The Role of Cytosolic Ca2+, Protein Kinase C, and Protein Kinase A in ...

CHEMISTRY AL THEJOURNAL OF B I O L ~ ~ I C 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 2, Issue of January 14, pp. 849-859, 1994 Printed in U.S.A.

The Role of Cytosolic Ca2+,Protein Kinase C, and Protein Kinase A in Hormonal Stimulation of Phospholipase D in Rat Hepatocytes* (Received forpublication, June 2, 1993, and in revised form, August 30, 1993)

Lena GustavssonS, Gisela Moehren, M. Eugenia Torres-Marquez, Christine BenistantB, Raphael Rubin, and JanB. Hoekll From the Department of Pathology and Cell Biology, Thomas Jefferson Uniuersity, Philadelphia, Pennsylvania 19107

receptors in the plasma membrane(for reviews, see Refs. 1-41, but the mechanisms linking receptor occupation to activationof this enzyme in intact cells remain controversial. From studies on isolated plasma membranes and permeabilized cells, PLD appears to be coupled to some receptors through an as yet unidentified G protein (1-5). However, other mechanisms of activationexist. Hormones operatingthrough polyphosphoinositide-specific phospholipase C (PLC) increase cellular levels of inositol 1,4,5-trisphosphate (InsP3) and 1,2-diacylglycerol (DAG), thereby increasing cytosolic free Ca2+ concentrations ([Ca2+],) and activating protein kinase C. Both of these could contribute to PLD activation. Protein kinase C activation by phorbol esters stimulatesPLD in different cell types (1-46). In several studies, inhibitorsof protein kinase C or down-regulation of the enzyme prevented receptor-mediated activation of PLD (7-13). Overexpression of protein kinase C P l enhanced thrombin-stimulated PLD activation in Rat 6 fibroblasts (14). However, other studies reported thatPLD activation by different agonists was notaffected by inhibitors of protein kinase C or by down-regulation of the enzyme (15-19). Purified protein kinase C directly activated PLD in fibroblast plasma membrane preparations, by a mechanism that appeared to be independent of the protein kinaseactivity of the enzyme (20). Kiss and Anderson (16) found that a phosphatidylethanolamine-specific PLD was selectively enhanced inNIH-3T3 cells by inhibitors of protein kinase C. Changes in [Ca2+lc,, induced by Ca2+ ionophores also activate PLD (21-24). Some of the effects of Ca2+ionophores may be indirect, since other phospholipases are activated by high [Ca2+l,, causing an accumulation of DAG, arachidonate, or other activators of protein kinase C. Inhibition of receptormediated PLD activation by depleting intracellularCa2+stores orbuffering[Caz+lcfi wasreportedinneutrophils (25,26), erythroleukemia cells (241, and other cell types (27). By conPhospholipase D (PLD)l activity in hepatocytes and other trast, purinergic stimulation of PLD in a murine macrophage cell types is stimulatedby a variety of agonists acting through cell line was independentof changes in[Ca2+Ic,, (28). Genyand in * This workwas supported in part by the United States Public Health Cockcroft (29) reported that Ca2+ was required micromolar Service Grants AA07186, AA07215, AA08714, AA07309,and AA00123, concentrations for GTPyS-dependent stimulation of PLD in the Swedish Medical ResearchCouncil(ProjectNo. 088071, and the permeabilizedHL-60 cells. Liscovitch and Eli (30) observed Swedish Alcohol Research Fund. The costsof publication of this article inhibition of GTPyS-stimulated PLD by Ca2+ inpermeabilized were defrayed in part by the payment of page charges. This article must NG108-15 cells and PLD activation by GTPyS in liver plasma therefore be hereby marked “aduertisernent”in accordancewith 18 membranes wasfound not to require Ca2+ (21,31,32). A soluble U.S.C. Section 1734 solely to indicate this fact. $ Permanent address: Dept. of Psychiatry and Neurochemistry, Uni- PLD, detected in bovine tissues by Wang et al. (33), was stimuversity of Lund, P. 0. Box 638, 5-22009 Lund, Sweden. lated by millimolar Ca2+levels. Similarly, a soluble PI-selective 3 Permanent address: Unite INSERM 205, Laboratoire de Chimie Biologique INSA, 20 Av. A. Einstein, 69621 Villeurbanne, Cedex, PLD activity in permeabilized Madin-Darbey canine kidney a (MDCK) cells wasstrictlyCaz+-dependent(34),whereas France. ll To whom correspondence should be addressed: Dept. of Pathology and Cell Biology, Thomas Jefferson University, 1020 Locust St., Philadelphia, PA 19107. Fax: 215-923-2218. methylpiperazine;H89, N-(2-((3-(4-bromophenyl)-2-propenyl)-amino)The abbreviations used are: PLD, phospholipase D; [Ca2+],, cytoethyl)-5-isoquinolinesulfonamide; InsP,,inositol1,4,5-trisphosphate; solic free Ca2+ concentration; [Ca2+],,,, extracellular CaZ+ concentra- PA, phosphatidic acid; PC,phosphatidylcholine; Peth, phosphatition;CPT-CAMP, 8-(4-~hlorophenylthio)adenosine3’:5’-cyclic mono- dylethanol; PI, phosphatidylinositol; phospholipase C; RpPLC, phosphate; DAG, 1,2-diacylglycerol; D’TT, dithiothreitol; GTPyS, cAMP[SI, (Rp)-adenosine cyclic 3‘,5’-phosphorothioate;TPA, 12-0-tetguanosine-5’-0-(3-thiotriphosphate);H7, l-(5-isoquinolylsulfonyl)-2- radecanoylphorbol 13-acetate.

Ca2+-dependent and protein kinase C-dependent mechanisms of phospholipase D (PLD) activation were studied in rat hepatocytes by measuring phosphatidylethanol (Peth) formation in thepresence of ethanol. Stimulation of Peth formation by 12-0-tetradecanoylphorboll3-acetate (TPA),vasopressin, orA23187 wasinhibited by multiple protein kinase C inhibitors or by protein kinase C down-regulation, indicating that this enzyme is involved in the action of all these agents. A controlled elevation of the cytosolic Ca2+ concentration ([Ca2+Icfl) over the range of0.1-2.0 p~ activated Peth formation in theabsence of other agonists. Staurosporin potentiated Ca2+-inducedPeth formation by shifting the [Ca2+],,, dose-response curve to the left. Other protein kinase C inhibitors (calphostin C, bisindolylmaleimide) inhibited Ca2+-mediatedPeth formation, but thisinhibition was reduced in staurosporin-treated cells. Okadaic acid potentiated PLD activation by TPA, but suppressed PLD activation by elevated [Ca2+lCyt. Desensitization of TPA-induced PLDactivity did not affect PLD activation by Ca2+. Thesedata indicate that [Ca2+lcyt and protein kinase C control distinct pathways ofPLD activation, but theCa2+-mediated pathway is suppressed by a staurosporin-sensitive protein kinase.Bothmechanisms contribute to vasopressin-inducedPeth formation in intact hepatocytes. Activation of protein kinase A enhanced vasopressininduced Peth formation, but not TPA-stimulatedor Ca2+stimulated Peth formation. Protein kinase A acted by enhancing hormonal Ca2+ mobilization,rather than by directly activating PLD, and thereby shifted the balance of Ca2+-dependentand protein kinase C-dependent activation mechanisms of PLD in intactcells.

