Phosphatidylinositol 3,4,5-Trisphosphate Mediates Aldosterone Stimulation of Epithelial Sodium Channel (ENaC) and Interacts with γ-ENaC*

Whole cell voltage clamp experiments were performed in a mouse cortical collecting duct principal cell line using patch pipettes back-filled with a solution containing phosphatidylinositol 3,4,5-trisphosphate (PIP3). PIP3 significantly increased amiloridesensitive current in control cells but not in the cells prestimulated by aldosterone. Additionally, aldosterone stimulated amiloridesensitive current in control cells, but not in the cells that expressed a PIP3-binding protein (Grp1-PH), which sequestered intracellular PIP3. 12 amino acids from the N-terminal tail (APGEKIKAKIKK) of γ-epithelial sodium channel (γ-ENaC) were truncated by PCRbased mutagenesis (γT-ENaC). Whole cell and confocal microscopy experiments were conducted in Madin-Darby canine kidney cells co-expressing α- and β-ENaC only or with either γ-ENaC or γT-ENaC. The data demonstrated that the N-terminal tail truncation significantly decreased amiloride-sensitive current and that both the N-terminal tail truncation and LY-294002 (a PI3K inhibitor) prevented ENaC translocation to the plasmamembrane. These data suggest that PIP3 mediates aldosterone-induced ENaC activity and trafficking and that the N-terminal tail of γ-ENaC is necessary for channel trafficking, probably channel gating as well. Additionally, we demonstrated a novel interaction between γ-ENaC and PIP3.

Studying the mechanisms that regulate ENaC function is important because abnormal channel activity leads to several severe diseases. Constitutive activation of any component of ENaC subunits can cause Liddle's syndrome, an autosomal dominant inherited disease that causes excessive sodium retention and hypertension. Conversely, loss of function mutations in ␣-, ␤-, or ␥-ENaC causes pseudohypoaldosteronism type I, a hypotensive condition characterized by an inability to retain salt. These syndromes highlight the importance of normal ENaC activity in the kidney to maintain fluid and sodium homeostasis. The proper regulation of ENaC activity is also very important in the lung, because transgenic mice lacking functional channels die within 40 h of birth from fluid filled airways (10). Additionally, increases in intracellular Cl Ϫ concentrations that secondarily lead to changes in ENaC activity play an important role in the pathophysiology of cystic fibrosis (11).
Anionic phospholipids, such as phosphatidylinositol 4,5-bisphosphate (PIP 2 ) and phosphatidylinositol 3,4,5-trisphosphate (PIP 3 ), are normally located in the inner leaflet of the plasma membrane and are emerging as important regulators of ion transporters and channels. Although basal levels of both anionic phospholipids are generally very low, several models for the regulation of channels and transporters by PIP 2 and PIP 3 have been proposed. For example, all members of the inward rectifier potassium channel family (K ATP , IRK, GIRK, and ROMK) are thought to be positively regulated by PIP 2 interaction (reviewed in Ref. 12). Of these, the best characterized PIP 2 -binding domain is that of the K ATP channels. PIP 2 binds directly to the C terminus of the K ATP channel, which contains multiple positively charged lysine and arginine residues and maintains an open conformation by preventing ATP binding (13)(14)(15). Classically, PIP 3 is considered to be the lipid product generated when activated phosphoinositide 3-OH kinase (PI3K) phosphorylates PIP 2 at the 3Ј position and is the principle mediator of PI3K effects. Although little is known for its role in regulating the open state of channels, PIP 3 does exhibit binding specificity and may be important in ion channel regulation by hormones and growth factors. It has recently been reported that PIP 3 binds reversibly to regulators of G protein signaling molecules in cardiac cells to regulate K ϩ channel activity in response to changes in intracellular calcium levels. In a resting (low Ca 2ϩ ) state, the action of regulators of G protein signaling is thought to be allosterically inhibited by PIP 3 (16). These studies demonstrate that anionic phospholipids can regulate various ion channels in many different systems and serve as possible analogous models for PIP 3 regulation of ENaC activity in Na ϩ transporting epithelia.
