Voltage-Dependent Gating Underlies Loss of ENaC Function in Pseudohypoaldosteronism Type 1

Abstract

Here we explore the mechanism and associated structure-function implications of loss of function for epithelial Na⁺ channel (ENaC) containing a pseudohypoaldosteronism type 1 (PHA-1)-causing missense point mutation. As expected, human ENaC that contained subunits harboring PHA-1-causing substitutions within an absolutely conserved, cytosolic Gly residue (e.g., βG37S) had significantly less activity. Unexpectedly, though, such substitution also results in voltage sensitivity with greater activity at hyperpolarizing potentials. This is a consequence of voltage-dependent changes in the single-channel open probability and is not species- or subunit-dependent. Voltage sensitivity in PHA-1 mutants stems from the disruption of critical structure, rather than the development of new properties resulting from the introduction of novel side chains. Residues near the conserved His-Gly sequence of G95 in α-mENaC are particularly important for voltage sensing. Although substitution of I93 in α-mENaC results in voltage sensing, it also slows the activation and deactivation kinetics enough to enable capture of the dynamic changes in single-channel open probability that account for changes in macroscopic activity. This provides definitive proof of the mechanism that underlies loss of function. In addition, the voltage dependence of ENaC with PHA-1 substitutions is akin to that which results from substitution of a critical, interfacial Trp residue conserved at the intracellular base of TM1 (e.g., W112 in α-mENaC). Dynamic interactions between similarly positioned His and Trp residues are essential for gating and the girdle-like structure that lines the intracellular mouth of the M2 proton channel. The similar residues in ENaC may serve a shared function, suggesting the possibility of an intracellular girdle just below the mouth of the ENaC pore.

Introduction

The epithelial Na⁺ channel (ENaC) is a member of the ENaC/Degenerin ion channel family (1–3). Acid-sensing ion channels (ASICs) of central and peripheral neurons involved in touch and pain sensation are also members of this channel family (4–6). ENaC is common to the apical membrane of reabsorptive epithelial cells. It is selective for cations over anions and greatly favors Na⁺ over other cations. ENaC gates in a constitutive manner and is not overtly sensitive to voltage. Similarly to all ENaC/Deg channels, ENaC is trimeric, being a heterotrimer comprised of three similar subunits (α, β, and γ) encoded by distinct genes (2,7–9). All ENaC/Deg subunits share a common secondary and tertiary structure, i.e., a large extracellular domain separated from shorter cytosolic NH₂- and COOH-tails by two transmembrane domains (TM1 and TM2).

ENaC serves an important physiological function because its activity is limiting for Na⁺ transport across epithelial barriers, particularly in the kidney (4–6). Consequently, ENaC is a critical end-effector of homeostatic control systems such as the rennin-AngII-aldosterone system, which governs blood pressure through feedback regulation of systemic Na⁺ levels. The critical importance of this channel for proper control of blood pressure is apparent when one considers that gain and loss of ENaC function result in increases and decreases in blood pressure, respectively, associated with improper Na⁺ conservation and wasting by the kidney (10–13).

Pseudohypoaldosteronism type 1 (PHA-1) is an inherited renal Na⁺ wasting disorder that results from loss of ENaC function (11–13). Every member of the ENaC/Deg family contains an absolutely conserved His-Gly sequence in its NH₂-terminal cytosolic region just preceding TM1 (11,14–16). These residues are critical for channel function and their loss leads to decreased channel activity resulting from a decrease in the channel open probability (P_o) (11,15). A point mutation resulting in substitution of the critical Gly residue in human ENaC (hENaC) causes PHA-1 (11). However, beyond the fact that this decreases the P_o, as tested at a single membrane potential, little is known about the mechanism that causes loss of function in ENaC harboring this PHA-1-causing substitution.

The recent solving of the crystal structure for the transmembrane and extracellular domains of chicken ASIC1 shed much light on the structure of these portions of ENaC/Deg channels (8,9). In comparison, less is known about the structure and relative position of the cytosolic portions of ENaC to include the necessary His-Gly functional element that is critical for normal P_o and is disrupted in some forms of PHA-1.

Kir channels contain a girdle-like structure that fits into the intracellular mouth of their pore (17–19). This structure consists of G-loops at the apex of the cytoplasmic domains and functions as a flexible diffusion barrier, or gate, in the K⁺ conduction pathway. Mutations that disrupt G-loops result in loss of function (18,19). Indeed, such a mutation is causative for Andersen-Tawil syndrome, a form of inherited periodic paralysis and ventricular arrhythmias. The intracellular mouth of the pore of the M2 proton channel is also ringed by a girdle-like structure (20–24). In this channel, the indole rings in the side chains of Trp residues located at the base of transmembrane domains form cation-π interactions with the imidazolium rings in side chains of critical His residues to stabilize the girdle and thus affect gating. The side chains of these critical Trp residues in M2 proton channels jut into the aqueous environment in the mouth of the pore. Of interest, ENaC/Deg channels contain a conserved Trp at the intracellular base of TM1 (25). In the crystal structure of cASIC1, the side chains of these Trp residues jut into the aqueous mouth of the pore (9). Considering the structure of Kir and M2 proton channels, it is possible that the homologous Trp residues in ENaC interact with critical His residues to form a girdle-like structure that is important for gating.

To explore this hypothesis, we studied the mechanism that underlies loss of function in PHA-1-causing substitutions in ENaC. We found that as a result of PHA-1-causing substitutions, ENaC becomes voltage-dependent and the P_o is significantly decreased at physiological potentials. It is the loss of a critical functional element upon substitution rather than the introduction of novel structure that results in this phenotype. Moreover, the voltage sensitivity that results from mutation of the conserved His-Gly residues is akin to that observed upon substitution of the conserved Trp at the base of TM1. These data are consistent with the notion that ENaC also has a girdle-like structure just below the intracellular mouth of its pore that plays an important role in channel gating.

Materials and Methods

In this study we used standard reagents, practices, and electrophysiology methods to investigate wild-type and mutant ENaC (25–27). A complete description of the methods used is provided in the Supporting Material.

Results

PHA-1-causing mutations in hENaC decrease channel activity

Fig. 1 A shows representative macroscopic Na⁺ currents before and after treatment with 10 μM amiloride (amil., indicated by arrow) for CHO cells expressing wild-type hENaC (top, α + β + γ) and hENaC containing PHA-1-causing missense substitutions, such as G70S (middle) and G37S (bottom) in the α-hENaC and β-hENaC subunits, respectively. Currents were evoked with a voltage ramp from 60 to −100 mV from a holding potential of 40 mV, which is near E_Na for the physiological pipette and bath solutions used in these experiments. Amiloride is a broad-spectrum, open-channel inhibitor for most channels, particularly ENaC, in the ENaC/Deg family (4,5). As shown in Fig. 1 B, the introduction of a PHA-1-causing mutation into at least one hENaC subunit resulted in a significant decrease in channel activity, as expected. This finding is consistent with previous studies showing that PHA-1-causing substitutions (e.g., G37S in β-hENaC) in hENaC result in a loss of channel function (11,12). In addition, it is consistent with previous investigations of the conserved His-Gly sequence in which substitution of these residues decreased channel activity (11,16).

Figure 1.

PHA-1-causing mutations in hENaC decrease channel activity. (A) Representative macroscopic Na⁺ currents from CHO cells expressing wild-type hENaC α-, β-, and γ-subunits (top) or subunits engineered to contain PHA-1-causing mutations (middle and bottom) coexpressed with the complementary wild-type subunits before and after addition of 10 μM amiloride (arrow). Whole-cell, macroscopic currents evoked with a voltage ramp (shown in inset) from 60 mV to −100 mV from a holding potential of 40 mV. For these experiments, the bath and pipette [Na⁺] were asymmetrical at 150 and 5 mM, respectively. (B) Summary graph of (amiloride-sensitive) current density at −80 mV (from voltage ramps) for voltage-clamped CHO cells, similar to that in panel A, expressing wild-type hENaC and hENaC harboring PHA-1-causing mutations in either the α- or β-subunit. The number of experiments for each group is indicated. Current was measured with voltage ramps. ^∗ Significant (P < 0.05) decrease versus wild-type hENaC.

hENaC containing PHA-1-causing mutations senses voltage as a result of voltage-dependent changes in P_o

In previous studies (11,15), the decrease in activity for hENaC containing PHA-1-causing mutations was investigated at steady state, at a single holding potential. Thus, little was revealed about the mechanism that causes loss of function, and how these mutant channels behave over a range of voltages. To address these issues, we generated a whole-cell macroscopic current-voltage (I/V) relation for hENaC harboring a PHA-1-causing substitution, αG70S, and compared it with that of wild-type hENaC. Fig. 2 A shows representative families of macroscopic Na⁺ currents in symmetrical NaCl solutions, evoked by 20 mV voltage steps up to 100 mV and down to −200 mV from a holding potential of 0 mV for CHO cells expressing wild-type (top, α + β + γ) and mutant hENaC containing an α-subunit with the G70S substitution coexpressed with wild-type β- and γ-subunits (bottom). Fig. 2 B shows the resulting I/V relations for these channels at steady state. As expected, and consistent with previous findings (4,25), wild-type hENaC has a linear I/V relation in symmetrical NaCl solutions. Surprisingly, the I/V for hENaC containing the PHA-1-causing substitution has a noticeable voltage dependence. The voltage dependence of steady-state currents (at time 1) and tail currents (at time 2, at 80 mV) for wild-type (open boxes) and mutant hENaC containing the αG70S substitution (solid boxes and open circles) is quantified in Fig. 2 C. Clearly, hENaC does not respond to voltage. In contrast, hENaC harboring a PHA-1-causing mutation has pronounced voltage dependence, with activity increasing significantly at hyperpolarizing potentials. We found no difference when we quantified the voltage dependence with voltage steps versus tail currents. The half-activation potential for hENaC containing αG70S is estimated to be −236 ± 38 mV. As shown in Fig. S1, the time constants of activation and deactivation for αG70S channels are 764.4 ± 66.4 and 374.1 ± 104.2 ms, respectively, at −100 mV.

Figure 2.

hENaC harboring the PHA-1-causing mutation, G70S, senses voltage. (A) Families of macroscopic Na⁺ currents from representative CHO cells expressing wild-type (top) and hENaC harboring the αG70S PHA-1 mutation (bottom). Currents were evoked by progressive 20 mV voltage steps from a holding potential of 0 mV up to 100 mV and down to −200 mV (voltage protocol shown in inset). For these experiments, the bath and pipette [Na⁺] were symmetrical at 150 mM. (B) Macroscopic I/V relations for CHO cells expressing wild-type (black line) and hENaC with the αG70S mutation (gray line). The I/V relations were generated from experiments similar to that in panel A. For presentation, the current was normalized to current at −100 mV. (C) Steady-state G-V relations fitted with the Boltzmann equation for CHO cells expressing wild-type (open squares) and hENaC containing αG70S (solid squares: developed from steady-state current at time 1 using voltage steps; open circles: developed from instantaneous tail currents at 80 mV at time 2). For presentation, conductance was normalized to the maximum of the fit. For B and C, n ≥ 9 for each group.

Voltage dependence at a macroscopic level can arise from differences in permeability, selectivity, single-channel conductance, or an effect of voltage on channel gating leading to changes in P_o. To distinguish among these possibilities, we next studied wild-type and hENaC containing the αG70S substitution at the single-channel level in excised, outside-out patches. Representative, continuous current traces for hENaC (top) and hENaC containing the αG70S substitution (bottom) at voltages spanning 80 to −80 mV are shown in Fig. 3 A. Traces were obtained from patches pulled from CHO cells exposed to symmetrical LiCl solutions. As is clearly demonstrated by the I/V relations shown in Fig. 3 B, wild-type and mutant hENaC have similar single-channel conductance (8.34 ± 0.13 and 7.93 ± 0.16 pS, respectively) and, to the extent tested, selectivity and permeability values. The cause of the voltage dependence is revealed by the P_o-V curves shown in Fig. 3 C: the P_o of wild-type hENaC does not change over the range of voltages probed here, whereas that for hENaC containing αG70S increases as a function of hyperpolarizing potentials. The potential where P_o = 0.50 for this mutant is estimated to be −145 ± 4 mV. These single-channel results support the idea that voltage-sensitive changes in P_o are the primary mechanism underlying the voltage sensitivity and lower resting activity at physiological potentials for hENaC that harbor this PHA-1 substitution.

Figure 3.

hENaC containing αG70S gates in a voltage-dependent manner with increased P_o at hyperpolarizing potentials. (A) Families of representative single-channel current traces for wild-type (top) and mutant (bottom) hENaC in outside-out patches stepped from 80 mV to −80 mV. The bath and pipette [Li⁺] were symmetrical for these experiments. Inward Na⁺ current is downward; C, closed state. (B) Single-channel I/V relations for wild-type (black squares) and αG70S (gray circles) hENaC in outside-out patches. Data were obtained from experiments identical to that in panel A; n ≥ 3 for each group. (C) Plot showing the ENaC P_o as a function of voltage for wild-type (black squares) and mutant channels containing the αG70S mutation (gray circles). Data were fit by the Boltzmann equation. Data were obtained from experiments identical to that in panel A; n ≥ 3 for each group.

The voltage sensitivity in ENaC caused by PHA-1 substitutions of the critical His-Gly sequence is species- and subunit-independent

After establishing the mechanism underlying loss of function for hENaC containing the αG70S substitution, we wondered whether the properties introduced by this substitution were unique to ENaC from a specific species or to a particular channel subunit. To investigate this issue, we introduced homologous PHA-1-causing mutations into individual mouse ENaC (mENaC) subunits and expressed them with complementary wild-type mENaC subunits. Fig. 4 shows representative macroscopic Na⁺ currents (A) and the resulting I/V relations (B) from CHO cells expressing mENaC containing αG95S (top), βG37S (middle), or γG40S (bottom). Symmetrical NaCl solutions were used in these experiments. Fig. 4 C compares the steady-state activity (at −80 mV) of mutant mENaC containing homologous PHA-1-causing mutations in a single subunit with that of wild-type mENaC. Similarly to hENaC, mENaC containing a PHA-1 substitution has decreased activity. As is clear from the conductance-voltage (G-V) curves shown in Fig. 4 D (which were developed from the respective macroscopic I/V relations in Fig. 4 B), the underlying cause of this decrease in activity, at least for mENaC containing αG95S (estimated V_1/2 = −192 ± 13 mV) and γG40S (estimated V_1/2 = −185 ± 11), is the voltage sensitivity introduced by substitution. This voltage sensitivity is akin to that seen in hENaC containing αG70S. It is less certain that this also is the underlying cause of decreased activity in mENaC containing βG37S, since our investigation of the mechanism for this mutant was constrained by experimental limitations (we were unable to probe potentials more hyperpolarized than −200 mV). However, voltage dependence and clear activation kinetics begin to emerge at potentials near −200 mV for this mutant. It is possible that the activity of mENaC containing βG37S would increase upon further hyperpolarization. Nevertheless, these results are most consistent with the idea that the effects of substituting the homologous Gly residues in ENaC are responsible, in a species- and subunit-independent manner, for the PHA-1 phenotype leading to voltage sensitivity, resulting in less activity at physiological potentials.

Figure 4.

PHA-1 mutation results in a consistent ENaC phenotype across species and channel subunits. (A) Families of macroscopic Na⁺ currents from representative CHO cells expressing mENaC harboring PHA-1-causing mutations in the α (top; G95S), β (middle; G37S), and γ (bottom; G40S) subunits. Current was evoked by progressive 20 mV voltage steps from a holding potential of 0 mV up to 100 mV and down to −200 mV. For these experiments, the bath and pipette [Na⁺] were symmetrical at 150 mM. (B) Macroscopic I/V relations for CHO cells expressing wild-type (black lines) and mutant mENaC (gray lines) containing αG95S (top), βG37S (middle), or γG40S (bottom). The I/V relations were generated from experiments similar to that in panel A. For presentation, current was normalized to current at −100 mV. (C) Summary graph of (amiloride-sensitive) current density at steady state at −80 mV for CHO cells, similar to that in panel A, expressing wild-type mENaC and mENaC harboring PHA-1-causing mutations in the α-, β-, or γ-subunit. The number of experiments for each group is indicated. ^∗ Significant (P < 0.05) decrease versus wild-type mENaC. (D) Steady-state G-V relations were fitted with the Boltzmann equation for CHO cells expressing wild-type (open boxes) and mENaC containing αG95S (black squares) and γG40S (open circles). Data were obtained from experiments identical to that in panel A. For presentation, conductance was normalized to the maximum of the fit. For B and D, n ≥ 6 for each group.

Loss of the critical Gly residue rather than introduction of novel side chains results in the PHA-1 phenotype

As shown in Fig. 5, a His-Gly sequence is common to the NH₂-terminal cytoplasmic region just preceding TM1 in every ENaC/Deg subunit except human ASIC2b, which contains an Arg in place of His. This conservation led us to ask whether the loss-of-function phenotype in channels harboring subunits with the PHA-1-causing substitution of the critical Gly arises from loss of this residue or from the introduction of a specific type of side chain when the residue is replaced. To investigate this issue, we assayed the activity and voltage dependence of mENaC containing α-subunits in which G95 was substituted with distinct amino acids bearing widely different side chains. Fig. 6, A and B, show representative macroscopic currents and the associated I/V relations for CHO cells expressing mENaC that contain α-subunits with G95 substituted with different amino acids (alanine and cysteine). Studies were performed with symmetrical NaCl. Inward rectification is apparent in both mutants. Fig. 6 C compares the steady-state activity (at −80 mV) of mENaC containing differential substitution of αG95 with S, A, C, and W compared with wild-type mENaC. Every mutant had significantly less activity than the wild-type, and αG95W was nonfunctional (as far as tested). These findings are consistent with the notion that loss of the critical Gly residue, rather than the introduction of a novel residue with a specific type of side chain, results in this phenotype. Thus, Gly at this position is an absolute requirement for normal channel activity. Moreover, as indicated by the G-V curves in Fig. 6 D, substitution of G95 in α-mENaC always resulted in a similar voltage dependence when a functional channel was produced (estimated V_1/2 for G95S, G95A, and G95C = −192 ± 13 mV, −222 ± 27, and −203 ± 5 mV, respectively). This voltage dependence underlies the notable inward rectification for these mutants that is apparent in Fig. 6 B. In addition, the dependence on hyperpolarizing voltages for significant activity in these mutant channels is consistent with voltage dependence leading to low activity at physiological potentials.

Figure 5.

Consensus sequence around the conserved HG motif in the cytosolic amino-terminus of ENaC/Deg subunits. The sequences shown for α-, β-, and γ-ENaC subunits are absolutely conserved among ENaC orthologs but not ENaC/Deg paralogs. Examples of paralogs that contain the different amino acids at −3, −2, and +1 reported to date are shown below. Numbering for α-mENaC.

Figure 6.

Loss of a critical Gly residue in the cytosolic portion of mENaC results in the PHA-1 phenotype. (A) Families of macroscopic Na⁺ currents from representative voltage-clamped CHO cells expressing mENaC containing αG95 substituted with Ala (top) or Cys (bottom). Currents were evoked by progressive 20 mV voltage steps from a holding potential of 0 mV up to 100 mV and down to −200 mV. For these experiments, the bath and pipette [Na⁺] were symmetrical at 150 mM. (B) Macroscopic I/V relations for CHO cells expressing wild-type (black lines) and mutant (gray lines) mENaC containing αG95A (top) or αG95C (bottom). The I/V relations were generated from experiments similar to that shown in Fig. 5A. For presentation, the current was normalized to current at −100 mV. (C) Summary graph of (amiloride-sensitive) current density at steady state at −80 mV for voltage-clamped CHO cells (as in Fig. 5A) expressing wild-type mENaC and mENaC containing αG95S, αG95A, αG95C, or αG95W. The number of experiments for each group is indicated. ^∗ Significant (P < 0.05) decrease versus wild-type mENaC. (D) Steady-state G-V relations were fitted with the Boltzmann equation for CHO cells expressing mENaC containing αG95S (open squares), αG95A (gray circles), or αG95C (black squares). Data are from experiments identical to that in panel A. For B and D, n ≥ 6 for each group.

