14-3-3 Proteins Modulate the Expression of Epithelial Na+ Channels by Phosphorylation-dependent Interaction with Nedd4-2 Ubiquitin Ligase*

The ubiquitin E3 protein ligase Nedd4-2 is a physiological regulator of the epithelial sodium channel ENaC, which is essential for transepithelial Na+ transport and is linked to Liddle's syndrome, an autosomal dominant disorder of human salt-sensitive hypertension. Nedd4-2 function is negatively regulated by phosphorylation via a serum- and glucocorticoid-inducible protein kinase (Sgk1), which serves as a mechanism to inhibit the ubiquitination-dependent degradation of ENaC. We report here that 14-3-3 proteins participate in this regulatory process through a direct interaction with a phosphorylated form of human Nedd4-2 (a human gene product of KIAA0439, termed hNedd4-2). The interaction is dependent on Sgk1-catalyzed phosphorylation of hNedd4-2 at Ser-468. We found that this interaction preserved the activity of the Sgk1-stimulated ENaC-dependent Na+ current while disrupting the interaction decreased ENaC density on the Xenopus laevis oocytes surface possibly by enhancing Nedd4-2-mediated ubiquitination that leads to ENaC degradation. Our findings suggest that 14-3-3 proteins modulate the cell surface density of ENaC cooperatively with Sgk1 kinase by maintaining hNedd4-2 in an inactive phosphorylated state.

The epithelial Na ϩ channel (ENaC) 1 is a heteromultimeric protein comprising three subunits, ␣, ␤, and ␥, each of which is composed of two transmembrane domains, an extracellular loop, and short N and C termini (1). This channel is distributed widely in salt-reabsorbing epithelia and represents the first and rate-limiting step for transepithelial Na ϩ transport (2,3). Previous studies have shown that mutations in ENaC subunits are associated with several heritable human disorders exhibiting abnormal regulation of blood pressure and electrolyte balance. For example, dominant gain-of-function mutations in ENaC cause Liddle's syndrome, an inherited form of hypertension (4). Conversely, loss-of-function mutations cause pseudohypoaldosteronism type I, a salt-wasting hypotensive condition (5). Thus, proper regulation of ENaC activity is critical for maintenance of whole body Na ϩ homeostasis and for blood pressure control.
The activity of ENaC is regulated by a variety of mechanisms including ubiquitination and phosphorylation. At least two regulatory proteins participate directly in this regulation: the neural precursor cell-expressed, developmentally down-regulated gene 4 isoform (Nedd4-2) and serum, glucocorticoid-inducible kinase 1 (Sgk1) (6). Nedd4-2 is a ubiquitin E3 protein ligase composed of four WW domains and a ubiquitin ligase Hect domain (7,8). Nedd4-2 interacts directly with ENaC by association of its WW domains with the PY motifs present in each ENaC subunit to ubiquitinate the ENaC channel (9 -12). Ubiquitination accelerates the rate of ENaC degradation and reduces the channel copy number at the cell surface (13), ultimately resulting in a drastic decrease in Na ϩ current in Xenopus laevis oocytes (10 -12). The importance of this regulation is also suggested by findings that Liddle's syndrome is linked to mutations in ENaC which invariably cause either deletion or alteration of the ␤ or ␥PY motif (14,15). Sgk1 is an early aldosterone-induced protein kinase that relays signals from the pathways that are transmitted via phosphatidylinositol 3-kinase and cAMP-dependent protein kinase (protein kinase A). Sgk1 induces cell surface expression of ENaC and increases Na ϩ current when both Sgk1 and ENaC are coexpressed in Xenopus oocytes (16,17). Recently, it was reported that Nedd4-2 is a substrate of Sgk1 and, importantly, that the Sgk1-dependent phosphorylation of Nedd4-2 reduces its affinity for ENaC, thereby reducing the inhibitory effect of Nedd4-2 on ENaC (18,19). Thus, it appears that Sgk1 and Nedd4-2 are in a common cell signaling pathway that regulates the cell surface density of ENaC by a ubiquitination-dependent mechanism, although Sgk1 may also regulate ENaC activity by phosphorylating its ␣ subunit directly (20). However, the details of the mechanism remain unknown (6). The 14-3-3 proteins recognize the phosphopeptide consensus sequence motifs RXXpS/TXP and RXXXpS/TXP (where pS/T is phospho-Ser/Thr and X is any amino acid; see Refs. 21 and 22) and bind a variety of enzymes and cell signaling molecules to modulate their activity, conformation, stability, intracellular localization, and function (23)(24)(25)(26). Here we report that a 14-3-3 protein binds Nedd4-2 (a human KIAA0439 gene product; termed hNedd4-2), when it is phosphorylated at Ser-468 by Sgk1 kinase. The interaction inhibits the dephosphorylation of hNedd4-2, reduces the rate of Nedd4-2-catalyzed ubiquitination of ENaC subunits, and maintains the Sgk1-stimulated increase in ENaC-mediated current. Based on these findings, we propose that a 14-3-3 protein is a cofactor for the regulation of ENaC function via an Sgk-1/Nedd4-2-dependent mechanism.
