Peptide Inhibition of Constitutively Activated Epithelial Na+ Channels Expressed in Xenopus Oocytes*

The hypothesis that 30-amino acid peptides corresponding to the C-terminal portion of the β- and/or γ-rat epithelial sodium channel (rENaC) subunits block constitutively activated ENaC was tested by examining the effects of these peptides on wild-type (wt) rENaC (αβγ-rENaC), truncated Liddle's mutants (αβTγ-, αβγT-, and αβTγT-rENaC), and point mutants (αβYγ-, αβγY-rENaC) expressed inXenopus oocytes. The chord conductances of αβTγ-, αβγT-, and αβTγT-rENaC were 2- or 3-fold greater than for wt αβγ-rENaC. Introduction of peptides into oocytes expressing αβTγ-, αβγT-, and αβTγT-rENaC produced a concentration-dependent inhibition of the amiloride-sensitive Na+ conductances, with apparent dissociation constants (K d ) ranging from 1700 to 160 μm, depending upon whether individual peptides or their combination was used. Injection of peptides alone or in combination into oocytes expressing wt αβγ-rENaC or single-point mutants did not affect the amiloride-sensitive whole-cell currents. The single channel conductances of all the mutant ENaCs were the same as that of wild type (αβγ-). The single channel activities (N·P o) of the mutants were ∼2.2–2.6-fold greater than wt αβγ-rENaC (1.08 ± 0.24,n = 7) and were reduced to 1.09 ± 0.17 by 100 μm peptide mixture (n = 9). The peptides were without effect on the single channel properties of either wt or single-point mutants of rENaC. Our data demonstrate that the C-terminal peptides blocked the Liddle's truncation mutant (αβTγT) expressed in Xenopusoocytes but not the single-point mutants (αβYγ or αβγY). Moreover, the blocking effect of both peptides in combination on αβTγT-rENaC was synergistic.

Hypertension is a common multifactorial disease imparting an increased risk of myocardial infarction, stroke, and endstage renal disease. Epidemiological studies suggest that up to 30% of human hypertension may have a genetic basis (1). Epithelial sodium channels (ENaC) 1 play a key role in regulating salt and water homeostasis by controlling sodium reabsorption in the distal nephron. The cDNAs encoding ENaC have been identified, and a heteromultimeric structure of the channel, comprised of three homologous ␣, ␤, and ␥ subunits, has been proposed (2)(3)(4). Each subunit contains a large extracellu-lar loop, located between two transmembrane domains, and two short intracellular N-and C-terminal domains. ENaC activities are controlled by insulin, corticosteroids, aldosterone (Ref. 5; for reviews see Refs. 6 and 7), protein kinases (8), proteases (9), cations (10 -12), the cytoskeleton (13), and osmotic pressure (14). Shimkets et al. (8) confirmed that protein kinase A, kinase C, and insulin could phosphorylate the Cterminal regions of both ␤and ␥-ENaC subunits. It is likely that the cytoplasmic N-and C-terminal segments of each subunit may contain the regulatory sites of many of the other aforementioned ENaC modulators.
The role of ENaC in the pathogenesis of genetic hypertension, namely Liddle's syndrome (pseudo-hyperaldosteronism), has been demonstrated by linkage analyses of the genes encoding the ENaC subunits in families with a history of salt-sensitive, low-renin, volume-expanded hypertension. These studies suggested that Liddle's syndrome resulted from truncating or missense mutations deleting a PPPXY motif in the cytoplasmic C terminus of either the ␤- (15)(16)(17)(18)(19)(20) or ␥-ENaC subunits (21). Systemic mutagenesis of the rat ENaC homolog, rENaC, suggested that normal ENaC activity could be modified by altering the consensus PPPXY sequence in an individual ␤or ␥-ENaC subunit (22). Each of these mutations resulted in increased channel activity when co-expressed with the other wt rENaC subunits in Xenopus oocytes, consistent with increased sodium reabsorption in the distal nephron (22,23). Similarly, stimulated amiloride-sensitive Na ϩ currents in oocytes expressing the Liddle's mutants of human ENaC were also observed (20,24).
At present, two non-mutually exclusive mechanisms have been posited to account for the enhanced macroscopic Na ϩ channel activity resulting from Liddle's mutations, namely an increased surface density of channel protein and/or increased channel open probability (P o ) (24 -26). While an increased residence time of ENaC at the cell surface has been attributed to defective internalization of the channel protein for ENaC constructs containing Liddle's mutations, the mechanisms by which these mutations in the ␤or ␥-ENaC subunits lead to an increase in single channel P o are not well understood. Interestingly, the currents associated with Liddle's mutants expressed in oocytes were not down-regulated by elevated intracellular Na ϩ concentration in contrast to wt ENaC currents (27). Furthermore, Dinudom et al. (28) suggested that elevated intracellular Na ϩ promotes the interaction of Nedd4 with the PY motif of ␤-rENaC. Hence, feedback inhibition of ENaC by intracellular Na ϩ would be lost with Liddle's mutations. The differences between the regulation of wt ENaC and Liddle's mutants have not been explored in detail. However, like other cation channels (29 -31), the activity of truncated Liddle's ␤-ENaC mutants immunopurified from human lymphocytes were also inhibited by synthesized peptides when reconstituted into planar lipid bilayers (32).
