J Biol Chem, Vol. 274, Issue 43, 30345-30348, October 22, 1999

COMMUNICATION
Regulation of the Amiloride-sensitive Epithelial Sodium Channel by Syntaxin 1A^*

Juanjuan Qi, Kathryn W. Peters, Chongguang Liu, Jun-Min Wang, Robert S. Edinger, John P. Johnson, Simon C. Watkins, and Raymond A. Frizzell^§

From the Departments of Cell Biology and Physiology and Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania 15261

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The first step in transepithelial sodium absorption lies at the apical membrane where the amiloride-sensitive, epithelial sodium channel, ENaC, facilitates sodium entry into the cell. Here we report that the vesicle traffic regulatory (SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor)) protein, syntaxin 1A (S1A), inhibits ENaC mediated sodium entry. This inhibitory effect is selective for S1A and is not reproduced by syntaxin 3. The inhibition does not require the membrane anchoring domain of syntaxin 1A. It was reversed by the S1A-binding protein, Munc-18, but not by a Munc-18 mutant, which lacks syntaxin affinity. Immunostaining of epitope-tagged ENaC subunits showed that syntaxin 1A decreases ENaC current by reducing the number of ENaC channels in the plasma membrane; S1A does not interfere with ENaC protein expression. Immunoprecipitation of syntaxin 1A from the sodium-transporting epithelial cell line, A6, co-precipitates ENaC. These findings indicate that syntaxin 1A and other members of the SNARE machinery are involved in the control of plasma membrane ENaC content, and they suggest that SNARE proteins participate in the regulation of sodium absorption in relation to agonist mediated vesicle insertion-retrieval processes.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Sodium entry across the apical membranes of most high resistance epithelia is facilitated and regulated by the epithelial sodium channel, ENaC¹ (reviewed in Refs. 1 and 2). These channels can be identified operationally by their high affinity for the diuretic, amiloride, and by their high selectivity for sodium over potassium. Cloned ENaC channels, when expressed in Xenopus oocytes, exhibit properties expected for the highly selective amiloride-sensitive channel. This expression system has been employed extensively for functional studies of ENaC genetic mutations (3), subunit stoichiometry (4), and plasma membrane ENaC turnover (5).

Acute hormonal regulation of sodium entry is provided by vasopressin, prostaglandins, and insulin (2). Vasopressin and prostaglandin E increase sodium entry by promoting the delivery of new ENaC channels to the apical surface (6, 7). The biochemical events that govern many membrane trafficking processes conform to a similar paradigm and involve interactions among proteins that comprise the SNARE fusion complex (8, 9). These interactions govern the insertion and retrieval of membrane vesicles that contain secretory products or integral membrane proteins, like ENaC, that will become constituents of the plasma membrane.

Evidence of ENaC regulation by the SNARE machinery is provided by our finding that exogenous expression of syntaxin 1A selectively inhibits ENaC currents. Syntaxin expression decreases the plasma membrane content of ENaC, consistent with its role as a traffic regulatory protein. Biochemical studies suggest a physical interaction between ENaC and syntaxin 1A, which may reflect a direct role for the SNARE machinery in regulated Na entry.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Oocyte Expression-- Oocyte isolations and RNA injections were performed as described previously (10). Stage 5-6 oocytes were maintained in a modified Barth's solution overnight before injection (50 nl) with cRNA for ENaC , , and  subunits with or without syntaxin 1A, 2, or 3. Noninjected oocytes served as controls. Expression proceeded at 18 °C for 1-3 days in a sodium-free ND-96 solution before current recordings or immunofluorescence measurements. Human ENaC cDNAs were kindly provided by Dr. Michael Welsh (University of Iowa), syntaxin constructs by the laboratory of Dr. Richard Scheller (Stanford University), and Munc-18 constructs by Dr. Jonathan Pevsner (Johns Hopkins University).

Electrophysiological Measurements-- The ND-96 solution utilized for current measurements contained (mM): 96 NaCl, 1 KCl, 1.8 CaCl₂, 1.0 MgCl₂, and 5 HEPES, pH 7.2. In low sodium ND-96, N-methyl-D-glucamine Cl replaced NaCl. Recordings of ENaC-mediated sodium current (I_Na) were performed by double electrode voltage clamp as described (10). Steady-state currents recorded at 100 mV are given in the figures; they reflect inward flow of Na through ENaC, as reflected by their amiloride sensitivity and augmentation when bath sodium was replaced by lithium.

