A mutation of the epithelial sodium channel associated with atypical cystic fibrosis increases channel open probability and reduces Na⁺ self inhibition

Abstract

Increased activity of the epithelial sodium channel (ENaC) in the respiratory airways contributes to the pathophysiology of cystic fibrosis (CF), a genetic disease caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene. In some patients suffering from atypical CF a mutation can be identified in only one CFTR allele. We recently identified in this group of CF patients a heterozygous mutation (W493R) in the α-subunit of ENaC. Here, we investigate the functional effects of this mutation by expressing wild-type αβγENaC or mutant α_W493RβγENaC in Xenopus oocytes. The αW493R mutation stimulated amiloride-sensitive whole-cell currents (ΔI_ami) by ∼4-fold without altering the single-channel conductance or surface expression of ENaC. As these data suggest that the open probability (P_o) of the mutant channel is increased, we investigated the proteolytic activation of ENaC by chymotrypsin. Single-channel recordings revealed that chymotrypsin activated near-silent channels in outside-out membrane patches from oocytes expressing wild-type ENaC, but not in membrane patches from oocytes expressing the mutant channel. In addition, the αW493R mutation abolished Na⁺ self inhibition of ENaC, which might also contribute to its gain-of-function effects. We conclude that the αW493R mutation promotes constitutive activation of ENaC by reducing the inhibitory effect of extracellular Na⁺ and decreasing the pool of near-silent channels. The resulting gain-of-function phenotype of the mutant channel might contribute to the pathophysiology of CF in patients carrying this mutation.

Introduction

The epithelial sodium channel (ENaC) is a member of the ENaC/degenerin family of non-voltage gated ion channels. It is found in the apical membrane of sodium-absorbing epithelial cells e.g. in the respiratory tract, distal nephron, distal colon, sweat and salivary ducts. In these epithelia, ENaC is the rate-limiting step of sodium absorption (Garty & Palmer, 1997; Kellenberger & Schild, 2002). In the lung, fluid absorption critically depends on ENaC function (Hummler et al. 1996). ENaC forms a heteromeric channel composed of three homologous subunits α, β and γ (Canessa et al. 1994). Each subunit has two transmembrane domains, a large extracellular loop and cytosolic N- and C-termini. The recently published crystal structure of the related acid-sensing ion channel ASIC1 suggests that ENaC is probably a heterotrimer (Canessa, 2007; Jasti et al. 2007; Stockand et al. 2008).

The genetic disease cystic fibrosis (CF) is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene (Riordan et al. 1989). The CFTR gene encodes a Cl⁻ channel found in the apical membrane of many epithelial cells (Sheppard & Welsh, 1999). Thus, CF epithelia have a reduced apical Cl⁻ conductance. Some patients may have all the classical symptoms of CF, while others have milder or even atypical disease manifestations. The CF phenotype is a continuum of symptoms and cannot be easily defined in two distinct disease categories (typical versus atypical CF) (De Boeck et al. 2006). The large variability of CF symptoms suggests that CF modifier genes exist. Interestingly, the genes encoding the β- and γ-subunits of ENaC have been reported to be potential modifier genes in CF (Stanke et al. 2006). Moreover, missense mutations in the β-subunit of ENaC causing abnormal channel function have been identified in two patients with CF-like syndromes without CFTR mutations (Sheridan et al. 2005). In addition, the tumour necrosis factor α (TNFα) receptor located next to the gene encoding the α-subunit of ENaC has been suggested as a potential modifier gene (Stanke et al. 2006).

In respiratory epithelia, a fine balance between Cl⁻ secretion via CFTR and Na⁺ absorption via ENaC is necessary to maintain an appropriate airway surface liquid (ASL) volume which is important for pulmonary mucus clearance. CF airway epithelia are characterized by a combined defect of accelerated Na⁺ transport and the failure to secrete Cl⁻ (Boucher, 2004, 2007). These abnormal ion transport properties are thought to cause ASL volume depletion leading to ‘thickened’ mucus plaques and plugs which become the nidus for bacterial infections.

In some patients with CF(-like) disease a mutation cannot be identified on both CFTR alleles. Mutations in the genes encoding ENaC may potentially explain disease in these atypical CF patients. Indeed, sodium hyper-absorption through ENaC is thought to contribute to CF pathophysiology (Boucher, 2004, 2007) and mice overexpressing βENaC in the airways present with CF-like pulmonary disease (Mall et al. 2004; Zhou et al. 2008). In a recent study, we investigated the possibility that atypical CF patients might harbour mutations in their ENaC genes (Azad et al. 2009). αW493R-ENaC, one of the identified mutations, stimulated ENaC currents when expressed in Xenopus oocytes. The size of this gain-of-function effect was similar to that described for an hereditary form of salt-sensitive arterial hypertension (Rossier & Schild, 2008). The latter mutations affect a PY-motif in the C-termini of the β- or γ-subunit of ENaC. The PY-motif is critical for the interaction of the channel with the regulatory protein Nedd4-2 which promotes channel internalization and degradation. Mutations affecting the PY-motifs lead to higher channel activity at the plasma membrane due to a rise in cell surface expression and an impaired Na⁺-dependent feedback inhibition of the mutant channel (Staub & Verrey, 2005). In contrast to the mutations causing Liddle's syndrome, the αW493R mutation is localized in the extracellular loop of the α-subunit. Thus, the gain-of-function effect of this mutation is likely to be mediated by a different mechanism. The aim of the present study was to investigate the mechanisms underlying the gain-of-function phenotype of the mutant channel.

Methods

Plasmids

Full length cDNAs for the wild-type α-, β- and γ-subunits and for the mutant αW493R subunit of human ENaC were in pcDNA3.1. The T663 polymorphism was present in the α-subunit, unless stated otherwise. Linearized plasmids were used as templates for cRNA synthesis using either T7 (αβγ-hENaC, αW493R-hENaC and αβγ-rENaC) or SP6 (β-FLAG-rENaC) RNA polymerases (mMessage mMachine, Ambion, Austin, TX, USA). To minimize the risk of expression artefacts that may arise from differences in cRNA quality, cRNAs for wild-type and mutant ENaC were synthesized in parallel and the experiments were performed using at least two different batches of cRNA. The αW493R, αW493A, αW493C and αW493E mutants were generated by site-directed mutagenesis (QuikChange Site-Directed Mutagenesis Kit, Stratagene, Amsterdam, the Netherlands). Mutations were confirmed by sequence analysis.

