J Biol Chem, Vol. 274, Issue 46, 32889-32896, November 12, 1999

The NH₂ Terminus of the Epithelial Sodium Channel Contains an Endocytic Motif^*

Michael L. Chalfant, Jerod S. Denton, Anne Lynn Langloh^§, Katherine H. Karlson, Johannes Loffing^¶, Dale J. Benos^§, and Bruce A. Stanton

From the  Department of Physiology, Dartmouth Medical School, Hanover, New Hampshire 03755, the ^§ Department of Physiology and Biophysics, University of Alabama, Birmingham, Alabama 35233, and the ^¶ Institute of Anatomy, University of Zurich, Zurich, Switzerland CH-8057

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

An epithelial sodium channel (ENaC) is composed of three homologous subunits: , , and . To elucidate the function of the cytoplasmic, NH₂ terminus of rat ENaC (rENaC) subunits, a series of mutant cDNAs was constructed and the cRNAs for all three subunits were expressed in Xenopus oocytes. Amiloride-sensitive Na⁺ currents (I_Na) were measured by the two-electrode voltage clamp technique. Deletion of the cytoplasmic, NH₂ terminus of  (2-109),  (2-49), or -rENaC (2-53) dramatically reduced I_Na. A series of progressive, NH₂-terminal deletions of -rENaC were constructed to identify motifs that regulate I_Na. Deletion of amino acids 2-46 had no effect on I_Na: however, deletion of amino acids 2-51, 2-55, 2-58, and 2-67 increased I_Na by ~4-fold. By contrast, deletion of amino acids 2-79, 2-89, 2-100, and 2-109 eliminated I_Na. To evaluate the mechanism whereby 2-67--rENaC increased I_Na, single channels were evaluated by patch clamp. The single-channel conductance and open probability of ,,-rENaC and 2-67-,,-rENaC were similar. However, the number of active channels in the membrane increased from 6 ± 1 channels per patch with ,,-rENaC to 11 ± 1 channels per patch with 2-67-,,-rENaC. Laser scanning confocal microscopy confirmed that there were more 2-67-,,-rENaC channels in the plasma membrane than ,,-rENaC channels. Deletion of amino acids 2-67 in -rENaC reduced the endocytic retrieval of channels from the plasma membrane and increased the half-life of the channel in the membrane from 1.1 ± 0.2 to 3.5 ± 1.1 h. We conclude that the cytoplasmic, NH₂ terminus of -, -, and -rENaC is required for channel activity. The cytoplasmic, NH₂ terminus of -rENaC contains two key motifs. One motif regulates the endocytic retrieval of the channel from the plasma membrane. The second motif is required for channel activity.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

An amiloride-sensitive, epithelial sodium channel (ENaC)¹ mediates Na⁺ transport across the apical membrane of a variety of epithelia including the kidney, lung, and intestine and, thereby, plays a vital role in maintaining Na⁺ and fluid homeostasis (1-4). ENaC is composed of three subunits: , , and  (5, 6). The expression of the  subunit in Xenopus oocytes produces very small currents, and the expression of  and/or  subunits generates no current (5, 6). However, coexpression of -, -, and -rENaC produces large Na⁺ currents in oocytes (5, 6). The ENaC subunits are members of a growing family of ion channels that include the FMRFamide-gated Na⁺ channel, Na⁺ channels in brain (BNC1 and BNC2), and the degenerins of Caenorhabditis elegans that encode mechanosensitive channels (e.g. DEG-1, MEC-4, and MEC-10) (4, 7).

Amino acid sequence analysis and biochemical studies suggest that the ENaC subunits have cytoplasmic NH₂ and COOH termini, two hydrophobic transmembrane domains (M1 and M2) and a large extracellular domain (4-8). Several lines of evidence suggest that the -, -, and -subunits interact to form a heteromultimeric channel complex and that this complex is required for maximum Na⁺ currents (4-8). However, relatively little is known about the function of the different domains of ENaC. The region immediately preceding the second transmembrane domain contains an amiloride-binding site and determines ion selectivity, suggesting that it forms part of the channel pore (9-11). The extracellular domain plays a role in subunit interaction, targeting channels to the plasma membrane and channel gating (4, 7-9, 11). The COOH terminus plays an important role in localizing ENaC to the apical membrane, and it contains an NPXY motif that is important in the endocytic retrieval of the channel from the plasma membrane (1-4, 12-15). For example, in Liddle's syndrome mutations and/or deletions of the NPXY motif in - or -ENaC reduces subunit ubiquitination and endocytic retrieval of ENaC from the plasma membrane. This results in an increase in the number of channels in the membrane, which causes hyperabsorption of Na⁺ and hypertension (12-15). Although recent studies suggest that the NH₂ terminus may play a role in channel assembly and gating (16-18), a complete understanding of the function of the NH₂ terminus of any ENaC subunit is not available. Thus, the goal of this study was to test the hypothesis that the NH₂ termini of ENaC's are critical for channel activity. To elucidate the function of the cytoplasmic, NH₂ terminus of each rENaC subunit, a series of mutant cDNAs was constructed and the cRNAs were expressed in Xenopus oocytes. Amiloride-sensitive Na⁺ currents (I_Na) were measured by the two-electrode voltage-clamp technique. We report that the cytoplasmic, NH₂ terminus of -, -, and -rENaC is required for channel activity. Our data demonstrate that the NH₂ terminus of -rENaC contains two key regions: one that regulates the half-life of the channel in the plasma membrane. Deletion of this motif increases I_Na by reducing the rate of channel endocytosis. The second region is required for normal channel activity. Deletion of this region eliminates I_Na.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

