A Novel Pathway of Epithelial Sodium Channel Activation Involves a Serum- and Glucocorticoid-inducible Kinase Consensus Motif in the C Terminus of the Channel’s (cid:1) -Subunit*

Aldosterone-induced serum- and glucocorticoid-in-ducible kinase isoform 1 (SGK1) contributes to the regulation of the epithelial sodium channel (ENaC), the activity of which is critical for long term blood pressure control. Aldosterone-induced SGK1 is thought to enhance ENaC surface expression by phosphorylating Nedd4-2 and thereby preventing ENaC retrieval and degradation. In outside-out membrane patches of Xenopus laevis oocytes heterologously expressing ENaC, amiloride-sensitive ENaC currents were enhanced by phosphatase inhibitors and were dependent on cytosolic Mg 2 (cid:2) . This indicates that a kinase is involved in channel regulation. Indeed, recombinant constitutively active SGK1, included in the pipette solution, caused a sustained 2- to 3-fold increase of ENaC currents. Deletion of the C terminus of (cid:1) ENaC largely reduced the stimulatory effect of SGK1, whereas stimulation by SGK1 did not require the presence of the C termini of the (cid:3) - or (cid:4) -subunits. Replacing the serine residue Ser 621 of the SGK1 consensus motif in the C terminus of the (cid:1) -subunit by an alanine specifically abolished the stimulatory effect of SGK. Our findings indicate that SGK1 can stimulate of the PPI2 stock solution (50 m M Tris-HCl, 1 m M EDTA, 50% glycerol) were added as vehicle controls to the pipette solution in control experiments. sulfhydryl reagent MTSET ((2-(trimethylammonium)eth- yl)methanethiosulfonate bromide) was obtained from Toronto Research Chemicals (Toronto, Canada). Amiloride hydrochloride was purchased from Sigma-Aldrich (Taufkirchen, Germany).

Aldosterone-induced serum-and glucocorticoid-inducible kinase isoform 1 (SGK1) contributes to the regulation of the epithelial sodium channel (ENaC), the activity of which is critical for long term blood pressure control. Aldosterone-induced SGK1 is thought to enhance ENaC surface expression by phosphorylating Nedd4-2 and thereby preventing ENaC retrieval and degradation. In outside-out membrane patches of Xenopus laevis oocytes heterologously expressing ENaC, amiloride-sensitive ENaC currents were enhanced by phosphatase inhibitors and were dependent on cytosolic Mg 2؉ . This indicates that a kinase is involved in channel regulation. Indeed, recombinant constitutively active SGK1, included in the pipette solution, caused a sustained 2-to 3-fold increase of ENaC currents. Deletion of the C terminus of ␣ENaC largely reduced the stimulatory effect of SGK1, whereas stimulation by SGK1 did not require the presence of the C termini of the ␤or ␥-subunits. Replacing the serine residue Ser 621 of the SGK1 consensus motif in the C terminus of the ␣-subunit by an alanine specifically abolished the stimulatory effect of SGK. Our findings indicate that SGK1 can stimulate ENaC activity independently of an inhibition of Nedd4-2-mediated channel retrieval. This defines a novel regulatory pathway likely to be relevant for aldosterone-induced stimulation of ENaC in vivo.
The appropriate regulation of the epithelial sodium channel (ENaC) 1 in the kidney is critically important for the maintenance of body sodium balance and hence for long term regulation of arterial blood pressure (1). Indeed, two human genetic diseases provide direct evidence that molecular dysfunction of ENaC has severe effects on arterial blood pressure. Loss-offunction mutations of ENaC cause urinary sodium loss, hyperkalemia, and low blood pressure in patients with pseudohy-poaldosteronism type 1 (2). In contrast, gain-of-function mutations of ENaC are found in patients with so-called Liddle's syndrome (pseudohyperaldosteronism) and result in increased renal sodium re-absorption, hypokalemia, and severe arterial hypertension (3).
ENaC is composed of three subunits called ␣, ␤, and ␥ (4). The C termini of the ENaC subunits each contain a proline-rich PPXY (PY) motif, which is believed to be important for interaction with the ubiquitin-protein ligases Nedd4 and Nedd4-2, promoting the ubiquitination, endocytosis, and proteasomal degradation of the channel (5)(6)(7)(8). The functional importance of the PY motif was recognized in Liddle's syndrome where mutations and/or deletions of the PY motif in ␤ or ␥ ENaC reduce the endocytic retrieval of ENaC from the membrane (9,10). This results in an increase in the number of ENaC channels in the membrane, which in turn is thought to cause hyperabsorption of Na ϩ and hypertension in patients with Liddle's syndrome (11,12).
The most important hormone to regulate ENaC activity is the mineralocorticoid aldosterone. The effects of aldosterone include transcriptional, translational, and post-translational modifications of ENaC and involve a complex system of aldosterone-induced and/or aldosterone-repressed regulatory proteins (13). Despite impressive progress in this field of research the molecular mechanisms that mediate the stimulatory effect of aldosterone on ENaC activity remain incompletely understood (14). However, there is a growing body of evidence that the serum-and glucocorticoid-inducible kinase isoform 1 (SGK1) is an important contributing factor in the signal transduction cascade of aldosterone action on epithelial sodium transport (15). In primary cultures of rabbit renal collecting duct cells (16) and in rat distal nephron, aldosterone was shown to stimulate the expression of SGK1 mRNA (17). This effect correlates well with the stimulatory effect of aldosterone on sodium transport and appears to be mediated by the mineralocorticoid receptor (18). Immunohistochemical studies have shown that in the renal collecting duct SGK1 is not expressed in intercalated cells but in principal cells consistent with a specific co-expression of SGK1 and ENaC (19). SGK1 mRNA and protein levels are also increased by aldosterone in distal colon (17,18,20). These findings suggest that SGK1 is a mediator of aldosterone action in classic mineralocorticoid target tissues.
