FoxO1 inhibits transcription and membrane trafficking of epithelial Na⁺ channel

The epithelial Na⁺ channel (ENaC) localizes to the apical membranes of principal cells in the distal nephron, where it contributes to the maintenance of whole-body Na⁺ homeostasis, and consequently extracellular fluid volume as well as blood pressure (Butterworth, 2010). Therefore, ENaC is the major pathway that determines the fine control over distal nephron Na⁺ reabsorption. ENaC dysregulation has been implicated in many other clinical conditions, such as pulmonary edema, acute respiratory distress syndrome and nephrosis, as well as salt-dependent hypertension (Pratt, 2005; Soundararajan et al., 2010a).

In the terms of ENaC structure, it comprises three homologous subunits, α, β and γ. Each subunit contains two transmembrane domains, an extracellular loop and intracellular N- and C-termini (Butterworth, 2010). During Na⁺ repletion, the β- and γ-subunits of ENaC are present in distal nephron principal cells, but the low expression of the ENaC α subunit (α-ENaC; also known as SCCN1A) is limiting for channel function because α-ENaC is essential for the assembly of functional Na⁺ channels that can traffic to the apical surface (Loffing et al., 2001). Thus, stimulation of transcription of α-ENaC by any transcription factors permits assembly of the heterotrimeric channel in the endoplasmic reticulum (ER) and its trafficking to the apical membrane, thereby providing an important control point for enhancement of Na⁺ absorption.

Various hormones, including aldosterone, vasopressin and insulin have been reported to regulate ENaC expression in the kidney to a high degree (Bubien, 2010). Aldosterone exerts its stimulation of Na⁺ transport in both early (minutes to hours) and late (hours to days) phases. There are different ways for aldosterone to increase Na⁺ transport. Cell localization studies show that aldosterone redistributes ENaC subunits from intracellular compartments to the apical membranes of principal cells, particularly in the early phase of the response (Ergonul et al., 2006; Frindt et al., 2001; Masilamani et al., 1999). Aldosterone increases the synthesis of a number of new proteins and represses others (Robert-Nicoud et al., 2001); these gene products modulate the number of Na⁺ channels in the apical membrane, as well as increase their activity. Increases in mRNA encoding ENaC and channel protein expression contribute to the increase in apical channel density. Among the most significant proteins induced by aldosterone are serum- and glucocorticoid-induced kinase (SGK1) and the glucocorticoid-induced leucine zipper protein (GILZ) (Bhalla et al., 2006). Both SGK and GILZ inhibit ubiquitin-dependent internalization of ENaC (Malik et al., 2006). SGK1 blocks the interaction of ENaC with the E3 ubiquitin ligase Nedd4-2 (also known as NEDD4L), which binds to the PY motifs at the ENaC subunit C-termini to permit channel ubiquitylation and internalization through clathrin-mediated adapter protein binding. GILZ acts by inhibiting ERK-mediated phosphorylation of ENaC, which results in a relative decrease in the affinity of Nedd4-2 for the channel. The processes of ENaC ubiquitylation and internalization account for the relatively short half-life of ENaC at the plasma membrane that is observed under non-stimulated conditions. The synergistic interference with the action of Nedd4-2 by these SGK1- and GILZ-mediated phosphorylation events leads to a decrease in channel ubiquitylation and retrieval, and increased Na⁺ reabsorption (Soundararajan et al., 2009, 2010a,b). Our previous work demonstrates that the phosphorylation-dependent inactivation of Nedd4-2 is stabilized by binding to the 14-3-3 protein, which assists in blocking ENaC–Nedd4-2 interactions (Liang et al., 2008, 2006). In fact, a significant fraction of the aldosterone-induced increase in ENaC expression results from stabilization of the channel at the apical surface, which protects it from degradation pathways.

Vasopressin, prostaglandins and insulin usually exert acute regulation on ENaC activity. Insulin is considered as one of the causal factors of hypertension, resulting from its activation of ENaC and renal Na⁺ retention. It has been reported that insulin induces the migration of ENaC from the cytoplasm to the apical and lateral membranes in A6 cells (Blazer-Yost et al., 2003), and it was observed that insulin stimulates ENaC activity in mCCD cells (Ilatovskaya et al., 2015). However, the underlying mechanism of ENaC regulation by insulin remains unclear.