cDNA Constructs-- Plasmids containing cDNAs encoding the wt -, -, and -subunits of rENaC, cloned into the pSport vector, were a generous gift of Dr. Bernard C. Rossier, Lausanne, Switzerland (5, 19). Deletions of the cytoplasmic, NH₂ termini of -, -, and -rENaC were made by PCR-based mutagenesis. Because all deletions were generated using a similar approach, details will be given for only one construct. The sequence of all PCR products and cDNAs with deletions was confirmed by dye terminator cycle sequencing (ABI PRISM: PE Applied Biosystems, Foster City, CA).

pSport/2-41--rENaC, encoding a protein in which the NH₂-terminal 41 amino acids of -rENaC were deleted, was constructed in two steps. First, a 476-bp PCR fragment was synthesized using pSport/-rENaC cDNA as a template, with a sense primer (CGACGTCGACCATGCAAGGACTGGGGAAGGGGGAC-3') corresponding to nucleotides 127-147 of -rENaC and an antisense primer (CTGGCGAGTGTAGGAAGAGTTGTA) corresponding to nucleotides 565-588. The sense primer also contained an upstream SalI restriction site, a partial Kozak consensus sequence, and an initiator methionine codon. The antisense primer was located downstream of a unique BsrGI restriction site. The 476-bp PCR product was isolated, purified (Wizard, Promega, Madison, WI), subcloned into pcR 2.1 (TA Cloning Kit, Invitrogen), and sequenced. In the second step, pSport/-rENaC and pcR 2.1 containing the 476-bp PCR product were digested with SalI/BsrGI, and the gel-purified 440-bp PCR fragment was ligated into digested pSport/-rENaC.

To monitor the expression of rENaC in the plasma membrane the 5' end of -rENaC and -rENaC cDNA were ligated in-frame to the 3' end of the cDNA encoding the enhanced green fluorescent protein (EGFP-C1 and EGFP-C2: CLONTECH, Palo Alto, CA). GFP--rENaC was constructed in two steps. First, -rENaC was excised from pSport/-rENaC with SalI/KpnI, and the excised fragment was ligated into SalI/KpnI digested pEGFP-C1 to construct pEGFP-C1--rENaC (i.e. GFP--rENaC). To remove two pre-existing stop codons between the SalI site and an initiator methionine in pEGFP-C1--rENaC a 563-bp PCR fragment was synthesized using -rENaC cDNA as a template with a sense primer (ACGCGTCGACGGTGCCACCATGCCAGTGAAGAAGTACCT) corresponding to nucleotides 1-20 of -rENaC and an antisense primer (GGTGCTTCCTGGGGCTGGGTTGCTGCTGTT) corresponding to nucleotides 514-543 of -rENaC. The sense primer also contained an upstream SalI restriction site and a Kozak consensus site sequence. The PCR product contained a unique BsmBI restriction site just upstream of the anitsense primer sequence. The 563-bp PCR product was isolated, purified (Wizard, Promega), and subcloned into pcR 2.1 (TA Cloning Kit, Invitrogen) for sequencing. In the second step, pEGFP-C1--rENaC and pcR 2.1 containing the 563-bp PCR product were digested with SalI/BsmBI and the gel-purified 316-bp PCR fragment was ligated into digested pEGFP-C1--rENaC. Subsequently, GFP--rENaC was subcloned from pEGFP-C1--rENaC into pcDNA3.1 (Invitrogen, Carlsbad, CA) using NheI/KpnI.

GFP--rENaC was constructed by excising -rENaC from pSport/-rENaC with SalI/KpnI, and ligating the excised fragment into SalI/KpnI digested pEGFP-C2. GFP--rENaC was subcloned from pEGFP-C2--rENaC into pcDNA3.1 using NheI. pcDNA3.1 was digested with NheI and calf intestinal alkaline phosphatase-treated to prevent self-ligation (to generate pcDNA3.1/EGFP/-rENaC).

cRNA Preparation-- pSport vectors were linearized with NotI and pcDNA3.1 vectors were linearized with AfflI. The linearized cDNAs were used as a template for cRNA synthesis using a kit containing T7 RNA polymerase, ribonucleotides, and a 7-methylguanosine cap analog following the manufacturer's instructions (mMessage mMachine, Ambion Inc., Austin TX).