Isolation of oocytes and injection of cRNA

Oocytes were obtained from adult female Xenopus laevis in accordance with the principles of German legislation, with approval by the animal welfare officer for the University of Erlangen-Nürnberg, and under the governance of the state veterinary health inspectorate (permit no. 621-2531.32-5/02). The authors confirm that the experiments complied with the ethical guidelines of The Journal of Physiology (Drummond, 2009). Animals were anaesthetized in 0.2% MS222 (Sigma, Taufkirchen, Germany) and ovarian lobes were obtained through a small abdominal incision. After suture the animals were allowed to recover fully in a separate tank before returning to the frog colony 1 day later. A minimum of 8 weeks was obligatory before next surgery on the same animal. Oocytes were isolated from the ovarian lobes by enzymatic digestion at 19°C for 3–4 h with 600–700 U ml⁻¹ type 2 collagenase from Cl. histolyticum (CLS 2, Worthington, Lakewood, NJ, USA) dissolved in a solution containing (in mm) NaCl 82.5, KCl 2, MgCl₂ 1, and Hepes 1 (pH 7.4 with Tris). Defolliculated stage V–VI oocytes were injected (Nanoject II automatic injector, Drummond, Broomall, PA, USA) with 0.5 ng cRNA per ENaC subunit, unless stated otherwise. The cRNAs were dissolved in RNase-free water and the total volume injected was 46 nl. Injected oocytes were stored at 19°C in high Na⁺ ND96 (NaCl 96, KCl 2, CaCl₂ 1.8, MgCl₂ 1, Hepes 5, pH 7.4 with Tris) or low Na⁺ solution (NMDG (N-methyl-d-glucamine)-Cl 87, NaCl 9, KCl 2, CaCl₂ 1.8, MgCl₂ 1, Hepes 5, pH 7.4 with Tris) supplemented with 100 units ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin.

Two-electrode voltage-clamp

Oocytes were routinely studied 1–2 days after injection using the two-electrode voltage-clamp technique (TEVC) as described previously (Rauh et al. 2006; Kraus et al. 2007; Wielpütz et al. 2007). The oocytes were placed in a small experimental chamber and constantly superfused at room temperature with ND96 supplemented with 2 μm amiloride (Sigma) at a rate of 2–3 ml min⁻¹. Bath solution exchanges were controlled by a magnetic valve system (ALA BPS-8) in combination with a TIB14 interface (HEKA, Lambrecht, Germany). Voltage-clamp experiments were performed using an OC-725C amplifier (Warner Instruments Corp., Hamden, CT, USA) interfaced via a LIH-1600 (HEKA) to a PC with PULSE 8.67 software (HEKA) for data acquisition and analysis. Oocytes were voltage-clamped at −60 mV and amiloride-sensitive whole-cell currents (ΔI_ami) were obtained by washing out amiloride with amiloride-free ND96 and subtracting the whole-cell currents measured in the presence of amiloride (2 μm) from the corresponding whole-cell currents recorded in the absence of amiloride. To determine the concentration dependence of the amiloride effect, oocytes were continuously clamped at −60 mV and superfused with ND96 containing amiloride in increasing concentrations of 0.01, 0.1, 1, 10 and 100 μm. Current values were determined after a stable current plateau was reached. For the evaluation of the concentration–response curve the initial current value in the absence of amiloride was set to 0 and the minimum current value measured in the presence of 100 μm amiloride was set to 1. The IC₅₀ value of amiloride and the Hill slope were determined by fitting the individual normalized current values to the Hill equation. For the detection of Na⁺ self inhibition, oocytes were continuously clamped at −100 mV and superfused at a rate of 8–10 ml min⁻¹, initially with a 1 mm Na⁺ solution (in mm: NMDG-Cl 109, NaCl 1, KCl 2, CaCl₂ 1.8, MgCl₂ 1, Hepes 5, pH 7.4 with Tris) followed by a rapid change to a 110 mm Na⁺ solution (NaCl 110, KCl 2, CaCl₂ 1.8, MgCl₂ 1, Hepes 5, pH 7.4 with Tris). To quantify Na⁺ self inhibition, we used the model described by Chraibi & Horisberger (2002) in which the current I_t is given as a function of time by

(1)

where k_a is the activation rate constant, k_i is the inactivation rate constant, k_down is the rundown rate constant and I_max is the maximum current at t= 0. Rate constants were determined with eqn (1) and the non-linear regression function of GraphPad Prism 4.03 for Windows (GraphPad Software, Inc., San Diego, CA, USA), starting from the inflection point of the current declining phase of each trace (Chraibi & Horisberger, 2002). I_max was determined by extrapolation to t= 0.

Patch-clamp

Single-channel recordings in outside-out membrane patches were performed as described previously using oocytes kept in low sodium modified Barth's saline after cRNA injection to prevent Na⁺ overloading (Diakov & Korbmacher, 2004; Diakov et al. 2008). Patch pipettes were pulled from borosilicate glass capillaries (GC150–15, I.D. 0.86 mm, O.D. 1.5 mm, Harvard Apparatus Ltd, Edenbridge, UK) using a DMZ-Universal puller (Zeitz Instrumente, Munich, Germany) and had a tip diameter of about 1 μm after fire polishing. 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 7.28 with Tris). Seals were routinely formed in a low sodium NMDG-Cl bath solution (in mm: NMDG-Cl 95, NaCl 1, KCl 4, MgCl₂ 1, CaCl₂ 1 and Hepes 10, pH 7.4 with Tris). In this bath solution the pipette resistance averaged about 10 MΩ. In the NaCl and LiCl bath solutions, 95 mm NMDG-Cl was replaced with either 95 mm NaCl or 95 mm LiCl. Downward current deflections correspond to cell membrane inward currents, i.e. the movement of positive charge from the extra- to the intra-cellular side of the membrane. Membrane patches were voltage-clamped at −70 mV, close to the calculated reversal potential of Cl⁻ (E_Cl=−77.2 mV) and K⁺ (E_K=−79.4 mV) under our experimental conditions; experiments were performed at room temperature. Single-channel currents were filtered at 500 Hz and digitized at 2 kHz before refiltering at 16 Hz to calculate single-channel current amplitude and channel activity. The latter was derived from binned amplitude histograms as the product NP_o, where N is the number of channels and P_o is the open probability (Korbmacher et al. 1995; Diakov & Korbmacher, 2004; Diakov et al. 2008). Continuous current traces of 60–130 s were selected from different experimental periods to analyse changes in NP_o.