Cell Surface ENaC Expression-- We used immunofluorescence techniques and confocal microscopy to monitor the expression of ENaC in the oocyte plasma membrane. Cells were injected with cRNA encoding ENaC subunits in which the FLAG epitope (DYKDDDDK) was introduced into the human ENaC , , and  subunits at positions used previously to monitor cell surface expression of rat ENaC (11). In our studies, the amiloride-sensitive currents of flag-hENaC were indistinguishable from wild-type: wt-ENaC = 1.5 ± 0.2 µA; FLAG-ENaC = 1.8 ± 0.4 µA; n = 6.

FLAG-ENaC-expressing oocytes were subjected to a staining protocol to quantitate ENaC protein expression at the cell surface using the monoclonal M2 antibody (Eastman Kodak Co.). Oocytes were cooled rapidly to 4 °C and incubated with M2 antibody overnight (1:1250 dilution). After washing with 5% fetal calf serum/ND-96, oocytes were incubated at 4 °C in fluorescein conjugated goat-anti-mouse IgG (1:100 dilution) and then washed five times as above. This protocol avoids permeabilization or fixation conditions that might expose intracellular FLAG-ENaC to the M2 antibody. Thus, the staining conditions and epitope positions are selected to permit detection of FLAG-ENaC only when it is localized in the plasma membrane.

In other surface expression studies, oocytes expressing FLAG-ENaC were fixed or fixed and permeabilized before M2 antibody labeling. Oocytes were cooled rapidly to 0-4° C and blocked for 15 min in 5% BSA-C (Aurion) in low sodium ND-96 (BSA-C/low sodium). They were fixed for 30 min at room temperature with Medium A (Caltag) and, if permeabilized, were bathed in Medium B for 15 min, then processed as described for nonfixed oocytes.

The vegetal poles of individual, labeled oocytes were scanned with a Molecular Dynamics Multi-probe 2001 laser confocal microscope at 10× magnification using 10-µm optical sections. Control, noninjected oocytes were scanned to set the fluorescence background and to obtain laser intensity and voltage settings within the linear camera range.

ENaC and Syntaxin Immunoprecipitation-- For metabolic labeling, oocytes were incubated overnight in 1.2 µCi/µl Tran³⁵S-label^TM (ICN, 10 µl/oocyte), washed in sodium-free ND-96, and solubilized (25 mM MES, pH 6.4, 200 mM NaCl, 1% Triton X-100, 60 mM n-octyl glucoside, 0.1% SDS, 0.5% Nonidet P-40, 0.02% deoxycholic acid (sodium salt), 1% digitonin, 0.5% Tween 20, 0.02% CHAPS, and 2 mM Empigen BB). Lysates were homogenized on ice, forced through an acrodisc filter, spun at 4 °C (13,800 × g, 10 min), and 200 µl of M2 affinity gel (Kodak) added to supernatants. After overnight rotation at 4 °C, samples were spun as above for 1 min. Bead complexes were washed and then 70 µl of Laemmli buffer was added. Samples were boiled 3 min, then resolved on a 7.5% polyacrylamide gel, which was subjected to fluorography and exposed to film. Xenopus Syntaxin 1A (Sigma HPC-1 monoclonal) or -ENaC were immunoprecipitated from A6 epithelia using procedures described previously (12). When using chicken antibodies, immobilized anti-chicken IgY (Promega, number G1191) was used in place of GammaBind^R Sepharose beads. Western blots were performed as described previously (12). Reactive proteins were detected using enhanced chemiluminescence (Pierce, ULTRA-ECL) followed by autoradiography. All results are expressed as the mean ± S.E.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Syntaxin 1A Selectively Decreases ENaC Currents-- Fig. 1 provides the results of ENaC current measurements in oocytes co-expressing the human , , and  ENaC subunits together with syntaxin 1A or 3. ENaC current is markedly attenuated in oocytes co-expressing ENaC and syntaxin 1A, but not in oocytes that co-express ENaC and syntaxin 3. The absence of syntaxin 3 inhibition indicates specificity among syntaxin isoforms and that the inhibitory effect of S1A is not due to translational competition. Co-expression of a truncated syntaxin 1A (S1AC) inhibited ENaC currents to a level similar to that observed with the full-length syntaxin. This mutant lacks the C-terminal transmembrane domain and is expected to result in the expression of a soluble syntaxin 1A. The inhibition of ENaC currents by syntaxin 1A was reversed by co-expression of the syntaxin binding protein Munc-18 (13). This gain-of-function effect also cannot be attributed to translational competition among the cRNAs, since ENaC currents increased. Co-expression of a mutant Munc-18 which lacks high affinity for syntaxin was ineffective in reversing the S1A inhibition.