Isolation of Xenopus Oocytes and Injection of cRNA-- Oocytes were isolated and injected with cRNA as described previously (20, 21). Briefly, ovarian lobes were removed from Xenopus laevis and stored in calcium-free OR-2 solution. Oocytes were isolated and defolliculated using a combination of enzymatic treatment and manual dissection. Defolliculated Stage V and VI oocytes were transferred to L-15 medium modified for use with amphibian cells and supplemented with gentamycin sulfate. cRNA transcribed from wt or truncated -, -, and -rENaC cDNAs was injected into oocytes as described under "Results."

Two-electrode Voltage Clamp-- As described in detail elsewhere (21), the amiloride-sensitive current (I_Na) was measured in oocytes 1-3 days after injection of cRNA using the two-microelectrode voltage-clamp technique in oocytes bathed in a solution containing (in mM): NaCl, 110; KCl, 2; CaCl₂, 0.4; MgCl₂, 1.0; HEPES, 5, pH 7.4 (21). All experiments were performed at 22-24 °C. Unless otherwise noted oocytes were co-injected with equal amounts of -, -, and -rENaC cRNA (2.5 ng/subunit). I_Na was measured as the difference in the whole cell current at 100 mV before and after amiloride (100 µM), a concentration sufficient to completely inhibit rENaC currents (22).

Single-channel Currents-- Single channel currents were measured in the cell-attached mode using the patch-clamp technique in de-vitellated oocytes as described in detail (21, 23). Currents were amplified, filtered with a low-pass, 4-pole Bessel filter with a cutoff frequency of 200 Hz, digitized at a sampling rate of 2 kHz, and stored on the hard disc of a DOS-based computer for subsequent analysis using pCLAMP software version 6.03. Single-channel currents were measured at voltages from 40 to 100 mV (21, 24). P_o, the single channel open probability, and n, the number of active channels in the membrane, were calculated as described in detail (24).

Immunofluorescent Localization of ENaC in Oocytes-- To examine the effect of NH₂-terminal truncations of -rENaC on the expression of rENaC in the plasma membrane, stage V or VI oocytes were injected with cRNAs coding for wt, 2-67--rENaC, or 2-109--rENa (4 ng/subunit) in combination with GFP--rENaC (4 ng/subunit) and GFP--rENaC (4 ng/subunit). Four groups of oocytes (10 oocytes/group) were studied: 1) wt--rENaC + GFP--rENaC + GFP--rENaC; 2) 2-67--rENaC + GFP--rENaC + GFP--rENaC; 3) 2-109--rENaC + GFP--rENaC + GFP--rENaC, and 4) water. 48 h after cRNA injection the cellular localization of GFP-tagged rENaC was determined by laser scanning confocal microscopy as described in detail previously (25). Proteins in the plasma membrane were biotinylated using EZ-Link Sulfo-NHS-LC-Biotin (Pierce, Rockford, IL). Subsequent to biotinylation, oocytes were washed in ND-48 and the biotin was labeled at 4 °C with stepavidin conjugated to Texas Red (Molecular Probes, Eugene, OR). Confocal images were acquired using an Olympus Flouview BX50 upright microscope equipped with an air-cooled krypton/argon laser scanning head and viewed with a UplanF1 × 10X/0.3 NA air objective. GFP fluorescence was excited using the 488-nm laser line and collected using a standard fluorescein isothiocyanate filter set (530 ± 30 nm). Texas red fluorescence was excited using the 568-nm laser line and collected using a standard Texas Red filter set (605 ± 32 nm). XY scans were obtained at 12-bit resolution at approximately the midpoint of each oocyte. All images were acquired, processed (Adobe Photoshop 5.0), and printed using the same settings in each of the three groups of oocytes.

Statistical Analysis-- Differences between means were compared by ANOVA and the Bonferroni post hoc comparison test, or the paired or unpaired Student's t test, as appropriate. Statistical analyses were performed with the InStat statistical software package (Graphpad, San Diego, CA). Data were expressed as the mean ± S.E. p < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

-, -, and -Subunits of rENaC Are Required for Maximum ENaC Currents-- As illustrated in Fig. 1, expression of -rENaC in Xenopus oocytes, produced a small amiloride-sensitive I_Na. Co-expression of -rENaC with -rENaC or -rENaC elicited larger currents compared with -rENaC alone. Maximum currents were expressed when oocytes expressed all three rENaC subunits. In the absence of -rENaC, -rENaC and -rENaC did not produce a current in oocytes (data not shown). These data confirm earlier studies (5, 22).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   All three subunits of rENaC are required for the expression of maximum, amiloride-sensitive, sodium currents in oocytes. Oocytes were injected with cRNA (2.5 ng/subunit/oocyte) and the INa (100 µM) was measured at 100 mV in experiments presented in this and all subsequent figures. The number of oocytes per group was between 8 and 14. Asterisks indicate p < 0.001 versus wt-, , and  and, in addition, that the current was significantly different from a value of 0 µA.

