The Na+:Cl– Cotransporter Is Activated and Phosphorylated at the Amino-terminal Domain upon Intracellular Chloride Depletion*

The renal Na+:Cl– cotransporter rNCC is mutated in human disease, is the therapeutic target of thiazide-type diuretics, and is clearly involved in arterial blood pressure regulation. rNCC belongs to an electroneutral cation-coupled chloride cotransporter family (SLC12A) that has two major branches with inverse physiological functions and regulation: sodium-driven cotransporters (NCC and NKCC1/2) that mediate cellular Cl– influx are activated by phosphorylation, whereas potassium-driven cotransporters (KCCs) that mediate cellular Cl– efflux are activated by dephosphorylation. A cluster of three threonine residues at the amino-terminal domain has been implicated in the regulation of NKCC1/2 by intracellular chloride, cell volume, vasopressin, and WNK/STE-20 kinases. Nothing is known, however, about rNCC regulatory mechanisms. By using rNCC heterologous expression in Xenopus laevis oocytes, here we show that two independent intracellular chloride-depleting strategies increased rNCC activity by 3-fold. The effect of both strategies was synergistic and dose-dependent. Confocal microscopy of enhanced green fluorescent protein-tagged rNCC showed no changes in rNCC cell surface expression, whereas immunoblot analysis, using the R5-anti-NKCC1-phosphoantibody, revealed increased phosphorylation of rNCC amino-terminal domain threonine residues Thr53 and Thr58. Elimination of these threonines together with serine residue Ser71 completely prevented rNCC response to intracellular chloride depletion. We conclude that rNCC is activated by a mechanism that involves amino-terminal domain phosphorylation.

The renal Na ϩ :Cl Ϫ cotransporter (NCC 4 or TSC, gene symbol SLC12A3, locus identification number 6559) that is expressed at the apical membrane of the mammalian distal convoluted tubule represents the major salt transport pathway in this segment of the nephron (1)(2)(3)(4). Its essential role in preserving the extracellular fluid volume and blood pressure has been established by the identification of inactivating mutations of the SLC12A3 gene as the cause of Gitelman's disease (5,6), an inherited disorder featuring arterial hypotension, renal salt wasting, hypokalemic metabolic alkalosis, hypocalciuria, and hypomagnesemia. In addition, a defect in NCC regulation by serine/threonine kinases WNK1 and WNK4 has been implicated in the pathogenesis of a salt-dependent form of human hypertension known as pseudohypoaldosteronism type II (PHAII) (7,8), which features marked sensitivity to hydrochlorothiazide and a clinical picture that is a mirror image of Gitelman's disease (9). NCC is the pharmacological target of thiazide-type diuretics that are currently recommended by the Joint National Committee VII for the detection, evaluation, and treatment of high blood pressure as the first line treatment of arterial hypertension either as the unique drug or in combination with other antihypertensive agents (10). Despite the importance of NCC for cardiovascular and renal physiology, pharmacology, and pathophysiology, little is known about the mechanisms by which NCC activity is regulated.
NCC belongs to the superfamily of electroneutral cationcoupled chloride cotransporters (SLC12) from which seven members have been identified at both the functional and molecular level. NCC, together with two isoforms of the Na ϩ : K ϩ :2Cl Ϫ cotransporter, NKCC1 and NKCC2, compose the sodium-driven branch (NKCCs), and four isoforms of the K ϩ :Cl Ϫ cotransporter compose the potassium-driven branch (KCCs). Because these cotransporters are involved in the regulation of cell volume and/or in clamping the intracellular chloride concentration [Cl Ϫ ] i , it has been proposed that their activity is regulated by changes in cell volume and/or [Cl Ϫ ] i by means of phosphorylation/dephosphorylation pathways (for review, see Refs. [11][12][13][14][15]. Several lines of evidence suggest that phosphorylation activates NKCCs and inhibits KCCs cotransporters, whereas dephosphorylation inhibits NKCCs and activates KCCs cotransporters. For instance, cell shrinkage, low intracellular chloride concentration, and protein phosphatase inhibitors activate NKCC1. Studies in this last cotransporter led to the identification of three amino-terminal threonine residues that become phosphorylated under such stimulatory conditions (16,17). These threonine residues participate also in the stimulation of NKCC1 and NKCC2 by serine/threonine kinase WNK3 (18,19) or WNK1-WNK4/STE20 kinases pathways (20,21) as well as NKCC2 by vasopressin in thick ascending limb cells (22). Little is known, however, about NCC regulation. We have shown that rNCC is partially inhibited by cell swelling (23) or WNK4 (7) and is remarkably activated by WNK3 (19), suggesting that, like NKCC1 and NKCC2, NCC could be regulated by cell volume, [Cl Ϫ ] i , or WNK kinases, at least in part through phosphorylation of the conserved amino-terminal domain threonine residues. Here we show that rNCC activity and amino-terminal domain phosphorylation is increased by [Cl Ϫ ] i depletion strategies. rNCC activation is completely prevented when the amino-terminal domain threonine residues Thr 53 and Thr 58 and serine residue Ser 71 are eliminated, suggesting that these amino acid residues are absolutely required for such regulation.