Nuclear transcriptional factor Forkhead box protein O1 (FoxO1) is one of the predominant members of the forkhead O gene family. FoxO1 has a conserved threonine residue and two conserved serine residues (Thr24, Ser256 and Ser319) (Barthel et al., 2005). The different kinases, including Akt1/protein kinase B (PKB) (Brownawell et al., 2001), SGK1 (Brunet et al., 2001), cyclin-dependent kinase 2 (CDK2) (Huang et al., 2006) and IkB kinase (IKK) (Hu et al., 2004) phosphorylate FoxO1, and subsequently modulate its nuclear exclusion. In the presence of insulin, FoxO1 is phosphorylated by activated Akt; resulting in the cytoplasmic retention of FoxO1. This prevents FoxO1 from accessing the nucleus and acting as a transcriptional factor (Kamagate and Dong, 2008). Phosphorylated FoxO1 usually interacts with 14-3-3 proteins in the cytoplasm, which helps to maintain FoxO1 in the phosphorylated state (Rena et al., 2001). In the absence of insulin, FoxO1 stays in a dephosphorylated state and relocates back into nucleus, where it exerts transcriptional activity.

FoxO1 is involved in diverse activities, including the response to modulation of Na_V1.5 (Mao et al., 2012) and K⁺ ATP channels (Philip-Couderc et al., 2008), oxidative stress, insulin-dependent regulation of metabolism (Puigserver et al., 2003) and commitment to apoptosis (Stahl et al., 2002). The growing evidence has demonstrated that the signaling pathway of insulin regulates FoxO1 and is involved in modulation of ENaC trafficking. Promoter analysis suggests that the promoter of α-ENaC contains the binding elements for FoxO1. Moreover, knockdown of aldosterone-induced 14-3-3 isoforms in our previous work almost completely suppressed aldosterone action and its effect on apical ENaC density, leading to the hypothesis that phosphorylated proteins other than Nedd4-2 bind to 14-3-3 proteins in order to promote regulation of ENaC trafficking.

Guided by this concept, we sought to determine whether FoxO1 participates in the regulation of ENaC by examining whether FoxO1 is one of the 14-3-3 binding partners that modulate ENaC. Accordingly, we examined the involvement of FoxO1 in the transcription and protein expression of ENaC, as well as in the insulin-stimulated forward trafficking of ENaC to the apical surface of mCCD epithelia.

We determined whether FoxO1 participates in the regulation of α-ENaC by infecting mCCDs with ADA-FoxO1, a constitutively active form of human FoxO1. FoxO1 cDNA between mouse and human has approximately 92% amino acid sequence similarity.

Real-time PCR and western blotting analyses were performed to determine the expression of α-ENaC in mCCD cells. The cDNAs generated from control mCCDs, empty vector-infected mCCDs and ADA-FoxO1-infected mCCDs were used with primers for α-ENaC and FoxO1 in real-time PCR analyses. FoxO1 mRNA expression was obviously increased by ADA-FoxO1 infection in mCCD cells, as shown in supplementary material Fig. S1, but ADA-FoxO1 infection significantly decreased the expression of α-ENaC mRNA in mCCD cells (Fig. 1A). In other experiments, cell lysates were obtained from control, vector only and ADA-FoxO1-infected mCCD monolayers, as described above, and the lysates were resolved by SDS-PAGE and probed for α-ENaC and FoxO1 proteins. We examined total and apical surface expression of α-ENaC as a biochemical marker of steady-state channel density. Representative blots are shown in Fig. 1B and the data are summarized in Fig. 1C. ADA-FoxO1 expression significantly decreased the total and apical α-ENaC expression. These results agree qualitatively with the evaluation of α-ENaC expression at the RNA level.

To verify the suppressive effect of ADA-FoxO1 on α-ENaC expression, we performed immunofluorescence staining using a rabbit polyclonal antiserum against α-ENaC in polarized mCCD cells with or without ADA-FoxO1 infection. As shown in Fig. 1D, α-ENaC expression was significantly decreased in mCCD cells that had been infected with ADA-FoxO1, which is consistent with the biochemical data described above.

The physiological significance of FoxO1 in the regulation of α-ENaC expression was evaluated under the basal condition in knockdown experiments performed with siRNA targeting FoxO1 expression (FoxO1 siRNA). The results were compared with those obtained with scrambled siRNA as a control. The transfection conditions permitted assays of protein expression and transepithelial current in polarized mCCD cells. FoxO1 mRNA expression following siRNA transfection was detected and confirmed by real-time PCR analysis (supplementary material Fig. S2). FoxO1 knockdown significantly increased α-ENaC mRNA expression, as shown in Fig. 2A. The siRNA-induced reduction in FoxO1 expression in different experiments averaged 60–90%, whereas the scrambled siRNA had no effect relative to untreated controls. We next examined the total and apical surface expression of α-ENaC as described above. Under basal (non-stimulated) conditions, FoxO1 knockdown significantly increased the levels of cell surface α-ENaC approximately fourfold, with a slight alteration of the total α-ENaC expression (Fig. 2B). The mean data for apical α-ENaC expression are shown in Fig. 2C; α-ENaC expression at the protein and RNA level was very consistent during FoxO1 knockdown.

We also determined the functional impact of reduced FoxO1 expression on ENaC-mediated transepithelial Na⁺ absorption by determining the amiloride-sensitive short-circuit current (I_sc) across mCCD epithelia. The mean data are shown in Fig. 2D. Transfection with siRNA targeting FoxO1 expression elicited an approximately 2.5-fold increase in the amiloride-sensitive I_sc across mCCD epithelia, whereas the control siRNA had no significant effect. These functional data are in agreement with the biochemical findings obtained above, and they highlight the physiological significance of FoxO1 in Na⁺ transport.

Furthermore, we investigated whether ADA-FoxO1 overexpression through viral vector infection rescued the effects of FoxO1 knockdown on α-ENaC expression in mCCD cells. As shown in Fig. 2E, increased apical surface and total expression of α-ENaC upon transfection with FoxO1 siRNA was blunted by ADA-FoxO1 infection in mCCD cells. Quantification of the immunoblot data for apical α-ENaC expression is shown in Fig. 2F.

Insulin is a well-known inducer of FoxO1 phosphorylation (probably mediated by the PI3K–Akt pathway) and results in FoxO1 exclusion from the nucleus. We first examined the expression of insulin receptor β (IRβ, encoded by Igf1r) in mCCD cells with or without insulin treatment (100 nM) for 2 h using western blotting. As shown in supplementary material Fig. S3, IRβ was physically expressed in mCCD cells, and expression of phosphorylated IRβ was markedly increased by addition of insulin. We further investigated whether FoxO1 mediated the expression of ENaC by using immunoblot analyses to examine the expression of α-ENaC, phosphorylated FoxO1, FoxO1, phosphorylated Akt and Akt as a function of time during the stimulation of mCCD epithelia with 100 nM insulin. As shown in Fig. 3A, the expression of α-ENaC, phosphorylated FoxO1 and phosphorylated Akt was increased during the action of insulin, whereas the expression of total FoxO1 and Akt remained unchanged. These data show that the increases in α-ENaC expression occurred in parallel with the increased phosphorylation of FoxO1. Quantification of the immunoblot data from all experiments, which included normalization to β-actin expression (Fig. 3B), indicate that α-ENaC, phosphorylated FoxO1 and phosphorylated Akt each increased about threefold during the insulin treatment period. Thus, the insulin-induced expression of α-ENaC is similar in time course to that of other crucial components of the regulatory pathways that mediate the steroid response. However, the increase in α-ENaC expression observed upon insulin treatment (100 nM for 2 h) was blunted by about 60% by infection with ADA-FoxO1, as shown in Fig. 3C. The mean data for α-ENaC expression are shown in Fig. 3D.

Our previous studies have shown that 14-3-3 proteins bind to proteins that have been phosphorylated in response to steroid hormones and that are, therefore, involved in the stimulation of transepithelial Na⁺ transport (Liang et al., 2010, 2006). We evaluated the mechanism of FoxO1 in ENaC regulation by examining the selectivity of its interaction with the five 14-3-3 isoforms that we have previously identified to be expressed in polarized mCCD epithelia (Liang et al., 2006). We used antibodies against FoxO1 to isolate protein complexes from epithelia that had or had not been treated with insulin. The isolated proteins were then examined by western blotting to detect the different 14-3-3 isoforms (Fig. 4A). IgG was used as a control in the immunoprecipitation experiments and yielded no 14-3-3 signal. Under basal conditions, no interactions were observed between FoxO1 and 14-3-3 isoforms. The treatment with insulin markedly increased the FoxO1 interaction with 14-3-3ε; no positive signals were detected for binding of FoxO1 to the 14-3-3β, 14-3-3γ, 14-3-3θ and 14-3-3ζ isoforms.

We next used an antibody against 14-3-3ε in competition experiments, in an attempt to further confirm 14-3-3ε binding to phosphorylated FoxO1. Pro-Ject™ Protein Transfection was used to deliver the antibody against 14-3-3ε into mCCD cells. As shown in supplementary material Fig. S4, the IgG heavy chain (IgG-H)-positive signal indicated successful transfection of the antibody into mCCD cells. In the subsequent experiments, cell lysates were obtained from mCCD epithelia under basal and insulin-treated conditions after transfection with the 14-3-3ε-targeting antibody. The protein complexes collected following immunoprecipitation of 14-3-3ε were blotted for phosphorylated FoxO1, anti-IgG as control for immunoprecipitations, as shown in Fig. 4B. As anticipated, the level of phosphorylated FoxO1 binding to 14-3-3ε was minimal under the basal, unstimulated conditions. Treatment with insulin increased the expression of phosphorylated FoxO1 and augmented its association with 14-3-3ε (Fig. 4B, Fig. 3A). However, pre-transfection of mCCD cells with an antibody against 14-3-3ε reduced the levels of phosphorylated FoxO1 binding to those observed in epithelia that had not been treated with insulin. The data from all experiments are quantified in Fig. 4C.

We further evaluated the role of the FoxO1 binding to 14-3-3ε using a 14-3-3ε-targeting antibody in competition experiments. As shown in Fig. 4D, consistent with data provided above, treatment with insulin increased the expression of apical α-ENaC, total α-ENaC and phosphorylated FoxO1. However, these increased expressions were substantially blunted in cells that had been pre-transfected with the antibody against 14-3-3ε.

To determine whether FoxO1 regulates α-ENaC expression transcriptionally, we cloned the mouse α-ENaC promoter into a luciferase reporter system and observed the effect of FoxO1 on α-ENaC promoter activity. As shown in Fig. 5A, the overexpression of ADA-FoxO1 led to a significant decrease of α-ENaC promoter activity with either 400, 200 or 100 multiplicity of infection (MOI) in HEK293 cells. Furthermore, we investigated the impact of insulin on FoxO1-mediated suppression of α-ENaC promoter activity. After transfection with the ADA-FoxO1 vector at a given titer, HEK293 cells were cultured in the presence or absence of insulin in the medium. ADA-FoxO1, a mutant of three conserved phosphorylation sites (Thr24, Ser256 and Ser319), was unable to undergo phosphorylation and cytoplasmic retention under insulin stimulation. Therefore, ADA-FoxO1 is constantly located in the nucleus, resulting in the constitutive expression of FoxO1 target genes. As shown in Fig. 5B, the relative levels of luciferase activity, used to reflect α-ENaC promoter activity, were significantly downregulated upon ADA-FoxO1 overexpression but were increased in HEK293 cells by treatment with insulin alone. However, the presence of insulin in the culture medium did not affect ADA-FoxO1-mediated suppression of α-ENaC promoter activity.

We defined the FoxO1 target region in the α-ENaC promoter by generating a series of truncated portions of the α-ENaC promoter. These constructs were then subcloned into the luciferase reporter system (pENaC2000, pENaC1500, pENaC1000, pENaC500 and pENaC200). Following transfection into HEK293 cells, the activity of each construct was measured with or without FoxO1 expression.

Deletion of the region −2000 nt to −500 nt relative to the α-ENaC promoter start site (pENaC2000, pENaC1500, pENaC1000 and pENaC500) did not change the promoter activity in the presence of FoxO1 in HEK293 cells (Fig. 6). However, the promoter variant (pENaC200), made by further deletion of up to −200 nt in the α-ENaC promoter, did not response to FoxO1-mediated suppression. These results confine the putative FoxO1 target site in the α-ENaC promoter to a small region (−500 to −200 nt).

Growing evidence suggests that the regulation of ENaC expression and trafficking is controlled by a series of kinase-mediated signaling pathways that respond to multiple hormonal stimulations. The mineralocorticoid hormone aldosterone regulates Na⁺ transport through both early (minutes to hours) and late (hours to days) components (Butterworth, 2010; Butterworth et al., 2009). Much of what we know about the regulatory mechanism of ENaC is associated with the effects of aldosterone. Vasopressin, prostaglandins and insulin usually undergo acute hormonal regulation of ENaC activity compared with chronic regulation of aldosterone on ENaC. (Butterworth et al., 2009).

It has been shown that insulin increases Na⁺ transport in both tissues and model cells, partly through the induction of ENaC translocation to the apical membrane. Blazer-Yost et al. indicated in 1998 that insulin activates ENaC (Blazer-Yost et al., 1998). An increase in the abundance of ENaC subunits has been reported among the plasma membrane proteins of kidneys isolated from insulin-treated mice, indicating that acute insulin stimulation causes ENaC trafficking to the apical membrane. However, the mechanisms of insulin-mediated regulation of ENaC expression and trafficking are still poorly characterized. How insulin and its downstream kinases participate in ENaC activation remains to be defined. Here, we provide evidence that FoxO1 is a key negative regulatory factor in the insulin-dependent control of ENaC expression and forward trafficking in mCCD epithelia.

FoxO1 has been recently demonstrated to suppress Na⁺ channel activity (Cai et al., 2014), but the mechanism of its effect on ENaC function has not been investigated. In the current study, we employed a combination of gain-of-function and loss-of-function approaches to determine whether FoxO1 participates in the regulation of ENaC. Our data support the concept that FoxO1, as a negative regulator, inhibits the expression of α-ENaC at the mRNA and subsequent protein levels in mCCD cells. Under basal conditions, the overexpression of ADA-FoxO1 suppressed α-ENaC mRNA products and apical α-ENaC density in mCCD epithelia, which indicates a physical role for FoxO1 in ENaC regulation because ADA-FoxO1 cannot be phosphorylated and, subsequently, cannot respond to insulin signaling. Treatment with insulin, however, was found to induce α-ENaC expression with a time course that is similar to that of phosphorylated FoxO1 and phosphorylated Akt. Moreover, steroid stimulation resulted in increased phosphorylation of FoxO1 in polarized mCCD epithelia, resulting in its exclusion from the nucleus. Knockdown of FoxO1 mimicked the action of insulin, permitting α-ENaC progression to the apical surface in the absence of steroid. These findings indicate that FoxO1 has an inhibitory action on ENaC transcription and protein expression under unstimulated conditions. In addition, FoxO1 knockdown compromised the ability of insulin to increase apical α-ENaC density and Na⁺ transport, consistent with the concept that one physiological action of insulin is the suppression of FoxO1 through Akt-mediated phosphorylation of FoxO1 and nuclear exclusion, which permits ENaC expression and membrane trafficking.

As transcription factors, Foxo proteins target either a conserved DNA binding sequence, 5′-GTAAA(C/T)A-3′ (Furuyama et al., 2000) or insulin response element (IRE), 5′-CAAAA(C/T)A-3′ (Ayala et al., 1999), in the promoter regions of their target genes and subsequently regulate the expression of certain genes. By analyzing of the sequence of the α-ENaC promoter region, we found a DNA sequence resembling that of an IRE (5′-CAAAACA-3′), located within the −2000 nt to −1000 nt region of the promoter. FoxO1 usually upregulates target gene expression through interaction with the target promoter (Kamagate et al., 2008; Philip-Couderc et al., 2008); however, negative regulation by FoxO1 has been reported for some target genes, which implies that FoxO1 has different effects on different downstream factors. FoxO1 inhibits basal transcription and gonadotropin-releasing hormone (GnRH)-mediated induction of the gene encoding luteinizing hormone β (Lhb) in pituitary gonadotrope cells. This response is likely to occur through protein–protein interactions between the FoxO1 DNA-binding domain and transcription factors and/or cofactors recruited to the proximal Lhb promoter (Arriola et al., 2012). In another study, FoxO1 is reported to downregulate Na_V1.5 expression and consequent cardiac Na⁺ channel activity through direct binding to the IRE in the SCN5A promoter region and suppressing its transcriptional activity (Mao et al., 2012).

In the current study, FoxO1 negatively regulated α-ENaC expression and function. Data from characterization of the FoxO1 target region confined the binding site to a small region (−500 to −200 nt) in the α-ENaC promoter. The results presented here suggest that FoxO1 inhibits α-ENaC promoter activity by directly binding to its promoter region. This inhibition decreases α-ENaC expression and Na⁺ transport. Further investigation is needed to identify the FoxO1 binding site in the α-ENaC promoter because this −500 to −200 nt region was not located in the assumed binding region of FoxO1 in the α-ENaC promoter, based on promoter analysis.

Our earlier work, as well as that of others, has shown that 14-3-3 proteins are essential components of ENaC regulation by aldosterone (Bhalla et al., 2005; Liang et al., 2006). The interaction between 14-3-3 proteins and their substrates is usually triggered by the phosphorylation of their targets at specific Ser/Thr residues (Yaffe, 2002), and in this manner, the 14-3-3 proteins stabilize key regulators in numerous signaling pathways. With two binding sites for protein interactions, 14-3-3 proteins are able to form dimers with their counterpart proteins. This structure enables them to bring different regions of the same protein, or two different proteins, into proximity (Aitken et al., 2002; Wilker and Yaffe, 2004). Structural studies also suggest that their high helix content make 14-3-3 dimers especially rigid, allowing them to physically deform their binding partners. Indeed, most 14-3-3 interactions are thought to be intramolecular. Thus, according to the ‘molecular anvil’ hypothesis, the formation of a dimer of 14-3-3 with Nedd4-2 might prevent 14-3-3 from interacting with ENaC (Yaffe, 2002).

The 14-3-3 proteins perhaps require selectivity of isoforms in different regulatory systems. We have defined previously that the seven known mammalian 14-3-3 isoforms are expressed in renal cells, and showed that the expression of 14-3-3β and 14-3-3ε are selectively induced by physiological levels of aldosterone (Liang et al., 2006). The substrate specificity of 14-3-3 interactions might arise from the conformation at the dimer interface (Chaudhri et al., 2003), so the discovery that the aldosterone-induced isoforms, 14-3-3β and 14-3-3ε, form an obligate heterodimer owing to the unique salt bridges that they form is particularly interesting (Liang et al., 2008). The physiological impact of these proteins in response to aldosterone was revealed in our previous work through knockdown of their expression, which markedly decreases the Na⁺ transport evoked by the steroid. In fact, a 50% reduction in either 14-3-3β or 14-3-3ε virtually eliminates the aldosterone response. We interpret this unexpected finding as indicating that the 14-3-3 proteins target key regulators, other than Nedd4-2, in the ENaC trafficking pathway. Previously, we have used affinity capture methods to identify AS160, an Akt/PKB phosphorylation substrate with the molecular signature of a Rab GTPase-activating protein (Rab-GAP), as a 14-3-3 binding protein in aldosterone-stimulated mCCD epithelia (Liang et al., 2010).

In the present paper, the insulin-dependent interaction of phosphorylated FoxO1 with 14-3-3ε was very similar to that observed for Nedd4-2 and AS160 induced by aldosterone in our previous work (Liang et al., 2010, 2008, 2006). Specifically, our antibody competition experiments showed that the increased apical α-ENaC density and the binding of phosphorylated FoxO1 to 14-3-3ε were substantially blunted in mCCD cells that had been pre-incubated with a 14-3-3ε-targeting antibody. These findings suggest that phosphorylation-dependent inactivation of FoxO1 is stabilized by binding to 14-3-3 proteins. Accordingly, in addition to its effect on α-ENaC transcription in the absence of FoxO1 inhibition, insulin enhances Na⁺ absorption and increases ENaC levels by stabilizing channels at the apical membrane.

Taken together, our data for the first time indicate that FoxO1 plays an important role in the regulation of ENaC, and these findings are consistent with a model (Fig. 7) in which the non-phosphorylated FoxO1 under basal conditions, located in the nucleus, physically inhibits ENaC transcription by binding to the ENaC promoter. Insulin stimulation of Na⁺ transport, associated with increased apical ENaC density, evokes the transcriptional induction of Akt, which phosphorylates FoxO1, thereby blocking its activity of transcriptional suppression of ENaC and permitting the translocation of phosphorylated FoxO1 to the cytoplasm. This relieves the FoxO1-mediated suppression of ENaC expression and forward trafficking to the apical surface. Selective binding of the insulin-induced 14-3-3ε to FoxO1 stabilizes FoxO1 in its phosphorylated form, thereby increasing the apical ENaC density and Na⁺ transport.

The antibodies specific for FoxO1, phosphorylated FoxO1 (Ser256), Akt1, phosphorylated Akt (Ser473), IRβ and phosphorylated IRβ were purchased from Cell Signaling Technology (Danvers, MA). The following rabbit anti-14-3-3 isoform antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA): β (C-20), γ (C-16), ε (T-16), θ (C-17) and ζ (C-16). The polyclonal antibody targeting an epitope at the extracellular loop of α-ENaC has been described previously (Liang et al., 2006). The mouse anti-β-actin antibody was purchase from Sigma-Aldrich (St Louis, MO). Other antibodies used included rabbit IgG and secondary antibodies against mouse or rabbit (Thermo Fisher, Waltham, MA).

The mCCD cells were a gift from A. Vandewalle and M. Bens (INSERM, Paris, France). The cells were seeded in plastic flasks in modified medium as previously described (Liang et al., 2010, 2006). The modified medium comprised equal volumes of Dulbecco's modified Eagle's medium (DMEM) and Ham's F12 containing 60 nM Na⁺ selenate, 5 mg/ml transferrin, 2 mM glutamine, 50 nM dexamethasone, 1 nM tri-iodothyronine, 10 ng/ml epidermal growth factor, 5 µg/ml insulin, 20 mM d-glucose, 2% FBS and 20 mM HEPES (reagents from Invitrogen and Sigma-Aldrich), pH 7.4. Cells were cultured at 37°C in 5% CO₂, 95% air atmosphere, and the media were changed every second day. The HEK293 cells were cultured in DMEM supplemented with 10% FBS and antibiotics (50 U/ml penicillin and 50 µg/ml streptomycin).

For electrophysiological experiments, mCCD cells were subcultured on a permeable filter (0.4-µm pore size, 0.33-cm² surface area; Transwell, Corning, Lowell, MA) in 24-well plates. They were then polarized after 7–10 days as assessed using ‘chopstick’ electrodes (Millipore). The open-circuit voltage was typically ∼50 mV and the transepithelial resistance was∼2 kΩ cm². The medium was replaced with a basal medium (lacking serum and hormone) for 24 h before the experiment.

The mCCD cells were transduced with adenoviral vectors at defined doses of 400, 200 or 100 plaque-forming units (pfu) per cell. The adenoviral vectors used in this study were as follows: Adv-CMV-FoxO1-ADA (ADA-FoxO1) expressing a constitutively active FoxO1-ADA allele (1.0×10¹¹ pfu/ml); and the null adenovirus, Adv-null (1.25×10¹¹ pfu/ml). Generation of the adenoviral constructs has been described previously (Altomonte et al., 2004). Those adenoviruses were a gift from Dr Henry Dong (University of Pittsburgh, Pittsburgh, PA). All adenoviral vectors were produced in HEK293 cells and purified as described previously (Kamagate et al., 2008).

The scrambled siRNA and siRNA against mouse FoxO1 (targeting the sequence 5′-GCGGGCUGGAAGAAUUCAAdtdt-3′) were commercially obtained from RIBOBIO. In vitro transfections of the mCCD cells with FoxO1 or scrambled siRNAs were performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. In brief, mCCD cells were seeded onto filters and then transfected with siRNAs when they were approximately 90% confluent. A total of 50 nM of siRNA was used for 5×10⁵ cells in 2 ml of culture medium. One day after transfection, cells were subcultured onto filter supports, where they polarized after 7–8 days, as detected with electrophysiological experiments. After the measurements, the cells were harvested from the filter supports to determine protein expression.

Total RNA was isolated from mCCD cells using Trizol reagent as per the manufacturer's protocol. Reverse transcription of 500 ng total RNA was performed using a kit (Takara). The mRNA expression of these genes was determined using real-time PCR amplification and detected using ABI Plus One Step (BioRad). The primer sequences used were: FoxO1 forward 5′-CCCAGGCCGGAGTTTAACC-3′ and reverse 5′-CCTTGTCCATCTGCATGCA-3′; α-ENaC forward 5′-CCTTCTCCTTGGATAGCCTGG -3′ and reverse 5′-CAGACGGCCATCTTGAGTAGC-3′; β-actin forward 5′-GCAAGTGCTTCTAGGCGGAC-3′ and reverse 5′-AAGAAAGGGTGTAAAACGCAGC-3′. All primers are commercially available (Invitrogen).

The use of equivalent amounts of protein for western blot analysis and immunoprecipitation was ensured by use of protein assays (BCA; Pierce). Precleared mCCD cells lysates (∼500 µg of protein) were incubated with the appropriate primary antibodies overnight at 4°C in lysis buffer (0.4% deoxycholate acid, 1% NP-40, 50 mM EDTA, 10 mM Tris-HCl at pH 7.4). Washed Protein-A–agarose beads (50 µl) were added to each sample and incubated for 3 h at 4°C with gentle rotation. Immunocomplexes were washed with lysis buffer four times and precipitated by centrifuging at 12,000 g for 5 min. The immunocomplexes were resuspended in SDS sample buffer and subjected to immunoblotting (see below). Controls for the immunoprecipitations were performed using IgG antibody at amounts equal to those of the primary precipitating antibody.

Equal amounts of protein from polarized mCCD cells, or the immunoprecipitates described above were resolved by 10% SDS-PAGE and transferred to PVDF membranes. Unbound sites were blocked with 5% non-fat milk in TBS with 0.1% Tween-20 (TBST) for 1.5 h at room temperature. The blots were incubated overnight with primary antibody (anti-α-ENaC, 1:2000; anti-FoxO1, 1:1000; anti-phosphorylated FoxO1, 1:2000; anti-Akt, 1:1000; anti-phosphorylated Akt, 1:1000; anti-14-3-3β, 1:2000; anti-14-3-3γ, 1:2000; anti-14-3-3ε, 1:2000; anti-14-3-3θ, 1:2000; anti-14-3-3ζ, 1:2000; or anti-β-actin, 1:5000) at 4°C. The blots were then washed three times for 5 min each with TBST and incubated at room temperature for 1 h with horseradish peroxidase (HRP)-conjugated secondary antibody. The blots were then washed three times (10 min each) with TBST. The reactive bands were detected by using enhanced chemiluminescence (PerkinElmer Life Sciences, Wellesley, MA) and exposed to X-ray film (Eastman Kodak Co.). Western blot data were scanned, and band densities were quantified using ImageLab software.

The mCCD cells that had been cultured on filter supports were transfected with siRNA for 48 h or transduced with adenoviruses for 24 h. The cells were then washed five times with ice-cold PBS (containing Mg²⁺ and Ca²⁺) on ice to remove the medium, and then incubated for 20 min with 0.5 mg/ml SS-Biotin (Thermo Fisher, Waltham, MA) in borate buffer (85 mM NaCl, 4 mM KCl, 15 mM Na₂B₄O₇, pH 9) at 4°C. The reaction was stopped by adding a double volume of FBS-containing medium. Monolayer cells were washed five times with ice-cold PBS, and the cells were harvested by scraping. A cell homogenate was obtained by lysing cells in lysis buffer (0.4% deoxycholic acid, 1% NP-40, 50 mM EGTA and 10 mM Tris-HCl pH 7.4) for 10 min, and then the lysate was centrifuged at 14,000 g for 5 min at 4°C. The supernatant was collected to assay for protein concentration. For the biotinylated sample, 200 µg of protein was mixed with streptavidin (Thermo Fisher) and incubated overnight at 4°C with gentle rotation. Samples from the streptavidin beads were centrifuged at 12,000 g for 3 min. The beads containing biotinylated proteins were washed three times with lysis buffer. The beads were incubated with 4× sample buffer containing 10% β-mercaptoethanol for 20 min at room temperature. Finally, samples were heated at 95°C for 7 min, separated by SDS-PAGE and blotted as described above to determine the density of ENaC at the apical membrane surface of mCCD cells.

The mCCD cells were grown on filter supports until a high-resistance monolayer was obtained. The filter supports were mounted in modified Ussing chambers (Costar) and continuously short circuited with an automatic voltage clamp, as previously described (Liang et al., 2010). Transepithelial resistance was calculated using Ohm's law from the current response to a periodic 2.5-mV bipolar pulse. The bathing solution buffer comprised 120 mm NaCl, 25 mm NaHCO₃, 3.3 mm KH₂PO₄, 0.8 mm K₂HPO₄, 1.2 mm MgCl₂, 1.2 mm CaCl₂ and 10 mm glucose. The chambers were maintained at 37°C and constantly gassed with a mixture of 95% O₂, 5% CO₂, which maintained the pH at 7.4. 10 µM Amiloride was added to the apical bath to determine ENaC-mediated transepithelial currents.

The mCCD cells cultured on filter supports, or HEK293 cells were fixed with cell stationary liquid (Beyotime) for 20 min at room temperature, followed by permeabilization with 0.3% Triton X-100 in PBS for 5 min. After rinsing with PBS, cells were incubated with 5% goat serum in PBS for 1 h. Cells were then incubated with specific antibodies against α-ENaC (1:200) overnight at 4°C, and reacted with cyanine FITC-conjugated donkey anti-rabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) at 37°C for 1 h. Stained mCCD cells were mounted with Vectashield mounting medium (Vector Laboratories, Burlingame, CA) and viewed with a Carl Zeiss LSM710 microscope equipped with a digital camera (Melville, NY).

Pro-Ject™ Protein Transfection Reagent Kit (Thermo Scientific) was used to deliver the 14-3-3ε-targeting antibody into mCCD cells according to the manufacturer's instructions. In brief, 5 µg of the anti-14-3-3ε antibody was diluted in PBS buffer (20 mM sodium phosphate, 150 mM NaCl, pH 7.4) and hydrated with prepared Pro-Ject™ Reagent at room temperature for 5 min. For a well of a 6-well plate, 1 ml of serum-free medium was added to the Pro-Ject™-Reagent–protein complex and added to the cells.

A 2000-bp DNA fragment containing the mouse α-ENaC promoter (pENaC2000) was amplified from C57BL/6J mouse genomic DNA by using PCR with the primers forward 5′-CGAGCTCTTACGCGTGCTAGCAGCAGCATAGGGGTGGCTG-3′ and reverse 5′-CAGTACCGGAATGCCAAGCTTGGTTCCTTTCCAGTTAAATC-3′. After verifying its nucleotide sequence, the α-ENaC promoter was cloned into the luciferase reporter pGL3-Basic vector (Promega). The promoter variants were generated by amplifying DNA fragments covering different lengths of the α-ENaC promoter by using PCR with primers. The forward primers used for all of the α-ENaC promoter variants were the same, and the reverse primers used were as follows: 5′-CAGTACCGGAATGCCAAGCTTGGTTCCTTTCCAGTTAAATC-3′ for pENaC1500; 5′-CAGTACCGGAATGCCAAGCTTGGTTCCTTTCCAGTTAAATC-3′ for pENaC1000; 5′-CAGTACCGGAATGCCAAGCTTGGTTCCTTTCCAGTTAAATC-3′ for pENaC500; 5′-CAGTACCGGAATGCCAAGCTTGGTTCCTTTCCAGTTAAATC-3′ for pENaC200.

The plasmids of different α-ENaC promoter variants (2 μg) were co-transfected with PGL4.75 encoding the Renilla luciferase gene (Promega) into HEK293 using Lipofectamine 2000 (Invitrogen), followed by infection with the Adv-FoxO1 or control Adv-null vector. After 24 h of incubation, the cells were subjected to a dual-luciferase activity assay to determine α-ENaC promoter activity.

Data were acquired from experiments performed two or three times. Mean±s.e.m. values were calculated using GraphPad Prism5 (GraphPad Software, San Diego, CA). Statistical significance between two groups was determined using Student's t-test. Curves were obtained using ANOVA. Significance was defined as P<0.05.

We thank Dr Henry Dong (University of Pittsburgh) and Dr Yongjian Liu (Nanjing Medical University) for their critical suggestions and proofreading of this manuscript.