Surface labelling

ENaC surface expression was determined using a chemiluminescence assay and a βENaC construct with a FLAG tag inserted into its extracellular loop (Firsov et al. 1996) as described previously (Zerangue et al. 1999; Konstas et al. 2001; Rauh et al. 2006; Wielputz et al. 2007). All steps were performed on ice and no glassware was used. Unspecific binding sites were blocked by 30 min incubation in ND96 supplemented with 1% bovine serum albumin (BSA, Sigma). Then oocytes were incubated for 1 h in primary mouse anti-FLAG M2 monoclonal antibody (Sigma), washed 6 × 3 min in ND96 + BSA, incubated for 45 min in secondary peroxidase-conjugated sheep anti-mouse IgG (Chemicon, Boronia Victoria, Australia), washed 6 × 3 min in ND96 + BSA and finally 6 × 3 min in ND96. Individual oocytes were placed in a white U-bottom 96-well plate and 50 μl of SuperSignal ELISA Femto Maximum Sensitivity Substrate (Pierce, Rockford, IL, USA) were added to each oocyte. Chemiluminescence was quantified with a Tecan GENios microplate reader (Tecan, Crailsheim, Germany). Results are given in relative light units (RLU).

Detection of ENaC cleavage products at the cell surface

Biotinylation experiments were performed essentially as described previously (Harris et al. 2007; Diakov et al. 2008) using 20 oocytes per group. All biotinylation steps were performed at 4°C. Oocytes were incubated in the biotinylation buffer (in mm: triethanolamine 10, NaCl 150, CaCl₂ 2, EZ-link sulfo-NHS-SSBiotin (Pierce) 1 mg ml⁻¹, pH 9.5) for 15 min with gentle agitation. The biotinylation reaction was stopped by washing the oocytes twice for 5 min with quench buffer (in mm: glycine 192, Tris-Cl 25, pH 7.5). Oocytes were lysed by passing them 5 times through a 27-gauge needle in lysis buffer (in mm: NaCl 500, EDTA 5, Tris-HCl 50, pH 7.4) supplemented with protease inhibitor cocktail (Complete Mini EDTA-free protease inhibitor cocktail tablets, Roche Diagnostics, Mannheim, Germany) according to the manufacturer's instruction. The lysates were centrifuged for 10 min at 1000 g. After addition of 1% Triton X-100 and 1% Igepal CA-630 (Sigma), supernatants were incubated for 20 min on ice and 100 μl of Immunopure immobilized Neutravidin beads (Pierce) were added. After overnight incubation at 4°C with overhead rotation, the tubes were centrifuged for 1 min at 1000 g. Supernatants were removed, and beads were washed 3 times with lysis buffer. A 100 μl aliquot of 2× SDS-PAGE sample buffer (Rotiload 1, Roth, Karlsruhe, Germany) was added to the beads. Samples were boiled for 5 min at 95°C and centrifuged for 3 min at 20,000 g before loading the supernatants on 10% SDS–PAGE. Monoclonal anti-V5 antibody (Invitrogen, Karlsruhe, Germany) was used at a dilution of 1:5000. Horseradish peroxidase labelled secondary sheep anti-mouse antibody (Sigma) was used at a dilution of 1:20,000. Absence of intracellular proteins was confirmed by absence of a β-actin signal. Densitometric analysis was done with ImageJ 1.38x (National Institutes of Health, Bethesda, MD, USA).

Statistical methods

Data are presented as mean ±s.e.m. N indicates the number of different batches of oocytes, n the number of individual oocytes studied. Statistical significance was assessed by the appropriate version of Student's t test (i.e. paired test for the comparison of values before and after chymotrypsin application and for densitometric analysis; unpaired test for all other comparisons) with GraphPad Prism 4.03 for Windows (GraphPad Software, Inc.).

Results

αW493R stimulates ENaC-mediated whole-cell currents in Xenopus laevis oocytes

To investigate the effects of the αW493R mutation on ENaC function, we measured macroscopic amiloride-sensitive whole-cell currents (ΔI_ami) in wild-type αβγENaC or mutant α_W493RβγENaC-expressing Xenopus laevis oocytes. Figure 1A shows representative whole-cell currents from oocytes expressing wild-type (left) and mutant ENaC (right). ΔI_ami was larger in oocytes expressing the mutant channel. With the common T663 polymorphism in the α-subunit, ΔI_ami was 3.8-fold larger in oocytes expressing αW493R compared to those expressing wild-type ENaC (Fig. 1B; T663). The stimulatory effect of the αW493R mutation was preserved (4.2-fold) when it was combined with the αT663A polymorphism (Fig. 1B; A663), a frequent variant that has been reported to increase ENaC activity in Xenopus oocytes (Tong et al. 2006). Since the αW493 residue is located 63 amino acids upstream of an important residue (αS556) for amiloride binding (Kellenberger & Schild, 2002), we also estimated the inhibitory constant (IC₅₀) of amiloride in wild-type and mutant ENaC-expressing oocytes which averaged 0.38 ± 0.03 μm and 0.16 ± 0.01 μm (P < 0.001), respectively (Fig. 1C). The Hill slope was similar in both groups (wt: 0.95 ± 0.05, mutant: 0.93 ± 0.03). The apparent minor difference in the IC₅₀ values cannot explain the larger ΔI_ami values observed in the oocytes expressing the mutant channel. Thus, other factors must be responsible for the observed ∼4-fold stimulation of ENaC whole-cell currents by the αW493R mutation.

Figure 1. αW493R mutation stimulates ENaC-mediated Na⁺ currents.

Oocytes were injected with 0.5 ng cRNA per subunit and incubated in ND96. Amiloride-sensitive whole-cell currents (ΔI_ami) were measured with the two-electrode voltage-clamp technique 2 days after cRNA injection. A, representative whole-cell current traces of a αβγENaC (wt) and a α_W493RβγENaC (αW493R)-expressing oocyte at a holding potential of −60 mV. Amiloride (2 μm) was present in the bath to specifically inhibit ENaC as indicated by the black bars. Washout of amiloride revealed a sizeable inward current component which corresponds to the ENaC-mediated Na⁺ inward current. Reapplication of amiloride returned the whole-cell current toward the initial baseline level. At the time points indicated by asterisks current responses from voltage-step pulses were removed for clarity. B, summary of similar experiments as shown in A performed in wild-type and mutant ENaC-expressing oocytes on the wild-type αT663 and the polymorphic αA663 background. To pool data from different batches of oocytes, the individual ΔI_ami values were normalized to the mean ΔI_ami value of the corresponding wild-type ENaC-expressing control group. C, average amiloride concentration–response curves measured in wild-type and mutant ENaC-expressing oocytes (9 oocytes per group). Error bars are partly covered by symbols. IC₅₀ inhibitory constants of amiloride are indicated. N indicates number of different batches of oocytes. ***P < 0.001.

Stimulatory effect of αW493R is independent of Na⁺ feedback inhibition

Since impaired Na⁺ feedback inhibition is a hallmark of the gain-of-function mutations causing Liddle's syndrome (Kellenberger et al. 1998), we tested whether impaired Na⁺ feedback inhibition also contributes to the stimulatory effect of the αW493R mutation. Oocytes heterologously expressing ENaC become severely sodium loaded when maintained in a bath solution containing a normal, i.e. high, extracellular Na⁺ concentration (Canessa et al. 1994). This activates Na⁺ feedback inhibition and promotes channel retrieval from the plasma membrane (Kellenberger et al. 1998; Konstas et al. 2002; Volk et al. 2004). Such Na⁺ feedback inhibition can be prevented by maintaining ENaC-expressing oocytes in a low Na⁺ solution. The stimulatory effect of ENaC mutations causing Liddle's syndrome is largely reduced in oocytes pre-incubated in low Na⁺ (Kellenberger et al. 1998). In contrast, as shown in Fig. 2, the stimulatory effect of the αW493R mutation was preserved in oocytes pre-incubated in low Na⁺. Pre-incubation in low Na⁺ increased ΔI_ami in wild-type and mutant ENaC-expressing oocytes to the same extent (Fig. 2A). Thus, the relative gain-of-function effect of the mutation was similar under high Na⁺ (4.4-fold) and low Na⁺ (4.1-fold) conditions (Fig. 2B). Moreover, the relative inhibitory effect of Na⁺ loading on ΔI_ami was similar in both groups of oocytes (Fig. 2C). This indicates that Na⁺ feedback inhibition is not reduced by the αW493R mutation and that impaired Na⁺ feedback inhibition does not contribute to the gain-of-function effect of the mutation.

Figure 2. The stimulatory effect of the αW493R mutation is independent of Na⁺ feedback inhibition.

αβγENaC (wt)- and α_W493RβγENaC (αW493R)-expressing oocytes were incubated in parallel in high Na⁺ and low Na⁺ solution. A, representative whole-cell current traces of individual oocytes at a holding potential of −60 mV. Amiloride (2 μm) was present in the bath as indicated by the black bars. B, summary of similar experiments as shown in A performed in oocytes from 3 different batches. C, to illustrate the relative effect of Na⁺ feedback inhibition on wild-type and mutant ENaC the individual ΔI_ami values summarized in B were normalized for each batch to the mean ΔI_ami value measured in the corresponding group of low Na⁺ oocytes. Numbers in or above columns indicate individual number of oocytes, N indicates number of different batches of oocytes. **P < 0.01, ***P < 0.001.

Surface expression is similar for wild-type and mutant ENaC

The stimulatory effect of the mutated α-subunit on ΔI_ami may be caused by an increase of the overall number of channels expressed at the cell surface. To assess channel surface expression, we inserted a FLAG epitope in the extracellular loop of the β-subunit and expressed αβ_FLAGγENaC and α_W493Rβ_FLAGγENaC in oocytes. We observed channel surface expression in oocytes expressing mutant ENaC to be similar to that in oocytes expressing wild-type ENaC. In contrast, a characteristic stimulatory effect of the mutated α-subunit on ΔI_ami was present in matched oocytes (Fig. 3). To confirm that the chemiluminescence assay can detect an increase in channel surface expression under our experimental conditions, we performed control experiments in which the amount of cRNA injected per oocyte was tripled. As expected, this increased both ΔI_ami and surface expression. We conclude that the stimulatory effect of the mutated α-subunit on ΔI_ami is not caused by an increase in ENaC surface expression.

Figure 3. The stimulatory effect of the αW493R mutation is not caused by an increase in channel surface expression.

Oocytes were injected with cRNA for αβ_FLAGγENaC (wt) or α_W493Rβ_FLAGγENaC (αW493R) and incubated for 2 days in a low Na⁺ solution. The insertion of a FLAG reporter epitope in the extracellular loop of the β-subunit enabled the detection of channel surface expression by a chemiluminescence assay (filled columns). For control experiments the amount of αβ_FLAGγENaC cRNA injected per oocyte was tripled (3× wt). ΔI_ami was determined in parallel in oocytes from the same experimental groups (open columns). To pool data from different batches of oocytes, individual current and surface expression values were normalized to the corresponding mean values of wild-type control oocytes. Numbers in columns indicate individual number of oocytes. N indicates number of different batches of oocytes. n.s., not significant. ***P < 0.001.

The αW493R mutation does not alter the single-channel conductance of ENaC in the presence of Na⁺ and Li⁺

Since the αW493R mutation did not increase ENaC surface expression, we considered the possibility that it altered the single-channel conductance of ENaC. We therefore recorded single-channel currents in outside-out membrane patches from oocytes expressing wild-type or mutant ENaC. Figure 4A demonstrates that the single-channel current amplitudes of wild-type and mutant ENaC were comparable at −70 mV, while Fig. 4B and C reveals that their I–V relationships were the same. The single-channel conductance of mutant ENaC (5.02 ± 0.08 pS; Fig. 4C) was not significantly different from that of wild-type ENaC (4.87 ± 0.14 pS; Fig. 4B) and similar to previously reported single-channel conductance values for wild-type human ENaC (Waldmann et al. 1995; Ji et al. 2004). ENaC is known to have a larger single-channel conductance for Li⁺ than for Na⁺. Figure 4 demonstrates that the single-channel current amplitude and I–V relationship of wild-type and mutant ENaC remained comparable when Na⁺ was replaced with Li⁺ in the extracellular solution. Thus, increased single-channel conductance does not explain the gain-of-function phenotype of the mutant channel.

