Regulation of a sodium channel-associated G-protein by aldosterone.

The action of aldosterone to increase apical membrane permeability in responsive epithelia is thought to be due to activation of sodium channels. This channel is regulated, in part, by G-proteins, but it is not known if this mechanism is regulated by aldosterone. We report that aldosterone stimulates the expression of the 41-kDa alphai3 subunit of the heterotrimeric GTP-binding proteins in A-6 cells. Both mRNA and the total amount of this protein are increased by aldosterone. The G-protein is palmitoylated in response to the steroid, and the newly synthesized subunit is found to co-localize with the sodium channel. Aldosterone stimulation of sodium transport is significantly inhibited by inhibition of palmitoylation. These results suggest that aldosterone regulates sodium channel activity in epithelia through stimulation of the expression and post-translational targeting of a channel regulatory G-protein subunit.

port. A6 cells are maintained in amphibian media with 10% fetal bovine serum in an atmosphere of humidified air-4% CO 2 at room temperature as described previously (13). All studies were carried out in cells grown on Millipore filters (Millipore, Bedford, MA). For electrophysiologic studies, A6 cells were grown on Millicell inserts and transepithelial potential difference and short circuit current were measured with a sterile, in-hood short-circuiting device as described previously (13). For biochemical studies, cells were grown on large Millipore filters (HAWP, 0.45 m) attached to rings made from acrylic tubing as described previously (14). For studies requiring cell disruption, A6 cells were scraped from filters and homogenized with 40 strokes of a Dounce homogenizer in an isotonic calcium-free Ringers solution. This technique results in disruption of Ͼ90% of cells. Crude cell homogenates were centrifuged at 1000 ϫ g for 10 min to remove intact cells or fragments of filter, and the resulting supernatant represented whole cell lysate.
Protein Separation Methods-Proteins were precipitated overnight with 10 volumes acetone at Ϫ20°C and collected by centrifugation. Precipitated proteins were dissolved in 200 l of sample buffer (62.5 mM Tris-HCl, 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, 0.025% bromphenol blue, pH 6.8) and heated at 100°C for 5 min. Aliquots were removed for protein determination (DC protein assay, Bio-Rad) and for counting in a liquid scintillation spectrometer. Protein-matched samples were subjected to SDS-PAGE 1 using 5% stacking gels and 15% separating gels made using the buffer system of Laemmli (15). Molecular weight standards (Bio-Rad) were run in adjacent lanes. Gels were run at constant current (35 mA), stained with Coomassie Blue dye, or silver-stained (Silver Stain Plus Kit, Bio-Rad, Hercules, CA) impregnated with fluorographic enhancer (Enlightening, DuPont NEN) and dried under reduced pressure. Gels were exposed to X-Omat AR film (Eastman Kodak) for 14 -30 days at Ϫ80°C. Fluorograms were scanned with an EC 910 transmission densitometer for quantification.
Immunoprecipitation-Experiments were carried out on whole cell lysates. Acetone-precipitated pellets were solubilized in 1% Triton X-100, 50 mM Tris, pH 8. Four l of rabbit polyclonal antibodies to either G-protein subunits (see below) or sodium channel antibodies were added to 200 g of protein (DC protein assay, Bio-Rad, Hercules, CA) and incubated overnight at 4°C. Then, 200 l of washed (0.1 M phosphate buffer, pH 7) GammaBind Plus-coated Sepharose beads (Pharmacia) were added and incubated at room temperature on a rotary stirrer for an additional 2 h. The samples were then centrifuged briefly to sediment the beads, and the supernatant was removed and counted in a liquid scintillation spectrometer prior to discarding. The beads were washed in 500 l of wash buffer (0.3% Nonidet P-40, 0.3% SDS, 0.9% NaCl, 50 mM Tris, pH 7.5). After a 5-min incubation with gentle agitation, at room temperature, the samples were centrifuged to sediment the beads, and the supernatant was removed and counted as before. Four additional washes were performed in an identical manner, and then the beads were washed an additional 2 times in 500 l of deionized water. To the sedimented beads, 100 l of sample buffer (2.5% SDS, 6 M urea, 5% 2-mercaptoethanol, 50 mM Tris, pH 6.8) was added and agitated at room temperature for 30 min. The beads were sedimented, and the supernatant was saved. This procedure was repeated once, and the two supernatants were pooled. 40 l of glycerol and 10 l * These studies were supported in part by National Institutes of Health Grants DK-37206 (to D. B.) and DK-47874 (to J. P. J.) and National Kidney Foundation Young Investigator Award (to M. D. R.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ʈ To whom correspondence should be addressed: Renal-Electrolyte Division, F-1159 Presbyterian University Hospital, Pittsburgh, PA 15213. Tel.: 412-648-9075; Fax: 412-647-6222. of bromphenol blue were added to the samples, and they were heated at 100°C for 5 min and then subjected to SDS-PAGE as described above.
