Characterization and cellular localization of the epithelial Na+ channel

Vol. 266, No. 6, Issue of February 25, pp. 3907-3915,1991 Printed in U.S. A.

OF BIOLOGICAL CHEMISTRY THEJOURNAL

Characterization and Cellular Localization of the Epithelial Na+ Channel STUDIES USING AN ANTI-Na' CHANNEL ANTIBODY RAISED BYAN

ANTIIDIOTYPIC ROUTE* (Received for publication, September 4, 1990)

Thomas R. KleymanSg, Jean-Pierre Kraehenbuhlll, and StephenA. Ernst11 From the $Departments of Medicine and Physiology, University of Pennsylvania and the Veterans Affairs Medical Center, Philadelphia, Pennsylvania 19104, the 7Department of Biochemistry, University of L u a n n e and the Swiss Institute for Experimental Cancer Research, Epalinges CH 1066, Switzerland, and the IlDepartment of Anatomy and Cell Biology, University of Michigan, Ann Arbor, Michigan 48109

Amiloride-sensitive Na+ channels are expressed at the apical membrane of high resistance, Na+-transporting epithelial. The specific interaction of amiloride with this transport protein suggested the feasibilityof raising anti-Na+ channel antibodies by an antiidiotypic approach designed togenerateantibodiesdirected against the amiloride-bindingdomain on the channel. Antiidiotypic monoclonal antibody RA6.3 mimicked the effect of amiloride by inhibiting Na+ transport across A6 cell monolayers when applied to the apical cell surface. Inhibition of transport required pretreatment of the apical cell surface with trypsin in the presence of amiloride in order to enhance accessibility of the antibody to theamiloride-binding site. This antibody specifically immunoprecipitated a large 750,000-700,000 Da protein from [36S]methioninelabeled A 6 cell cultures, which was resolved further under reducingconditions as a set of polypeptides with apparent molecular masses of 260,000-230,000, 180,000, 140,000-110,000, and 70,000 Da. The antibody recognized the 140,000-Da subunit, known to contain theamiloride-binding domain, on immunoblots of purified A6 cell Na+ channel. Immunoprecipitation of apical or basolateral plasma membrane proteins selectively labeled with "1 ' demonstrated that expression of the oligomeric Na+ channel was restricted to the apical plasma membrane. Immunocytochemical localization in A 6 cultures revealed apicalmembrane as well as cytosolic immunoreactive sites. Immunostaining was also observed at or near basolateral the plasma membrane.

review, see Gartyand Benos, 1988; Sariban-Sohrabyand Benos, 1986). The epithelial Na' channel differs from other Na+-selective ion channels found in nerve and muscle in that it is specifically inhibited by submicromolar concentrations of the diuretic amiloride, is insensitive to tetrodotoxin, andis not voltage gated. Conflicting results have been reported recently regarding subunit composition of the epithelial Na' channel isolated from an established epithelial cell line or from mammalian kidney using purification schemes based on binding of high affinity analogs of amiloride. Benos et al. (1987) isolated a large 730,000-Da protein that iscomposed of five distinct subunits, varying in molecular mass from 55,000 to 315,000 Da. Barbry et al. (1987, 1990) isolated a 185,000-Da protein that is composed of two, possibly identical, subunits. Both groups have observed amiloride-sensitive "Na flux after reconstitution of purified protein into lipid vesicles (Benos et al., 1987; Barbry et al., 1990). As an alternative approach, antiidiotypic antibodies raised against an anti-amiloride antibody could provide a tool to examine both biochemical characteristics of the Na' channel as well as its cellular localization. This method is based on the apparent structuralmimicry of drug-binding domains on receptor proteins and on antidrug antibodies (Sege and Peterson 1978; for review, see Erlanger et al. 1986) and does not require prior purification of the Na+ channelin order to raise antichannel antibodies. Antibodies directed against receptors for acetylcholine (Wassermann et al. 1982), &adrenergic agonists (Homcy et al., 1982), adenosine (Ku et al., 1987), glucocorticoids (Cayanis et al., 1986), insulin (Sege and Peterson, 1978), retinol-binding protein (Sege andPeterson, 1978), and thyroid-stimulating hormone (Hill and Erlanger, 1988) have been raised by generating antiidiotypic antibodies A Na+-selective cation channel is present in the apical against specific antidrug (or hormone) antibodies. plasma membrane of high resistance, Na+-transporting epiThe specific interaction of different Na+-selectivetransport thelia. This protein provides the rate-limiting step for reab- proteins with certain amiloride analogs suggested that the sorptive Na' transportandconstitutesa critical site for amiloride-binding site on these transport proteins might rechormonal regulation of transepithelial Na' transport (for ognize different regions of the amiloride molecule (for review, see Kleyman and Cragoe, 1988). For example, modification of *This workwas supported in part by grants from the Cystic amiloride by the addition of hydrophobic substituents on the Fibrosis Foundation and the Zyma Foundation (to T. R. K.), the terminal nitrogen of the guanidino moiety of amiloride inSwiss National Science Foundation (398086) (to J. P. K.), and the National Institutes of Health (DK27559) and Biomedical Research creased both the affinity and specificity of amiloride analogs Support (RR05385) (to S. A. E.). The costs of publication of this for the Na' channel. The concentrations of several of these article were defrayed in part by the payment of page charges. This analogs required to achieve a 50% inhibition of the Na+ article must therefore be hereby marked "advertisement" in accord- transport (IC,,,) were less than 10 nM (Cuthbert and Fanelli, ance with 18 U.S.C. Section 1734 solely to indicate this fact. 1978;Kleyman and Cragoe, 1988).Alternatively, modification Recipient of a Clinician-Scientist Award from the American Heart Association. To whom correspondence should be sent: Renal of amiloride by the addition of hydrophobic substituents on Section, Dept. of Medicine, 700 Clinical Research Bldg., University its 5-aminogroup decreased affinity of the ligand for the Na' of Pennsylvania, Philadelphia, PA 19104-6144. channel markedly although several of these analogs inhibited

3907

3908

Epithelial Na+ Channel

the Na+/H+ exchanger with ICsoin the range of between 10 and 100 nM (Zhuang et al., 1984). We havepreviously raised and characterized polyclonal antibodies that recognize distinct epitopes on amiloride (Kleyman et al., 1989b). One antibody was raised against amiloride coupled to carrier protein (bovine serum albumin (BSA)l) through the guanidino group in order to expose epitopes relevant in binding to the epithelial Na+ channel (Kleyman et al., 1986a). This antibody bound amiloride analogs in a manner similar to the Na' channel and suggested that the antibody was directed against an epitope on amiloride which is involved in specific binding to the Na+ channel. We felt that this polyclonal antibody would be suitable to screen for antiidiotypic antibodies that, like amiloride, recognize the amiloride-binding site of the Na+channel. We have raised antiidiotypic antibodies, directed against polyclonal anti-amiloride antibodies, which also recognize the epithelial Na+ channel. These antibodies were raised with an autoantiidiotypic approach in which mice are immunized with a hapten (amiloride) coupled to carrier protein and allowed to generate cells secreting antihapten (idiotypic) antibodies as well as cells secreting antiidiotypic antibodies (Cleveland et al., 1983). These antibodies were used to examine biochemical characteristics and cellular localization of the Naf channel in A6 cells, an established epithelial cell line derived from Xenopus kidney which expresses functional amiloride-sensitive Na+ channels at the apical plasma membrane (Perkins and Handler, 1981). EXPERIMENTAL PROCEDURES

