and hormonal responsiveness, the rectal gland of the dogfish shark has become one of the ..... phosphatase inhibitors (10-13) implicates regulation by anal-.
THEJOURNALOF BIOLOGICAL CHEMISTRY
Vol. 267, No. 35, Issue of December 15, pp. 25438-25443,1992 Printed in U.S.A.
0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.
The Na-K-Cl Cotransport Protein of Shark RectalGland 11. REGULATION BY DIRECT PHOSPHORYLATION* (Received for publication, June 19, 1992)
Christian LytleS and Bliss Forbush I11 From the Department of Cellular and Molecular Physiology, Yale University School of Medicine, New Haven, Connecticut 06510 and Mount Desert Island Biological Laboratory, Salsbury Cove, Maine 04672
We determined the relationship between the activaWhen exposed to agents that raise intracellular cAMP tion state and phosphorylation state of the Na-K-Cl (vasoactive intestinal peptide, adenosine, or forskolin), the cotransport protein in tubules isolated from the shark gland rapidly secretes a sea water-like fluid (-0.5 M NaCl). rectal gland, a prototypic chloride-secreting epithe- The transition to the secreting state involves a coordinated lium. In response to CAMP-dependent secretagogues activation of the pathways responsible for dissipative chloride (e.g. vasoactive intestinal peptide, adenosine, and for- exit and concentrative chloride entry (4), i.e. apical chloride skolin) orosmotically induced changes in cell volume, channels (2,3) and basolateral Na-K-Cl cotransporters( 5 , 6 ) . the activation state of the cotransportprotein (assessed The sensitivity of the secretory process to cAMP suggests from measurements of loop diuretic binding) increased that reactions catalyzed by protein kinase A initiate this 6-10 fold. The response was temporally associated with a comparable increase (3-9 fold) in cotransport transition (7). Direct activation of the putative secretory protein phosphorylation. Graded changes in cotrans- chloride channel by protein kinase A has been observed in porter activation evoked proportional changes in co- excised apical membrane patches (8) and inapical membrane transporter phosphorylation. Under the conditions of protein reconstitutedinto artificial bilayers (9).An unresolved our experiments, the 195-kDa cotransporter was the question is whether the Na-K-C1 cotransport protein is also only membrane protein whose phosphorylation state controlled by direct phosphorylation. In support of this posincreased conspicuously in response to both cAMP and sibility, inhibitors of protein kinases and phosphatases have cell shrinkage. Both stimuli promoted phosphorylation proved to be potent effectors of Na-K-C1 cotransport activity of the cotransport protein at serineand threonine res- in other cells (10-13). The entityresponsible for Na-K-Cl cotransport in the rectal idues.One of the CAMP-sensitive phosphoacceptors was found within a segment of the cotransportprotein gland is a 195-kDa basolateral membrane glycoprotein ( 5 , 6, comprised of a sequence (Phe-Gly-His-Asn-Thr*-Ile-14). When the cell is exposed to hormonal secretagogues or Asp-Ala-Val-Pro)that corresponds to a segment of the osmotically shrunken or swollen, the activation state of this Na-K-Cl cotransport protein predicted by cDNAanaly- protein increases 10-fold or more (5, 15). To evaluate the sis, where the phosphoacceptor (Thr*) is threonine hypothesis that thecotransport protein is regulated by direct 189. Incubation of rectal gland tubules with K-252a reversible phosphorylation, we determined the relationship or H-8, structurally differentprotein kinase inhibitors, between its activation state and phosphorylation state in rendered the cotransporter insensitive to both cAMP secretory tubules enzymatically liberated from thin rectal and cell shrinkage. We conclude that the rectal gland gland slices (15). Changes in the cotransporter’s activation Na-K-Cl cotransport protein is regulated by direct re- state were achieved by osmotic perturbation or by forskolin versible phosphorylation at serineand threoninesites. (CAMP) and quantified by specific loop diuretic ([3H]benzmetanide) binding. Phosphorylation was quantified by selective immunoprecipitation of the 32P-labeledcotransport protein with monoclonal antibodies (6). Due to its structural simplicity, experimental accessibility, Our results establish that the cotransport protein acquires and hormonal responsiveness, the rectal gland of the dogfish phosphate at serine and threonine residues in response to shark has become one of the foremost experimental models both CAMP-dependent (forskolin) and CAMP-independent of electrolyte secretion. The gland is comprised of radially (cell shrinkage) stimuli. The quantitative and temporal corbranching tubules formed by specialized chloride-transporting relation between the phosphorylation state and activation cells. These cells, like their counterparts in other secretory state of the Na-K-Cl cotransport protein supportsthe concept epithelia (e.g. the mammalian colonic crypt, pancreatic acithat the turnover rate of this protein is governed by direct nus, salivary gland, airway, and cornea), move chloride from phosphorylation. the blood to the duct (1-3).