The Amino Terminus of -, -, and -rENaC Is Required to Produce Maximum Amiloride-sensitive Na+ Currents-- The cytoplasmic, amino (NH₂) termini of -rENaC (amino acids 1-109), -rENaC (amino acids 1-49), and -rENaC (amino acids 1-53) have regions of highly conserved sequences among species. This suggests that the NH₂ terminus of all three subunits may subserve important functional roles. To test this hypothesis we deleted the NH₂-terminal, cytoplasmic domain (N) of each rENaC subunit and examined the effect of the truncation on I_Na. In oocytes expressing wt--rENaC, wt--rENaC, and wt--rENaC, the I_Na was 4.0 ± 0.7 µA (Fig. 2). Truncation of the NH₂ terminus of -, -, or -rENaC dramatically reduced I_Na (Fig. 2). For example, in oocytes expressing N-rENaC with wt--rENaC and wt--rENaC we could not detect an amiloride-sensitive current (Fig. 2). Similarly, in oocytes expressing wt--rENaC with N-rENaC and N-rENaC no amiloride-sensitive current could be detected (Fig. 2). Co-expression of wt--rENaC with N-rENaC and wt--rENaC dramatically and significantly reduced I_Na compared with oocytes expressing -, -, and -rENaC (p < 0.001). Finally, co-expression of wt--rENaC with wt--rENaC and N-rENaC also dramatically and significantly reduced I_Na compared with oocytes expressing -, -, and -rENaC (p < 0.001). These data demonstrate that the NH₂ termini of -, -, and -rENaC are required to produce maximum I_Na in oocytes.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Truncation of the NH2 terminus of -, -, or -rENaC reduces the amiloride-sensitive current. The number of oocytes per group was 5-14. Asterisks indicate INa significantly different from ,,-rENaC (p < 0.001). N lacks amino acids 2-109, N lacks amino acids 2-49, and N lacks amino acids 2 to 53.

NH₂-terminal Deletion of - and -rENaC Produces a Dominant Negative Mutant-- To determine if truncation of the NH₂ terminus of each rENaC subunit eliminated the contribution of that subunit to a functional channel or if the mutant subunit interacted with the wild type subunits, in a negative way, we examined the effect of co-expressing NH₂-terminal truncated subunits with only one or two wild type subunits. In oocytes co-expressing wt--rENaC and wt--rENaC, I_Na was 222 ± 25 nA (Fig. 3). Co-expression of N-rENaC with wt--rENaC and wt--rENaC significantly reduced I_Na to 135 ± 24 nA (p < 0.02: Fig. 3). By contrast, coexpression of wt--rENaC with wt--rENaC and wt--rENaC dramatically and significantly increased I_Na (Fig. 1). Thus, N-rENaC interacts negatively with wt--rENaC and wt--rENaC. In oocytes expressing wt--rENaC and wt--rENaC I_Na was 84 ± 18 nA (Fig. 3). Co-expression of N-rENa with wt--rENaC and wt--rENaC did not significantly alter I_Na (Fig. 3). Thus, N-rENaC did not functionally interact with wt--rENaC and wt--rENaC. By contrast, coexpression of wt--rENaC with wt--rENaC and wt--rENaC dramatically increased I_Na (Fig. 1). Finally, in oocytes expressing wt--rENaC alone I_Na was 19 ± 2 nA (Fig. 3). Co-expression of N-rENaC and N-rENaC with wt--rENaC reduced I_Na to 5 ± 2 nA (Fig. 3). By contrast, co-expression of wt--rENaC and wt--rENaC and wt--rENaC dramatically increased I_Na (Fig. 1). Thus, N-rENaC and N-rENaC interact negatively with wt--rENaC.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Co-expression of NH2-terminal truncated rENaC subunits with wt-subunits inhibits INa. The number of oocytes per group was 8-14. Asterisks indicate that INa is significantly different from the data immediately above (p < 0.001). Because deletion of the NH2 terminus of -rENaC (N) eliminated INa (5 ± 13 nA: p = NS) we did not co-express N with N and N.

Recently it was demonstrated that ,,-rENaC channels and ,-rENaC channels are more permeable to Li⁺ than to Na⁺, whereas ,-rENaC channels are more permeable to Na⁺ than to Li⁺ (22). To determine if co-expression of N-rENaC with wt-ENaC subunits not only reduced I_Na but also affected cationic permeability, the relative permeability of rENaC channels to Li⁺ versus Na⁺ (I_Li/I_Na) was determined by calculating the ratio of amiloride-sensitive current in a Li⁺-containing bath solution (100 mM LiCl) to the amiloride-sensitive current in a Na⁺-containing bath solution (100 mM NaCl). In oocytes expressing -, -, and N-rENaC, the I_Li/I_Na was 0.81 ± 0.06, indistinguishable from that observed with ,-rENaC (0.79 ± 0.04; p > 0.5). These observations are consistent with the conclusion that whereas -rENaC affects cation permeability, N-rENaC does not. In oocytes expressing -, N-, and -rENaC the I_Li/I_Na was 1.15 ± 0.06, indistinguishable from that observed with ,-rENaC (1.15 ± 0.05; p > 0.9). Thus, although truncation of the NH₂ terminus of -rENaC decreased I_Na when co-expressed with - and -rENaC, this truncation did not affect the permeability of the channel to Li⁺ over Na⁺. Taken together these observations are consistent with the conclusion that N-rENaC combines with the - and -subunit channel complex.

Overexpression of -, -, and -rENaC Modulates I_Na-- The data presented in Fig. 3 demonstrate that deletion of the cytoplasmic, NH₂ terminus of  and perhaps -rENaC produced a subunit that reduced I_Na. Moreover, studies on other multimeric ion channels have shown that the interaction of nonfunctional channel subunits with functional ones disrupts channel function and/or expression in the plasma membrane (26-28). In addition, overexpression of the cytoplasmic NH₂ terminus of degenerins, which are closely related to ENaC, inhibits channel activity (16). Thus, to determine if overexpression of wt or N subunits of rENaC affects I_Na, oocytes were injected with equal amounts of cRNA for the -, -, and -subunits (2.5 ng/subunit/oocyte) plus a 10-fold excess (25 ng/subunit/oocyte) of either a wild-type subunit or a corresponding NH₂-terminal truncated mutant (Fig. 4). Overexpression of either wt- or N-rENaC subunits had a significant effect on I_Na (Fig. 4). A 10 times excess of wt--rENaC significantly reduced I_Na (Fig. 4). By contrast, a 10 times excess of wt--rENaC significantly increased I_Na (Fig. 4). However, a 10 times excess of wt--rENaC had no effect on I_Na. On the other hand, 10-fold overexpression of NH₂-terminal truncated -, -, or -rENaC individually with wt partners completely eliminated I_Na (Fig. 4). These observations are consistent with the conclusion that deletion of the NH₂ terminus of -, -, and -rENaC produced a dominant negative mutant. In addition, the observation that an excess of wt--rENaC decreased I_Na whereas an excess of wt--rENaC increased I_Na suggests that the level of expression of the - and -subunit regulates I_Na.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Overexpression of wt--rENaC decreases and overexpression of wt--rENaC increases INa. By contrast, overexpression of the NH2-terminal truncated mutant of each subunit in oocytes also expressing wt-,,-rENaC eliminated INa. The number of oocytes per group was 5-8. Asterisks indicate that INa is significantly different from the data immediately above (p < 0.01). The number 1 indicates that we injected 2.5 ng/cRNA/oocyte of that subunit and 10 indicates that we injected 25 ng/cRNA/oocyte of that subunit.

The NH₂ Terminus of -rENaC Contains Two Domains That Regulate I_Na-- To begin to identify the key amino acids in the NH₂ terminus of -rENaC that are important for channel function, we made a series of NH₂-terminal, truncated -rENaC cDNAs and co-expressed truncated -rENaCs with wild type - and -rENaC. Deletion of amino acids 2-41 and 2-46 had no effect on I_Na (Fig. 5). However, deletion of amino acids 2-51 increased I_Na by almost 4-fold (Fig. 5). Thus, deletion of amino acids 47 through 50 had a dramatic and positive effect on I_Na. Deletion of amino acids 2-55, 2-58, and 2-67 had no additional effect on I_Na compared with deletion of amino acids 2-51. However, truncation of additional amino acids (2-79, 2-89, 2-100, and 2-109) reduced I_Na to 0. Thus, amino acids 68-109 of -rENaC are required for channel activity. Taken together, these results indicate that there are two domains in the NH₂ terminus of -rENaC that play key roles in channel activity. One domain between amino acids 47 and 50, that, when deleted increases I_Na. The second domain, located downstream of amino acid 67 is absolutely required for rENaC activity as assayed in the Xenopus oocyte expression system.

View larger version (35K): [in this window] [in a new window]   Fig. 5.   Progressive NH2-terminal deletions of -rENaC. Oocytes were co-injected with wt or mutant -rENaC and wt- and wt--rENaC (2.5 ng/cRNA/oocyte). Data are expressed as the ratio of INa mutant/INa wt (i.e. wt-,,-rENaC). The number of oocytes per group was 5-8. Asterisks indicate significantly different from wt-,,-rENaC (p < 0.001).

Deletion of Amino Acids 2-67 in -rENaC Enhances I_Na by Increasing the Number of Channels in the Plasma Membrane: Single Channel Analysis-- Deletion of amino acids 2-67 in -rENaC may enhance I_Na by increasing: the number of channels in the plasma membrane (N), the single channel open probability (P_o), and/or the single channel conductance (). To discriminate among these possibilities we conducted patch-clamp analysis on cell-attached membrane patches in oocytes expressing either wt-,,-rENaC or 2-67-,,-rENaC. Representative current records are depicted in Fig. 6. Deletion of amino acids 2-67 had no effect on  which, when measured using lithium as the permeant cation, was 7.6 ± 0.3 pS for ,,-rENaC (n = 6) and 7.2 ± 0.1 pS (n = 8: p > 0.2) for 2-67-,,-rENaC. The single channel conductance using Na⁺ as the primary cation was also not affected by truncation of the amino terminus of  (5.4 ± 0.2 pS for ,,-rENaC and 5.1 ± 0.1 pS for 2-67-,,-rENaC). Truncation of the NH₂ terminus of -rENaC also had no effect on the cation permeability of the channel. The P_Na/P_K was 82.0 ± 3.7 for ,,-rENaC and 97.1 ± 8.7 for 2-67-,,-rENaC (n = 4: p > 0.1).