SGK1 was the first aldosterone-induced gene shown to upregulate ENaC-mediated sodium transport. This was initially demonstrated by co-expression experiments in Xenopus laevis oocyte (16,18,21,22) and more recently in cultured renal epithelial cells (23,24). SGK1 is thought to be an important molecular target that integrates multiple endocrine inputs regulating epithelial sodium transport (24,25). The recent development of SGK1 null mice has confirmed that the presence of SGK1 is important for the maintenance of sodium balance, since these animals develop pseudohypoaldosteronism when kept on a low sodium diet (26).
The stimulatory effect of SGK1 occurs in the absence of an increase in ENaC expression levels and appears to be due, at least in part, to increasing the surface expression of ENaC. Recent evidence suggests that the increase in surface expression of ENaC is mediated by PY motif-dependent binding of SGK1 to Nedd4-2 and its subsequent phosphorylation. This has been reported to result in an inhibition of Nedd4-2-mediated ubiquitination, endocytic retrieval, and proteasomal degradation of ENaC, thereby increasing the number of functional ENaC channels in the plasma membrane (27,28). However, there is disagreement as to whether the stimulatory effect of SGK1 on ENaC surface expression is primarily mediated by a decrease in removal from or an increase in translocation to the plasma membrane (23,29). Finally, in addition to its stimulatory effect on ENaC surface expression, SGK1 may also increase ENaC open probability (30).
ENaC phosphorylation by kinases and dephosphorylation by phosphatases (31) has long been thought to contribute to ENaC regulation (1). Aldosterone, insulin, and protein kinases A and C have been shown to increase in vivo phosphorylation of the C termini of the ␤and ␥-subunits of ENaC (32). Moreover, the C termini of ENaC subunits expressed as glutathione S-transferase fusion proteins were found to be phosphorylated by cytosolic fractions derived from rat colon (33). This phosphorylation is thought to involve at least three different types of kinases, including the extracellular-regulated kinase (34) and casein kinase 2 (35).
In the present report we used outside-out membrane patch recordings to functionally confirm the involvement of kinases and phosphatases in the regulation of ENaC heterologously expressed in X. laevis oocytes. More specifically, we demonstrated a stimulatory effect of recombinant and constitutively active SGK1 on ENaC currents and identified the importance of an SGK consensus motif in the ␣-subunit for mediating this effect. These results suggest that SGK1 can directly stimulate ENaC activity independently of its effects on Nedd4-2-mediated channel retrieval.

EXPERIMENTAL PROCEDURES
Molecular Biology-The full-length cDNAs encoding the three subunits of wild-type (wt) rat ENaC (␣-, ␤-, and ␥-ENaC) (4) were in pGEM-HE. Those encoding the truncated rENaC subunits ␣ P646stop , ␤ R564stop , and ␥ F606stop (36) and the mutant subunit ␤ S518C (37) were in pSD5 and were a gift of Drs. Bernard C. Rossier and Laurent Schild (Lausanne, Switzerland). Linearized plasmids were used as templates for cRNA synthesis using either T7 (wt ␣␤␥-ENaC and truncated ␥-ENaC) or SP6 (truncated ␣␤-ENaC) RNA polymerases (mMessage mMachine, Ambion, Austin, TX). To replace the serine Ser 621 in the SGK consensus motif 616 RSRYWS 621 of rat ␣ENaC by an alanine (␣ S621A -ENaC) or by aspartate (␣ S621D -ENaC), site-directed mutagenesis extension overlap PCR was performed using T7 and SP6 as flanking primers. To generate ␣ S621A -ENaC, a mutagenic forward primer with the sequence 5Ј-CGG AGC CGG TAC TGG GCC CCA GGA CGA GGG GCC-3Ј and a reverse primer with the sequence 5Ј-GGC CCC TCG TCC TGG GGC CCA GTA CCG GCT CCG-3Ј were used to introduce a triplet mutation from TCT at nucleotides 1861-1863 to GCC. To generate ␣ S621D -ENaC a mutagenic forward primer with the sequence 5Ј-CGG AGC CGG TAC TGG GAC CCA GGA CGA GGG GCC-3Ј and a reverse primer with the sequence 5Ј-GGC CCC TCG TCC TGG GTC CCA GTA CCG GCT CCG-3Ј were used to introduce a triplet mutation from TCT at nucleotides 1861-1863 to GAC. Mutations were confirmed by sequence analysis.
Outside-out Macropatch Recordings-After a brief (1-2 min) exposure to hypertonic NMDG-Cl bath solution (in mM: 95 NMDG-Cl, 2 KCl, 1 MgCl 2 , 1 CaCl 2 , 200 sucrose, 10 HEPES adjusted to pH 7.4 with Tris) oocytes were stripped of the vitellin membrane using sharpened forceps and transferred to a bath chamber on a Leica DM IRB inverted microscope (Leitz Microsystems UK Ltd., Milton Keynes, UK). A computercontrolled EPC-9 patch clamp amplifier (HEKA Elektronik, Lambrecht, Germany) was used as described previously (38) to perform conventional outside-out membrane patch recordings (40). It was usually possible to obtain a second patch from the same oocyte for matched control experiments. Alternatively, a patch from another oocyte of the same batch served as control. For each experimental series oocytes from at least three different batches were used. Patch pipettes were pulled from borosilicate glass capillaries (catalog no. 1155150, inner diameter 0.87 mm, outer diameter 1.5 mm, Hilgenberg, Masfeld, Germany) using a DMZ-Universal puller (Zeitz Instrumente, Munich, Germany) and had a tip diameter of about 5-7 m after fire polishing to obtain macropatches. Unless stated otherwise pipettes were filled with potassium gluconate pipette solution (in mM: 90 potassium gluconate, 5 NaCl, 2 Mg-ATP, 2 EGTA, and 10 mM HEPES adjusted to pH 7.4 with Tris). A Na 2ϩ -free NMDG-Cl bath solution (in mM: 95 NMDG-Cl, 2 KCl, 1 MgCl 2 , 1 CaCl 2 , 10 HEPES adjusted to pH 7.4 with Tris) was the standard bath solution at the beginning of each experiment. In NaCl bath solution, NMDG-Cl was replaced by 95 mM NaCl. Downward current deflections correspond to cell membrane inward currents, i.e. movement of positive charge from the extracellular side to the cytoplasmic side. Outside-out patches were routinely held at a holding pipette potential of Ϫ70 mV, which was close to the reversal potential of Cl Ϫ (E Cl ϭ Ϫ77.2 mV) and K ϩ (E K ϭ Ϫ79.4 mV) under our experimental conditions. I-V plots were obtained from voltage step protocols, and ⌬I Ami values were obtained by subtracting the currents in the presence of amiloride (2 M) from the corresponding currents prior to addition of amiloride.
Single Channel Recordings in Outside-out Patches-Single channel recordings were performed essentially as described previously (38,41). Bath and pipette solutions were identical to those used for recordings in outside-out macropatches. The pipette had a tip diameter of about 1 m after fire polishing. Single channel current data were initially filtered at 250 Hz and sampled at 1 kHz. Data were analyzed using the program "Patch for Windows" written by Dr. Bernd Letz (HEKA Elektronik, Lambrecht/Pfalz, Germany). Using channel traces re-filtered at 15 Hz, channel activity was estimated from binned amplitude histograms as the product NP o , where N is the number of channels and P o is single channel open probability. The program for calculating NP o from integration of the areas under the peaks of amplitude histograms uses the following equation (42): NP o ϭ ⌺(n j ⅐ ⌬I j )/(i ⅐ ⌺n j ), where i refers to the mean single-channel current and where j refers to the j th current amplitude bin, and j ranges from 1 to the total number of bins; n j is the number of events within bin j; ⌬I j ϭ I j Ϫ I c , where I j is the current of bin j, and I c is the current at which all channels are closed. I c was determined in the presence of 2 M amiloride. Single channel P o was estimated by dividing NP o by the maximal apparent number of channel levels in the current trace analyzed.
Solutions and Chemicals-Recombinant constitutively active SGK1(⌬1-60, S422D) was purchased from Biomol GmbH (Hamburg, Germany) as 2-g vials in 50 l of stock solution containing as main components 50 mM Tris-HCl, 0.1 mM EGTA, 0.1% 2-mercaptoethanol, 0.15 mM NaCl, and 270 mM sucrose. SGK1 pipette solution was freshly prepared on the day of the experiment by adding 2 l of the SGK1 stock solution to 1 ml of the potassium-gluconate pipette solution giving a final SGK1 concentration of 80 ng/ml. To preserve SGK1 activity, the pipette solution was supplemented with dithiothreitol (1 mM). For control experiments the pipette solution also contained dithiothreitol (1 mM) and a vehicle control, including the main components of the SGK1 stock solution as indicated above. In addition control experiments were performed using heat-inactivated SGK1 stock solution, which had been incubated at 68°C for 45 min. Okadaic acid was purchased from Sigma-Aldrich (Taufkirchen, Germany) and was dissolved in Me 2 SO as a stock solution with a concentration of 0.1 mM. Recombinant protein phosphatase inhibitor type 2 (PPI2) was obtained from Merck Biosciences GmbH (Schwalbach, Germany) as a stock solution containing 1 mg/ml PPI2. Okadaic acid and PPI2 were added to the pipette solution to give final concentrations of 100 nM and 1 g/ml, respectively. Appropriate amounts of Me 2 SO or of a buffer corresponding to the main components of the PPI2 stock solution (50 mM Tris-HCl, 1 mM EDTA, 50% glycerol) were added as vehicle controls to the pipette solution in control experiments. The sulfhydryl reagent MTSET ((2-(trimethylammonium)ethyl)methanethiosulfonate bromide) was obtained from Toronto Research Chemicals (Toronto, Canada). Amiloride hydrochloride was purchased from Sigma-Aldrich (Taufkirchen, Germany).

ENaC Currents Can Be Recorded in Outside-out Macropatches from X. laevis Oocytes Heterologously Expressing
ENaC-In oocytes injected with ENaC cRNA, channel expression was routinely confirmed by using two-electrode voltage clamp measurements, which revealed amiloride (2 M)-sensitive currents (⌬I Ami ) averaging 10.3 Ϯ 1.1 A at a holding potential of Ϫ60 mV (n ϭ 43 in 14 batches of oocytes). From these oocytes it was possible to obtain stable outside-out macropatches in about 50% of attempts. Fig. 1A shows a typical current recording starting about 5 min after patch excision. In the absence of extracellular Na ϩ in NMDG-Cl bath solution, only a minor inward current component was initially observed. Changing to NaCl bath solution resulted in an immediate increase of the inward current consistent with the occurrence of a current component carried by Na ϩ influx. After an initial transient peak this Na ϩ current component relaxed to a slightly lower quasi-steady-state current. The current peak with subsequent relaxation is a well known phenomenon and is most likely due to so-called Na ϩ self-inhibition by extracellular Na ϩ thought to reduce the open probability of ENaC through interaction with an extracellular Na ϩ binding site (43). As shown in the current trace in Fig. 1A, application of amiloride, in a concentration of 2 M known to specifically inhibit ENaC, instantaneously inhibited the Na ϩ inward current component and the effect was readily reversible upon washout of amiloride. In similar experiments as the one shown in Fig. 1A, ⌬I Ami averaged 364 Ϯ 88 pA (n ϭ 22 in 9 batches of oocytes) at a holding potential of Ϫ80 mV. Amiloride-sensitive currents could be recorded in all successful outside-out patches from ENaC-expressing oocytes. Moreover, in about 60% of cases it was possible to obtain a second patch with amiloride-sensitive currents from the same oocyte to serve as a direct control experiment. We never detected amiloride (2 M)-sensitive currents in outside-out patches of non-injected control oocytes (n ϭ 12). Fig. 1B shows I/V plots obtained from voltage step protocols performed in the absence and presence of amiloride during the experiment shown in Fig. 1A. A subtracted I/V plot representing the average ⌬I Ami values of similar experiments is shown in Fig. 1C. A Goldman-Hodgkin-Katz (GHK) fit of these data reveals that ⌬I Ami is highly Na ϩ selective as expected for a current mediated by ENaC.
ENaC Activity Is Stimulated by Phosphatase Inhibitors and Is Dependent on the Presence of Cytosolic Magnesium-As shown in Fig. 2, we were able to obtain continuous current recordings from outside-out macropatches that were stable for at least 20 -30 min. This enabled us to continuously monitor ENaC activity in these patches by repeatedly measuring ⌬I Ami . As illustrated in the representative current trace shown in Fig.  2A, ⌬I Ami remained largely unchanged throughout the experiment. Maintenance of channel activity may require a balance between channel phosphorylation and dephosphorylation mediated by kinases and phosphatases, respectively, which may be associated with ENaC outside-out macropatches. We therefore tested the effect of phosphatase inhibitors that were included in the pipette solution. ENaC currents increased over time when okadaic acid (100 nM), a known inhibitor of a range of phosphatases, was included in the pipette solution. Simi- larly, inclusion of PIP2 (1 g/ml), a more specific inhibitor of phosphatase 1, had a stimulatory effect on ENaC currents as shown in the representative current recording in Fig. 2B. On average, both okadaic acid and PIP2 stimulated ENaC currents by 2-to 3-fold with a similar time course (Fig. 2D). These findings indicate that continuous phosphorylation and dephosphorylation reactions occur in the outside-out macropatches and that the phosphatase inhibitors shift the equilibrium toward phosphorylation resulting in enhanced channel activity. To confirm the kinase dependence of ENaC channel activity we performed additional experiments in which Mg 2ϩ was omitted from the pipette solution, which in addition contained the divalent cation-chelating substance EDTA (10 mM). As shown in the representative current trace in Fig. 2C, ENaC currents decreased to a very low level within 10 -15 min under these conditions. Data from similar experiments are summarized in Fig. 2D and indicate that ENaC activity is dependent on the presence of cytosolic Mg 2ϩ . This is likely to be due to an Mg 2ϩdependent kinase that appears to be necessary for maintaining ENaC channel activity.
Recombinant SGK1 Stimulates ENaC Currents in Outsideout Patches-To test whether SGK1 can directly affect ENaC activity in outside-out patches, constitutively active recombinant SGK1 (80 ng/ml) was included in the pipette solution. As shown in Fig. 3B, this resulted in a substantial increase of ⌬I Ami over time reaching a steady-state level that was on average about 2-to 3-fold higher than that of the initial ⌬I Ami (Fig. 3C). In contrast, including heat-inactivated recombinant SGK1 (n ϭ 8, Fig. 3A) or a vehicle control for the SGK1 buffer (n ϭ 22, Fig. 3C) did not stimulate ENaC currents. Data are summarized in Fig. 3C and indicate that recombinant SGK1 applied from the cytosolic side leads to a sustained stimulation of ENaC currents. The average I/V plots shown in Fig. 3D were derived from the set of experiments shown in Fig. 3C with SGK1 in the pipette solution. The I/V plots were constructed using ⌬I Ami values that were measured 5 and 28 min after patch excision. They demonstrate that stimulation of the currents by SGK1 does not shift the reversal potential, which indicates that the stimulated currents remain highly Na ϩselective. Thus, stimulation by SGK1 does not alter the selectivity of ENaC.
