Inhibition of the Epithelial Na+ Channel by Interaction of Nedd4 with a PY Motif Deleted in Liddle’s Syndrome*

The epithelial Na+ channel (ENaC) plays a critical role in Na+ absorption in the kidney and other epithelia. Mutations in the C terminus of the β or γENaC subunits increase renal Na+ absorption, causing Liddle’s syndrome, an inherited form of hypertension. These mutations delete or disrupt a PY motif that was recently shown to interact with Nedd4, a ubiquitin-protein ligase expressed in epithelia. We found that Nedd4 inhibited ENaC when they were coexpressed in Xenopusoocytes. Liddle’s syndrome-associated mutations that prevent the interaction between Nedd4 and ENaC abolished inhibition, suggesting that a direct interaction is required for inhibition by Nedd4. Inhibition also required activity of a ubiquitin ligase domain within the C terminus of Nedd4. Nedd4 had no detectable effect on the single channel properties of ENaC. Rather, Nedd4 decreased cell surface expression of both ENaC and a chimeric protein containing the C terminus of the β subunit. Decreased surface expression resulted from an increase in the rate of degradation of the channel complex. Thus, interaction of Nedd4 with the C terminus of ENaC inhibits Na+ absorption, and loss of this interaction may play a role in the pathogenesis of Liddle’s syndrome and other forms of hypertension.

The epithelial Na ؉ channel (ENaC) plays a critical role in Na ؉ absorption in the kidney and other epithelia. Mutations in the C terminus of the ␤ or ␥ENaC subunits increase renal Na ؉ absorption, causing Liddle's syndrome, an inherited form of hypertension. These mutations delete or disrupt a PY motif that was recently shown to interact with Nedd4, a ubiquitin-protein ligase expressed in epithelia. We found that Nedd4 inhibited ENaC when they were coexpressed in Xenopus oocytes. Liddle's syndrome-associated mutations that prevent the interaction between Nedd4 and ENaC abolished inhibition, suggesting that a direct interaction is required for inhibition by Nedd4. Inhibition also required activity of a ubiquitin ligase domain within the C terminus of Nedd4. Nedd4 had no detectable effect on the single channel properties of ENaC. Rather, Nedd4 decreased cell surface expression of both ENaC and a chimeric protein containing the C terminus of the ␤ subunit. Decreased surface expression resulted from an increase in the rate of degradation of the channel complex. Thus, interaction of Nedd4 with the C terminus of ENaC inhibits Na ؉ absorption, and loss of this interaction may play a role in the pathogenesis of Liddle's syndrome and other forms of hypertension.
The epithelial Na ϩ channel (ENaC) 1 is expressed at the apical membrane of a variety of epithelia, including the kidney, lung, and colon (1,2). Because Na ϩ entry through ENaC is the rate-limiting step for Na ϩ absorption, regulation of this channel plays a critical role in controlling extracellular fluid volume and blood pressure. The channel consists of three homologous subunits (␣, ␤, and ␥) (3)(4)(5)(6)(7). Mutations that truncate the cytoplasmic C terminus of the ␤ or ␥ subunits increase Na ϩ absorption in the kidney collecting duct, resulting in Liddle's syndrome, an autosomal dominant form of hypertension (8,9). Previous studies found that Liddle's syndrome-associated mutations increased Na ϩ current compared with the wild-type channel when expressed in Xenopus oocytes or renal epithelia (10,11). By truncating the C terminus, these mutations delete a sequence that is conserved in all three subunits (PP-PXYXXL). Two findings suggest that this tyrosine-based sequence plays an important role in controlling Na ϩ absorption; missense mutations in this sequence increased Na ϩ current similar to the C-terminal truncations (11,12), and mutations in this sequence have been identified in families with Liddle's syndrome (13,14). Deletion or disruption of the tyrosine-based sequence increased Na ϩ current at least in part by increasing the number of Na ϩ channels at the cell surface (10,11). Although it is not yet known how these mutations increase cell surface expression, similarity of the sequence to previously described protein motifs suggest two potential mechanisms.