PC12 Cell Culture-PC12 cells (a gift from Dr. L. A. Greene) were grown on collagen-coated tissue culture plates in RPMI 1640 medium supplemented with 10% heat-inactivated donor horse and 5% fetal bovine serum, as described previously (31). Retroviral infection was carried out as described previously (29) by adding 0.5-1 ml of a viruscontaining supernatant, recovered from packaging cells (Bosc23), to 50% confluent PC12 cell cultures. Infected cells were cultured for 1 week and then subjected to selection with G418 for at least 1 month. Surviving colonies were isolated and clonally expanded for characterization. The PC12-Mh clone was chosen for biochemical assessment on the basis of protein expression (10 -20% endogenous 14-3-3 proteins; Supplemental Fig. S2) and similarity with parental cells with regard to cellular morphology (data not shown).
MEF Purification-PC12-Mh cells in six 10-cm dishes (ϳ5 ϫ 10 7 cells) were lysed in 3 ml of lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10% (w/v) glycerol, 100 mM NaF, 10 mM EGTA, 1 mM Na 3 VO 4 , 1% (w/v) Triton X-100, 5 M ZnCl 2 , 2 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, and 1 g/ml leupeptin). The lysate was centrifuged at 100,000 ϫ g for 20 min at 4°C. The supernatant was passed through a 5-m filter, incubated with 150 l of Sepharose beads for 60 min at 4°C, and then passed through a 0.65-m filter. The filtrated supernatant was mixed with 150 l of anti-myc-conjugated Sepharose beads for the first immunoprecipitation. After incubation for 90 min at 4°C, the beads were washed five times with 1.5 ml of TNTG buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10% (w/v) glycerol, 0.1% (w/v) Triton X-100), twice with buffer A (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 0.1% (w/v) Triton X-100), and finally once with TNT buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% (w/v) Triton X-100). The washed beads were incubated with 15 units of TEV protease (Invitrogen) in 150 l of TNT buffer to release bound materials from the beads. After incubation for 60 min at room temperature, the supernatant was pooled, and the beads were washed twice with 75 l of buffer A. The resulting supernatants were combined together and incubated with 25 l of FLAG-Sepharose beads for the second immunoprecipitation. After incubation for 60 min at room temperature, the beads were washed three times with 500 l of buffer A, and proteins bound to the immobilized 14-3-3 on the FLAG beads were dissociated by incubation with 1 mM synthetic phosphopeptides (LSQRQRSTpSTPNVHA, based on amino acids 250 -265 of cRaf-1; Ref. 20) in buffer A for 120 min at 4°C. Approximately 3 g of protein (0.01% of stating materials) was routinely recovered by this procedure.
Tandem Mass Spectrometry-Proteins were separated by 7.5% SDS-PAGE and visualized by silver staining. The stained bands were excised and digested in the gel with Lys-C, and the resulting peptide mixtures were analyzed by the direct nanoflow LC-MS/MS system, equipped with an electrospray interface reversed phase column, a nanoflow gradient device, a high resolution Q-TOF hybrid MS (Q-TOF2, Micromass), and an automated data analysis system (32). All MS/MS spectra were searched against the nonredundant protein sequence data base maintained at the National Center for Biotechnology Information using the Mascot program (Matrixscience) to identify proteins. The MS/MS signal assignments were also confirmed manually.