View larger version (26K): [in this window] [in a new window]   Fig. 6.   Representative single channel current records of wt-,,-rENaC and 2-67-,,-rENaC in cell-attached patches of oocytes. The pipette potential was 60 mV. There were 4 channels in the membrane patch expressing wt-,,-rENaC and 9 channels in the membrane patch in oocytes expressing 2-67-,,-rENaC. Dashed line indicates the 0 current level (i.e. channels closed).

Deletion of amino acids 2-67 in -rENaC produced a striking and significant increase in the number of active channels in the membrane (N) but had no effect on P_o. N was 6.3 ± 1.2 for ,,-rENaC (n = 8) and 11.3 ± 1.2 for 2-67-,,-rENaC (n = 7: p < 0.01). Furthermore, the P_o was similar in both groups of oocytes (0.32 ± 0.06 (n = 8) for ,,-rENaC and 0.43 ± 0.03 for 2-67-,,-rENaC (n = 7: p > 0.1)). Thus, truncation of amino acids 2-67 in -rENaC enhanced I_Na primarily by increasing the number of channels in the plasma membrane.

Deletion of Amino Acids 2-67 in -rENaC Enhances I_Na by Increasing the Number of Channels in the Plasma Membrane: Laser Scanning Confocal Microscopy-- To provide independent support for the patch-clamp studies, presented above, demonstrating that deletion of amino acids 2-67 in -rENaC increased the number of channels in the plasma membrane we examined plasma membrane expression of wt and mutant rENaC by laser scanning confocal microscopy of GFP-tagged rENaC. We co-expressed GFP--rENaC and GFP--rENaC with either wt--rENaC or 2-67--rENaC. In a previous study we demonstrated that the GFP tag on - and -rENaC does not have a positive or negative effect on I_Na when co-expressed with -rENaC.² Moreover, I_Na was also significantly higher in oocytes expressing 2-67-, GFP-, and GFP--rENaC compared with oocytes expressing wt-, GFP-, and GFP--rENaC (5.23 ± 0.84 µA versus 1.67 ± 0.35 µA: n = 6, p < 0.005). Thus, GFP had no effect on I_Na.

As illustrated in Fig. 7, oocytes expressing 2-67, GFP-, and GFP--rENaC had dramatically more rENaC in the plasma membrane than oocytes expressing wt-, GFP-, and GFP--rENaC. This observation confirms our patch-clamp studies demonstrating that deletion of amino acids 2-67 in -rENaC increased the number of channels in the plasma membrane.

View larger version (25K): [in this window] [in a new window]   Fig. 7.   Representative images of oocytes injected with water (Row 1) or oocytes expressing GFP-tagged rENaC (Rows 2-4). Row 2: wt-, GFP-, GFP--rENaC. Row 3: 2-67, GFP-, GFP--rENaC. Row 4: 2-109, GFP-, GFP--rENaC. Proteins in the plasma membrane were labeled with biotin which was detected with strepavidin conjugated to Texas Red (images in column labeled a). The location of GFP-labeled rENaC is depicted in green in panel b. Panel c is an overlay of panels a and b. Co-localization of rENaC and the plasma membrane are indicted in yellow. There was no visible green fluorescence in water-injected oocytes. There was significantly more wt-, GFP-, GFP--rENaC in the membrane than 2-109, GFP-, and GFP--rENaC. However, there was more 2-67, GFP-, and GFP--rENaC in the membrane than wt-, GFP-, GFP-rENaC.

We have also begun to address the question why does deletion of amino acids 2-109 in -rENaC eliminate I_Na? We considered two possibilities. First, 2-109-,,-rENaC may not traffic to the plasma membrane. Second 2-109-,,-rENaC may traffic to the plasma membrane but does not form functional channels. To discriminate between these possibilities we co-expressed 2-109--rENaC with GFP-tagged - and -rENaC subunits in oocytes and examined the cellular distribution by laser scanning confocal microscopy. 2-109-, GFP--rENaC, and GFP--rENaC was expressed in the plasma membrane; however, the level of expression was less than wt-, GFP-, and GFP--rENaC (Fig. 7). Because we could not detect I_Na in oocytes expressing 2-109-,,-rENaC this observation suggests that 2-10-,,-rENaC does not form functional channels in the membrane, as assayed in the Xenopus expression system, or that expression is too low to detect by the two-electrode voltage clamp technique.