We had previously demonstrated in planar lipid bilayer studies that truncation of the C termini of ␤or ␥-rENaC led to increased channel P o (32,33). We wished to test further the hypothesis that the C-terminal peptides of these subunits could function as inhibitory peptides of rENaC expressed in Xenopus oocytes. In the present study, we investigated the effect of the 30-amino acid C-terminal peptides of ␤or ␥-rENaC subunits on the currents associated with the truncation and point mutants found in Liddle's syndrome using whole-cell and single channel recording techniques. The results showed that both the individual peptides SP 30 ␤ or SP 30 ␥ and a mixture of both peptides decreased the macroscopic currents of truncated Liddle's mutants expressed in oocytes but had no effect on the currents associated with expression of the Liddle's point mutants. The single channel activity (N⅐P o ) of the truncated Liddle's mutant was down-regulated by the peptides due to decreases in P o and the number of channels per patch. The conclusions of the present study were that the increased channel P o of the truncated Liddle's mutants expressed in Xenopus oocytes could be reversed by the synthetic peptides, thus the C termini cytoplasmic tails of the ␤ and ␥ subunits act in concert to form part of the normal gating mechanism of a functional ENaC.
Peptides SP 30 ␤, SP 30 ␥, and Pro-SP 30 ␤ were synthesized by Research Genetics (Huntsville, AL). Following synthesis, the peptides were purified (to Ͼ90% purity) by reversed-phase high performance liquid chromatography. The amino acid sequence of each peptide was verified by mass spectroscopy. The primary amino acid sequences of the synthetic peptides were as follows: SP 30  Oocyte Expression of ENaC-The methods for isolation of oocytes and expression of rENaC cRNA were as described previously (34). Briefly, ovarian tissue was removed from frogs (Xenopus Express, Burley Hills, FL) under anesthesia through a small incision in the lower abdomen. Oocytes were digested with 2-4 mg/ml collagenase in Ca 2ϩ -free OR-2 solution for 2 h on a shaker. Following several washes in Ca 2ϩ -free or regular OR-2 media, oocytes were manually defolliculated. Healthy oocytes at maturation stage V and VI were identified and maintained in L-15 medium at 18°C. Oocytes were incubated overnight before cRNA injection. A Nanoject microinjector (Drummond Scientific Co., Broomall, PA) was used for injection of 50 nl of rENaC cRNA per oocyte in the ration of 1:1:1 (␣:␤:␥) or the same volume of RNase free water (control). The total cRNA of all 3 subunits injected into each oocyte varied from 5 to 25 ng. Injected oocytes were transferred to MBS medium or L-15 medium and incubated at 18°C for 48 h before recording whole-cell and single channel data.
Solutions and Chemicals-Ca 2ϩ -free OR-2 medium was used to wash oocytes before and after digestion with collagenase. This medium contained (in mM) 85.2 NaCl, 2.5 KCl, 10 NaHEPES, 1.0 Na 2 HPO 4 , 0.5% streptomycin, pH 7.5. Calcium chloride (1.8 mM) was added to Ca 2ϩ -free OR-2 medium to make regular OR-2 medium. The incubation medium was modified Barth's solution (MBS) with the following composition (mM): For patch clamp experiments, we used a hypertonic solution for stripping the vitelline membrane from the oocytes composed of (in mM), 200 potassium aspartate, 20 KCl, 1 MgCl 2 , 10 EGTA, 10 NaHEPES, pH 7.4. The basal extracellular medium contained (in mM) 100 LiCl, 10 HEPES, pH 7.4. To obtain outside-out patches, 1.8 mM CaCl 2 was included in the extracellular medium. In order to measure the permeability ratio of Li ϩ :K ϩ , equimolar potassium was used to replace lithium. Peptides were dissolved in water to 100 mM and stored at 4°C until used. All other reagents were obtained from Sigma.
Two-electrode Voltage Clamp-Whole-cell currents were measured in oocytes held at room temperature (22-25°C) 24 h after cRNA injection. The currents were analyzed using pCLAMP software (Axon Instruments, Foster City, CA). Voltage clamp potentials were evoked using a TEA-200 voltage clamp (Dagan Corp., Minneapolis, MN) controlled by a personal computer connected via a TL-1 interface (Axon Instruments). The injected oocyte was placed in a small holding chamber (Ͻ50 l) and was initially superfused with the desired solution at a flow rate of 1.0 ml/min for at least 5 min before recording. Microelectrodes were filled with 3 M KCl and had a tip resistance of 0.5-2.0 M⍀. The bath electrodes consisted of two chloride-coated silver wires that made electrical contact with the bath through 3% agar bridges immersed in 3 M KCl. For peptide injection, a third glass electrode (Drummond Scientific Co.) filled with a stock solution of peptide (usually 100 mM) was inserted into an oocyte subsequent to impalement with the two recording electrodes. The desired concentration of peptides within an oocyte was defined by injecting a known volume of the peptide stock using an adjustable nanoliter injector (WPI, Sarasota, FL) and calculating the oocyte volume by assuming the oocyte was a prolate spheroid. The oocytes were clamped at a holding potential of Ϫ100 mV. Data were filtered at 0.5-1 KHz, digitized, and stored on hardware for off-line analysis. For ramp I-V recordings, test voltages were elicited from a holding potential of Ϫ100 to Ϫ140 mV through ϩ60 mV at a rate of 10 mV/25 ms for 500 ms. Currents at Ϫ100 mV were sampled at 20-s intervals. Capacitance neutralization was done for each recording using the circuit built into the TEA-200 voltage-clamp amplifier. In addition, the small volume of the holding chamber (Ͻ50 l) also helped minimize the capacitance at the input of the head stage. The resting membrane potentials were read directly from the monitor window of the voltage clamp before and after application of amiloride.
Single Channel Recordings-The oocytes used for patch clamp experiments were first placed in a hypertonic medium at room temperature for 5-20 min. The vitelline membrane was then manually removed from the shrunken oocytes using fine forceps (35). The devitellinized oocyte was immediately transferred to a recording chamber (Warner Instruments, Hamden, CT) and perfused with isoosmotic extracellular solution before patch clamping the cell.