* This research was supported by National Institutes of Health Grant DK17433, a postdoctoral fellowship from the American Heart Association (Connecticut Affiliate), and a Blum-Halsey Scholar Award. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. -$ To whom correspondence should be addressed Dept. of Cellular and Molecular Physiology, Yale University School of Medicine, 333 Cedar St., New Haven, CT 06510. Tel.: 203-785-4068; Fax: 203-7856834.
EXPERIMENTAL PROCEDURES
Materials-[’H]Benzmetanide (11.8 Ci/mmol) was synthesized from its precursor amine (kindly provided by Dr. P. W. Feit, Leo Pharmaceuticals, Ballerup, Denmark) and purified as previously described (16). K-252a and H-8l were obtained from Calbiochem, protease inhibitors and collagenase were from Boehringer Mannheim, The abbreviations used are: H-8, N-[2-(methylamino)ethyl]-5isoquinolinesulfonamide; CHAPS, 3-[(3-cholamidopropyl)dimethylammoniol-1-propanesulfonic acid; HPLC, high performance liquid chromatography.
25438
PhosphorylationNa-K-C1 ofCotransport the Protein and forskolin, CHAPS, Hepes, and protein A-Sepharose were from Sigma. Rabbit anti-mouse IgG was from Jackson Immunochemicals, polyvinylidene difluoride (Immobilon-P) was from Millipore, and cellulose thin-layer plates were from Merck (EM Reagents 5617-7). An immunoaffinity matrix was constructed from antibody 54 (6) using the two-layer method of Schneider et al. (17). [32P]Orthophosphoric acid was from Du Pont-New England Nuclear. Tubule Preparation-Experiments were performed on suspensions of secretory tubules enzymatically liberated from thin sections of the dogfish rectal gland as described previously (15). Those reported here were performed during the summers of 1989 (Fig. 2), 1990 (Figs. 24),and 1991 (Figs. 1, 5, and 6). The results obtained during these seasons were quantitatively similar, except that tubules prepared in 1991 consistently exhibited aweaker response to hypertonicity than those prepared previously (see "Results"). Endogenous Protein Phosphorylation-To incorporate 32P into the cellular ATP pool, -400 mg of tubules were incubated for 40 min at 15 "C with 5 ml of phosphate-free elasmobranch Ringer (15) containing 1mCi of [32P]orthophosphate.After two rapid washes, the labeled tubules were incubated (15 "C, 6.7% cytocrit) in fresh solution supplemented with an activator of Na-K-C1 cotransport (20 pM forskolin or 580 mM sucrose). Aliquots were periodically transferred to equal volumes of a "stop" solution containing 150 mM NaC1, 20 mM Hepes (pH 7.5 at 3 "C), 10 mM Na2ATP, 6% CHAPS or 4% SDS, and a battery of protease and phosphatase inhibitors: 5 mM EDTA, 300 PM phenylmethylsulfonyl fluoride, 100 p~ N-tosyl-L-phenylalanine chloromethyl ketone, 1.5 pM pepstatin, 1.5 p M leupeptin, 50 mM sodium fluoride, 15mM sodium pyrophosphate, and 100 p~ sodium vanadate. The samples were rapidly frozen by immersion in liquid nitrogen and stored at -70 "C. Quantification of Phosphorylation-Incorporation of 32Pinto the cotransport protein was measured by two methods. In early experiments, 32P-labeled tubuleswere solubilized in SDS and subjected to gel electrophoresis. After staining the gel with Coomassie Blue, the 32P contentof the 195-kDa band was quantified by (a) autoradiography and film densitometry using a Visage 2000 scanner (Bioimagel Millipore) or ( b ) counting the Cerenkov emission from the excised gel band. In later experiments, cotransporter phosphorylation was quantified by Cerenkov analysis of the immunoprecipitated 195-kDa protein. In this case, the 32P-labeled tubules were solubilized in the solution described above containing 3%CHAPS. After sedimentation of insoluble debris (40,000 X g for 10 min at 3 "C), the detergent extract was incubated for 3 h at 3 "C with monoclonal antibody 53 or 54 (6). Immune complexes were then precipitated by the addition of Protein-A Sepharose beads precoated with affinity-purified rabbit anti-mouse IgG. Proteins retained by the washed matrix were eluted with 4% SDS. Phosphopeptide and Phosphoamino Acid Analysis-Cotransporter fragments were generated from the purified 195-kDa protein using cyanogen bromide and trypsin andthen isolated by HPLCand
conlCAKMPH-8
con1
K
H-8
hypertonic
FIG. 1. Protein kinase inhibitors suppress activation of NaK-Cl cotransport by cAMP and cell shrinkage. Suspensions of rectal gland tubules were incubated for 25 min with 40 ~ L MK-252a ( K ) , 400 p~ H-8, or 0.2% Me2S0 vehicle alone (cont)andthen transferred to an identical medium containing 0.5 p~ [3H]benzmetanide and either 30 p~ 8-(4-~hlorophenylthio)-cAMP(CAMP) or 580 mM sucrose (hypertonic). The amount of ['Hlbenzmetanide specifically bound to the tubules after 30 min, relative to the paired control, is plotted on the ordinate (mean +. S.E., n = 4). Quiescent tubules bound 1.8 0.28 pmol/mg of protein, whereas those stimulated by cAMP or hypertonicity bound 17.0 f 0.83 or 13.8 2 1.93, respectively.
