Defining an inhibitory domain in the gamma subunit of the epithelial sodium channel

Abstract

Proteases activate the epithelial sodium channel (ENaC) by cleaving the large extracellular domains of the α- and γ-subunits and releasing peptides with inhibitory properties. Furin and prostasin activate mouse ENaC by cleaving the γ-subunit at sites flanking a 43 residue inhibitory tract (γE144-K186). To determine whether there is a minimal inhibitory region within this 43 residue tract, we generated serial deletions in the inhibitory tract of the γ-subunit in channels resistant to cleavage by furin and prostasin. We found that partial or complete deletion of a short segment in the γ-subunit, R158-N171, enhanced channel activity. Synthetic peptides overlapping this segment in the γ-subunit further identified a key 11-mer tract, R158-F168 (RFLNLIPLLVF), which inhibited wild-type ENaC expressed in Xenopus laevis oocytes, and endogenous channels in mpkCCD cells and human airway epithelia. Further studies with amino acid-substituted peptides defined residues that are required for inhibition in this key 11-mer tract. The presence of the native γ inhibitory tract in ENaC weakened the intrinsic binding constant of the 11-mer peptide inhibitor, suggesting that the γ inhibitory tract and the 11-mer peptide interact at overlapping sites within the channel.

the epithelial sodium channel (ENaC) resides in the apical membrane of high-resistance epithelia in a wide array of tissues including the airway and kidney (4). In the aldosterone-sensitive distal nephron, ENaC participates in the fine control of sodium and fluid balance (16, 41). ENaC also participates in the maintenance of airway surface liquid volume, aiding mucociliary clearance of inhaled particles and pathogens (7, 30). Perturbations of ENaC activity have been described in pathophysiologic states that have an altered extracellular volume. For example, volume regulatory hormone excess (hyperaldosteronism) and ENaC mutations that impair channel endocytosis (Liddle syndrome) heighten ENaC activity in the kidney leading to volume expansion and hypertension (5, 16, 42).

ENaC is cleaved and activated by specific serine proteases including furin (a proprotein convertase that resides primarily in the trans-Golgi network), prostasin [a GPI-anchored serine protease also known as channel-activating protease 1 (CAP1)], TMPRSS4 (known as CAP2), matriptase (CAP3), trypsin, neutrophil elastase, pancreatic elastase, and plasmin (1, 6, 8, 9, 14, 20, 21, 23, 36, 43). ENaC modulation by proteases may be relevant to specific disease states including volume depletion, glomerular injury, and cystic fibrosis (26). Proteases and protease inhibitors have been shown to regulate surface liquid volume in airway epithelial cell cultures by modulating ENaC activity (8, 33, 44). Furthermore, excessive proteolytic activation of ENaC is believed to promote unregulated Na⁺ and water absorption in the cystic fibrosis airway and thereby compromise effective mucociliary clearance (4, 29, 30). Enhanced cleavage of ENaC subunits in the aldosterone-sensitive distal nephron occurs in animal models of sodium avid states such as hyperaldosteronism, low-salt diet, and proteinuria (19, 25, 31). Plasmin, a serine protease that can cleave and activate ENaC, has been implicated in the pathogenesis of volume expansion in nephrotic syndrome and was found in the urine of humans and animals with proteinuria, but was absent in the urine of control subjects (35, 36, 43).

Furin and prostasin activate ENaC by liberating inhibitory tracts from the extracellular regions of the α- and γ-subunits (6, 13, 23, 24). Within the α-subunit, furin cleaves ENaC twice at sites flanking a 26 residue inhibitory tract (13, 23). Within the γ-subunit, furin cleaves ENaC once following γR143 (in the mouse sequence) and a number of other serine proteases have been shown to cleave the γ-subunit at sites distal to the furin site. For example, γ-subunit processing by furin and prostasin results in the release of a 43-mer residue inhibitory tract (γE144-K186 in mouse ENaC) (6). Proteolytic processing does not affect the number of channels at the membrane. Instead, processing of ENaC increases the open probability (P_O) of the channel. Fully processed channels, lacking α and γ inhibitory tracts, have an P_O approaching 1.0 when expressed in Xenopus laevis oocytes, in contrast to unprocessed channels that have a P_O below 0.1 (39). Channels with only the α-subunit processed have an intermediate P_O of ∼0.3–0.4 (6, 28), whereas channels that have the γ inhibitory tract removed but retain the α inhibitory tract are highly active (11). Other factors such as phosphatidylinositides also have important roles in modulating ENaC P_O (27, 37).

ENaC activity is greatly reduced in mutant channels that lack furin-dependent cleavage sites in the α-subunit, but can be restored to wild-type levels if the intervening 26 residue inhibitory tract (αD206-R231) is deleted (13). We identified an eight amino acid minimal inhibitory tract within this 26 residue fragment (αL211-L218) (12). Synthetic peptides that correspond to either the 26 residue α inhibitory tract (αD206-R231) or to the eight key residues within the α inhibitory tract, LPHPLQRL, reduced ENaC currents in X. laevis oocytes, mouse cortical collecting duct cells (mpkCCDs), and human airway epithelia (HAE) (12, 13). In an analogous manner, a synthetic peptide corresponding to the inhibitory tract released from the γ-subunit by furin and prostasin (γE144-K186) inhibited endogenous ENaC activity in mpkCCD cells and HAE (6).

In this report, we identified an 11-mer residue tract (γR158-F168) that retains the inhibitory properties of the region extending between the furin- and prostasin-dependent cleavage sites in the γ-subunit (γE144-K186). Further studies with synthetic peptides bearing substitutions at specific sites defined important residues required for inhibition of wild-type channels by the 11-mer peptide. Our studies suggest that the tract γR158-F168, when present, interacts with the extracellular region of the γ-subunit reducing the P_O of the channel.

EXPERIMENTAL PROCEDURES

DNA constructs, site-directed mutagenesis, and cRNAs.

We used a PCR-based approach to generate serial deletions in mouse γ-ENaC within the γE144-K186 region (2, 40). The previously described γ-ENaC construct that possesses mutations at the furin- and prostasin-dependent cleavage sites (R143A and RKRK186QQQQ) was used as a template to create these deletion mutations (6). The γ-ENaC (γR143AΔE144-K186 and γΔE144-K186) constructs and α-ENaC construct that has furin-dependent cleavage site mutations (αR205A/R231A) were previously described (6, 12). cRNAs for α-, β-, and γ-ENaC subunits (wild-type and mutant) were synthesized using T3 mMessage mMachine (Ambion, Austin,TX).

Peptides.

Synthetic peptides with acetylated NH₂ termini and amidated COOH termini were produced and HPLC purified (>90% purity) by Genscript (Piscataway, NJ) or by the Genomics and Proteomics Core Laboratories of the University of Pittsburgh (Pittsburgh, PA).

Functional expression in X. laevis oocytes.

The protocol for harvesting oocytes from X. laevis was approved by the University of Pittsburgh's Institutional Animal Care and Use Committee. α-, β-, and γ-ENaC cRNAs were injected at a concentration of 2 ng·subunit⁻¹·oocyte⁻¹ into stage V–VI X. laevis oocytes (10). Electrophysiological measurements were performed using two-electrode voltage clamp (TEV) as previously described (10). The recording solution contained (in mM) 110 NaCl, 2 KCl, 1.54 CaCl₂, 10 HEPES, pH 7.4. For Na⁺ self-inhibition experiments, the low-Na⁺ TEV solution contained (in mM) 1 NaCl, 109 N-methyl-d-glucamine, 2 KCl, 2 CaCl₂, and 10 HEPES, pH 7.4. Na⁺ self-inhibition was measured by rapidly increasing the bath [Na⁺] from 1 to 110 mM (28). The amiloride-insensitive component of the whole cell current was determined by perfusion with recording solution supplemented with amiloride (10 μM). For the peptide dose-response curves, currents were measured at −60 mV. In all other experiments, currents were measured at −100 mV before and after amiloride. ENaC-mediated Na⁺ currents were defined as the amiloride-sensitive component of the current.

mpkCCD cell culture and short-circuit current measurements.

mpkCCD(cl4) cells were grown on permeable filter supports (0.4-μm-pore size, polycarbonate membrane, 1-cm² surface area Snapwell filters; Corning, Corning, NY). Medium was changed every 2 days and short-circuit currents (I_sc) were measured 4 to 6 days after cells were placed on filters (3). Filters containing the cell monolayers were mounted in modified Ussing chambers and continuously short-circuited using a VCC MC6 or MC8 voltage/current clamp system (Physiologic Instruments, San Diego, CA) (13). The bath solution contained (in mM) 120 NaCl, 25 NaHCO₃, 3.3 KH₂PO₄, 0.8 K₂HPO₄, 1.2 MgCl₂, 1.2 CaCl₂, and 10 glucose, pH 7.4. A gas mixture of 95% O₂-5% CO₂ at 37°C was constantly infused into the bath solution. The resistance of the monolayers was measured every 60 s using a 5-mV bipolar pulse.

Primary cultures of HAE cells.

As approved by the University of Pittsburgh Institutional Review Board, HAE were obtained from excess tissue remaining from lung transplantation or organ donor lungs not suitable for transplant. Cells were isolated from second through sixth generation bronchi and grown on permeable Transwell filters (Corning) as previously described (15). HAE cultures were used for I_sc experiments following 3–6 wk of culturing on an air-liquid interface.

Data and statistical analyses.

Data are expressed as means ± SE (n), where n is the number of independent experiments analyzed. Statistical comparisons between groups were performed using Kruskal-Wallis nonparametric ANOVA followed by Dunn's multiple comparisons posttest. A P < 0.05 was considered statistically significant. The IC₅₀ was expressed as a mean with 95% confidence interval. The IC₅₀ was estimated using normalized amiloride-sensitive currents and plotted as a function of peptide concentration fitted by the equation y = t[1+10^{(log X−log IC50)}], where X is the concentration of peptide. The IC₅₀ is the concentration of peptide that inhibits amiloride-sensitive currents (y) halfway between the baseline (t) and maximal response. Statistical comparisons and fitting of data points to dose-response curves were performed using GraphPad 5.0 (GraphPad Software, San Diego, CA) and Igor Pro (Wavemetrics, Portland, OR) (12).

RESULTS

Activating deletions in protease-resistant channels implicate a small inhibitory region within γE144-K186.

ENaC activity is modulated by proteolytic processing of its α- and γ-subunits (6, 13, 14, 23, 24). Mutations at the furin- and prostasin-dependent cleavage sites in the γ-subunit (γR143A and γRKRK186QQQQ) prevented the activation of the channel that results from cleavage of the γ-subunit (6). However, deletion of the 43-mer residue intervening inhibitory tract between the furin- and prostasin-dependent cleavage sites resulted in channel activation (6, 11). We hypothesized that the 43-mer residue tract interacts with the extracellular region of the γ-subunit, affecting the gating of the channel. To identify the key functional residues within the γ 43-mer inhibitory tract necessary for channel inhibition, we generated serial deletions within this tract (γE144-K186) in channels bearing mutations at the furin- and prostasin-dependent cleavage sites. As these γ constructs had furin- and prostasin-dependent cleavage sites mutated, only deletion of key residues required for inhibition, and not cleavage, should activate the channel (6, 23). ENaCs lacking furin-dependent cleavage sites in the α- and γ-subunits have very low activity when expressed in X. laevis oocytes and deletion of the γ inhibitory tract in these mutant channels induces a large increase in ENaC activity (11). To increase the sensitivity of the assay, mutant γ-subunits were coexpressed with α-subunits that had both furin-dependent cleavage sites mutated (αR205A/R231A), and wild-type β-subunits.

Channels composed of αR205A/R231A, wild-type β and wild-type γ retained their α and γ inhibitory tracts and have low activity (Fig. 1, top bar, n = 55). αR205A/R231Aβγ (control) had similar levels of functional expression compared with channels containing protease resistant α- and γ-subunits, αR205A/R231AβγR143A/RKRK186QQQQ (7 ± 2 and 4 ± 2% of wild-type αβγ current, respectively, Student's t-test, P = 0.14, αR205A/R231Aβγ vs. αR205A/R231AβγR143A/RKRK186QQQQ). Channels that retained their α inhibitory tract but had the γ -subunit inhibitory tract deleted as well as a mutated γ-subunit furin site (αR205A/R231AβγR143A/ΔE144-K186) had significantly higher activity than controls (αR205A/R231Aβγ; P < 0.001, n = 61; Fig. 1). In the setting of protease-resistant α- and γ-subunits, channels with γ-subunit deletions ΔE144-S150, ΔE144-P157, and ΔE172-G182 did not exhibit enhanced ENaC activity, suggesting that the tracts E144-P157 and E172-G182 were not required for inhibition (Fig. 1). In contrast, channels with the γ-subunit deletions ΔE144-F168, ΔT151-N161, ΔP157-L166, ΔP164-T181, and ΔN161-E170 showed significantly enhanced ENaC activity, compared with αR205A/R231Aβγ (*P < 0.001; see Fig. 1; Kruskal-Wallis nonparametric ANOVA followed by Dunn's multiple comparisons posttest, n = 10–61 oocytes). Our data suggest that partial or complete deletion of the intervening tract γR158-N171 is sufficient to activate the channel (Fig. 1). For instance, deletions of γT151-N161 and γP164-T181 do not overlap, but both resulted in enhanced activity.

Fig. 1.

Activating deletions in protease-resistant channels implicate a short inhibitory tract within γE144-K186. Two-electrode voltage clamp was performed in Xenopus laevis oocytes expressing ENaCs 36–48 h postinjection. Wild-type or mutant γ-subunits were coinjected with αR205A/R231A and wild-type β-subunits. Data are expressed as the ratio of the amiloride-sensitive currents of ENaCs containing mutant γ-subunits relative to controls (αR205A/R231Aβγ; I/I_control). Amiloride-sensitive currents in oocytes expressing αR205A/R231Aβγ and αR205A/R231AβγR143AΔ144–186 were −500 ± 55 nA (n = 55) and −3,270 ± 500 nA (n = 61), respectively. Statistically significant differences in the relative activity compared with controls are indicated as *P < 0.001 (Kruskal-Wallis nonparametric ANOVA followed by Dunn's multiple comparisons posttest). Experiments were performed with 10–61 oocytes per experimental group. Black bars within the illustrations on the left indicate residues that were deleted within the tract γR144-K186. Hatched bars indicate residues that were retained within the tract γR144-K186. Activating deletions involved all or part of the tract γR158-N171.

11-mer peptide corresponding to the tract γR158-F168 inhibits endogenous channels in mpkCCD and primary HAE cells.

To define the minimal tract in the γ-subunit required for ENaC inhibition, we examined the effect of a series of synthetic peptides corresponding to all or part of the region γR158-N171 on endogenous ENaCs in mpkCCD monolayers (Fig. 2). I_sc was determined before and following the addition of the peptides at a final concentration of 44 μM. DMSO (vehicle alone) served as a negative control in these experiments. The 43-mer peptide corresponding to the entire γ inhibitory tract, released by furin and prostasin cleavage, inhibited endogenous channels in mpkCCD cells (6). We found that a short peptide corresponding to γP157-N171 (PRFLNLIPLLVFNEN), encompassing the smaller region identified in the deletion experiments above, also significantly inhibited endogenous channels in mpkCCD cells (P < 0.001, n = 26; Fig. 2). The shortest peptide that inhibited endogenous channels was RFLNLIPLLVF corresponding to γR158-F168. Peptides lacking the proximal R158 or distal F168, such as FLNLIPLLVF and PRFLNLIPLLV, lacked inhibitory properties. The 11-mer RFLNLIPLLVF inhibited amiloride-sensitive currents in mpkCCD cells with an IC₅₀ of 3.5 μM (95% CI 2.7–4.5 μM, n = 22; Fig. 3C). This value is similar to the previously described IC₅₀ of the 43-mer peptide measured in mpkCCD cells (6). A peptide containing a scrambled version of this sequence (ILLVPFFLNLR) had no inhibitory effect (Figs. 2 and 3B). Representative tracings of the I_sc of an mpkCCD monolayer at increasing concentrations of RFLNLIPLLVF or scrambled peptide are shown in Fig. 3, A and B. We also tested the peptide RFSHRIPLLIF, corresponding to the 11-mer human ortholog, on cultured HAE cells (Table 1). At a single dose of 100 μM, this peptide reduced amiloride-sensitive currents in HAE cells by 95.2 ± 2% (Fig. 3D). By comparison, the mouse 11-mer peptide, RFLNLIPLLVF, inhibited I_sc in HAE cells by 46.0 ± 4.3%, suggesting a degree of species specificity. A scrambled version of the mouse 11-mer peptide had no effect in HAE monolayers.

Fig. 2.

11-mer peptide RFLNLIPLLVF inhibits ENaC. Synthetic peptides were generated based on the tract γR158-N171 identified by the activating deletions in protease-resistant channels. Synthetic peptides were applied apically at 44 μM to mpkCCD cells grown on permeable supports and mounted in an Ussing chamber. The relative inhibition of the ENaC-mediated I_sc (I/I_basal) was estimated as the ratio of the amiloride-sensitive component of I_sc following application of peptide, relative to the basal amiloride-sensitive I_sc. The amiloride-sensitive I_sc after application of the 43-mer peptide was 6.7 ± 0.6 μA/cm², n = 24 (I/I_basal was 0.26 ± 0.03), after RFLNLIPLLVF was 5.8 ± 0.7 μA/cm², n = 22 (I/I_basal was 0.28 ± 0.02), after scrambled peptide (ILLVPFFLNLR) was 24.5 ± 2.2 μA/cm², n = 6 (I/I_basal was 0.94 ± 0.02), and after an equivalent volume of DMSO (vehicle alone) was 27.1 ± 2.0 μA/cm², n = 25 (I/I_basal was 0.95 ± 0.02). Statistically significant differences in the response to the inhibitory peptides compared with control (DMSO) are indicated as *P < 0.001(Kruskal-Wallis nonparametric ANOVA followed by Dunn's multiple comparisons posttest). Experiments were performed with 6–26 filters per experimental group. Light gray letters mark the beginning and end of the 11-mer RFLNLIPLLVF.

Fig. 3.

Inhibition of endogenously expressed ENaCs in mpkCCD and human airway epithelia (HAE) cells by the 11-mer peptide RFLNLIPLLVF. Representative recordings of the effect of the 11-mer (A) and scrambled peptide (B) on I_sc measured in mpkCCD monolayers. Arrows indicate the addition of peptide (in μM) or amiloride (Amil) at 10 μM. C: inhibition of ENaC in mpkCCD cells by the 11-mer peptide RFLNLIPLLVF. A dose-response curve for the peptide RFLNLIPLLVF was generated in mpkCCD cells. I/I_basal represents the ratio of the amiloride-sensitive component of the I_sc following apical application of peptide, relative to the basal amiloride-sensitive current (I/I_basal). ENaC-mediated I_sc was inhibited with an IC₅₀ of 3.5 μM, 95% confidence interval (CI) 2.7–4.5 μM (n = 22). D: human ortholog of γR158-F168 inhibits endogenous ENaC in HAE cells. Inhibition was estimated as the ratio of the amiloride-sensitive I_sc following apical application of peptide (100 μM) relative to the basal amiloride-sensitive I_sc (I/I_basal). Human and mouse peptides significantly inhibited amiloride-sensitive I_sc compared with scrambled peptide (*P < 0.01 and **P < 0.001; Kruskal-Wallis nonparametric ANOVA followed by Dunn's multiple comparisons posttest). Experiments were performed with 3–11 filters per experimental group.

Table 1.

Alignment of orthologous 11-mer peptide sequences

Species	11-Mer Sequence
Mus musculus	(158) RFLNLIPLLVF (168)
Rattus norvegicus	(153) RFFKLIPLLVF (163)
Homo sapiens	(153) RFSHRIPLLIF (163)
Equus caballus	(155) KFLNLVPLLAF (165)
Canis lupus	(155) KFLRLVPLMVF (165)
Monodelphis domestica	(155) KFINSLPLVVF (165)
Oryctolagus cuniculus	(155) KFLNLVPLLIF (165)
Bos taurus	(155) KFLNLAPLMAF (165)
Xenopus laevis	(158) IFLKQIPLFRL (168)

The Xenopus laevis sequence is listed for comparison.

Substitutions of residues within the 11-mer peptide reveal structural requirements for inhibition.

Table 1 shows an alignment of 11-mer sequences of γ-subunits from various species. The mouse and human sequences have notable differences at positions 3–5 (LNL vs. SHR). A peptide corresponding to the human tract sequence, RFSHRIPLLIF, inhibited endogenous mouse ENaCs in mpkCCD cells with an IC₅₀ of 6.7 μM (95% CI 4.9–9.2 μM), suggesting that residues LNL at positions 3–5 are not required to inhibit the mouse channel (Table 2). The orthologous 11-mer peptide of X. laevis, IFLKQIPLFRL, showed reduced apparent affinity for endogenous channels in mpkCCD cells with an IC₅₀ of 17.4 μM (95% CI 11.2–27.1 μM), compared with the mouse 11-mer peptide (Table 2). To further examine structural requirements for channel inhibition, we generated peptides with substitutions at specific positions within the 11-mer peptide. We investigated the effect of the peptides (44 μM) on the amiloride-sensitive component of the I_sc in mpkCCD cells (Fig. 4). The 11-mer (RFLNLIPLLVF) reduced currents by 72 ± 2% at 44 μM. Peptides with substitutions of L by R at positions 3 or 9 retained inhibitory properties. In contrast, peptides with substitutions at positions 6, 7, 8, or 11 did not inhibit ENaC currents in mpkCCD monolayers and were not statistically different than DMSO (vehicle alone). The substitutions at positions 2 and 10 reduced the efficacy of the 11-mer peptide, compared with RFLNLIPLLVF (Student's t-test, P < 0.001). Taken together, our data suggest that the key residues for inhibition are RFXXXIPLXVF, where X is a site where an amino acid substitution did not significantly reduce the efficacy of the 11-mer peptide to inhibit ENaC. A review of mammalian orthologs listed in Table 1 suggests that positions 6 and 10 may also allow for other nonpolar amino acid substitutions, while the first position may require a basic residue.

Table 2.

Summary of peptide IC₅₀s

Cell System	ENaC	11-Mer Peptide (Species)	IC50, μM (CI)
mpkCCD	Endogenous channels	RFLNLIPLLVF (mouse)	3.5 (2.7–4.5) (n = 22)
mpkCCD	Endogenous channels	RFSHRIPLLIF (human)	6.7 (4.9–9.2) (n = 6)
mpkCCD	Endogenous channels	IFLKQIPLFRL (X. laevis)	17.4 (11.2–27.1) (n = 7)
X. laevis oocytes	αβγ	RFLNLIPLLVF (mouse)	2.7 (1.3–3.0) (n = 10)
X. laevis oocytes	αβγR143A	RFLNLIPLLVF (mouse)	2.3 (0.9–2.9) (n = 10)
X. laevis oocytes	αβγΔE144-K186	RFLNLIPLLVF (mouse)	3.5 (3.1–3.9) (n = 10)
X. laevis oocytes	αβγR143AΔE144-K186	RFLNLIPLLVF (mouse)	3.2 (1.8–3.1) (n = 10)
X. laevis oocytes	αR205A/R231AβγΔE144-K186	RFLNLIPLLVF (mouse)	0.56 (0.43–0.60) (n = 10)

Fig. 4.

Structure-activity analyses of the inhibitory peptide RFLNLIPLLVF. Synthetic peptides with single amino acid substitutions (in light gray letters) were applied apically at a concentration of 44 μM to mpkCCD cells grown on permeable supports and mounted in Ussing chambers. The magnitude of inhibition was estimated as the ratio of the amiloride-sensitive I_sc following application of peptide relative to the basal amiloride-sensitive I_sc (I/I_basal). An equivalent volume of DMSO (vehicle alone) had no effect on amiloride-sensitive currents (27.1 ± 2.0 μA/cm², n = 25) compared with amiloride-sensitive currents remaining after RFLNLIPLLVF (5.8 ± 0.7 μA/cm², n = 22). Responses were compared with the DMSO control using Kruskal-Wallis nonparametric ANOVA followed by Dunn's multiple comparisons posttest (*P < 0.001, **P < 0.01, +P < 0.05). Experiments were performed with 6–25 filters per experimental group.

Effects of proteolytic processing on 11-mer peptide efficacy.

When expressed in X. laevis oocytes, ENaC γ-subunits are primarily cleaved at the furin-dependent cleavage site (γR143), resulting in retention of the γ inhibitory tract (23). Wild-type channels exhibited an P_O of ∼0.3–0.4 (6, 28). Deletion of the γ 43-mer inhibitory tract (γΔE144-K186) resulted in channels with an P_O approaching 1, representing fully processed channels (6, 11). We examined the effects of the 11-mer peptide (RFLNLIPLLVF) on ENaC expressed in oocytes, where furin cleavage of the γ-subunit was blocked (γR143A mutant). In addition, the γ 43-mer inhibitory tract was either retained or deleted (γ vs. γΔE144-K186). In the presence or absence of the γ inhibitory tract, peptide inhibition was unaffected by cleavage at γR143, with the four possible combinations of conditions having IC₅₀ values of 2.3–3.5 μM (Table 2).

The proteolytic processing of the γ-subunit could have an effect on the measured 11-mer peptide efficacy by several mechanisms. First, excising the innate inhibitory tract may enhance accessibility to the peptide-binding site. Second, changing the open state-closed state equilibrium of the channel may affect the efficacy of the peptide. Determining the effect of release (or deletion) of the innate γ inhibitory tract on 11-mer peptide efficacy is complicated by the effect that tract release has on channel P_O. As the 11-mer peptide is an allosteric inhibitor that drives the channel to the closed state, increasing the channel P_O through γ inhibitory tract release will weaken measured 11-mer peptide affinity independent of its “true” binding affinity. In a pathway-independent equilibrium scheme

O and C are the open and closed states, respectively, P is the peptide, and O·P and C·P are peptide-bound states to the open (O·P) and closed (C·P) channel. The apparent peptide affinity K′ is given by

$$K\prime =\frac{[\mathrm{P}]([\mathrm{O}]+[\mathrm{C}])}{[\mathrm{O}\cdot \mathrm{P}]+[\mathrm{C}\cdot \mathrm{P}]}$$ (1)

Eq. 1 simplifies to eq. 2 assuming that [C·P] is much greater than [O·P] at equilibrium

$$K\prime ={\mathrm{K}}_P^C(1+K_o)$$ (2)

If the open state is highly favored, in the absence of peptide and giving a large K_o value, the value of K′ is also large despite the value of the intrinsic binding constant K_P^C. Since P_O = [O]/([O] + [C]), K_o can be expressed as P_O/(1 − P_O). Thus, small changes in P_O near P_O values of 1 will have dramatic effects on measured K′. Conversely, changes in P_O below values of ∼0.8 will have more modest effects.

To determine the P_O effect on measured 11-mer peptide affinity, we compared 11-mer inhibition in two groups of channels lacking the γ inhibitory tract that have different P_O (αβγΔE144-K186 vs. αR205A/R231AβγΔE144-K186). We measured both Na⁺ self-inhibition and 11-mer inhibition in these channels (Fig. 5, A and B) and estimated the intrinsic binding constant K_P^C based on these data. Measurements of Na⁺ self-inhibition can be used to approximate P_O (28). Na⁺ self-inhibition represents a reduction in channel P_O in response to external Na⁺. In general, this response is observed when channels are initially in a low-[Na⁺] bath, where we showed that they have a high P_O (39). When the bath [Na⁺] is acutely raised, measured current immediately increases commensurate with the increased driving force for Na⁺ entry reaching a peak current value (I_peak), which subsequently declines as the channel arrives at a new equilibrium at a lower P_O [current steady-state (I_ss)] due to Na⁺ self-inhibition. The ratio of (I_peak − I_ss)/I_peak represents the magnitude of the Na⁺ self-inhibition response.

Fig. 5.

Effects of cleavage, open probability, and presence of the inhibitory tract on 11-mer peptide affinity. A: representative tracing showing the Na⁺ self-inhibition and peptide response in an oocyte expressing αR205A/R231AβγΔE144-K186. Oocytes were perfused with two-electrode voltage clamp (TEV) solution containing 1 mM Na⁺ and then rapidly switched to a solution containing 110 mM Na⁺ to determine Na⁺ self-inhibition. The oocytes were subsequently perfused with TEV solution containing 110 mM Na⁺ and 2 μM peptide, followed by TEV solution containing 110 mM Na⁺ and 10 μM amiloride. B: analysis of the relationship between Na⁺ self-inhibition and peptide inhibition of mutant ENaC channels. Circles represent individual X. laevis oocytes expressing αβγΔE144-K186 (●) and αR205A/R231AβγΔE144-K186 (○). Inhibition by the peptide RFLNLIPLLVF (2 μM) was determined as the percent reduction in amiloride-sensitive whole cell Na⁺ current at the plateau. Na⁺ self-inhibition was estimated as previously described (28). A relationship exists between Na⁺ self-inhibition (ratio of I_peak − I_ss/I_peak), a surrogate for P_O, and the magnitude of peptide inhibition. Data were fitted with eq. 3. C: dose-response curves showing peptide block of αβγΔE144-K186 channels (●, IC₅₀ = 3.5 μM, 95% CI 3.1–3.9 μM, n = 10) and of αR205A/R231AβγΔE144-K186 channels (○, IC₅₀ = 0.56 μM, 95% CI 0.43–0.60 μM, n = 10). The relative inhibition was estimated as the ratio of the amiloride-sensitive component of the whole cell Na⁺ current following application of peptide relative to the basal amiloride-sensitive current (I/I_basal).

Na⁺ self-inhibition was measured by rapidly increasing the bath [Na⁺] from 1 to 110 mM, while 11-mer inhibition was measured at a single dose of 2 μM after whole cell currents had stabilized in the 110 mM Na⁺ bath (Fig. 5A). Oocytes expressing αβγΔE144-K186 had minimal Na⁺ self-inhibition (6 ± 2%, n = 10), reflecting the expected high P_O for these channels, and 38 ± 2% 11-mer inhibition of whole cell Na⁺ currents by 2 μM peptide. We also performed these experiments with channels that lacked the γ inhibitory tract but retain the α inhibitory tract (αR205A/R231AβγΔE144-K186), under the premise that this α-subunit mutation would reduce P_O without altering the 11-mer putative binding site in the γ-subunit. We found that these channels had increased Na⁺ self-inhibition (25 ± 4%, n = 10), consistent with a decrease in P_O, and increased 11-mer inhibition (84 ± 2%), consistent with the relationship described by eq. 2. We then estimated P_O based on the Na⁺ self-inhibition response, using the previously described linear relationship between P_O and Na⁺ self-inhibition (28). We fit these data to eq. 3 derived from the equilibrium scheme:

$$I_{\mathrm{P}}=I_{\text{max}}\frac{[P]}{[P]+\frac{K_P^C}{1+P_o}}$$ (3)

where I_P and I_max are the observed peptide inhibition and maximal predicted peptide inhibition, respectively, and calculated the intrinsic binding constant (K_P^C) of the 11-mer peptide for channels lacking the γ inhibitory tract at 0.17 ± 0.04 μM. This relationship between Na⁺ self-inhibition and 11-mer inhibition in channels lacking the γ inhibitory tract (αβγΔE144-K186 vs. αR205A/R231AβγΔE144-K186) is illustrated in Fig. 5B.

We estimated apparent 11-mer peptide affinity by measuring the dose response in these mutant channels (Fig. 5C). The measured IC₅₀ values of the 11-mer for αβγΔE144-K186 and αR205A/R231AβγΔE144-K186 were 3.5 μM (95% CI 3.1–3.9 μM) and 0.56 μM (95% CI 0.43–0.60 μM), respectively. Based on these values and the K_P^C value for channels lacking the γ inhibitory tract determined above, we estimated the P_O values of αβγΔE144-K186 and αR205A/R231AβγΔE144-K186 as 0.95 and 0.70, respectively. By comparison, wild-type ENaC expressed in oocytes had an 11-mer IC₅₀ value of 2.7 μM (95% CI 1.3–3.0 μM). Since wild-type ENaC has a P_O of ∼0.3–0.4, its K_P^C value is estimated at ∼1.8 μM, which is 10-fold weaker than for channels lacking the γ inhibitory tract (0.17 ± 0.04 μM). As the presence of the γ inhibitory tract weakens the K_P^C, these data suggest that the γ inhibitory tract and the 11-mer occupy overlapping sites on ENaC.

DISCUSSION

We identified an essential 11-mer residue tract (γR158-F168) within the γ-subunit of ENaC that encompasses the key functional residues of the γ inhibitory tract (γE144-K186) (6). Whole cell Na⁺ currents were enhanced in channels bearing partial or complete deletion of the 11-mer residue tract (γR158-F168), compared with controls. The synthetic peptide, RFLNLIPLLVF, corresponding to this tract inhibited ENaC expressed endogenously in mpkCCD cells and heterologously in X. laevis oocytes (IC₅₀ = 3.5 and 2.7 μM, respectively). The inhibitory efficacy of the 11-mer peptide is similar to the previously described γ-subunit derived 43-mer peptide (corresponding to γE144-K186) (6). The orthologous human 11-mer, RFSHRIPLLIF, inhibited endogenous channels in HAE cells as well.

The region between the furin- and prostasin-dependent cleavage sites is poorly conserved in mammalian γ-subunits. Within the 11-mer tract, all mammalian sequences have an initial basic residue, F at positions 2 and 11, and PL at positions 7–8. There is otherwise species variability within the 11 amino acid sequence, particularly at positions 3–5. Our data suggest that the residues at positions 3–5 may not have an important role regarding channel inhibition, as both human and mouse 11-mer peptides, which diverge at positions 3–5 (SHR for human and LNL for mouse), had similar IC₅₀ values when applied to mpkCCD cells. Eleven residues is the minimal length required for inhibition as determined by sequence-activity analyses. Deletion of R in position 1 or F in position 11 resulted in loss of inhibition. Comparison of mammalian sequences revealed that the R in position 1 could be replaced with another basic residue, K. Replacement of F at position 11 with a charged residue disrupted peptide function. The IPL tract at positions 6–8 was essential for inhibition. As this P is highly conserved, we speculate that this residue introduces a functionally important bend in the peptide. There appears to be only limited sequence similarities between this 11-mer inhibitory peptide derived from the γ-subunit and the previously defined α-subunit derived, eight residue inhibitory tract (αL211-L218), LPHPLQRL (12), suggesting distinct peptide-specific binding sites within the channel. Both peptides have P residues that are required for their inhibitory properties (12).

Although the X. laevis ortholog of the 11-mer peptide (IFLKQIPLFRL) inhibited ENaC with reasonable efficacy, it shares only partial sequence similarity with the mouse 11-mer peptide. Residues at position 2, 7, and 8 that are conserved within mammals are also present in the X. laevis ortholog (see Table 1). At position 11, F is substituted by L in X. laevis. Interestingly, we found that replacement of L by F at position 1 in the 8-mer α inhibitory peptide did not alter its efficacy (12). The sites within the channel that bind the 11-mer γ peptide are unknown. However, if the 11-mer peptide binds to a hydrophobic pocket, replacement of F by L may only have a modest effect on the efficacy of inhibition as both amino acids are relatively hydrophobic.

The presence or absence of the inhibitory tract in the γ-subunit did not significantly alter the measured IC₅₀ for the 11-mer peptide (Table 2). However, channels that lacked the γ inhibitory tract have a high P_O that will reduce the 11-mer apparent affinity. Taking into account the effect of the channel P_O on the efficacy of peptide inhibition, our results suggest the presence of the γ inhibitory tract weakened the intrinsic peptide affinity by ∼10-fold. This observation is consistent with the notion that the 43-mer γ inhibitory tract and the 11-mer peptide have overlapping binding sites. However, we cannot rule out the possibility of noncompetitive binding, as the concentration of the γ inhibitory tract within the γ-subunit cannot be experimentally altered.

Differences in the response to the external application of proteases have been observed in isolated collecting ducts when compared with cultured kidney epithelial cell lines. For example, trypsin did not affect amiloride-sensitive currents in A6, M1, or mpkCCD cell lines, suggesting that ENaC subunits are fully processed under the conditions that these cultured cell lines were maintained (34, 45). In contrast, trypsin increased ENaC activity and Na⁺ absorption in isolated CCDs (18, 32, 34), suggesting that a population of channels with subunits that are unprocessed or partially processed by proteases is expressed at the apical membrane in vivo. While γ-subunits that migrate at a molecular mass consistent with proteolytic processing have been described in rodent kidney (18, 31), evidence of multiple cleavage events leading to inhibitory tract release is lacking.

In conclusion, we identified an 11-mer residue stretch within the γ inhibitory tract that is essential for its inhibitory function. Aerosolized amiloride analogs offer an approach to inhibit ENaC in the airways of individuals with disorders such as cystic fibrosis (22, 38, 46), where aberrant proteolytic activation of ENaC has been implicated in the pathogenesis of mucociliary dysfunction (7, 8, 29, 30, 33). External peptides may provide an alternative therapeutic approach to inhibit ENaC.

GRANTS

This work was supported by National Institutes of Health Grants R01-DK-065161, P30-DK-079307, F32-DK-080574, K01-DK-078734, and K08-HL-087932. M. D. Carattino was supported by a Carl W. Gottschalk award from the American Society of Nephrology.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).