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Effect of syntaxin co-expression on ENaC currents. Current inhibition by syntaxin 1A (Syn1A), with or without (S1AC, amino acids 4-267 of S1A) its membrane anchoring C terminus, but not by syntaxin 3 (Syn3). Inhibition by S1A is reversed by Munc-18, but not by a Munc-18 mutant. cRNA amounts: ENaC, 5 ng total; syntaxins, 5 ng; Munc-18s, 5 ng per oocyte. Amiloride-sensitive, steady-state currents recorded at 100 mV are given; data from 5-11 oocytes per group.

Syntaxin 1A Does Not Alter ENaC Protein Expression-- After metabolic labeling, cell lysates from oocytes expressing flag-ENaC were precipitated using M2 antibodies and the precipitate subjected to SDS-polyacrylamide gel electrophoresis. We observed no difference in the amount of ENaC in cells co-expressing syntaxin 1A relative to control (Fig. 2). In parallel current measurements, ENaC inhibition with syntaxin 1A co-expression was similar to that shown in Fig. 1.

View larger version (85K): [in this window] [in a new window]   Fig. 2.   Immunoprecipitation of FLAG-ENaC from oocytes with and without S1A co-expression. Oocytes were injected with 1.25 ng of each FLAG-ENaC subunit. Results are typical of those from three experiments. The inset shows increased resolution of individual subunits separated on a 10% polyacrylamide gel.

Syntaxin 1A Decreases Plasma Membrane ENaC-- The inhibitory effect of S1A could result from an effect on ENaC channel gating or from a decrease in the number of plasma membrane resident sodium channels. To examine this issue, we expressed epitope-tagged , , and  ENaC subunits and monitored their cell surface expression using immunofluorescence and confocal microscopy. For the initial studies, FLAG-ENaC expressing oocytes were labeled with primary and secondary antibodies at 4 °C without cell permeabilization or fixation. Composite confocal images of intact, nonpermeabilized oocytes expressing FLAG-ENaC are shown (Fig. 3, B and C). The background fluorescence intensity of noninjected oocytes was low (Fig. 3A), whereas surface fluorescence of FLAG-ENaC oocytes was readily detected (Fig. 3B). Co-expression with syntaxin 1A markedly reduced ENaC expression in the plasma membrane (Fig. 3C). As observed for ENaC currents (Fig. 1), syntaxin 3 co-expression had no effect on cell surface ENaC staining (data not shown). These findings suggest that syntaxin 1A reduces ENaC currents by decreasing the number of sodium channels in the plasma membrane.

View larger version (44K): [in this window] [in a new window]   Fig. 3.   Effect of syntaxin on cell surface ENaC. Composite confocal images of oocytes stained at 4 °C without fixation or permeabilization. A, noninjected control; B, FLAG-ENaC alone; C, FLAG-ENaC plus syntaxin 1A. cRNA amounts: 5 ng of total ENaC, 5 ng of S1A. Data are representative of results from three experiments. Sections were scanned at 512 × 512 pixels (488 laser, 510 nm primary beam splitter, 510 secondary beam splitter). Twenty image planes through the specimen were collected. A quantitative measure of protein expression was derived from a maximal intensity rendered image. The periphery of the image was delineated and mean pixel intensity calculated.

Similar experiments were carried out in oocytes subjected to fixation and/or permeabilization conditions prior to antibody labeling, the latter to detect intracellular ENaC. Figs. 4, A and B, illustrate the effect of syntaxin 1A on ENaC cell surface expression in fixed oocytes. The results are similar to those obtained when ENaC was labeled at 4 °C without fixation (Fig. 3C), indicating that the fixation conditions do not affect antibody labeling. Fig. 4, C and D, provide composite images of permeabilized cells expressing ENaC or ENaC plus S1A. The detection of ENaC after permeabilization in oocytes co-expressing S1A (Fig. 4D) indicates that syntaxin does not compromise ENaC protein expression. The average fluorescence intensity and sodium current data from all experiments of this type are provided in Fig. 4E. The data indicate that syntaxin 1A inhibits ENaC current by reducing plasma membrane ENaC content. The effect of S1A is on ENaC location, not on the level of protein expression.

View larger version (50K): [in this window] [in a new window]   Fig. 4.   Effect of syntaxin on cell surface ENaC. Composite confocal images of oocytes expressing flag-ENaC with and without S1A as indicated, stained after fixation (A and B) or fixation and permeabilization (C and D). E, left ordinate, mean fluorescence intensity values for the conditions indicated, corrected for background intensity of noninjected oocytes; right ordinate, corresponding amiloride-sensitive currents. cRNA amounts: FLAG-ENaC, 7.5 ng total, S1A, 5 ng. Data are from at least three oocytes in each of three experiments.