Deletion of Amino Acids 2-67 in -rENaC Increases the Half-life of the Channel in the Plasma Membrane-- To determine if the increased expression of 2-67-,,-rENaC channels in the plasma membrane is due to an increase in the rate of delivery of channels to the membrane and/or to a decrease in the rate of endocytic retrieval of channels from the membrane we inhibited delivery of channels to the plasma membrane with brefeldin A (BFA). BFA is a fungal metabolite that inhibits the anterograde transport of newly synthesized proteins from the endoplasmic reticulum to the Golgi apparatus (18, 30-32). Although BFA has many effects on vesicle transport it does not affect clathrin-mediated endocytosis in Xenopus oocytes. Moreover, BFA has no direct effect on rENaC channel activity and it is fully reversible in oocytes (30). Addition of BFA reduced I_Na in oocytes expressing ,,-rENaC and 2-67-,,-rENaC (Fig. 8). We determined the half-life of the channel in the membrane by measuring the rate of decay of I_Na as a function of time after addition of BFA to the incubation media. This half-life reflects the rate of endocytic retrieval of rENaC from the membrane.³ The half-life of I_Na was 1.1 ± 0.2 h for oocytes expressing ,,-rENaC (n = 7) and 3.5 ± 1.3 h for oocytes expressing 2-67-,,-rENaC (n = 7, p < 001). These data are most consistent with the conclusion that deletion of amino acids 2-67 in -rENaC increased the number of rENaC channels in the membrane by reducing endocytic retrieval of channels from the plasma membrane.

View larger version (15K): [in this window] [in a new window]   Fig. 8.   The effect of brefeldin A (BFA) on INa in oocytes injected with wt-,,-rENaC or 2-67-,,-rENaC. cRNA for wt-,,-rENaC or 2-67-,,-rENaC was injected into oocytes. Two days later, BFA (10 µg/ml) was added to the bath solution at time 0 and INa was measured at the time points indicated. The mean INa for wt-,,-rENaC and 2-67-,,-rENaC were significantly different at each time point (p < 0.01). Standard error bars for some means are smaller than the symbol. The half-life of INa was calculated using PRISM software (Graphpad Software). The data were best fit to a single exponential. In the absence of BFA, INa was constant for 8 h in oocytes expressing both wt-,,-rENaC and 2-67-,,-rENaC.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The major new finding in this report is that the cytoplasmic, NH₂ termini of -, -, and -rENaC are required for rENaC activity. Our data also demonstrate that the NH₂ terminus of -rENaC contains two key domains. One domain, amino acids 47 to 50, regulates the endocytic retrieval of the channel from the plasma membrane. The second domain, located downstream of amino acid 67, is required for channel function. Our data also reveal that overexpression (10-fold) of -rENaC inhibits I_Na in oocytes expressing wild-type -, -, and -channel subunits, whereas overexpression (10-fold) of -rENaC enhances I_Na in oocytes expressing wild-type -, -, and -channel subunits. Lastly, we demonstrate that deletion of the cytoplasmic, NH₂ terminus of , , and  produces a dominant negative mutant.

The Cytoplasmic, NH₂ Terminus of -, -, and -rENaC Is Required for Maximum Channel Activity-- The current view on ENaC channel assembly and function is that all three subunits are required for maximal channel activity and that the three subunits assemble in the endoplasmic reticulum, are post-translationally modified in the Golgi network, and traffic to the plasma membrane. Consistent with this view we demonstrated that maximum I_Na was observed when all three wild-type rENaC subunits were co-expressed in oocytes. However, we also showed that substitution of any wt-rENaC subunit with a N-truncated subunit reduced I_Na dramatically. Thus, we conclude that the cytoplasmic, NH₂ terminus of -, -, and -rENaC is required for maximum channel activity. Our data confirm and extend observations reporting the importance of the cytoplasmic, NH₂ terminus in rENaC function. For example, Grunder et al. (17) demonstrated that point substitutions of a highly conserved glycine residue with a serine in the NH₂ termini of -rENaC (G95S), -rENaC (G37S), or -rENaC (G40S) dramatically reduced I_Na. Adams et al. (16) reported that N-hENaC interacted with -hENaC but failed to stimulate I_Na when coexpressed with -hENaC. In fact, co-expression of N-hENaC with -hENaC decreased I_Na slightly, and reduced protein levels of -hENaC (16) suggesting that interaction of N-hENaC with -hENaC facilitated its degradation. It is interesting to note that Adams et al. (16) also reported that at least two domains of -hENaC interact with -hENaC: a domain located within amino acids 3-53 and another, unidentified domain. Accordingly, the ability of -hENaC to contribute maximally to rENaC function may require that -hENaC interacts with at least two domains in  and perhaps with domains in -hENaC. Abrogation of either interaction may be sufficient to reduce channel activity. Finally, in a previous study we demonstrated that deletion of amino acids 2-109 in -rENaC had no effect on  or P_o when the truncated subunit was studied in lipid bilayers (33). However, deletion of the NH₂ terminus changed the kinetic properties of the channel (33). Given these observations it was somewhat surprising that we could not measure an amiloride-sensitive current in oocytes expressing 2-109-,,-rENaC (Fig. 2) even though the channel was expressed in the plasma membrane, although the expression was less than wt-, GFP-, and GFP--rENaC (Fig. 7). The most parsimonious explanation for these observations is that 2-109 assembles with - and -rENaC and traffics to the plasma membrane but the current produced by the mutant channel is too low to detect by the two-electrode voltage clamp technique.