849

850

Mechanisms of Activation of Phospholipase D

membrane-bound PC-specific PLD in the same study was in- length pair of 340 n d 3 9 5 nm (indo 1) or 506 n d 5 2 6 nm (fluo 3) in a Perkin-Elmer Cetus 44B-MPF spectrofluorometer. After each incubahibited by millimolar Ca2+ levels. Thus, it remains unclear whether changes in [Ca2+],,, in the physiologically relevant tion, the fluorescence scale wascalibrated by adding excess CaCl, and [Ca*+],,, was range affect PLDactivity in intact cells. How the different ionomycin,followedby quenching with MnC1, (1 m). calculated as described previously (38). modes of activation contribute tohormonal stimulation of PLD in intact cells also remains tobe defined. RESULTS Activation of PLD in thepresence of ethanol typically results Contributions of [Ca2+Ic,,,a n d Protein Kinase C to Phosphoin the formationof phosphatidylethanol (Peth),at the expense of PA accumulation (3,31, 35, 36).In theaccompanying paper lipase D Activation in IntactHepatocytes-Initial studies indi(371, we demonstrated that the initial rate of vasopressin- cated that both [Ca2+l,,, and protein kinase C contribute to stimulated or TPA-stimulated PLD in intact hepatocytes is not hormonal stimulation of PLD in intact hepatocytes. PLD is itself affected by the presence of ethanol. Peth is metabolized stimulated both by a direct activation of protein kinase C by only slowly, in contrast toPA, and Peth accumulation is there- TPA or by the Ca2+ ionophore A23187 (37). When cells were fore a better indicatorof PLD activity than the formation of PA depleted of Ca2+ by preincubation with EGTA, TPA-induced Peth formation was only slightly depressed (20-30%), but both or DAG. In thispaper, we used Peth formation as an indicator of PLD the ionophore-induced and the receptor-mediatedresponses 95%inhibition for that the enzyme were almostcompletely inhibited (more than activity in intact hepatocytes and demonstrate vasopressin andA23187; data not shown). A23187 also disrupts is stimulated independently by protein kinase C and by an elevation of[Ca2+lc,, in the range of0.2-2 w. Both mecha- cellular M e distribution. However, Mg2+was not effective in nisms of activation contribute to the activation of PLD by va- activating Peth formation in ionophore-treatedcells in theabsopressin. The Ca2+-stimulatedPLD activation was enhanced sence of Ca2+ andomission of Mg2+ from the medium did not by the protein kinase inhibitor staurosporin and suppressed by affect the A23187-induced Peth formation in the presence of the protein phosphatase inhibitor okadaic acid,suggesting that Ca2+. In contrast to the differential sensitivity to Ca2+ depletion, this pathway of PLD activation was suppressed by a staurosporin-sensitive protein kinase. Other selective protein kinase protein kinase C down-regulation by overnight culture of heC inhibitors (calphostinC, bisindolylmaleimide)inhibited PLD patocytes with a highconcentration of TPA uniformly deactivation by Ca", indicating a n additional role for protein creased Peth formation by 60-70%, irrespective of the agonist kinase C in the Ca2+-mediatedpathway. Protein kinase A po- used for stimulation (Fig. 1,left panel ). Even withTPA the PLD tentiated PLD activation by vasopressin, but this effect could activity was notcompletely suppressed, indicatingthat protein be accountedfor entirely by its ability to enhance the hormone- kinase C down-regulation was only partial. This may be accounted for by a variable extent of down-regulation of protein induced [Ca2+Ict increase in intacthepatocytes. kinase C isozymes; in our hands, overnightTPA pretreatment appeared to down-regulate predominantly the a-isoform of proEXPERIMENTALPROCEDURES tein kinase C (data not shown). PA formation in the protein Materials kinase C down-regulated cells was also partially suppressed H7 was obtained from Seikagaku America Inc.; A23187, ionomycin, with TPA or A23187 as the agonist (Fig. 1,right panel ), but not H89, okadaic acid, thapsigargin, W7, staurosporin, and the bisindolyl- with