In Vitro Phosphorylation and Binding Assay-Phosphorylation of GST-hNedd4-2 and the hNedd-S468A mutant (1 g each) was performed at 30°C for 30 min in a reaction containing 20 mM MOPS, pH 7.2, 15 mM MgCl 2 , 20 mM ␤-glycerol phosphate, 1 mM EGTA, 1 mM Na 3 VO 4 , 1 mM dithiothreitol, 0.1 mM ATP, 25 ng of Sgk1 (⌬1-59, S422D; Upstate Biotechnology), and 2 g of 14-3-3 in a final volume of 200 l. When necessary, 5 Ci of [␥-32 P]ATP was added to the mixture. For binding experiments, glutathione-agarose (ϳ50 l) was added to the reaction and incubated for 60 min at room temperature. The protein complexes bound to the beads were washed five times with 20 mM Tris-HCl, 150 mM NaCl, 0.1 mM dithiothreitol, pH 7.5, and then analyzed by SDS-PAGE followed by Coomassie Brilliant Blue staining or Western blotting with antibodies to 14-3-3.
Metabolic 32 P Labeling and Immunoprecipitation-HEK293 cells transiently expressing ME-hNedd4-2 were incubated in phosphate-free Dulbecco's modified Eagle's medium for 20 min and then labeled in the same medium containing 40 Ci/ml 32 P i for 2 h. After incubation, cells were lysed in lysis buffer, and the expressed ME-hNedd4-2 was immunoprecipitated with anti-myc-Sepharose. The immunocomplexes were washed five times in buffer A and once in TNT buffer, and the ME-hNedd4-2 was released from the beads by digestion with TEV protease. For the experiment shown in Fig. 2B (lanes 1-4), the cells were also placed in fresh serum-free medium for 18 h prior to initiation of the 32 P labeling. Dephosphorylation Assay-GST-hNedd4-2 or the GST-hNedd-S468A mutant (1 g each) was phosphorylated by Sgk1 in the presence of [␥-32 P]ATP and then pulled down with ϳ50 l of glutathione-agarose. After washing, the beads were suspended in 200 l of 50 mM Tris-HCl, 20 mM MgCl 2 , 1 mM dithiothreitol, pH 7.5, and incubated with 2 g of 14-3-3 for 10 min at room temperature. The dephosphorylation reaction was performed at room temperature for 30 min by the addition of 0.5 unit of PP1 catalytic subunit (Sigma).
Electrophysiological Measurements and Immunofluorescence Studies-Electrophysiological studies in Xenopus oocytes were performed as described previously (33,34). In brief, cRNAs (three human ENaC subunits, 0.1 ng each; Sgk1, 0.2 ng; wild type or V180D mutant 14-3-3, 0.5 ng) were injected into Xenopus oocytes. Electrophysiological recordings were taken 24 h after injection using a two-electrode voltage clamp technique. The recording solution had an ionic composition of 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , and 5 mM HEPES, adjusted to pH 7.5 with NaOH. All electrophysiological recordings were performed at room temperature (25 Ϯ 1°C) at a holding potential of Ϫ60 mV. The expression of ENaC was examined by immunohistochemistry using rat ␣ ENaC-FLAG cRNA as reported previously (18). In brief, FLAG-tagged ␣ENaC and nontagged ␤ and ␥ENaC were expressed in oocytes together with Sgk1 and either 14-3-3 or its V180D mutant. Cryosections (10 m) of Xenopus oocytes were fixed with 4% paraformaldehyde in 0.1 M phosphate buffer at room temperature for 30 min and then washed with 10 mM phosphate-buffered saline. These sections were pretreated with 5% normal goat serum (Dako) in phosphate-buffered saline at room temperature for 30 min. ENaC-FLAG was detected with 3.3 g/ml anti-FLAG M2 (Kodak) for 12 h at 4°C. Sections were rinsed with phosphate-buffered saline and subsequently incubated with 2.0 g/ml Alexa Fluor 594 goat anti-mouse IgG (Molecular Probes) for 1.5 h at room temperature. These sections were rinsed with phosphate-buffered saline, mounted in Aquatex (Merck), and observed using an Olympus AX70 fluorescent microscope.
In Vivo Ubiquitination Assay-The ubiquitination assay was performed according to the procedure described by Staub et al. (13) with slight modifications. HEK293 cells were transiently coexpressed with FLAG-tagged ␣ENaC or ␥ENaC (35) together with other protein components as indicated in Fig. 4. After 48 h, cells were lysed in 1% SDS and boiled for 5 min. Samples were subsequently diluted with 9 volumes of lysis buffer and then incubated with anti-FLAG-Sepharose beads for 90 min at room temperature. The beads were washed five times with lysis buffer containing 0.1% SDS, and bound proteins were analyzed by Western blotting with anti-HA (Upstate) or anti-FLAG (Kodak).

Identification of hNedd4-2-KIAA0439 as a 14-3-3-interacting
Partner-We developed a novel tandem affinity purification approach coupled with MS-based proteomics technology to identify cellular proteins associated with the 14-3-3 isoform in PC12 cells. The purification strategy employed an original Nterminal affinity tag (termed MEF) composed of two independent epitope tags, myc and FLAG, that were fused in tandem and separated by a spacer sequence containing a TEV protease cleavage site (myc-TEV-FLAG; Supplemental Fig. S1). The MEF cassette was fused to the N terminus of the bovine 14-3-3 isoform, and PC12 cells were transformed with the cDNA to generate a monoclonal cell line (termed PC12-Mh) that stably expressed the MEF-fused version of the 14-3-3 isoform (see "Experimental Procedures" and Supplemental Fig. S2).
After PC12-Mh cells were lysed, the expressed MEF-14-3-3 was recovered with its binding proteins (see "Experimental Procedures" and Supplemental Fig. S3). Ultimately, proteins bound to MEF-immobilized on the FLAG beads were dissociated with a synthetic phosphopeptide (LSQRQRSTpSTPN-VHA) designed to mimic the 14-3-3-binding site of Raf kinase (21). Fig. 1A (lane 2) shows the SDS-PAGE profiles of the dissociated proteins. About 40 proteins with molecular masses between 40 and 300 kDa were reproducibly detected by this procedure, whereas no proteins were detected using a lysate obtained from control cells transformed with vector alone (lane 1). Two controls were used to assess the specificity of this interaction. First, we used a nonphosphorylated peptide (LSQRQRSTST-PNVHA) instead of its phosphorylated form to dissociate the complex (lane 3). In the second control, we performed the procedure using PC12 cells expressing a V180D point mutant of 14-3-3 that replaced Val-180 with Asp, a mutation known to disrupt the interaction of 14-3-3 with a variety of targets including ExoS (36), ASK1 (37) and KLC2 (27) (lane 4). As shown in Fig. 1A (lanes 3 and 4), no protein bands were detected in these controls. The V180D mutant was expressed in PC12 cells at a level similar to the wild type protein (Supplemental Fig. S2 and Fig. 1B, arrow) and appeared to have a structural fold similar to the wild type protein as judged by its ability to form heterodimers with the endogenous wild type 14-3-3 subunits (␤, ␥, ⑀, , , and isoforms, as analyzed by SDS-PAGE and LC-MS/MS; see Fig. 1B and data not shown). Thus, we assumed that the proteomic procedure using a MEF tag would yield specific targets of 14-3-3 proteins in PC12 cells. In fact, we found that many known targets of 14-3-3 proteins, including IRS1 and 2, phosphatidylinositol 3-kinase class 1, Wee-1, KSR-1, A-raf, B-raf, C-raf, Cdc25C, KHC, KLC2, PFK-2 and tyrosine hydroxylase, were simultaneously isolated by this procedure as revealed by Western blot analyses (data not shown).
To identify the proteins recovered by the procedure described above, silver-stained bands were excised from the gel, digested in the gel by Lys-C, and analyzed by a direct nanoflow LC-MS/MS system (32). This procedure identified more than 100 individual proteins, of which ϳ40 were known and ϳ60 were unknown targets of 14-3-3 (data not shown), including hNedd4-2 reported here. The identification of hNedd4-2 was based on six independent MS/MS spectra derived from six doubly charged peptide ions, each corresponding to the unique amino acid sequence within KIAA0439 (Supplemental Table  S1). Western blotting using Nedd4 antibodies confirmed that the protein preparation contained hNedd4-2 (Fig. 1C, arrow), whereas another member of the Nedd4 family, Nedd4 -1 (11), was not detected in this study (data not shown). We confirmed that the 14-3-3/hNedd4-2 complex was endogenous to PC12 cells because the immunoprecipitate pulled down from the cell extract with anti-Nedd4 contained endogenous 14-3-3 (Fig.  1D), and a number of 14-3-3 isoforms other than , such as ␤, ␥, ⑀, and , also formed a complex with hNedd4-2 (Fig. 1E).