The region that includes the conserved His-Gly sequence is important for normal channel function

We next scanned through the region that included the conserved His-Gly sequence to determine whether nearby residues/positions are also critical for normal ENaC function. Fig. 7 shows representative macroscopic currents (A) and the corresponding I/V relations (B) for mENaC containing α-subunits with T92C (left), I93C (left middle), H94C (right middle), or A96C (right) substitutions. As shown in Fig. 5, neither the Thr at position 92, the Ile at 93, nor the Ala at 96 is conserved in ENaC/Deg paralogs, although these residues are conserved within the individual subunit orthologs. As expected, substitution of H94 resulted in inward rectification. Substitution of the Thr and the Ala three positions upstream and one downstream, respectively, of the critical G95 in α-mENaC had no effect on I/V relations. Substitution of I93, though, also resulted in inward rectification. The steady-state activities for mutant channels at −80 mV are summarized in Fig. 7 C. In contrast to the conserved His-Gly sequence, although substitution of I93 resulted in inward rectification, it did not have a major effect on steady-state activity. Fig. 7 D reports the G-V relations for these mutants. As for G95C, significant voltage dependence is the underlying cause of the inward rectification of H94C (estimated V_1/2 = −179 ± 9 mV) and I93C mutants (estimated V_1/2 = −166 ± 10 mV). Also included in Fig. 7 D is the G-V curve for mENaC bearing the W112C mutation in the α-subunit. As reported previously (25), this mutation, similarly to PHA-1-causing substitutions of G95, leads to voltage dependence and decreased activity. Of interest, the activation time upon hyperpolarization for the I93C substitution (Fig. 7 E) is >10 times slower than that for either the H94C or G95C mutation. At −60 mV, τ_act for I93C, H94C, and G95C = 3.4 ± 0.3, 0.2 ± 0.02, and 0.2 ± 0.03 s, respectively.

Figure 7.

Disruption of the Ile (αI93) and His (αH94) residues immediately upstream of the critical PHA-1 Gly (αG95) also leads to voltage dependence reminiscent of mutation of the conserved Trp (αW112) residue at the intracellular base of TM1. (A) Families of macroscopic Na⁺ currents from CHO cells expressing mENaC containing α-subunits with T92C (left), I93C (left middle), H94C (right middle), or A96C (right) substitution. The relative position to the critical Gly (G95) is indicated. Currents evoked by progressive 20 mV voltage steps from a holding potential of 0 mV up to 100 mV and down to −200 mV. For these experiments, the bath and pipette [Na⁺] were symmetrical at 150 mM. (B) Macroscopic I/V relations for CHO cells expressing wild-type (black lines) and mutant (gray lines) mENaC containing α-subunits with T92C (left), I93C (left middle), H94C (right middle), or A96C (right) substitution. The I/V relations were generated from experiments similar to that in panel A. For presentation, current was normalized to current at −100 mV. (C) Summary graph of (amiloride-sensitive) current density at steady state at −80 mV for voltage-clamped CHO cells (as in Fig. 6A) expressing wild-type mENaC and mENaC containing α-subunits with T92C, I93C, H94C, G95C, or A96C substitution. ^∗ Significant (P < 0.05) decrease versus wild-type mENaC. (D) Steady-state G-V relations were fitted with the Boltzmann equation for CHO cells expressing mENaC containing α-subunits with I93C (gray triangle), H94C (open circle), G95C (black box), A96C (open box), or W112C (dashed line) substitution. Data are from experiments identical to that shown in panel A. (E) Typical activation kinetics at −200 mV for normalized (to maximums at steady state) macroscopic Na⁺ currents carried by mENaC containing α-subunits with I93C, H94C, or G95C substitution. Current for I93C is shown at both 1 s (black line) and 10 s (gray line; noted with arrow) timescales. Inset shows mean τ_activation at −60 mV for the respective mutants. For B–D and inset in E, n ≥ 6 for each group.

Fig. 8 shows representative, continuous single-channel current traces for αI93C (A) and αH94C (B) mutants in outside-out patches in symmetrical LiCl. As is clear in these representative traces and the summary P_o-V plot in Fig. 8 C, voltage-dependent changes in P_o underlie voltage sensitivity at the macroscopic level. The steady-state P_o at −80 mV for I93C mutants approaches that of the wild-type at this voltage, which explains the normal macroscopic activity.

Figure 8.

mENaC containing α-subunits with I93C and H94C substitutions gate in a voltage-dependent manner. Families of representative single-channel current traces from CHO cells for recombinant mENaC containing αI93C (A) and αH94C (B) in outside-out patches stepped from 80 mV to −80 mV. The bath and pipette [Li⁺] were symmetrical for these experiments. Inward Na⁺ current is downward; C, closed state. (C) Plot showing ENaC P_o as a function of voltage for wild-type (black squares) and mutant mENaC containing αW112C (open circles), αI93C (open squares), or αH94C (black circles). Data were fit by the Boltzmann equation and obtained from experiments identical to that shown in panels A and B; n ≥ 4 for each group.

The relatively slow time of activation for mENaC containing an αI93C substitution allows definitive determination of the mechanism underlying voltage sensitivity

Fig. 9 A shows four representative single-channel current traces from outside-out patches pulled from CHO cells expressing mENaC with the αI93C mutation. Experiments were performed in symmetrical LiCl. Patches were rapidly transitioned from 60 mV to −60 mV and back again, with instantaneous current captured directly after the voltage step. These representative experiments and the summary graph of instantaneous NP_o (channel activity where N is the number of functional channels in the patched membrane and P_o is the mean open probability these channels have) as a function of time directly after the step in voltage to 60 mV (Fig. 9 B) clearly demonstrate that changes in single-channel P_o occur immediately. The time constants of activation and deactivation for I93C exposed to this voltage protocol are 4.3 ± 0.6 and 4.0 ± 1.1 s, respectively. These time constants agree well with the macroscopic current results for this mutant presented in Fig. 7 E. Moreover, these results provide incontrovertible evidence of mechanism.

Figure 9.

The relatively slow times of activation and deactivation for mENaC containing αI93C substitution enables documentation at the single-channel level of voltage-dependent gating for a substitution similar to PHA-1-causing mutations. (A) Four representative single-channel current traces in outside-out patches held at 60 (left) and −60 mV (right) for mutant mENaC harboring the αI93C substitution. The displayed currents immediately follow changes in holding potential. The bath and pipette [Li⁺] were symmetrical for these experiments. Inward Na⁺ current is downward; C, closed state. (B) Dairy plot showing changes in instantaneous (2 s bins), mean NP_o for αI93C channels as they deactivate immediately after a voltage step from −60 to 60 mV. Data were obtained from n ≥ 4 experiments identical to those shown in panel A.

Discussion

The results of this study confirm that substituting a conserved Gly residue in the cytoplasmic pre-TM1 region of the NH₂-terminus of hENaC subunits causes loss of function (11,14–16). We further show that this effect is not subunit-dependent, since decreases in activity can arise from mutation of any one of the three component subunits of the channel. We interpret this as indicating that the conserved Gly residue within each subunit serves a common function. Moreover, loss of function arises from homologous mutations in mENaC subunits, suggesting that the effect is common to all ENaC/Deg channel proteins independently of species. These findings demonstrate the critical importance of this Gly residue for activity, and are consistent with the idea that it is a key functional element within ENaC/Deg channels. Having a Gly residue at this key position is critical for channel function because replacing this residue with different amino acids with distinct side chains leads to similar decreases in channel activity. Replacing the small, nonpolar Gly residue at this position with a similarly small, nonpolar Ala residue is equivalent to inserting bulkier, hydrophobic cysteine and tryptophan residues. Thus, the difference between having a Gly or Ala at this position is enough to cause loss of function, suggesting a rigid structure-function requirement at this site. A surprising finding of our study is that missense substitution of the conserved Gly residue results in inward rectification of macroscopic currents due to notable voltage dependence, such that the activity of mutant channels is low at physiological potentials but increases with hyperpolarization. The cause of this voltage sensitivity is voltage-dependent gating rather than differences in single-channel conductance or selectivity with P_o increasing as a function of membrane hyperpolarization. The conserved Gly mutated in some forms of PHA-1 is preceded in the primary sequence by a conserved His residue. In similarity to the critical role played by Gly, missense substitution of this conserved His results in voltage-dependent gating and decreases in channel activity. This finding is consistent with the previous observation that macroscopic activity is decreased for channels that are mutant in this His (15,16). Substitution of the Ile in α-mENaC two positions upstream of the critical Gly residue slows the activation and deactivation kinetics of voltage-dependent changes in P_o enough to allow quantitation of changes in instantaneous P_o immediately after changes in potential. This is important because it definitively establishes the mechanism that underlies loss of function: a voltage-dependent decrease in P_o. The decrease in activity and underlying mechanism of voltage-dependent gating for ENaC containing substitution of the key His-Gly residues is akin to that observed for ENaC containing a missense substitution of a conserved Trp residue at the base of TM1 (25), suggesting the possibility of functional convergence and/or interaction between these residues/regions of the channel. The mechanism underlying voltage dependence in the latter mutant form of the channel also includes voltage-dependent changes in P_o with hyperpolarizing voltages relieving block of the pore by intracellular Na⁺.

We are not the first to report that missense substitutions of conserved His and Gly residues in the cytoplasmic, pre-TM1 region of ENaC/Deg subunits cause loss of function. Indeed, this has been established (11,12,14,15). Of importance, though, we expand on this understanding by delving into the underlying mechanism that causes decreases in channel activity, and discuss the associated structure-function implications in the context of what is known about the structure of ENaC and other ion channels.

As noted in the Introduction, Kir and M2 proton channels contain a girdle-like structure that runs around the intracellular mouth of their pores (17–19,21–24). These structures play a role in gating and involve the interactions of cytoplasmic regions of the channels with residues in transmembrane domains. The results we obtained here and in a previous work concerning an interfacial Trp residue at the base of TM1 (25) suggest that ENaC may have a similar structure. This interpretation is just one of many possibilities, but to us it seems the most likely and well supported because it is consistent with all current findings. There are several commonalities in how ENaC function changes upon substitution of the residues in the conserved His-Gly sequence and the conserved Trp at the base of TM1. For instance, both lead to a decrease in channel activity as a consequence of the emergence of voltage-dependent gating, with P_o being low at physiological potentials and increasing with hyperpolarization. It is also likely that both share a common mechanism of activation involving relief of pore block upon hyperpolarization. Considering the structure-function relations in M2 proton channels, which also have similar conserved His and Trp residues that interact to form the girdle-like structure inside the mouth of this channel (which is important for gating) (21–24), a logical suggestion is that these residues in ENaC also interact in a manner reminiscent of interactions in M2 proton channels to influence gating. Such an interaction in ENaC/Deg proteins may structurally require a small, nonpolar residue, namely Gly, to immediately follow the conserved His to allow normal channel function and gating. This scenario is consistent with the recent finding that the side chains of the interfacial Trp residues at the base of TM1 protrude into the aqueous environment of the intracellular mouth of the pore, where they are available to interact with other residues (9). In the context of such a discussion, the novel (to our knowledge) findings presented here are the first to suggest that the cytoplasmic, intracellular NH₂-terminal regions of ENaC/Deg channel proteins may have a critical structure that is capable of impacting channel gating. One possibility is that the functional elements within this region of ENaC may be part of a structure reminiscent of the girdle-like structure that rings the intracellular mouths of Kir and M2 proton channels. The disruption of such a structure would then interrupt normal channel gating, leading to the loss of function that is apparent in PHA-1 mutants.