All three configurations (cell-attached, excised inside-out, and outside-out) of patch clamp recordings were used (36). The reference electrode was an Ag/AgCl pellet bathed in the same solution as the bath. For constructing current-voltage (I-V) curves and calculating P o , data obtained from cell-attached, inside-out, and outside-out patches were included. For plotting concentration dependence curves of amiloride inhibition, the outside-out patch configuration was employed, and a series (from 0 to 10 M) of concentrations of amiloride was applied. Patch pipettes were made of borosilicate glass, pulled in two steps with a PP83 vertical puller (Narishige, Japan). The tip resistance of the electrodes was 1-10 M⍀. Single channel currents were recorded with an Axopatch 1-B current-voltage clamp amplifier (Axon Instruments, Inc., Burlingame, CA). The current traces were displayed both by CLAMPEX version 7.0 associated Axoscope software on the monitor and on an oscilloscope and were stored on a hard drive connected with the working computer through the internet. By convention, for the outside-out configuration, the intracellular potential corresponds to the pipette potential (V p ), and negative (downward) single currents correspond to cation influx from the extracellular to the intracellular side of the membrane. For the cell-attached configuration, the membrane potential should be close to the actual potential across the membrane patch because the oocyte membrane potential is close to 0 mV measured by two-electrode voltage clamp under our experimental conditions. Current signals were filtered at 1 kHz with an 8-pole Bessel filter (Frequency Devices Inc., Haverhill, MA) and digitized at 2 kHz using a Digidata 1200 interface and CLAMPEX software (Axon Instruments, Inc., Foster City, CA). For display, currents were low-pass filtered at 100 -200 Hz.
Data Analysis-Amiloride-sensitive macroscopic whole-cell currents were calculated by subtracting the currents in the presence of 10 M amiloride from the currents in the absence of amiloride. Student's t test was used to analyze the differences of amiloride-sensitive currents among each group. The results were presented as mean Ϯ S.E., and the degree of significance was assessed using the Student's t test. Data from rENaC expressing oocytes for quantifying the inhibitory effect of peptide were normalized to the maximum current in the absence of peptide. The apparent inhibition dissociation constants K d were computed according to Equation 1, where K d is the apparent dissociation constant and n is the Hill coefficient. I and I 0 represent the amiloride-sensitive Na ϩ currents in the presence or absence of the C-terminal peptides, respectively. Analysis of single channel data was performed using FETCHAN and pSTAT program of software CLAMPEX version 7.0 (Axon Instruments). The holding potentials (or the polarity of the single channel currents) were corrected accordingly. Assuming the lowest level during more than 3 and up to 30 min of continuous recordings was the closed level, the single channel open probability (P o ) was calculated from using Equation 2, where N is total number of channels, Ī is the mean current over the period of observations, and i is the main state unitary current determined from all points current amplitude histograms produced by FETCHAN. The mean current over the period of observation was calculated using the events list files generated by CLAMPEX version 7.0 software and Equation 3, where i m is an event current (all level including the 0 level); t m is an event dwell time, and M is the total number of events. The parameters calculated according to the above method may be underestimated due to the long open and closed transitions and multichannel nature of the patches of ENaC expressed in oocytes. Data are expressed as mean value Ϯ S.E. for n patches. Because the pSTAT program could not used to analyze the results of multichannel patches with more than five conductance levels, the program GAUSS (supplied to us by Dr. James Kenyon of the Department of Physiology, University of Nevada) was used to analyze channel data consisting of six or more discrete current levels (38). The permeability ratios of wt ␣␤␥-rENaC and Liddle's mutants expressed in oocytes were computed from the measured reversal potentials using the Goldman-Hodgkin-Katz (GHK) equation.

Macroscopic Currents of Truncation Mutants Are Inhibited
by Peptide Injection-In order to test the hypothesis that peptides corresponding to the C termini of the ␤and/or ␥-rENaC subunits (SP 30 ␤ and SP 30 ␥) inhibit Liddle's mutations inserted into rENaC, the ␤-rENaC subunit truncated at amino acid position Arg-564 (␤ T ) and the ␥-rENaC subunit truncated at amino acid position Arg-574 (␥ T ) were expressed in oocytes in combination with the wt ␣-rENaC subunit (␣␤ T ␥ T ). Fig. 1 shows representative current traces in oocytes evoked by a voltage ramp from Ϫ140 mV to 60 mV. The chord conductance of ␣␤ T ␥ T -rENaC-associated currents at hyperpolarizing potentials was ϳ3-fold greater than that of wt ␣␤␥-rENaC. Injection of a mixture of ␤ ϩ ␥ C-terminal peptides (1:1, 138 M each) into oocytes expressing the truncated ENaC constructs (␣␤ T ␥ T -rENaC) decreased the current by approximately 63% at hyper-polarizing potentials. In contrast, the current in oocytes expressing wt ␣␤␥-rENaC was not affected by the same concentration of peptide mixture. The ␣␤ T ␥ T -rENaC-associated current was not decreased by water injection or by injection of either of two control peptides, one comprised of the last 13 amino acids of the C terminus of CFTR (KEETEEEVQDTRL, final concentration ϭ 1.7 mM; n ϭ 4) or a 30-mer peptide identical in all respects to SP 30 ␤ except for substitution of 3 prolines for the single valine and two isoleucines (Pro-SP 30 ␤). Recordings were made 24 h after cRNA injection for both ␣␤␥and ␣␤ T ␥ T -rENaC. Currents were elicited with a ramp voltage pulse from Ϫ140 to ϩ60 mV starting at a holding potential of Ϫ100 mV. Amiloride-sensitive Na ϩ currents shown in this figure were calculated by subtracting the currents recorded in the presence of amiloride from those in the absence of amiloride (10 M). The left panels show amiloride-sensitive currents of wt ␣␤␥and ␣␤ T ␥ T -rENaC, and the right panels show effect of peptide mixture (276 M; 1:1 of SP 30 ␤ ϩ SP 30 ␥) or water as a control on the corresponding currents in the same oocyte. Injection of the peptide mixture does not affect the current of ␣␤␥-rENaC (B) but significantly inhibits ␣␤ T ␥ T -rENaC-associated Na ϩ current by approximately 63% (F). The amplitude of ␣␤ T ␥ T -rENaC-associated Na ϩ current (C and E) at Ϫ100 mV is approximately 3-fold greater than that of ␣␤␥-rENaC (A). Injection of same volume of nuclease-free water (13.8 nl) into an oocyte expressing ␣␤ T ␥ T -rENaC (D) does not show an inhibitory effect on the currents of ␣␤ T ␥ T -rENaC. These experiments were each repeated four times. the bath (n ϭ 3, 500 M) or Pro-SP 30 ␤ (n ϭ 3, 1.7 mM) failed to alter the gating behavior of ␣␤ T ␥ T -rENaC in inside-out patches (data not shown).