To determine whether the effect of S1A on cell surface ENaC is associated with an inhibition of ENaC insertion into the plasma membrane or to its enhanced endocytic retrieval, we inhibited channel delivery to the plasma membrane using brefeldin A (BFA). The time-constant () describing the single exponential decay of amiloride-sensitive sodium current following BFA addition was used as a measure of ENaC retrieval from the plasma membrane (5). In oocytes expressing ENaC alone or ENaC plus S1A,  was 44 or 45 min¹, respectively. Pre-BFA current was reduced 60% by S1A co-expression in these experiments. The failure of syntaxin to enhance ENaC retrieval suggests that S1A reduces ENaC current and channel density by interfering with its insertion into the plasma membrane.

ENaC-S1A Interactions in A6 Epithelia-- To determine whether ENaC-syntaxin interactions are present also in epithelial cells, we asked whether a syntaxin 1A homolog is expressed in A6 cells, a cell line with distal nephron properties. As shown in Fig. 5A, antibodies against rat syntaxin 1A immunoprecipitated a 35-kDa protein from A6 cell lysates, which were prepared from cells grown as transport-competent monolayers on permeable supports. To assess putative interactions of syntaxin with ENaC in A6 epithelia, the syntaxin IP was blotted with antibodies to the , , and  Xenopus ENaC subunits (12). Each antibody recognizes its respective full-length subunit on immunoblots subsequent to ENaC in vitro translation, and this interaction is abolished by excess immunizing peptide (12). In addition, there is no subunit cross reactivity among these antibodies. As shown in Fig. 5B, the -xENaC antibody identifies a 97-kDa band in the syntaxin immunoprecipitate from A6 cells, which corresponds to the glycosylated form of -xENaC (12). The  and  xENaC antibodies did not detect these subunits in the syntaxin IP. Fig. 5C shows that the -xENaC antibody produces a similar result. In A6 cell lysates, anti- antibody precipitates ; under these experimental conditions,  and  are not detected significantly. The results of Fig. 5 are consistent with a physical association between syntaxin and the  subunit.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Syntaxin 1A-ENaC interaction in A6 epithelia. Immunoprecipitation or immunoblot of S1A and ENaC subunits from A6 monolayers grown on permeable supports. A, syntaxin IP, Coomassie-stained polyvinylidene difluoride. The dense bands at 50 and 25 kDa represent the syntaxin monoclonal heavy and light chains. The band(s) above 66 kDa may represent Munc-18 or a t-SNARE complex containing S1A. B, syntaxin IP and blot by ENaC , , or  subunit antibodies, as shown; C, anti- antibody IP followed by , , or  subunit blot. Similar results were obtained in four experiments.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The results of these studies provide evidence of functional and physical interactions between syntaxin 1A and epithelial sodium channels. They suggest that a functional interaction between syntaxin and ENaC, perhaps modulated by Munc-18, is involved in the control of sodium entry rate at the apical membranes of sodium-absorbing epithelial cells. Syntaxin expression selectively decreased ENaC currents, and cell surface labeling studies indicate that this inhibition of current reflects a decrease in the number of sodium channels present in the plasma membrane. A syntaxin 1A homolog is detected in A6 epithelia,² a sodium-transporting cell line that is used widely for studies of regulated sodium absorption, and a physical interaction between S1A and ENaC was evident from their co-precipitation. In prior experiments (14), we examined the influence of syntaxin 1A on the functional activity of several other membrane proteins to determine whether S1A has a generalized effect on plasma membrane protein expression in this system. The functional expression of several other transport or receptor proteins was not affected by S1A, indicating that syntaxin is not simply disrupting the expression of integral membrane proteins.

Physical and functional interactions of syntaxin 1A with ion channels have been observed in several other systems. Binding of syntaxin 1A to N-type calcium channels is thought to play an important role in the docking of presynaptic vesicles containing neurotransmitter at sites of calcium entry (15). Calcium current measurements suggest that syntaxin co-expression inhibits calcium entry (16, 17). Similarly, co-expression of syntaxin with CFTR inhibits cAMP-dependent chlorine currents in Xenopus oocytes and an interaction between these proteins is detected using in vitro protein binding assays (18, 19). In principle, syntaxin could alter ion channel currents by affecting channel open probability (gating) or channel number, but the above reports do not provide insight into this issue. The inhibition of plasma membrane ENaC content by syntaxin 1A (Figs. 3 and 4) implicates the SNARE machinery, and in particular S1A, in the control of plasma membrane sodium channel density. Results from the brefeldin A experiments are consistent with a primary effect of S1A on the rate of ENaC delivery to the plasma membrane.

