into PCR primers that were used to amplify WNK3 from the Origene WNK3 cDNA clone, and this product was subcloned into pCDNA3.1- (Invitrogen). QuikChange (Stratagene) was used to introduce the D294A or Q545E mutations into pGH19-WNK3 and pcDNA3.1-WNK3-HA. All constructs were verified by DNA sequencing.

pSPORT1-rNCC was used for ²²Na⁺ flux studies (20). pSPORT1-GFP-rNCC was used for quantitation of EGFP-NCC surface expression (6). pSPORT1-rNKCC2 (20) was used for ⁸⁶Rb⁺ influx studies.

Antibodies. Anti-WNK3 antibody was obtained from Alpha Diagnostics (San Antonio, TX). Other antibodies used were: anti-ZO-1 (21), anti-Megalin (gift of D. Biemesderfer, Yale University, New Haven, CT), anti-E-cadherin antibody (Santa Cruz Biotechnology), anti-Tamm-Horsfall protein, anti-Calbindin (Swant, Bellinzona, Switzerland), anti-Aquaporin-2 (Santa Cruz Biotechnology), anti-HA (Santa Cruz Biotechnology), anti-rabbit IgG (Zymed), anti-NKCC2 (T9 antibody), and anti-phospho NKCC2 (R5 antibody) (gifts of B. Forbush, Yale University), and affinity-purified secondary antibodies conjugated to the CY2, CY3, or CY5 fluors (Jackson ImmunoResearch).

Transfections, SDS/PAGE, and Immunoblotting. Transfections of WNK3 constructs into COS-7 cells were performed by using Lipofectamine 2000 (Invitrogen). Lysates of transfected COS-7 cells or lysates from mouse or human tissues were solubilized in sample buffer, and proteins were separated by SDS/PAGE, transferred to nitrocellulose, blocked, and probed with anti-WNK3 or anti-HA (each at 1:1,000 dilution), as described (6). Membranes were then washed, incubated with secondary antibody, and processed with the enhanced chemiluminescence system (Amersham Pharmacia), as described (6).

Autophosphorylation Assays. Autophosphorylation assays were performed as described (16). WNK3-HA was expressed in COS-7 cells, and total cell lysates were incubated with anti-HA agarose beads (Santa Cruz Biotechnology). Beads were washed in lysis buffer (6) and resuspended in kinase buffer (16). Reactions were initiated by adding beads with WNK3-HA and 5 μCi γ³²P ATP (1 Ci = 37 GBq) (Amersham Pharmacia) in kinase buffer. The reaction was incubated at 30°C for 30 min and stopped by addition of sample buffer (6). Beads with WNK3-HA were boiled in sample buffer, and phosphorylated proteins were resolved by SDS/PAGE and visualized by autoradiography.

Immunolocalization Studies. Studies were approved by the Yale University Animal Care and Use Committee. Mice were killed by cervical dislocation. Excised tissues were prepared and sectioned as described (21). Tissue sections were processed with primary and secondary antibodies and visualized by immunofluorescence or confocal microscopy (21). Results were consistent among three different mice. Anti-WNK3 immunostaining was competed with a 3-fold molar excess of the immunizing peptide; staining with secondary antibody alone revealed no signal.

Functional Assays with NCC and NKCC2. Xenopus laevis oocytes were harvested and injected with cRNA of NCC or NKCC2 alone or together with cRNA of wild-type, kinase-dead, or PHAII-like mutant WNK3, essentially as described (8). After 4 days of incubation, metolazone-sensitive ²²Na⁺ influx (for NCC; ref. 6) or bumetanide-sensitive ⁸⁶Rb⁺ influx (for NKCC2; ref. 22) was determined as described. NCC and NKCC2 measurements were performed in isotonic conditions (200-210 mM). In each experiment, ≈15 oocytes were tested in each group; results were highly reproducible across at least four independent experiments for each condition. The significance of differences between groups of oocytes was assessed by two-tailed Student's t test or one-way ANOVA with multiple comparisons using Bonferroni correction, as appropriate.