Summarized data showing inhibition of wt and ␣␤ T ␥ T -rENaC-induced macroscopic Na ϩ currents both for the individual peptides and a 1:1 mixture of SP 30 ␤ and SP 30 ␥ peptides (138 M each) are plotted in Fig. 3. The mean chord conductance of ␣␤ T ␥ T -rENaC-expressing oocytes was 123.7 Ϯ 12.6 S (n ϭ 11). This value was more than twice that of wt ␣␤␥-rENaC-expressing oocytes (51.9 Ϯ 50 S, n ϭ 5; p Ͻ 0.001). Injection into oocytes of the mixture of SP 30 ␤ plus SP 30 ␥ peptides decreased the ␣␤ T ␥ T -rENaC-associated Na ϩ conductance to 37.6 Ϯ 10.7 S (n ϭ 11), a value statistically indistinguishable from wild-type (p Ͼ 0.4). The individual peptides, at the same final concentration as the mixture, i.e. 276 M, also inhibited the macroscopic inward Na ϩ currents but were not nearly as effective as the peptide mixture. In contrast, neither the individual peptides nor the peptide mixture inhibited wt ␣␤␥-rENaC-associated Na ϩ conductance (Fig. 3).
To study the inhibitory kinetics of the individual peptides and their mixture, increasing concentrations of the SP 30 ␤ and/or SP 30 ␥, varying from 92 to 1732 M, were injected into the cytoplasm of oocytes expressing ␣␤ T ␥-, ␣␤␥ T -, and ␣␤ T ␥ T -rENaC. The concentration-response curves of peptide(s) on amiloride-sensitive Na ϩ currents in oocytes injected with cRNA of rENaC constructs associated with Liddle's syndrome (Fig. 4A) were well described by the Hill equation (Fig. 4B).
The K d values and the corresponding Hill coefficient are summarized in Table I. The Hill coefficients for the individual peptides and the peptide mixture range from 0.7 to 1. These results suggested that the individual C-terminal peptides and their mixture interact with one rENaC subunit or with a single binding site.
Macroscopic Currents of Point Mutants Are Not Inhibited by Peptides-Among the Liddle's ENaCs, both ␤ and ␥ C-terminal truncated (nonsense) mutants and the single amino acid point mutants resulted in greater amiloride-sensitive Na ϩ currents compared with the wild-type construct when expressed in Xenopus oocytes (22). Because the C-terminal peptides inhibited the currents associated with ␣␤ T ␥-, ␣␤␥ T -, and ␣␤ T ␥ T -rENaC expressed in oocytes, the question arose as to whether this specific effect was based on the ability of the peptides to restore the wild-type nature of these truncation mutants through the specific interaction with the truncated channel or by involvement of intracellular secondary components. To test this hypothesis, subunits containing point mutations associated with Liddle's syndrome, ␤ Y618A (␤ Y ) or ␥ Y628A (␥ Y ), combined with the two complementary subunits of wt rENaC, were expressed in oocytes, and the effect of the peptides was examined.
In oocytes injected with wt ␣␤␥-rENaC and point mutant constructs, ␣␤ Y ␥ and ␣␤␥ Y , three parallel groups of oocytes from the same frog were injected on the same day. The amiloride-sensitive conductance in oocytes injected with ␣␤ Y ␥-rENaC and ␣␤␥ Y -rENaC were 45.77 Ϯ 3.8 S (n ϭ 6) and 43.5 Ϯ 3.9 S (n ϭ 7), respectively. In comparison, the chord conductance of wt ␣␤␥-rENaC was 15.7 Ϯ 4.0 S (n ϭ 6, p Ͻ 0.001). The reversal potentials among each group did not show a statistically significant difference, indicating that the cationic selectivity was not modified by point mutagenesis of the tyrosines in the ␤-(in position 618) or ␥-(in position 628) rENaC subunit. The reason why the absolute value of Na ϩ conductance in this batch of oocytes was lower than that measured in the experiments presented in Fig. 3 was because less ENaC cRNA was injected (7.5 versus 25 ng/oocyte).
As shown in Fig. 5, following an injection of either 276 M SP 30 ␤ or SP 30 ␥ into oocytes expressing ␣␤ Y ␥-rENaC, the current amplitudes were slightly increased to 108.91 Ϯ 2.5 and 114.07 Ϯ 1.83%, respectively. For oocytes expressing ␣␤␥ Y -rENaC, the Na ϩ currents were decreased slightly to 85.31 Ϯ 2.4 and 91.53 Ϯ 2.4%, respectively, following ␤or ␥-peptide injection. No significant inhibitory effects of the individual peptides were found (variation less than 15%) on ␣␤ Y ␥and ␣␤␥ Y -rENaC-associated amiloride-sensitive Na ϩ currents in oocytes (Fig. 5). Similarly, injection of the peptide mixture into oocytes expressing either ␣␤ Y ␥or ␣␤␥ Y -rENaC did not significantly affect the current amplitudes (Fig. 5). The above experiments were repeated using the peptides at a concentration of 1732 M; comparable results, i.e. no significant current inhibition (within 15% of the preinjection control), were observed (n ϭ 3 each; data not shown).
Amiloride Blockade of rENaCs Expressed in Oocytes-One of the defining features of ENaC is its sensitivity to the potassium sparing diuretic amiloride, which reversibly blocks the channel at nanomolar to micromolar concentrations (for review see Ref. 7). To identify that the currents were carried by exogenous rENaC expression, inhibition of current by amiloride was studied in rENaC-injected oocytes and in H 2 O-injected oocytes as a control. Under voltage clamp conditions, amiloride-blockable Na ϩ currents could not be detected in H 2 O-injected oocytes (I Ϫ140 ϭ Յ100 nA). In comparison to the small currents observed in uninjected or water-injected oocytes, large amiloridesensitive inward Na ϩ currents of 5-10 A in oocytes injected with cRNA of wt rENaC were seen, indicative of the fact that the currents recorded were indeed produced by exogenous rENaC expression (data not shown).
In order to verify that the currents we observed were produced by exogenous cRNAs of rENaC rather than by activation of an endogenous cation channel, amiloride was applied to outside-out patches. Outside-out patches from oocytes expressing rENaC (wt or mutant varieties) invariably contained multiple channels. Similar continuous recordings were made on oocytes expressing all of the other Liddle's truncation constructs used in these experiments (Fig. 6). When perfused with 100 nM amiloride, the multichannel currents of rENaC and the Liddle's mutants became flickery. After washing out amiloride, the original multichannel traces were recovered, indicating that the inhibitory effect of amiloride was reversible. In contrast, no amiloride-sensitive single channel currents could be seen in 10 outside-out patches of H 2 O-injected oocytes.
In order to determine whether the efficacy of amiloride was changed by any of these truncation mutations, we constructed dose-response curves for amiloride inhibition of channel activity (N⅐P o ) in these outside-out patches (Fig. 6B).  Table I. assay the inhibitory effect of the C-terminal peptides on wt ␣␤␥-rENaC and the Liddle's constructs when expressed in oocytes. Fig. 7 shows the single channel current traces for ␣␤␥ (top) and ␣␤ T ␥ T -rENaC (middle) and their respective I-V curves (bottom) in inside-out patches of oocytes membrane under bi-ionic conditions. The unitary conductances of both wt ␣␤␥ (n ϭ 9) and ␣␤ T ␥ T -rENaC (n ϭ 6) averaged around 7 pS. The reversal potentials under bi-ionic conditions averaged 97 mV, indicating a P Li /P K of 42. The effects of 1000 M peptide mixture on the kinetics of the double-truncated Liddle's mutant (␣␤ T ␥ T ) are shown in Fig. 8. It can be seen that the peptide mixture decreased NP o , but not the conductance of ␣␤ T ␥ T -rENaC, and was completely reversible (Fig. 8C). The inhibition produced by the peptides on ␣␤ T ␥ T -rENaC was very different than that seen for amiloride (cf. Fig.  6A), i.e. the peptides did not appear to act as open channel blockers.
If the C-terminal regions of the ␤and ␥-ENaC subunits act as intrinsic gating particles, exogenously added ␤ and ␥ Cterminal peptides should not block wt ␣␤␥-rENaC because of the presence of intact ␤and ␥-ENaC C termini. The effect of the peptide mixture on wt ␣␤␥-rENaC was thus explored following application of peptides to inside-out patches (Fig. 8A).
In contrast to ␣␤ T ␥ T -rENaC, the C-terminal peptides had no effect on the gating of wt ␣␤␥-rENaC currents nor did 1000 M Pro-SP 30 ␤ peptide produce any inhibition of ␣␤ T ␥ T -rENaC (Fig.  8B), consistent with the macroscopic current recordings (Fig.  1). The single channel activity (N⅐P o ) and P o of each group are summarized in Fig. 9. N⅐P o of ␣␤ T ␥ T was approximately 3-fold greater than wt ␣␤␥-rENaC at a holding potential of Ϫ40 mV, consistent with whole-cell current measurements. The corresponding P o of ␣␤ T ␥ T -rENaC increased by 41.5% compared with wt ␣␤␥-rENaC.
Effect of SP 30 ␤:SP 30 ␥ Peptides on Single Channel Current of Point Mutants- Fig. 10 presents representative single channel current traces for wt ␣␤␥-, ␣␤ Y ␥-, and ␣␤␥ Y -rENaC in cellattached patches of Xenopus oocyte membranes. The Liddle's point mutants expressed in oocytes had similar biophysical characteristics to those of wt ␣␤␥-rENaC. The channel number (N), the single channel activity (N⅐P o ), and mean P o are summarized in Table II. N⅐P o of ␣␤ Y ␥ and ␣␤␥ Y were 2.65-and 2.18-fold over that of wt ␣␤␥-rENaC, respectively, at a holding potential of Ϫ40 mV. The corresponding P o of ␣␤ Y ␥ and ␣␤␥ Y increased only slightly, namely by 6.8 and 4.5%, respectively (Table II), values not significantly different than that for the wild-type channel. The number of channels per patch and the single channel activity of ␣␤ Y ␥and ␣␤␥ Y -rENaC on the other hand were significantly increased compared with wt rENaC, p Ͻ 0.01.
The effect of peptides on ␣␤ Y ␥and ␣␤␥ Y -rENaC was also assessed using the inside-out patch configuration. As shown in Fig. 11, no change in channel activity was recorded for both ␣␤ Y ␥and ␣␤␥ Y -rENaC. Unlike the truncated Liddle's mutant, the number of channels per patch, the single channel activity, and the P o were not affected by application of the peptide mixture to the cytosolic side of the patch from the oocytes expressing ␣␤ Y ␥or ␣␤␥ Y -rENaC ( Fig. 12 and Table II). Taken together, these data indicated that the C-terminal peptides inhibited the amiloride-sensitive macroscopic Na ϩ currents and the single channel activity of ␣␤ T ␥ T -rENaC but not those of ␣␤ Y ␥and ␣␤␥ Y -rENaC, consistent with a role for the C terminus acting as an intrinsic inhibitory particle of ENaC. DISCUSSION The hypothesis that the cytosolic tails of ␤or ␥-ENaC subunits may function as inherent inhibitory particles involved in the normal gating behavior of epithelial Na ϩ channels arose from our previous observations made in planar lipid bilayers that synthetic peptides made up of the last 10 or the last 30 amino acids of either ␤or ␥-ENaC could inhibit Na ϩ channels whose P o was increased because they possessed Liddle's truncation mutations (40,41). In the present study, we extended these initial observations to the oocyte expression system, using a combination of two-electrode voltage clamping and excised patch clamping, to confirm the inhibitory actions of these peptides on ENaCs containing truncations in their ␤ and/or ␥ subunits when expressed in a more native environment.
Expression of Liddle's Mutants-Two non-mutually exclusive mechanisms have been proposed to account for the basalactivated, macroscopic amiloride-sensitive Na ϩ currents induced by expression of the Liddle's mutants: the retrieval theory (i.e. decreased channel internalization) and/or enhanced single channel activity (42). Identification of specific Nedd4 and spectrin binding domains in the cytosolic region of the ENaC subunits (43,44) and the interactions of ENaC with actin and other cytoskeletal elements (45) supported the idea that the fundamental problem in Liddle's disease could be inefficient internalization of the channel from the cell surface. Furthermore, observations by Snyder et al. (24) that suggested that Ͼ90% of the basal-activated Na ϩ currents are correlated to an increase in ENaC expression at the cell surface. This finding was verified in oocytes expressing a rENaC construct containing a truncated version of the ␤ subunit (R564X). Cell surface expression of ␣␤ R564X ␥-rENaC was increased over wt rENaC expression, corresponding to a 5.6-fold increase in ENaC currents (25). However, we previously reported that a 35% increase in the fluorescence intensity in oocytes injected with ␣␤ R564X ␥-rENaC was accompanied by a 4.4-fold increase in amiloride-sensitive Na ϩ current (26). Thus, a smaller increase in cell surface channel density was associated with a large increase in current, suggesting that the retrieval theory alone did not adequately explain the large macroscopic Na ϩ current seen in Liddle's mutants (46). It is possible that the differences between these studies could be due to different responses produced by rENaC and hENaC (24).
The idea that ϳ90% of the ENaC channels at the cell surface are inactive and only 10% are active (27) underscores the methodological limitations in quantitatively determining not only active ENaC channel number at the cell surface but also single channel P o . Electrophysiological recordings, including two-electrode voltage and patch clamping, report only electrically active ENaC channels at the cell surface and cannot reveal the existence of quiescent channels within the plasma membrane. The assessment of single channel activity, specifically open probability, can only be done if the total number of channels (active plus inactive) per patch is known. The relationship between macroscopic currents and single channel activity is given by Ī ϭ i⅐N⅐P o where Ī is the mean current amplitude; i indicates the amplitude of the unitary current; N represents the number of active channels at the cell surface, In the absence of amiloride, each patch contained around five channels for the Liddle's mutations but less than three channels for wt ␣␤␥-rENaC. Superfusion of amiloride from 0 to 10 M to the extracellular side of rENaC channels caused the current traces to flicker due to decrease in open time constant, especially when the concentrations of amiloride are above 100 nM. The blocking of amiloride on rENaCs was reversible and concentration-dependent. For illustration purposes, the effects of only a single concentration of amiloride are shown, and only partial records are presented for simplicity. B, log dose-response curves of amiloride inhibition of wt ␣␤␥ and Liddle's mutations. The data are expressed as percent inhibition of NP o determined from channel activity in outside patches of oocyte membrane. The data points and the error bars represent the mean and the S.E. for at least 3 separate experiments. and P o is the open probability of the channel. Therefore, in a case where the majority of channels are quiescent, the true P o will be very low. Thus, a significant component of the basalactivated ENaC currents that is not due to increased ENaC channel expression at the cell surface can result from altered single channel behavior. Such alterations may include an increase in P o or an increase in single channel unitary conductance.
The findings of the present study showed that the unitary conductances for both the truncation mutants (␣␤ T ␥ T -rENaC) and the missense Liddle's mutants (␣␤ T ␥and ␣␤␥ T -rENaC) did not differ from that of wt ␣␤␥-rENaC, in accordance with the observations of others (23,24). However, the electrically active ENaC number at the cell surface of Liddle's mutants described as the number per patch were 1.8 -2.5-fold over that of wt ␣␤␥-rENaC. Significantly, in our experiments, P o for ␣␤ T ␥ T -rENaC was 40% higher as compared with wt ␣␤␥-rENaC (Table II). The single channel activity (N⅐P o ) for all the mutant channels showed a ϳ3-fold increase, paralleling the macroscopic ENaC currents in oocytes expressing ENaC-containing Liddle's mutations. Moreover, the P Li /P K permeability ratios for all of the mutant channels were comparable to those of wt ␣␤␥-rENaC expressed in oocytes or of the native ENaC in epithelia (47)(48)(49), as was the amiloride sensitivity. Thus, like Firsov et al. (25) and Ismailov et al. (32), we propose the open probability of Liddle's mutations producing truncations of the ␤ or ␥ subunit carboxyl tails is also an important parameter contributing to enhanced macroscopic ENaC currents seen in this disease.
It is important to reconcile our findings with those of other laboratories claiming that single channel P o does not change when a ␤ or ␥ subunit truncating Liddle's mutation is present in ENaC (24, 50). As indicated above, analyses of P o from single channel data recorded by patch clamp are difficult because the true number of channels in the patch must be known. Moreover, appropriate observation times must also be attained. Fyfe and Canessa (50) concluded that subunit composition of ENaC primarily determines channel kinetics, yet only two, not three, subunit channels were used in their studies. As they and others (50 -53) indicate, the P o of ENaC is quite variable, ranging from 0.01 to 0.9. This variability in basal P o reflects uncertainties in channel number within any given patch. If the channel number is underestimated, for example, due to the presence of quiescent channels, then higher values of P o will be calculated. More reliable estimates of P o for wild-type ENaC would be obtained because there would be on average fewer quiescent channels present. In our experiments, wide variability in P o was not observed from patch-to-patch. Although our experiments also suffer from lack of certainty of the absolute number of channels per patch, we only performed our analyses on multichannel patches (2-7 channels), not "single" channel patches, so that a more representative (and potentially more reliable) determination of channel number could be made.
Peptide Inhibition of Liddle's Mutants-By using the wholecell configuration of the patch clamp technique, the ability of a 10-amino acid polypeptide (200 M) identical to either the cytosolic tail of ␤-rENaC or ␤-hENaC to inhibit amiloride-sensitive inward Na ϩ currents in lymphocytes of Liddle's disease was found (41). This finding was confirmed in planar lipid bilayers reconstituted with the Na ϩ channel complex immunopurified from lymphocytes derived from patients affected with Liddle's disease (32). Unresolved questions concern the molecular mechanism underlying peptide inhibition of ENaC and whether peptides comprising the missing segments of the intracellular C-terminal tails of the ␤ or ␥ subunit can inhibit basal-activated Na ϩ currents caused by both single and double truncation mutations (␣␤ T ␥-rENaC, ␣␤␥ T -rENaC, and ␣␤ T ␥ T - rENaC) and missense mutations (␣␤ Y ␥and ␣␤␥ Y -rENaC) identified in Liddle's syndrome cases. In the present study, a combination of the two peptides markedly inhibited ␣␤ T ␥ T -rENaC-associated amiloride-sensitive Na ϩ currents in a dosedependent fashion. Whereas the individual peptides were also inhibitory, their efficacy was much reduced (Fig. 4). Expression of only one truncation mutant of either the ␤ subunit (␤ T ) or the ␥ subunit (␥ T ) combined with complementary wt ENaC subunits also induced a 3-fold increase in macroscopic amiloridesensitive Na ϩ currents (Refs. 23, 24, and 27 and present study). The significantly lower inhibitory effect of the individual peptides compared with the peptide mixture on ␣␤ T ␥ T -rENaC suggests that both the ␤ and ␥ subunit C termini are required for efficient inhibition of an activated ENaC. This idea is supported by our data showing that the efficacy of either SP 30 ␤ or SP 30 ␥ used individually was enhanced in the single subunit truncation mutant (␣␤␥ T and ␣␤ T ␥ constructs as compared with the double truncation constructs (␣␤ T ␥ T )). Furthermore, the fact that these peptides did not block the point mutant channels (␣␤ Y ␥ or ␣␤␥ Y ) and the fact that these channels did not have an increased P o is consistent with the argument that the cytoplasmic C-terminal tails of ␤and ␥-ENaC are involved in the normal gating of ENaC. We hypothesize that there is ␤-sheet formation between the C-terminal tails of ␤ and ␥ that, in essence, forces the tails to act as a single binding peptide. Evidence supporting such an interaction comes from intrinsic fluorescence and circular dichroism studies of these synthetic peptides in solution (33). Measured Hill coefficients of 1 for the peptide mixture are also consistent with this hypothesis.