*
25439
sequenced as described in the preceding article (6). The phosphoamino acid composition of the 195-kDa protein was analyzed by the procedure of Kamps and Sefton(18). Protein isolated by the 54 immunoaffinity matrix was fractionated by SDS gel electrophoresis and transferred to polyvinylidene difluoride using a Bio-Rad TransBlot apparatus. The glycine transfer buffer contained 10% ethanol and 0.1% SDS. Theregion of the blot containing the 195-kDa protein was excised, washed, and incubated in boiling hydrochloric acid (5.7 M ) for 1 h. The hydrolysate was lyophylized, supplemented with nonradioactivephosphoamino acid standards,and separated by crossed electrophoresis at pH 1.9 and pH 3.5 on cellulose thin-layer plates (19). Phosphoamino acids were detected by autoradiography using preflashed film, and the standardsby ninhydrin staining.
RESULTS
Effect of Protein Kinase Inhibitors on Nu-K-C1 Cotransport-To determine if protein phosphorylation isinvolved in the activation of Na-K-C1 cotransport by cAMP and cell shrinkage, we tested the effects of two protein kinase inhibitors, K-252a and H-8, on [3H]benzmetanide binding to isolated rectal gland tubules. Benzmetanide, like its structural analog bumetanide,is a potent inhibitor of Na-K-Cl cotransport whose binding to intactcells parallels Na-K-C1 cotransport activity (5, 15, 20, 21). As found previously (15), specific [3H]benzmetanide binding increased profoundly in response to cAMP(9.6 fold) and hypertonicity(7.8 fold). When present , blunted the response to each stimulus by at 40 p ~ K-252a -78%. The inhibition does not appear to involve a direct interaction of K-252a with the cotransporteritself, since this agent does notinterferewithion-dependent loop diuretic binding toisolated duck redcell membranes (21). We obtained similar results with the structurally unrelated kinase inhibitor H-8. When present at400 p ~H-8 , rendered the cotransporter insensitive to cAMP and inhibited response its to cell shrinkbe age by 70%. Thus,proteinphosphorylationappearsto involved in the activation of Na-K-Cl cotransport by both cAMP and cell shrinkage. Protein Phosphorylation in Rectal Gland Tubules-To evaluate the effect of cotransport stimuli on membrane protein phosphorylation, thecellular ATP pool of rectal gland tubules was labeled by incubation with [32P]orthophosphate. During a 40-min incubation, 32Pbecame incorporated into numerous membrane proteins (Fig. 2). When the labeled tubules were then exposed for 15 min to activators of Na-K-Cl cotransport (20 p M forskolin or 580 mM sucrose), the phosphorylation state of just one protein changedconspicuously; both activators evoked a large (-4-fold) increase in the 32Pcontent of the -195 kDa band. The fact that54, a monoclonal antibody that selectively recognizes the rectal gland Na-K-C1 cotransporter (6), efficiently immunoprecipitates the 195-kDa protein and its phosphorylated form (Fig. 2) verifies that this phosphoprotein is the cotransporter. Equivalent resultswere obtained infive similar experimentsemploying 54,and intwo others using 53, amonoclonal antibody that recognizes a different structural domain of the 195-kDa cotransport prowas the tein (6). In each experiment, the cotransport protein major component and thesole radioactive constituent of the immunoprecipitate (e.g. see Fig. 2). We detected no change in the electrophoretic mobility of the cotransport protein upon phosphorylation, nor differences in the protein composition of the immunoprecipitate (Fig. 2). Time Course of Cotransporter Phosphorylation and Dephosphorylation-Analyses of loop diuretic bindingrates have established that cotransporter activation becomes maximal 5-10 min afterexposure to forskolin or hypertonicity (15). To determine if cotransporter phosphorylation follows a similar time course, the 32P content of the195-kDaprotein was monitored under equivalent experimental conditions. Both
25440
Phosphorylation of the Nu-K-Cl Cotransport Protein \
FIG.2. Activators of cotransport promote phosphorylation of the NaR-Cl cotransport protein. Rectal eland tubules were labeled with 32Pfor 1 h, washed quickly, and then exposed for 15 min to fresh medium containing forskolin (10 p M ) or 580 mM sucrose (hypertonic). The 195-kDa cotransport protein was isolatedfrom each CHAPSsolubilized sample using the 54 immunomatrix. Immunopurified proteins, along with samples of cellular protein obtained before (total) and after (residual) incubation with the immunoaffinity matrix were separated by gel electrophoresis and stained with Coomassie Blue (protein).Phosphoproteins were visualThe '*P ized by autoradiography ("P). content of the cotransport protein (quantifiedasthe radioactivity in the excised 195 kDa band) increased 4.3- or 3.3-fold in response to forskolin or hypertonicity, respectively. Representative of five similar experiments.
