Q 1993 by The American Society for Biwhemiatry and Molecular Biology, Inc. THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 268, No. 25, Issue of September 5, pp. 19070-19075, 1993

Printed in U.S.A.

Mechanism of Inhibition of CAMP-dependent Epithelial Chloride Secretion by Phorbol Esters*

(Received for publication, March 17, 1993, and in revised form, May 6, 1993)

Ben-Quan ShenS, Roger A. BarthelsonS, William SkachQ, Dieter C. Gruenertl, Elliott SigalQ, Randall J. Mrsnyq, and Jonathan H. Widdicombe+II** From the $Cystic Fibrosis Research Center and the Cardiovascular Research Institute and the Departments of (1 Physiology and $Medicine, University of California, San Francisco, California 94143 and llGenentech Znc., South San Francisco, California 96080

In Tar cells, we investigated how stimulation of pro- tein kinase C leads to an inhibition of CAMP-dependent chloride secretion. Specifically, we tested the hypoth- esis that the inhibition was caused by loss of the cystic fibrosis transmembrane regulator (CFTR), an apical membrane chloride channel. As described by others (Trapnell, B. C., Zeitlin, P. L., Chu, C.-S., Yoshimura, K., Nakamura, H., Guggino, W. B., Bmgon, J., Banks, T. C., Dalemans, W., Pavirani, A., Lecocq, J.-P., and Crystal, R. G. (1991) J. Biol. Chern. 266, 10319- 10323), we found that treatment with the phorbol ester, phorbol myristate acetate (PMA), reduced CFTR mRNA levels by -80% with a tlIz of -2 h. Chloride secretion, measured as forskolin-induced short circuit current, was also abolished by PMA with a tllz of -2 h. Levels of mature glycosylated CFTR measured by Western blotting also declined to 60 f 8% (n = 7) of control after a 12-h PMA treatment. However, a 12-h exposure to PMA did not affect the forskolin-stimu- lated efflux of '''1 into high potassium medium, a meas- ure of apical membrane CFTR activity. We conclude that increased turnover of apical membrane CFTR in PMA-treated cells compensates for the decline in anion channel numbers. By contrast to its lack of effect on la61 effluxes, PMA reduced the CAMP-induced increase in '"Rb efflux, suggesting that it inhibits chloride se- cretion mainly by an action on basolateral potassium channels.

Chloride secretion by vertebrate epithelia involves uptake of chloride across the basolateral membrane by cotransport with sodium and potassium, followed by its diffusional move- ment through anion channels in the apical membrane (1,2). Both calcium- and CAMP-activated anion channels have been identified (3, 4), and recent work strongly suggests that the protein that is defective in cystic fibrosis (the cystic fibrosis transmembrane conductance regulator or CFTR)' is a CAMP- activated apical membrane chloride channel (5-8).

* This work was supported by National Institutes of Health Spe- cialized Center for Research Grant HL-42368. The costs of publica- tion of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertise- ment'' in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

**To whom correspondence and reprint requests should be ad- dressed Cardiovascular Research Institute, 513 Parnassus Ave., Box 0130, San Francisco, CA 94143, Tel.: 415-476-1296; Fax: 415-476- 9986.

The abbreviations used are: CFTR, cystic fibrosis transmembrane conductance regulator; PMA, phorbol myristate acetate; BSA, bovine serum albumin; cpt-CAMP, 8-(4-~hlorphenylthio)-cAMP; Ab, anti- body; TBS, Tris-buffered saline; Zw, short-circuit current.

In several epithelia (9-ll), stimulation of protein kinase C by phorbol esters inhibits CAMP-dependent chloride secre- tion. In TM cells, a human colon carcinoma cell line, this inhibition is associated with reduction in the levels of mRNA for CFTR (11). Thus, following the addition of 40 nM phorbol myristate acetate (PMA), CFTR mRNA levels were -80% of control at 1 h, fell to -10% of control after 4-12 h, and recovered to 40 and 60%, respectively, after 24 and 48 h. Thus the t l /z for the reduction of CFTR mRNA by PMA is -2 h. In the same study, increases in 36Cl efflux in response to forskolin were measured at 1-, 24-, and 48-h exposures to 40 nM PMA or 24- and 48-h exposures to 100 nM PMA. There was a rough correlation between levels of CFTR mRNA and forskolin-induced 36Cl efflux for the different exposures. With 40 nM PMA, the forskolin-stimulated efflux was significantly less than that of control tissues at 24 h, but not at 1 or 48 h. From these results Trapnell et al. (11) suggested that the PMA-induced decline in mRNA led to reductions in CFTR which accounted for the inhibition of CAMP-dependent chlo- ride efflux. For this hypothesis to be plausible, CFTR would have to have a relatively rapid turnover with a tlI2 in the range of hours.

The purpose of the present study was to test directly the hypothesis that the PMA-induced loss of chloride transport is caused by a decline in CFTR levels. Accordingly, we have compared the effects of PMA on CFTR mRNA, CFTR pro- tein, CAMP-dependent chloride secretion, and CAMP-depend- ent apical membrane anion permeability. Despite a reduction in levels of mature CFTR protein, our results suggest that the main inhibitory action of PMA on chloride secretion is via an action on basolateral potassium channels.

MATERIALS AND METHODS

Cell Culture-Tsr cells were obtained from the University of Cali-

mixture of Dulbecco's modified Eagle's medium and Ham's F-12 fornia-San Francisco Cell Culture Facility and grown in a 50:50

medium containing 5% fetal calf serum. They were maintained in T75

flasks (Corning Glass Works, Corning, NY) in a 6% Con, air incubator a t 37 "C and 98% relative humidity. When 80% confluent they were detached with saline-trypsin-versine and passaged to new flasks at a 1:5 dilution. Cells were plated at 106/cm2 in T76s for mRNA measure- ment, in 12-mm Millicell-HA inserts (Millipore, Bedford, MA) or 15- mm inserts (Costar, Cambridge, MA) for Ussing chamber studies (the two types of insert gave identical results), in 25-mm Millicell-HA inserts for lZ6I and %Rb efflux studies, and in Costar inserts for CFTR determination. Estimates of CFTR levels were made on tissues a t the end of Ussing chamber studies.

Ussing Chamber Studies-Standard techniques were used. About 1 week after plating, inserts were mounted in Ussing chambers and bathed in bicarbonate-buffered Krebs-Henseleit solution (pH 7.4), bubbled with 95% 02, 5% Con, at 37 "C. Transepithelial potential difference was clamped to zero, and the resulting short-circuit current (Zx) was continuously displayed on a pen recorder. Transepithelial resistance was determined from the size of the current deflections

19070

Protein Kinase C and Chloride Secretion 19071

caused by 0.2-5 voltage pulses of constant amplitude (0.2-1 mV) imposed on short-circuited cell sheets every 20 s. Drugs were added as 100-fold concentrated stock solutions to both sides of the tissue.

Radiotracer Eflures-Cells were loaded with isotope (l%I or %b) in normal physiological saline. However, to minimize changes in rate of efflux caused by changes in membrane potential, effluxes were performed in high potassium medium. Under these conditions, apical and basolateral membrane potentials should both be close to zero. Furthermore, as the membrane potentials will be close to the equilib- rium potential for potassium, and as the basolateral potassium con- ductance should predominate under thew conditions, changes in either potassium or chloride conductance should have little effect on apical or basolateral membrane potentials. Thus the membrane po- tentials should be clamped at -0 mV, and alterations in efflux will reflect changes in permeability rather than changes in membrane potential. To avoid large changes in membrane potential at the start of the efflux, the cells were transferred to high potassium medium for the final 10 min of loading with isotope.

The porous bottoms of the inserta with their attached confluent cells were removed from 25-mm HA inserts and placed in serum-free culture medium containing Na'=I (10 pCi/ml; 3 nM) or BBRbCl (10 pcifml, 85 pm) for 2 h. Filters and attached cells were then placed for a further 10 min in high potassium medium (in mM: 120 KCl, 25 KHepes, 1.2 KHzP04, 1.2 MgSO4, 2.5 CaCl2, 5.9 glucose) containing the same levels of isotope. Tissues were then rinsed for 15 s in 500 ml of high potassium medium to remove extracellular counts and placed in 2 ml of high potassium medium in a scintillation vial at 37 "C. At 1-min intervals the medium in the vial was removed and replaced with 2 ml of fresh, prewarmed (37 'C) and oxygenated (100% 02) solution. A standard protocol was followed. After the first five efflux samples had been taken, the fresh medium added to the cells contained a CAMP-elevating cocktail consisting of 200 pm 8-(4- chlorpheny1thio)-CAMP (cpt-CAMP), 10 p~ forskolin, and 1 mM 3- isobutyl-1-methylxanthine. Following five efflux samples containing this cocktail, the tissues were returned to drug-free medium for a further five samples. The efflux samples and the tissue itself were then counted in a y- ('%I) or @-counter (@Rb). The total counts in the tissue at each sample point were calculated, and the efflux of tracer was expressed as a fractional rate of loss (i.e. total tissue counts lostfmin of sampling period divided by the average total counts present in the tissue during the period).

Measurement of CFTR-Peptides corresponding to regions of the N-terminal (amino acids 45-65), the C-terminal (amino acids 1458- 1479) or the R-domain (amino acids 680-700) of CFTR (12) were synthesized on an Applied Biosystems model 430 using tertbutylox- ycarbonyl amino acids and phenylacetamidomethyl resin. The pep- tides were conjugated to keyhole limpet hemocyanin with m-maleim- idobenzoyl-N-hydroxysulfosuccinimide ester (13) and were used to generate rabbit antisera (Berkeley Antibody Co., Richmond, CA). The antibodies (Ab-A to residues 45-65, Ab-C to residues 680-700, and Ab-E to residues 1458-1479) were purified by ammonium sulfate precipitation and affinity chromatography on Sepharose G (Phar- macia LKB Biotechnology Inc.) and stored at -70 "C at -5 mg/ml. Antibodies a-102 and a-1468 to the extracellular and C-terminal domains of CFTR, respectively, were obtained from Jonathan Cohn at Duke University and prepared as described previously (14, 15).

The specificity of antisera raised against the CFTR peptide was tested in one of two ways. First, we assessed the ability of antibody to precipitate CFTR generated in a cell-free translation system. Second, CFTR was immunoprecipitated and phosphorylated with [y- '*P]ATP using the catalytic subunit of protein kinase A, and phos- phopeptide maps were generated.

To prepare CFTR in cell-free lysates, CFTR cDNA, engineered behind the SPS promoter, was used to generate mRNA and full-length radiolabeled CFTR in a transcription-linked translation rabbit retic- ulocyte system supplemented with canine pancreas microsomes and

pl of T x S W (0.1 M Tris-HC1, 0.1 M NaCl, 2 mM NazEDTA, 1% [%]methionine. One pl of translation mixture was solubilized in 400

Triton X-100 (pH 8.0)) containing 5 pl pf protein A-Affi-Gel beads (Affi-Gel-A; Bio-Rad) and 1 pl of either antiserum, nonimmune serum, or antiserum plus excess peptide against which the antiserum was raised. Following incubation for 6 h at 40 "C, beads were washed three times with TxSWB and twice with 0.1 M NaCl, 0.1 M Tris-HCl (pH 8.0). Translation products and immunoprecipitated material were analyzed by electrophoresis on 6% SDS-polyacrylamide gels and autoradiography.

To generate phosphopeptide maps, CFTR was first immunoprecip- itated as follows. Confluent cell sheets were rinsed twice with ice-cold

phosphate-buffered saline. Ice-cold buffer A (1% Triton X-100, 100 mM NaC1,lOO mM Tris-HC1, 10 mM NaBDTA, 1 mM phenylmeth- anesulfonyl fluoride (pH 8.0)) was then added (-1 ml/cm2 of cell surface). After a 30-min incubation on ice, the cells were scraped and collected. Cell debris was dispersed by repeated pipetting and centri- fuged at 9,000 X g for 2-4 min at 4 "C in a Beckman 12 Microfuge. The protein concentration of the supernatant was measured by the BCA assay (16). To eliminate cellular components that bind nonspe- cifically to protein A, the lysate (800 pg protein in 600 pl of buffer A) was incubated with 50 pl of Affi-Gel-A and 10 pl of rabbit serum/800 pg of protein at 4 "C for 1 h of gentle agitation. Affi-Gel-A was equilibrated before use by three washings in buffer A. The Affi-Gel- A was removed by centrifugation for 1 min in a Microfuge, and the supernatant was stored on ice until further use. Antibody (5 pl) and Affi-Gel-A (5 pl) were added to each 600-pl sample and incubated for 4-5 h at 4 'C with gentle agitation. The Affi-Gel-A pellet was collected by centrifugation in a Microfuge for 30 s and washed three times with buffer A, two times with buffer B (100 mM NaCl, 100 mM Tris-HC1 (pH 8.0)) and once with buffer C (10 mM MgClz, 100 pg/ml bovine serum albumin (BSA), 50 mM Tris-HC1 (pH 7.5)). All buffers were ice-cold. The pellet was then resuspended in 50 pl of phosphorylation buffer (50 mM Tris-HC1, 10 mM M&L, 100 pg/ml BSA (pH 7.5)). For the phosphorylation reaction, 10 pl of catalytic subunit of protein kinase (5 unitslpl; Sigma) in 50 mg/ml dithiothreitol and 6 pl [y-"PI ATP (30 Ci/mmol, 2 mCi/ml; DuPont-New England Nuclear) were added to the solution and incubated at 30 "C for 45 min. The reaction was stopped with 500 pl of ice-cold buffer A. The pellet was washed twice with buffer A, suspended in SDS-sample buffer, run on a 6% SDS-polyacrylamide gel, and labeled protein was then detected by autoradiography. Phosphopeptide maps of 32P-labeled protein were then generated as described previously (17). In brief, gel pieces containing labeled proteins were lyophilized and digested with tryp- sin. Phosphopeptides were then separated by thin-layer two-dimen- sional chromatography and detected by autoradiography.

For quantitation of CFTR, we used Western blots. Cells were lysed as described above, and 100 pl of the lysate (2 mg/ml) was mixed with an equal volume of 2 X SDS-sample buffer. Following incubation at 30 'C for 10 min, the proteins were separated on an 8% SDS gel. The gel and nitrocellulose membrane (Schleicher & Schuell) were first incubated in transfer buffer (25 mM Tris-HC1, 150 mM glycine, 0.05% SDS, 10% methanol (pH 8.3)) for 30 min at room temperature, then were sandwiched between filter papers (Bio-Rad) prewetted in transfer buffer (two each side). Proteins were transferred (100 V; 2 h) with a Transblot cell (Bio-Rad) a t 4 "C. After transfer, the mem- brane was taken off (gel side face up) and rinsed once with Tris- buffered saline (TBS; 140 mM NaC1,30 mM Tris-HC1 (pH 8.0)) and then incubated in blocking buffer (1% nonfat dried milk, 1% BSA, 1% polyvinylpyrrolidone 10, 10 mM NazEDTA in TBS) either over- night at 4 "C or a t room temperature for 4 h. Following two 5-min washes (TBS + 0.05% Tween 201, the membrane was washed twice in wash buffer (1% nonfat milk, 0.5% BSA, 0.05% Tween 20 in TBS). Then the membrane was incubated in primary antibody (1:1,OOO) in incubation buffer (1% nonfat milk, 5% BSA, 0.05% Tween 20 in TBS) at room temperature for 2 h. The membrane was washed four times for 5 min each in wash buffer and then incubated with horse- radish peroxidase-conjugated secondary antibody (1:2,000) in incu- bation buffer a t room temperature for 1.5 h. Then the membrane was washed three times for 5 min each in wash buffer and twice for 5 min in TBS + 0.05% Tween 20. CFTR was detected with an enhanced chemiluminescence kit (Amersham Corp.). Levels of CFTR were quantified with a Gilford multimedia densitometer (Ciba-Corning, Oberlin, OH).

Digestion with N-glycanase (Genzyme, Cambridge, MA) or binding to Sambucus nigra agglutinin agarose beads (SNA beads; E-Y Labo- ratories, San Mateo, CA) was used to determine if the CFTR detected in Western blots was the mature glycosylated form. For digestion of CFTR by glycanase, 5% SDS (67 pl) was added to lysate (200 pg in 100 pl of lysis buffer), and the mixture was incubated at room temperature for 15 min. Ten volumes of digestion buffer (50 mM Tris-HC1, 25 mM NazEDTA, 1% Triton X-100, 1% 2-mercaptoetha- no1 (pH 8.0)) was then added to the mixture. N-Glycanase (2.5 units in 10 pl) was added, and the mixture was incubated at 37 "C for 20 h. Cold ethanol (4 volumes at -80 "C for 2 h) was used to precipitate protein, which was pelleted in a Beckman Microfuge for 15 min. The pellets were Suspended in SDS-sample buffer and CFTR detected by Western blotting. Control samples were treated exactly as described above, except glycanase was not added. We studied binding to lectin- agarose beads by an established method (18). Lysate (200 p1) was

19072 Protein Kinase C and Chloride Secretion

incubated with 100 p1 SNA beads for 2 h a t 4 "C. The beads were pelleted and bound proteins eluted with 0.15 M lactose in lysis buffer. Samples of the supernatant and eluate were run on SDS-polyacryl- amide gels, and CFTR was determined by Western blotting.

Northern Blot Hybridization-RNA was isolated by a protocol adapted from Laski et al. (19). Cells were washed twice with ice-cold phosphate-buffered saline and treated with trypsin. Trypsin treat- ment was stopped with an equal volume of medium containing 10% fetal calf serum, and the single cell suspension was centrifuged (1,000 X g, 10 min). The pellet was washed twice with ice-cold phosphate- buffered saline. Cells were resuspended in lysis buffer (0.65% Nonidet P-40,lO mM Tris-HC1,150 mM NaCl, 1.5 mM M&12 (pH 7.8)) on ice for 10 min and centrifuged at 1,500 X g for 5 min. The cytoplasmic supernatant was then separated from the nuclear pellet. The super- natant, containing RNA, was diluted with an equal volume of urea buffer (7 M urea, 10 mM Tris-HC1,lO mM Na2EDTA, 350 mM NaCl, 1% SDS (pH 7.5)). The solution was extracted one time with an equal volume of phenol. The RNA was precipitated with 2.5-3 volumes of 100% ethanol a t -70 "C for 3-4 h. After centrifugation for 30 min at 15,000 X g the pellet was washed once with 70% ethanol, dried, and resuspended in water treated overnight with 0.5% diethyl pyrocar- bonate and autoclaved. The absorbance measurements a t 260 nm were used to determine the level of RNA in each sample, and samples were then stored at -70 "C for further use.

For electrophoresis, 20 pg of total RNA from each sample was lyophilized and resuspended in 3.7 pl of diethyl pyrocarbonate-treated water. Following the addition of 2.7 pl of 40% aqueous solution of glyoxal, 8 pl of dimethyl sulfoxide, and 1.6 pl of 100 mM sodium phosphate (pH 6.5), the sample was heated at 50 "C for 15 min. The RNA was separated on a 1.2% agarose gel in 10 mM sodium phosphate (pH 6.5) (20). The gel was run at 50 V for 4 h, soaked in 50 mM NaOH for 20 min, stained in 100 mM Tris-HC1 (pH 7.5) containing 0.5 pg/ml ethidium bromide for 10 min, and finally washed in 25 mM sodium phosphate (pH 6.5) for 15 min. To check the integrity of the mRNA, the mRNA-ethidium bromide complexes were visualized un- der ultraviolet light. RNA was transferred onto a GeneScreen Plus membrane (Du Pont-New England Nuclear) by capillary blot. Hy- bridization at 42 "C using formamide was carried out as recommended by Du Pont-New England Nuclear protocols. Blots were hybridized with both CFTR and y-actin probes and analyzed autoradiographi- cally. The probe for y-actin was obtained from Sigma. The CFTR probe was a 4.7-kilobase fragment of full-length CFTR cDNA ob- tained by agarose purification of a PstI-digested fragment from plas- mid pBQ4.7, obtained from the American Type Culture Collection (Rockville, MD). The probes were labeled by random priming with a multiprime DNA labeling kit (Amersham Corp.). The labeled probe was separated from the radiolabeled bases by passage through an exclusion column (Minispin; Worthington).

RESULTS

We confirmed the results of Trapnell et al. (11) that PMA reduces levels of the mRNA for CFTR (see Fig. 1). With continuous exposure to 100 nM PMA, the mRNA for CFTR declined to -50% of control after -2 h (see Fig. 2) and reached a minimum value of -15% of control at 4 h. At longer exposure times there was some slight recovery of the mRNA levels. Again in agreement with Trapnell et dl. (ll), we found that

CFTR 7-act in - * - +

- r m

6.5* A kb

FIG. 1. Levels of transcripts for CFTR (6.5 kilobases (Kb)) and y-actin (2.3 kilobases) determined in control cells (-) or cells treated with 100 nM PMA for 8 h (+). RNA (20 pgllane) was separated on an agarose gel and transferred by capillary blot to a GeneScreen Plus membrane. Blots were hybridized to 32P-labeled cDNA probes and analyzed by autoradiography. Results are repre- sentative of several similar experiments.

- 0 10 20 30 4 0

TIME (hr)

FIG. 2. Time course of action of PMA (100 nM) on mRNA for CFTR and y-actin. Plots are of peak heights from densitometric scans of autoradiographs from Northern blots. Results are represent- ative of three identical experiments.

0 0 6 12 18 24 40

TIME (hr)

FIG. 3. Effects of PMA on forskolin-induced increases in I,. The Z, response to forskolin is plotted against the duration of exposure to PMA. Results are expressed as a percent of the control response of tissues unexposed to PMA. Open circles, 40 nM PMA; closed circles, 100 nM PMA. Means f S.E., n = 5.

PMA caused small increases in the message for y-actin (Figs. 1 and 2).

Control transepithelial resistance of 156 f 32 il cm2 (n = 10) compared with 132 f 17 Q.cm2 (n = 5 ) and 138 f 17 il. cm2 (n = 5 ) after a 24-h exposure to 40 and 100 nM PMA, respectively, demonstrating that the transepithelial resistance was not affected by treatment with PMA. Fig. 3 shows the maximal increases in I, induced by forskolin after incubations with PMA (40 or 100 nM) of various durations. It is clear that PMA inhibits the forskolin-induced increases in I, with a t1I2 of -2 h, not noticeably different from the t1/2 for reduction in CFTR mRNA.

Of our three antibodies (Ab-A, Ab-C, Ab-E), only Ab-E was effective at immunoprecipitating CFTR. We compared the immunoprecipitate produced by Ab-E with that obtained with a-102. Three bands of -135, 145, and 165 kDa were brought down by a-102, with the band at -165 kDa being the most intense. These presumably correspond to the unglycosylated, core-glycosylated, and fully glycosylated forms of CFTR (21). Ab-E brought down the same three bands as a-102, although now the band at 165 kDa was very faint, and the predominant band was at -135 kDa. Thus we conclude that Ab-E pref- erentially immunoprecipitates the unglycosylated form of CFTR. Several lines of evidence suggested that the material immunoprecipitated by Ab-E was CFTR. First, incubation of Ab-E with 10 mg/ml of the peptide used to generate it blocked immunoprecipitation, although BSA or a 20-amino acid pep- tide corresponding to residues 1317-1336 of CFTR (both at

Protein Kinase C an

10 mg/ml) was without effect. Second, the major 35S-labeled protein produced by incubating CFTR mRNA with reticulo- cyte lysates had an apparent molecular mass of 140 kDa. This band was immunoprecipitated by Ab-E. Again, this immuno- precipitation was prevented by incubating the antibody in the presence of the peptide sequence used to generate it (data not shown). Third, the phosphopeptide map of the bands immu- noprecipitated by Ab-E was identical to that of the diffuse 170-kDa protein precipitated by a-1468 (see Fig. 4); a-1468 has earlier been shown to precipitate a protein whose phos- phopeptide map is identical to that of the CFTR R-domain (17).

In Western blots, Ab-A, Ab-C, and Ab-E all recognized a 165-kDa protein (see Fig. 5), which was not seen when blots were exposed to preimmune serum. In addition, Ab-A recog- nized a diffuse band at -210 kDa. However, it is unlikely that this is a form of CFTR, as this band was present at the same intensity in fibroblast lysates, although the 165-kDa band was absent (data not shown).

N-Glycanase treatment reduced the molecular mass of the protein detected by Ab-E in Western blots to -150 kDa (see Fig. 6), indicating that this antibody recognizes a glycosylated form of CFTR (21). Furthermore, when Western blots were performed on lysate that had been preincubated with SNA beads, the 165-kDa band was virtually eliminated from West- ern blots, whereas CFTR was readily detectable in the lactose eluate from the SNA beads. SNA is a lectin with a primary affinity for sialic acid linked to galactose, the carbohydrates

d Chloride Secretion 19073

116- " P - 84- D .

FIG. 6. Effects of glycanase on apparent molecular mass of CFTR. Lysate (200 pg of protein) was incubated (37 "C, 20 h) with or without N-glycanase (2.5 units in 10 pl). Protein was precipitated with ethanol and run on SDS-polyacrylamide gels. CFTR was de- tected by Western blotting. Left lune, control lysate receiving no glycanase. Right lane, lysate treated with glycanase.

lSOY

kDa 0 1 3 6 12 24

1161 84

FIG. 7. Levels of mature CFTR determined by Western blot- ting. Top panel shows Western blots from a typical experiment. Durations of exposure to PMA in hours are indicated above the lunes, each of which received 200 pg of lysate protein. Bottom panel shows densitometer scans of the band at 165 kDa for each lune.

aor r

9 FIG. 4. Peptide digests of immunoprecipitated proteins.

Proteins were immunoprecipitated with Ab-E and labeled with 32P04 using [y-S2P]ATP and the catalytic subunit of protein kinase A. Gel pieces containing labeled protein were lyophilized and digested with 1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin (50 pg/ml, 20 h at 37 "C) . Phosphopeptides were resolved by electropho- resis at 400 V for 90 min in 10% acetic acid, 10% pyridine (pH 3.5) followed by chromatography in 1510:123 (v/v) pyridine/l-butanol/ water/acetic acid on 20 X 20-cm cellulose sheets. They were then detected by autoradiography. Panel A , combined 135- and 145-kDa bands precipitated by Ab-E. Panel B, 155-170-kDa band precipitated by a-1468. Panel C, immunoprecipitates of Ab-E and a-1468 com- bined.

kDo 200-

A C E

dm

97-

69-

FIG. 5. Western blots with Ab-A, Ab-C, and Ab-E. Lysate (200 pg of proteinllane) was separated on 8% SDS-polyacrylamide gels. Proteins were transferred with a Transblot cell to a nitrocellulose membrane. Following incubation in blocking buffer containing 1% nonfat dried milk, primary and horseradish peroxidase-conjugated secondary antibodies were applied. Secondary antibody was detected by an enhanced chemiluminescence kit (Amersham Corp.). All anti- bodies recognized a band at -165 kDa.

0" 10 15 0 10 15

TIME (mfn) TIME (mln)

FIG. 8. Effects of PMA (100 nM, 12 h) on CAMP-induced '"1 efflux. Left panel , forskolin (10 p ~ ) , cpt-CAMP (200 p ~ ) , and 3-isobutyl-1-methylxanthine (1 mM) were applied as indicated by the horizontal bar. Right panel, time controls not receiving CAMP-elevat- ing cocktail. Solid symbols, control cells; open symbols, PMA-treated cells. Values are means & S.E., n = 3-5.

associated with mature glycosylation. Using Ab-E, we measured mature CFTR in lysates from

Tu cells exposed to PMA (100 nM) for varying lengths of time. A typical experiment is illustrated in Fig. 7. In seven similar experiments, CFTR levels were 75 f 9% of control at 1-h PMA exposure, 62 f 7% at 3 h, 57 f 9% at 6 h, 50 f 8% at 12 h, and 41 f 4% at 24 h. Values for levels of CFTR were significantly less than control ( p < 0.05, paired t test) for all exposure times.

Pretreatment with PMA for 12 h did not affect the CAMP- induced increase in '''1 efflux. Thus, as shown in Fig. 8, efflux increased immediately on addition of CAMP-elevating cock- tail, reaching maximal values after 2 min. At this time point, the increase in the rate of loss over the value immediately before the addition of cocktail was 11.30 f 2.29% .min" for control and 18.68 f 2.45% emin-' for PMA-treated cells. These values were not statistically different from one another as determined by paired t test. During the same time interval, the efflux of '"I declined slightly in cells that did not receive the CAMP-elevating cocktail (-1.61 f 0.50% .min", n = 3 for control; -1.59 f 1.16% .min", n = 3 for PMA-treated cells).

19074 Protein Kinase C and Chloride Secretion

The forskolin-induced increase in I, in the same cultures was inhibited 5040% by PMA.

In contrast to its lack of effect on CAMP-induced '''9 efflux, PNA markedly inhibited the CAMP-induced increase in =Rb efflux (see Fig. 9). Stimulation of =Rb efflux by cAMP was transient reaching a peak 2 min after the addition of cocktail. At this point =Rb efflux had increased by 2.93 k 0.39%. min" in control cells and significantly less ( p < 0.005, unpaired t test) in PMA-treated cells (1.08 -C 0.31% .min"). Over the same time interval =Rb efflux decreased by -0.26 f 0.54%. min" (n = 5 ) in control cells that did not receive the cocktail and by -0.07 f 0.19%.min" (n = 6) in PMA-treated time controls.

DISCUSSION

Here we have confirmed previous results (9-11) showing that stimulation of protein kinase C by PMA inhibits CAMP- activated chloride secretion across vertebrate epithelia. Also in agreement with others (ll), we found it to cause a rapid (tl,z -2 h) decline in CFTR mRNA.

It seems very probable that CFTR is the apical membrane chloride channel responsible for CAMP-induced chloride se- cretion across epithelia affected in cystic fibrosis. Thus, when expressed in nonepithelial cells, CFTR induces a 10-pico- Siemen, nonrectifying chloride channel which is opened by the catalytic subunit of CAMP-dependent protein kinase (5, 6, 8) and has electrical properties very similar to those of CAMP-activated chloride channels in the apical membrane of several epithelia (22, 23). Therefore, a plausible explanation for the PMA-induced reduction in CAMP-dependent chloride efflux described by Trapnell et al. (11) is that the decline in its mRNA leads to a loss of CFTR.

In these studies we have tested this hypothesis by measur- ing total CFTR levels by Western blotting. We found that a 12-h exposure to PMA reduced the levels of mature CFTR to 50 f 8% of control. However, CAMP-dependent apical mem- brane anion permeability, as measured by the CAMP-me- diated loss of into high potassium medium, was not altered. The most likely explanation for these findings is provided by

0-

lor T

I 1 I 0 S 10 1s

TIME (min)

FIG. 9. Effects of PMA (100 n ~ , 12 h) on CAMP-induced wRb efflux. Top panel , forskolin (IO p ~ ) , cpt-CAMP (200 p ~ d , and 3-isobutyl-l-methylxanthine (1 mM) were applied as indicated by the horizontal bar. Bottom p a n e l , time controls not receiving CAMP- elevating cocktail. Solid symbols, control cells; open symbols, PMA- treated cells. Values are means f S.E., n = 5-7.

the patch-clamp data of Tabcharani et al. (6). Using patches of membrane from CFTR-expressing Chinese hamster ovary cells, these authors showed that phosphorylation of CFTR by protein kinase C enhanced its subsequent opening in response to protein kinase A. Thus, loss of apical membrane CFTR is compensated for by increased responsiveness of CFTR to protein kinase A. An alternative possibility, that our assay is not an accurate measure of CFTR function, can be ruled out on several grounds. First, iodide is not transported by the NaK2Cl cotransporter in the basolateral membrane (24), so its only route of exit from the cells is via apical membrane anion channels. Second, confluent cell sheets of Tsc cells, grown on filters, possess only CAMP-, not Ca-, activated apical membrane anion conductance (3), and single channel data have demonstrated that the CAMP-activated channels of Tsc cell apical membranes have electrical properties similar to those of CFTR (23). Finally, CFTR is permeant to iodide (5, 8, 25, 26). Thus CAMP-activated loss of '%I is presumably mediated by CFTR but could be increased either by hyper- polarization of the apical membrane or by increases in the number of open channels. To control for the former possibil- ity, our experiments were performed in high potassium me- dium. In this medium, the potassium conductance should be increased and the apical membrane potential clamped close to the potassium equilibrium potential, which should be ap- proximately zero. Changes in either potassium or chloride conductance should, therefore, have relatively minor effects on the apical membrane potential difference. Thus, we are confident that our assay accurately reflects chloride move- ment through CFTR and that PMA had no effect on the CFTR-mediated increase in chloride permeability in response to cAMP even though the level of mature CFTR protein declined.

Our finding that PMA did not inhibit the forskolin-induced increase in '%I efflux is in apparent disagreement with the results of others (11) who have shown that PMA inhibits forskolin-induced V I efflux from TU cells. One obvious dif- ference is in the isotope used. Radioactive chloride will leave the cell through both anion channels in the apical membrane and by the basolateral NaK2Cl cotransporter; lZ6I exits only via apical membrane anion channels. Of relevance here is the report that treatment with phorbol esters inhibits the NaK2C1 cotransporter of Balb/c 373 cells (27). Another important difference between our study and that of Trapnell et al. (11) is in our use of high potassium medium, which as discussed above, should clamp the membrane potential close to zero.

Thus, although PMA did reduce levels of mature CFTR protein, we suggest that this effect was offset by a synergism between protein kinases C and A on channel opening, such that the CAMP-dependent increase in apical membrane anion permeability was not altered by PMA. How then does PMA inhibit epithelial chloride secretion? There are several trans- port proteins involved in chloride secretion, including apical membrane chloride channels, as well as NaK2Cl cotransport- ers, NaKATPase and potassium channels in the basolateral membrane. Also, there are several steps in the signaling pathways regulating these proteins. However, in dog tracheal epithelium, PMA inhibits CAMP-dependent chloride secre- tion without altering the isoproterenol-induced increase in cAMP or the properties of protein kinase A (9). Thus its actions are most probably at the level of the transport proteins themselves. Rubidium is effectively transported by both PO- tassium channels and the basolateral NaK2CI cotransporter (24, 28). Here, we found that PMA inhibited the CAMP- dependent increase in rubidium efflux. Bumetanide, the in- hibitor of basolateral NaK2C1 cotransport, does not alter the

Protein Kinase C and Chloride Secretion 19075

forskolin-induced increase in =Rb efflux into Na-free (high K) medium? Thus, the inhibition of CAMP-dependent trans- epithelial chloride secretion by PMA is best explained by an inhibition of basolateral potassium channels. Opening of basolateral potassium channels contributes to chloride secre- tion by hyperpolarizing the cell and increasing the driving force for chloride exit across the apical membrane. The inhi- bition of forskolin-induced %C1 efflux by PMA in the studies of Trapnell et al. (11) can thus be accounted for by a failure of potassium channels to open in PMA-treated cells.

In conclusion, we confirm that PMA markedly inhibits epithelial chloride secretion, thereby producing a phenotype resembling cystic fibrosis. This is accompanied by declines in CFTR mRNA and in mature CFTR protein. However, CAMP- activated apical membrane anion permeability does not change, presumably because of a synergistic action of protein kinases A and C on the opening probability of CFTR (6). Instead, our results suggest that inhibition of chloride secre- tion by PMA is mediated by an action on basolateral potas- sium channels.

AcknowZedgments-We are grateful to Dr. Jonathan Cohn for providing antibodies a-102 and a-1468 and for performing the peptide digests. We thank Jocelyn Matsumoto-Pon for technical assistance and Rachel Kline for manuscript preparation.

REFERENCES 1. Frizzell, R. A., Field, M., and Schultz, S. G. (1979) Am. J . Physiol. 236,

l71 -RQ 2. Welsh, M. J. (1987) PhysWL Rev. 67,1143-1184 3. Anderson, M. P and Welsh, M. J. (1991) Proc. NatL Acad. Sci. U. S. A.

" & "

RR. ma&nni' 4. Cliff, W. H., and Frizzell, R. A. (1990) Proc. Natl. Acad. Sei. U. S. A. 87,

4956-4960 6. Kartner, N., Hanrahan, J. W., Jensen, T. J. Naismith, A. L., Sun S.,

Acker!ey, C. A., Reyes, E. F., Tsui, L.-C., dmmens, J . M., Bear, C: E., and Rlordan, J. R. (1991) Cell 64,681-691

", -__- "".

B.-Q. Shen and J. H. Widdicombe, unpublished data.

6. Tahcharani, J. A., Chang, X.-B., Riordan, J. R., and Hanrahan, J. W.

7. Anderson, M. P., Rich, D. P., Gregory, R. J., Smith, A. E., and Welsh, M. (1991) Nature 362,628-631

J. (1991) Scaence 251 ,679682 8. Berger, H. A., Anderson, M. P., Greg", R. J , Thompson, S., Howard, P.

J. Clin. Invest. 88, 1422-1431 W., Maurer, R. A., Mulligan, R., mlth, A E., and Welsh, M. J. (1991)

9. Barthelson, R. A., Jacoby, D. B., and Widdicombe, J. H. (1987) Am. J.

10. Welsh, M. J. (1987) Am. J. PhysioL 2 6 3 , C828-CS34 11. Trapnell, B. C., Zeitlin, P. L., Chu, C.-S., Yoshimura, K., Nakamura, H.,

Lecocq, J.-P., and Crystal, R. G. (1991) J. BioL Chem. 266,10319-10323 Guggino, W. B., Bargon, J., Banks, T. C., Dalemans, W., Pavirani, A.,

12. Riordan, J. R., Rommens, J. M., Karem, B., Alon, N., Rozmahel, R., Grzelczak, Z., Zielenski, J., Lok, S., Plavsic, N., Chou, J.-L., Drumm, M. L., Iannuzzl, M. C., Collins, F. S., and Tsui, L.-C. (1989) Science 246 ,

13. Kitagawa T., and Aikawa, T. (1976) J. Bioehem. 79,233-236 1066-1073

14. Marino, 6. R., Matovcik, L. M., Gorelick, F. S., and Cohn, J. A. (1991) J. Clin. Invest. 88,712-716

15. Cohn, J, A., Melhus, O., Pa e, L. J., Dittrich, K. L., and Vigna, S. R. (1991) Bioehem. BW hys Res &mmun. 181,36-43

16. Smith, P. K., Zrohn, R.'I., Hermanson, G. T., Mallia, A. K., Gartner, F. H., Provenzano, M. D., Fujimoto, E. K., Goeke, N. M., Olson, B. J., and

17. Cohn, J. A., Naim, A.,C., Marino, C. R., Melhus, O., and Kole, J. (1992) Klenk, D. C. (1985) Anal. Bioehem. 71,1392-1401

18. Sarkadi, B., Bauzon, D., Huckle, W. R., Earp, H. S. Berry, A., Suchindran, Pmc. NatL Acad. SCL U. S. A. 89,2340-2344

H., Price, E. M., Olsen, J. C., Boucher, R. C., &d Scarborough, G. A. (1992) J. BWL Chem. 267,2087-2095

19. Laski F. Alzner-DeWeerd, B., RajBhandary, U., and Sharp, P. (1982) N&kd Acids Res. 10,4609-4626

20. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecuhr Cloni Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring H s k

Physml. 2 6 3 , C802-C808

N v 21. Cd& S. H., Gregory, R. J., Marshall, J., Paul, S., White, G. A., O'Riordan,

22. Gray, M. A., Harris A Coleman, L., Greenwell, J. R., and Argent, B. E.

23. Tabcharani J. A., Low, W., Elie, D., and Hanrahan, J. W. (1990) FEBS

24. O'Grady, S. M., Palfrey, H. C., and Field, M. (1987) Am. J. Physiol. 263,

25. Anderson, M. P., Gregory, R. J., Thompson, S., Souza, D. W., Paul, S., Mulligan, R. C., Smith, A. E., and Welsh, M. J. (1991) Science 263,203- 2M

C. k., and Smith, A. E. (1990) CeU 63,827-834

(1989) Am. J. Ph;sidi 2 6 8 , C240-C251

Lett. 270,157-164

C177-C192

26. Dalemans, W., Barbry, P., Champigny, G., Jallat, S., Dott, K., Dreyer, D., Crystal, R. G., Pavirani, A., Lecocq, J.-P., and Lazdunski, M. (1991)

27. Sussman, I., Prettyman, R., and OBrian, T. G. (1985) J. Cell Biol. 101, Nature 354,526-528

28. McCann, J. D., and Welsh, M. J. (1990) Am. J. Physiol. 268, L34SL348 2316-2323