Open probability of the epithelial sodium channel is regulated by intracellular sodium

Abstract

The regulation of epithelial Na⁺ channel (ENaC) activity by Na⁺ was studied in Xenopus oocytes using two-electrode voltage clamp and patch-clamp recording techniques. Here we show that amiloride-sensitive Na⁺ current (I_Na) is downregulated when ENaC-expressing cells are exposed to high extracellular [Na⁺]. The reduction in macroscopic Na⁺ current is accompanied by an increase in the concentration of intracellular Na⁺ ([Na⁺]_i) and is only slowly reversible. At the single-channel level, incubating oocytes in high-Na⁺ solution reduces open probability (P_o) approximately twofold compared to when [Na⁺] is kept low, by increasing mean channel closed times. However, increasing P_o by introducing a mutation in the β-subunit (S518C) which, in the presence of [2-(trimethylammonium) ethyl] methane thiosulfonate (MTSET), locks the channel in an open state, could not alone abolish the downregulation of macroscopic current measured with exposure to high external [Na⁺]. Inhibition of the insertion of new channels into the plasma membrane using Brefeldin A revealed that surface channel lifetime is also markedly reduced under these conditions. In channels harbouring a β-subunit mutation, R564X, associated with Liddle's syndrome, open probability in both high- and low-Na⁺ conditions is significantly higher than in wild-type channels. Increasing the P_o of these channels with an activating mutation abrogated the difference in macroscopic current observed between groups of oocytes incubated in high- and low-Na⁺ conditions. These findings demonstrate that reduction of ENaC P_o is a physiological mechanism limiting Na⁺ entry when [Na⁺]_i is high.

The physiological role of the epithelial Na⁺ channel (ENaC) has been extensively characterized in the salt-reabsorbing epithelia of the kidney, urinary bladder, intestines, sweat and salivary ducts, and the lung. In these tissues, ENaC serves to mediate Na⁺ transport across membranes, to regulate salt balance or, in the case of the lung, to maintain an appropriate level of hydration (Garty & Palmer, 1997; Kellenberger & Schild, 2002). The channel itself is composed of three subunits, α, β and γ, each of which contains two membrane-spanning (M1 and M2) regions, and relatively short N- and C-terminal intracellular regions. The channel is characterized biophysically by its slow kinetics of gating, lack of strong voltage dependence and small unitary conductance (Canessa et al. 1994; Garty & Palmer, 1997; Kellenberger & Schild, 2002; Snyder, 2002). Pharmacologically, it is identified by its sensitivity to the potassium-sparing diuretic, amiloride, and its analogues. Channels expressed in Xenopus oocytes, the most common expression system used to study ENaC behaviour, are biophysically and pharmacologically similar to those expressed endogenously in native tissue (Garty & Palmer, 1997; Kellenberger & Schild, 2002).

The epithelial Na⁺ channel regulates the entry of Na⁺ ions into cells in which it is expressed. Conversely, entry of Na⁺ into cells can itself modulate channel activity by a process generically described as feedback inhibition (Garty & Palmer, 1997; Kellenberger & Schild, 2002). The importance of cellular Na⁺ regulation in fluid homeostasis is emphasized by the findings that deletion or missense mutations of residues after the second transmembrane region of native β or γ subunits in the kidney lead to an increase in renal Na⁺ reabsorption seen in an inheritable form of hypertension called Liddle's syndrome (Shimkets et al. 1994; Hansson et al. 1995; Snyder et al. 1995; Schild et al. 1996; Snyder, 2002) and that this increase may be at least in part attributable to the suppression of feedback inhibition (Kellenberger et al. 1998).

While the general features of ENaC feedback inhibition are well characterized (e.g. reduction of whole-cell Na⁺ currents when intracellular Na⁺ levels are high), the mechanisms underlying the inhibition of Na⁺ entry are still not entirely clear. Some reports have suggested that high intracellular [Na⁺] ([Na⁺]_i) acts directly on ENaC to curtail Na⁺ movement into cells (Ishikawa et al. 1998; Awayda, 1999). There is also evidence to suggest that inhibition of ENaC activity by intracellular Na⁺ is effected via cellular mediators (Frindt et al. 1993; Silver et al. 1993; Komwatana et al. 1996; Abriel & Horisberger, 1999). When intracellular [Na⁺] is elevated, the rate of endocytic channel retrieval is enhanced (Volk et al. 2004). Channel internalization itself is stimulated by the actions of ubiquitin protein-ligases, such as Nedd4, which are purported to recognize ‘PY’ motifs present in the cytoplasmic, post-M2 region of ENaC subunits and to target them for degradation (Schild et al. 1996; Staub et al. 1997; Dinudom et al. 1998, 2001; Konstas et al. 2002; Fotia et al. 2003). When missense or deletion mutations occur, as in Liddle's syndrome, the interaction between Nedd4 and ENaC is disrupted, leading to increased surface expression of the channel and, finally, to increased absorption of Na⁺ (Schild et al. 1996; Shimkets et al. 1997; Abriel et al. 1999; Dinudom et al. 2001; Snyder, 2002). In oocytes, the increased activity of Liddle's syndrome channels resulted from relief of inhibition of the channels by high [Na⁺]_i (Kellenberger et al. 1998), again suggesting that intracellular Na⁺ may activate channel internalization through the Nedd4-dependent mechanism. However, studies comparing the activity of wild-type and Liddle's mutant ENaC in Na⁺-loaded oocytes showed that increases in amiloride-sensitive macroscopic current in the mutant were not matched by commensurate increases in surface channel expression (Firsov et al. 1996; Kellenberger et al. 1998). This suggests that at least some of the increase in whole-cell current in the mutant is due to an increase in current passed per channel in the membrane. However, direct measures of single-channel kinetics with high or low [Na⁺]_i in oocytes are lacking.

The aim of this study was to describe the effects of high and low intracellular [Na⁺] on the single-channel properties of wild-type and Liddle's mutant ENaC expressed in oocytes. Our findings support the idea that, in wild-type channels, the process of feedback inhibition relies on reducing both channel open probability and channel number to limit Na⁺ entry when [Na⁺]_i is high; in Liddle's syndrome ENaC, feedback inhibition is achieved primarily by reducing single-channel open probability.

Methods

Channel expression

The plasmids containing rat ENaC α, β and γ subunits were linearized with NotI restriction enzyme (New England Biolabs, Ipswish, MA); cRNAs were transcribed with T7 RNA polymerase using the mMESSAGE mMACHINE kit (Ambion, Austin, TX). Pellets of cRNA were dissolved in nuclease-free water and stored at −70°C before use.

Oocytes were harvested from Xenopus laevis. Animals were anaesthetized in water containing tricaine methanesulphonate (1.5 g l⁻¹, adjusted to pH 7.0). Oocytes were removed through a 1 cm incision in the abdomen. After the final collection of oocytes, animals were killed by double pithing while under anaesthesia. All animal procedures were carried out according to the guidelines of and with the approval of the Institutional Animal Use and Care Committee of Weill Medical College of Cornell University. After removal, oocytes were incubated in OR2 solution (in mm: NaCl 82.5, KCl 2.5, CaCl₂1, MgCl₂1, Na₂HPo₄1, HEPES 5 pH7.4) with 2 mg ml⁻¹ collagenase type II (Worthington) and 2 mg ml⁻¹ hyaluronidase type II (Sigma-Aldrich) and incubated with gentle shaking for 60 min and then for another 30 min (if necessary) in a fresh enzyme solution at room temperature. Before injection, oocytes were incubated in OR2 solution for 2 h at 19°C. Defolliculated oocytes were selected and injected with 8 ng cRNA of each the α, β and γ subunits in all experiments. During the expression phase, the oocytes were incubated in either a high- or low-Na⁺ modified Barth's saline (MBS), pH 7.4. The high-Na⁺ MBS contained (mm): 85 NaCl, 1 KCl, 0.7 CaCl₂l, MgCl₂l, Na₂ HPo₄l, HEPES5 pH¹ 7.4) 0.8 MgSO₄ and 5 Hepes. The low-Na⁺ MBS contained (mm): 1 NaCl, 40 KCl, 60 NMDG (N-methyl d-glucamine), 0.7 CaCl₂, 0.8 MgSO₄ and 5 Hepes. All chemicals were from Sigma-Aldrich unless otherwise noted.

Site-directed mutagenesis

Site-directed mutagenesis was performed on rat β ENaC cDNA using the Pfu enzyme (Stratagene, La Jolla, CA) according to the manufacturer's instructions. Primers were synthesized by Operon Technologies, Inc. (Huntsville, AL) To confirm that the point mutation had been successfully made, sequences of cDNAs were checked by the Cornell University Bio Resource Center (Ithaca, NY, USA).

Electrophysiology

Two-electrode voltage clamp (TEVC)

Oocytes were bathed in either a high-Na⁺ extracellular recording solution with (mm): 110 NaCl, 2 CaCl₂, 1 MgCl₂, 2 KCl and 5 Hepes, pH 7.4, or a low-Na⁺ extracellular recording solution with (mm): 5 NaCl, 105 NMDG, 2 CaCl₂, 1 MgCl₂, 2 KCl and 5 Hepes, pH 7.4. Whole-cell currents were measured in intact oocytes using a two-electrode voltage clamp (OC-725, Warner Instrument Corp.) with ITC-16 interface (Instrutech) running Pulse software (Heka Elektronik). Pipettes were made from haematocrit capillary tubes (Fisher Scientific) with a three-step vertical pipette puller (Kopf). Pipette resistances were 0.5–1 MΩ, when filled with 3 m KCl. Steady-state current–voltage curves were generated from a step-voltage protocol consisting of 15 pulses lasting 50 ms from −100 to +40 mV, from a holding potential of 0 mV. Measurements of I_Na, the amiloride-sensitive Na⁺ current (difference between Na⁺ currents obtained in the presence of 10 μm amiloride from those obtained in the absence of amiloride at −100 mV) were made 16–48 h postinjection.

Intracellular [Na⁺] was calculated using the Nernst equation:

where E_rev is the voltage at which I_Na reverses. F is the Faraday constant, R is the gas constant and T is the absolute temperature. This calculation assumes that the channels are perfectly selective for Na⁺ ions. When the renal outer medullary K⁺ channel (ROMK2) was coexpressed with ENaC, 5 mm BaCl₂ was included in the high-Na⁺ recording solution to assess Ba²⁺-sensitive current.

The effects of the methanethiosulphonate (MTS) reagent were studied by TEVC. [2-(Trimethylammonium)ethyl]methanethiosulphonate bromide (MTSET, Toronto Research Chemicals) was dissolved in the high-Na⁺ extracellular recording solution to 1 mm and bath applied to the oocyte, as previously described (Snyder et al. 2000). The percentage change in amiloride-sensitive Na⁺ current was calculated as ((I_MTS − I_basal)/I_basal) × 100, where I_MTS is the amiloride-sensitive current after treatment with MTSET and I_basal is the amiloride-sensitive current before treatment.

Brefeldin A (Sigma) was added to the MBS incubation solution at a final concentration of 5 μm.

Patch clamp

Prior to patch clamping the oocyte, the vitelline membrane was mechanically removed in a hypertonic solution containing 200 mm sucrose. The bath solution contained (mm): 110 NaCl, 2 CaCl₂, 1 MgCl₂, 2 KCl and 5 Hepes, pH 7.4. The pipette solution contained (mm): 110 LiCl, 1 MgCl₂ and 5 Hepes, at pH 7.4. Patch-clamp pipettes were prepared from Fisherbrand haematocrit capillary glass (Fisher Scientific) using a vertical puller (Kopf Instruments), coated with Sylgard® (Dow Corning) and fire polished with a microforge to yield resistances of 3–10 MΩ. Currents from patches containing one to eight channels were recorded with an EPC-7 patch-clamp amplifier (Heka Elektronik) for a duration of 1–10 min, and digitized with a Digidata 1332A interface (Axon Instruments). Data were filtered at 1 kHz and analysed with pCLAMP8 software (Axon Instruments).

Data analysis

The probability of a channel being open was calculated from the equation:

where n is the number of channels open at any given time, N is the total number of channels in the patch, and P_n, the probability that n out of N channels are open. N was estimated from the number of discrete current levels observed. This will tend to underestimate the actual number of channels, since not every level will necessarily be visited during the finite recording period. To check whether the estimate of N was reasonable, we compared the measured values of P_n with the theoretical values predicted from the binomial distribution for N channels all having the same P_o. Although this assumption is probably not strictly true, in general the agreement was satisfactory (see Fig. 5).

Figure 5. Comparison of P_n measured and P_n predicted by binomial distribution.

The probability that a channel resides at a particular open level (n) was either determined directly by dividing level dwell time by the total time of recording, or calculated from the binomial distribution as described in the text. A, comparison of measured and predicted P_n values for a recording with 4 channels in the patch, taken from an oocyte incubated in high-Na⁺ solution. B, comparison of measured and predicted P_n values for recording with 5 channels in the patch, taken from an oocyte incubated in low-Na⁺ solution.

Mean open and closed times were estimated in patches containing one or more channels as previously described (Palmer & Frindt, 1986; Frindt et al. 1993). We again assumed that the patch contained N identical channels that operated independently. We further assumed that each channel had a single open and closed state and that transition rates between these were identical for every channel in the patch. Channel opening and channel closing were denoted by the rate constants k₊ and k₋, respectively. We selected two levels with n and m channels open, respectively, that had the largest number of events. The rates of leaving this level are given by the reciprocal of the mean dwell times:

Solving the two equations gave values for the two unknowns, k₊ and k₋.

Data are presented as means ± s.e.m. Student's two-tailed t test was used to determine whether differences between groups were significant (P < 0.05) using Microcal Origin 6.0.

Results

In accordance with published reports (Firsov et al. 1996; Kellenberger et al. 1998; Konstas et al. 2002; Volk et al. 2004; Rauh et al. 2006), we observed a marked reduction in macroscopic Na⁺ current (I_Na) when oocytes injected with wild-type αβγ ENaC mRNA and incubated overnight in a low-Na⁺ MBS were placed in the high-Na⁺ recording solution (110 mm; Fig. 1A). Inhibition of ENaC current was monitored by measuring currents every 5 min and correcting for amiloride-insensitive currents measured at the end of the time course. Over 80 min, I_Na fell to ∼40% of its initial value. The reversal potential of the Na⁺ current was noted in each experiment and used to calculate the [Na⁺]_i using the Nernst equation (Fig. 1B and C). Intracellular Na⁺ was initially ∼10 mm, and rose in a sigmoidal manner to approximately 50 mm over the time course of the experiment. The effect of high [Na⁺] on oocytes expressing a mutant channel with a premature stop codon in the β-subunit, R564X, was also assessed. Despite the fact that intracellular Na⁺ levels were roughly the same in mutant and wild-type EnaC-expressing oocytes over the time course of the experiment (Fig. 1C), I_Na in the mutant stayed the same or even rose slightly (Fig. 1A). Other groups have reported similar differences in the behaviour of Liddle's mutant and wild-type ENaC in response to increasing [Na⁺]_i, presumably reflecting a physiological downregulation of channel activity, which is lacking in the mutant (Kellenberger et al. 1998). However, longer exposure to high extracellular [Na⁺] did diminish I_Na even in the Liddle's mutant channels (see below).

Figure 1. Increased [Na⁺]_i inhibits ENaC currents.

Oocytes were injected with either wild-type (αβγ; wt) or mutant (αβR564Xγ) ENaC and incubated in low-Na⁺ (1 mm) MBS overnight. The following day, Na⁺ current was measured by TEVC following the voltage-step protocol described in the Methods. A, normalized change in I_Na over 8 experiments with assessment of 2 batches of oocytes (wt, I₀ = 14 ± 3 μA; mutant, I₀ = 18 ± 3 μA). Current measurements were taken at intervals shown while oocytes were maintained in 110 mm NaCl bath solution. Amiloride (10 μm) was bath applied to the oocyte at the end of every experiment to ensure that currents measured were amiloride sensitive. After 80 min, wt currents had fallen to ∼40% of their initial value, while currents in the mutant stayed the same or increased. B, Na⁺ current reversal potentials (E_rev) were noted, and used to calculate [Na⁺]_i by the Nernst Equation. C, [Na⁺]_i rises over the course of the experiment. Error bars represent the s.e.m.

We next tried to reverse the inhibitory effect of raising [Na⁺]_i on wild-type ENaC current, by simply lowering [Na⁺]_i. This was accomplished in the following manner. Oocytes were injected with αβγ ENaC cRNA and incubated overnight in low-Na⁺ solution. The next day, as shown in Fig. 2A, they were pre-incubated in a high-Na⁺ solution for approximately 50 min. Over this time, I_Na declined to approximately 40% of its starting value. The oocytes were subsequently flushed with low-Na⁺ solution, and changes in I_Na monitored roughly every 30 min by briefly switching to the high-Na⁺ recording solution. With this protocol, even after a 60 min low-Na⁺ perfusion, I_Na did not recover to its starting value. At best, then, the downregulation of ENaC by high [Na⁺] is slowly reversible even though the change in [Na⁺]_i induced by the high-Na⁺ pre-incubation could be entirely reversed within 30 min, achieving levels between 10 and 20 mm over 65 min (Fig. 2B). To examine reversibility over longer time periods, oocytes were first exposed to a high-Na⁺ solution overnight, and I_Na measured by TEVC. These oocytes were subsequently switched to a low-Na⁺ medium, and I_Na was measured again the following day. Following the low-Na⁺ incubation, macroscopic current increased but was still lower than in those oocytes that had been incubated continuously for two nights in a low-Na⁺ solution (Fig. 2C).

Figure 2. Reversal of [Na⁺]_i does not reverse effects of high [Na⁺] on macroscopic ENaC current.

Oocytes were injected with αβγ ENaC and incubated in low-Na⁺ MBS overnight. A, normalized change in macroscopic I_Na in response to changing extracellular Na⁺ solutions is monitored by TEVC. Filled and shaded bars indicate periods of high (110 mm) and low (5 mm) bath [Na⁺], respectively. At approximately 90 and 120 mins, bath solution was briefly switched from low to high [Na⁺], and I_Na measured. B, while intracellular Na⁺ levels return to, or go below, their pre-high-Na⁺ levels, currents do not significantly recover (see partA), or recover slowly (n = 8). C, oocytes expressing ENaC were incubated overnight in high-Na⁺ solution. The next day, after amiloride-sensitive Na⁺ currents were measured, oocytes were transferred to low-Na⁺ MBS. However, Na⁺ currents were still not as high in this group (17 ± 2 μA) as in those oocytes incubated continuously for 2 days in low [Na⁺] (28 ± 3 μA; n = 19–28). *P < 0.05 compared to high-Na⁺ condition; †P < 0.05 compared to high-Na⁺, then low- Na⁺ condition.

To test whether high [Na⁺]_i specifically reduces ENaC current, ROMK2, an inward-rectifying K⁺ channel, and αβγ ENaC cRNA were co-injected into oocytes, which were then incubated overnight in a low-Na⁺ MBS. The following day, macroscopic Na⁺ and potassium currents were measured by TEVC as amiloride-sensitive and Ba²⁺-sensitive currents, respectively. After a further overnight incubation in high-Na⁺ solution, whole-cell currents were measured and normalized to the initial current (i.e. current after the first incubation). As Fig. 3 illustrates, high [Na⁺]_i specifically downregulated ENaC current, while the magnitude of ROMK2 current remained unchanged.

Figure 3. Feedback inhibition is specific for ENaC current.

Rat ROMK2 cRNA was co-injected into oocytes with αβγ ENaC cRNA, and incubated in low-Na⁺ MBS. The next day, amiloride- and Ba²⁺-sensitive currents were assessed. After a further overnight incubation in high-Na⁺ MBS, currents were again measured and normalized to current measured the previous day. While ENaC current declined over this time, ROMK current was not affected (n = 10; means ± s.e.m.).

We next sought to determine to what extent changes in macroscopic current observed with high [Na⁺]_i could be accounted for by changes in single-channel kinetics. For these experiments, oocytes were injected with cRNA for each of the α, β and γ wild-type ENaC subunits. Injected oocytes were subsequently incubated overnight in an MBS solution containing high [Na⁺], low [Na⁺], or high [Na⁺] with 200 nm ouabain to completely block the endogenous Na⁺,K⁺-ATPase (Horisberger & Kharoubi-Hess, 2002). Whole-cell I_Na was recorded the following day before any attempt was made to patch oocytes. As shown in Fig. 4A, the decrease in ENaC activity after overnight incubation in high [Na⁺] was similar to that seen after 1 h (Fig. 1A). The I_Na recorded from oocytes in high [Na⁺] was −6.9 ± 0.7 μA, while current from those in low [Na⁺] was −15 ± 2 μA, about a twofold difference. Ouabain application reduced macroscopic currents in oocytes even further to −1.6 ± 0.3 μA. Presumably, these cells suffer a particularly severe rise in [Na⁺]_i as a consequence of inhibiting the Na⁺,K⁺ pump which would otherwise work to help extrude intracellular Na⁺. As a control, ENaC-expressing oocytes were also incubated in a low-Na⁺ solution containing ouabain. The Na⁺ current in these cells was not significantly different from those that were incubated in low [Na⁺] alone (−14 ± 4 μA; n = 6).

Figure 4. Effects of high- and low-Na⁺ MBS on ENaC kinetics.

Oocytes injected with αβγ ENaC cRNA were incubated overnight in MBS solution containing high [Na⁺], low [Na⁺], or high [Na⁺] with 200 nm ouabain. A, macroscopic amiloride-sensitive Na⁺ current is inhibited by high [Na⁺] (high [Na⁺], n = 53; low [Na⁺], n = 52; ouabain, n = 30). B, exemplar cell-attached patch recordings taken from oocytes incubated in high- or low-Na⁺ MBS, with Li⁺ as the charge carrier. The number (N) of channels in each patch and estimated P_o are indicated to the left of each trace. Segments of recording are 1 min in length and correspond to a pipette voltage of +90 mV. Downward deflections show inward current from the pipette to the cell. C, single-channel i–V relationships were determined by measuring amplitude of transition levels at the voltages indicated (n = 19–146 for each point; means ± s.e.m.). Single-channel conductance (g) of ENaC was 7.4 ± 0.1 pS in low-Na⁺ solution and 7.5 ± 0.3 pS in high-Na⁺ solution. D, estimated P_o of ENaC in high-Na⁺ solution (0.30 ± 0.03; n = 13) was lower than P_o of ENaC in low-Na⁺ solution (0.57 ± 0.05; n = 11). *P < 0.05 compared to high-Na⁺ condition.

Having shown that the feedback inhibition mechanism was working in these cells, we analysed their single-channel currents by patch-clamp. In Fig. 4B, representative 1 min cell-attached recordings from patches containing one or more channels expressed in oocytes incubated in high- and low-Na⁺ solutions are shown. The recordings were made in a high-Na⁺ medium and with 110 mm Li⁺ in the pipette, irrespective of the pre-incubation conditions. The pipette voltage was +90 mV, with downward deflections corresponding to movement of positive charge from the pipette to the cell (inward current). To the left of the recordings, the estimated number of channels in the patch (N) and open probability (P_o) are indicated. Although we attempted to obtain cell-attached recordings from oocytes treated with ouabain, we were unable to form stable seals on these cells. Occasionally, brief (< 1 ms) large-amplitude (single-channel conductance, g ≫ 10 pS) transitions were noted during the recordings. We believe that these are unlikely to represent ENaC, and they were ignored in the data analysis.

Single-channel i–V curves were obtained by measuring the amplitude of the current deflections (openings) at various voltages. The single-channel conductance (g) was calculated from the slope of the single-channel i–V curves for each recording. This corresponded to 7.5 ± 0.3 pS for ENaC expressed in oocytes in a high-Na⁺ medium and 7.4 ± 0.1 pS for those incubated in a low-Na⁺ medium (Fig. 4C); there was no difference in single-channel conductance between the groups. The conductances were similar to values previously reported for ENaC when Li⁺ is the charge carrier (Sheng et al. 2000). The shift in i–V relationship is presumably the result of a more positive membrane potential secondary to a larger Na⁺ gradient in the case of the oocytes which were incubated in low-Na⁺ medium.

We next calculated N and P_o for each of the high-Na⁺- and low-Na⁺-treated groups as described in the Methods. N was not significantly different in the two groups (high Na⁺, 3.8 ± 0.7, n = 13; low Na⁺, 3.1 ± 0.5, n = 11). This parameter does not necessarily reflect the overall channel density in the membrane, since silent patches or patches with too many channels to resolve were excluded from the analysis. As illustrated in Fig. 4D, there was a significant difference in estimated P_o between the two groups; the P_o of ENaC expressed in oocytes incubated in high-Na⁺ solution overnight was 0.30 ± 0.03 (n = 13), while the P_o of ENaC in oocytes incubated in low-Na⁺ solution was 0.57 ± 0.05 (n = 11). In other words, the single-channel open probability appears to be halved as a consequence of Na⁺ loading.

In our analysis, the biggest assumption was that the number of channels in each patch was known. To check that our estimates for the number of channels in each patch (determined by the number of current levels observed) were reasonable, the aggregate dwell time for each open level was used to compute P_n, or the probability that n channels are open (as described in Methods). These measured, or experimental, P_n values were subsequently compared to P_n values predicted by the binomial distribution. Measured and predicted P_n values for selected traces were directly compared as depicted in Fig. 5A and B, respectively, showing that they are in good agreement. This supports our initial estimates of channel number and, by extension, the estimates of mean single-channel open probability in the high- and low-Na⁺ groups.

A reduction in P_o may be achieved by either decreasing channel open time or increasing channel closed time. A kinetic analysis of opening and closing rate constants was performed as described in the Methods; the results are shown in Table 1. While the values obtained for the rate constant for channel closing were similar in oocytes incubated in either high- or low-Na⁺ solution, the rate constant for opening was significantly different between the two groups (0.16 ± 0.03 s⁻¹ for high-Na⁺ solution and 0.81 ± 0.20 s⁻¹ for low-Na⁺ solution). Thus, it appears that a rise in intracellular [Na⁺] increases the amount of time ENaC spends in the closed state.

Table 1.

Effect of intracellular Na⁺ on single-channel properties

Channel	Condition	n	N	Po	g (pS)	k+ (s−1)	k− (s−1)
Wild type	High Na+	13	3.8 ± 0.7	0.30 ± 0.03	7.5 ± 0.3	0.16 ± 0.03	0.70 ± 0.09
Wild type	Low Na+	11	3.1 ± 0.5	0.57 ± 0.05*	7.4 ± 0.1	0.81 ± 0.20†	0.83 ± 0.16
βR564X	High Na+	7	3.4 ± 0.7	0.52 ± 0.05*	8.5 ± 0.6	0.73 ± 0.19*	0.90 ± 0.37
βR564X	Low Na+	6	2.8 ± 0.8	0.81 ± 0.04†	8.1 ± 0.1*	2.2 ± 0.5†	0.30 ± 0.17

n, number of trials; N, apparent number of channels estimated from observed current level transitions; P_o, estimated open probability; g, single-channel conductance; and k₊ and k₋, channel opening and closing rate constants, respectively. Data are presented as means ± s.e.m.

^*Statistically significant (P < 0.05) compared to wild-type ENaC high-Na⁺ condition;

^†statistically significant (P < 0.05) compared to βR564X high-Na⁺ condition.

It has been previously suggested that Liddle's mutations not only increase the number of Na⁺ channels at the membrane, but also increase their open probability and/or conductance (Firsov et al. 1996; Kellenberger et al. 1998). While quantitative studies on Liddle's channel surface expression exist, direct estimates of P_o by patch clamp have not been reported (Kellenberger et al. 1998). Because our data show that high intracellular [Na⁺] inhibits macroscopic wild-type ENaC current at least partly by reducing single-channel P_o, we hypothesized that in Liddle's syndrome, the part of the feedback process regulating P_o might also be disrupted.

For these experiments, wild-type (αβγ) or mutant (αβR564Xγ) ENaC cRNA was injected into oocytes; oocytes were incubated overnight in either a high-Na⁺ MBS, low-Na⁺ MBS, or high-Na⁺ MBS containing ouabain (200 nm). The next day, whole-cell currents were assessed by TEVC; these results are shown in Fig. 6A. The βR564X channels still responded to changes in intracellular Na⁺, although the fractional decrease in I_Na was significantly smaller than that in wild-type channels. Since there was not evidence for a Na⁺-dependent decline in current after a 1 h challenge (Fig. 1), this decline reflects a slower mechanism of feedback inhibition. We also analysed oocytes expressing βR564X channels by patch clamp. Figure 6B shows representative 1 min cell-attached patch recordings from oocytes incubated in high- and low-Na⁺ solution. Single-channel i–V curves were fitted by measuring the amplitude of the current deflections (openings) at various voltages. The single-channel conductance (g) calculated from the slope of the i–V curves was 8.1 ± 0.1 pS for low-Na⁺ solution and 8.5 ± 0.6 pS for oocytes incubated in high-Na⁺ solution overnight (Fig. 6C). The difference in single-channel conductance in cells expressing mutant channels was not significant under conditions of high and low [Na⁺]. However, the conductance of the mutant channels under conditions of low [Na⁺] was significantly higher that of the wild-type channels (P < 0.05). We next calculated NP_o and P_o for each of the high-Na⁺- and low-Na⁺-treated groups as described in the Methods. Single-channel open probability of βR564X ENaC expressed in oocytes incubated in low-Na⁺ solution was 0.81 ± 0.04 (n = 6), and for channels expressed in oocytes incubated in high-Na⁺ solution it was 0.52 ± 0.05 (n = 7; Fig. 6D; P < 0.05). Further analysis of these recordings showed that, while rate constants for channel closing were not significantly different between wild-type and mutant ENaC in high-Na⁺ solution, rate constants for opening were (0.73 ± 0.19 s⁻¹ for mutant versus 0.16 ± 0.03 s⁻¹ for wild type; Table 1). Moreover, there was also a significant difference in the calculated rate constant for opening for the mutant channels expressed in oocytes incubated in low- compared to high-Na⁺ solution (2.2 ± 0.5 s⁻¹ for low-Na⁺ solution versus 0.73 ± 0.19 s⁻¹ for high-Na⁺ solution). Since the rate constant for opening is inversely proportional to mean channel closed time, it therefore appears that the higher open probability of βR564X ENaC reflects the shorter amount of time that the channel spends in the closed state.

Figure 6. Effects of high and low [Na⁺] on single-channel kinetics of ENaC Liddle's mutant αβR564Xγ.

Oocytes injected with αβR564Xγ ENaC cRNA were incubated overnight in MBS solution containing high [Na⁺], low [Na⁺], or high [Na⁺] with 200 nm ouabain. A, I_Na in wild-type (wt) and βR564X mutant channels in the same batches of oocytes. Mutant channels are still sensitive to high [Na⁺], but fractional decrease in macroscopic amiloride-sensitive Na⁺ current is less than in wild-type channels (high [Na⁺], n = 12 wt, n = 25 mutant; low [Na⁺], n = 16 wt, n = 26 mutant; ouabain, n = 11 wt, n = 15 mutant). B, exemplar cell-attached patch recordings taken from oocytes incubated in high- or low-Na⁺ MBS, with Li⁺ as the charge carrier. Segments of recording are 1 min in length and correspond a pipette voltage of +90 mV. Downward deflections show movement of inward current from the pipette to the cell. C, single-channel i–V relationships were determined by measuring amplitude of transition levels at the voltages indicated (n = 7–84 for each point; means ± s.e.m.). Single-channel conductance (g) of ENaC was 8.5 ± 0.6 pS in low-Na⁺ solution and 8.1 ± 0.1 pS in high-Na⁺ solution. D, estimated P_o of ENaC in high-Na⁺ solution (0.52 ± 0.05; n = 7) was lower than P_o of ENaC in low-Na⁺ solution (0.81 ± 0.04; n = 6); P < 0.05. *P < 0.05 compared to high-Na⁺ condition; †P < 0.05 compared to wt channels under the same condition.

The fact that single-channel open probability of ENaC expressed in oocytes incubated in high-Na⁺ solution (0.30) was about one-half of that of those incubated in low-Na⁺ solution (0.57), and that whole-cell currents from oocytes in high-Na⁺ solution were one-third to one-half of those in low-Na⁺ solution suggested that changes in P_o might account for a significant component of the reduction in macroscopic current observed in oocytes in high-Na⁺ solution. Thus, we asked whether ‘activating’, or increasing the P_o of ENaC expressed in oocytes, would reduce or abolish the inhibitory effects of high-Na⁺ solution. Earlier studies have shown that a β-subunit with a serine 518 to cysteine mutation, when expressed with wild-type α and γ subunits, can, in the presence of the sulfhydryl reagent MTSET, form a channel that is persistently open; that is, when modified with the sulfhydryl reagent, the P_o of these channels is almost 1 (Snyder et al. 2000). Oocytes were injected with αβS518Cγ cRNA and incubated in either high- or low-Na⁺ MBS overnight. Whole-cell I_Na was measured by TEVC the following day, and was found to be similar to those of wild-type channels subjected to the same incubation conditions (Fig. 7A and B). Subsequent external application of MTSET to the oocyte bath solution upregulated whole-cell currents of oocytes incubated in low-Na⁺ solution by approximately 210% and in high-Na⁺ solution by 140%. Despite the increased activity of the single channels, whole-cell currents in oocytes incubated in high-Na⁺ solution remained substantially lower than those incubated in low-Na⁺ solution (Fig. 7B). This suggests that a reduced open probability of ENaC accounts for some but not all of the reduced macroscopic current observed in oocytes with high intracellular Na⁺ compared to low intracellular Na⁺.

Figure 7. Effects of increasing single-channel P_o on macroscopic currents.

A, oocytes expressing wt ENaC were incubated in high- or low-Na⁺ solution, and I_Na measured by TEVC. MTSET was then dissolved in 110 mm NaCl to a final concentration of 1 mm immediately before use. Currents in the presence of MTSET were measured every 1–2 min until they stabilized; amiloride-sensitive currents were subsequently assessed by perfusing 10 μm amiloride into the bath (n = 3 high [Na⁺]; n = 3 low [Na⁺]). B, oocytes expressing βS518C ENaC were incubated in high- or low-Na⁺ solution, and I_Na measured by TEVC. MTSET (1 mm) was then bath applied to oocytes, and amiloride-sensitive currents assessed (n = 14 high [Na⁺]; n = 12 low [Na⁺]). C, representative recordings taken from oocytes expressing βS518C ENaC incubated in either high- or low-Na⁺ solution overnight, showing response to MTSET and amiloride application. *P < 0.05 compared to wt high-Na⁺ condition; †P < 0.05 compared to mutant high-Na⁺ condition.

In order to test the possibility that Na⁺-dependent downregulation of current results from a difference in the number of conducting channels at the membrane, trafficking of new channels from the endoplasmic reticulum to the Golgi apparatus was inhibited using the fungal metabolite brefeldin A (BFA; Shimkets et al. 1997; Konstas et al. 2002; Volk et al. 2004). Sixteen to 24 hours postinjection, amiloride-sensitive Na⁺ current of oocytes maintained in either high- or low-Na⁺ solution was assessed (time 0). BFA (5 μm) was then added to the incubation media, and I_Na measured every 2 h for 10 h by TEVC; these results are shown in Fig. 8. In oocytes treated with BFA, the data were fitted by a single-exponential decay with a time constant in both high- and low-Na⁺ solution of 2.1 h s⁻¹. However, while current in high-Na⁺ solution decays to essentially zero over 8–10 h, in low-Na⁺ solution, it approaches a positive asymptote. Presumably, this reflects a pool of channels that are much more stable at the membrane (i.e. have a longer half-life), although we cannot rule out the possibility that in low-Na⁺ solution some channels bypass the Golgi apparatus on their way to the membrane and are thus resistant to BFA treatment (Hughey et al. 2004). In either case, the presence of more conducting channels at the membrane surface could contribute to the increased I_Na in oocytes maintained in low-Na⁺ solution.

Figure 8. Effect of brefeldin A on macroscopic Na⁺ current.

Oocytes injected with αβγ ENaC cRNA were incubated overnight in MBS solution containing high or low [Na⁺]. Sixteen to 24 hours after injection, macroscopic amiloride-sensitive Na⁺ current was assessed by TEVC. Brefeldin A (BFA; 5 μm) was then added to the incubation media, and Na⁺ current assessed every 2 h by TEVC. Each point represents the mean of 4–11 oocytes. Data were fitted by an exponential decay function (I = Ae^−t/τ + B). In high-Na⁺ solution, A = 3.0 μA, τ = 2.1 h and B = 0.3 μA; in low-Na⁺ solution, A = 5.0 μA, τ = 2.1 h and B = 1.9 μA. Error bars represent the s.e.m.

Feedback inhibition of macroscopic current in βR564X ENaC-expressing oocytes in high-Na⁺ solution can be attributed to a reduction in single-channel open probability, channel number, or some combination of the two. We thus employed a strategy similar to that described earlier, and introduced an additional mutation (S518C) into the β-subunit of ENaC harbouring the Liddle's syndrome truncation in order to assess whether simply activating these channels with MTSET would abrogate the effects of Na⁺-dependent downregulation. As shown in Fig. 9A, in the absence of MTSET, there is a significant difference in macroscopic Na⁺ current in βS518C–R564X ENaC-expressing oocytes in high- and low-Na⁺ solutions. When MTSET is washed into the bath, currents in both groups rise considerably, stabilizing at a level where they are no longer different in magnitude between incubation conditions. This demonstrates that with Liddle's mutant channels, a reduction in single-channel open probability accounts for the inhibition of current when [Na⁺]_i is high.

Figure 9. Effects of increasing single-channel P_o of βR564X ENaC on macroscopic currents.

A, oocytes were injected with αβS518C–R564Xγ ENaC and incubated in high- or low-Na⁺ solution. The following day, after I_Na was measured, MTSET was then dissolved in 110 mm NaCl to a final concentration of 1 mm immediately before use and applied to oocytes. Currents in the presence of MTSET were measured every 1–2 min until they stabilized; amiloride-sensitive currents were subsequently assessed by perfusing 10 μm amiloride into the bath (n = 11 high [Na⁺]; n = 15 low [Na⁺]). B, representative recordings taken from oocytes expressing βS518C–R564X ENaC incubated in either high- or low-Na⁺ solution overnight, showing response to MTSET and amiloride application. *P < 0.05 compared to high-Na⁺ condition.

Discussion

The main conclusions from this study are: (1) the single-channel open probability of epithelial Na⁺ channels in cells that are allowed to accumulate intracellular Na⁺ is reduced compared to those expressed in cells maintained in a low-Na⁺ milieu; (2) the single-channel open probability of ENaC harbouring a truncation mutation associated with Liddle's syndrome (a genetic disorder causing excessive renal Na⁺ absorption) is higher than the open probability of wild-type channels in cells maintained in either a high- or low-Na⁺ environment; (3) locking channels in a high-P_o state does not abrogate the effects of incubating cells in high-Na⁺ solution; (4) surface lifetime of wild-type channels in high-Na⁺ solution is reduced compared to low-Na⁺ solution; and (5) activating Liddle's mutant channels by increasing single-channel P_o eliminates the difference observed in macroscopic current between oocytes incubated in high- and low-Na⁺ solution. The simplest interpretation of these results is that in wild-type channels, the process of feedback inhibition relies on reducing both channel open probability and channel density to limit Na⁺ entry when [Na⁺]_i is high; in Liddle's syndrome ENaC, feedback inhibition is achieved primarily by reducing single-channel open probability.

As mentioned in the Introduction, ENaC activity is regulated by a negative feedback mechanism that enables the cell to curtail inward Na⁺ movement as intracellular Na⁺ rises. At the whole-cell level, feedback inhibition is characterized by a reduction of amiloride-sensitive current that is associated with elevated levels of intracellular Na⁺. Incubation of oocytes in a high-Na⁺ medium for 1 h or longer results in macroscopic I_Na that is two- to threefold lower than that of oocytes incubated in low-Na⁺ solution. It has been suggested that current ‘run-down’, or the short-term decline (within 1–2 h) of macroscopic current with TEVC, may be related to experimental treatment of the oocytes, such as physical perturbation of the cell membrane by the impaling electrodes (Volk et al. 2004). However, our results suggest that the response is specifically associated with an increase in [Na⁺]_i and may be physiologically relevant.

There are at least three ways in which feedback inhibition might work, none of which is exclusive of the others. First, a cell may reduce the number of active channels at the cell surface. Second, channels which exist in the plasma membrane in an active form can be converted to an inactive state. Third, the percentage of time that an active channel spends in the open state (i.e. the open probability) can be decreased. Unlike the effect of high [Na⁺] on channel number (N), the effect of high [Na⁺] on ENaC kinetics or P_o has not been studied in oocytes, although qualitatively our results are similar to those reported by Ishikawa and colleagues using excised patches in MDCK cells (Ishikawa et al. 1998). Previous attempts to study ENaC by patch clamp in both native and heterologous systems have been complicated by the fact that from patch to patch, channels display different kinetics and open probabilities (Garty & Palmer, 1997; Kellenberger & Schild, 2002). In our experiments, we decided not to restrict our analysis to patches with a single channel but to estimate the average P_o of patches containing one to eight channels. This allowed us to study a larger and perhaps more representative sample of patches. This approach has the disadvantage of uncertainties in the estimates of the number of channels in a patch, which will also affect the estimates of P_o. However, we do not think that this potential problem affects the main conclusion that P_o is lower in oocytes incubated in high-Na⁺ solution than those kept in low-Na⁺ solution. When channels have a P_o < 0.5, as in the high-Na⁺ group, it is more likely that the state in which all the channels are closed is visited than the state in which all channels are open. Thus N will be underestimated while P_o would be overestimated. In contrast, when P_o > 0.5, as in the low-Na⁺ group, the opposite occurs, and underestimates of N will lead to underestimates of P_o.

The possibility that we had either underestimated or overestimated P_o was assessed quantitatively following Marunaka & Eaton (1991). For all the recordings taken from wild-type ENaC-expressing oocytes incubated in low-Na⁺ solution, the probability that we had missed the level with all channels closed was < 0.05; the probability that we missed the level with all channels open was also < 0.05 for all recordings except one, where it was 0.13. Therefore, it is apparent that N was determined accurately for this group. In oocytes expressing wild-type ENaC incubated in high-Na⁺ solution, the probability that we missed the level with all channels closed was again very unlikely for all recordings (< 0.05). However, the probability that we missed a level with all the channels open in this group varied from 0.01 to 0.99, with a mean of 0.42. Thus, N is almost certainly underestimated, with the consequence that P_o is almost certainly overestimated in this group. If we assume that in every other recording in this batch, one opening was missed, the estimated single-channel P_o then falls from 0.30 to 0.26. Consequently, the difference in P_o between the high- and low-Na⁺ groups is likely to be greater, if anything, than what we have estimated.

In oocytes expressing mutant channels, N, and hence P_o, was estimated well. In low-Na⁺ solution, the probability that we missed the level with all channels open or all channels closed was < 0.05 for all recordings. In the high-Na⁺ group, the probability that we missed a level with all the channels closed was again < 0.05 for all recordings; there was only one instance of a probability greater than 0.05 that we missed the level with all channels open, where it was 0.26.

Further analysis of wild-type Na⁺ channel rate constants determined that while the rate constants of closing did not vary significantly between high- and low-Na⁺ incubation groups, rate constants of opening did. Specifically, the rate constant for channel opening of ENaC expressed in oocytes incubated in high-Na⁺ solution was 0.16 ± 0.03 s⁻¹ compared to 0.81 ± 0.20 s⁻¹ for low-Na⁺ solution. Thus, it appears that the lower open probability of channels in high-Na⁺ solution reflects an increased mean closed time for ENaC.

While the decline in I_Na appears to be triggered by increased Na⁺ entry, the time course of the decline is not related in a simple way to the intracellular Na⁺ concentration. This was particularly evident in experiments where attempts were made to reverse inhibition, when intracellular [Na⁺] fell quickly while current recovered very slowly (Fig. 2). This suggests that the high [Na⁺]_i does not act directly on the channels, which is in agreement with a study by Abriel and Horisberger using the ‘cut-open’ oocyte preparation (Abriel & Horisberger, 1999). Here, the intracellular contents of the oocyte were perfused with solutions containing varying concentrations of Na⁺, Ca²⁺ and ATP. In fact, none of these agents was able to induce rundown of macroscopic ENaC current. Thus, rather than Na⁺ interacting directly with the channel, it seems more likely that modulation of channel activity involves the activity of intracellular factors or proteins which can impinge on channel gating, trafficking, or both (Frindt et al. 1993; Silver et al. 1993; Komwatana et al. 1996).

If a change in ENaC gating could completely account for the difference in macroscopic Na⁺ current observed when oocytes are incubated in either high- or low-Na⁺ solution, then activating the channels (i.e. increasing P_o) by chemical modifiers might produce greater effects on the channels in high-Na⁺ solution, diminishing the effects of Na⁺ loading. To test this hypothesis, an MTS reagent was externally applied to oocytes expressing a mutant channel. MTSET stabilizes the open state of the channel such that the P_o approaches one. We found that this manoeuvre reduced but did not abolish the effects of Na⁺ loading, indicating that changes in P_o account for only part of the effects of changing [Na⁺]_i. Indeed, our data support the idea that Na⁺ entry is also limited by a decrease in channel number at the membrane. In particular, in experiments where BFA was applied externally to oocytes maintained in either high- or low-Na⁺ solution, it was apparent that although the rate of decay of current in the two conditions was similar, in high-Na⁺ solution the current decayed to essentially zero, while in low-Na⁺ solution, even after 10 h, a significant portion of the current remained. This is consistent with a decreased rate of internalization of at least a portion of the channels when intracellular [Na⁺] remains low.

The time course of decline in channel activity after BFA treatment suggests the presence of two populations of channels, at least with low-Na⁺ solution. One pool of channels is retrieved or inactivated rapidly, while the other is long lived. This could reflect post-translational processing of the channels, such as phosphorylation of one of the subunits. Previous results showed that phosphorylation of C-terminal residues could alter the affinity of ENaC for ubiquitin protein ligase (Nedd4) binding, which would in turn affect rates of internalization (Shi et al. 2002; Dinudom et al. 2004). It is possible that phosphorylation could also alter P_o, providing a link between the two mechanisms of feedback control. Alternatively, the two populations could represent channels with different subunit compositions that could also have different internalization rates. The observations of Weisz and colleagues, who reported differences in lifetimes of different subunits on the surface of A6 cells, are consistent with this possibility (Weisz et al. 2000). If these channels also had different open probabilities, selective depletion of one population from the membrane could give rise to the observed changes in mean P_o. Previous results support the possibility that subunit composition can affect single-channel properties (Fyfe & Canessa, 1998).

When compared with wild-type channels in oocytes, Liddle's mutants have a greater macroscopic current amplitude and are less sensitive to downregulation of current in high-Na⁺ solution. Several studies have shown that the increased ENaC surface expression, presumably caused by impaired internalization of the channel, contributes to increased whole-cell current (Snyder et al. 1995; Firsov et al. 1996; Kellenberger et al. 1998; Volk et al. 2004). Recent data suggest that Liddle's mutations may also increase the fraction of channels at the membrane that are proteolytically cleaved (Knight et al. 2006). Cleavage of ENaC subunits is thought to alter channel gating and increase single-channel P_o (Caldwell et al. 2004). This hypothesis is consistent with our experimental data, which show that the increased activity of one Liddle's mutant, βR564X, results, in part, from a higher single-channel P_o. When rate constants for the high- and low-Na⁺ conditions were analysed, it was apparent that, again, although differences in rates of channel closing were not statistically significant between wild-type and mutant ENaC groups, rates of channel opening were (see Table 1). Taken together, these observations suggest that the increased open probability of channels expressed in cells incubated in low-Na⁺ solution results from a decrease in the amount of time they spend in the closed state.

Since the Liddle's truncation mutants lack the C-terminal consensus motif for Nedd4-targeted ubiquitination, this suggested to us that the Na⁺-dependent downregulation of macroscopic current observed in oocytes expressing these channels may result entirely from a reduction in open probability. We thus reasoned that increasing the P_o of βR564X ENaC to 1 might eliminate differences in macroscopic current observed between groups of oocytes incubated in high- and low-Na⁺ solution. A cysteine at serine 518 was introduced in the ENaC β-subunit already harbouring the R564X truncation mutation, and MTSET was bath applied to oocytes expressing αβS518C–R564Xγ channels. Indeed, we found this manoeuvre to increase I_Na such that the difference in whole-cell current between incubation groups was abolished, indicating that a reduction in P_o can account for most, if not all, of the downregulation of Na⁺ current through these mutant channels observed when cells are Na⁺ loaded.

In summary, in this study we show directly for the first time in intact cells that changes in single-channel kinetics of ENaC can contribute to channel regulation in the case of elevated intracellular [Na⁺]. Specifically, under these conditions, channels demonstrate a reduced open probability induced by increased channel closed times. Increases in P_o can also contribute to channel dysregulation in the case of Liddle's syndrome, although these channels are still susceptible to a Na⁺-dependent fall in P_o. Our results support a physiological scheme where reduced channel P_o acts in concert with other processes, such as increased endocytic channel retrieval, to limit Na⁺ entry when [Na⁺]_i is high.