6 ".
40
to'
*Q
*Q $ .G
x0
8 0 Qet $0
8
Go
\
$0
cG
x0
\
to F+
*e
,$ d
$0 Go to
Q ' .
c"
x0
Q ' .
"
"
69 46
-
21 14-
-
"
total
residual
immunopurified
protein
forskolin and hypertonicity promoted a progressive increase tion, respond in the same complex way to cAMP (forskolin) in cotransporterphosphorylation that peaked after 8-13 min and to changesin cell volume (external osmolality). The (Fig. 3). In five similar experiments,maximal phosphorylation experiment alsoshows that ourtwo measuresof cotransporter occurred -9 min after treatmentwith forskolin and -12 min phosphorylation (autoradiography of the 195 kDa gel band after osmotic perturbation. In each experiment, the cotrans- and Cerenkov emission from the excised 195 kDa band) yield port protein appeared to be the only membrane protein to equivalent results. Further quantification of the phosphorylation response was gain large amounts of 32Pin response to both forskolin and obtained in two series of immunoprecipitation experiments. hypertonicity (e.g. see Fig. 2). period from The first employed monoclonal antibody 53 (6) toselectively By reducing the [32P]orthophosphate-labeling 40-90 min to 10-20 min, a lower base-line level of protein immunoprecipitate thecotransport protein from CHAPSphosphorylation could be achieved. In some experiments of solubilized tubules. In threesuch experiments,the 32Pcontent this design, the cotransporterappeared to be the only cellular of the cotransporter protein increased 8.8 f 1.73 or 4.4 f protein to become labeled during the period of stimulation 0.56-fold after stimulation with forskolin (20 p ~ or) hyper(Fig. 3C). Thus, theprotein kinase reaction(s)that act on the tonicity (+580 mM sucrose), respectively, for 8 min. Similar cotransporter in response to cAMP and cell shrinkage would changes inthe activation state of the cotransportprotein were appear to be among the most active in the rectal gland cell. observed in measurements of [3H]benzmetanide binding to With such a low background phosphorylation, the effect of tubules prepared during the same season. Interestingly, the forskolin and hypertonicity become particularly conspicuous. tubules prepared during thisseason were only half as responA comparison of Fig. 3 with our previous time course experi- sive to hypertonicity as they were to forskolin, both with respect to phosphorylation and activation. Our finding that ments (e.g. Fig. 5 and Fig. 6 of Ref. 15) shows that the cotransport protein is activated and phosphorylated a t similar these parameters co-vary further substantiates the concept rates. that phosphorylation causes activation. Phosphoamino Acid Analysis of the Cotransport ProteinThe synchrony was also observed when the hypertonic stimulus was removed. After restoration of cell volume, the In order to identify sites of reversible phosphorylation, we cotransport protein returned to its basal state of activation determined the phosphoamino acid composition of the puri(cf.Fig. 3B and Fig. 6B in Ref. 15)and phosphorylation fied cotransport protein. After exposing 32P-labeledtubules to within 8-14 min (Fig. 3B). Thus, changes in cotransporter forskolin or hypertonicity for 8 min, the 195-kDa protein was activation and phosphorylation appear to be temporally cor- isolated using the 54 monoclonal antibody matrix (6). To obtain essentially pure cotransport protein, the immunoprerelated. Activation and Phosphorylation Are Qualitatively Corre- cipitate was then separated by SDS gel electrophoresis, and lated-Measurements of [3H]benzmetanide binding have es- the 195-kDa protein band was isolated by excision and electablished that the Na-K-C1 cotransport protein is activated troelution. After isolation, forskolin-stimulated cotransportby secretagogues, cell shrinkage, and cell swelling (5, 15). An ers contained 4.3 f 0.2 times as much 32Pas did quiescent example of this response is seen in the top panel of Fig. 4, ones, and those stimulatedby cell shrinkage contained 2.9 f where [3H]benzmetanidebinding, a sensitive index of cotrans- 0.4 times as much (mean f S.E., n = 4). The protein was porter activation, is plottedas a function of external osmolal- partially hydrolyzed in acid, and its constituents were sepaity, a parameter known to be inversely proportional to rectal rated by two-dimensional thin-layer electrophoresis on cellugland cell volume (15). To determine if the phosphorylation lose plates. As shown in Fig. 5, the radioactivity was resolved state of the cotransportprotein parallels its activation state, into spots corresponding to unhydrolyzed protein, phosphowe repeated the experiment on 32P-labeledtubules and meas- peptides,phosphoamino acids, and free phosphate. In reured the radioactivity of the 195-kDa protein band on SDS sponse to forskolin (CAMP),the phosphoserine and phospho1.3 gels. A comparison of the top and bottom panels of Fig. 4 threonine content of the cotransport protein rose 3.4 indicates that both parameters, phosphorylation and activa- and 5.6 k 1.1fold, respectively. A similar increase (3.6 0.5
* *
Phosphorylation of the Na-K-Cl Cotransport Protein
m
Y
0
,
5
1
2
0.0
isotonic
-
9 O
25441
0
I
I
5
10
I n -
I
15
isotonic
+
0.0
, 0
I 5
,
I
,
I I5
10
,
I
,
I
,
I
,
30
25
20
. 35
time (min) C 0 2 . 4 .6 8 1 0 1 2 1 4 " . 200
0
2
.
.
-
69 46 -
FIG.3. Rapidity and reversibility of Na-K-CI cotransport protein phosphorylation. A suspension of 3ZP-labeledrectal gland tubules was incubated at 15 "C with a phosphate-free isotonic solution containing or lacking an activator of Na-K-C1 cotransport. Samples were removed a t the indicated times and denatured with SDS. After electrophoretic separation of the cellular proteins, the 3zPcontent of the 195-kDa protein band was quantified by autoradiography and , densitometry. Inpaneld, theactivator was forskolin (20 p ~ )whereas in panel B it was hypertonicity (+580 mM sucrose); after 18 min, isotonicity was restored. Representative of three experiments. Panel C portrays an experiment in which a shorter (15 min) [32P]orthophosphate labeling period yielded an unusually low base-line level of protein phosphorylation; in this circumstance, the 195-kDa cotransport protein is the only detectable cellular protein to be phosphorylated in response to forskolin or hypertonicity.
and 5.8 & 0.7 fold, respectively) was evoked by cell shrinkage. No incorporation of 32Pinto tyrosine residues was detected. Localization of a CAMP-sensitive Phosphothreonine within the Primary Sequence-To identify sites of regulatory phosphorylation, we isolated and sequenced small fragments of the cotransport proteinthat acquire 32Pin response to CAMP. The cotransport protein was isolated from forskolin-stimulated 32P-labeledtubules by immunoaffinity purification and gel electrophoresis, as described above. After exhaustive digestion of the protein with cyanogen bromide and trypsin, the resulting fragments were separated by two steps of reversephase HPLC (6). The chromatogram shown in Fig. 6 illustrates the elution profile obtained in the first HPLCseparation using a CIS column. Most of the recovered radioactivity emerged within two peaks (denoted a and b in Fig. 6). After re-purification on ac8 column, the radioactive peptides within peaks a and b were sequenced by automated Edman degradation. Because their phenylthiohydantoin adducts are not detected in an Edman cycle, phosphoserine or phosphothre-
900
1100
1300
1500
osmolality (mom)
4 6 8 1 0 1 2 1 4
93
30
700
minutes with fonkolin
minutes in hypertonic medium
FIG.4. Correlation between cotransporter activation and cotransporter phosphorylation. Graded changes in cotransporter activation were produced by osmotically shrinking or swelling the rectal gland cell (filled symbols). Maximal activation was elicited by 20 PM forskolin (open symbols). Top, effect of osmotic perturbation on [3H]benzmetanidebinding site density, a reflection of the cotransporter's regulatory state.Departures from isotonicity (915 mosm, denoted as an arrow on the abscissa) were achieved by the addition of sucrose (1065-1567 mosm) or by the omission of NaCl (690-775 mosm). An inverse relationship was found between external osmolality and rectal gland cell water content (kg of cell HnO/kg of cell solid = (3, 237/mosm) - 0.752). The data in this panel, adapted from Fig. 3 of Ref. 15, represent the mean S.E. of four experiments. Bottom, effect of osmotic perturbation on the "P content of the 195-kDa cotransport protein. Phosphorylation was ascertained by two methods: (a) scintillation counting of the 195 kDa gel band (circles); or (b) autoradiography and densitometry (squares). Data in the lower panel are representative of three experiments.
*
CONTROL
FORSKOLIN
HYPERTONIC
FIG.5. Phosphoamino acid analysis of the Na-K-CI cotransport protein. Rectal gland tubules were labeled with "P for 20 min a t 15 "C, washed, and then exposed for 8 min at 15 "Cto a medium lacking (CONTROL)or containing an activator of cotransport, either 20 p M forskolin (FORSKOLIN)or 580 mM sucrose (ffYPER7'ONzc). The 195-kDa cotransport protein was isolated by 54 immunoaffinity purification followed by SDS gel electrophoresis. Acid hydrolysates of the protein were spotted onto cellulose plates (lower left corner) and separated by thin-layer electrophoresis in two dimensions (first toward the top a t pH 1.9, then toward the right at pH 3.5). Unlabeled and phosphocarrier phosphoserine ( P - S ) ,phosphothreonine (P-T), tyrosine ( P - Y )were detected by ninhydrin staining (traced outlines). Radioactive phosphoamino acids were visualizedby autoradiography. Representative of four experiments.
25442
PhosphorylationNa-K-C1 ofCotransport the Protein
will make this a challenging undertaking. Definitive proof that cotransporter phosphorylation causes cotransport will ultimately require identification and genetic modification of the serine and threonine phosphoacceptors that are sensitive to cAMP and cell volume. We have located one of the CAMP-sensitive phosphothreonines (Thr*) within a segment of the protein (Phe-Gly-His-Asn-Thr*-Ile-AspAla-Val-Pro) that corresponds to positions 185-194of the Na-K-C1 cotransport protein encoded by a cloned cDNA (22). The region encompassing this threonine does not constitute an optimal consensus target for well characterized protein 50 100 150 kinases (23), though it does conform to the broad requireelution time (min) ments of casein kinase I1 (Ser*/Thr*-[Asp/Gl~/Ser-P]~-~-X~FIG. 6. SeparationofNa-K-Clcotransportprotein fragwas o); this alone, however, does not assurethat casein kinase acts ments by reverse-phase HPLC. Thecotransportprotein isolated from forskolin-stimulated 32P-labeled tubules by immunoaf- on Thr’*’, nor does it exclude the possibility that otherknown gel electrophoresis and thenexhaustively finitypurificationand serine/threonine kinases act there. digested with CNBr and trypsin. Peptide fragments were isolatedby Under the conditions of our assay, the 195-kDa cotransport HPLC on a narrow bore CIScolumn. The chromatogram illustrates protein proved to be the only membrane protein to undergo the elution profile of the peptide fragments (solid truce) and a blank large changes in phosphorylation state in response to cAMP digestion (broken truce). Most of the recovered 32Pwas found within the peaks labeled a and b. The peptides in the peak labeled a were and cell shrinkage. Though not apparent in our experiments, repurified on a Cs column and successfully sequenced.The small phosphorylation of the cotransport protein during CAMParrows denote peaks corresponding tothose analyzed in the preceding induced secretion probably follows or accompanies phosreport (6). phorylation of the apical chloride channel (8, 9), a 160-kDa membrane protein homologous to the human cystic fibrosis onine residues appear as gaps in the derived peptide sequence. transmembrane conductance regulator (24). Because these Unambiguous sequence information was obtained for peptide channels probably constitute a minute fraction of the memA: Phe-Gly-His-Asn-(gap)-Ile-Asp-Ala-Val-Pro. This seg- brane protein, our analysis of total cellular phosphoprotein ment is found as Phe-Gly-His-Asn-Thr-Ile-Asp-Ala-Val-Pro probably lacks the sensitivity necessary to detect them. in thesequence of the Na-K-Cl cotransport proteinpredicted We have postulated that theactivation of Na-K-C1cotransby cDNA analysis (22). We infer that the fifth position of port during secretion representsa corrective response to this segment, which corresponds to position 189 of the intact CAMP-induced chloride loss throughthese apical chloride protein, represents a threonine residue that serves as a CAMP- channels ( 5 , 15). The “chloride-coupling’’ hypothesis is supsensitive phosphoacceptor. ported by our observation that the cotransportprotein is phosphorylated and activated by maneuvers known to reduce tubule cell chloride concentration (i.e. exposure to CAMP,low DISCUSSION external chloride, or hypotonicity) and that its response to Our results establish that the activation state and phos- cAMP is blocked by maneuvers known to prohibit conductive phorylation state of the rectal gland Na-K-C1 cotransport chloride loss ( i e . exposure to elevated extracellular potassium protein are closely correlated. Both states increase propor- or barium) (5, 15, 25). An inverse relationship between intrationately when rectal gland tubules are osmotically shrunken cellular chloride concentration and Na-K-C1 cotransport acor swollen, or when they are exposed to a hormonal secretativity has also been found in the squid giant axon (26). The gogue. We infer that phosphorylation of the cotransport promechanism by whichcytoplasmic chloride influences cotranstein enhances its intrinsic activity. In