Figure 4. The αW493R mutation does not alter the single-channel conductance of ENaC.

A, representative single-channel current recordings obtained at a holding potential of −70 mV from an outside-out patch of an oocyte expressing wild-type (upper trace) or mutant ENaC (lower trace). NaCl or LiCl were present in the bath solution as indicated above the trace by white and grey bars, respectively. B and C, in similar experiments as shown in A single-channel current voltage (I–V) values were obtained at different holding potentials (V_hold) with NaCl or LiCl in the bath solution from outside-out patches of wt (B, n= 4, N= 2) and mutant ENaC (C, n= 6, N= 2)-expressing oocytes. The I–V data were fitted using the Goldman–Hodgkin–Katz equation.

Chymotrypsin activation is reduced in mutant ENaC

Since surface expression and single-channel conductance are similar for wild-type and mutant ENaC, the overall activity of the mutant channels in the plasma membrane must be increased to explain the larger ENaC whole-cell currents in oocytes expressing the mutant channel. ENaC can be activated by extracellular proteases without an increase in channel surface expression (Chraibi et al. 1998; Hughey et al. 2007; Diakov et al. 2008; Kleyman et al. 2009; Haerteis et al. 2009; Rossier & Stutts, 2009). Therefore, we tested whether the mutation alters the responsiveness of the channel to extracellular proteases. We used chymotrypsin as a prototypical protease, which has a robust stimulatory effect on ENaC. Unlike trypsin, it does not stimulate transiently Ca²⁺-activated Cl⁻ currents in Xenopus oocytes (Durieux et al. 1994; Chraibi et al. 1998; Diakov et al. 2008). Chymotrypsin activated the steady state ΔI_ami in wild-type ENaC-expressing oocytes by ∼3-fold. In contrast, in oocytes expressing the mutant ENaC, chymotrypsin activated ΔI_ami only by ∼1.2-fold (Fig. 5). Thus, the αW493R mutation largely reduces the ability of chymotrypsin to activate ENaC. Such reduced activation by chymotrypsin supports the conclusion that the baseline activity of the mutant ENaC is increased.

Figure 5. The αW493R mutation reduces the stimulatory effect of chymotrypsin on ΔI_ami.

Oocytes were injected with cRNA (0.2 ng per subunit) for αβγENaC (wt) or α_W493RβγENaC (αW493R) and incubated for 2 days in low Na⁺. A, representative current traces demonstrating the effect of chymotrypsin (2 μg ml⁻¹) on wild-type and mutant ENaC. B, summary of ΔI_ami values determined before and after exposure to chymotrypsin in similar experiments as shown in A. C, same data as in B but with ΔI_ami values after chymotrypsin application normalized to the corresponding control value before chymotrypsin application. Numbers in columns indicate individual number of oocytes. N indicates number of different batches of oocytes. *P < 0.05, ***P < 0.001.

Chymotrypsin activates near-silent channels in outside-out patches from oocytes expressing wild-type ENaC but not in those from oocytes expressing the αW493R mutant channel

From the reduced stimulatory effect of chymotrypsin on mutant ENaC whole-cell currents we concluded that the baseline activity, i.e. the average baseline P_o, of the αW493R mutant channel is higher than that of wild-type ENaC. To complicate matters, ENaC at the plasma membrane does not constitute a homogenous channel population with a uniform P_o. Instead, ENaC populations with different P_o values exist in the plasma membrane (Firsov et al. 1996; Rossier, 2002), including a pool of so-called near-silent channels (Caldwell et al. 2004, 2005) that can be acutely activated by extracellular proteases (Chraibi et al. 1998; Diakov et al. 2008). We hypothesized that the reduced stimulatory effect of chymotrypsin on ENaC whole-cell currents may be related to a reduced pool of near-silent channels in oocytes expressing the mutant channel. To test this possibility, we investigated the effect of chymotrypsin on wild-type and mutant ENaC in outside-out membrane patches. Figure 6A and B show representative recordings of wild-type and mutant ENaC, and Fig. 6C and D quantifies the number of apparent channels and their activity. In Fig. 6A, on washout of amiloride, a single wild-type ENaC channel was active. However, application of chymotrypsin activated at least four additional wild-type ENaC channels, consistent with the recruitment of near-silent channels (Caldwell et al. 2004, 2005; Diakov et al. 2008). Following chymotrypsin treatment, the single-channel current amplitude of wild-type ENaC was unchanged at −0.38 pA. However, chymotrypsin treatment increased the NP_o of wild-type ENaC 4-fold (Fig. 6C) and the number of active channels 2.5-fold (Fig. 6D).

Figure 6. Chymotrypsin fails to activate near-silent channels in outside-out patches expressing the mutant α_W493RβγENaC.

A and B, representative single-channel current recordings obtained at a holding potential of −70 mV from an outside-out patch of oocytes expressing wild-type (A) or mutant ENaC (B). Amiloride (Ami, 2 μm) and chymotrypsin (2 μg ml⁻¹) were present in the bath solution as indicated above the trace by black and grey bars, respectively. The current level at which all channels are closed (C) was determined in the presence of amiloride. The different channel open levels are indicated by horizontal lines. Time periods used to calculate NP_o values are indicated by broken lines below the traces. C, amplitude histograms were used to calculate NP_o values from similar outside-out patch-clamp recordings as shown in A and B. D, the maximal number of apparent channel levels was determined by visual inspection of single-channel current traces as shown in A and B. C and D summarize data from n= 13 and n= 14 outside-out patches with wild-type or mutant ENaC, respectively. Please note that some open circles and lines represent 2–3 individual experiments. *P < 0.05, **P < 0.01, n.s., not significant.