14-3-3 Binds hNedd4-2 Phosphorylated by Sgk1 Kinase-Previous studies showed that Nedd4-2 is a substrate of Sgk1 kinase (18,19). Because 14-3-3 binds phosphorylated targets in most known cases, we examined whether the interaction be- . The position of hNedd4-2 is indicated by an arrow. B, endogenous 14-3-3 isoforms that coprecipitated with the expressed MEFwere dissociated from the FLAG beads with 80 g/ml FLAG peptides and analyzed by SDS-PAGE and silver staining (10% gel). The arrow indicates the position of MEF-and MEF-V180D, and arrowheads indicate the 14-3-3 isoforms. C, the proteins detected in A were analyzed by Western blotting with polyclonal anti-Nedd4. The arrow indicates the position of hNedd4-2. D, endogenous Nedd4-2 was immunoprecipitated from PC12 cells with control or Nedd4 antibodies (IP), and the immnocomplexes were analyzed by immunoblot (IB) with antibodies specific to Nedd4 and 14-3-3. As a reference, Nedd4 and 14-3-3 immunoblotting of PC12 cell extract is included. E, control and Nedd4 immunocomplexes were analyzed as in D using antibodies to 14-3-3␤, ␥, ⑀, or isoforms. Arrows indicate the positions of the specific protein recognized by each antibody. tween 14-3-3 and hNedd4-2 depends on the Sgk1-mediated phosphorylation of hNedd4-2. A GST-hNedd4-2 fusion protein was expressed in Escherichia coli, and the purified protein was assayed for 14-3-3 binding before and after phosphorylation by Sgk1 kinase. As shown in Fig. 2A (lanes 4 and 6, bottom panel), the phosphorylated form of GST-hNedd4-2 bound directly to 14-3-3, whereas the nonphosphorylated protein did not.
To examine further whether the hNedd4-2/14-3-3 interaction is physiologically relevant, we prepared HEK293 cells transfected with a construct expressing ME-tagged hNedd4-2. The transformed cells were labeled with 32 P i and stimulated by insulin or insulin-like growth factor I, both of which provide an upstream signal to activate Sgk1 kinase (38,39). The ME-hNedd4-2 protein was then recovered from the cells, separated by SDS-PAGE, and analyzed by autoradiography to measure   FIG. 2. Interaction of hNedd4-2 and 14-3-3 depends on Sgk1-catalyzed phosphorylation. A, purified GST-hNedd4-2, GST-hNedd-S468A mutant, and 14-3-3 were analyzed by SDS-PAGE and Coomassie Brilliant Blue (lanes 1-3). The purified hNedd4-2 proteins were incubated with [␥-32 P]ATP and 14-3-3 in the absence or presence of Sgk1, pulled down with glutathione-Sepharose, and the precipitates were analyzed by 10% SDS-PAGE and Coomassie Brilliant Blue staining (lanes 4 -7, top), antoradiography (middle), or immunoblot (IB) with anti-14-3-3 (bottom). The arrow indicates the position of 14-3-3. N4 indicates GST-hNedd4-2 proteins. B, HEK293 cells transiently expressing ME-hNedd4-2 were metabolically labeled with 32 P i and stimulated with insulin or insulin-like growth factor I (IGF1) as indicated. After immunoprecipitation with anti-myc-Sepharose, the ME-hNedd4-2 was released from the beads by digestion with TEV protease and then analyzed by 10% SDS-PAGE and Coomassie Brilliant Blue staining (top), antoradiography (middle), or immunoblot with antibodies to PAN-14-3-3 proteins (bottom) (lanes 1-4). Similar experiments were performed with HEK293 cells expressing ME-hNedd4-2 plus Sgk1 ( lanes  5-7). Arrows indicate the positions of endogenous 14-3-3s. C, HEK293 cells transiently expressing ME-hNedd4-2 were lysed, and the ME-hNedd4-2 was recovered by immunoprecipitation after digestion with TEV protease as in B. The recovered protein was subjected to SDS-PAGE, digested in the gel with trypsin, and then analyzed by direct nanoflow LC-MS/MS to identify phosphoserine 468. The y-ion fragment series generated by collision-induced dissociation is indicated. pS, phosphoserine 468. the extent of ME-hNedd4-2 phosphorylation and by Western blotting with antibodies to PAN-14-3-3 proteins to detect the interaction between ME-hNedd4-2 and endogenous 14-3-3s. As shown in Fig. 2B (lanes 1-4), both the phosphorylation reaction and complex formation were enhanced significantly by stimulation with insulin and insulin-like growth factor I. hNedd4-2/ 14-3-3 complex formation was enhanced after transfection of Sgk1 cDNA into the cell (Fig. 2B, lanes 5-7), suggesting that Sgk1 is responsible for triggering the interaction. Furthermore, ME-hNedd4-2 was found to be indeed phosphorylated at Ser-468 in HEK293 cells as demonstrated by mass spectrometric analysis (Fig. 2C). However, we could not exclude the possibility that other protein kinases activated by insulin or insulinlike growth factor I, such as Sgk2 and Sgk3 (39), also contribute to the complex formation.
As shown in Fig. 3A, the expression of ␣ ␤ ␥ ENaC induced the 10 M amiloride-sensitive current at the holding potential of Ϫ60 mV in Xenopus oocytes (683 Ϯ 36 nA, n ϭ 9). This current was increased significantly by coexpression of Sgk1 kinase with ␣ ␤ ␥ ENaC (1,617 Ϯ 115 nA, n ϭ 9, p Ͻ 0.01;   Fig. 3B), which is consistent with the previous observation that Sgk1 interferes with Nedd4-2-dependent suppression of ENaC activity (18,19). We found, however, that coexpression of the V180D mutant completely abolished the stimulatory effect of Sgk1 on the ENaC-dependent Na ϩ current (311 Ϯ 33 nA, n ϭ 9, p Ͻ 0.01 versus ENaC ϩ Sgk1; Fig. 3D), whereas little effect was observed with the wild type 14-3-3 (1,457 Ϯ 103 nA, n ϭ 9; Fig. 3C) (see also Fig. 3E). We further examined whether the interaction of 14-3-3 and Nedd4-2 affects the cell surface density of ENaC by means of immunohistochemical detection of FLAG-tagged ENaC on the cryosections of Xenopus oocytes (see also "Experimental Procedures"). Consistent with the electrophysiological experiments, the cell surface density of ENaC was increased significantly by coexpression with Sgk1, whereas the effect of Sgk1 was completely abolished by coexpression of the V180D mutant but not the wild type 14-3-3 (data not shown). Taken together, these findings suggest that 14-3-3 participates in Sgk1-dependent control of the ENaC level through a phosphorylation-dependent interaction with hNedd4-2.
A Dominant Negative 14-3-3 Mutant Stimulates Nedd4-2catalyzed Ubiquitination of ENaC-Because hNedd4-2 is a ubiquitin E3 protein ligase, the binding of 14-3-3 to hNedd4-2 might affect the extent of ENaC ubiqutination. To test this possibility, we employed an in vivo coexpression assay in HEK293 cells. This assay is effective and used widely to characterize Nedd4-catalyzed ubiqutination of many proteins, including ENaC (13). HEK293 cells were cotransformed with FLAG-tagged ␣ENaC subunit and HA-tagged ubiquitin, together with the wild type or V180D 14-3-3 mutant. After 2 days in culture, the FLAG-␣ENaC was immunoprecipitated from the cells with monoclonal anti-FLAG, and the precipitate was analyzed by Western blotting with an HA polyclonal anti- body to detect HA-ubiquitinated FLAG-␣ENaC. As shown in Fig. 4A (lane 4), FLAG-␣ENaC was heavily ubiquitinated when expressed with the dominant negative V180D mutant of 14-3-3, as observed by a smear of high molecular weight species of HA-ubiquitin. However, the HA-ubiquitinated FLAG-␣ENaC was not detected or was observed at much lower levels when expressed with the wild type 14-3-3 (lane 6), as well as in a series of control cells transfected with HA-ubiquitin plus V180D without ␣ENaC, ␣ENaC alone, or HA-ubiquitin plus ␣ENaC without V180D (lanes 1-3, n ϭ 4). We found that the V180D mutant also stimulated the ubiquitination of ␥ENaC ( Fig. 4B; n ϭ 2). Taken together, these results suggest that 14-3-3 modulates the ubiquitination of ENaC subunits by associating with phosphorylated Nedd4-2 ubiquitin ligase.