This effect of syntaxin expression has been observed previously for membrane trafficking processes. For example, exogenous expression of the Golgi t-SNARE syntaxin 5 inhibits protein traffic from ER to Golgi, a step in which this syntaxin isoform functions (20). Expression of syntaxin 1A, but not 1B, blocks glucose stimulated insulin secretion in pancreatic  cells (21). Likewise, expression of syntaxin 4 blocks insulin-stimulated GLUT-4 trafficking in adipocytes (22), a process mediated by syntaxin 4 and other SNARE proteins (23). In MDCK cells, syntaxin 3 is apically localized, and its expression selectively inhibits apical targeting and recycling of the polymeric immunoglobulin receptor (27). Together, these findings indicate that exogenous expression of a specific syntaxin isoform disrupts the pathway in which that isoform normally plays a role in membrane trafficking events. This inhibition occurs presumably because overexpression of a single component of the fusion complex disrupts the stoichiometric interactions among SNARE proteins that are required for normal membrane trafficking (20-24).

As for other syntaxin-sensitive ion channels, our findings raise questions about the molecular mechanism of these effects, and in particular, their relation to the physical interactions with syntaxin that are detected in protein binding assays. Immunoprecipitation of S1A from A6 cells co-precipitated -ENaC, but the IP did not contain detectable  or  subunit. This may result from dissociation of the ENaC subunits under the immunoprecipitation conditions employed, since precipitation performed using -ENaC antibody produced an identical result (Fig. 5C). The forces that govern subunit associations have not been defined, and they may be of relatively low affinity. A selective association of S1A with -ENaC could lead to its sequestration and degradation as a means of reducing ENaC currents, but we did not detect a reduction in protein levels during syntaxin 1A co-expression (Figs. 2 and 4). In addition, our data cannot distinguish a direct interaction of S1A and -ENaC from the possibility that these proteins are part of a macromolecular complex where their interaction is conferred by other proteins.

The actions of several sodium transport agonists are thought to involve the delivery of additional ENaC channels to the apical cell surface (above discussion). Shimkets et al. (25) identified similar sites in the C termini of the  and  subunits that are phosphorylated in response to aldosterone, insulin, and protein kinases A and C, suggesting that these regulatory pathways may converge at a common control point. Thus, interactions of ENaC with SNARE proteins may be influenced by the state of ENaC regulation (e.g. phosphorylation), which would permit SNARE protein interactions with ENaC to govern the apical sodium channel density in response to agonists. Consistent with this view, phosphorylation of the syntaxin binding domain of the N-type Ca channel was found to markedly alter its affinity for syntaxin 1A (26). Further understanding of ENaC regulation by syntaxin will require identification and manipulation of the other SNARE constituents that lie at the apical membrane domain of sodium-transporting epithelia.

	ACKNOWLEDGEMENTS

We thank Megan Weiss and Lisa Tkach for technical assistance and Patricia Connelly for typing the manuscript. We thank the laboratories of Drs. Jonathan Pevsner, Richard Scheller, and Michael Welsh for cDNAs.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grants DK54814 (to R. A. F.) and DK47874 (to J. P. J.) and by the Cystic Fibrosis Foundation.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

These authors contributed equally to this work.

^§ To whom correspondence should be addressed: Dept. of Cell Biology and Physiology, 3500 Terrace St., S362 BST, Pittsburgh, PA 15261. Tel.: 412-648-9498; Fax: 412-648-2004; E-mail: frizzell+@pitt.edu.

² The HPC-1 antibody used in these studies immunoprecipitates rat syntaxin 1A, but not rat syntaxin 3, from Xenopus oocytes expressing these proteins; it does not immunoblot rat syntaxins 2, 3, or 4 in lysates from MDCK cells expressing these proteins (data not shown; MDCK cells provided by S. H. Low and T. Weimbs, Cleveland Clinic).

	ABBREVIATIONS

The abbreviations used are: ENaC, epithelial sodium channel; M2, monoclonal antibody detecting flag epitope sequence (DYKDDDDK); SNARE, soluble N-ethylmaleimide-sensitive factor attachment protein receptor; BSA-C, acetylated bovine serum albumin; S1A, syntaxin 1A; S1AC, truncated S1A lacking a membrane anchor; MES, 4-morpholineethanesulfonic acid; CHAPS, 3-[(3-cholamindopropyl)dimethylammonio]-1-propanesulfonate; IP, immunoprecipitation; BFA, brefeldin A; MDCK, Madin-Darby canine kidney.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES