Serum and Glucocorticoid-regulated Kinase Modulates Nedd4-2-mediated Inhibition of the Epithelial Na (cid:1) Channel*

The epithelial Na (cid:1) channel (ENaC) forms the pathway for Na (cid:1) absorption across epithelia, including the kidney collecting duct, where it plays a critical role in Na (cid:1) homeostasis and blood pressure control. Na (cid:1) absorption is regulated in part by mechanisms that control the expression of ENaC at the apical cell surface. Nedd4 family members ( e.g. Nedd4, Nedd4-2) bind to the channel and decrease its surface expression by catalyzing its ubiquitination and degradation. Conversely, serum and glucocorticoid-regulated kinase (SGK), a downstream mediator of aldosterone, increases the expression of ENaC at the cell surface. Here we show that SGK and human Nedd4-2 (hNedd4-2) converge in a common pathway to regulate epithelial Na (cid:1) absorption. Consistent with this model, we found that SGK bound to hNedd4-2 and hNedd4. A PY motif in SGK mediated the interaction and was required for SGK to stimulate ENaC. SGK phosphorylated hNedd4-2 (but not hNedd4), altering hNedd4-2 function; phosphorylation reduced the binding of hNedd4-2 to (cid:2) ENaC, and hence, the hNedd4-2-mediated inhibition of Na (cid:1) absorption. These data suggest that SGK regulates epithelial Na (cid:1) absorption in part by modulating the

The epithelial Na ϩ channel (ENaC) 1 forms the pathway for Na ϩ absorption across a variety of epithelia, including the kidney collecting duct, airway, and distal colon (reviewed in Refs. 1 and 2). Three homologous subunits (␣-, ␤-, and ␥ENaC) form the channel complex (3,4). Dominant gain of function mutations in ENaC cause Liddle's syndrome, an inherited form of hypertension (5). Conversely, loss of function mutations cause salt wasting and hypotension (pseudohypoaldosteronism type I) (5). Thus, the regulation of ENaC is critical for the maintenance of Na ϩ homeostasis and for blood pressure control.
In the kidney collecting duct, Na ϩ absorption must vary over a wide range in response to conditions of Na ϩ depletion or Na ϩ excess. This occurs in large part by mechanisms that modulate the expression of ENaC at the apical cell surface (reviewed in Ref. 6). For example, the Nedd4 family of ubiquitin proteinligases (including Nedd4 and Nedd4-2) reduce ENaC surface expression (7)(8)(9). They contain multiple WW domains that bind to PY motifs in the C termini of ␣-, ␤-, and ␥ENaC (10 -12). This interaction facilitates the ubiquitination of ENaC, catalyzed by a ubiquitin ligase domain at the C terminus of Nedd4 family members (7,8). Ubiquitination reduces ENaC at the cell surface by increasing the rate of channel degradation (7). Liddle's syndrome is caused by defects in this regulatory pathway; mutations in the ENaC PY motifs disrupt their interaction with Nedd4 family members, resulting in increased expression of ENaC at the cell surface, and hence, excessive Na ϩ absorption (10,13,14). Thus, the Nedd4 family of proteins is critically important in reducing Na ϩ absorption.
Conversely, the renin-angiotensin-aldosterone pathway increases renal Na ϩ absorption, in part by increasing the expression of ENaC at the cell surface (15). This pathway plays a key role in responding to Na ϩ depletion and hypovolemia. Moreover, disruption of this pathway underlies several acquired and genetic disorders of blood pressure control, including primary aldosteronism and glucocorticoid-remediable aldosteronism (5). An important downstream mediator of aldosterone is serum and glucocorticoid-regulated kinase (SGK), (16,17). SGK transcription is induced by aldosterone over a very rapid time course (30 -60 min) (16,17), and it is post-translationally activated in response to insulin and other stimuli by phosphorylation through the phosphoinositide 3-kinase pathway (18). Thus, it has been proposed that SGK integrates a variety of signals that modulate renal Na ϩ absorption (19). SGK increases the expression of ENaC at the cell surface (20), but little is known about the mechanisms involved. Previous work reported that SGK phosphorylates Ser/Thr residues within the sequence RXRXXS/T (18,21). However, ENaC subunits are not phosphorylated by SGK (19). Thus, it seems likely that SGK phosphorylates one or more proteins involved in controlling ENaC surface expression, although such SGK substrates have not yet been identified.
Two observations suggest the possibility that the Nedd4 family and SGK might converge in a common pathway to regulate ENaC surface expression. First, SGK contains a PY motif (see Fig. 1A), suggesting that it might bind directly to WW domains in Nedd4 or Nedd4-2. Second, Nedd4-2 (but not hNedd4) contains three sequences that fit the consensus for phosphorylation by SGK, suggesting it might be a SGK substrate. The goal of this work was to test the hypothesis that SGK modulates ENaC surface expression in part through the phosphorylation of Nedd4-2.

EXPERIMENTAL PROCEDURES
DNA Constructs-hNedd4 was cloned as described previously (12). hNedd4-2 and human SGK were cloned by PCR of cDNA reverse transcribed from kidney poly(A) ϩ RNA (CLONTECH). Mutations were created using QuikChange Kit (Stratagene) and each cDNA was sequenced in the University of Iowa DNA Sequencing Core. Human ␣-, ␤-, and ␥ENaC in pMT3 or pcDNA3 were cloned as described previously (4). The PY motifs of each subunit were mutated to PLP motifs (P 7 LP 6 ) to * 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.
Expression and Electrophysiology in FRT Epithelia-FRT cells were grown on permeable filter supports as described (22). One day after seeding, cells were cotransfected with ␣-, ␤-, and ␥ENaC (0.07 g each) and hNedd4-2, SGK, or GFP as a negative control (0.02-0.8 g using TFX 50 (22)). The total DNA was held constant by varying the ratio of hNedd4-2 or SGK to GFP. Expression of GFP did not alter ENaC Na ϩ currents.
Na ϩ transport was measured 2-3 days after transfection in modified Ussing chambers (Warner Instrument Corporation). The apical and basolateral surfaces were bathed in 135 mM NaCl, 1.2 mM CaCl 2 , 1.2 mM MgCl 2 , 2.4 mM K 2 HPO 4 , 0.6 mM KH 2 PO 4 , 10 mM dextrose, 10 mM HEPES, pH 7.4, at 37°C and bubbled with O 2 . Amiloride-sensitive short-circuit current was determined as the difference in current with and without amiloride (10 M) in the apical bathing solution.

RESULTS
SGK Binds to hNedd4-2 and hNedd4 -The PY motif (PPXY) is a sequence that mediates protein interactions through its binding to type I WW domains (23). SGK contains a sequence that fits the PY motif consensus (PPFY, amino acids 295-298, Fig. 1A). Members of the Nedd4 family contain multiple WW domains, which bind to PY motifs in ENaC. We tested the hypothesis that their WW domains might also bind to SGK. SGK (containing a FLAG epitope at the C terminus) was expressed in COS-7 cells. We detected SGK protein by Western blot (using anti-FLAG M2 antibody) in cells expressing SGK, but not in cells expressing GFP (Fig. 1B). To test for interactions, we incubated immunoprecipitated SGK with hNedd4-2 or hNedd4 (generated and [ 35 S]methionine-labeled by in vitro translation). Both hNedd4-2 and hNedd4 bound to SGK, but not to immunoprecipitated protein from cells expressing GFP (Fig. 1C). The PY motif of SGK mediated these interactions; mutation of a critical residue within the motif (Y298A) abolished SGK binding to hNedd4-2 and hNedd4 (Fig. 1C). This did not result from decreased protein production; the mutant and wild-type constructs generated similar amounts of SGK protein (Fig. 1B).
To investigate the functional role of this interaction, we asked whether the SGK PY motif was required to stimulate ENaC. Expression of ␣-, ␤-, and ␥ENaC in FRT epithelial cells generated transepithelial short-circuit Na ϩ currents that were blocked by amiloride (22). Coexpression of ENaC with SGK increased Na ϩ current 2.7-fold (compared with ENaC expressed with GFP) (Fig. 1D). In contrast, SGK Y298A did not stimulate ENaC (Fig. 1D), suggesting that the SGK PY motif is required for stimulation.
SGK Phosphorylates hNedd4-2-We tested whether SGK kinase activity is required for it to stimulate ENaC in epithelia. In a previous study, mutation of a residue in the ATP binding site (SGK K127M ) abolished the ability of SGK to phosphorylate a peptide substrate (18). We found that this mutation prevented SGK from stimulating ENaC expressed in FRT cells (Fig. 1D), but it did not alter levels of SGK protein (Fig. 1E). This suggests that SGK stimulates ENaC by phosphorylating one or more substrates.
SGK phosphorylates serine or threonine residues in the context of the sequence RXRXXS/T (18,21). Interestingly, hNedd4-2 contains three sequences that fit this consensus (Ser 221 , Thr 246 , and Ser 327 ), but they are not conserved in hNedd4 ( Fig. 2A). We therefore tested the hypothesis that hNedd4-2 is a substrate for SGK phosphorylation. hNedd4-2, hNedd4, or GFP (negative control) were expressed in COS-7 cells, immunoprecipitated, and incubated with [␥-32 P]ATP with or without an activated form of SGK (⌬1-60, S422D). We found that SGK phosphorylated hNedd4-2 but not hNedd4 (Fig. 2B). As a control for expression, we labeled cells with [ 35 S]methionine and immunoprecipitated with antibodies against the WW domains. Bands of the appro- SGK Modulates Nedd4-2 6 priate size were detected in cells transfected with hNedd4-2 and hNedd4, but not in cells transfected with GFP (Fig. 2C).
SGK Modulates the Function of hNedd4-2-We tested the hypothesis that phosphorylation alters hNedd4-2 function. ENaC was coexpressed with SGK or GFP (negative control) in FRT cells, along with increasing amounts of hNedd4-2 cDNA. When ENaC was expressed with GFP, hNedd4-2 decreased Na ϩ current in a dose-dependent manner (Fig. 3A). In contrast, hNedd4-2 reduced current to a lesser extent when expressed with SGK (Fig. 3A). However, a kinase inactive mutant (SGK K127M ) did not decrease inhibition (Fig. 3A), suggesting that phosphorylation was required. Thus, SGK-mediated phosphorylation modulates hNedd4-2 function, decreasing its ability to inhibit ENaC.
Phosphorylation Alters hNedd4-2 Binding to ENaC-For hNedd4-2 and hNedd4 to inhibit Na ϩ current, their WW domains must bind to PY motifs in ENaC (7)(8)(9). Interestingly, the SGK consensus sites are located between the WW domains of hNedd4-2, suggesting that phosphorylation might alter its binding to ENaC. To test this hypothesis, we immunoprecipitated hNedd4-2 or hNedd4 from COS-7 cells, followed by incubation with one of the ENaC subunits (␣ENaC, translated in vitro and labeled with [ 35 S]methionine). We chose to use the ␣ subunit, since previous work suggested that the three ENaC subunits are equivalent in their binding to WW domains (9,11,12,24). ␣ENaC bound to hNedd4-2 and hNedd4, but not to immunoprecipitated lysates from cells expressing GFP (Fig. 3B). Phosphorylation of hNedd4-2 by SGK decreased the binding of hNedd4-2 to ␣ENaC (Fig. 3B). In contrast, SGK caused minimal change in the binding of hNedd4 to ␣ENaC (Fig. 3B), consistent with our finding that SGK did not phosphorylate hNedd4.
The data suggest that SGK might increase Na ϩ current in part by decreasing the binding of hNedd4-2 to ENaC. Such a mechanism should therefore be disrupted by mutation of the ENaC PY motifs, which are required for this interaction. To test this hypothesis, we expressed ENaC (wild-type or PY motif mutations) with SGK (or GFP as negative control) in FRT cells and measured amiloride-sensitive short-circuit Na ϩ currents. SGK increased Na ϩ current in cells expressing wild-type ENaC (167%) but not when the PY motifs were mutated in ␣-, ␤-, and ␥ENaC (Fig. 3C). Thus, in epithelial cells, stimulation by SGK was dependent on the PY motifs of ENaC.

DISCUSSION
The regulation of ENaC is critically important to maintain Na ϩ homeostasis. Disruption of this regulation causes genetic and acquired forms of hypertension and hypotension (5). Two important regulators of ENaC are aldosterone and its downstream mediator SGK, and the Nedd4 family of ubiquitin protein ligases, which modulate ENaC surface expression in a reciprocal manner. Our data suggest that these two pathways intersect, regulating ENaC in part through a common pathway.
The data support a model in which SGK regulates Na ϩ absorption in part by modulating the inhibition of ENaC by hNedd4-2 (Fig. 4). Under basal conditions (low aldosterone), hNedd4-2 is unphosphorylated and represses epithelial Na ϩ transport; its WW domains bind to ENaC PY motifs, resulting in ubiquitination, endocytosis, and degradation of the channel (Fig. 4, left panel). As a result, hNedd4-2 decreases Na ϩ current by reducing the expression of ENaC at the cell surface. In response to salt deprivation or in pathological states, aldosterone releases this repression of Na ϩ absorption. Aldosterone stimulates the transcription of SGK (16,17), which binds to hNedd4-2 and phosphorylates one or more Ser/Thr consensus sites (Fig. 4, right panel). The binding of SGK to hNedd4-2 appears to be essential, since mutation of the SGK PY motif prevented it from stimulating ENaC. Phosphorylation of hNedd4-2 reduces its binding to ENaC, resulting in increased ENaC at the cell surface, and hence, increased Na ϩ absorption.
How might phosphorylation alter the binding of hNedd4-2 to ENaC? Although we do not yet have direct evidence to support a mechanism, it is interesting to note that the potential phosphorylation sites are located in segments between WW domains. For example, Ser 221 and Thr 246 are located between WW domains 1 and 2, and Ser 327 is located between WW domains 2 and 3 ( Fig. 2A). Perhaps phosphorylation induces a conformational change in hNedd4-2, altering the accessibility of the WW domains to bind ENaC. Interestingly, GenBank TM contains at least three different hNedd4-2 splice forms, differing in the number of potential SGK phosphorylation sites. The splice form we studied (Nedd4La, also known as Nedd18) corresponds most closely to mouse Nedd4-2 and has three potential phosphorylation sites. A second form, KIAA0439, lacks a 20-amino acid segment between WW domains 1 and 2, which deletes one of the SGK sites (Thr 246 ). A third variant, DKFZp234p2422, lacks the second and third SGK sites, as well as WW domain 2. Importantly, each splice form appears to be  3. SGK modulates hNedd4-2 binding and function. A, amiloride-sensitive short-circuit Na ϩ current (relative to 0 g of hNedd4-2) in FRT cells expressing ENaC (0.07 g of each subunit), SGK (wild-type or SGK K127M , 0.6 g), and hNedd4-2 (0 -0.2 g). Total cDNA was held constant using GFP cDNA (mean Ϯ S.E., n ϭ 9 -15). Asterisks indicate p Ͻ 0.02 versus SGK K127M . B, autoradiogram of ␣ENaC (translated and [ 35 S]methionine-labeled in vitro) bound to immunoprecipitated hNedd4-2, hNedd4, or GFP. The hNedd4 proteins and GFP were treated or not treated with SGK (as indicated) prior to binding to ␣ENaC. Data are representative of three experiments. C, percent stimulation of amiloride-sensitive short-circuit Na ϩ current when ␣-, ␤-, and ␥ENaC (0.07 g each) were expressed in FRT cells with SGK (compared with GFP, 0.8 g) (mean Ϯ S.E., n ϭ 20 -22). Asterisk indicates p Ͻ 0.0001 versus wild-type ENaC. The three ENaC subunits were wild-type or contained mutations in the PY motifs, as indicated.
SGK Modulates Nedd4- 2 7 expressed in collecting duct epithelia, 2 and all three inhibit ENaC ( Fig. 3 and Ref. 25). Thus, it seems possible that the splice forms could be differentially phosphorylated, and thus, differentially regulated by SGK. Alternatively, perhaps only the site present in all three forms is phosphorylated (Ser 221 ).
What is the function of the interaction between SGK and hNedd4-2? A recurring theme is that many signaling molecules assemble into complexes to target their activity to specific substrates. Interactions between signaling molecules can occur through adapter proteins, such as protein kinase A anchoring proteins (26). Interactions can also be direct; phosphorylation of the N-methyl-D-aspartic acid receptor by calcium-and calmodulin-dependent protein kinase II is mediated by a direct interaction (27). Thus, perhaps SGK binding targets its kinase activity to hNedd4-2. Binding can also be required to activate a kinase; N-methyl-D-aspartic acid receptor binding activates calcium-and calmodulin-dependent protein kinase II, a mechanism thought to be important for learning and memory (28). Similarly, SGK might be activated through its binding to hNedd4-2.
In addition to hNedd4-2, the function of ENaC may also be regulated by other ubiquitin protein-ligases. Nedd4 is expressed in the renal collecting duct (29) and inhibits ENaC (7,8,30). However, SGK consensus phosphorylation sites are not present in Nedd4, suggesting that SGK does not modulate the activity of this ubiquitin protein-ligase. Consistent with such a hypothesis, we found that SGK did not phosphorylate hNedd4 or alter its binding to ␣ENaC. This raises several possibilities. Perhaps Nedd4 inhibits ENaC in a constitutive manner. More likely, the inhibition of ENaC by Nedd4 might be regulated by different mechanisms. In this regard, it has been reported that cytosolic Ca 2ϩ alters the localization of Nedd4, resulting in its translocation to the cell surface (31). Alternatively, Nedd4 might be phosphorylated by a different kinase or might be transcriptionally regulated. Finally, it is possible that Nedd4 does not regulate Na ϩ transport in epithelia, despite its ability to inhibit ENaC expressed in heterologous cells.
Could SGK modulate ENaC function by additional mechanisms? Several observations raise this possibility. For example, it is possible that the SGK PY motif competes with ENaC for binding to the WW domains of hNedd4-2 (or other ubiquitin protein-ligases). In this way, SGK could increase Na ϩ current by directly blocking the binding of hNedd4-2 to ENaC. However, competition for hNedd4-2 binding does not appear to be the sole mechanism, since SGK kinase activity was also re-quired for ENaC stimulation and for SGK to modulate hNedd4-2 function. In addition, SGK might stimulate ENaC through mechanisms independent of hNedd4-2. In support of this possibility, in Xenopus oocytes, mutation of the ENaC PY motifs did not prevent SGK from increasing Na ϩ current (20), in contrast to our data in epithelia. It was also reported that SGK could bind directly to the C terminus of ␣and ␤ENaC (19). However, SGK did not phosphorylate ENaC, and the functional role and the sequences that mediate this interaction are not yet known.
Our data suggest that aldosterone, SGK, and hNedd4-2 converge into a common pathway to regulate Na ϩ absorption. Within this pathway, hNedd4-2 and related members of the Nedd4 family might form a central point of convergence to regulate ENaC surface expression, and hence, to maintain Na ϩ homeostasis. FIG. 4. Model for modulation of hNedd4-2 by SGK. hNedd4-2 binds and ubiquitinates ENaC, targeting the channel for endocytosis and degradation (left panel). Aldosterone increases transcription of SGK, which binds to hNedd4-2 (via SGK PY motif, right panel). SGK phosphorylates hNedd4-2, which decreases its binding to ENaC. As a result, hNedd4-2-mediated degradation of ENaC is reduced, leading to increased expression of ENaC at the cell surface. SGK Modulates Nedd4-2 8