Materials-Amiloride and benzamil were a gift from Merck. Biotinconjugated anti-mouse Ig raised in sheep and peroxidase-conjugated streptavidin were purchased from Amersham Corp. Reagents for immunocytochemistry were from Jackson Immunoresearch Laboratories (West Grove, PA). All other compounds were reagent grade. Amiloridecoupled to BSAwas prepared as described previously (Kleyman et al., 1986a). Anti-amiloride antibodies were raised and affinity purified as described previously (Kleyman et al., 1986a). Cell Culture Conditions and Physiologic Measurements-A6 cells derived from Xenopus laevis kidney were obtained from the American Tissue Type Collection and cloned by limiting dilution (Verrey et al., 1989). Cell culture conditions and measurements of potential difference and short circuit current were as described previously (Kleyman et al., 1989a). A6 cells (passages 80-95)were seeded on collagencoated polycarbonate filters (Nucleopore) at lo6 cells/cm2. Collagen coating was performed as follows.Collagen type 1 (Vitrogen 100, Collagen Corp.) was applied as a thin layer over polycarbonate filters, polymerized with NH,OH, and then cross-linked with 2.5% glutaraldehyde in 100 mM NaCl, 10 mM NaP04, pH7.4. Filters were washed extensively with water followed by tissue culture medium prior to use. Cells were kept in a humidified 28 "C incubator with 5% CO, for 9 days and exposed to media supplemented with aldosterone (300 nM) for 16 h prior to biochemical or electrophysiologic studies in order to maximize Na+ transport (Kleyman et al., 1989a). For studies utilizing apical cell surface trypsinization, A6 cells were washed in serum-free medium, and aftera 5-min incubation in serum-free medium with 50 p~ amiloride added to thesolution bathing the apical surface, trypsin was added to apical solution a t a final concentration of 1pg/ml (Tousson et al., 1989). After a 30-min incubation at 28"C, the cells were washed extensively with media supplemented with 5% fetal calf serum. Cells were then placed in amodified Ussing chamber for electrophysiologicstudies. Antiidiotypic Antibodies-Monoclonal antiidiotypic antibodies directed against polyclonal anti-amiloride antibodies were generated using the method of Cleveland et al. (1983). BALB/c micewere immunized with an intraperitoneal injection of 0.1 ml of the amiloride-BSA conjugate (Kleyman et al., 1986a) in complete Freund's The abbreviations used are: BSA, bovine serum albumin; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; EGTA, [ethylenbis(oxyethylenenitrilo)]tetraaceticacid PBS, phosphate-buffered saline.

adjuvant (1mg/ml). Mice wereboosted 23 days later with an identical intraperitoneal injection of the amiloride-BSA conjugate in incomplete Freund's adjuvant (1 mg/ml). After 5 days spleen cells were isolated and fused with a nonsecreting mouse myeloma cell line as described previously (Kleyman et aL, 1989b). Hybridoma supernatants were screened for binding to affinity-purified rabbit anti-amiloride antibody using an enzyme-linked solid phase immunoadsorbent assay (see below). Cells from positive wells were subcloned by the method of limiting dilution. Four separate clones, RA2.4,RA3.1, RA6.3, and RA9.3, were isolated. Antibody isotypes were determined with an isotyping kit purchased from Amersham Corp. The isotypes of RA2.4, RA3.1, RA6.3, and RA9.3 were IgG,, IgG2b, IgG1,and IgG1, respectively. Enzyme-linked Zmmunosorbent Assay-Microtiter wells (96 well plates) were coated overnight with 1pglwell of affinity-purified rabbit anti-amiloride antibody in 50 pl of 0.1 M Na2C03/NaHC03,pH 9.5. Wells were rinsed and then incubated for 1 h at 37 "C (or overnight at 4 "C) with 100 p1 of 0.15 M NaC1,lO mM NaP04, pH 7.4, containing 0.1% Tween 20 and 5% (w/v) nonfat dry milk (buffer A) to prevent nonspecific adsorption of antibodies. Wells were rinsed and then incubated for 2 h at 37 "C with antibody (with or without amiloride analogs) in 50 plof buffer A without 5% nonfat dry milk. The dilution of the antibody and concentrations of the amiloride analogs used are listed in the figures or figure legends. Bound antibody was detected using a secondary biotinylated anti-mouse Ig antibody and streptavidin-conjugated peroxidase as described previously (Kleyman et al., 1989b). Peroxidase assay was initiated with the addition of 0.07% (w/v) o-phenylenediamine in 0.1 M sodium citrate, pH4. The reaction was stopped by the addition of 0.05% NaN3 in 0.1 M sodium citrate, pH 4, and absorbance a t 492 nm was measured. Zmnunoblots-The epithelial Na+ channel was purified from A6 cells as described previously, solubilized with SDS, and subjected to 10% SDS-PAGE (350 ng of purified epithelial Na+ channel protein/ lane). Proteins were then transferred to nitrocellulose (Towbin et al., 1979), and binding of anti-idiotypic antibody was then determined as described by Sorscher et al. (1988). Biosynthetic Labeling-A6 cells were seeded on collagen coated polycarbonate filters (5 cm'), kept in culture for 9 days, and then incubated overnight in the presence of 300 nM aldosterone. The monolayers were rinsed three times for 5 min in methionine-free serum-free medium supplemented with 300 nM aldosterone. The filter supports were inverted, and 200 p1 of methionine-free serum-free medium supplemented with 300 nM aldosterone and 1 mCi/ml [35S] methionine was added to thebasolateral surface of the filter. Following a 15-minincubation at 28 "C,filters were washed once and placed for up to 1h (see legend to Fig. 4) in serum-free medium supplemented with 300 nM aldosterone and 10 mM methionine. The monolayers were then washed once at 4°C with amphibian Ringer's solution containing 1 mM CaC12and protease inhibitors (1p M antipain, 1 p M leupeptin, 1 p~ pepstatin A, and 0.1 mM phenylmethylsulfonyl fluoride). Apical or Basolateral Cell Surface Radiowdination-A6 cells were seeded on collagen-coated polycarbonate filters (5 cm'), kept in culture for 9 days, and then incubated overnight in the presence of 300 nM aldosterone. The monolayers were rinsed three times for 5 min at 4 "C with amphibian Ringer's solution containing 11mM glucose and 1 mM CaCl' andthen radioiodinated with an enzyme-catalyzed method. Amphibian Ringer's solution containing 11 mM glucose, 1 mM CaC12, 50 plof lactoperoxidase (1 mg/ml, 60-80 units/mg protein), and 1 mCi of NalZ5 (carrier free) was added to the apical or basolateral surface. The reaction was started with the addition of 20 plof glucose oxidase (diluted 1:lOO from stock, Sigma type 5), and after 15 min at 4 "C, 10 pl of 20% NaN3 (w/v) was added to stop the reaction. Aliquots removed from apical and basolateral solutions were counted. Monolayers in which the leak of lZ5Iacross the monolayer was less than 1%were used subsequently. The apical and basolateral surfaces were washed five times with iced amphibian Ringer's solution containing 1 mM CaC12, 0.05%NaN3, and protease inhibitors. Zmmunoprecipitation-A6 cell proteins were labeled with 35Sor 'I as described above and then lysed and cellular proteins collected in 0.5mlof a buffer containing 0.4% (w/v) sodium deoxycholate, 1% (v/v) Nonidet P-40, 50 mM EGTA, 10 mM Tris-HC1, pH 7.4, and protease inhibitors. After a 2-min centrifugation (12,000 X g) to pellet-insoluble material, 10-pl aliquots were removed from the supernatant for SDS-PAGE to determine protein recovery andto quantitate lZ5I or 35Sincorporation. Following trichloroacetic acid precipitation, the pellet was resuspended in formic acid. Samples were removed for y- (or liquid scintillation) counting and protein deter-

3909

Epithelial Nuc Channel mination. Equal amounts of solubilized proteins were then diluted to 3 ml with 150 mM NaC1, 5 mM EGTA, 1%(v/v) Triton X-100, and 50 mM Tris, pH 7.4, containing protease inhibitors. After a 48-h incubation at 4°C with a 1:lOO dilution ofRA6.3 (in ascites) or a control monoclonal antibody (anti-amiloride BA1.l raised in ascites), 10 pg of a rabbit antibody directed against mouse IgG (Sigma) was added and incubated for 2 h a t 4 "C. Protein A-Sepharose was added (50 ~ l )and , after a 1-h incubation a t room temperature, the Sepharose beads were collected by brief centrifugation and washed four times. The initial two washes werewith a solution containing 150 mM NaC1, 5 mM EGTA, 1%(v/v) Triton X-100, and 50 mM Tris, pH 7.4, with protease inhibitors. The final two washes were with a solution containing 0.1% (w/v) SDS, 2 mM EGTA, and 10 mM Tris-HC1, pH 7.4, with protease inhibitors. Bound protein was eluted into 125 p1 of a solution containing 3% (w/v) SDS, 15%(w/v) sucrose, and 92.5 mM Tris-HC1, pH 6.9, by heating the sample a t 90 "C for 5 min. Samples for SDS-PAGE under nonreducing conditions were applied directly to 3% gels (2.4% stacking gel). Samples for SDS-PAGE under reducing conditions were incubated for 15 min with 1 mM dithiothreitol andthen with 2 mM iodoacetamide for 15 min. Sample pH was adjusted with 1 p1 of375 mM Tris-HC1, pH 8.9, and samples were applied to 5-1376 linear gradient gels (3.9%stacking gel). SDS-PAGE was performed as described previously (Maizel, 1971). Equal amounts of protein were used for SDS-PAGE when comparing the pattern of total cell surface-labeled proteins under different conditions. Gels containing [35S]methionine-labeledproteins were incubated for 30 min in Amplify (Amersham Corp.) prior to autofluorography. Immunocytochernistry-A6 cells weregrown on collagen-coated semipermeable supports essentially as described above. Cultures were fixed at room temperature with 4% paraformaldehyde in phosphatebuffered saline (PBS) for 2-4 h. For comparison, some monolayers werefixed with paraformaldehyde/lysine/perodiate for 6 h. After rinsing in PBS, cultures were infiltrated with graded sucrose/PBS solutions to a final concentration of2.3 M sucrose, cut into strips, and then frozen in liquid nitrogen. Semithin frozen sections (0.5-1 pm thick) were prepared following the methods of Tokuyasu and Singer (1976) using the Reichert-Jung FC 4E cryokit attachment to the Reichert Ultracut E ultramicrotome. After rinsing in PBS, sections were preincubated with 20% normal rabbit serum in PBS.After a 1-h exposure to monoclonal antibody RA6.3 diluted appropriately with 2% normal rabbit serum/PBS, sections were rinsed with PBS and 2% normal rabbit serum/PBS and then incubated for 1 h with rabbit anti-mouse IgG/IgM conjugated to fluorescein isothiocyanate (diluted 1:lOO with 2% normal rabbit serum/PBS). After rinsing, sections were mounted with 0.1% p-phenylaminediamine in 20% PBS/glycerol and examined and photographed with a Leitz Aristoplan fluorescent microscope. For control studies, RA6.3 was replaced with a monoclonal antibody directed against amiloride (BAl.1). Similar immunofluorescence localizations were carried out with semithin frozen sections of outer medulla from rabbit kidneys fixed by perfusion via the renal artery with 4% paraformaldehyde for 10 min and subsequent immersion in fixative for 3 h. Statistics-Results are expressed as the mean k S.E. RESULTS

0.00

$:lo

1:100 1:1.000 1:10.000 1:100.000

Antlbody dilution

b.

U

c

P

;jP

"1

04 -8

1

-?

-6

-5

.4

-3

lOg I M b r l d .M d O O l (u) FIG. 1. Binding of monoclonal antiidiotypic antibody RA6.3 to rabbit polyclonal anti-amiloride antibody. Amiloride coupled to bovine serum albumin was used as animmunogen to generate antiamiloride (i.e. idiotypic) antibodies. BALB/c mice were immunized intraperitoneally and boosted at day 21. Five days later spleens were excised and cells fused with nonsecreting myeloma cell line. Hybridomas secreting antiidiotypic antibodies were identified by the ability to bind to heterologous (rabbit) idiotype (polyclonal anti-amiloride antibody), as determined by enzyme-linked solid phase immunosorbent assay (see "Experimental Procedures"). a, binding of varying dilutions of RA6.3 (0)and of a control (anti-amiloride) monoclonal antibody (m) to rabbit polyclonal anti-amiloride antibody. b, inhibition of RA6.3 (1:lOO dilution) binding to a polyclonal anti-amiloride antibody by varying concentrations of amiloride (m) and theamiloride analog benzamil (0).Binding was measured by enzyme-linked immunosorbent assay. Results are listed as percent of RA6.3 bound in the absence of amiloride or benzamil.

The binding of the antiidiotypic antibodies to a polyclonal Identification of Hybridomas Secreting Antiidiotypic Antibodies-Antiidiotypic monoclonal antibodies were generated anti-amiloride antibody was measured in the presence or with the protocol of Cleveland et al. (1983) using amiloride absence or varying concentrations of amiloride or of the coupled to BSA through its guanidino moiety (Kleyman et ai., amiloride analog benzamil, in order to determine whether 1986a) as an immunogen. This protocol is designed to allow antiidiotypic antibodies bound at or near the amiloride-bindfor the generation of populations of spleen cells that secrete ing site on the anti-amiloride antibody. Both amiloride and antibodies that bind amiloride (idiotypic antibodies) as well benzamil inhibited the binding of antibody RA6.3 to the as those that secrete antibodies that are directed against the idiotypic antibody (Fig. l b ) . No inhibition of binding was variable region on the anti-amiloride antibodies (antiidiotypic observed with the otherantiidiotypic antibodies. antibodies). Hybridoma cells secreting antiidiotypic antibodAntibody RA6.3 Inhibits Transepithelial Nu' Transporties were identified by screening hybridoma supernatants with Transepithelial transport ofNa' across A6 monolayers is rabbit polyclonal anti-amiloride antibodies, using a solid mediated by an epithelial Na+ channel on the apical cell phase immunoassay. Four clones secreting antiidiotypic an- surface and a Na+/K+-ATPase on the basolateral cell surface. tibodies were isolated, and cells were then subcloned by the Amiloride both bindsto theepithelial Na' channel and inhibmethod of limiting dilution and used to raise antibodies in its Na+ transport mediated by this channel. When A6 cells ascites. The binding of one of the antiidiotypic antibodies, are grown on a semipermeable support, net transepithelial RA6.3, to a rabbit polyclonal anti-amiloride antibody is illus- Na+ transport can be determined by measurement of short trated in Fig. la. No binding was observed with a control circuit current (Perkins and Handler, 1981). As the antiidimonoclonal antibody raised against amiloride in ascites. otypic antibodies in this study are intended to be directed

3910

Channel

Na+

Epithelial

against the amiloride-binding site on the Na’ channel, studies were performed to examine whether these antibodies could mimic the effect of amiloride by inhibiting Na’ transport across A6 cell monolayers. No inhibition of Na’ transport, as measured by short circuit current, was observed with a 1:lOO dilution of RA6.3 (Fig. 2a, column 1 ( n = 5)). Subsequent

a

2

1

3

4

b. 120

-

BA 1.1

E

100-

H

Arnlloride

-ae 6

;

RA 6.3

*O-

c

60-

e

2

40-

B

E

20-

fn

I

-

0

-25

0

5 02 5

7 5 125 100

Time (mln)

FIG.2. Antibody RA6.3 inhibits transepithelial Na’ transport. a, effect of RA6.3 and control monoclonal antibody on short circuit current. Short circuit current was measured across A6 cells grown on collagen-coated filters as described under “Experimental Procedures.” Column I , RA6.3 (1:lOO dilution) was added tothe solution bathing the apical plasma membrane. No inhibition of short circuit current was observed over 90 f 8 min ( n = 5). Columns 2-4, apical cell surface was exposed to 1 pg/ml trypsin in the presence of 50 p~ amiloride in order to improve access of antibodies to the Na’ channel. Column 2, RA6.3 (1:lOO dilution) was added to the solution bathing the apical plasma membrane after apical cell surface trypsinization. 26 t 2% inhibition of short circuit current was observed over 70 f 4 min ( n = 5). Column 3, control anti-amiloride monoclonal antibody BA1.l (1:lOO dilution) was added to the solution bathing the apical plasma membrane after apical cell surface trypsinization. No inhibition of short circuit current was observed over 74 f 9 min ( n = 5). Column 4, RA6.3 (1:OO dilution) was added to the solution bathing the basolateral plasma membrane after apical cell surface trypsinization. No inhibition of short circuit current was observed over 77 t 6 min ( n = 5). b, time course of the effect of RA6.3 (1:lOO dilution) and control BA1.l (1:lOO dilution) on short circuit current. A6 cells were grown on collagen-coated filters, and apical cell surface proteins were exposed to trypsin in the presence of amiloride in order to improve access of antibodies to theNa’ channel as described under “Experimental Procedures.” Cells were then placed in a modified Ussing chamber, and short circuit current was monitored continuously. Short circuit current measurements taken a t 2-min intervals are plotted. Addition of RA6.3 (boM line) at time = 0 resulted in a slowly progressive inhibition of short circuit current after a delay of 12 min. The shortcircuit current fell 42% of its initial value over 120 min. Subsequent addition of 10 p~ amiloride (arrow) resulted in a rapid inhibition of short circuit current to14% of its initial value. No inhibition of short circuit current was observed over 120 min after the addition of anti-amiloride BA1.l (thin line) a t time = 0.

addition of amiloride (lo-$ M ) inhibited short circuit current as expected. The lack of inhibition of short circuit current by RA6.3 was not surprising, as Hamilton and Eaton (1985) and Palmer (1984) had suggested previously that the amiloridebinding site on the intact channel at thecell surface is within the channel’s pore. Accordingly, this site might not be accessible to thebinding site on the antiidiotypic antibody. Proteolysis of apical cell surface proteins in the presence of amiloride mightenhance accessibility of the antiidiotypic antibody for the amiloride-binding site. Previous studies by Garty and Edelman (1983) and Tousson et al. (1989) demonstrated thatwhen trypsin is added to thesolution bathing the apical cell surface of toad urinary bladder or A6 cells in the presence of amiloride, minimal proteolysis of the Ne’ channel occurs, as determined by measurement of amiloride-sensitive Na’ transport or by immunofluorescence. The effect of RA6.3 (1:lOO dilution) on Na’ transport across A6 cell monolayers was measured after exposure of the apical cell surface to 1pg/ ml trypsin for 30 min in thepresence of 50 p~ amiloride. The cells were washed and allowed to establish a new steady-state short circuit current before application of the antibody. As shown in Fig. 2a (column Z),short circuit current declined to 76 k 4% ( n = 5) of the initial value after a 70-min exposure to RA6.3. In contrast, no inhibition was observed after the addition of a control anti-amiloride monoclonal antibody to the solution bathing theapical cell surface, nor was inhibition of short circuit current noted after the addition of RA6.3 to the solutionbathing the basolateral cell surface (Fig. 2a, columns 3 and 4 ) . The time course of inhibition ofNa’ transport byRA6.3 is shown in Fig. 2b. The short circuit current began a steady decline 5-15 min after addition of the antibody, with up to 42% of the currentinhibited by 120 min, the longest period measured. Subsequent additionof amiloride ( M ) reduced the short circuit current to 14% of the initial value measured prior to theaddition of RA6.3. The relatively slow reduction of short circuit current after the addition of RA6.3 differed from that seen with amiloride in which inhibition occurred within seconds after the addition of the drug and reached plateau levels within several minutes. No decline in Na’ transport was seen after apical exposure to a control monoclonal antibody over the 120-min time period (Fig. 2b). Antibody RA6.3 Recognizes the Amiloride-binding Subunit of the Purified Na’ Channel-The antiidiotypic antibody RA6.3 inhibited Na’ transport across A6 cells, presumably by binding to the amiloride-binding site on the channel, and therefore might recognize the amiloride-binding subunit of the Na’ channel by other methods, such as immunoblot. Benos et al. (1987) described recently the purification and reconstitution of the epithelial Na+ channelfrom both bovine kidney papilla and the epithelial cell line A6. This is a large 730,000-Da protein that is composed of five distinct subunits. Benos et al. (1987) and Kleyman et al. (1989a) used photoactive amiloride analogs to identify the amiloride-binding subunit of the Na’ channel, which is a polypeptide with an apparent molecular ‘massof between 130,000 and 180,000 Da. Weak labeling of a 50,000-Da polypeptide has also been reported by both groups. An immunoblot of purified Na’ channel was probed with antiidiotypic antibodies or with a control monoclonal antibody in order to determine whether the antiidiotypic antibodies recognized the amiloride-binding subunit of the channel. The antiidiotypic antibody RA6.3 bound to the 140,000-Da subunit of the purified epithelial Na’ channel whereas no binding was observed with a control monoclonal antibody (Fig. 3). Weak binding of the 50,000-Da subunit was also observed. Biochemical Characterization of the A6Cell Na+ Channel-

Epithelial Na’ Channel

3911

quently incubated with excess unlabeled methionine for 1 h (chase). Cellswere then detergent solubilized, and labeled Na’ channels immunoprecipitated with RA6.3 were reduced with dithiothreitol and identified by 5-13% SDS-PAGE and autofluorography. A complex of polypeptides was immunoprecipitated from solubilized cells after the l-h chase, with apparent molecular masses of 260,000-230,000, 180,000, 140,000, and 70,000 Da (Fig. 5, lane 2). A large 300,000-Da polypeptide was also observed, although not consistently. This complex of polypeptides was not seen after the 15-min pulse (Fig. 5, lane 1), nor was an obvious precursor of the amiloridebinding subunit (140,000 Da) resolved. This suggests that greater than15 min was required for assembly and/or efficient recognition by the antiidiotypic antibody. Thelack of recognition of a complex of polypeptides or precursor polypeptides by RA6.3 after the 15-min pulse wasa result not of insufficient labeling of cell proteins. The profile of total A6 cell proteins RA6.3 CONTROL biosynthetically labeled after the 15-min pulse (Fig. 5, lane 3 ) FIG. 3. Binding of RA6.3 to purified Na+ channel isolated and l-h chase (Fig. 5, lane 4 ) demonstrated that thespecific from A 6 cells. Na+ channel was purified from A6 cells, and immu- activity of labeled proteins was clearly greater after the 15noblots (350 ng of Na’ channel/lane) were probedwith a 1:50 dilution min pulse as compared with thel - h chase. of RA6.3 raised in ascites. RA6.3 recognized the 140,000-Dasubunit Polarized Cell SurfaceExpression of theEpithelial Na+ of the Na+ channel (arrow). Bound antibody was detected with a Channel-The Na+ channel is functionally restricted to the secondary peroxidase-conjugated anti-mouse antibody and a peroxidase assay, as described under “Experimental Procedures.”No bind- apical plasma membrane of A6 cells (Perkins and Handler, a control monoclonal antibody raised in ascites. ing was observed with 1981), and it is likely that Na’ channel protein expressedat

1

669

-

2

1

3

2

4

c. 205

-

t t

335 -

443

205

-

FIG. 4. Immunoprecipitation of [“S]methionine-labeled A 6 cell proteins with RA6.3; SDS-PAGEunder nonreducing conditions. A6 cells grown on collagen-coated filters were labeled with [:“S]methioninefor 15 min and chased for 1 h in media supplemented with 10 mM unlabeled methionine as described under “Experimental

Procedures.” Labeled proteins were immunoprecipitatedwith RA6.3 (lane 2 ) or with control monoclonal antibody (lane I ) and detected by 3% SDS-PAGE under nonreducing conditions followed by autofluorography. A 750,000-700,000-Da protein specifically immunoprecipitated is indicated by the arrow. Migration of various molecular mass standards is indicatedto the left of the figure. In orderto estimate thesize of the Na’ channel complex, A6 cell proteins were labeled biosynthetically with [35S]methionine. Na’ channel proteinwas then immunoprecipitated with either RA6.3 or a control monoclonal antibody. Labeled proteins were identified by SDS-PAGE under nonreducing conditions and autofluorography. A protein with an apparent molecular mass in the range of 750,000-700,000 Da was identified (Fig. 4, lane 2). T o examine subunit structure of the Na+ channel, newly synthesized A6 cell proteins were labeled with [35S]methionine for 15 min (pulse) or were pulsed for 15 min and subse-

97 -

116

66

-

45

-

30

-

t

FIG.5. Immunoprecipitation of A 6 cell proteins labeled with [S5S]methionine;SDS-PAGE under reducing conditions. A6 cells grown on collagen-coated filters were incubated with [3sS] methionine for 15 min (lunes 1 and 3 ) or subsequently chased for1 h (lanes 2 and 4 ) in media supplemented with 10 mM unlabeled methionine as described under “Experimental Procedures.” Labeled proteins were immunoprecipitatedwith RA6.3 (lanes 1 and 2 ) and

detected by 5-13% SDS-PAGE under reducing conditions followed by autofluorography.Total labeled A6 cell proteins were alsodetected by 5-13% SDS-PAGE under reducing conditions followed by autofluorography (lane 3, 15-min pulse; lane 4, l-h chase). Migration of various molecular massstandards is indicated to the left of the figure. Labeled polypeptides immunoprecipitated with RA6.3 are indicated by arrows to the right of the figure.

Epithelial Na+ Channel

3912

the plasma membrane is restricted to the apical cell surface. A method of enzyme-catalyzed radioiodination was used to label specifically proteins expressed at either the apical surface or basolateral surface of cultured A6 cells (Kleyman et al., 1989a).To identify apical membrane Na' channels, detergent-solubilized cells were immunoprecipitated with the antiidiotypic antibody RA6.3 or with a control monoclonal antibody. RA6.3 specifically immunoprecipitated a complex of polypeptideswith apparent molecular masses of260,000230,000, 180,000,140,000,110,000,and 70,000 Da (Fig. 6,lane 4 ) . Both a 300,000- and a 50,000-Da subunit were also observed, although not consistently. The 50,000-Dapolypeptide might represent proteolysis of a higher molecular mass subunit. The complex of polypeptides was found to be restricted to the apical plasma membrane, as no labeled polypeptides were recovered followingimmunoprecipitation of basolateral radioiodinated cell surface proteins (Fig. 6,lane 3). No labeled polypeptideswereresolvedfollowing immunoprecipitation with control monoclonal antibody (Fig. 6, lane 1). The pattern of total radioiodinated apical cell surface proteins (Fig. 6, lane 7) suggests that theepithelial Na' channel is a minor component of the pool of labeled apical cell surface proteins. The pattern of total labeled basolateral cell surface proteins (Fig.6, lane 6 ) differedfrom apical cell surface proteins although the specific activity of labeling was similar. Confluent A6 cell monolayersgrown on permeable support express Na' channels with electrophysiologic properties that differ from Na' channels expressed in cells grownon perme-

1 2

3 4

5

6

able support (Hamilton and Eaton,1986). Alterations in both cation selectivity as well as mean open time have been reported. To determine whether these differences are reflected in changes in the subunit structure of Na' channels, iodinated apical membranes of A6 cells grown on a nonpermeable support were immunoprecipitated with RA6.3. As withcells grown on a permeable support, a similar complex of polypeptides with apparent molecular masses of260,000-230,000, 180,000,140,000, 110,000,and 70,000 Da was resolved (Fig. 6, lane 2). Although the specific activity of labeled proteins was similar, the patternof total labeled apical cell surface proteins differed when cellsgrown on a nonpermeable support (Fig. 6, lane 5) were compared with cells grown on a permeable support (Fig. 6, lane 7). ImmunocytochemicalDistribution of Nu' Channel Protein in Confluent A6 Monolayers-Immunofluorescence studies were carried out with RA6.3 on semithin (0.5-1pm thick) frozen sections of formaldehyde-fixed A6 cell cultures. The general distribution of staining is shown at low power in Fig.

7

c 11697 66

-

45

-

29-

FIG.6. Specific A6 cell surface radioiodination and immunoprecipitation withRA6.3. A6 cells were grownon either plastic support (lanes 2 and 5 ) or collagen-coated filters (lanes I , 3, 4, 6, and 7). Apical (lanes I , 2, 4, 5, and 7) or basolateral (lanes 3 and 6 ) membranes were radioiodinated (as described under "Experimental Procedures"), and labeled proteins immunoprecipitated with RA6.3 (lanes 2-4) or with a control antibody ( l a n e I ) . Immunoprecipitated proteins were analyzed by 5 1 3 % SDS-PAGE under reducing conditions and autoradiography (lane I , apical/filter/control monoclonal antibody; lane 2, apical/plastic/RA6.3; lane 3, basolateral/filter/ RA6.3; lane 4 , apical/filter/RA6.3). Total labeled A6 cell proteins were also detected by 513% SDS-PAGE under reducing conditions followed by autoradiography (lane 5, apical/plastic; lane 6,basolateral/filter; lane 7 apical/filter). Migration of various molecular mass standards is indicated to the left of the figure. Labeled polypeptides immunoprecipitated with RA6.3 are indicated by arrows.

FIG. 7. Immunofluorescence localizationof Na+ channels in A 6 cellmonolayerswith RA6.3. A , the apical membrane is immunostained (arrowheads). In addition, accumulation of cytosolic immunofluorescence is seen in subapical areas (arrows)and in deeper regions of the cells where the staining is often close to or at the basolateral cell borders. Ascites fluid was diluted 1:200. The collagen layer was variably stained with this antibody. Although no staining of the collagen layer was observed with the control monoclonal antiamiloride antibody, staining of this layer was noted with a control polyclonal anti-amiloride antibody. B , the immunostaining associated with the basolateral plasma membrane often appears punctate (arrow). Note apical membrane staining (arrowheads) and subapical immunofluorescence. Hybridoma supernatant was diluted 1:2. C, immunofluorescence staining is absent when a monoclonal antiamiloride antibody (BA1.l) replaces RA6.3. Ascites fluid was diluted 150. Bar, 10 pm.

Epithelial Nu' Channel 7A. The apical membrane was immunoreactive, and immunofluorescence was present in the subapical cytoplasm where it varied in amount from cell to cell and oftenappeared punctate. In deeper regions of the cytosol, immunostaining was frequently distributed at or near thebasolateral plasma membrane (Fig. 7, A and B).Similar patterns of localization were seen with RA6.3 raised in ascites (Fig. 7A) or as hybridoma supernatant (Fig. 7B),and withsections fixed with paraformaldehyde-lysine-periodate. Controls in which a monoclonal anti-amiloride antibody(or buffer alone) replaced RA6.3 were unstained (Fig. 7C). To testfor cross-reactivity and specificity of RA6.3 for cell types in heterogeneous tissues known from functional studies to express amiloride-sensitive Na' transport (Koeppen, 1986; Light et dl., 1988), immunostaining was carried out on semithin frozen sections of rabbit kidney outer medulla. Immunofluorescence was localized preferentially to apical membranes of cells in collecting tubules of inner and outer stripes (Fig. 8, A and B).The intensity of immunofluorescence was greater incollecting tubules in the inner stripe when compared with that of collecting tubules in the outerstripe. Occasional staining a t or near the basolateral plasma membrane was observed as well as a weak cytoplasmic fluorescence. There was no specific membrane staining associated with proximal straight tubules (thick descending limb) or thick ascending limbs. In controls, there was no specific staining associated with any of the medullary tubules when a monoclonal antibody against amiloride was substituted for RA6.3.

3913

goe, 1988). The sensitivity and specificity of this transport protein for amiloride and specific amiloride analogs allowed for generation of anti-Na' channel antibodies via an antiidiotypic approach. The antiidiotypic antibody RA6.3 appears to bind in proximity to the amiloride-binding site on polyclonal anti-amiloride antibodies and can mimic the effect of amiloride by inhibiting Na' transport across A6 cells (Fig. 2). Interestingly, pretreatment of apical cell surface proteins with trypsin (in thepresence of amiloride) was required to observe an inhibitory effect of RA6.3 on Na' transport. Thissuggests that RA6.3 has poor access to the amiloride-binding site on the native Na' channel and is in agreement with previous observations that the amiloride-binding site on the intact channel at thecell surface is within the channel's pore (Hamilton and Eaton,1985; Palmer, 1984) and therefore might not be accessible to thebinding site on the antiidiotypic antibody. The antibody RA6.3 recognized the Na' channel purified from A6 cells by Benos et al. (1987). As predicted, this antibody specifically bound the 140,000-Da subunit of the epithelial Na' channel (Fig. 3). This subunitwas shown with photoactive amiloride analogs to contain theamiloride-binding site of the Na' channel (Benos et al., 1987; Kleyman et al., 1989a). This monoclonal antibody is notspecies restricted and recognizes Na' channelin A6 cells orrabbit kidney medulla. The efficiency of binding to SDS-denatured Na+ channel is poor. Although the antibody recognizes the amiloride-binding subunit of the purified Na' channel (Fig. 3) on immunoblots, binding was not detected on immunoblots to Na' channels present in A6 cell extracts (data not shown). DISCUSSION Poor efficiency of antibody binding to SDS-denatured Na' Na' transport mediated by the epithelial Na' channel is channel is predicted from previous observations that radioinhibited by nanomolar concentrations of the diuretic amil- labeled amiloride analogs do not bind Na' channels solubioride and specific analogs of amiloride which have hydropho- lized with SDS (Kleyman et al., 1986b). bic substitutents on itsguanidino moiety (Kleyman and CraThe antiidiotypicantibody RA6.3 was used to examine biochemical characteristics of Na' channels in A6 cells that were either metabolically labeled or radiolabeled at the cell surface by enzyme-catalyzed radioiodination. We observed four to five subunits, with apparent molecular masses of 260,000-230,000,180,000,140,000-110,000, and 70,000 Da. The subunits labeled with [""Slmethionine were also identified at the apical cell surface using '''1 cell surface labeling (Figs. 5 and 6). Variation in the apparent molecular masses of subunits in the range of 110,000-180,000 Da was noted, with occasionally more than two subunits present. This variability may reflect post-translational modification of one or more subunits. Although we have attempted tolimit proteolysis by use of protease inhibitors, the possibility that variability in apparent molecular masses is a result of proteolysis cannot be excluded. In biosynthetic labeling studies, the antibody did not recognize polypeptides after labeling for 15 min with ["Slmethionine. Between 15 and 75 min (60-min chase) was required for efficient recognition of the Na' channel complex by the antibody (Fig. 5). These data suggest that the antibody does not recognize newly synthesized Na' channels. Rather, posttranslational modification or assembly of the Na+ channel complex is a prerequisite for antibody binding. Alternatively, FIG.8. Immunolocalization of Na+ channels in semithin froheat-shock proteins such as BiP (Flynn et aZ., 1989) might zen sections of outer medulla of rabbit kidney with RA6.3 (ascites fluid was diluted 1:50). Immunofluorescence staining is associate with polypeptide components of the Na' channel localized primarily to the apicalmembranes of principal cells of prior to assembly and interfere with antibody binding to the collecting tubules (C)in inner stripe (Fig. 8A) and outer stripe (Fig. channel. 8B).The intensity of staining was greater in the inner stripe. Cells The stoichiometry of subunits required for assembly of the in collecting tubules with unreactive apical surfaces are likely inter- Na' channel is difficult to estimatefrom these studies. Based calated cells (arrowheads). Plasma membranes of proximal straight tubules (descending thick limbs) (P)and ascending thicklimbs (D) on present andprevious studies, the apparentmolecular mass are unstained, except for some diffuse cytoplasmic fluorescence. Bar, of the Na' channel complex is approximately 750,000-700,000 10 pm. Da (Fig. 4; Benos et al., 1987). Assuming a 1:l stoichiometry

3914

Epithelial Na+ Channel

of subunits identified in this study, the apparent molecular mass of the channel complex would be 650,000-620,000 Da. The difference between these values may simply reflect anomalous migration under differing electrophoretic conditions, that being reducing uersus nonreducing gels, or may reflect the presence of additional phospholipids. Alternatively, the stoichiometry may not be 1:l. Different subunits might be utilized to generate epithelial Na' channels with distinct transportcharacteristics as has been reported in A6 cells (Hamilton and Eaton,1986). Two groups have described previously biochemical characteristics of Na' channels isolated from urinary epithelia. The results of the present study arein close agreement with those of Benos et al. (1987), whoalso reported that theNa' channel isolated from A6 cells or bovine kidney papilla consists of a high molecular mass (730,000-Da) complex. Benos and coworkers identified channel subunits with apparent molecular masses of 320,000-300,000, 186,000-150,000, 110,000-95,000, 85,000-70,000, and 55,000 Da. Apparent molecular masses of subunits identified with the antiidiotypic antibody, in particularthe 260,000-230,000-,180,000-,140,000-110,000and 70,000-Da subunits, are inreasonable agreement with data of Benos et al. (1987), especially in view of the inherent variability of subunit size noted previously by these authors(Benos et al., 1987; Sorscher et al., 1988). We have also noted a 50,000-Da polypeptide, although this polypeptide was not present consistently. In contrast, Barbry et al. (1987, 1990) reported isolation and characterization of an epithelial Na+ channel from pig kidney cortex. This channel consists of a postulated 185,000-Da protein composed of disulfide-linked 90,000-105,000-Da polypeptides. This structure differs from the oligomeric complex noted in both the present study and that of Benos et al. (1987). This discrepancy remains to be resolved. It is possible that thedifferences simply reflect Na+ channel isoforms isolated from different cell types. The Na+ channel is functionally restricted to the apical plasma membrane of A6 cells. Plasma membrane localization of the Na+ channel in A6 cell monolayers was examined by radioiodination of either apical or basolateral plasma membraneproteinsand subsequent immunoprecipitation. The Na+ channel complex was restricted to the apical plasma membrane (Fig. 6). Thesedata also suggest that there is minimal antibody cross-reactivity with other amiloride-sensitive transport proteinslikely present on basolateral plasma membranes. These include an amiloride-sensitive Na+/H+ exchanger for regulation of intracellular pH (Seifter and Aronson, 1986), Na+/Ca2+exchanger (Chase and Al-Awqati, 1981), and Na'/K+-ATPase (Verrey et al., 1989). Previous studies examining the effect of different amiloride analogs on these transport proteins suggest that the sites on amiloride recognized by the Na+/H+ exchanger, Na+/Ca2+exchanger, and Na'/K+-ATPase are probably distinct from the site on amiloride recognized by the Na' channel (Kleyman and Cragoe, 1988), suggesting that the likelihood of antibody crossreactivity with amiloride-sensitive transport proteins (other than the Na' channel) was low. Na' channel localization was also examined by immunostaining sections of A6 cell monolayers and of outer medulla of rabbit kidney. Three observations emerge from the immunofluorescent staining pattern in A6 cells. First, the apical membrane was clearly immunoreactive (Fig. 7), which is consistent with the dataobtained from surface-labeling studies (Fig. 6) and correlates with physiological data showing the presence of Na' channels in thismembrane domain (Perkins and Handler, 1981).Second, immunofluorescencewas present in subapical cytoplasm. This distribution is consistent with

the presence of a cytosolic Na+ channel pool, possibly vesicular in nature. Evidence exists for the presence of a pool of intracellular Na' channels which is recruited to the apical plasma membrane in a regulated manner (Lewis and de Moura, 1982; Garty and Edelman, 1983; Asher and Garty, 1988). Third, immunoreactivity was seen in deeper regions of the cytoplasm where it was frequently observed a t or near the basolateral cell border (Fig. 7B). The presence of immunoreactivity at thebasolateral membrane itself cannot be ruled out at this level of resolution. The observation that iodinated basolateral membrane proteins were not immunoprecipitated with the antiidiotypic antibody (Fig. 6) suggests that if the Na+ channelis expressed at thebasolateral plasma membrane, it is either expressed at low levels relative to its expression a t the apical plasma membrane, is expressed in a manner such that its tyrosine residues are not readily accessible for iodination, or that the protein is modified in amanner that confers a lowered affinity for the antibody following detergent solubilization. Na' channel expression in basolateral plasma membrane of a high resistance, Na+-transporting epithelium has been reported recently although its IC5ofor amiloride is 13 PM, almost 2 orders of magnitude higher than the apical plasma membrane Na+channel(Garty et al., 1987). This suggests that this protein might differ structurally from the apical plasma membrane Na+ channel. Distribution of Na+ channelslocalized with RA6.3 in proximity to the lateral plasma membrane may represent Na' channel-containing vesicles associated with the microtubular network. If microtubular bundles are organized along the apical-basal axis in proximity to lateral plasma membrane in A6 cells, as have been described in Madin-Darby canine kidney cells (Bacallao et al., 1989), the localization noted in deeper regions of the A6 cell cytoplasm may represent channels near the lateral cell border. A nonpolarized intracellular distribution of Na+/K+-ATPase in A6 cells was reported recently although the Na'/K'-ATPase was expressed at the cell surface in a polarized fashion (Verrey et al., 1989). High resolution immunocytochemical studies arerequired to delineate whether these transport proteins (e.g. Na' channel and Na'/K+-ATPase) are presentin distinct populations of intracellular vesicles or whether they reside.in common a intravesicular pool. Na+ channelswere localizedpreviously in A6 cells (Tousson et al., 1989) and in rat kidney outer medulla (Brown et al., 1989) by immunofluorescence using polyclonal antibodies raised against Na' channel purified from bovine kidney papilla (Sorscher et al., 1988). The immunolocalization to apical plasma membrane of principal cells of collecting tubules in the rat kidney is similar to thatobtained in the present study in rabbit medulla using RA6.3 (Fig. 8) and is consistent with the expected distribution ofNa' channels in this heterogeneous tissue based on functional studies (Koeppen, 1986; Light et al., 1988). The absence of immunoreactive sites along basolateral and apical plasma membranes in proximal straight tubules (descending thick limbs) suggests that these antibodies do not cross-react with the amiloride-sensitive Na+/H+ exchanger (Seifter and Aronson, 1986). The localization in A6 cell monolayers observed with RA6.3 differed from that reported by Tousson et al. (1989) in that the immunofluorescent stainingwith the anti-bovine Na' channel antibody was restricted to theapical plasma membrane in cryostat sections of intact A6 cell monolayers. Staining associated with the basolateral plasma membrane was, however,noted in A6 cells scraped off semipermeable supports although this was attributed to redistribution of apical Na' channels. Tousson et al. (1989) did not observe cytoplasmic localization of Na+ chan-

Epithelial Na+ Channel nels. This apparentdiscrepancy in results regarding intracellular immunoreactivity might reflect a limited access of antibodies to intracellular sites with use of conventional cryostat sections (e.g 5-10 Frn thick). In the present study, semithin frozen sections were used to observe the staining pattern shown in Fig. 7. We have observed recently intracellular localization in semithin sectionsof A6 cell monolayers with a polyclonal antibody (Sorscher et al., 1988) to purified bovine Na+ channel.' In the presentstudy, RA6.3 was used to immunoprecipitate Na' channel from A6 cells grown on either permeable or nonpermeable supports, and the channel complex was identified at the apical plasma membrane in cells grown under both conditions (Fig. 6). Although Sariban-Sohraby et al. (1983) initially suggested that thechannel is not expressed in cells grown ona nonpermeable support, functional apical Na+ channels have been identified in A6 cells grown under these conditions (Hamilton and Eaton, 1985, 1986). Electrophysiologic characteristics of cells grown on permeable supports differed from cells on nonpermeable support. Channels with high Na' to K+ selectivity and low conductance were expressed predominantly in cells grown on permeable supports whereas channels expressing poor Na+ to K+ selectivity and high conductance were expressed predominantly in cells grownon nonpermeable supports. Both types of channels were inhibited by amiloride (Hamilton and Eaton,1986; Ling and Eaton, 1989). These data suggest that RA6.3 might recognize channels with differing electrophysiologic properties and that these channels have similar biochemical characteristics. Acknowledgments-We are grateful to Eugenia Rauccio-Farinon, James Foster, James Schreiber, and Joseph Zebrowitz for excellent technical assistance, and to Dr. Dale Benos for immunoblots of purified Na' channel protein. We thank Drs. Bernard Erlanger and Qais Al-Awqatiforadvice and support, Dr. Mortimer Civan for critically reviewing the manuscript, and JamesSchreiber for valuable discussions. REFERENCES Asher, C., and Garty, H. (1988) Proc. Natl. Acad. Sci. U. S. A,. 85, 7413-7417 Bacallao, R., Antony, C., Dotti, C., Karsenti, E., Stelzer, E. H. K., and Simons, K. (1989) J . Cell Biol. 1 0 9 , 2817-2832 Barbry, P., Chassande, O., Vigne, P., Frelin, C., Ellory, C., Cragoe, E. J., Jr., and Lazdunski, M. (1987) Proc. Natl. Acad. Sci. U. S. A. 84,4836-4840 Barbry, P., Chassande, O., Marsault, R., Lazdunski, M., and Frelin, C. (1990) Biochemistry 29, 1039-1045 Benos, D. J., Saccomani, G., and Sariban-Sohraby, S. (1987) J. Biol. Chem. 2 6 2 , 10613-10618 Brown, D., Sorscher, E. J., Ausiello, D. A., and Benos, D. J. (1989) Am. J. Physiol. 256, F366-F369 Cayanis, E., Rajagopalan, R., Cleveland, W. L., Edelman, I. S., and Erlanger, B. F. (1986). J . Biol. Chem. 2 6 1 , 5094-5103

* T. R. Kleyman, J.-P. Kraehenbuhl, and S. A. Ernst, unpublished observations.

3915

Chase, H. S., and Al-Awqati, Q. (1981) J. Gen. Physiol. 77, 693-712 Cleveland, W. L., Wassermann, N. H., Sarangarajan, R., Penn, A. S., and Erlanger, B. F. (1983) Nature 305,56-57 Cuthbert, A. W., and Fanelli, G.M. (1978) Br. J. Pharmacol. 63, 139-149 Erlanger, B.F., Cleveland, W. L., Wassermann, N. H., Ku, H. H., Hill, B.L., Sarangarajan, R., Rajagopalan, R., Cayanis, E., Edelman, I. S., and Penn, A. S. (1986) Zmmunol. Reu. 94,23-37 Flynn, G. C., Chappell, T. G., and Rothman, J. E.(1989) Science 245,385-390 Garty, H., and Benos, D. J. (1988) Physiol. Reu. 6 8 , 309-373 Garty, H., and Edelman, I. S. (1983) J. Gen. Physiol. 81,785-804 Garty, H., Warncke, J., and Lindemann, B. (1987) J. Membr. Biol. 95,91-103 Hamilton, K. L., and Eaton,D. C. (1985)Am. J. Physiol. 249, C200C207 Hamilton, K. L., and Eaton, D. C. (1986) Fed. Proc. 45, 2713-2717 Hill, B. L., and Erlanger, B. F. (1988) Endocrinology 122,2840-2850 Homcy, C. J., Rockson, S. G., and Haber, E. (1982) J . Clin. Znuest. 69,1147-1154 Kleyman, T. R., and Cragoe, E. J., Jr. (1988) J. Membr. Biol. 105, 1-21 Kleyman, T. R., Rajagopalan, R., Cragoe, E. J., Jr., Erlanger, B. F., and Al-Awqati, Q. (1986a) Am. J. Physiol. 2 5 0 , C165-Cl70 Kleyman, T. R., Yulo, T., Cragoe, E., Jr., and Al-Awqati, Q. (1986b) Kidney Znt. 2 9 , 400 (abstr.) Kleyman, T. R., Cragoe, E. J., Jr., and Kraehenbuhl, J. P. (1989a) J. Biol. Chem. 264, 11995-12000 Kleyman, T. R., Kraehenbuhl, J. P., Rossier, B. C., Cragoe, E. J., Jr., and Warnock, D. (198913) Am. J. Physwl. 257, C1135-Cl141 Koeppen, B. M. (1986) Am. J. Physiol. 2 5 0 , F70-F76 Ku, H. H., Cleveland, W. L., and Erlanger, B. F. (1987) J . Zmmunol. 1 3 9 , 2376-2384 Lewis, S. A., and de Moura, J. (1982) Nature 297, 685-688 Light, D. B., McCann, F. V., Keller, T. M., and Stanton,B. A. (1988) Am. J. Physiol. 255, F278-F286 Ling, B. N., and Eaton, D. C. (1989) Am. J. Physiol. 256, F1094F1103 Maizel, J. V. (1971) Methods Virol. 5, 179-246 Palmer, L. G. (1984) J. Membr. Biot. 80, 153-165 Perkins, F. M., and Handler, J. S. (1981)Am. J. Physiol. 2 4 1 , C154C159 Sariban-Sohraby, S., and Benos, D. J. (1986) Am. J. Physiol. 250, C175-C190 Sariban-Sohraby, S., Burg, M. B., and Turner, R. J. (1983) Am. J . Physiol. 245, C167-Cl71 Sege, K., and Peterson, P. (1978) Proc. Natl. Acad. Sci. U. S. A. 7 5 , 2443-2447 Seifter, J. L., and Aronson, P. S. (1986) J. Clin. Znuest. 78,859-864 Sorscher, E., Accavitti, M. A., Keeton, D., Steadman, E., Frizzell, R., and Benos, D. J. (1988) Am. J. Physiol. 255, C835-C843 Tokuyasu, K. T., and Singer, S. J. (1976) J . Cell Biol. 7 1 , 894-906 Tousson, A., Alley, C. D., Sorscher, E. J., Brinkley, B. R., and Benos, D. J. (1989) J . Cell Sci. 93, 349-362 Towbin, H., Staehelin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci. U. S. A. 76, 4350-4354 Verrey, F., Kairouz, P., Schaerer, E., Fuentes, P., Geering, K., Rossier, B. C., and Kraehenbuhl, J . P. (1989) Am. J. Physiol. 256, F1034F1043 Wassermann, N. H., Penn, A. S., Freimuth, P. I., Treptow, N., Wentzel, S., Cleveland, W.L., and Erlanger, B. F. (1982) Proc. Natl. Acad. Sci. U. S. A. 79,4810-4814 Zhuang, Y., Cragoe, E. J., Jr., Shaikewitz, T., Glaser, L., and Cassel, D. (1984) Biochemistry 23, 4481-4488