First, the PPPXYXXL sequence is similar to two different tyrosine-based internalization motifs found in proteins such as the LDL receptor (NPXY) (15) and the transferrin receptor (YXXL/hydrophobic) (16). If this sequence functions as an internalization signal in ENaC, then Liddle's mutations might lead to an accumulation of ENaC at the cell surface by disrupting channel internalization.
Second, the tyrosine-based sequence fits the consensus of a recently defined motif that plays a role in protein-protein interactions, the PY motif (PPXY) (17). PY motifs were identified by their ability to bind to a 35-40-amino acid sequence containing two conserved tryptophan residues (WW domain) (17). Although the functional effects of these interactions are not yet known, this observation suggests that an interacting protein might control the surface expression of ENaC. A candidate is Nedd4, a protein expressed in epithelia that contains three WW domains (18 -20). Using a biochemical assay, Staub and co-workers found that Nedd4 interacted with the PY motifs of ENaC (21). In addition, mutations that cause Liddle's syndrome abolished this interaction. Thus, it is tempting to speculate that Nedd4 inhibits Na ϩ absorption through ENaC and that loss of this inhibition plays a role in Liddle's syndrome. However, direct evidence to support this hypothesis is lacking. Here we coexpress Nedd4 with ENaC to directly test the hypothesis that Nedd4 inhibits ENaC, demonstrating the functional importance of this WW domain-PY motif interaction.

EXPERIMENTAL PROCEDURES
DNA Constructs-Nedd4 (rat) was cloned by polymerase chain reaction of cDNA reverse transcribed from rat lung poly(A) ϩ RNA (CLON-TECH). Two independent clones were generated that produced identical functional results. A Nedd4 construct containing only the three WW domains (Nedd4 WW ) was generated by polymerase chain reaction (amino acids 233-500). Mutation of Cys 854 to alanine (Nedd4 C854A ) was performed by site-directed mutagenesis (QuickChange, Stratagene). cDNAs were subcloned into pMT3 for expression in Xenopus oocytes.
Expression and Electrophysiology in Xenopus Oocytes-␣, ␤, and ␥hENaC in pMT3 (0.2 ng each) were coexpressed with either wild-type or mutant Nedd4 or an irrelevant protein (secreted alkaline phosphatase, SEAP) (0.2 or 0.8 ng) in Xenopus oocytes by nuclear injection of cDNA. Expression of SEAP with ENaC (␣, ␤, and ␥) did not alter Na ϩ current compared with expression of ENaC alone. One to two days after injection, whole cell current was measured by two-electrode voltage-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Supported by a Student Research Fellowship from the American Heart Association.
ʈ Supported by the Roy J. Carver Charitable Trust and by the NHLBI and NIDDK, National Institutes of Health. To whom correspondence should be addressed: 200K EMRB, Dept. of Internal Medicine, University of Iowa College of Medicine, Iowa City, IA 52242. E-mail: psnyder@blue.weeg.uiowa.edu. 1 The abbreviations used are: ENaC, epithelial Na ϩ channel; SEAP, secreted alkaline phosphatase; MTS, methanethiosulfonate. clamp at Ϫ60 mV with oocytes bathed in 116 mM NaCl, 2 mM KCl, 0.4 mM CaCl 2 , 1 mM MgCl 2 , and 5 mM HEPES (pH 7.4). Single channel currents were recorded by cell-attached patch-clamp in devitellinized oocytes one day after injection as described previously (11). The extracellular solution was identical to that described above, and LiCl replaced NaCl in the pipette solution. Statistical significance was assessed using a Student's unpaired t test.
Surface Expression of ENaC in Xenopus Oocytes-␣ S549C , ␤ S520C , and ␥ S529C (22) were coexpressed (ENaC Cys ) in Xenopus oocytes (0.2 ng each) with either Nedd4 or SEAP (0.8 ng). As a negative control, we injected oocytes with the vector (pMT3) alone. 1 h after injection, we treated oocytes with 1 mM (2-aminoethyl)methanethiosulfonate hydrobromide (Toronto Research Chemicals) for 10 min at room temperature to modify endogenous sulfhydryls. Similar results were also obtained using [2-(trimethylammonium) MTS-4-fluorescein irreversibly stimulates ENaC Cys but not wild-type ENaC. Nedd4 did not alter the ability of MTS-4-fluorescein to stimulate ENaC Cys , suggesting that Nedd4 did not alter accessibility of the introduced cysteines. Cell surface fluorescence was quantitated by confocal microscopy (Bio-Rad MRC-600, krypton/argon laser). Optical sections obtained at 30-m steps were superimposed, mean fluorescence at the plasma membrane was determined, and background (pMT3 alone) was subtracted.
Expression of ENaC with Nedd4 in COS-7 Cells-COS-7 cells were electroporated with ␣, ␤, and ␥ENaC (10 g each) with or without Nedd4 (10 g), pulse-labeled with 100 Ci/ml [ 35 S]methionine (New England Nuclear) at 37°C for 30 min, and then chased for 0 -4 h. Proteins were solubilized in cold lysis buffer containing 1% Triton X-100, 0.4 mM phenylmethylsulfonyl fluoride, 20 g/ml aprotonin, 20 g/ml leupeptin, and 10 g/ml pepstatin A. Protein insoluble to Triton X-100 was solubilized in 2% SDS at 95°C and then diluted 10-fold in lysis buffer containing 1% Triton X-100. ␣ENaC in the Triton X-100 soluble and insoluble fractions was immunoprecipitated with anti-Flag M2 antibody (Eastman Kodak), which recognized a Flag epitope in the extracellular domain, as described previously (11). Immunoprecipitated proteins were separated by SDS-polyacrylamide gel electrophoresis and imaged and quantitated by phosphorimaging.

Inhibition of ENaC by Nedd4 -To test the hypothesis that
Nedd4 inhibits ENaC Na ϩ current, we coexpressed Nedd4 with ENaC in Xenopus oocytes and measured the whole cell Na ϩ current blocked by amiloride. When we injected equal amounts of cDNA encoding Nedd4 and each ENaC subunit, Nedd4 decreased amiloride-sensitive Na ϩ current 53% (Fig. 1, A-C). Injection of a 4-fold higher amount of Nedd4 cDNA decreased current 88% (Fig. 1B). In contrast to its effect on Na ϩ current, Nedd4 did not alter several characteristic biophysical properties of ENaC, including its voltage independence (Fig. 1C), its high selectivity for Na ϩ over K ϩ , and its sensitivity to amiloride (not shown).
Biochemical studies showed that Nedd4 binds to the PY motif in the C terminus of ENaC subunits (21). The Liddle's syndrome mutation ␤ R566X deletes most of the cytoplasmic C terminus of the ␤ subunit, including the PY motif. We therefore asked whether this mutation would disrupt inhibition by Nedd4. Coexpression of wild-type ␣ and ␥ENaC with ␤ R566X increased Na ϩ current compared with wild-type ␤ (Fig. 2), as previously reported (10,11). However, Nedd4 did not inhibit the mutant channel, suggesting that the C terminus of ␤ENaC was required for inhibition by Nedd4. Mutation of Tyr 620 in the PY motif to alanine also increased Na ϩ current and disrupted inhibition by Nedd4 (Fig. 2). This tyrosine is required for WW domain-PY motif interactions (17), and an equivalent mutation in a peptide derived from the PY motif of ␤ENaC abolished binding of the peptide to Nedd4 (21). These data suggest that a direct interaction between Nedd4 and the PY motif in ␤ENaC is required for channel inhibition by Nedd4.
Role of Nedd4 Ubiquitin Ligase Domain-Although the WW domains of Nedd4 are sufficient for binding to the C-terminal PY motifs of ENaC (21), Nedd4 also contains a calcium phospholipid binding domain (C2 domain (23)) and a ubiquitin ligase domain (20,21). To determine whether these domains are required for inhibition by Nedd4, we tested a construct containing only the three WW domains (Nedd4 WW ). In contrast to wild-type Nedd4, expression of Nedd4 WW with ENaC did not decrease Na ϩ current (Fig. 3A). Thus, binding of the WW domains to the PY motif alone is not sufficient to inhibit ENaC; this suggests that the C 2 domain, ubiquitin ligase domain, or both are also required.
Ubiquitin ligase domains transfer ubiquitin to target substrates. A conserved cysteine within the domain forms a thioester bond with ubiquitin, and mutation of this cysteine abolishes its ability to accept ubiquitin (24). To test the hypothesis that the ubiquitin ligase domain is required for ENaC inhibition by Nedd4, we mutated the conserved cysteine to alanine (Nedd4 C854A ). This mutation disrupted inhibition of ENaC by Nedd4; expression of equal amounts of cDNA encoding Nedd4 C854A and each ENaC subunit did not significantly decrease Na ϩ current (Fig. 3B). A 4-fold higher amount of Nedd4 C854A cDNA decreased Na ϩ current 33% (Fig. 3B), although this was significantly less than the decrease produced by wild-type Nedd4. Therefore, the ubiquitin ligase domain and Cys 854 play an important role in the inhibition of ENaC by Nedd4. However, the small amount of inhibition produced by higher concentrations of Nedd4 C854A indicates that the mutant has partial activity. Thus, part of the inhibitory effect of Nedd4 may not require its ubiquitin ligase activity.
Single Channel Properties of ENaC Coexpressed with Nedd4 -Nedd4 could inhibit ENaC by decreasing open-channel probability (P o ), single channel conductance, or the number of channels at the cell surface. To test the first two possibilities, we measured the single channel properties of ENaC expressed with Nedd4 in Xenopus oocytes. ENaC is characterized by very slow kinetics, with long channel openings and closings (2,5,11). Fig. 4A shows a patch containing a single channel from a cell coexpressing ENaC and Nedd4. Nedd4 did not alter the slow kinetics of ENaC or produce a detectable decrease in P o (Fig. 4B). Fig. 4C shows the single channel current-voltage relationship for ENaC with and without Nedd4; Nedd4 did not decrease the single channel conductance. Therefore, by exclusion these results suggest that Nedd4 inhibits ENaC by decreasing the number of channels at the cell surface.
Nedd4 Binding Decreases Cell Surface Expression of ENaC-To test the hypothesis that Nedd4 decreased the number of channels at the cell surface, we measured surface expression of ENaC in Xenopus oocytes using a fluorescence assay. Exposed cysteines in a protein at the cell surface can be covalently modified by MTS compounds added to the extracellular bathing solution (25,26). To fluorescently label ENaC, we substituted a cysteine in the ␣ (␣ S549C ), ␤ (␤ S520C ), and ␥ (␥ S529C ) subunits at a position that can be covalently modified from the extracellular side of the plasma membrane (22). Following modification of the introduced cysteines with MTS-4fluorescein, a fluorescent MTS derivative, we detected and quantitated ENaC at the cell surface using confocal microscopy, similar to a method previously described (27). To minimize background, we used a nonfluorescent MTS compound ((2-aminoethyl)methanethiosulfonate hydrobromide) to modify cysteines in endogenous cell surface proteins 1 h after cDNA injection (16 -20 h before study). Oocytes expressing the cysteine-tagged ENaC (ENaC Cys ) had increased cell surface fluorescence compared with oocytes injected with vector cDNA lacking ENaC (Fig. 5A). Wild-type ENaC did not increase fluorescence relative to background, indicating that the introduced cysteines were required for the fluorescent signal. Coexpression of ENaC Cys with Nedd4 decreased fluorescence (Fig.  5), suggesting that Nedd4 decreased cell surface expression of ENaC. In contrast, Nedd4 did not decrease surface expression when the ␤ subunit contained the Y620A mutation within the PY motif (Fig. 5), consistent with our finding that this mutation prevented Nedd4 from decreasing ENaC Na ϩ current. These data suggest that binding of Nedd4 to the C-terminal PY motif inhibits ENaC by decreasing the number of channels at the cell surface.
Nedd4 Decreases Cell Surface Expression of a Chimeric Protein Containing the C Terminus of ␤ENaC-To determine whether the interaction of Nedd4 with the C-terminal PY motif alone is sufficient to decrease surface expression or whether other ENaC domains are also required, we asked whether Nedd4 would decrease surface expression of an unrelated protein containing the C-terminal PY motif. We fused the extra-cellular and membrane-spanning domains of A2, an HLA protein, to the cytoplasmic C terminus of ␤ENaC (A2-␤, Fig. 6A). Following expression in COS-7 cells, we biotinylated cell surface proteins and detected and quantitated biotinylated A2-␤ as an assay of its surface expression. Expression of A2-␤ produced a 55-kDa biotinylated band (Fig. 6B) corresponding to the expected size of A2-␤ (11). This band was not seen in cells that were not biotinylated (Fig. 6B) or in untransfected cells (not shown). Coexpression with Nedd4 produced a significant decrease in biotinylated A2-␤, indicating that Nedd4 decreased the amount of A2-␤ at the cell surface (Fig. 6B). In contrast, Nedd4 did not decrease surface expression of a construct containing the Y620A mutation (Fig. 6C). Thus, interaction of Nedd4 with the C-terminal PY motif was sufficient to decrease surface expression of an unrelated protein.
Decreased Stability of ENaC Expressed with Nedd4 in COS-7 Cells-Our finding that Nedd4 decreased the number of ENaC channels at the cell surface, coupled with the requirement for the ubiquitin ligase activity of Nedd4, suggests that Nedd4 might target ENaC for degradation. To test this hypothesis, we coexpressed ␣, ␤, and ␥ENaC in COS-7 cells with or without Nedd4. Following pulse labeling with [ 35 S]methionine (0 h chase) and solubilization with Triton X-100, we immunoprecipitated the ␣ subunit as a marker for the channel complex. This produced two predominant bands corresponding to the unglycosylated and glycosylated forms of ␣ENaC (Fig. 7A). When we coexpressed ENaC with Nedd4, an additional band was observed, corresponding to Nedd4 that coprecipitated with ENaC. Nedd4 did not decrease the amount of 35 S-labeled ␣ENaC immediately after pulse labeling, suggesting that Nedd4 did not decrease the rate of synthesis of ␣ENaC. In contrast, Nedd4 decreased the amount of ␣ENaC protein dur- ing the 4-h chase (Fig. 7B). Nedd4 also decreased soluble ␤ENaC that coprecipitated with ␣ENaC (not shown). Thus, Nedd4 decreased the stability of the channel complex, perhaps by targeting the channel for degradation. DISCUSSION In the collecting duct of the kidney, Na ϩ absorption through ENaC is highly regulated, varying from maximal absorption at times of Na ϩ depletion to abolition of absorption in states of Na ϩ excess. However, in contrast to voltage-gated and ligandgated ion channels, ENaC is constitutively active (5,7). This suggests that control of ENaC surface expression may be critical in the regulation of Na ϩ absorption. The finding that Liddle's syndrome mutations increase Na ϩ absorption at least in part by increasing the number of channels at the cell surface supports this hypothesis. Our results suggest that Nedd4 may play an important role in the regulation of Na ϩ absorption by controlling the cell surface expression of ENaC.
Nedd4 interacts with ENaC through the binding of its WW domains to PY motifs in the C terminus of ENaC. This interaction is required for Nedd4 to inhibit ENaC; mutations that abolished Nedd4 binding prevented inhibition of ENaC by Nedd4. The PY motif-WW domain interaction may provide specificity, targeting the activity of Nedd4 to ENaC and possibly other proteins containing PY motifs. However, interaction between ENaC and the WW domains alone was not sufficient to inhibit ENaC, suggesting that other sequences within Nedd4 are required for decreased surface expression.
Several observations suggest that the ubiquitin ligase domain of Nedd4 plays an important role in the inhibition of ENaC. First, this domain is homologous to E3 ubiquitin-protein ligases, which transfer ubiquitin to target proteins (24), increasing their rate of degradation. Recent work indicates that Nedd4 has ubiquitin ligase enzymatic activity (28). Second, it was recently reported that lysines within the N termini of ␣ and ␥ENaC are substrates for ubiquitination and that mutation of these lysines increased ENaC Na ϩ current in Xenopus oocytes (29). However, the role of Nedd4 in ubiquitination of these lysines is unknown. Finally, we found that mutation of a residue in Nedd4 required for ubiquitin conjugation (Cys 854 ) significantly decreased the ability of Nedd4 to inhibit ENaC. This suggests a model in which the WW domain-PY motif interaction directs the ubiquitin ligase domain to the channel, where it can ubiquitinate one or more subunit(s) to target the complex for degradation (Fig. 8). Consistent with this model, we found that Nedd4 increased the rate of degradation of ENaC. It is also possible that Nedd4 interacts with ENaC in an intracellular compartment, preventing transport of the channel complex to the cell surface.
Nedd4 might also decrease ENaC at the cell surface by increasing the rate of channel internalization by two potential mechanisms. First, ubiquitination by Nedd4 might stimulate channel internalization (Fig. 8); ubiquitination was recently shown to be required for internalization of the yeast pheromone receptor Ste2p (30). Second, the Nedd4 C 2 domain could play a role in the internalization of ENaC; a C 2 domain within synaptotagmin I mediates endocytosis of synaptic vesicles by binding to the clathrin AP-2 complex (31). Our finding that Nedd4 C854A retained partial activity supports such a functional role for the C 2 domain in the inhibition of ENaC. A recent report that a dominant negative dynamin increased ENaC current in Xenopus oocytes (32) also supports a role for endocytosis in controlling Na ϩ absorption. Although our results indicate that the PPPXYXXL sequence functions as a PY motif, similarity of this sequence to tyrosine-based internalization motifs suggests that it could also play a role in ENaC internalization independent of its function as a PY motif (11).
A recent report by Dinudom et al. (33) suggests that Nedd4 or a related WW domain protein may be involved in inhibition of Na ϩ current in response to high levels of intracellular Na ϩ in mouse mandibular salivary duct cells. They found that a WW domain peptide or a polyclonal antibody against Nedd4 disrupted inhibition of Na ϩ current in response to 72 mM intracellular Na ϩ . A dominant-negative ubiquitin also disrupted inhibition by high Na ϩ , consistent with our finding that the ubiquitin ligase activity of Nedd4 plays an important role in its ability to inhibit ENaC. However, it is not known whether ENaC is responsible for Na ϩ current in mouse mandibular salivary duct cells, and the molecular identity of the WW domain protein involved in inhibition is also unknown. Interestingly, a number of other WW domain-containing proteins have been identified (20,34), including several with significant homology to Nedd4, and some have been shown to bind to the PY motifs of ENaC (34). Thus, the WW domain might serve as an adapter to allow a variety of regulatory proteins with diverse functions to bind to the C terminus of ENaC, regulating channel function and, hence, Na ϩ absorption.
Nedd4 contains three WW domains, and the ␣, ␤, and ␥ENaC FIG. 7. Nedd4 decreases stability of ENaC. ␣, ␤, and ␥ENaC were expressed with or without Nedd4 (as indicated) in COS-7 cells, pulse labeled with [ 35 S]methionine, chased for 0 -4 h, and solubilized with Triton X-100, and ␣ENaC was immunoprecipitated. A, immunoprecipitation immediately after pulse labeling. ␣ENaC and Nedd4 are indicated. At later time points, ␤ENaC that coprecipitated with ␣ENaC could also be identified on the gel based on its different mobility. ␥ENaC could not be detected because its migration is similar to that of ␣ENaC. B, quantitation of total ␣ENaC (Triton X-100 soluble and insoluble) by phosphorimaging (relative to t ϭ 0) at 0 -4 h chase. Similar results were obtained in three experiments. subunits each contain a PY motif. It therefore seems likely that Nedd4 binds to more than one subunit within a channel complex. Alternatively, Nedd4 might cluster channels by binding to PY motifs from two or three different channels. Interestingly, we found that mutation or deletion of the ␤ENaC PY motif prevented inhibition by Nedd4, suggesting that interaction with the ␤ subunit is required for channel inhibition. One interpretation of this result is that the ␤ subunit is the functionally relevant substrate for Nedd4. Alternatively, the functional interaction between Nedd4 and ENaC might require a specific pattern of binding to PY motifs in more than one subunit.
In Liddle's syndrome, deletion or mutation of conserved amino acids in the PY motif in ENaC increases Na ϩ absorption, resulting in hypertension. Because Liddle's syndrome-associated mutations abolished binding and inhibition of ENaC by Nedd4, our data suggest that loss of control of ENaC surface expression by Nedd4 might be involved in the pathophysiology of Liddle's syndrome. It seems possible that loss of function mutations in Nedd4 could also underlie other forms of hypertension. Thus, an understanding of the control of ENaC by Nedd4 and other related proteins may provide new insight into electrolyte homeostasis and the pathophysiology of hypertension.