vasopressin. This observation agrees with ourfinding in maleimide GF 109203X were from Calbiochem;calphostin C and chelthe preceding paper (37) that thePA formation in response to erythrin were from LC Corporation; Williams E medium and fetal calf vasopressin is largely due to other reactions than PLD. These serum were from Life Technologies, Inc. Radiolabeled compounds were C is an importantcomponent purchased from Amersham Corp. Peth standardwas prepared from egg data suggest that protein kinase PC as described in (37). Other lipid standards were obtained from in all thesedifferent modes of activation of PLD. Avanti Lipids. Other chemicals were from Sigmaor Fisher. Although the experiments on protein kinase C down-regulated cells are compatible with a uniform involvement of proMethods tein kinaseC in PLD activation, theeffects of protein kinase C Cell Suspension Experiments-Hepatocytes were isolated from male inhibitors suggested a more differentiated picture. PretreatSprague-Dawley rats (20CL300 g) by collagenase perfusion and incu- ment with H7 (200 w) had a differential effect on the time bated as described in the accompanying paper (37). For the determination of PLD activity by Peth formation, a saturating ethanol concentra- course of PLD activity, depending on the agonist (Fig. 2). The tion (300 m) was added 5 min prior to the agonist. Other inhibitors or initial rate of Peth formation was inhibited by approximately 50% with bothTPA and with vasopressin, but with either agoactivators were added 5-30 min before the agonists. Reactions were stopped with HCIO, and phospholipids wereextracted with chloroform: nist the rapiddecline in Peth formation was prevented by the methanol, 1:2 (v:v) and analyzed as described elsewhere (37). inhibitor. Hence, after a 30-min reaction, there was no differExperiments with Cultured Hepatocytes-Freshly isolated hepato- ence between incubations with or without H7. This finding cytes were plated in 35-mm Petri dishes (9 x lo6 cellddish) in Williams suggests that the desensitization of the response to vasopressin E medium containing 10%fetal calf serum, 20 unitsfliter insulin, 100 or TPA, discussed in the accompanying paper (37), is dependent unitsfliter streptomycin, 100 unitsfliter penicillin, 13 mg/ml gentamycin, and 4 mg/ml dexamethasone and incubated in a humidified atmo- on protein kinase C activity. By contrast, the partial inhibition sphere with COdair (5%/95%)at 37 "C. After 2 h the medium was by H7 of Peth formation in response ATP to or with A23187 in changed and TPA (100 m) added. After overnight treatment with TPA, the presence of Ca2+ wasnot accompanied by a change in the cellswerewashedtwice and labeled with [3H]arachidonicacid (2.5 time course, indicating that the PLD inactivation with these pCi/dish) in serum-free Williams E medium for 2 h. Before agonist addition, cells were washed twice with Krebs-Ringer buffer, pH 7.4, agonists did not involve protein kinase C. Differential effects onPeth formation with different agonists containing 2 mM CaCl, and 0.2% bovine serum albumin and preincuwere alsofound with other protein kinase C inhibitors (Table I). bated at 37 "C for 10 min. Ethanol (150 m) was added 5 min prior to the agonist. Reactions were stopped by washing twice with ice-cold Staurosporin and its more selective analog, the bisindolylmabuffer, scraping cells into 1ml of buffer, and transferring the suspension leimide GF 109203X, act on the catalytic subunit of protein into 3.75 ml of chloroform:methanol, 1:2 (v:v). kinase C (39,40). CalphostinC binds covalently to theenzyme's Intracellular ea2+Measurements-Isolated hepatocytes were loaded regulatory subunit (41). All inhibitors (with the exception of with indo 1 or fluo 3 as described previously (38), stored on ice, and H7) wereused at a maximally effective concentration of 10 w. washed once with Krebs-Ringer bicarbonate buffer containing 2 m CaC1, just prior to use. Fluorescence intensity was measured at a wave- Staurosporin and bisindolylmaleimide inhibited Peth forma-

Mechanisms of Activation of Phospholipase D

85 1

0 control 0 Control

TPA-pretreated

@ TPA-pretreated

T T

Ethanol Ethanol Ethanol Ethanol

+

+

+

:hanol

+

r

thanol Et1hanol

+

+

TPA A23187 VP TPA A231 87 VP FIG.1. Effect of TPA pretreatment on agonist induced Peth and PA formation in cultured hepatocytes. Protein kinase C was down-regulated by treatment of the cells in culture with TPA (100 nM) for 16 h. Controlcells were treated with a corresponding volume of dimethyl sulfoxide. The cells were washed and labeled with [3H]arachidonic acid for 120 min before the addition of ethanol (150 m)and agonists. The concentrations of agonists used were100 nM TPA, 100 nM vasopressin, and 20 wA23187. The data are presented as mean S.E. from four separate experiments. Statistical significanceof differences: *, p < 0.05; **, p < 0.01.

*

tion with all three agonists, but were significantly more potent dium rapidly restored [Ca2+lc,, to restinglevels (0.1-0.2 w);at with TPA and vasopressin than with A23187. By contrast, cal- thesametime,intracellular hormone-sensitive Ca2+stores phostin C was less effective in inhibiting TPA-stimulated Peth were refilled, as indicated by the Ca2+elevation in response to formation (less than 30% inhibition), but inhibited Peth forma- vasopressin (Fig. 3A). However, when thapsigargin (1J ~ M )was tion in A23187-stimulated cells by 70%. Thus, although these added first, Ca2+ was prevented from entering the hormonefindings support the involvement of protein kinase C in PLD sensitive Ca2+ stores, andvasopressin (or other agonists) was activation with all these agonists, the mode of action of protein unable to elevate [Ca2+],,, (Fig. 3B ). The resulting steady state kinase C is not uniform. Only bisindolylmaleimide inhibited [Ca2+l,,, was now determined primarily by Ca2+fluxes across the modest formation of Peth occurring with ethanol itself; by the plasma membrane and was an almost linear function of the contrast, its analog staurosporin induced a small enhancement extracellular Ca2+ concentration ([Ca2+],,,) between 0 and 2 of basal Pethformation. Astimulation of PLD activity by stau- mM. Determination of [Ca2+l,, in indo 1-loaded cells in the rosporin was also reportedrecently in other cell types (42,43). presence of thapsigargin indicateda steady stateconcentration None of these inhibitorsaffected GTPyS-stimulated PLD activ- ranging from 0.05 to >2 p ~ when , [Ca2+lOut was variedfrom 0 to ity in isolated liver plasma membrane preparations (data not 4 mM (Fig. 4). Although vasopressin did not further change shown). [Ca2+l,, under these conditions (Fig. 3B), it did induce InsPs Of other protein kinase C inhibitors tested, chelerythrin (10 formation comparable with untreated cells when [Ca2+l,,, was p ~had ) no detectable effect on Peth formation with any of these restored to 100 p~ or above (not shown). agonists. Sphingosine alsodid not inhibit TPA- or vasopressinThis protocol was used to determine the effects of vasopresinduced PLD activity, but instead stimulated Peth formation sin and TPA on Peth formation at different [Ca2+Icfl, under slightly (20%) at high concentrations (100 w).At this level, conditionswhere theagoniststhemselves didnot affect sphingosine also increased the PA/l,S-DAG ratio by about 50%, [Ca2+lc,, (Fig. 5). An increase in [Ca2+l,,, from 0 to 4 mM (corin agreement with its reported inhibition of PA phosphohydro- responding to[Ca2+l,,, of 0.05-2 w) enhanced Peth formation lase (44). more than 10-fold in the absence of any other stimulus. HalfIncubating liver cells withsaturatingconcentrations of maximal stimulation was obtained at about 1.0 mM [Caz+]out, A23187 a t millimolar CaC1, concentrations floods the cytosol corresponding to a [Ca2+l,,,of approximately 1.0 p~ (left with unphysiologically high [Ca2+l and activates multiple other panel). TPA activated Peth formation even in the absence of lipases and proteases that could affect protein kinaseC activity. added CaCl,, and the addition of CaC12stimulated PethformaIn order toprobe a more direct role of [Ca2+],,, in PLD activa- tion only by about 40%. Bycontrast, Pethformation in response tion, we used thapsigargin, an inhibitor of the endoplasmic to vasopressin(100 n ~was ) entirely dependenton CaClz addireticular Ca2+ pump (45), to elevate [Ca2+],,, in a controlled tion. Both agents shifted theECS0for CaC12 slightly to theleft manner, independent of the additionof hormones or other ago- to approximately 0.5 mM, (corresponding to a [Ca2+], of 0.5 nists. After depletion of intracellular Ca2+ stores by incubation w),but also enhanced Peth formation at saturating Ca2+conwith EGTA, reincubation of the cells in a Ca2+-containing me- centrations by 40-80%. The 3H/'4C ratio of Peth formed was

852

Mechanisms of Activation of Phospholipase D

H

i

1

Irn

'6

I

c 9

53

:

f

a 0

lime (minutes)

+ n7 lime (minutes)

+ H7

FIG.2. Effect of H7 on the time course of agonist-induced Peth formation in intact hepatocytes.TPA (100I"), vasopressin (100I"), ) added where indicatedto cells pretreated with ethanol (300m)for 5 min. H7 (200 p a ) was added 10 min ATP (20 VM), and A23187 (20 p ~ were before the agonists. The time courses are presented as the percent of radioactivity in Peth compared with the maximal responsefor each agonist in cells prelabeled with [3Hlarachidonic acid. The bar graphsto the right presents the effect of H7 on Peth formed (dpm/106cells) after 2.5 min of stimulation with agonists. The results in the bar graphs are presentedas mean t S.E. calculated from six separate experiments. H7 did not affect the specific radioactivity of the Peth formed. Statistical significance of differences: ***, p < 0.001;**, p < 0.01. TABLE I Effect of protein kinase C inhibitors on phosphatidylethanol formation induced by different agonists in intact hepatocytes [3HlArachidonate-labeledcells were preincubatedfor 30 min with the inhibitors indicated, prior to stimulation with vasopressin (100 I"), TPA (100 I"), or A23187 (20 p ~ in) the presence of ethanol (300m). Control incubations received ethanol alone. All inhibitors were added from a stock solutionin dimethyl sulfoxide toa final concentrationof 10 p ~ Reaction . time 5 min. Results are mean 2 S.E. for three or four individual experiments. Inhibitor

EtOH

r3H1Peth formation Vasopressin TPA + EtOH +EtOH % of

No inhibitor Calphostin C Staurosporin Bisindolylmaleimide

0.09 t 0.04 0.10-c 0.03 0.37t 0.06 0.05 2 0.10

A23187

+ EtOH

PHlPC 0.92 t 0.11 2.28 t 0.31

1.15t 0.17 0.66 t 0.22 0.472 0.11 0.132 0.04

0.67t 0.11 0.69t 0.14 0.27 0.07 1.312 0.22 0.112 0.04 0.68* 0.15

unaffected by [Ca2+lor by the presence of agonists (not shown), indicating that the nature of the PLD substrate was not affected. In parallel experiments, 0.5 mM CaC1, maximally restored vasopressin-induced InsP3 formation in thapsigargintreated cells (not shown). Thus, the [Ca2+lcfirequirement for vasopressin stimulation of polyphosphoinositide-specific PLC is less than for PLD. The effect of staurosporin (10p ~ on) Peth formationin thapsigargin-treated cells was investigated in the same experiments (Fig. 5). Staurosporin inhibited PLD activation by TPA (middle panel)and by vasopressin (right panel 1, probably due to inhibition of protein kinase C. By contrast, staurosporin markedly enhanced the Ca2+-dependent Peth formation (left panel ). Staurosporin decreased the ECS0for added CaC12 from about 1 mM to