Although the regulation of ENaC has been extensively studied, the specific regulation of ENaC by phosphoinositides remains largely unexplored. However, we have recently demonstrated that application of PIP 2 as well as PIP 3 to the cytoplasmic surface of apical membranes of A6 cells and injected into Xenopus oocytes heterologously expressing ENaC prevented run down of ENaC activity and increased amiloridesensitive channel activity in voltage clamp recordings (17,18). Additionally, Tong et al. (19) demonstrated that PIP 2 and PIP 3 increased the open probability of reconstituted ENaC in excised patches of Chinese hamster ovary cells. Blazer-Yost and Nofziger have recently compared and contrasted the multiple effects of phosphoinositide lipids on ENaC in A6 and Chinese hamster ovary cells in a recent review (20).
In the present study, we examined the direct influence of PIP 3 on ENaC activity in mpkCCD c14 clones, a mouse collecting duct principle cell line, which maintains aldosterone responsiveness and express functional ENaC endogenously (21,22). We also truncated 12 amino acids from the N-terminal tail of ␥-ENaC subunit, suggesting that full-length expression of this subunit is required for normal ENaC trafficking or stability at the plasma membrane. We also demonstrate a novel interaction between ␥-ENaC and phospholipids, including PIP 2 and PIP 3 .

MATERIALS AND METHODS
Cell Culture-The mouse cortical collecting duct principal cell line (mpkCCD c14 ) is often employed in the study of aldosterone-induced ENaC activity because of their specific responsive to physiological concentrations of mineralocorticoid hormone (21, 22). The mpkCCD c14 cells were incubated in a 1:1 mix of Dulbecco's modified Eagle's medium/Ham's F-12 medium (Invitrogen) supplemented with 50 nM dexamethasone, 1 nM triiodothyronine, 20 mM HEPES, 2 mM L-glutamine, 0.1% penicillin/streptomycin, and 2% heat-inactivated fetal bovine serum. Madin-Darby canine kidney (MDCK) and Chinese hamster ovary cells (CHO) (obtained from ATCC, Manassas, VA) are also routinely used in investigating the regulation of sodium channel activity by exogenously expressing ␣-, ␤, and ␥-ENaC subunits (19,23), because these cells do not express a significant amount of endogenous ENaC. MDCK cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) with 10% fetal bovine serum, and CHO cells were cultured in Hain's F-12-Kaighn medium (Invitrogen) with 10% fetal bovine serum. All of the mammalian cell lines were maintained in plastic tissue culture flasks at 37°C with 5% CO 2 in air.
Generation of cDNA Constructs of Tagged ENaCs-Original plasmids containing cDNAs encoding the wild type ␣-, ␤-, and ␥-rENaC in pSport vectors were provided by Dr. Bernard C. Rossier (University of Lausanne, Lausanne, Switzerland). A ␥-rENaC-pDsRed2-N1 construct was created in three steps. First, a 1833-base pair fragment was excised from ␥-rENaC-pSport using EcoRI and ApaI restriction enzymes, which was then subsequently ligated into the corresponding cloning sites of pDsRed2-N1 vector (BD Biosciences, Palo Alto, CA). In the second step, a 242-base pair PCR fragment was synthesized using pcDNA3-␥-rENaC cDNA as a template, with a sense primer (TTGTT-GGGCCCGTAGGCAGA) corresponding to nucleotides 1814 -1833 and an antisense primer (ACCGGTCCCAACTCATTGGTCAACT) corresponding to 2048 -2032. This primer pair was chosen because unique ApaI and AgeI restriction sites are located within the primer sequences. The 242-base pair amplicon, lacking stop codons, was then QIAquick purified (Qiagen, Valencia, CA) and subsequently subcloned into the PGEM-TEasy vector (Promega, Madison, WI). In the third step, PGEM-TEasy containing the 237-base pair PCR product was digested with ApaI/AgeI and then ligated into the pDsRed2-␥-rENaC vector described above. In this way, we eliminated endogenous stop codons and cloned ␥-rENaC in frame with pDsRed vector. The ␤-rENaC-pEYFP-N1 and ␣-rENaC-pECFP-N1 constructs were generated using similar strategies.
The ␥ T -ENaC-pDsRed construct, encoding a protein in which the N-terminal tail of ENaC (APGEKIKAKIKK) is truncated, was created in three steps. First, a 36-nucleotide segment from the 5Ј end of ␥-ENaC was removed by EcoRI and BamHI enzyme digestion. Then a 393-PCR synthesized fragment was generated by using pcDNA3-␥-ENaC as a template and the primer pair 5Ј-ACCATGGCTCTGCCGGTTCGA and 3Ј-GGACGGCATGGATCCTGCTT), to recreate the ATG start site and Kozak sequences in the truncated ␥-ENaC-pDsRed construct. This was achieved by expressing the 393-base pair PCR fragment in pGEM-TEasy construct and subsequently cloning it in frame with pDsRed EcoRI and BamHI sites.