The failure of N-rENaC subunits to support maximum channel activity could be related to their inability to interact with wild-type subunits and/or to traffic to the plasma membrane. Alternatively, or in addition, it is possible that N-rENaC subunits interact with wild-type subunits and traffic to the membrane: however, the NH₂ terminus of each subunit may be essential for channel activity. For example, the NH₂ terminus is important in the assembly of other multimeric ion channels, including the acetylcholine receptor and voltage-gated K⁺ channels (27, 28, 34, 35). Moreover, the NH₂ terminus of some channels, including -rENaC, is involved in channel gating (33, 36, 37). Additional studies, beyond the scope of the present report, are required to elucidate the mechanism whereby deletion of the NH₂ termini reduces I_Na.

Overexpression of wt- and -rENaC Affects I_Na-- Overexpression of -rENaC in oocytes expressing -, -, and -rENaC inhibited I_Na whereas overexpression of -rENaC in oocytes expressing -, -, and -rENaC enhanced I_Na. Assuming that the amount of cRNA injected into oocytes correlates with the amount of protein expressed, these observations are consistent with the view that the relative levels of expression of ,,-rENaC regulates I_Na. Thus, our data suggest that overexpression of -rENaC disrupts the optimal stochiometry of ,,-rENaC subunits whereas overexpression of -rENaC enhances the optimal stochiometry of ,,-rENaC subunits. The view that varying the ratio of -, -, and -rENaC subunit expression affects I_Na is supported by two recent publications (22, 38). For example, inspection of individual data points in Fig. 1 in Firsov et al. (38) suggests that overexpression of -rENaC reduces I_Na in oocytes expressing ,,-rENaC. In addition, in oocytes expressing - and -rENaC or - and -rENaC, variations in the ratio of cRNA injection from 1:1 significantly decreased I_Na (22).

Deletion of the NH₂ Terminus of -, -, and -rENaC Produces a Dominant Negative Mutant-- When co-expressed with wild-type -, -, and -rENaC subunits, NH₂-terminal mutants of each rENaC subunit dramatically reduced I_Na. Thus, NH₂-truncated-rENaC subunits appear to be dominant negative mutants that displace wt subunits from the multimeric ion channel complex. Overexpression of the cytoplasmic NH₂ termini of degenerins, which are closely related to rENaC, also inhibits degenerin channel activity. Thus, the NH₂ terminus of the superfamily of ion channels that includes the FMRFamide-gated Na⁺ channel, Na⁺ channels in brain (BNC1 and BNC2), and the degenerins of C. elegans that encode mechanosensitive channels (e.g. DEG-1, MEC-4, and MEC-10) plays an important role in channel function (4, 7).

A Domain between Amino Acids 47 and 50 (KGDK) in -rENaC Regulates rENaC Endocytosis-- Our data suggest that amino acids 47-50 (KGDK) in -rENaC may be an endocytic motif that regulates the number of channels in the plasma membrane. We report that deletion of amino acids 2-41 or 2-46 in -rENaC had no effect on I_Na; however, deletion of four additional amino acids (i.e. KGDK 47-50) increased I_Na by ~4-fold (Fig. 5). Additional deletions, including 2-55, 2-59, and 2-67 failed to increase I_Na further. Thus, deletion of amino acids 47-50 (KGDK), a region which is conserved across species (11), increased I_Na. Several lines of evidence in this report suggest that deletion of amino acids 47-50 enhances I_Na by increasing the number of channels in the plasma membrane and that the increase in the number of channels results from a decline in the endocytic retrieval of rENaC from the plasma membrane. Single channel patch-clamp studies demonstrated that the increase in I_Na with 2-67-,,-rENaC compared with ,,-rENaC-rENaC was referable to an increase in the number of channels in the membrane and not due to an increase in  or P_o. We confirmed the electrophysiological data by laser scanning confocal microscopy of GFP-tagged - and -rENaC subunits. We observed a dramatic increase in the amount of 2-67, GFP-, and GFP--rENaC in the plasma membrane compared with wt-, GFP-, and GFP--rENaC. Finally, inhibition of channel insertion into the membrane with BFA revealed that deletion of amino acids 2-67 in -rENaC decreased the rate of decline of I_Na compared with wt-rENaC, an observation consistent with the view that the amino terminus of -rENaC, in particular KGDK, is involved in the endocytic retrieval of channels from the plasma membrane. Truncation of a similar motif (i.e. RGER) in furin, a mammalian endopeptidase, impairs its endocytosis (39). RGER is isoelectrically identical to KGDK, consistent with the view that these amino acids may be an endocytic motif which is important for rENaC internalization from the plasma membrane.

One splice variant of human -ENaC subunit (-hENaC2), extends the length of the NH₂ terminus by 59 amino acids, which includes a second copy of the KGDK motif (29, 40). Thus, it can be predicted that -hENaC2 may have a different half-life than -hENaC. The KGDK motif is present in rat, human (KDNK), bovine, and mouse -ENaC, but is absence in chicken and frog -ENaC (11, 40).