14-3-3 Keeps hNedd4-2 Phosphorylated-14-3-3 exerts a variety of effects on phosphorylated targets, including protection from dephosphorylation (24,25). For example, the binding of 14-3-3 to Raf-1 maintains Raf-1 in a phosphorylated, inactive form by reducing the rate of dephosphorylation (42). Likewise, the binding of 14-3-3 to NUDEL prevents the dephosphorylation (and inactivation) of NUDEL, a protein associated with the pathology of Miller-Dieker syndrome (43). To test the effect of 14-3-3 on cellular levels of phosphorylated hNedd4-2 protein, a series of transformants was prepared with HEK293 cells expressing ME-tagged hNedd4-2, ME-Nedd4-2 plus 14-3-3, or ME-Nedd4-2 plus V180D 14-3-3 mutant. Each group of cells was labeled metabolically with 32 P i , and the level of phosphorylation was monitored by measuring 32 P incorporation into ME-hNedd4-2 that was pulled down from the transformed cells. As shown in Fig. 5A, the level of 32 P incorporated into ME-hNedd4-2 recovered from cells overexpressing wild type 14-3-3 was approximately equivalent to that recovered from control cells, whereas the same protein recovered from V180D-expressing cells con-tained only ϳ50% of the level of 32 P in six independent analyses (compare lane 3 with lanes 1 and 2; p Ͻ 0.01). In addition, we found that the levels of phosphorylated ME-hNedd4-2 increased severalfold in response to 1 M okadaic acid, a potent inhibitor of PP1 and PP2A, included in the culture medium during the metabolic labeling step (data not shown).

DISCUSSION
The activity of the ENaC sodium channel is regulated by a variety of mechanisms including ubiquitination and phosphorylation (6). A ubiquitin E3 protein ligase, Nedd4-2, mediates the ubiquitination of ENaC to target it for degradation. This reduces the copy number of the ENaC channel at the plasma membrane and drastically decreases the Na ϩ current in X. laevis oocytes or Fischer rat thyroid cells (10 -12). On the other hand, the Ser/Thr kinase Sgk1 induces the cell surface expression of ENaC and increases the Na ϩ current in Xenopus oocytes (16,17). It was shown recently that Sgk1 and Nedd4-2 converge into a common cell signaling pathway and regulate the cell surface density of ENaC, where Sgk1 phosphorylates Nedd4-2, thereby reducing both its binding affinity to ENaC and the rate of ENaC degradation (18,19). This finding was supported further by observations that the mutation of the Sgk1 phosphorylation sites in Nedd4-2 abrogates the effect of Sgk1 on ENaC-dependent Na ϩ currents in X. laevis oocytes and Fischer rat thyroid cells (18,19,41,44). We report here that 14-3-3 participates in this regulatory process and modulates the Na ϩ current cooperatively with the Sgk1-mediated phosphorylation of hNedd4-2. We propose that the formation of a complex between 14-3-3 and phosphorylated hNedd4-2 maintains the ubiquitin ligase in a phosphorylated, inactive form possibly by blocking the access of phosphatases to the phosphorylated enzyme, thereby increasing the density of ENaC on the surface of the plasma membrane. Recently, it was reported that Nedd4-2 is also phosphorylated by protein kinase A and that the protein kinase A-mediated phosphorylation inhibits Nedd4-2 by decreasing its binding to ENaC (45). Importantly, overexpression of Sgk1 blunts ENaC stimulation by cAMP, and conversely, cAMP agonists decrease ENaC stimulation by Sgk1 (45). Thus, it appears that 14-3-3 also participates in the protein kinase A-mediated control of ENaC level through an interaction with the phosphorylated form of Nedd4-2, although this awaits further investigation.
Our strategy to identify hNedd4-2 as a 14-3-3-binding partner was based on the MS-based proteomics technology com- HEK293 cells transiently cotransfected with the indicated constructs were lysed, and the lysates were incubated with FLAG-agarose beads to precipitate the FLAG-tagged ENaC subunits. The precipitated proteins were separated by 10% SDS-PAGE, transferred to polyvinylidene difluoride, and blotted with anti-FLAG to detect total ENaC subunits or with anti-HA to detect specifically the ubiquitinated ENaC subunits. Ubiquitinated species are shown as high molecular mass smeared bands, as indicated with a bar, and the arrows indicate the ENaC subunits. Similar results were obtained for at least two independent experiments. bined with a novel tandem affinity purification tag, called MEF. The general strategy of MEF-mediated purification is similar to previously published protocols using TAP (46) and CHH (47) and has the following characteristics: (i) the MEF tag introduces a relatively small polypeptide into the bait protein; (ii) the specific antibodies are commercially available to pull down both myc-and FLAG-tagged protein complexes; and (iii) the purification procedure does not require specific divalent cations such as Ca 2ϩ and Ni 2ϩ for affinity selection. These factors may be efficient for purification of multiprotein complexes to homogeneity and to avoid contaminants caused by nonspecific interaction. In fact, no background proteins were detectable in the final fraction of the candidates of 14-3-3binding partners as analyzed by SDS-PAGE and silver staining (Fig. 1A, lane 1). Because the FLAG peptide is available to dissociate specifically the resultant FLAG-tagged protein complex from an immobilized support, 2 the MEF purification procedure, coupled with highly sensitive direct nanoflow LC-MS/MS technology (32), will be similarly useful to characterize many physiological multiprotein complexes in vivo beyond the 14-3-3 complexes reported here.
Nedd4-2 associates with ENaC via direct interaction of its WW domains with the PY motif located in the C-terminal portion of each ENaC subunit (9 -12, 18, 48). The WW domain is a widespread, tryptophan-based protein-protein interaction module comprising ϳ40 amino acids that serve as a surface for hydrophobic interaction with a variety of proline-containing motifs, including the PY motif (49,50). Previous studies showed that the interaction between a subset of WW domains and proline-containing motifs is regulated by site-specific phosphorylation of a Ser/Thr or Tyr residue within the WW domain or the proline-containing motif. For example, the interaction of Wwox, a tumor suppressor protein containing two WW domains, with the p53 homolog p73 is enhanced significantly by Tyr phosphorylation within the first WW domain (51). It is also known that the mitotic prolyl isomerase Pin1 can bind the microtubule-associated protein tau only when tau is phosphorylated at Thr within the p(S/T)P motif (52,53). From this perspective, the regulatory mechanism of the Nedd4-2/ENaC interaction reported here appears unique in that the phosphorylation occurs at a residue located between the WW domains. Moreover, this phosphorylation triggers the complex formation with a 14-3-3 protein. We propose that 14-3-3 stabilizes the conformation of phosphorylated Nedd4-2 or induces a conformational change in phosphorylated hNedd4-2 by associating with phosphoserine 468. A previous crystallographic study indicated that the complex between 14-3-3 and the synthetic phosphopeptide MARSHpSYPAKK (based on the polyomavirus middle T antigen) is stabilized by salt bridges or hydrogen bonds between four residues in the 14-3-3 dimer (Arg-56, Arg-127, Lys-49, and Tyr-128) and phosphoserine, Lys-120 and a proline at the ϩ2 position in the peptide (22). Interestingly, the proline residue is fixed in all-cis configuration within the complex (22). These data suggest that 14-3-3 might exert its biological activity by inducing a conformational change in hNedd4-2 through proline 470 isomerization or by stabilizing this proline in a cis configuration against its cis-trans equilibrium. Future studies will explore this hypothesis.
Many ion channels or proteins other than ENaC have been identified as substrates of Nedd4-2 ubiquitin ligase, including voltage-gated Na ϩ channel (Nav1.5; Ref. 54 FIG. 5. 14-3-3 maintains hNedd4-2 phosphorylation and protects it from phosphatase attack. A, HEK293 cells transiently expressing ME-hNedd4-2 (control), ME-hNedd4-2 plus wild type 14-3-3, or ME-hNedd4-2 plus the V180D 14-3-3 mutant, were metabolically labeled as in Fig. 2B. The expressed ME-hNedd4-2 was immunoprecipitated from the lysate and then analyzed by SDS-PAGE and Coomassie staining (top), autoradiography (middle), and scintillation counting (bottom). Radiation was normalized as percentage of the control. n ϭ 6. B, GST-hNedd4-2 and the phosphorylation mutant (GST-hNedd-S468A) were phosphorylated with Sgk1 and [␥-32 P]ATP. The 32 P-labeled proteins were incubated with 0.5 unit of PP1 in the presence or absence of 14-3-3 or V180D mutant, as indicated. Each reaction was analyzed by SDS-PAGE and Coomassie staining (top two panels), autoradiography (middle), and scintillation counting (bottom). Data are presented as the percentage of 32 P i incorporated into GST-hNedd4-2 or the mutant prior to the addition of PP1. n ϭ 2. each case, the Nedd4-2 WW domains mediate substrate binding. Our study identifies 14-3-3 as a binding partner of Nedd4-2; however, 14-3-3 appears not to be a substrate but rather a cofactor for this ubiquitin ligase. In fact, the interaction between these proteins was strictly regulated by sitespecific phosphorylation of Nedd4-2, and the binding site was distinct from the WW domains, that is, within the RXRXXpSXP segment including phosphoserine 468, spanning domains WW2 and WW3. Thus, the 14-3-3 family represents a novel class of Nedd4-2-binding proteins that affect how Nedd4-2 interacts with ENaC, a paradigm that may hold for other potential Nedd4-2 substrates as well.