The specificity of the peptide inhibition on Liddle's ENaC (truncation mutants) was verified in experiments using non-ENaC peptides (e.g., a 14-amino acid-mer of CFTR or Pro-SP 30 ␤). Cytoplasmic injection or perfusion of either of these peptides at the same concentrations used for the peptide mixture did not cause any decrease of ␣␤ T ␥ T -rENaC-associated amiloride-sensitive Na ϩ currents. Because rENaC activity is down-regulated by hypo-osmotic stress (14) and peptide injection may have increased the oocyte volume (the cell volume increased ϳ13% assuming the mean liquid volume of oocyte is 500 nl), it is also possible that the decreased channel activity could be due to oocyte swelling. However, injection of the same volume of distilled water had no effect on the current amplitude (Fig. 1). The observations of Kellenberger and colleagues (27) excluded the possibility that the inhibitory effects of the Cterminal peptides of rENaC on the basal-activated currents of Liddle's mutations result from rundown. Their examination of the rundown phenomenon suggested that the wt rENaC-associated, rather than the Liddle's mutant-associated, Na ϩ cur- Continuous single channel records (inside-out patches) of wt ␣␤␥-rENaC (A) or ␣␤ T ␥ T -rENaC (B and C) expressed in oocytes are shown. Excised inside-out patches were prepared in order to apply the C-terminal peptides to the cytosolic side of the rENaC channels. The desired concentration of peptide was added directly to a small (0.3 ml) chamber. The effect of SP 30 ␤ ϩ SP 30 ␥ on the wt (A) and ␣␤ T ␥ T channel (C) and the lack of effect of 200 M Pro-SP 30 ␤ peptide on ␣␤ T ␥ T -rENaC (B) are depicted. The holding potential was Ϫ40 mV. Each condition was replicated six times. rent shows a time-dependent decrease (by 27% in 25 min). In the present study, the C-terminal peptides did not exert any effect on wt rENaC or the point mutant constructs (␣␤ Y ␥, ␣␤␥ Y -rENaC), indicating that they did not have untoward nonspecific effects in this expression system. The amiloride-sensitive Na ϩ currents generally achieved stabilization after ϳ10 min of perfusion under our experimental conditions, and the rundown of wt rENaC currents was correlated to the accumulation of intracellular Na ϩ (27). Comparable experiments were performed on excised inside-out patches of oocyte membranes containing active ENaC.
Our data show that the ␣␤ T ␥ T -rENaC number per patch, P o , and thus overall channel activity (N⅐P o ) were inhibited to wildtype levels by the peptide mixture, similar to that seen in the macroscopic experiments. The inhibition of the channel by peptide occurs only from the cytoplasmic side and is dissimilar to that produced by amiloride, in that a "flickering"-type block was not evident. These results are essentially identical to those seen in the bilayer, in that increasing concentrations of the peptides increase a second slow component of channel closed time distribution, leaving the open time distribution unaffected (33), again consistent with the reconstituted planar lipid bilayer system (33).
Deletion of the C termini of the ␤and ␥-ENaC subunits leads to loss of the PPPXY conserved sequence. The PPPXY motifs located within the C termini have been found to play an  important physiological role in controlling Nedd4-induced Ca 2ϩdependent retrieval and clathrin-mediated endocytosis of ENaC (44,53). Thus, the possibility that the applied C-terminal peptides can function as a sensor for ENaC internalization signaling pathways and thereby result in a decrease in the number of channels per unit at the cell surface cannot be excluded. More likely, however, is that the decrease in the active channel number per patch seen subsequent to peptide addition is due to a decrease in the fraction of the electrically detected rENaC per unit of cell surface, simply by a reduction in P o . Because the C-terminal domains of the ␤and ␥-ENaC subunits contain phosphorylation sites for protein kinase A and protein kinase C (8), in vivo the C-terminal regions may interact with some target for protein kinase A-or protein kinase C-mediated phosphorylation on rENaC (subunits or associated cytoskeletal elements) and therefore regulate the activity of rENaC by altering its phosphorylation state (32,46). The enhanced sensitivity to the activating effects (via protein kinase A) of normal endogenous cAMP levels in affected lymphocytes of Liddle's disease (41) is consistent with this possibility. Also, based on the finding that the channel activity of truncated Liddle's mutants is no longer down-regulated in a feedback manner by the increasing intracellular Na ϩ concentration and that Ͻ10% of ENaC population located at the cell surface is in the active channel pool, it has been proposed that a greater fraction of "silent" rENaCs has escaped from the inactive pool due to the rise in cytosolic Na ϩ concentration and thus has contributed to the increased electrically detected number of channels per patch (27).
In summary, our results demonstrate that both a greater ENaC number of channels at the cell surface and an increased P o contribute to the increased basal ENaC currents associated with expression of ENaC truncation mutants identified in Liddle's disease. In addition, C-terminal peptides structurally identical to the cytoplasmic tails of individual ENaC subunits afforded us an opportunity to examine the regulatory function imposed by the C-terminal tails of rENaC and thus further elucidate the pathogenesis of Liddle's disease. The results in the present study confirmed that the constitutively activated Na ϩ absorption through truncation mutants of the ENaC channel located on the apical membrane of kidney collecting duct correlates to the loss of the inhibitory tails of the ␤ and/or ␥ subunits. Physical addition of the missing regions for both subunits of the truncation Liddle's mutations could achieve successful correction of currents in vitro, both at the macroscopic and single channel levels. Combined with similar observations obtained from patch clamping on human B lymphocytes (41), dual-electrode voltage clamping on oocytes expressing ENaC (26), and planar lipid bilayer incorporation experiments (32,33), our results provide an experimental basis for understanding inherent ENaC function. Our results also support the idea that different Liddle's mutations, e.g. truncation versus point mutations, produce enhanced macroscopic Na ϩ currents through different mechanisms.  Fig. 10. The data are expressed as mean Ϯ S.E. The averaged results of N⅐P o , P o , and N in the absence of the peptide mixture (control) for ␣␤ Y ␥-rENaC (left panels) are 2.87 Ϯ 0.37, 0.44 Ϯ 0.03, and 6.53 Ϯ 0.36, respectively (n ϭ 5). Similar results are observed for ␣␤␥ Y -rENaC (right panels). The cytoplasmic membrane in inside-out patches perfusion with the peptide mixture (100 M) seems not to influence the preceding parameters of ␣␤ Y ␥-(left) and ␣␤␥ Y -rENaC (right).