Western Blotting-Equal amounts of solubilized protein were separated by SDS-PAGE as described above and transferred to nitrocellulose membranes in transfer buffer (39 mM glycine, 48 mM Tris base, 0.037% SDS, 20% methanol). Membranes were stained with Ponceau S (0.2% Ponceau S, 3% trichloroacetic acid, 3% sulfosalicylic acid). Nonspecific binding sites in the transfers were blocked by a 1-h exposure to blocking solution (5% nonfat dried milk in Tris, pH 7.4). Transfers were overlaid with 1:1000 dilutions of specific antibody in blocking solution and incubated overnight at 4°C. The membrane was washed 3 times in phosphate-buffered saline with 0.025% Tween 20, pH 7.4. A horseradish peroxidase-conjugated secondary antibody, diluted 1:10,000 in 5% nonfat dry milk, 5% fetal bovine serum, in phosphate-buffered saline, pH 7.4, was allowed to incubate for 1 h at room temperature. The membrane was again washed, and signals were detected by an enhanced chemiluminescence system (Amersham) and exposed to Kodak X-Omat AR film (Eastman Kodak Co.).
Northern Blotting-A6 cells were serum-depleted overnight and then exposed to 10 Ϫ7 M aldosterone for 12-16 h. Poly(A ϩ ) mRNA was isolated, and 5 g of mRNA was electrophoresed through 1% formaldehyde-agarose gels. Gels were rinsed several times in RNA-free water prior to mRNA transfer to nitrocellulose. After transfer, the nitrocellulose filters were prehybridized at 42°C for 2 h. The prehybridization solution used is given in Sambrook et al. (19). A specific 5Ј-32 P-labeled ␣ i3 oligonucleotide probe was designed from conserved sequences for G␣ i3 showing least homology with G␣ i1 and G␣ i2 (CCTGGCAGCTTC-CCCAAA). This probe does not recognize ␣ i1 or ␣ i2 mRNA (data not shown). Approximately 500,000 cpm/ml 32 P-labeled probe was added to the medium along with 10% dextran sulfate, and the filter was incubated at 42°C for 16 h. The filter was washed for 20 min at room temperature in 1 ϫ SSC, 0.1% SDS, followed by three washes of 20 min each at 68°C in 0.2 ϫ SSC, 0.1% SDS. Filters were exposed at Ϫ70°C for 48 h to Kodak X-Omat film. The autoradiograph for ␥-actin mRNA was obtained by stripping the ␣ i3 probe and rehybridizing with the second probe to detect ␥-actin, a housekeeping gene, for quantification. To control for variability, the G␣ i3 mRNA density values were normalized to the value for the ␥-actin hybridization within each sample. Three different autoradiograph exposures were analyzed to ensure linearity. Three 150-mm filters of confluent A6 cells were used for mRNA isolation under each condition for each experiment.
Materials-G-protein antibodies raised to amino acid sequence G␣ z 40 -45 (GTSNSGKSTIVKQMK) part of the GTP binding domain (clone GA/1) (16) or to the carboxyl-terminal decapeptide (KNNLKECGLY) of ␣ i3 (EC/2) were purchased from DuPont NEN (7,33). G-protein standards were purchased from Calbiochem. Polyclonal antibody to the sodium channel complex was prepared as described previously (17). All other chemicals were purchased from Sigma unless specifically indicated.

RESULTS
To determine if G-protein content was increased by aldosterone, A6 cells were labeled with [ 35 S]methionine in the presence and absence of 0.1 M aldosterone. Whole cell lysates were first incubated with an antibody directed against the common GTP binding site of G-proteins (GA/1), and the immunoprecipitated proteins were subjected to SDS-PAGE and autoradiography. Densitometry revealed that aldosterone enhanced metabolic labeling of a 41-45-kDa GTP-binding protein 3-fold compared to control (Fig. 1a). Other GTP-binding proteins were also labeled. To identify the protein labeled at 41 kDa, cell lysates were then subjected to immunoprecipitation with an affinitypurified antibody (EC/2) specific for G␣ i3 (DuPont NEN) (7,16), directed against the carboxyl-terminal decapeptide of ␣ i3 , analyzed by SDS-PAGE and fluorography (Fig. 1b). Aldosterone enhanced metabolic labeling of G␣ i3 by 3-4-fold compared to controls. Next, whole cell lysates were subjected to Western blot analysis with EC/2 and visualized using peroxidase-conjugated second antibodies and chemiluminescence conjugates. Aldosterone increased the amount of the 41-kDa ␣ i3 G-protein 2-3-fold by densitometry (Fig. 1c).
To determine whether the increased expression of G␣ i3 was due to an increase in mRNA levels, A6 cells were serum-depleted overnight, then exposed to 0.1 M aldosterone for 12-16 h, poly(A ϩ ) mRNA was isolated, and Northern blot analysis was performed. The mRNA was probed with a specific G␣ i3 probe and a probe for ␥-actin, a housekeeping gene (Fig. 2). Densitometry revealed that aldosterone induced a 1.6 -2-fold specific increase in G␣ i3 message. These results indicate that aldosterone may increase Na ϩ transport via an increase in Fluorography was carried out for 2 weeks at Ϫ80°C. c, A6 cells were treated with 1 ϫ 10 Ϫ6 M aldosterone or diluent for 18 h, and whole cell lysates were protein-matched, subjected to SDS-PAGE, and transferred to nitrocellulose for Western blotting. Membranes were probed with EC/2 antibody and goat anti-rabbit second antibody, and visualization was accomplished using an ECL kit (Amersham), an enhanced chemiluminescence system and fluorography. Shown is a representative fluorogram of 4 experiments. Molecular mass marker positions are noted on the left.

G-protein mRNA expression.
Experiments were designed to determine whether the newly synthesized G-protein becomes associated with the sodium channel. A6 cells were metabolically labeled with [ 35 S]methionine in the presence and absence of 10 Ϫ6 M aldosterone. Whole cell lysates were protein-matched and incubated with polyclonal antisera raised against a highly purified preparation of sodium channel isolated from bovine renal papilla (17) or preimmune rabbit serum using the same conditions which had previously described the association of G ai3 with the sodium channel complex (7). Immunoprecipitated proteins were subjected to SDS-PAGE under reducing conditions and autoradiographed. Fig. 3a demonstrates that aldosterone enhances the co-localization of the labeled 41-45-kDa ␣ 13 G-protein with the sodium channel (2.5-2.6-fold increase by densitometry). Consistent with previous observations (33), most other channel subunits are not metabolically labeled with [ 35 S]methionine over the time course employed here. In order to ensure that aldosterone selectively increases the expression of this subunit of the channel complex, the experiment was repeated with an extended period of metabolic labeling with [ 35 S]methionine prior to aldosterone exposure and immunoprecipitation with the sodium channel antibody performed as described above. Fig. 3b demonstrates that channel subunits are metabolically labeled at 50, 95, 70, 55, and 41-45 kDa. Only the 41-45-kDa subunit is significantly enhanced in labeling by aldosterone. This finding is consistent with previous electrophysiological and biochemical evidence that aldosterone acts primarily by activating pre-existing channels (1,2). To determine whether post-translational modifications with lipids target the induced G-protein, we examined the effects of aldosterone on palmitoylation and myristoylation of G␣ i3 . A6 cells were labeled with [ 14 C]palmitate in the presence of 1 M aldosterone or diluent. Cells were homogenized and crude membrane fraction was isolated by centrifugation at 100,000 ϫ g for 1 h and subjected to SDS-PAGE. As shown in Fig. 4a, aldosterone stimulated palmitoylation of several membrane proteins including a 41-45-kDa protein. There were also enhanced labeling of a broad band around 30 kDa, although palmitoylation of a smaller molecular mass protein at 18 kDa was not enhanced by aldosterone. This pattern of palmitoylation of membrane proteins was similar with both the 4-and 18-h exposure to aldosterone. In order to determine whether the 41-45-kDa palmitoylated protein was in fact a G-protein associated with the channel, we undertook immunoprecipitation of cellular proteins with both G-protein and Na ϩ channel antibodies following incubation with isotopically labeled acyl groups in the presence or absence of aldosterone. Whole cell lysates from A6 cells metabolically labeled with [ 14 C]palmitate in the presence of 0.1 M aldosterone or diluent were protein-matched and immunoprecipitated with with GA/1 G-protein antibody (Fig. 4b). Treatment with aldosterone resulted in a 2-fold enhancement of labeling of the 41-kDa G-protein. Several other G-proteins also appear to be palmitoylated in response to aldosterone. Similar experiments were performed with [ 3 H]myristate but failed to show incorporation of the isotopically labeled lipid into the 41-kDa G-protein FIG. 2. Northern blot hybridization of A6 mRNA for G␣ i3 and ␥-actin. C, control; A, aldosterone-treated. 5 g of mRNA was electrophoresed through 1% formaldehyde-agarose gels. A specific G␣ i3 oligonucleotide probe was synthesized to examine if the mRNA for the ␣ i3 subunit of the Na ϩ channel complex was regulated by aldosterone. The blot was probed with a specific 5Ј-32 P-labeled ␣ i3 oligonucleotide probe (CCTGGCAGCTTCCCAAA) and exposed at Ϫ80°C for 48 h to Kodak X-Omat film. This probe does not recognized G␣ i1 or G␣ i2 mRNA (data not shown). The autoradiograph for ␥-actin mRNA was obtained by stripping the G␣ i3 probe and rehybridizing with the second probe to detect actin for quantification. To control for variability, the G␣ i3 mRNA density values were normalized to the value for the ␥-actin hybridization within each sample. This is a representative autoradiograph of 3 experiments. Three 150-mm filters of confluent A6 cells were used for mRNA isolation under each condition for each experiment.  Fig. 1. 200 g of whole cell lysate was subjected to immunoprecipitation with the sodium channel antibody. Immunoprecipitated proteins were resolved on 15% polyacrylamide gels under reducing conditions. Proteins were visualized by fluorography. Molecular mass marker positions are shown on the left. A, aldosteronetreated. C, control. n ϭ 4 for each (4 -8-week exposure) experiment. In aldosterone-treated cells only, the 41-kDa band is metabolically labeled and co-localized with the sodium channel complex. b, aldosterone selectively increases the expression the 41-45-kDa G␣ i3 subunit relative to the other channel subunits. A6 cells were labeled with [ 35 S]methionine for 24 h and then exposed to 10 Ϫ6 M aldosterone or diluent for an additional 4 h. 200 g of whole cell lysate was then subjected to immunoprecipitation with the sodium channel antibody or preimmune rabbit serum. Immunoprecipitated proteins were resolved on 5-15% SDS-PAGE gels and visualized by autoradiography. Shown is a representative experiment of 4 experiments. A, immunoprecipitate from aldosterone-treated cells. C, immunoprecipitate from control cells. The last lane on the right labeled (Ϫ) represents whole cell lysates from metabolically labeled cells treated with aldosterone which were immunoprecipitated with preimmune rabbit serum. Numbers shown to the left demonstrate the molecular weight of the resolved channel subunits as determined from migration of molecular mass standards. Aldosterone increases the metabolic labeling only of the 41-45-kDa subunit of the sodium channel (average 2.5 times by densitometry). in either the presence or absence of aldosterone, although other proteins were clearly labeled (data not shown). In A6 cells, it appears that the 41-kDa G-protein is palmitoylated but not myristoylated. A6 cells were next metabolically labeled with [ 14 C]palmitate in the presence and absence of aldosterone, and whole cell lysates were subjected to immunoprecipitation with the Na ϩ channel antibody previously used to localize the Gprotein to the channel (7). When protein-matched samples were immunoprecipitated with this antibody, aldosterone specifically increased palmitoylation of a 41-kDa protein (2-3-fold by densitometry) (Fig. 4c).
The physiological relevance of this observation was examined using an inhibitor of palmitoylation, 2-fluoropalmitic acid (18). 2-Fluoropalmitic acid (Biomol, Plymouth Meeting, PA) had no effect on basal sodium transport over short time courses at the concentration of 37.5 M (Fig. 5), but markedly inhibited aldosterone-induced stimulation of sodium transport. In order to demonstrate that the action of this inhibitor might in fact be related to inhibition of G-protein palmitoylation, cells were metabolically labeled with[ 14 C]palmitate in the presence and absence of aldosterone and 37.5 M 2-fluoropalmitic acid. Protein-matched whole cell lysates were subjected to immunoprecipitation with GA/1 and subjected to SDS-PAGE and autoradiography. Fig. 6 demonstrates that aldosterone stimulates palmitoylation of several G-proteins including a band at 41 kDa and another around 30 kDa, and that labeling of these proteins with the acyl group is markedly inhibited by 2-fluoropalmitic acid.

DISCUSSION
The 41-kDa G␣ i3 subunit has been shown previously to increase the open time of the apical sodium channel in excised apical membrane patches from A6 cells (3) and to be a component of the 700-kDa sodium channel complex (7). Our results demonstrate that aldosterone stimulates the expression of mRNA for this subunit and results in an increase in the total amount of this protein, which becomes associated with the sodium channel complex. Since this G-protein is thought to have a gating effect on the sodium channel, it seems reasonable to propose that increased expression and localization of this subunit may be one mechanism whereby aldosterone activates quiescent channels already present at or near the apical membrane (1). It also seems likely that aldosterone may direct a mechanism that localizes this G-protein to a site adjacent to the channel.
Several types of post-translational modifications have been described that are associated with membrane targeting or attachment of G-proteins. A number of G-protein ␣ subunits have been shown to be post-translationally acylated with either palmitate or myristate at sites near their amino termini, and these modifications promote membrane attachment (8 -10). Smaller molecular weight G-proteins are targeted to membranes by a sequence of events involving highly conserved carboxyl-terminal cysteine residues. This pathway involves first prenylation of a cysteine residue, cleavage of the terminal amino acids, and subsequent carboxylmethylation and/or acylation (20 -22). Several previous observations suggest that these of molecular mass markers are indicated on the left. A 41-kDa G-protein is palmitoylated in response to aldosterone. Shown is a representative fluorograph of 4 separate experiments. c, sodium channel associated G␣ i3 is palmitoylated in aldosterone-treated cells. A6 cells were labeled with [ 14 C]palmitate in the presence or absence of 10 Ϫ6 M aldosterone. 200 g of whole cell lysate was subjected to immunoprecipitation with sodium channel antibody. Immunoprecipitated proteins were resolved under reducing conditions on 15% polyacrylamide gels as in Fig. 3 and exposed at Ϫ80°C for 10 weeks. Shown is a representative fluorogram of 3 separate experiments. First, aldosterone stimulates acylation-deacylation reactions (23), and inhibition of these reactions blocks both the transport response and the localization of aldosterone-induced proteins to membranes (24). Second, aldosterone stimulates carboxylmethylation reactions which result in increased channel activation (25)(26)(27). The results presented here indicate that aldosterone stimulates the palmitoylation, but not myristoylation, of the 41-kDa G␣ 13 protein. These studies have also shown that aldosterone induces increased metabolic labeling and palmitoylation of several small GTP-binding proteins most notably at 30 and 18 kDa (Figs. 1 and 6 and Ref. 32). The identity of these G-proteins and their relation to aldosterone stimulation of sodium transport is not known. Inhibitor studies suggest that palmitoylation of one or more G-proteins may be required for the full expression of the early transport response to aldoster-one. As G␣ i3 is the only G-protein currently known to be associated with a channel regulatory function, the simplest hypothesis to explain our results is a model whereby aldosterone stimulates localization of the G-protein by lipidation reactions.
The specific site of post-translational lipidations of the Gprotein subunit have not been identified in these studies. It seems unlikely that this represents the pathway of carboxylterminal lipidation and carboxylmethylation described for the targeting of small molecular weight GTP-binding proteins for several reasons. First, palmitoylation of the ␣ subunit of Gproteins has been described primarily as an amino-terminal modification (8,9). Second, although carboxylmethylation has been implicated in the action of aldosterone on the sodium channel (25)(26)(27), recent studies with purified channel complex indicate that the 95-kDa subunit is carboxylmethylated rather than the 41-kDa subunit (28). Carboxylmethylation of this 95-kDa subunit results in rapid activation of the channel (28). While these in vitro studies do not necessarily rule out carboxylmethylation of the 41-kDa subunit in intact cells, they are consistent with the recent observations of Sariban-Sohraby et al. (29), which suggest that aldosterone stimulates the methylation of a 90 -95-kDa membrane protein in A6 cells.
Taken together, the observations that aldosterone stimulates localization of a 41-kDa subunit of the channel and carboxylmethylation of a 90 -95-kDa subunit (30,31), suggest the possibility that there may be more than one action at the channel. Such a suggestion has previously been made by Asher and Garty (30) on the basis of vesicle studies. They described an early stimulation of transport by aldosterone that was not stable to vesicle preparation and a prolonged effect that was stable to vesicle preparation. Carboxylmethylation reactions of membrane-bound proteins are known to be reversible reactions due to the presence of methylesterases in the membrane (31). We speculate that aldosterone may activate sodium channels already residing in the apical membrane, both through an early, reversible carboxylmethylation and a later, more stable association of a gating G-protein.