The Effect of SGK1 Is Preserved in the Presence of Colchicine and of Low Cytosolic Ca 2ϩ -An SGK1-mediated increase in ENaC currents may be due to an increase in channel open probability or an increase in the number of channels in the plasma membrane. In outside-out macropatches some cytoskeletal elements may remain attached to the excised plasma membrane. Thus, some of the machinery required for ion channel trafficking may be functional even in excised patches. Channel trafficking to the plasma membrane is likely to involve the microtubule system, whereas the final fusion step of sub-membranous vesicles containing ion channels with the plasma membrane is likely to be calcium-dependent. To test a possible involvement of the microtubule system, we investigated the effect of SGK1 on ENaC currents in the presence of 20 M colchicine, which was included in the pipette solution and has previously been used in similar concentrations to inhibit microtubule-dependent recycling (44). As shown in Fig. 4A, colchicine had no effect on ENaC currents in control experiments and did not prevent the stimulatory effect of SGK1, which was fully preserved in the presence of colchicine. To investigate the calcium dependence of ENaC stimulation by SGK1, we tested the effect of 10 mM EGTA included in a nominally calcium-free pipette solution, which should maximally reduce the cytosolic and sub-membranous free Ca 2ϩ concentration without having a major effect on the free Mg 2ϩ concentration. In these exper- iments we also used a nominally calcium-free bath solution containing 1 mM EGTA to prevent Ca 2ϩ influx from the extracellular solution. As shown in Fig. 4B this resulted in a continuous rundown of ⌬I Ami in control experiments to less than 50% of the initial current value indicating that a minimal cytosolic Ca 2ϩ concentration is probably required to maintain stable ENaC currents. Importantly, even in the presence of 10 mM EGTA in the pipette solution and in the absence of extracellular Ca 2ϩ a substantial stimulatory effect of SGK1 was preserved (Fig. 4B), indicating that Ca 2ϩ -dependent mechanisms are unlikely to be essential for mediating the effect of SGK1. Taken together these findings suggest that the stimulatory effect of SGK1 observed in outside-out patches is not due to an increase in microtubule-mediated or calcium-dependent channel insertion into the plasma membrane.
Single Channel Recordings Demonstrate That SGK1 Increases the Number of Apparent Channels in the Patch-If the stimulatory effect of SGK1 is not mediated by insertion of new channels in the plasma membrane, it is likely to be mediated by an effect of SGK1 on channel open probability (P o ). In this case SGK1 may either uniformly stimulate P o of all ENaC channels present in the plasma membrane or, alternatively, may activate silent channels that are thought to be present in the plasma membrane of ENaC-expressing oocytes (11). To investigate this issue we performed single channel recordings in outside-out patches in the presence or absence of SGK1 in the pipette solution. We were able to detect amiloride-sensitive single channel current events in outside-out patches with a single channel current amplitude of 0.67 Ϯ 0.03 pA (n ϭ 11) at a holding potential of Ϫ100 mV. This corresponds to a single channel conductance of 6.7 Ϯ 0.3 pS (n ϭ 11) typical for ENaC. Importantly, we observed a gradual increase of the number of apparent single channel current levels in experiments in which SGK1 was included in the pipette solution (n ϭ 4). This is illustrated by the current traces shown in Fig. 5, which were recorded in the same patch at various times after patch excision. In the particular experiment shown in Fig. 5, we observed maximally two channel levels at the beginning of the experiment. Amplitude histograms revealed NP o values of 1.13 for the current trace starting 4 min after excision and of 0.95 for the current trace starting 8 min after excision, corresponding to a P o of 0.57 and 0.48, respectively, assuming that only two channels were active in the patch during this period. These estimates for P o are in good agreement with the average P o of 0.5 reported for ENaC in the native collecting duct (45). In the traces starting 12 and 16 min after patch excision, additional channel levels appeared and NP o increased to 3.84 and 4.23, respectively. Dividing these NP o values by the maximal number of channels observed during the corresponding recording period we can estimate P o values of 0.64 and 0.53, respectively. Importantly, in matched control experiments without SGK1 in the pipette solution stable single channel currents or continuous channel rundown was observed (n ϭ 6). Taken together, these findings indicate that SGK1 does not uniformly increase the P o of ENaC channels that are already active in the plasma membrane. Indeed, the gradual appearance of additional channel levels suggests that SGK1 may convert silent ENaC channels resident in the plasma membrane (11) to active channels with a single channel P o of about 0.5-0.6. Although this is a plausible interpretation, it has to be noted that our single channel recordings cannot distinguish between the possibility of an activation of silent channels resident in the plasma membrane and the insertion of new channels into the plasma membrane.
The Stimulatory Effect of SGK1 Is Not Due to a Uniform Increase in ENaC Open Probability-To confirm the finding of our single channel data that SGK1 increases the apparent number of ENaC channels in the patch without having a major effect on the single channel open probability, we performed additional experiments using the S518C mutant of the ␤-subunit of rat ENaC (␤ S518C ). This mutant channel can be converted from a channel with a normal open probability to a channel with an open probability of nearly one by exposing the channel to the sulfhydryl reagent MTSET, which destabilizes the channel's closed state (37).
In outside-out patch recordings from oocytes expressing the mutant channel application of MTSET increased ENaC currents as expected by a factor of about two consistent with an increase in P o from about 0.5 to 1 (Fig. 6, A and C). We also demonstrated that the mutant channel was stimulated by recombinant SGK1 to a similar degree as the wild-type channel (Fig. 6B). Importantly, after stimulation of ENaC by SGK1 the subsequent exposure to MTSET resulted in an additional increase of ENaC currents by a factor of about two (Fig. 6, B and  C). This demonstrates that MTSET also increases the open probability of ENaC channels from about 0.5 to 1 after they have been activated by SGK1. This finding is nicely consistent with our single channel data demonstrating that SGK1 does not uniformly increase the open probability of ENaC channels but increases the number of apparent ENaC channels in the patch.
The effect of MTSET on the ␤ S518C mutant ENaC has been FIG. 5. SGK1 increases the number of apparent ENaC channels. ENaC single channel currents were recorded at a holding potential of Ϫ100 mV in an outside-out patch with SGK1 in the pipette solution. Four sections of a continuous current recording are shown with traces starting 4, 8, 12, and 16 min after patch excision (from top to bottom). To confirm that the single channel activity was due to ENaC channels and to determine the current level at which all channels are closed, amiloride (Ami; 2 M) was applied as indicated by the black bars. For the periods recorded in the absence of amiloride channel activity (NP o ), the number of apparent channel levels were estimated using binned amplitude histograms as shown on the right side of each trace.
reported to be state-dependent, and the modification of the channel by MTSET is thought to occur only when the channel is in an open state (37). Thus, MTSET may not affect silent channels resident in the plasma membrane. Therefore, our findings are consistent with the interpretation that SGK1 increases the number of apparent ENaC channels by converting silent channels into active channels, which then become accessible to MTSET. Alternatively, SGK1 may stimulate insertion of ENaC channels with a normal P o that subsequently are converted into channels with a P o of about one by exposure to MTSET.
Deletion of the C Termini of All Three ENaC Subunits Abolishes the Stimulatory Effect of SGK1 on ENaC Currents-ENaC stimulation by SGK1 may be due to an inhibition of ENaC channel retrieval from the plasma membrane. Indeed, SGK1 has been reported to stimulate ENaC by phosphorylating Nedd4-2 thereby preventing Nedd4-2-dependent channel ubiquitination, retrieval, and proteasomal degradation (27,28). Because the Nedd4-2-mediated inhibition of ENaC requires the presence of PY motifs in the C termini of ENaC, we tested the effect of SGK1 on mutated ENaC channels with C-terminal deletions in all three subunits. The sites of the C-terminal truncations were analogous to the original Liddle's syndrome mutation of the ␤-subunit and resulted in the loss of all PY motifs (36). As shown in Fig. 7, deletion of the C termini of ␣␤␥ENaC largely abolished the stimulatory effect of SGK1 on ENaC currents. This finding indicates that the C termini of one or more of the ENaC subunits are important in mediating the stimulatory effect of SGK1.
The C Terminus of the ␣-Subunit Is Critical for the Stimulatory Effect of SGK1-To determine the relative importance of the C termini of the different subunits for the stimulatory SGK1 effect, we performed an additional series of experiments in which we tested the effects of C-terminal truncations of individual ENaC subunits. As shown in Fig. 8 the stimulatory effect of SGK1 was largely preserved when the C termini of ␤ENaC or of ␥ENaC were deleted (Fig. 8, B and C). However, in ENaC channels with a C-terminally truncated ␣-subunit the stimulatory effect of SGK1 was largely abolished. In humans C-terminal truncations of either the ␤or the ␥-subunit are known to cause Liddle's syndrome. Because Liddle's syndrome is thought to occur because the mutations reduce the PY motif-dependent Nedd4-2-mediated retrieval of ENaC, our finding of a preserved stimulatory SGK1 effect in ENaC channels with C-terminally truncated ␤and ␥-subunits indicates that in outside-out macropatches ENaC stimulation by SGK1 is not due to an inhibition of Nedd4-2-mediated channel retrieval.
The C-terminal Sgk1 Consensus Motif in the ␣-Subunit Is Required for ENaC Stimulation-Our findings suggest that the stimulatory effect of SGK1 may be mediated by a direct effect of SGK1 on ENaC. Sequence analysis of the ENaC subunits indicated the presence of two SGK consensus motifs corresponding to a sequence of RXRXX(S/T) (Fig. 9) (27,46). One consensus motif (RKRKIS) is localized in the extracellular loop of the ␥-subunit at the amino acid position 178 -183 shortly after the first transmembrane domain. This motif is unlikely to be involved in mediating the SGK1 effect, because SGK1 acts from the cytosol. The other consensus motif is localized in the C terminus of the rat ␣-subunit at amino acid position 616 -621 just after the second transmembrane domain and is well conserved in mammals. To test whether this SGK1 consensus motif is involved in mediating the effect of SGK1, we mutated its serine to an alanine changing the motif from 616 RSRYWS 621 to 616 RSRYWA 621 . Two-electrode voltage clamp experiments confirmed that the mutated channel was functional and that its expression resulted in amiloridesensitive whole cells currents that averaged 11.1 Ϯ 1.0 A (n ϭ 8) and were well within the range of currents observed with wild-type ENaC. However, as shown in Fig. 10, in outside-out patches the stimulatory effect of SGK1 was completely absent in ENaC channels with a mutated SGK1 consensus site in the ␣-subunit. Similarly, SGK1 failed to stimulate ENaC currents in outside-out patches when the serine residue 621 of the ␣-subunit was mutated to an aspartate (␣ S621D -ENaC; n ϭ 10). These findings indicate that in outside-out macropatches the stimulatory effect of SGK1 on ENaC currents is critically dependent on the presence of the SGK1 consensus site in the C terminus of ␣ENaC.

DISCUSSION
The key findings of the present study are the following: (i) ENaC channel activity in outside-out macropatches is enhanced by phosphatase inhibitors and is dependent on the presence of cytosolic Mg 2ϩ indicating the involvement of a kinase in channel regulation, (ii) recombinant constitutively active SGK1 included in the pipette solution causes a sustained increase of the ENaC current to a level about 2-to 3-fold that of its initial value, (iii) deletion of the C terminus of the ␣-subunit of ENaC largely reduces the stimulatory effect of SGK1, whereas ENaC stimulation by SGK1 does not require the presence of the C termini of the ␤or ␥subunit, and (iv) replacing the serine of the SGK consensus motif in the C terminus of the ␣-subunit by an alanine specifically abolishes the stimulatory effect of SGK1.
These findings define a novel pathway by which SGK1 acti-FIG. 8. The C terminus of ␣ENaC is critical for the stimulatory effect of SGK1. The effect of SGK1 included in the pipette solution was tested in outside-out patches from oocytes expressing a C-terminally truncated ␣-subunit and wild-type ␤␥-ENaC (␣ T -ENaC) (with SGK1, n ϭ 8; without SGK1, n ϭ 6) (A), a C-terminally truncated ␤-subunit and wild-type ␣␥-ENaC (␤ T -ENaC) (with and without SGK1, n ϭ 7) (B), and a C-terminally truncated ␥-subunit and wild-type ␣␤-ENaC (␥ T -ENaC) (with SGK1, n ϭ 5; without SGK1, n ϭ 4) (C). vates ENaC. This pathway is likely to contribute to the aldosterone-induced stimulation of ENaC in vivo. So far it has been thought that the main effect of aldosterone-induced SGK1 is mediated by its inhibitory action on Nedd4-2. Indeed, phosphorylation of Nedd4-2 by SGK1 has been reported to reduce its ability to interact with ENaC thereby preventing ubiquitination, retrieval, and proteasomal degradation of the channel (27,28). Furthermore, SGK1 has been shown to increase surface expression of ENaC, which is consistent with the concept that it reduces the rate of ENaC retrieval by inhibiting the interaction of Nedd4-2 and ENaC. However, it has been demonstrated that the stimulatory effect of SGK1 on ENaC whole cell currents and surface expression was preserved in oocytes expressing mutated ENaC channels consisting of ␣␤␥-subunits that were C-terminally truncated (29). Under these conditions all PY motifs of the channel were missing preventing an interaction of ENaC with the WW domains of Nedd4-2. Thus, in these experiments the stimulatory effect of SGK1 could not be due to an inhibition of Nedd4-2-mediated channel retrieval. Moreover, in the same study it was demonstrated that the rate of channel endocytosis was similar in oocytes co-expressing ENaC and SGK1 compared with control oocytes expressing ENaC alone. Recently, another group has reported that a point mutation (Y618A) in the ␤-subunit of ENaC, which is known to impair regulation of ENaC by Nedd4, had no significant effect on the stimulatory effect of SGK on ENaC currents (18). Taken together, these findings argue against an exclusive effect of SGK1 on the rate of Nedd4-2-mediated channel retrieval. Indeed, it has been proposed that the stimulatory effect of SGK1 on ENaC currents and channel surface expression is probably due to an increased channel insertion rate rather than to an inhibition of channel retrieval (23,29).
Our experiments using outside-out patch recordings cannot resolve the question of whether SGK1 enhances channel insertion or reduces channel retrieval in the intact oocyte, because these processes are unlikely to be fully operational in the outside-out patches used in our study. Interfering with the microtubular system by using colchicine or preventing calcium-mediated vesicle insertion by using high concentrations of calcium buffer in the pipette solution did not prevent the stimulatory effect of SGK1 in outside-out macropatches. This suggests that, in outside-out macropatches, channel trafficking and insertion of additional channels into the plasma membrane are unlikely to contribute to the SGK1-mediated increase in ENaC currents. On the other hand inhibition of Nedd4-2-mediated channel retrieval is also unlikely to contribute to the effect of SGK1 in our experiments for the following reasons. First, channel retrieval from the plasma membrane and subsequent intracellular processing also involve the microtubule system, but colchicine failed to affect ENaC stimulation by SGK1. Second, although Na ϩ feedback inhibition is thought to be mediated by an activation of Nedd4-2-mediated ENaC retrieval (47), we can assume that under our experimental conditions the Nedd4-2 pathway is largely suppressed due to the low cytosolic Na ϩ concentration (5 mM) of the pipette solution. Third, the finding that SGK1 stimulation of ENaC was not dependent on the presence of the C terminus of the ␤or ␥-subunit also argues against an involvement of the Nedd4-2 pathway, because Nedd4-2 is thought to interact with the PY motifs of these subunits. Thus, the stimulatory effect of SGK1 in outside-out membrane patches is probably not due to an inhibition of Nedd4-2-mediated channel retrieval.
Our single channel recordings suggest that SGK1 recruits silent ENaC channels resident in the plasma membrane (11) and/or promotes channel insertion rather than uniformly increasing single channel P o of ENaC channels that are already active. Our experiments, using MTSET as a tool to increase the open probability of mutant ␤ S518C ENaC channels (37), demonstrated that the stimulatory effect of MTSET is preserved in SGK1-treated patches and was similar to the effect in nontreated patches. This is consistent with the interpretation that SGK1 increases the number of active channels in the patch without having a major effect on the open probability of the individual channels. Taken together our data suggest that in addition to increasing channel surface expression (29) SGK1 increases the number of active channels in the plasma membrane possibly by converting silent channels into an active state. This interpretation is consistent with a recent report that coexpression of SGK1 increased ENaC whole cell currents 2.5-fold but channel surface expression only 1.6-fold, indicating a dual effect on the number of channels present in the membrane and on their activity (30). This concept of a dual effect of SGK1 is also supported by our finding that a C-terminal truncation of ␣ENaC largely inhibits the stimulatory effect of SGK1 in outside-out membrane patches probably by impeding the SGK1-induced conversion of silent into active channels, whereas in the intact oocyte the C-terminal truncation of all three subunits does not prevent a stimulatory effect of SGK1 on ENaC surface expression (29).
How then does SGK1 affect ENaC function in the plasma membrane of outside-out patches? Our findings clearly demonstrate the functional importance of the C terminus of the ␣-subunit and more specifically of the SGK consensus motif. In this context it should be pointed out that the C-terminal truncation of the ␣-subunit used in our study did not ablate the SGK consensus motif, which is found in the initial portion of the C terminus between the second transmembrane domain and the site at which the C terminus was truncated. Interestingly, as shown in Fig. 8A, the C-terminal truncation of the ␣-subunit did not completely abolish the stimulatory effect of SGK1 on ENaC currents. This may be due to the fact that the SGK motif remains intact and still allows a partial, albeit very small, activation of the channel. Thus, the presence of the SGK1 motif in the ␣-subunit is not sufficient for ENaC activation but in addition requires an intact C terminus. In contrast, direct mutation of the SGK consensus motif completely prevented SGK1-mediated ENaC activation even in the presence of an intact C terminus. Taken together these findings indicate that, although the consensus motif is the critical site for the stimulatory effect of SGK1 on ENaC, the more distal C terminus, including the PY motif, is also functionally important for ENaC activation. It is tempting to speculate that phosphorylation of the serine residue within the SGK consensus motif results in a conformational change of the more distal C terminus leading to an activation of ENaC.
As stated in the introduction, phosphorylation and dephosphorylation of ENaC seem to play a major role in the regulation of channel activity. Our finding that phosphatase inhibitors increase ENaC currents in outside-out patches of oocytes is in good agreement with a recent report of the stimulatory effect of okadaic acid on ENaC-mediated transepithelial sodium transport in A6 cells (31). The Mg 2ϩ dependence of ENaC activity further confirms the presence of a kinase in the outside-out macropatch. Taken together, our findings demonstrate that a kinase-mediated phosphorylation step is important for maintaining or stimulating ENaC activity. ENaC phosphorylation in Madin-Darby canine kidney cells, stably transfected with the three subunits of ENaC, has been shown to involve the C termini of the ␤and ␥-subunits, whereas no phosphorylation of the ␣-subunits was found (32). In A6 cells a baseline phosphorylation of the ␣and ␤-subunits of ENaC was detected, whereas the ␥-subunit was found to be either weakly phosphorylated or not phosphorylated (48). More recently, in vitro studies have confirmed the phosphorylation of certain threo-nine and serine residues of the ␤and ␥-subunits, whereas no significant phosphorylation was found in the C terminus of the ␣-subunit (33)(34)(35). In this context it is interesting to note that SGK1 has been reported to physically interact with the Cterminal tails of ␣-ENaC and ␤-ENaC in vitro. However, there was no evidence that this association resulted in ENaC phosphorylation (25). Thus, as far as we know, specific phosphorylation of the serine residue 621 within the SGK consensus motif of ␣-ENaC has not yet been reported. Nevertheless, according to the functional analysis of our present study it is likely but not yet proven that phosphorylation of this serine residue plays an important role in the process of SGK1-mediated ENaC activation. It is not yet clear whether SGK1 directly mediates the phosphorylation of this residue or whether this requires an additional not yet identified SGK1-dependent kinase. Indeed, the SGK consensus motif is not uniquely specific for SGK1 and may also be phosphorylated by other kinases like, e.g., protein kinase B known to mediate many of the metabolic actions of insulin (46). Thus, the emerging picture of ENaC regulation by various phosphorylation and dephosphorylation events may turn out to be highly complex, involving a network of interacting kinase cascades. Indeed, following induction of its expression by aldosterone, SGK1 itself requires activation by a pathway that involves the phosphatidylinositol 3-kinase and the 3-phophoinositide-dependent-kinases 1 and 2 (46,49). Protein kinase A has also been reported to activate SGK, which may contribute to the stimulatory effect of vasopressin on Na ϩ transport (50).
In conclusion our study provides evidence for a novel pathway of SGK1 stimulation of ENaC involving the C terminus of the ␣-subunit and, more specifically, a serine residue in its SGK consensus motif close to the second transmembrane domain. Thus, in addition to the known stimulatory effect of SGK1 on ENaC surface expression, which may be due to an inhibition of Nedd4-2-dependent channel retrieval or to a stimulation of channel insertion, this direct pathway is likely to contribute to aldosterone-induced activation of ENaC channels that are already present in the plasma membrane. At present we cannot rule out the possibility that this pathway may also contribute to channel insertion. Interestingly, it has recently been demonstrated that the stimulatory effect of mineralocorticoids is preserved in a mouse model for Liddle's syndrome (51) in which the traditional Nedd4-2-mediated stimulatory pathway of SGK1 should be compromised due to the deletion of the PY motif of ␤ENaC. Under these conditions aldosterone-induced SGK1 may still enhance ENaC activity by the additional pathway involving the SGK consensus motif in the C terminus of ␣ENaC. These findings suggest that the novel stimulatory pathway defined in this study is likely to be relevant for ENaC stimulation by SGK1 in vivo.