DNA Transfections-mpkCCD c14 cells were transfected with GFPfused pleckstrin homology (PH) domains of either Grp1 or dynamin construct (obtained from Dr. Mark A. Lemmon, University of Pennsylvania School of Medicine, Philadelphia, PA); MDCK cells were transfected with fluorescently labeled ␣-, ␤-, and ␥-ENaC constructs (described above), and CHO cells were transfected with either fulllength ␥-ENaC or the N-terminal tail truncated ␥ T -ENaC construct or pDsRed vector alone. Each cell line was transfected using Lipofectamine Plus reagent (Invitrogen) in accordance with the manufacturer's recommended protocol. Briefly, the cells were seeded at subconfluent densities 1 day before the transfection. DNA constructs were diluted with serum-free medium (1 g DNA/50 l medium), mixed with Plus reagent, and incubated at room temperature for 15 min. Then Lipofectamine reagent was diluted with serum-free medium (1 l of Lipofectamine/25 l of medium), mixed with the DNA/Plus solution, and incubated at room temperature for an additional 15 min. Finally, the transfection solution containing DNA, Plus reagent, and Lipofectamine reagent was applied to the cells and allowed to incubate for 4 -6 h at 37°C before the transfection solution was replaced with regular growth medium.
Patch Clamp Recording and Analysis-For patch clamp experiments, either MDCK cells or mpkCCD c14 were grown to confluent densities on permeable polyester membranes. The permeable support allowed patch pipette access to the apical membrane, as well as a physical separation of the apical and basolateral bath compartments. Immediately before use, the cells were thoroughly washed with NaCl bath solution (145 mM NaCl, 5 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , and 10 mM HEPES, at a pH of 7.4), transferred into the patch recording chamber mounted on a Nikon Eclipse TE200 inverted microscope, and then visualized with Hoffman modulation optics. The whole cell configuration was established with polished patch pipettes with a tip resistance of 2 M⍀. Only patches with a seal resistance above 10 G⍀ were used for the experiments. Pipette solution contained 145 mM KCl, 5 mM NaCl, 50 nM free Ca 2ϩ (after titration with 2 mM EGTA), 1 mM MgCl 2 , 2 mM K 2 -ATP, and 10 mM HEPES, at a pH of 7.2. A voltage step protocol from Ϫ120 mV to ϩ40 mV (in 20-mV intervals) was used to monitor the current using an Axopatch 1-D (Axon Instruments, Union City, CA). The data were acquired using TL-1 acquisition hardware and analyzed with pClamp software (Axon Instruments). Patch clamp recordings were performed at room temperature.
Localization of PIP 3 and ENaC Subunits with Laser Confocal Microscopy-According to techniques that have been previously established, the localization of PIP 3 can be visualized after transfecting cells with a GFP-fused PH domain (24 -27). Therefore, mpkCCD c14 cells were cultured on glass coverslips and then transiently transfected with the GFP-fused Grp1-PH construct to localize endogenous PIP 3 in a collecting duct cell line. Fluorescently labeled ENaC subunits (␥-rENaC-pDsRed2-N1, ␤-rENaCpEYFP, and ␣-rENaCpCFP-N1) were also transfected into MDCK cells grown on permeable supports, as described above. Cells transfected with either the GFP-fused Grp1-PH construct or fluorescent vectors containing ENaC subunits were visualized using Zeiss LSM 510 NLO META confocal microscope (Zeiss, Thornwood, NY).
Protein-Lipid Overlay-To test the binding properties of ␥ T -ENaC and wild type ␥-ENaC, a protein-lipid overlay was performed using PIP MicroStrips, commercially available from Echelon Biosciences Inc. (Salt Lake City, UT). These strips contain 100 pmol of various phospholipids (listed in Fig. 7), spotted and immobilized on a nitrocellulose membrane. Cell lysate from CHO cells that were transiently transfected with full-length ␥-ENaC or N-terminal tail truncated ␥ T -ENaC subunits (in the presence of ␣and ␤-ENaC) were overlaid onto the PIP MicroStrips. As a negative control, we also transfected cells with the pDsRed construct only and proceeded with the same experimental protocol. CHO cells were thoroughly rinsed with phosphate-buffered saline before lysing with hypotonic gentle lysis buffer containing: 10 mM Tris HCl (pH 7.5), 10 mM NaCl, 2 mM EDTA, 0.5%Triton X-100, and freshly prepared 1ϫ protease inhibitor mixture. The nitrocellulose strips were blocked in TBS-T 3% fatty acid-free BSA (Sigma) for 1 h at room temperature. Then ϳ26 g/ml protein from CHO cell lysate (expressing either ␣,␤,␥ T -ENaC or ␣,␤,␥-ENaC) was incubated with the strips in TBS-T 3% fatty acid-free BSA at 4°C overnight. The strips were then washed with TBS-T/BSA three times with gentle agitation, for 10 min each wash, at room temperature. ␥ T -ENaC and ␥-ENaC interaction with spotted phospholipids were detected by subsequently blocking the strips in TBS-T buffer (10 mM Tris, pH 7.5, 70 mM NaCl, and 0.1% Tween) with 5% dry milk and then incubating the strips in a 1:10,000 dilution of polyclonal rabbit anti-DsRed-antibody (BD Bioscience, Palo Alto, CA) in blocking buffer for 1 h. IgG-alkaline phosphatase-labeled secondary antibody (KPL, Gathersburg, MD) was added (1:5,000 in blocking buffer) and incubated for an additional hour at room temperature. After thorough washes, alkaline phosphatase signal was detected using Nitroblock chemiluminescence enhancer (Tropix, Bedford, MA) and CDP-Star substrate (Tropix) in combination with Kodak Image Station 200MM and Kodak 1D software (Kodak, New Haven, CT).
Protein Pull-down Assay-To demonstrate effective ␥ T -ENaC and ␥-ENaC expression and interaction with PIP 3 in our cell model, we performed additional protein pull-down assays. CHO cells expressing ␣,␤,␥ T-ENaC or ␣,␤,␥-ENaC were lysed as described above in gentle lysis buffer. 2 mg of CHO cell lysate incubated overnight, at 4°C, with 75 l of phosphatidylinositol 3,4,5-triphosphate bound beads (Echelon Biosciences Inc.). The protein-bound beads were then thoroughly washed with TBS-T 3% fatty acid-free BSA three times. The protein was eluted from the beads by adding 2ϫ Laemmli sample buffer and heated at 95°C. Standard PAGE and immunoblot techniques, described above, were used to detect ␥ T -ENaC and ␥-ENaC pull-down with PIP 3 after transfer to membrane. We subjected CHO cell lysate expressing DsRed only to the same protein pull-down assay as an appropriate negative control.
Statistical Analysis-The data are reported as the mean values Ϯ S.E. Statistical analysis was performed with SigmaPlot and SigmaStat software (Jandel Scientific, CA). Paired or unpaired t tests were used to determine statistical significance between two groups. Analysis of variance was used for multiple comparisons. The results were considered significant if p Ͻ 0.05, as we described previously (28).

PIP 3 Stimulates ENaC in Control mpkCCD c14 Cells but Not in Aldo-
sterone-treated mpkCCD c14 Cells-It is well known that blocking production of PIP 3 by inhibiting PI3K blocks the effect of aldosterone on sodium transport in renal cells (29,30). However, if an aldosteroneinduced increase in PIP 3 is the major cause of the initial hormoneinduced increase in ENaC activity, then the addition of PIP 3 to the cytosolic surface of renal cells should increase ENaC activity in the absence of aldosterone. Using the inside-out configuration, we recently demonstrated that PIP 3 did not elevate but only maintained ENaC activity in aldosterone-treated A6 cells (17). We hypothesized that the stimulatory effect of PIP 3 on ENaC activity may be already saturated in A6 cells that are continuously cultured in the presence of a high concentration of aldosterone. To test this hypothesis, in the present study we performed whole cell voltage clamp experiments in mpkCCD c14 cells, which do not require a high dose of aldosterone for growth and differentiation. Whole cell currents in response to a voltage step protocol (see "Methods and Materials") were recorded from control mpkCCD c14 cells and aldosterone-treated mpkCCD c14 cells. The patch pipettes were back-filled with a solution containing 10 M PIP 3 . Compared with the current immediately after forming the whole cell configuration, amiloride-sensitive current was significantly increased at 5 min (PIP 3 had already diffused into the cells) in an aldosterone-free control cell (Fig.  1A) but was not increased or was increased to a lesser degree in an aldosterone-treated cell (Fig. 1B). Amiloride-sensitive currents at Ϫ100 mV in control cells were Ϫ0.30 Ϯ 0.06 nA (control) and Ϫ1.30 Ϯ 0.27
Sequestering PIP 3 Reduces Aldosterone-stimulated ENaC Current-PH domains are small stretches of 100 -120-amino acid sequences found in many cell signaling and cytoskeletol proteins. PH domains bind with high specificity and affinity to phosphoinositides. In this way, the PH domain directly targets the "host" signaling or cytoskeletol protein to the cellular membrane. Also, the specific binding characteristics of PH domains are utilized in studying PIP 2 and PIP 3 activity in vivo. For example, the PH domain of Grp1 (Grp1-PH) binds to PIP 3 with high specificity and affinity (K d ϭ 32 nM), (26,31) and can be used to greatly reduce endogenous PIP 3 levels of activity. However, the PH domain of dynamin (Dyn-PH) only weakly associates with PIP 3 (K d ϭ 1.4 -4 M) (32,33). Therefore, Dyn-PH is an appropriate negative control for experiments sequestering PIP 3 with the Grp1-PH expression.
The effects of Grp1-PH and Dyn-PH expression in mpkCCD c14 cells on whole cell current are shown in Fig. 2. Very low basal levels of current were recorded from cells before aldosterone treatment, as expected, whether they were untransfected, expressed Grp1-PH domain, or expressed Dyn-PH domain ( Fig. 2A). However, after 1 M aldosterone treatment, whole cell current increased in the untransfected and Dyn-PH control cells but not in mpkCCD c14 , in which endogenous PIP 3 activity had been sequestered by Grp1-PH domain. The mean amiloride-sensitive current at Ϫ100 mV in the presence (black bars) and absence of aldosterone (white bars) is shown in Fig. 2C. PIP 3 significantly increased amiloride-sensitive current 5.6-and 8.0-fold in control (Ϫ0.26 Ϯ 0.04 nA versus Ϫ1.48 Ϯ 0.34 nA, n ϭ 5; p Ͻ 0.01) and Dyn-PH-transfected cells (Ϫ0.21 Ϯ 0.04 nA versus Ϫ1.68 Ϯ 0.30 nA, n ϭ 4; p Ͻ 0.01), respectively. Aldosterone, however, did not significantly increase the current in Grp1-PH expressing mpkCCDc14 cells (Ϫ0.34 Ϯ 0.07 nA versus Ϫ0.50 Ϯ 0.14 nA, n ϭ 6; p Ͼ 0.05).
In our studies, aldosterone failed to increase ENaC activity in cells that expressed the Grp1-PH domain (which has high binding specificity and affinity to PIP 3 ) compared with cells transfected with another type of PH domain, Dyn-PH, which has low binding affinity for PIP 3 . This is strong evidence that the PI3-K product, PIP 3 , specifically mediates the stimulation of ENaC by aldosterone. 3 Concentration via PI3K-As we have described above, the Grp1-PH domain binds strongly and specifically to PIP 3 and is therefore commonly employed in the study of PIP 3 in vivo. By fusing Grp1-PH to GFP, we were able to localize PIP 3 expression in mpkCCD c14 cells. In the absence of serum and aldosterone, we expect GFP-Grp1-PH expression to be evenly distributed across the cytoplasm with low fluorescence intensity, because PIP 3 levels under resting conditions are very low. However, aldosterone-induced activation of PI3K should enhance PIP 3 levels at the plasma membrane.

Aldosterone Elevates Membrane PIP
To test our hypothesis that aldosterone elevates membrane PIP 3 concentrations via PI3K, we first transfected mpkCCD c14 cells with the GFP-fused Grp1-PH domain (Fig. 3). The cells that were deprived of serum and hormone did not contain significant amounts of PIP 3 . In this basal state, the expressed GFP-fused Grp1-PH domain was distributed with an even intensity across the whole cytoplasm of mpkCCD c14 cells as expected (left panel). Cells treated with 1 M aldosterone for 30 min (middle panel) displayed predominant GFP-fluorescence intensity at the plasma membrane. This effect was prevented by pretreating the cells with 5 M LY294002, a specific PI3K inhibitor (right panel), indicating that aldosterone can elevate the concentration of PIP 3 in the plasma membrane (where functional ENaC resides) by stimulating PI3K. These data suggest that PIP 3 is an important regulator of aldosterone-induced sodium channel activity.
Truncation of ␥-ENaC N-terminal Tail Decreases Amiloride-sensitive Sodium Current-The N termini of ENaC subunits are very important in normal ENaC function. It has been shown that deletion of positively charged motifs in the cytoplasmic N termini of ␤-(⌬2-49) and ␥-ENaC (⌬2-53) dramatically reduces ENaC activity (34). We determined whether removal of 12 amino acids (⌬2-13), which include several conserved lysine residues (shown in Fig. 4D) from the N-terminal tail of ␥-ENaC, would lead to a reduction in amiloride-sensitive current and alter subunit translocation in MDCK cells.
Truncation of ␥-ENaC N-terminal Tail and Inhibition of PI3K Impede ENaC Translocation to the Plasma Membrane-Because ␣-, ␤-, ␥-, and ␥ T -ENaC subunits were cloned into pECFP, pEYFP, pDsRed, and pDsRed vector, respectively, we were able to perform confocal microscopy experiments to determine the effect of the N-terminal tail truncation and ␥-ENaC deletion on sodium channel translocation to the plasma membrane in MDCK cells. The localization of ␣-ENaC (blue), ␤-ENaC (yellow), or ␥-ENaC (red) subunit in the cells is shown separately in the first three panels of Fig. 5 and is then superimposed in the last panel. Compared with that under control conditions (Fig. 5A), wild type ␣␤␥-ENaC subunits all translocated to the plasma membrane at 2 h after 1 M aldosterone treatment at 37°C (Fig. 5B). Aldosteroneinduced trafficking of wild type ENaC subunits to the plasma membrane was abolished by 5 M LY294002 (Fig. 5C), strongly suggesting that PI3K-generated lipids (such as PIP 3 ) provide a recruitment mechanism for ENaC to the apical membrane. Importantly, expression of an N-terminal truncated ␥-ENaC (␥ T -ENaC) with full-length ␣and ␤-ENaC subunits prevented ENaC trafficking to the plasma membrane (Fig. 5D) and is consistent with our finding that ␣␤␥ T -ENaC expression in MDCK cells leads to decreased amiloride-sensitive current.
We performed additional experiments in which only ␣and ␤-ENaC subunits were expressed in MDCK cells. Fig. 6 shows that the ␣,␤-ENaC subunits can be efficiently expressed in the absence of ␥-subunit. Using the same excitation wavelengths and gain settings in confocal analysis, it appears that ␣and ␤-ENaC expression levels are similar to the levels reached by transfecting all three ENaC subunits in MDCK cells, shown in Fig. 5. However, no detectable amiloride-sensitive current was observed in ␣,␤-ENaC only, co-transfected cells (data not shown). Furthermore, Fig. 6 shows that the ␣ and ␤ subunits do not effectively traffic to the plasma membrane in response to 1 M aldosterone treatment in the absence of ␥-ENaC expression. Although ENaC trafficking is greatly limited in these cells, there is still some membrane localization of the ␣and ␤-subunits. Our findings show that complete, full-length expression of the ␣-, ␤-, and ␥-ENaC subunits are requisite for the formation as well as effective translocation of functional sodium transporting channels at the apical membrane of kidney cells.   Phospholipid Binding Specificity of ␥-ENaC-Because our data show that anionic phospholipids can mediate ENaC activity, we next tested the ability of ␥-ENaC to bind to various phospholipids using a proteinlipid overlay method. Phospholipids, including phosphoinositides, were spotted onto a membrane (Echelon Bioscience Inc.) and incubated with either cell lysate from ␣,␤,␥-ENaCor ␣,␤,␥ T -ENaC-transfected CHO cells. The membranes were then washed and immunoblotted using anti-DsRed antibody to detect ␥-rENaC-pDsRed2-N1 binding to the membrane, via direct interactions with the lipids. As shown in Fig. 7A, both ␥ T -ENaC and ␥-ENaC interacted with 100 pmol of PI(3,4)P 2 , PI(3,5)P 2 , PI(4,5)P 2 , PI(3,4,5)P 3 , phosphatic acid, phosphatidylserine, PI, PI(3)P, PI(4)P, and PI(5)P. The ENaC subunits did not bind to sphingosine-1-phosphate, lysophosphatidic acid, lysophosphocholine, phosphatidylethanolamine, phosphatidylcholine, nor the control in which no lipid was spotted onto the membrane (position 8 on the strips). As an additional control, we also demonstrated that DsRed protein alone did not bind nonspecifically to membrane that had only been incubated with protein lysate from pDsRed-transfected CHO cells. Fig. 7B (left panel) also confirms effective ␥and ␥ T-rENaC-pD-sRed2-N1 expression in CHO cells and that these ENaC subunits can pull down with 1.5 nmol of PIP 3 bound beads. We show in the right panel, as a negative control, that the DsRed label alone does not contribute to PIP 3 protein pull-down. Together, these data show that both the full-length and N-terminal tail truncated forms of ␥-ENaC bind to phospholipids spotted on a nitrocellulose membrane and can pull-down with PIP 3 immobilized on beads.