Surface Expression Studies. Oocytes were injected with EGFP-tagged NCC cRNA alone or together with wild-type or mutant WNK3 cRNAs, incubated for 3-4 days, and membrane surface expression of GFP-NCC was assayed by laser-scanning confocal microscopy as described (6, 23). Total membrane fluorescence intensity was calculated for each imaged oocyte by using sigmascan pro software (Jandel, San Rafael, CA; ref. 6). GFP-NCC results are data combined from four experiments; >12 oocytes were injected per experimental group, and each experiment used oocytes from a different frog. For each injection series, the mean fluorescence value for GFP-NCC alone was set at 100%, and other values were expressed as percentage of this value. The significance of differences between groups was assessed by two-tailed Student's t test.

NKCC2 Phospho-Protein Studies. For phospho-protein analysis, oocytes injected with indicated constructs were incubated as described above and exposed to an extracellular tonicity of either 200 or 380 mosM; the higher osmolarity was achieved by addition of sucrose to the medium. At the end of the incubation period, four oocytes per group were immediately homogenized by pipetting in 100 μl of ice-cold antiphosphatase solution: 150 mM NaCl/30 mM NaF/5 mM EDTA/15 mM Na₂HPO₄/15 mM pyrophosphate/20 mM Hepes, pH 7.2) with 1% Triton X-100 and a protease inhibitor mixture. The homogenate was cleared by centrifugation and supernatants collected for Western blot analysis. The equivalent of 6 μl of lysate was loaded per lane. The previously characterized anti-NKCC2 antibody T9 and the antiphospho-NKCC2 antibody R5 (24, 25) were used to detect total and phosphorylated NKCC2, respectively.

Results

WNK3 Localizes to Intercellular Junctions Throughout the Nephron. WNK3 transcripts are expressed in kidney (19). We explored the cellular and subcellular renal localization of WNK3. Light immunofluorescence microscopy of kidney sections stained with an antibody specific for WNK3 (Fig. 5, which is published as supporting information on the PNAS web site) revealed that WNK3, like WNK1 and WNK4, is confined to nephrons and localizes predominantly to intercellular junctions, as demonstrated by its colocalization with zona-occludens-1 (ZO-1), a tight junction protein (Fig. 1). Although WNK1 and WNK4 are confined to the distal nephron (DCT and CD) (5), WNK3 is present in all nephron segments (Fig. 1 A), with highest expression in the proximal convoluted tubule (PCT, Fig. 1B) and TAL (Fig. 1 C-E), and lower levels of expression in the DCT and CD (Fig. 1 F-H), as demonstrated by costaining experiments with antibodies that serve as markers of these nephron segments (see Methods). Confocal immunofluorescence microscopy of anti-WNK3-stained kidney sections demonstrates that WNK3 expression extends along the lateral membrane from the level of the tight junction to the adherens junction in all nephron segments (Fig. 6, which is published as supporting information on the PNAS web site).

Fig. 1.

WNK3 is expressed at intercellular junctions along the nephron. Frozen mouse kidney sections were stained with anti-WNK3 antibody as described in Methods and visualized with immunofluorescence light microscopy. (A) WNK3 is present in all nephron segments. Low-power view of renal tubules in cross section stained with anti-WNK3 antibody reveals expression of WNK3 in all nephron segments. It is apparent that WNK3 predominantly localizes to intercellular junctions. (Original magnification, ×200.) (B) WNK3 expression in PCT. Tubule segments were stained with anti-WNK3 antibody and identified by costaining adjacent sections with anti-Megalin antibody, a PCT marker (not shown). (Original magnification, ×630.) (C-E) WNK3 (red) and ZO-1 (green) immunostaining in the TAL. Tubule segments were determined by costaining experiments with anti-WNK3 and anti-Tamm-Horsfall protein, a TAL marker (data not shown). WNK3 staining overlaps with ZO-1, a marker of tight junctions. (Original magnification, ×630.) (F) WNK3 expression in the DCT and connecting tubule (CNT), as determined by costaining experiments with anti-WNK3 and anti-Calbindin D-28K, a DCT/CNT marker (data not shown). (Original magnification, ×400.) (G) WNK3 expression in the cortical CD (CCD), determined by costaining with anti-WNK3 (red) and anti-Aquaporin-2 (AQP2), a CCD marker (blue). (Original magnification, ×630.) (H) WNK3 (red) and ZO-1 (green) immunostaining in the CCD. Tubule segments were determined by costaining experiments with anti-WNK3 and anti-AQP2. WNK3 staining overlaps with ZO-1, a marker of tight junctions. (Original magnification, ×700.)