EXPERIMENTAL PROCEDURES
Clones and Mutagenesis-We used the rat NCC and human KCC2 cDNAs that we cloned previously from rat kidney and human brain, respectively (24,25). All site directed mutations were introduced by using the QuikChange site directed mutagenesis system (Stratagene). Automatic DNA sequencing was used to confirm all mutations. All primers used for mutagenesis were custom made (Sigma).
Western Blotting-Western blot was used to compare the amount of NCC protein in cRNA-injected oocytes exposed to the intracellular chloride depletion maneuvers described above. Immunoblots were performed using a rabbit polyclonal anti-NCC antibody kindly provided by Mark Knepper, National Institutes of Health, (27) following our previously published protocol (28). In brief, groups of 15 oocytes exposed to each maneuver were homogenized in 4 l/oocyte of homogenization buffer, centrifuged twice at 100 ϫ g for 10 min at 4°C, and supernatant was recollected. Oocyte protein (equivalent to one oocyte per lane) was heated in sample buffer containing 6% SDS, 15% glycerol, 0.3% bromphenol blue, 150 mM Tris, pH 7.6, and 2% ␤-mercaptoethanol, resolved by SDS-PAGE (7.5%). Proteins were transferred to a polyvinylidene difluoride membrane and exposed overnight at 4°C to the rabbit polyclonal anti-NCC antibody diluted 1:1500 in blocking buffer, TTBS (2.24 g/liter Tris-base, 8 g/liter NaCl, 0.1% Tween, pH 7.6) 0.2%. Membranes were washed and incubated with horseradish peroxidase-conjugated secondary (anti-rabbit) antibody (Alpha Diagnostic Intl.) diluted 1:2000 in blocking buffer and washed again. Bands were detected by using ECL plus Western blotting detection system (Amersham Biosciences).
Assessment of the rNCC Expression at the Oocytes Plasma Membrane-Surface expression of wild type or mutant NCC (see below) was determined with confocal microscopy by assessing the surface fluorescence in Xenopus oocytes using an amino terminus enhanced green fluorescent protein (EGFP)-NCC fusion construct that we have validated previously (7, 19, 28 -30). In this construct, the EGFP was fused in frame to the amino-terminal domain of NCC. Xenopus oocytes were microinjected with water as control or with EGFP-wild type-rNCC or EGFP-mutant-rNCC cRNA. Four days later, oocytes were monitored for EGFP fluorescence in the surface of the oocytes using a Zeiss laser scanning confocal microscope (objective lens ϫ10, Nikon). Excitation and emission wavelengths used to visualize EGFP fluorescence were 488 and 515-565 nm, respectively. For densitometry analysis, the plasma membrane fluorescence was quantified by determining the pixel intensity around the entire oocytes circumference using SigmaScan Pro image analysis software. Western blot was performed in proteins extracted from the EGFP-NCC-injected oocytes following the procedures described above.
NCC Phospho-antibody Studies-We used the previously characterized R5 antibody (17) that was raised to detect phosphorylation of residues Thr 212 and Thr 217 in human NKCC1 (shark Thr 184 and Thr 189 ), a generous gift from B. Forbush, Yale University. R5-antibody is also useful to detect phosphorylation of corresponding residues Thr 99 and Thr 104 in rabbit NKCC2 (31). Through sequence comparison and mutation analysis (see Fig. 3 in "Results and Discussion"), we found that for rNCC to be recognizable by the R5 antibody at least tyrosine at position 56 in rat NCC must be converted to the histidine found in NKCCs. Therefore, using custom primers we constructed an rNCC single mutant Y56H-NCC or a quadruple mutant L49Y-Y56H-I59M-V61A, this latter containing all significant differences between NCC and NKCC2. For functional analysis oocytes injected with appropriate cRNA constructs were incubated in the same experimental solutions and for the same time as those described above. At the end of the incubation period, 4 oocytes per group were homogenized in 100 l of ice-cold antiphosphatase solution (150 mM NaCl, 30 mM NaF, 5.0 mM EDTA, 15 mM Na 2 HPO 4 , 15 mM pyrophosphate, 20 mM HEPES, pH 7.2) with 1% Triton X-100 and a protease inhibitor mixture. Then, homogenate was cleared by centrifugation, and supernatants were subjected to Western blotting. The equivalent to 6 l of lysate was loaded per lane. The analysis was repeated four times, and functional expression was assessed in parallel for each experiment.