DISCUSSION
We previously demonstrated that anionic phospholipids including PIP 3 maintained ENaC activity in inside-out patches excised from aldosterone-conditioned A6 cells (17). Although results from this previous publication convincingly showed that PIP 3

increases the likelihood that
ENaC would be open, these single channel studies could not discern whether an increase in the number of channels trafficked to the apical membrane could also be responsible for maintaining ENaC activity in PIP 3 -treated cells. Our current study coupled with our previous observations suggest that PIP 3 serves as an effective regulator of ENaC trafficking (from the cytoplasmic pool to the surface membrane), promotes an open state of the channel, or maintains channel stability through direct interactions with the ENaC subunit at the plasma membrane.
Using the mouse mpkCCD cell line, which does not require aldosterone for growth and differentiation, we demonstrated that PIP 3 stimulated ENaC activity in the absence of aldosterone but could not further increase ENaC activity in the presence of aldosterone. We also showed that aldosterone elevated the concentration of PIP 3 in the plasma membrane and that sequestering PIP 3 with Grp1-PH domain prevented aldosterone activation of ENaC. Although the ␣and ␤-subunits were expressed at high levels in MDCK cells, ␣/␤-ENaC complexes could not completely traffic to the plasma membrane in the absence of ␥-ENaC expression, because we could measure no amiloride-sensitive current (data not shown). Truncation of lysine-rich residues in the N-terminal end of ␥-ENaC similarly inhibited aldosterone-induced increases in current and prevented appropriate channel translocation to the plasma membrane after aldosterone treatment. We originally thought that this might be due to a failure of ␥ T to bind PIP 3 , but our lipid overlay assays revealed that the ␥ T -ENaC binds to phospholipids, including PIP 3 . It appears that the binding ability of the N-terminally truncated form of ␥-ENaC was slightly higher than that of wild type. Therefore, it is possible that the N-terminal tail is required for channel gating or trafficking for reasons other than lipid binding. A report just came out suggesting that the region immediately following the second transmembrane spanning domain of ␥-ENaC acts as part of a functional PIP 3 -binding site (35). We are currently investigating additional arginine-and lysine-rich domains in ENaC subunits, which may directly interact with anionic phospholipids, as hypothesized in our recent review article (36).
However, our current model for the regulation of ENaC by anionic phospholipids does not exclude a role for the serum and glucocorticoidinducible kinase (SGK1), an immediate aldosterone induced kinase that increases the activity of ENaC (37)(38)(39)(40)(41)(42)(43). The upstream regulators of SGK1 enzyme activity are 3-phosphoinositide-dependent kinase-1 and Ϫ2 (PDK1 and PDK2, respectively); thus SGK1 is also inhibited by PI3K inhibitors such as LY294002 and is dependent upon PIP 3 for complete activation (44,45). PIP 3 may enhance ENaC function by associating with  SGK1 and recruit this kinase to the inner leaflet of the plasma membrane. Once at the appropriate site of PDK1 and PDK2 activation, SGK1 could then inhibit ubiquitin ligase Nedd4 -2 activity (as we currently understand it to). Because normal ENaC function is so important in maintaining fluid and ion homeostasis, it makes sense that tight epithelial cells would utilize multiple pathways to ensure net Na ϩ re-uptake.