support of this concept, port protein phosphorylation is not known; perhaps the cowe found that structurally unrelated proteinkinase inhibitors (K-252a and H-8) block the cotransporter’s response to cAMP transporter is phosphorylated more rapidly (or dephosphorylated moreslowly) when one or both of its own internal and cell shrinkage, and that stimulus-induced cotransporter chloride transport sites become unoccupied. phosphorylation persists after inhibitionof Na-K-C1 cotransThe finding that forskolin and cell shrinkage together do port with bumetanide. Thus, phosphorylation appears to be not produce additive effects on cotransporter activation (15) the cause, rather than the consequence, of Na-K-C1 cotranssuggests that ( a ) both stimuli promote phosphorylation of the port activation. The cellular response to secretagogues, seen as a profound same set of serine and threonine residues or ( b )cotransporters increasein apical chloride conductance, specific [3H]benz- activated by phosphorylation at one site are not further actimetanide binding, oxygen consumption, and chloride secre- vated by phosphorylation at another. To clarify this issue, it tion, could reflect an acceleration of basolateral Na-K-C1 will be important to isolate and sequence fragments of the cotransporters and/or a recruitmentof an additional one from cotransport protein that become phosphorylated in response intracellular repositories (5, 15). The close relationship be- to different stimuli. The identity of these sites may provide tween this response and cotransport protein phosphorylation insight into the specific kinase(s) and phosphatase(s) which suggests that direct phosphorylation of the cotransporter act on them (23). Recent studies on another chloride-secreting epithelium, enhances its catalytic rate. Nevertheless, our results do not exclude a potentialinvolvement of protein migration. Cycling the duck salt gland, have implicated a 170-kDa membrane of the cotransport protein between intracellular membranous protein in the cotransport process (27). The mutual recognipools and the basolateral membrane, should it occur, might tion of the duck and shark Na-K-Cl cotransporters by four be detected by immunoelectron microscopy using the mono- different monoclonal antibodies (27)indicates that these proteins are structurally related. As in the rectal gland, hormonclonal antibodies described in the preceding report (6); yet even with the spacial resolution afforded by this technique, ally triggered secretion by the duck salt gland is temporally the morphological complexity of the rectal gland secretory associated with a prominent increase (3-8 fold) in the phoscell, with its tortuously invaginated basolateral membrane, phorylation state of the cotransport protein (27, 28).
Phosphorylation of the Nu-K-C1 Cotransport Protein We presume that the proteins responsible for Na-K-C1 cotransport in other animal cells (29) are also homologous. Their mutual sensitivity to a diverse family of loop diuretic inhibitors suggests that they possess similar functionally important domains, and their sensitivity to protein kinase and phosphatase inhibitors (10-13) implicates regulation by analogous phosphorylation reactions. A recent study employing these inhibitors (11) suggests that atleast two kinases participate in the regulation of Na-K-Cl cotransport in the duck red cell. In this cell, as in the shark rectal gland cell, Na-KC1 cotransport and loop diuretic binding are stimulated by CAMP-dependent (forskolin) and CAMP-independent (cell shrinkage) reactions blocked by the protein kinase inhibitor K-252a. Neither calmodulin-dependent kinases, which have been found in rectal gland membranes (30), nor protein kinase C appear tobe major effectors of rectal gland Na-K-Cl cotransport, since agents that lead to theiractivation (phorbol esters or calcium ionophores), regardless of whether they areapplied alone or in combination, fail to promote chloride secretion by the perfused gland (31), oxygen consumption by thin slices of the gland (32), or [3H]benzmetanidebinding to isolated rectal gland tubules.' These kinases appear to play a much more important regulatory role in otherchloride-secreting epithelia (28, 33). Our results provide the first evidence that the Na-K-C1 cotransport protein is regulated by direct phosphorylation.In the rectal gland, the phosphorylation state of the cotransport protein appears to be governed in a complex way by stimulatory (vasoactive intestinal peptide and C-type natriuretic peptide) and inhibitory(somatostatin and A1 adenosine) receptor-mediated biochemical signals. Our ultimate objective is to determine how these signals allow chloride entry (via basolateral Na-K-Cl cotransporters)to keep perfect pace with chloride exit (via apical chloride channels). Acknowledgments-We thank Grace Jones for skilled technical assistance, Kathy Stoneand Kenneth Williams (Yale Protein Chemistry Facility) for phosphopeptide isolation and sequencing, and Dr. C. Lytle and B. Forbush 111, unpublished observations.
25443
David Stern (Department of Pathology, Yale University) for help with phosphoamino acid analyses. REFERENCES 1. Burger, J. W., and Hess, W. N. (1960)Science 131,760-761 2. Silva, P., Stoff, J., Field, M., Fine, L., Forrest, J. N., Jr., and Epstein,F. H. (1977)Am. J. Physiol. 233, F298-F306 3. Greger, R., Schlatter, E., Wong, F., and Forrest, J. N., Jr. (1984)Pflugers Arch. 402,376-384 4. Welsh, M., Smith, P. L., and Frizzell, R. (1983)Am. J. Physiol. 245,F640FGAA
5. FoiC~ki,B., 111, Haas, M., and Lytle, C. (1992) Am. J. Physiol. 262, C1000-C1008 6. Lytle, C.,Xu, J.-C., Biemesderfer, D., Haas, M., and Forbush, B., 111 (1992) J. Bwl. Chem. 267,25428-25437 7. Palfrey, H. C.,Alper, S. L., and Greengard, P. (1980)J. Exp. Biol. 89,103115
8. Griper, R., Gogelein, H., and Schlatter, E. (1988)Comp. Biochem. Physiol. A Comp. Physiol. BOA, 733-737 9. La. B.4.. Carosi. S. L.. Valentlch. J.. Shenolikar., S.., and Sansom. S. C. (1991jkm.J. Physiol: 260, C1217-C1233 10. OGrady, S. M., DeJonge, H. R., Vaandrager, A. B., and Field, M. (1988) Am. J. Physiol. 264, C115-C121 11. Pewitt, E. B., Hegde, R. S., Haas, M., and Palfrey, H. C. (1990)J. Biol. Chem. 265.20747-20756 12. Altamirano, A. A., Breitwieser, G. E., and Russell. J. M. (1988)Am. J. Physiol. 254, C582-C586 13. Kaji, D. (1991)FASEB J. 6 , A739 (abstr.) 14. Forbush, B., 111, and Haas, M. (1988)Biophys. J. 53,222a (abstr.) 15. Lytle, C.,and Forbush, B., 111 (1992)Am. J. Physiol. 262, C1009-C1017 16. Forbush, B., 111, and Palfrey, H. C.(1983)J. Biol. Chem. 258,11787-11792 17. Schneider, C., Newman, R., Sutherland, D. R., Asser, U., and Greaves, M. F. (1982)J. Biol. Chem. 257,10766-10769 18. Kamps, M. P., and Sefton, B. M. (1989)Anal. Biochem. 176,22-27 19. Cooper, J. A., Sefton, B. M., and Hunter,T. (1983)Methods Enzymol. 3874n3
20. Hais: M., and Forbush, B., 111 (1986)J. Biol. Chem. 261,8434-8441 21. Pewitt, E. B., Hegde, R. S., and Palfrey, H. C. (1990)J. Biol. Chem. 265, 14364-14370 22. Xu,J., Lytle, C.,Zhu, T.,Benz, E. J., and Forbush, B., 111 (1992)J. Gen. Physiol. 100,39a (abstr.) 23. Kennelp P. J., andKrebs, E. G. (1991)J. Biol. Chem. 266,15555-15558 24. Marsha I, J., Martm, K. A,, Picciotto, M., Hockfield, S., Nairn, A. C.,and Kaczmarek, L. K. (1991)J. Biol. Chum. 266,22749-22754 25. Lytle, C., and Forbush, B., I11 (1992)Bioph s J 61,A384 (abstr.) 26. Breltwleser, G. E., Altamirano, A. A,,andYRussel1, J. M. (1990)Am. J. Physiol. 258, C749C753 27. Torchla, J., Lytle C.,Pon, D. J., Forbush, B., 111, and Sen, A. K. (1992)J. Biol. Chem. 26?, 25444-25450 28. Torchia, J., Qu, Y., Francis, J., Pon, D. J., and Sen, A. K. (1991)Am. J. Physiol. 261, C543-C549 29. Haas, M. (1989)Annu. Reu. Physiol. 61,443-457 30. Palfrey, H. C.,Epstein, F. H., Silva, P., and Greengard, P. (1980)Bull. Mt. Desert Isl. Biol. L a b . 20, 146-151 31. Silva, P., Solomon, R. Brignull, H., Hornung, S., Landsberg J. Solomon, H., Wolff, D., and Epstein, F. (1992)Bull. Mt. Desert I d . Bid. Lab. 31, 65-67 32. Shuttleworth, T. J. (1983)Am. J. Physiol. 246, R894-R900 33. Kachintorn, U.,Vongkovit, P., Vajanaphanich, M., Dinh, S., Barret, K. E., and Dharmsatbaphorn, K. (1992)Am. J. Physiol. 262, C15-C22