To test for the specificity of R5 antibody signal with NCC samples, we carried out alkaline phosphatase treatment of oocyte homogenates. Briefly, after 16-h incubation in ND96 or Cl Ϫ -free solutions, oocytes were lysed in AlkPhos solution (50 mM Tris, pH 8.0, 1 mM EDTA, 100 mM NaCl, 2 mM MgCl 2 , 1% Triton, 0.2% SDS) to which protease inhibitors and calyculin A were added to avoid protein degradation and PP1-mediated dephosphorylation. Then, homogenates were split in two, and 5 units of phosphatase alkaline were added to one of the samples. Homogenates were incubated for 1 h at 37°C. The dephosphorylation reaction was terminated by adding the same volume of 2x sample buffer. 12 l of this solution were loaded per lane and was subjected to SDS-PAGE electrophoresis and Western blotting with R5 antibody.
In Vitro cRNA Translation-To prepare cRNA for microinjection, each of the wild-type or mutant cDNA was digested at the 3Ј end using NotI or NheI from New England Biolabs (Carlsbad, CA), and cRNA was transcribed in vitro using the T7 RNA polymerase mMESSAGE mMACHINE TM (Ambion) transcription system. cRNA product integrity was confirmed on agarose gels, and concentration was determined by absorbance reading at 260 nm (DU 640, Beckman Coulter, Fullerton, CA). cRNA was stored frozen in aliquots at Ϫ80°C until used.
Data Analysis-All results presented are based in a minimum of three different experiments with at least 10 oocytes per group in each experiment. Statistical significance is defined as two-tailed, with p Ͻ 0.05, and the results are presented as mean Ϯ S.E. The significance of the differences between groups was tested by one-way analysis of variance with multiple comparisons using Bonferroni's correction.

NCC Activity Is Increased by Intracellular Chloride Depletion
Maneuvers-Because the regulatory mechanisms controlling electroneutral cotransporters in their native tissues and in transfected cells seem to operate in response to intracellular chloride concentration or to correct changes in intracellular chloride concentration (11), the present study tested such mechanisms of regulation for NCC. We used two different experimental strategies to induce a depletion of [Cl Ϫ ] i . The first protocol was to compare the activity of rNCC in X. laevis oocytes injected with rNCC cRNA alone or together with of K ϩ :Cl Ϫ cotransporter KCC2 cRNA. We chose KCC2 because it is the K ϩ :Cl Ϫ cotransporter isoform that is significantly active in isotonic conditions when expressed in X. laevis oocytes (11,25), thus maintaining a continuous K ϩ :Cl Ϫ efflux over the incubation days before the uptake assays were performed. A similar approach was used in HEK-293 cells by Gillen and Forbush (32) that analyzed the regulation of NKCC1 activity by intracellular chloride depletion induced by cotransfecting the cells with KCC1. The second protocol was the "low Cl Ϫ hypotonic stress" in which rNCC-injected oocytes were incubated in a Cl Ϫ -free, slightly hypotonic medium (170 mosM/kg H 2 O) for several hours before the uptake assay. Low Cl Ϫ hypotonic stress is known to induce a decrease in [Cl Ϫ ] i in several cells (16,17), including oocytes from Rana pipens (33). In addition, low Cl Ϫ hypotonic stress in X. laevis oocytes induces the opening of Cl Ϫ channels that promote Cl Ϫ efflux (34). Fig. 1 depicts the effect of each protocol separately, or together, upon tracer 22 Na ϩ uptake in H 2 O-or rNCC-injected oocytes. As we have shown previously (1,23,30), 22 Na ϩ uptake in water-injected oocytes was very small and not thiazide sensitive, indicating that X. laevis oocytes do not express endogenous activity of a Na ϩ :Cl Ϫ cotransporter. As shown in Fig. 1, the minimum uptake observed in water-injected oocytes was not affected by KCC2 cRNA injection or by low Cl Ϫ hypotonic stress. 22 Na ϩ uptake in rNCC-injected oocytes incubated overnight in regular ND96 was 2840 Ϯ 154 pmol oocyte Ϫ1 h Ϫ1 . In contrast, 22 Na ϩ uptake in oocytes coinjected with rNCC and KCC2 cRNA and that were incubated in similar isotonic conditions was 11,757 Ϯ 514 pmol oocyte Ϫ1 h Ϫ1 ( p Ͻ 0.001). In oocytes that were injected with rNCC alone but were exposed to low Cl Ϫ hypotonic stress, 22 Na ϩ uptake was 12,449 Ϯ 974 pmol oocyte Ϫ1 h Ϫ1 ( p Ͻ 0.001). Combination of both experimental protocols resulted in synergistic effect because 22 Na ϩ uptake in rNCC ϩ KCC2 cRNA-injected oocytes incubated overnight in the Cl Ϫ -free hypotonic medium was 16,469 Ϯ 669 pmol oocyte Ϫ1 h Ϫ1 ( p Ͻ 0.001). As shown in Fig. 2, in oocytes incubated in isotonic conditions or exposed to low Cl Ϫ hypotonic stress, the effect of KCC2 cRNA coinjection upon rNCC activity was dose-dependent. Increased amount of KCC2 cRNA injected was associated with increased activity of rNCC. Oocytes exposed to low Cl Ϫ hypotonic stress exhibited higher basal activity and reached the plateau phase at lower KCC2 cRNA concentrations. Thus, promoting intracellular chloride depletion by two different strategies resulted in increased 22 Na ϩ uptake by the renal Na ϩ :Cl Ϫ cotransporter. The increased uptake could be caused by the augmented driving force by the intracellular chloride depletion. Alternatively, as has been shown to occur with NKCC1, NCC can also be regulated by chloride-sensitive mechanisms. Therefore, we analyzed whether the intracellular chloride depletion maneuvers modulate the phosphorylation status of the conserved NCC amino-terminal domain threonines that in NKCC1 have been shown to be involved in its regulation by intracellular chloride (16).
Effect of Intracellular Chloride Depletion Protocols upon rNCC Phosphorylation-Data with kinase/phosphatase pharmacological inhibitors suggest that members of the SLC12A family of cotransporters are regulated by phosphorylation/ dephosphorylation pathways. Phosphorylation induced by cell shrinkage, low intracellular Cl Ϫ , and protein phosphatase inhibitors stimulates NKCCs and inhibits KCCs, whereas dephosphorylation induced by cell swelling, high intracellular Cl Ϫ , and protein phosphatases stimulates KCCs and inhibits NKCCs (11, 13, 14, 33, 34). However, direct phosphorylation has only been demonstrated so far for NKCC1 and NKCC2 (17,22,35). By using R5 antibody, a phospho-antibody that recognizes the phosphorylation of two amino-terminal domain threonine residues in NKCC1 (human sequence, Thr 212 , Thr 217 ; shark, Thr 184 , Thr 189 ) (17), it has been shown that activation of NKCC1 by low Cl Ϫ hypotonic stress, cell shrinkage, or coexpression with the WNK3 kinase is associated with phosphorylation of these conserved threonine residues (17-19, 36, 37). R5 antibody recognizes phosphorylation of NKCC1 across species (shark versus human) or isoforms (NKCC1 and NKCC2 (17,22)) despite small amino acid divergences within the sequence used to raise the antibody (see alignment in Fig. 3A in which the 16 amino acid residues from human NKCC1 that were used to raise the R5-phosphoantibody (17) are indicated). In the close relative NCC, the two phosphoacceptor residues corresponding to Thr 53 and Thr 58 in rNCC are conserved. Because NCC retains the potential phosphoacceptor sites in this region of the amino-terminal domain, it was speculated previously that NCC could also be regulated by phosphorylation of these threonines (16,22). As shown in Fig. 3B, however, the R5-phosphoantibody is not readily able to detect the wild-type rNCC under any experimental conditions either because these threonines are not phosphorylated in rNCC or because the epitope for R5 is somewhat altered, reducing its affinity for the antibody. This latter explanation is favored by the fact that in some instances we can observe a very faint band in wild-type NCC under low chloride conditions. Thus, we analyzed the sequence divergence among NKCCs and NCC (Fig. 3A) and introduced mutations in rNCC accordingly to render it recognizable by R5 antibody. We found (Fig. 3B) that this can be accomplished by simply replacing tyrosine Tyr 56 by histidine (rNCC-Y56H in Fig. 3A), which is the major nonconservative change between NKCCs and NCC in the R5-16-amino acid residues peptide. R5 recognizes a band of 120 kDa corresponding to NCC only in the protein bearing the Y56H substitution, which is not observed in samples from wild-type NCC injected oocytes. More importantly, incubating the oocytes in Cl Ϫ -free solution overnight or coinjecting Y56H-NCC with KCC2, conditions that remarkable increased transport activity (Figs. 1 and 2), caused a parallel stimulation of R5 signal in samples from Y56H-injected oocytes. We also made a quadruple mutant bearing all substitutions corresponding to sequence differences between NCC and NKCC2 (Rat NCC-4M in Fig. 3A: L49Y-Y56H-I59M-V61A). As shown in Fig. 3B, the quadruple mutant rNCC-4 M was also detected by R5-antibody, and the signal was increased after exposing oocytes to intracellular chloride depletion maneuvers. To answer the question of whether increased signal in oocytes exposed to Cl Ϫ hypotonic stress or coinjected with KCC2 was caused by increased amount of NCC protein, extracts from oocytes were analyzed by Western blot with polyclonal anti-NCC antisera. As shown in Fig. 3C, no difference in the amount of NCC was observed in proteins extracted from NCC-injected oocytes in control conditions when compared with those either exposed to Cl Ϫ hypotonic stress or coinjected with KCC2 or both. In addition, to answer the question of whether what we observed with R5 antibody is actual phosphorylation of NCC, we treated the oocyte lysates with alkaline phosphatase. As shown in Fig. 3D, the R5 signal disappeared or was greatly reduced after this treatment. As shown in Fig. 4, functional expression analysis revealed that single substitution of the tyrosine 56 for histidine (rNCC-Y56H) or the quadruple substitution (rNCC-4M) did not affect either the basal activity of rNCC or its activation by the coinjection with KCC2 or low Cl Ϫ hypotonic stress.
Our results in the present study show that activity of rNCC in basal conditions is associated with phosphorylation of the amino-terminal domain threonine residues Thr 53 and Thr 58 . The phosphorylation level was significantly increased by intracellular chloride depletion strategies that simultaneously increased the cotransporter activity. These observations suggest that increased uptake by NCC is not only a consequence of the augmented driving force by the intracellular chloride depletion. Supporting this conclusion, we have observed that protein phosphatase inhibition is coupled with increased activity of rNCC. Incubation of rNCC-injected X. laevis oocytes with the protein phosphatase 1 and 2A inhibitor calyculin A (100 nM) resulted in significant increase in rNCC-mediated 22 Na ϩ uptake from a value of 3961 Ϯ 209 pmol oocyte Ϫ1 h Ϫ1 in the rNCC control group in the absence of calyculin A to a value of 5888 Ϯ 310 pmol oocyte Ϫ1 h Ϫ1 in its presence (n ϭ 45; p Ͻ 0.001). Taken together, our results demonstrate that, as for NKCCs, the activity of NCC correlates with phosphorylation of NH 2 -terminus threonine residues at position 53 and 58.
Effect of Intracellular Chloride Depletion Protocols upon rNCC Surface Expression-Increased activity of rNCC by coinjection of KCC2 or low Cl Ϫ hypotonic stress could be caused by the activation of cotransporter units that are already in the plasma membrane or by an increase in the amount of transporter units that reach the plasma membrane. To analyze these possibilities, we assessed in X. laevis oocytes the surface expression of the EGFP-rNCC construct that we have validated previously (29). We have shown that EGFP-rNCC fluorescence in the oocytes surface co-localizes with the F-404 specific plasma membrane dye and that oocytes injected with EGFP-rNCC exhibit significant thiazide sensitive 22 Na ϩ uptake, indicating the EGFP-NCC fluorescence is located in the plasma mem- The arrows show the putative phosphorylation sites. B, a representative immunoblot analysis of proteins extracted from X. laevis oocytes injected with water, wild-type rNCC cRNA, the mutant rNCC-Y56H cRNA, or the mutant rNCC-4M cRNA at 0.2 g/l using the R5-phosphoantibody raised against the 16-residue phosphorylated peptide of NKCC1 shown in A. The band corresponding to phosphorylated NCC is shown as phospho-NCC. C, a representative Western blot of proteins extracted from rNCC-injected oocytes using polyclonal anti-NCC antibody (27). D, a representative immunoblot analysis of proteins extracted from X. laevis oocytes injected with water, wild-type rNCC cRNA, or the mutant rNCC-4 M cRNA at 0.2 g/l using the R5-phosphoantibody. The immunoblot was performed in control conditions or after exposing proteins to alkaline phosphatase as described under "Experimental Procedures" and is shown in the upper panel. The corresponding Coomassie Blue image is shown in the lower panel. For immunoblots in B, C, and D, proteins were extracted from oocytes in control conditions and either exposed to low chloride hypotonic stress or coinjected with KCC2 cRNA or subjected to both maneuvers together. Similar results were observed in five different experiments.
brane. Thus, we have successfully used the EGFP-rNCC construct to assess the effect of elimination of N-glycosylation sites (29), Gitelman's type mutations (28), single nucleotide polymorphisms (30), and the kinases WNK3 (19) and WNK4 (7) upon both the surface and functional expression of renal Na ϩ : Cl Ϫ cotransporter. Because the fluorescence intensity is assessed in the confocal microscope using live oocytes, the same oocytes were then used to assess functional expression. Oocytes injected with EGFP-rNCC cRNA were exposed to both strategies as mentioned above, and the surface expression was assessed by monitoring the EGFP fluorescence with a confocal microscope. As shown in Fig. 5, A and B, surface expression analysis revealed similar fluorescence intensity at the surface of oocytes in all groups. No difference was observed in the amount of EGFP-NCC proteins by Western blot analysis (data not shown). However, as shown in Fig. 5C, thiazide sensitive 22 Na ϩ uptake in the EGFP-rNCC-injected oocytes was increased by the intracellular Cl Ϫ depletion strategies similar to that observed for wild-type rNCC (Fig. 1). The uptake observed in EGFP-rNCC cRNA-injected oocytes of 987 Ϯ 105 pmol oocyte Ϫ1 h Ϫ1 increased by coinjection with KCC2 cRNA (3518 Ϯ 687 pmol oocyte Ϫ1 h Ϫ1 , p Ͻ 0.01), by low Cl Ϫ hypotonic stress (3525 Ϯ 301 pmol oocyte Ϫ1 h Ϫ1 , p Ͻ 0.05), or by both maneuvers together (5096 Ϯ 401 pmol oocyte Ϫ1 h Ϫ1 , p Ͻ 0.01). These observations suggest that increased activity of rNCC induced by coinjection with KCC2 or low Cl Ϫ hypotonic stress is caused by stimulation of the cotransporter intrinsic activity rather than by an increase in rNCC exocytosis containing vesicles.
Effect of replacing Thr 53 , Thr 58 , and Ser 71 of rNCC with Alanine-To define the role of each of the three putative phosphorylation sites of the amino-terminal domain in NCC basal activity and the response to intracellular chloride depletion protocols, we substituted these residues with alanine by site directed mutagenesis to create the single mutants T53A, T58A, and S71A, the double mutants T53-58A, T58-S71A, and T53-S71A, and the triple mutant T53-T58-S71A. Fig. 6 shows the average results from several experiments in which X. laevis oocytes were injected with similar amounts of wild-type rNCC cRNA or each of the cRNA mutants. 22 Na ϩ uptake in water-injected oocytes was 249 Ϯ 18 pmol oocyte Ϫ1 h Ϫ1 , whereas that in rNCC-cRNA-injected oocytes was 3432 Ϯ 156 pmol oocyte Ϫ1 h Ϫ1 . Elimination of any of the three potential phosphorylation sites resulted in a significant reduction of rNCC functional expression to various degrees. 22 Na ϩ uptake in oocytes injected with the single mutants T53A, T58A, or S71A cRNA was 2486 Ϯ 155, 452 Ϯ 39, and 851 Ϯ 89 pmol oocyte Ϫ1 h Ϫ1 , respectively. Thus, the percentage of reduction was 27% for T53A, 75% for S71A, and 100% for T58A. The effect of T53A and S71A was synergistic as shown by the fact that any combination to produce double mutants resulted in complete abolishment of basal rNCC activity. Consequently, the triple mutant also exhibited no activity. These observations suggest that Thr 53 and Ser 71 are required to achieve the full basal activity of the cotransporter, whereas Thr 58 appears to be absolutely necessary for rNCC to express basal activity. Interestingly, our observations with the rNCC single mutants are similar to those of shark NKCC1 but different from observations with rabbit NKCC2. The NKCC1 activity was completely prevented by elimination of the threonine residue Thr 189 and partially reduced by elimination of Thr 184 and Thr 202 (16). In contrast, in rabbit NKCC2 elimination of each of these threonines only reduced the activity by ϳ20% (31). Therefore, the amino-terminal domain phosphorylation sites are required to achieve basal activity in NKCC1 and NCC but not in NKCC2. Fig. 7 shows the effect of single, double, or triple mutations upon the response of rNCC to intracellular chloride depletion after coinjection with KCC2 plus the exposure to low Cl Ϫ hypotonic stress. In these experiments, the basal activity of T53A and S71A was reduced and in any other mutant was completely prevented. However, the single and double mutants still exhibited increased thiazide sensitive Na ϩ uptake after intracellular chloride depletion. In contrast, the triple mutant in which Thr 53 , Thr 58 , and Ser 71 were substituted with alanine was not activated even by both maneuvers together. These observations suggest that elimination of the three phosphorylation sites resulted in an rNCC that was neither functional in basal conditions nor activated by the intracellular Cl Ϫ depletion strategies. In contrast, as long as one of the three phosphorylation sites is present, rNCC could be non-functional in basal conditions but still be activated by intracellular chloride depletion.
Effect of Replacing Thr 53 , Thr 58 , and Ser 71 upon EGFP-rNCC Surface and Functional Expression-Because elimination of the three phosphorylation sites in the amino-terminal domain alone or in combination resulted in a significant reduction of the cotransporter basal activity, we wanted to know the effect of these mutations upon surface expression of the cotransporter. Thus we introduced the triple mutation T53-T58-S71A into the EGFP-rNCC cDNA. X. laevis oocytes were injected with 25 ng of each clone cRNA, and 4 days later the surface fluorescence intensity was assessed under confocal microscopy. Then, the same oocytes were used for functional analysis by assessing the thiazide sensitive 22 Na ϩ uptake. As shown in Fig. 8, the surface expression of the triple mutant EGFP-rNCC was similar to wild-type EGFP-rNCC, whereas the functional expression was completely abolished. As a positive control, in the same experiment a group of oocytes was injected with 25 ng of cRNA transcribed from the EGFP-rNCC double mutant N404,424Q, in which both N-glycosylation sites of EGFP-rNCC were eliminated. As we have reported previously (29), elimination of the N-glycosylation sites resulted in a cotransporter in which the surface expression, and thus its activity, is seriously reduced. The observation that elimination of phosphorylation sites resulted in reduction of the cotransporter activity without affecting the surface expression is consistent with the results in Fig. 5 in which activation of rNCC by intracellular chloride depletion was not associated with surface expression changes. Taken together, the results of the present study show for the first time that the renal Na ϩ :Cl Ϫ cotransporter is regulated by intracellular chloride concentration through phosphorylation of the amino-terminal domain and that the mechanism seems to be caused by an increased turnover rate of the cotransporter rather than by an increase in its expression in the plasma membrane.
In the present study we used the heterologous expression system in X. laevis oocytes to assess the regulation of NCC by intracellular chloride depletion. This expression system has FIGURE 5. Effect of intracellular chloride depletion protocols upon EGFP-rNCC surface expression. Oocytes were injected with water or 25 ng of wildtype EGFP-rNCC cRNA with or without 10 ng/oocyte of KCC2 cRNA. Four days later oocytes were incubated overnight in either isotonic ND96 or Cl Ϫ -free hypotonic medium. The next day, the surface fluorescence of the oocytes was visualized through a laser scanning confocal microscope, and activity was assessed by tracer 22 Na ϩ uptake assay as described under "Experimental Procedures." A shows a representative confocal image of each group, as stated. B, fluorescence intensity is expressed as mean Ϯ S.E. of 30 oocytes from three different experiments exposed to isotonicity (open bars) or hypotonicity (closed bars) with or without coinjection with KCC2 cRNA, as stated. C, 22 22 Na ϩ uptake in control oocytes, e.g. oocytes injected with rNCC cRNA or mutant rNCC cRNA and incubated the night before the uptake assay in isotonic ND96. Black bars represent rNCC or mutant NCC cRNA-injected oocytes exposed to low Cl Ϫ hypotonic stress and coinjected with KCC2 cRNA. *, significantly different from the uptake observed in the corresponding control. Each bar represents at least the mean Ϯ S.E. of 20 oocytes from three different frogs.
shown to be an excellent tool for a robust and reproducible expression of NCC in our hands (1,7,19,23,24,29,35) and in other laboratories (8, 36 -38). In contrast, NCC expression in transfected mammalian cells has not been successful in many laboratories, including our own. In this regard, however, the prediction conducted using X. laevis oocytes by two different groups (7,8) indicating that WNK4 down-regulates the surface expression (and thus activity) of NCC has recently been confirmed by assessing surface expression of NCC in renal epithelial cells transfected with NCC and WNK4 cDNA (39).
Because of the gradient of accompanying cation, the Na ϩcouple chloride cotransporters NKCCs mediate Cl Ϫ influx, whereas K ϩ -coupled chloride cotransporters KCCs mediate Cl Ϫ efflux. Because intracellular concentration of Na ϩ and K ϩ are quickly restored by Na ϩ :K ϩ :ATPase, what the activity of electroneutral cation chloride cotransporters seems to modulate is the [Cl Ϫ ] i . For instance, in most cells, [Cl Ϫ ] i , and thus cell volume, is maintained by coordinated activity of NKCC1 and KCC1. It has been demonstrated in duck red blood cells that activity of these cotransporters is regulated by [Cl Ϫ ] i in such a way that low [Cl Ϫ ] i activates NKCC1 and inhibits KCC1, whereas high [Cl Ϫ ] i activates KCC1 and inactivates NKCC1 (40,41). In neurons, intracellular chloride concentration is defined by the ratio of NKCC1/KCC2, and these in turn define the type of response to neurotransmitters such as ␥-aminobutyric acid that acts by opening Cl Ϫ channels in postsynaptic membranes. The excitatory effect of ␥-aminobutyric acid in the prenatal period is caused by an increased NKCC1/KCC2 ratio, which results in increased [Cl Ϫ ] i , whereas the inhibitory effect of ␥-aminobutyric acid in the postnatal period can be explained by decreased NKCC1/KCC2 ratio that results in decreased [Cl Ϫ ] i (42)(43)(44)(45). It has been clearly shown that modulation of NKCC1 by [Cl Ϫ ] i is mediated, at least in part, by phosphorylation of the amino-terminal domain threonine residues shown in Fig. 3 (17, 46). In addition, recent studies have shown that electroneutral cation chloride cotransporters are regulated also through phosphorylation of the amino-terminal domain by members of two serine/ threonine kinase families known as WNK and STE20 (18 -20, 47, 48). First it was shown that STE-20-related kinases such as SPAK or OSR1 mediate functional activation of NKCC1 (47, 49) by a process in which phosphorylation of the amino-terminal domain threonine residues is implicated. Later, Vitari et al. (20) observed that WNK1 and WNK4 seem to activate SPAK and OSR1, which in turns phosphorylated the amino-terminal domain of NKCC1. At the functional level, Gagnon et al. (48) also showed that WNK4 is able to increase the activity of NKCC1 only in the presence of SPAK or OSR1. Distorted modulation of NCC activity by WNK1 and WNK4 is implicated in the pathophysiology of arterial hypertension in patients with PHAII (7,8). WNK4 inhibits the activity of NCC, and the PHAII-type mutations prevent this inhibition, thus releasing NCC, which by its increasing activity produces arterial hypertension. This hypothesis can explain the exquisite sensitivity to thiazide-type diuretics in PHAII patients (9). WNK1 seems to produce PHAII by its ability to regulate WNK4 activity (8,50,51). We have recently shown that WNK3, another member of the WNK family, is a powerful activator of NCC as well as NKCC1 and NKCC2 (18,19). The activation of NKCC1 and NKCC2 induced by WNK3 is also associated with increased phosphorylation of the amino-terminal domain threonine residues shown in Fig. 3 (18,19). In the present study we show that NCC activity is modulated by [Cl Ϫ ] i and that phosphorylation of the amino-terminal domain threonine residues Thr 53 and Thr 58 , and probably Ser 71 , is implicated, suggesting that modulation of NKCCs by [Cl Ϫ ] i and WNKs could be mediated by a common pathway. Supporting this possibility, it has been suggested that activity of WNK kinases is modulated by intracellular chloride. Xu et al. (52) identified WNK1 at the molecular level and observed that its activity was remarkably increased in the presence of NaCl. Then, Lenertz et al. (51) showed that WNK1 is activated by NaCl and other osmotic challenges in a variety of cell lines, including distal convoluted kidney tubule cells, although it is not affected by short or long exposure to several hormones and by agents that modulate cell proliferation, suggesting that the major regulator of WNK kinases is cell ion strength, volume, and/or intracellular ion concentrations. In this regard, Moriguchi et al. (21) have recently shown in HEK-293 cells that both WNK1 and SPAK/ OSR1 kinases activity is increased when cells were exposed to low Cl Ϫ hypotonic stress, suggesting that WNK1 behaves as an activator of SPAK/OSR1 in response to a decrease in the [Cl Ϫ ] i . Oocytes were injected with the wild-type or mutant EGFP-rNCC cRNA as stated. Functional expression assays were performed with the same oocytes that were used for fluorescence analysis in B. The EGFP-N404,424Q rNCC mutant, in which the two N-glycosylation sites have been eliminated, was used as control and was described previously (29). *, significantly different from wild-type EGFP-rNCC.
Supporting the observations in the present study, Moriguchi et al. (21) also observed in HEK-293 cells transfected with T7-tagged mouse NCC that the amino-terminal domain of NCC can be phosphorylated in vitro by the WNK1/SPAK/ OSR1 pathway and that the phosphorylation level increases when cells are incubated under low Cl Ϫ hypotonic stress and decreases when the residues Thr 53 , Thr 58 , and Ser 71 are eliminated. Thus, taking all these observations together it is reasonable to speculate that [Cl Ϫ ] i probably regulates WNKs and STE20 kinase activity that in turn modulates electroneutral cotransporters function.
In summary, we present evidence that rNCC can be activated by intracellular chloride depletion strategies such as co-expression with the K ϩ :Cl Ϫ cotransporter KCC2 and/or exposure to low Cl Ϫ hypotonic stress. The rNCC activation is associated with increased phosphorylation of the threonine residues 53 and 58 at the amino-terminal domain without changing the surface expression of the cotransporter. In addition, substituting Thr 53 , Thr 58 , and Ser 71 residues with alanine renders rNCC inactive in basal conditions and insensitive to intracellular chloride depletion, indicating that in addition to Thr 53 , Thr 58 , the serine at the position 71 is also critical for rNCC regulation.