WNK3 Is an Active Kinase. To examine WNK3's potential kinase activity, HA-tagged WNK3 was expressed in COS-7 cells and immunoprecipitated from cell lysates (see Methods). Incubation of immunoprecipitated WNK3 with γ³²P-labeled ATP, followed by electrophoresis, revealed phospholabeling of a ≈200-kDa protein, the expected size of WNK3 (Fig. 2A). The experiment was repeated with a WNK3 mutant harboring a missense mutation in its catalytic domain (WNK3-D294A); aspartate at this position is highly conserved among kinases because of its role in Mg²⁺ binding, and alanine substitution at this site impairs the catalytic activity of WNK1 and other kinases (16, 17). WNK3-D294A shows virtually no autophosphorylation, indicating the dependence of phospholabeling on WNK3's kinase activity (Fig. 2 A). In contrast, a Q545E missense mutation in WNK3 that mimics a PHAII-causing mutation in WNK4 (5) does not alter the phospholabeling of WNK3 (Fig. 2 A). These observations establish WNK3's kinase activity and validate the use of these kinase-active and -inactive (“dead”) forms of WNK3 in subsequent experiments.

Fig. 2.

WNK3 kinase regulates the renal NCC and NKCC2 cotransporters. (A) WNK3 is an active kinase. Wild-type WNK3 (WT), kinase-dead WNK3 (kin-), and PHAII-like mutant WNK3 constructs were tagged with HA and expressed in COS-7 cells, purified by immunoprecipitation, and incubated with γ-³²P ATP. The products were separated by SDS/PAGE, exposed to film, and also separately stained with anti-HA antibodies. Phosphorylated WNK3 (³²P-WNK3) species were seen at 200 kDa (the size of WNK3-HA) in assays with WT and PHAII-like mutant WNK3 but were markedly reduced in assays with kinase-dead WNK3. Western blots with anti-HA and IgG demonstrate equivalent protein loading. (B) WNK3 regulates NCC. Xenopus oocytes were injected with cRNAs encoding NCC alone or in combination with wild-type WNK3, kinase-dead WNK3 (kin-), or PHAII-like WNK3. Metolazone-sensitive ²²Na⁺ influx was measured as described in Methods. Results are expressed as mean ± SE of metolazone-sensitive ²²Na⁺ influx. ^*, P < 0.0001 vs. NCC alone. WNK3 markedly increases metolazone-sensitive ²²Na⁺ influx, kinase-dead WNK3 marked inhibits this activity, whereas PHAII-WNK3 behaves like wild-type WNK3. (C and D) WNK3 regulates NCC surface expression. (C) Oocytes were injected with cRNAs encoding EGFP-tagged NCC alone or in combination with wild-type or mutant WNK3. Surface expression of NCC was quantitated by confocal microscopy as described in Methods. Mean fluorescence seen in oocytes expressing GFP-NCC alone is expressed as 100%; other groups are expressed as a percentage of this value. Effects of WNK3 constructs on NCC surface expression closely parallel its effects on ²²Na⁺ flux. ^*, P < 10^-6 vs. NCC alone. (D) Examples of confocal microscopy of oocytes expressing GFP-tagged NCC alone or with wild-type or kinase-dead WNK3. (E) WNK3 regulates NKCC2. Oocytes were injected with cRNAs encoding NKCC2 alone or in combination with wild-type or mutant WNK3; bumetanide-sensitive ⁸⁶Rb⁺ influx was measured as described in Methods. Mean ± SE of bumetanide-sensitive ⁸⁶Rb⁺ influx is shown for each group in a representative experiment. As for NCC, WNK3 increases NKCC2 activity, kinase-dead WNK3 inhibits NKCC2, activity and PHAII-like WNK3 behaves like wild-type kinase.

WNK3 Regulates NCC and NKCC2 by Altering Their Surface Expression. The discrete localization of WNK3 to nephrons, along with its localization to Cl^- transporting epithelia outside the kidney (see Kahle et al., ref. 26) and the genetic and physiologic evidence that WNK1 and WNK4 modulate the activity of a number of mediators of electrolyte flux in vivo, suggests specific potential targets of WNK3 function. In particular, members of the SLC12A family of cation/Cl^- cotransporters are of interest, because they are known to play important roles in the entry and/or exit of Na⁺, K⁺, and Cl^- in kidney epithelia, and WNK4 regulates a number of different SLC12A members (13).

Guided by these considerations, we investigated the effect of WNK3 on NCC and NKCC2, kidney-specific SLC12A transporters that mediate apical NaCl reabsorption in the TAL and DCT, respectively, sites that express WNK3. In each case, wild-type WNK3 and WNK3 harboring the D294A mutation (kinase-inactive WNK3) were tested for their effects using coexpression studies in X. laevis oocytes, a well characterized system that has been used for the functional cloning and physiologic characterization of these transport proteins and has also proved useful for defining the mechanisms of their regulation (20). We also tested the effect of a WNK3 mutant harboring the Q565E mutation, which mimics a mutation that causes PHAII in WNK4 (PHAII-like WNK3).

In contrast to WNK4's inhibitory effect on NCC in oocytes (6, 7, 12), coexpression of WNK3 with NCC led to a dramatic >3-fold increase in NCC activity, as measured by metolazone-sensitive ²²Na⁺ uptake (P < 10^-6, Fig. 2B). Surprisingly, kinase-dead WNK3 not only failed to increase NCC activity but also markedly inhibited NCC activity by ≈85% (P < 10^-6; Fig. 2B). In contrast, a PHAII-like WNK3 mutant behaved like wild-type WNK3 on NCC (Fig. 2B). Similar effects on ²²Na⁺ influx were seen when WNK3 constructs were coexpressed with NCC tagged with EGFP-NCC (data not shown).

WNK3's modulation of NCC activity is achieved by altering the localization of NCC at the plasma membrane. Coexpression of WNK3 with GFP-NCC increased GFP-NCC surface expression ≈3-fold (P < 10^-6, Fig. 2C), whereas kinase-dead WNK3 had the opposite effect, decreasing GFP-NCC by 80% (P < 10^-6, Fig. 2 C and D). Thus, WNK3 is a potent regulator of NCC activity, increasing NCC surface expression in its kinase-active state but decreasing NCC surface expression when catalytically inactive.

Similarly, coexpression of WNK3 with NKCC2 in oocytes resulted in a ≈2-fold increase in NKCC2 activity, as measured by bumetanide-sensitive ⁸⁶Rb⁺ influx (P < 10^-6, Fig. 2E). Similar to WNK3's effect on NCC, kinase-dead WNK3 inhibited NKCC2 activity ≈70% (P < 10^-6, Fig. 2E), whereas PHAII-like WNK3 activated NKCC2 similar to wild-type WNK3 (Fig. 2E).

Together, these data reveal that WNK3 potently increases the activity of the mediators of apical NaCl entry in the DCT and TAL. The magnitudes of these effects are previously undescribed for any other protein kinase; moreover, WNK3 is shown to regulate both of these transporters (20). WNK3's activation is specific, because its activity is reversed from activation to inhibition by a single amino acid substitution in WNK3's kinase domain. WNK3's specificity is further demonstrated by the fact it has no effect on the activity of ENaC in Xenopus oocytes, or on paracellular Cl^- flux in Madin-Darby canine kidney II cells (data not shown). These other renal NaCl transport processes are regulated by WNK1 and WNK4, respectively (10, 11, 15).

WNK3 Regulates NKCC2 Phosphorylation. To further explore the mechanism underlying WNK3's action on these transporters, we focused on NKCC2, whose activation in response to vasopressin is associated with phosphorylation of Thr-184 and Thr-189 in its cytoplasmic N terminus (24, 25). Phosphorylation of the paralagous residues is necessary for and closely parallels NKCC1 activation and is also conserved in NCC (27). We monitored phosphorylation of NKCC2 at these sites with R5, an antibody that specifically recognizes phosphorylation of Thr-184 and Thr-189 (24, 25). In the absence of WNK3, NKCC2 phosphorylation increases from negligible levels under hypotonic conditions (200 mosM) to high levels under hypertonic conditions (380 mosM). In contrast, coexpression with WNK3 results in robust NKCC2 phosphorylation under hypotonic and hypertonic conditions (Fig. 3). Conversely, when kinase-dead WNK3 is expressed with NKCC2, there is a dramatic reduction in NKCC2 phosphorylation compared with the level seen with NKCC2 expression alone in each condition of tonicity (Fig. 3). This effect can account for the observed inhibition of NKCC2 by kinase-dead WNK3. These findings indicate that the effects of WNK3 on NKCC2 can be accounted for by altered phosphorylation of regulatory sites of NKCC2.

Fig. 3.

WNK3 modulates the phosphorylation of NKCC2. Xenopus oocytes were injected with the indicated constructs and incubated at varying extracellular osmolarities as indicated. After incubation oocytes were lysed and Western blotting was performed by using the R5 (anti-phospho-NKCC2) or T9 (anti-NKCC2) antibodies as described in Methods. Phosphorylation of NKCC2 normally increases from negligible levels under hypotonic conditions (200 mM) to complete phosphorylation under hypertonic conditions (380 mM). In contrast, coexpression of NKCC2 with kinase-active WNK3 results in robust phosphorylation of NKCC2 at all osmolarities. Expression of kinase-dead WNK3 results in marked reduction of NKCC2 phosphorylation under hyperosmolar conditions.

Discussion

We have shown that WNK3, a member of the WNK kinase family, has effects that are distinct from those of WNK1 and WNK4, demonstrating that its activity is not redundant to these other family members. WNK3 is expressed at intercellular junctions along the length of the nephron, including the TAL and DCT, in contrast to WNK1 and WNK4, which are restricted to the aldosterone-sensitive distal nephron (5). Unlike the inhibitory effect of WNK4, kinase WNK3 is a potent activator of NCC and also NKCC2; these distinct effects suggest that either the upstream regulators of these kinases or the timing of the effects must be distinct to avoid a futile cycle. Kinase-inactive WNK3's action is reversed, strongly inhibiting NKCC2 and NCC activity. Wild-type and kinase-inactive WNK3 regulate transporter activity by altering transporter expression at the plasma membrane.

WNK3 regulates the phosphorylation of Thr-184 and Thr-189 of NKCC2; these sites are conserved in NCC. The phosphorylation state of these threonines correlates with NKCC2's activity and plasmalemmal surface expression in vitro and in vivo (24, 25). Our experiments suggest WNK3 regulates transporter activity by altering protein trafficking via modulation of the transporter phosphorylation state. The reduced phosphorylation of NKCC2 induced by kinase-dead WNK3 cannot be attributed to simple loss of function, because the level of phosphorylation is lower than that seen in the absence of WNK3. This effect of kinase-dead WNK3 could be accounted for by direction of a phosphatase activity to the target protein, inhibition of a kinase that normally maintains phosphorylation of the target, or both. Further experiments will be needed to clarify WNK3's mechanism of action.

WNK3 is not only the most potent activator reported for NKCC2 or NCC; it is also the first kinase reported to regulate both of these transporters (20). In the kidney (Fig. 4), kinase-active WNK3's activities in the TAL and DCT are inferred to promote increased renal NaCl reabsorption in nephron segments that normally mediate the reabsorption of ≈30% and ≈7% of the filtered NaCl load, respectively. The absence of an effect of WNK3 on ENaC and paracellular Cl^- flux, processes that occur in the more distal CD, suggest that WNK3 activity can shift NaCl reabsorption toward more proximal nephron segments, a key difference from WNK1 and WNK4, whose action has been shown in part to lie downstream of aldosterone in more distal parts of the nephron (14). Nonetheless, the absence of effect on these latter pathways must be interpreted with some caution, because it is possible that additional cellular components required for a regulatory effect of WNK3 are absent from the oocyte and Madin-Darby canine kidney cell systems used in this study. The activity of WNK3 on NKCC2 and its presence in the TAL, where vasopressin is known to regulate NKCC2 activity (24, 28, 29), plus the fact that WNK3 and vasopressin induce phosphorylation of the same N-terminal threonines (24), suggests that WNK3 might lie downstream of vasopressin signaling. This may apply to NCC in the DCT as well (29). Because the reabsorptive capacity of the TAL/DCT for NaCl via NKCC2/NCC is large, activation of this pathway by WNK3 under conditions of intravascular volume depletion and/or high serum osmolarity could promote increased NaCl and water reabsorption, thereby defending intravascular volume and plasma tonicity. The effects of the WNK3 kinase-inactivating mutation on NKCC2 and NCC are intriguing and may mimic a normal in vivo phenomenon, perhaps achieved by phosphorylation or altered interaction with other proteins/ligands. In its inactive state, basal repression of NKCC2 and NCC might take place, whereas in its active state, WNK3 could facilitate robust NaCl and water reabsorption.

Fig. 4.

Proposed role of WNK3 in the kidney. By regulating apical NaCl entry in the TAL and DCT of the nephron, WNK3 could modulate NaCl and water reabsorption and therefore blood pressure. Kinase-active WNK3 might increase NaCl reabsorption, whereas kinase-inactive WNK3 (WNK3 kin-) might inhibit NaCl reabsorption. These active/inactive states of WNK3 may be achieved dynamically by ligands (e.g., downstream of vasopressin) binding to or dissociating from the kinase.

WNK3's regulation of transporters that are expressed in other nephron segments remains to be fully defined. In contrast to WNK1, WNK3 does not appear to regulate ENaC, the main mediator of NaCl reabsorption in the connecting tubule and CD. Potential targets in the CD include the K⁺ channel ROMK1 and the water channel aquaporin 2. The targets of WNK3 in the PCT are also unknown. KCC3 and KCC4, K-Cl cotransporters in the SLC12A family highly expressed in the PCT (20), are prime candidates, as is the Cl^-/formate exchanger CFEX (9). Furthermore, the functional significance of WNK3 at the tight junction is unclear; whereas WNK4 facilitates paracellular Cl^- flux in Madin-Darby canine kidney II cells (10, 11), WNK3 has no effect in the same model system. These issues require future investigation.

Because individual WNK kinases have different expression profiles, unique target specificities, and opposing effects at common targets, it is tempting to speculate that different WNKs may be recruited individually or in combination, enabling modulation of the activities of their targets over a wide range in response to physiological stimuli. For example, it would be logical if WNK4 signaling was downstream of angiotensin II signaling and WNK3 was downstream of vasopressin signaling. Mutations in humans in WNK1 and WNK4 have large effects to alter the balance between NaCl reabsorption and K⁺ secretion, demonstrating their physiologic relevance. Our findings suggest that WNK3 is integral to the regulation of NKCC2 and NCC, proteins necessary for normal blood pressure homeostasis. These observations add to the growing recognition that WNK family members play diverse and important roles in integrated electrolyte homeostasis.