A minority of outside-out patches from oocytes expressing wild-type ENaC contained only a single-channel (online Supplementary Fig. S1). These patches were particularly instructive since they illustrate the variable baseline P_o of ENaC which is a known feature of this channel (Garty & Palmer, 1997; Rossier, 2002). They confirm that proteolytic channel activation involves an increase in P_o of channels that are already active in the membrane and the recruitment of near-silent channels (Diakov et al. 2008). This dual stimulatory effect is consistent with our observation that in multi-channel patches NP_o of wild-type ENaC increased more than the number of apparent channels upon application of chymotrypsin (Fig. 6C and D).

Importantly, in patches from oocytes expressing the αW493R mutant channel we did not observe an increase in the number of apparent channel levels upon application of chymotrypsin (Fig. 6B and D). The average NP_o was ∼3-fold larger in patches from oocytes expressing the mutant channel than in patches expressing wild-type ENaC (Fig. 6C). This suggests that more mutant channels than wild-type channels are active under baseline conditions and that the population of near-silent channels that can be proteolytically activated is reduced in oocytes expressing the mutant channel. Indeed, channels with very low P_o were routinely observed in outside-out patches with wild-type but not with mutant ENaC (see Supplementary Fig. S2). As illustrated in the representative recording in Fig. 6B, chymotrypsin had a subtle effect on the gating kinetics of the mutant channel, which resulted in a small increase of NP_o from 1.41 to 1.73. On average chymotrypsin increased NP_o of mutant ENaC by ∼25% (Fig. 6C), which is in good agreement with the observed stimulatory effect of chymotrypsin on ENaC whole-cell currents in oocytes expressing the mutant channel (see Fig. 5). In the absence of an increase in the apparent number of active channels in the patch, this indicates that application of chymotrypsin slightly increased the P_o of mutant channels that were already active in the patch at the time of chymotrypsin application. Importantly, our single-channel recordings in outside-out patches demonstrate that chymotrypsin fails to recruit additional near-silent channels in patches from oocytes expressing the mutant channel (Fig. 6D). Taken together, these findings indicate that the αW493R mutation causes a constitutive activation of the channel and reduces the pool of near-silent channels that can be activated by extracellular proteases.

αW493R does not alter constitutive channel cleavage

The reduced responsiveness of the mutant channel to chymotrypsin could be the result of an increased level of channel activation by enhanced constitutive channel cleavage (Diakov et al. 2008; Rossier & Stutts, 2009). To detect cleaved channel fragments at the plasma membrane, we tagged the α-subunit of ENaC with an N-terminal HA-tag and a C-terminal V5-tag. In the biotinylated cell surface fraction we detected the ∼95 kDa full length αENaC together with an expected N-terminal ∼24 kDa fragment and the corresponding C-terminal ∼68 kDa fragment (Fig. 7A), both representing typical αENaC cleavage products (Bruns et al. 2007; Harris et al. 2007; Rossier & Stutts, 2009). Importantly, there was no significant difference in the cleavage pattern between wild-type and mutant αENaC (Fig. 7B). We also failed to detect an effect of the αW493R mutation on the constitutive cleavage of the γ-subunit (data no shown), which is thought to be important for channel activation by extracellular proteases (Bruns et al. 2007; Harris et al. 2007; Carattino et al. 2008; Diakov et al. 2008). From these experiments we conclude that the stimulatory effect of the αW493R mutation is not caused by enhanced endogenous channel cleavage.

Figure 7. The αW493R mutation does not alter endogenous channel cleavage.

Oocytes were injected with 2 ng per subunit of cRNA for wt and mutant (αW493R) ENaC and incubated in low Na⁺ for 2 days. A, surface-expressed protein was isolated by biotinylation and separated by PAGE. Full length αENaC (at ∼95 kDa) and expected cleavage products (Rossier & Stutts, 2009) of the α-subunit were detected by using an N-terminal HA-tag (left blot) or a C-terminal V5-tag (right blot). Non-injected (non inj.) oocytes served as control. B, cleavage was quantified by densitometric analysis. The N-terminal ∼24 kDa and the ∼95 kDa bands were used to determine the amount of cleaved and non-cleaved ENaC, respectively. N represents number of blots from different batches of oocytes.

Sodium self inhibition is abolished and contributes to the gain-of-function effect of the W493R mutation

In Fig. 2, we excluded the possibility that impaired Na⁺ feedback inhibition contributes to the gain-of-function effect of the αW493R mutation. An alternative mechanism to repress ENaC activity is Na⁺ self inhibition, which attenuates the P_o of ENaC within seconds when the extracellular Na⁺ concentration is acutely increased (Chraibi & Horisberger, 2002). To test whether the mutation impairs Na⁺ self inhibition, we performed experiments as illustrated in Fig. 8. The representative whole-cell current trace shown in Fig. 8A (left) demonstrates that in wild-type ENaC-expressing oocytes, changing the bath solution from a low (1 mm) to a high (110 mm) Na⁺ solution resulted in a transient inward current peak. The rapid current decay after the peak increase represents Na⁺ self inhibition (Chraibi & Horisberger, 2002). We noted that this initial peak response and the subsequent current decay were absent in oocytes expressing the mutant channel (Fig. 8A, right). Data from similar experiments as shown in Fig. 8A are summarized in Fig. 8B. These findings indicate that the mutation abolishes Na⁺ self inhibition. To investigate further how much this contributes to the gain-of-function effect of the mutation, we estimated the current component sensitive to Na⁺ self inhibition in oocytes expressing wild-type ENaC using a previously described model (Chraibi & Horisberger, 2002) (see eqn (1) in Methods). According to this model, the predicted amiloride-sensitive maximum current in the absence of Na⁺ self inhibition averaged 6.5 ± 0.7 μA, whereas steady state ΔI_ami averaged 3.0 ± 0.3 μA (n= 16). Thus, the current component sensitive to Na⁺ self inhibition is about 3.5 μA. From this we can conclude that the complete loss of Na⁺ self inhibition contributes about ∼40% to the observed increase of ΔI_ami to 11.3 ± 0.8 μA (n= 16) in oocytes expressing the mutant channel.

Figure 8. The αW493R mutation abolishes Na⁺ self inhibition.

Oocytes were injected with 0.5 ng per subunit of cRNA for αβγENaC (wt) or α_W493RβγENaC (αW493R). After 2 days of incubation in low Na⁺, oocytes were clamped at −100 mV. The bath solution initially contained 1 mm Na⁺ and was subsequently changed to a bath solution containing 110 mm Na⁺ as indicated. After 60 s, amiloride (2 μm) was applied to determine the ENaC-mediated current component. A, representative current traces of a wild-type (left) and a mutant ENaC (right)-expressing oocyte. Downward deflections represent inward currents. B, averaged current traces from similar experiments as shown in A, starting from the change to 110 mm Na⁺. Lines represent mean, error bars s.e.m. n indicate number of oocytes.

Substitution of αW493 by alanine, cysteine or glutamate also stimulates ENaC

Arginine carries a positive charge whereas aromatic tryptophan is uncharged. Therefore, the stimulatory effect of the αW493R mutation may depend on the charge of the amino acid that replaces the tryptophan residue in this position. To test this, we replaced tryptophan by the negatively charged amino acid glutamate (αW493E), by the uncharged amino acid alanine (αW493A), or by the uncharged but polar amino acid cysteine (αW493C). All three substitutions resulted in a ∼2.5-fold stimulation of ΔI_ami (Fig. 9). This stimulatory effect is smaller than that observed with the αW493R mutation. However, these results suggest that the gain-of-function effect of the mutation might not critically depend on the introduction of a positively charged arginine residue at this site, but instead that the loss of the tryptophan residue per se leads to ENaC activation. It should be noted that this tryptophan residue is highly conserved across species (see Supplementary Fig. S3A). In an additional set of experiments we demonstrated that a corresponding mutation in the α-subunit of rat ENaC (αW520R) had a similar stimulatory effect as the human αW493R mutation (see Supplementary Fig. S3B). Taken together these results confirm the functional importance of this highly conserved tryptophan residue in the α-subunit of ENaC.

Figure 9. Amino acid substitution at residue α493 stimulates ENaC.

Oocytes were injected with cRNA for αβγENaC (wt), α_W493RβγENaC (αW493R), α_W493AβγENaC (αW493A), α_W493CβγENaC (αW493C) or α_W493EβγENaC (αW493E). A, representative whole-cell current traces of individual oocytes at a holding potential of −60 mV. Amiloride (2 μm) was present in the bath as indicated by the black bars. B, summary of similar experiments as shown in A. To pool data from different batches of oocytes, individual ΔI_ami values were normalized to the mean ΔI_ami of the corresponding wild-type ENaC-expressing control oocytes. N indicates number of different batches of oocytes. Numbers in columns indicate total number of oocytes. ***P < 0.001.

Discussion

The present study demonstrates that the ENaC mutation αW493R associated with atypical CF stimulates ENaC currents when heterologously expressed in Xenopus oocytes. The magnitude of this gain-of-function effect is similar to that described for ENaC mutations causing Liddle's syndrome (Schild et al. 1995; Schild, 1996). However, the mechanisms responsible for the gain-of-function effect of the αW493R mutation are distinct from those of the Liddle's syndrome mutations. Liddle's syndrome mutations prevent ENaC retrieval from the cell surface and impair Na⁺ feedback inhibition (Rossier & Schild, 2008, 2009). By contrast, our data demonstrate that the gain-of-function effect of the αW493R mutation is caused by an increase in average channel P_o.

For ENaC the single-channel P_o values observed in patch-clamp experiments are known to be highly variable with a rather wide range and an average single-channel P_o of ∼0.5 (Waldmann et al. 1995; Garty & Palmer, 1997). However, an average P_o of ∼0.5 in the absence of an increase in channel surface expression could explain a doubling of ΔI_ami but not the ∼4-fold increase observed with the αW493R mutant channel. Thus, we have to assume that the average P_o of wild-type ENaC is much lower than ∼0.5. This could be explained by the presence of a pool of so called near-silent channels which can be activated by extracellular proteases (Caldwell et al. 2004, 2005; Diakov et al. 2008; Rossier & Stutts, 2009). The concept that a pool of near-silent channels is present in the plasma membrane is supported by our finding that additional channel levels appear after chymotrypsin application to outside-out patches from wild-type ENaC-expressing oocytes. In contrast, chymotrypsin failed to activate near-silent channels in outside-out patches with the mutant channel. This suggests that the mutation reduces the pool of near-silent channels and mimics proteolytic activation of ENaC. Interestingly, constitutive activation of the mutant channel was not associated with increased endogenous channel cleavage at the cell surface. Thus, the mutation seems to convert near-silent channels into active channels by an alternative mechanism. The residual effect of chymotrypsin on the mutant channel indicates that the mutant channel is not completely unresponsive to proteolytic activation. However, the failure of chymotrypsin to activate near-silent channels in outside-out patches from oocytes expressing the αW493R mutant channel clearly distinguishes the mutant from the wild-type channel.

The αW493R mutation introduces a positive charge (arginine) at a site where normally an uncharged aromatic amino acid (tryptophan) is present. The introduction of a negatively charged (glutamate), a polar (cysteine) or an uncharged amino acid (alanine) also resulted in a stimulation of ENaC. Therefore, the introduction of a positive charge does not appear to be essential for the stimulatory effect. Similarly, the length of the substituting amino acids does not seem to be critical for mediating the stimulatory effect, since tryptophan and arginine are of similar size, whereas alanine, cysteine and glutamate are much shorter. This suggests that the W493 residue in wild-type αENaC mediates an inhibitory effect on the channel and is essential to maintain the channel in a near-silent state. In this context our observation that the αW493R mutation leads to a loss of Na⁺ self inhibition is of interest. The loss of the tryptophan reduces the channel's ability to decrease its activity within seconds following a sudden increase of the extracellular [Na⁺]. This suggests a role of this residue in mediating Na⁺ self inhibition.

In the absence of a crystal structure for ENaC it is difficult to predict a structural effect of the W493R mutation on ENaC. However, the recently published crystal structure of the related acid-sensing ion channel ASIC1 can be used to generate a homologous ENaC model (Jasti et al. 2007; Stockand et al. 2008). Using this modelling approach, we found that W493 is located in a region of the ENaC structure which binds a monovalent cation in the ASIC channel (Gonzales et al. 2009) (Supplementary Fig. S4). The structural similarity between ASIC and ENaC suggests that this region represents an ion binding site in ENaC and may play a role as sodium sensor or as a site for sodium entry. Mutation of W493 to other amino acids will probably alter the structure and, hence, channel properties. The fact that the type of amino acid substitution appears to be rather unimportant suggests that the tryptophan side chain might itself play a key role in the structural integrity of this region.

In the oocyte expression system the overall stimulatory effect of the αW493R mutation is similar to that of mutations identified in patients with Liddle's syndrome. However, in the latter patients this increased channel function results in a severe form of salt-sensitive arterial hypertension, while to our knowledge pulmonary symptoms have not been described in patients with Liddle's syndrome. In contrast, the αW493R mutation has been identified in patients with atypical CF, but as far as we know, these patients do not suffer from arterial hypertension. What could be the reason for the different pathophysiological consequences of the αW493R mutation?

It is noteworthy that after chymotrypsin stimulation ΔI_ami reached similar levels in wild-type and in αW493R mutant ENaC-expressing oocytes. Assuming that chymotrypsin leads to a maximal stimulation of the channels in the plasma membrane, this is in good agreement with our finding that channel surface expression was similar in wild-type and mutant ENaC-expressing oocytes. Interestingly, the constitutive level of proteolytic ENaC activation appears to be high in the native kidney (Nesterov et al. 2008). Thus, in the kidney, the stimulatory effect of the αW493R mutation may be limited. Moreover, it may be compensated by an appropriate down-regulation of the renin–angiotensin–aldosterone system. In contrast, protease inhibitors are normally present in the ASL and tonically inhibit the proteolytic activation of ENaC in the lung (Myerburg et al. 2006). Thus, in the lung, the αW493R mutation, which mimics proteolytic channel activation, might have a bigger stimulatory effect on ENaC function than in the kidney. This may explain why the αW493R mutation is associated with a pulmonary rather than a renal phenotype.

It also remains an unresolved issue why certain conditions with enhanced ENaC function lead to pulmonary symptoms while others do not. A possible explanation is the different molecular mechanism by which ENaC mutations in patients with Liddle's syndrome lead to a gain-of-function effect. In patients with Liddle's syndrome, increased surface expression of ENaC may not cause pulmonary symptoms as long as ENaC activation by endogenous proteases in the ASL is maintained at a low level (Myerburg et al. 2006). Moreover, mechanisms may exist in the respiratory epithelium to compensate for the retrieval defect of ENaC with Liddle's syndrome mutation by reducing channel insertion into the plasma membrane. Finally, we cannot rule out the possibility that patients with Liddle's syndrome have a minor pulmonary phenotype which so far has escaped clinical diagnosis. Indeed, it has been reported that alveolar fluid absorption is increased and the severity of hydrostatic pulmonary oedema is reduced in a mouse model for Liddle's syndrome (Randrianarison et al. 2007). Like patients with Liddle's syndrome, these animals have no apparent pulmonary symptoms under normal conditions. In contrast, overexpression of the β-subunit of ENaC in airway epithelia leads to CF-like pulmonary symptoms in mice (Mall et al. 2004).

It has been speculated that proteolytic activation of ENaC may aggravate symptoms of cystic fibrosis during acute respiratory infections associated with the generation of local proteases, e.g. neutrophil elastase (Boucher, 2004; Caldwell et al. 2005; Harris et al. 2007). Moreover, in CF the protease–protease inhibitor balance appears to be altered which may favour proteolytic ENaC activation in the lung. Consistent with this, increased proteolytic processing of the α-subunit of ENaC has been demonstrated in human CF airway epithelial cells (Myerburg et al. 2006). Thus, the gain-of-function effect of the αW493R mutation, in particular in association with one mutated CFTR allele, may be pathophysiologically relevant in patients with atypical CF carrying this mutation (Azad et al. 2009).

The concept of sodium hyperabsorption via hyperactive ENaC as a pathophysiological basis for CF-like lung symptoms is intriguing. However, in patients with atypical CF, stimulatory as well as inhibitory ENaC mutations have been identified (Sheridan et al. 2005; Azad et al. 2009; Huber et al. 2010). CF-like pulmonary symptoms with chronic airway infections have also been described in patients with pseudohypoaldosteronism type 1 caused by loss-of-function ENaC mutations (Kerem et al. 1999). The rhinorrhea and increased mucociliary clearance observed in these patients is consistent with reduced ENaC-mediated sodium absorption in the airways (Kerem et al. 1999). These findings illustrate the complex effects that ENaC might exert in the pathophysiology of CF or CF-like pulmonary disease.

In summary, we have shown that the stimulatory effect of the αW493R mutation is based on constitutively high P_o of the mutant channel with a reduced pool of near-silent channels. Thus, the mutation functionally mimics proteolytic channel activation without changing the cleavage pattern of the mutant channel. This interpretation is supported by the reduced stimulatory effect of chymotrypsin on the mutant channel, by the failure of chymotrypsin to recruit near-silent channels in outside-out patches containing the mutant channel, and by the finding that the mutant channel lacks Na⁺ self inhibition. Thus, our results suggest an important role of the highly conserved αW493 residue for ENaC function and regulation of channel P_o. Our results also suggest that this gain-of-function mutation may contribute to the pathophysiology of atypical CF.

Glossary

Abbreviations

ASL

airway surface liquid

CF

cystic fibrosis

CFTR

cystic fibrosis transmembrane conductance regulator

E

equilibrium potential

ENaC

epithelial sodium channel

ΔI_ami

amiloride-sensitive current

P_o

open probability

Author contributions

R.R., A.K.A., H.C., J.J.C., J.D. and C.K. participated in the conception of the study; R.R., A.D., A.T. and C.K. designed the experiments; R.R., A.D., A.T. and J.K. performed the experiments and data analysis; R.R., A.D. and C.K. interpreted the data; H.S. did the modelling. R.R. and C.K. wrote the manuscript; all authors were involved in critically revising the manuscirpt and approved its final version. Experiments were performed at the Department of Cellular and Molecular Physiology of the Friedrich-Alexander University of Erlangen-Nürnberg, Erlangen, Germany.

Supplemental material

Figure S1

Figure S2

Figure S3

Figure S4

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors