Activation of Protein Kinase C by Oleic Acid

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY by The American Society for Biochemistry and Molecular Biology, Inc Vol. 267, No. 6, Issue of February 25, pp Printed in U. S. A. Activation of Protein Kinase C by Oleic Acid DETERMINATION AND ANALYSIS OF INHIBITION BY DETERGENT MICELLES AND PHYSIOLOGIC MEMBRANES: REQUIREMENT FOR FREE OLEATE* (Received for publication, August 21, 1991) Wasiuddin A. Khan, Gerard C. Blobe, and Yusuf A. Hannunt From the Departments of Medicine and Cell Biology, Duke University Medical Center, Durham, North Carolina Sodium oleate is able to activate soluble protein ki- tioned into platelet membranes with the majority of nase C (Murakami, K., Chan, S. Y., and Routtenberg, arachidonate (70%) remaining in the cytosolic frac- A. (1986) J. Biol. Chem. 261, ) but is tion. These studies strongly suggest that only free fatty unable to activate membrane-bound enzyme (El Touny, acids are able to activate soluble protein kinase C. It is S., Khan, W., and Hannun, Y. (1990) J. Biol. Chem. proposed that physiologic activation of protein kinase 265, ). Because physiologic interactions C may occur in two compartments: soluble protein of fatty acids with protein kinase C occur in the pres- kinase C would be a primary target for cis-unsaturated ence of membranes, the following studies were con- fatty acids, especially arachidonate, while membraneducted to evaluate the effects of surfaces (detergent associated protein kinase C is a target for activation micelles or platelet membranes) on the activation of by diacylglycerols. protein kinase C by oleate. At concentrations at or above the critical micellar concentration (CMC) of Triton X-100, oleate was present primarily in Triton X- 100/oleate-mixed micelles, as determined by gel per- Protein kinase C, a family of closely related isoenzymes, meation chromatography and equilibrium dialysis plays an important role in signal transduction and cell regubinding studies. At concentrations slightly below the lation (1, 2). Enzyme activity is dependent on phospholipids CMC for Triton X-100, the presence of oleate caused with some members of the family being Ca2+ dependent and the formation of a limited number of mixed micelles. others calcium-independent (1). Physiologically, the enzyme Studies of the dose-dependent activation of purified is activated by diacylglycerols which are generated from the platelet protein kinase C by sodium oleate in the presence of different concentrations of Triton X- 100 indicated that only unbound oleate was able to activate protein kinase C. Platelet protein kinase C was resolved into two major isoenzymes (types I1 (B) and I11 (a)) which displayed nearly identical interaction with oleate. Activation of protein kinase C by oleate in a physi- ologic setting employing platelet substrates endog- and enous platelet protein kinase C was investigated. Oleate potently activated protein kinase C in the cytosolic compartment. In platelet homogenates as well as in a reconstituted platelet cytosol and membrane system, the dose dependence of protein kinase C on oleate showed a significant shift to the right. Approximately 30% of oleate was associated with platelet cytosol and 70% was associated with platelet membranes. Partitioning of oleate into the two platelet compartments showed little change with ph, temperature, or duration of incubation. When corrected for free oleate concentration, activation of protein kinase C by oleate showed identical dose dependence in cytosol and homogenate. Arachidonate, a potential physiologic activator of protein kinase C, showed similar behavior as oleate although only 30% of arachidonate parti- * This work was supported in part by National Institutes of Health Grant HL and by National Cancer Institute Grant CA The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked aduertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Established investigator of the American Heart Association and Mallinckrodt Scholar. To whom correspondence should he addressed Dept. of Medicine, Div. of Hematology/Oncology, P. 0. Box 3355, Duke University Medical Center, Durham, NC breakdown of membrane phospholipids upon activation of phospholipases C with different specificities for phospholipid species (1, 3,4). Protein kinase C is also activated by phorhol esters (5) and related tumor promoters, and it serves as the main intracellular receptor for active phorbol esters (6). Protein kinase C is also activated in vitro by arachidonic, oleic, and other cis-unsaturated fatty acids (7-9). This activation does not require phospholipids (8, 9) but is dependent on calcium for the LY and p isoenzymes of protein kinase C (9, 10). Activation of protein kinase Cy by arachidonic acid appears to be mostly independent of calcium (10,ll) although other studies show a stimulatory effect of calcium (12). A number of features distinguish the mechanism of activation of protein kinase C by fatty acids from activation by diacylglycerol (DAG). For example, although DAG can inhibit phorbol ester binding to protein kinase C, oleic acid is unable to do so (9). Arachidonic acid was found to inhibit phorbol binding by a mechanism distinct from that of DAG (13). Both studies suggest that fatty acids do not interact with protein kinase C at the phorbol ester/dag-binding site. DAG can also induce autophosphorylation of protein kinase C while oleate cannot. Furthermore, activation of protein kinase C by sodium oleate is less sensitive to inhibition by sphingosine and NaCl but is equally susceptible to inhibition by H7 when compared to activation by PS/DAG (9). Finally, whereas DAG requires membrane association of protein kinase C as an essential step in the activation mechanism, oleate appears to activate preferentially soluble protein kinase C in the absence of membranes or detergent/lipid mixed micelles (8, 9). The abbreviations used are: DAG, sn-1,2-diacylglycerol; PS, phosphatidylserine; DiCIR:l, sn-1,2-dioleoylglycerol; CMC, critical micellar concentration; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid.

2 3606 Activation of Protein Kinase The observation that sodium oleate is unable to activate protein kinase C in the presence of Triton X-lOO/PS mixed micelles (9) raises important questions as to the proposed physiologic relevance of activation of protein kinase C by cisunsaturated fatty acids (7,8): does the presence of physiologic membranes negate the effects of fatty acids on protein kinase C? The following study was undertaken to clarify the role of membranes in modulating the ability of sodium oleate to activate protein kinase C. Two approaches were taken. First, the effects of detergent micelles on activation of protein kinase C by sodium oleate were investigated in uitro using purified protein kinase C and exogenous histone substrate. Next, the physiologic significance of these interactions were investigated by utilizing membranes from human platelets and endogenous substrates for platelet protein kinase C. We find that oleate in the form of Triton X-lOO/oleate mixed micelles is unable to activate protein kinase C. In uitro activation of protein kinase C requires the presence of soluble enzyme and free oleate. In platelets, up to 70% of oleate appears to partition into membranes with this pool of oleate being unavailable for activation of protein kinase C. Only the 30% remaining soluble oleate is competent to activate protein kinase C. The implications of these studies on regulation of protein kinase C by fatty acids is discussed. EXPERIMENTAL PROCEDURES Materials [y 32P]ATP was from Du Pont-New England Nuclear. [9,10-3H] Oleic acid and [5,6,8,9,11,12,14,15-3H]arachidonic acid were from Amersham Corp. Oleic acid, phorbol myristate acetate, Triton X-100, and histone type 111s were from Sigma. 1,2-Dioleoyl-sn-glycerol-3- phosphoserine was from Avanti Polar Lipids, Inc. 8-Octyl glucoside, deoxycholate, and Zwittergent 3-14were purchased from Behring Diagnostics. Methods Purification of Protein Kinase C-Protein kinase C was purified from platelet cytosol to homogeneity as described (14) to a specific activity of X lo3 nmol/min/mg. Separation of Protein Kinase C Isoenzymes-Platelet protein kinase C isoenzymes were separated by hydroxyapatite chromatography as described (15). Two peaks were resolved corresponding to brain types I1 (8) and I11 (a). Using isoenzyme-specific antibodies, human platelets were found to contain protein kinase C Dl, p2, and a. Protein kinase C y, 6, c, and { were not detected. Upon chromatography on hydroxyappatite, type I1 was found to contain protein kinase C pl and p2 while type I11 contained protein kinase C 01.' Assay for Protein Kinase C-Protein kinase C activity was assayed using Triton X-lOO/PS/DAG mixed micelles as previously described (16, 17); maximal activity was considered as that measured in the presence of 10 mol% PS and 2 mol% DiC18:1. When sodium oleate was used as an activator, the reaction assay contained 400 pm calcium, 100 pm EGTA, 200 pg/ml histone, 10 mm magnesium, 10 pm ATP, and 20 mm Tris-HC1, ph 7.5. Oleic acid was neutralized in ethanol with sodium hydroxide and further diluted to appropriate concentrations in aqueous solutions. These assays were done either in the absence or in the presence of the indicated concentrations of Triton X-100 f PS as described in the figure legends of the specific experiments. Chromatography of Triton X-lOO/Oleate Mixed Micelle-The determination of existence of Triton X-lOO/oleate mixed micelles was performed by chromatography on Sephadex G-25. For these experiments, columns of 1 ml of gel were equilibrated in buffer containing 20 mm Tris-HC1, ph 7.5, and 0.012% Triton X-100. Samples (10011) containing Triton X-100 at the indicated concentrations in the presence of50 pm sodium oleate, specific activity (1 pci/nmol) were loaded and diluted in equilibration buffer. The voidvolumewas collected in 1 ml of equilibration buffer; 2 fractions of 0.5 ml each, and the retained amount was collected in 5 ml. In the absence of W. A. Khan, G.C. Blobe, and Y. A. Hannun, manuscript in preparation. C by Oleic Acid Triton X-100, 11% oleate was eluted in the void volume. This was subtracted from the amount of sodium oleate that eluted in the void volume in the presence of the indicated concentrations of Triton X Equilibrium Dialysis Studies-To determine equilibrium partitioning of sodium oleate into Triton X-100 micelles, equilibrium dialysis experiments were performed in 20 mm Tris-HC1, ph 7.5. The dialysis bag contained the indicated concentrations of Triton X-100. For these experiments, sodium oleate was added either inside the dialysis bag or outside, and equilibrium was achieved between h. After 72 h, 1-ml aliquots were counted in an LKB scintillation counter. The concentrations of sodium oleate in micelles and free sodium oleate were calculated as follows [Oleatei,l~i,,~~,. = [Oleatei,]~~~ - [Oleatei,]~,. and [Oleatei,]~,, = [Oleate,,tl~ree where [Oleatei.]Miceell.. is the concentration of oleate in Triton X-100 micelles; [Oleatei,,]~ot is the concentration of oleate in the dialysis chamber at equilibrium; [ Ole~te~~]~~~ and [Oleateo&, are the free concentrations of oleate inside and outside the dialysis chamber at equilibrium, respectively. ole ate out]^, and [Oleatei,]Tot were measured at equilibrium and [Oleatein]Mieelles was calculated. Platelet Preparation-Fresh blood (from human donors) was diluted 9:l with ACD buffer (85 mm sodium citrate, 111 mm dextrose, and 71 mm citric acid), and platelets were pelleted at 200 X g for 20 min at room temperature with brakes off. Supernatant (platelet rich plasma) was drawn off and 5 ng/ml PGI, was added. The plateletrich plasma was pelleted down at 800 X g for 15 min at room temperature with brakes off. Supernatant was drawn off and discarded. The platelet pellet was resuspended in Tyrodes buffer (134 mm NaCl, 12 mm NaHC03, 2.9 mm KCl, 0.36 pm NaH2PO4.H20, 5 mm HEPES, 5 mm glucose) with 300 ng/ml PG12, and an aliquot was counted for total number of platelets. Preparation of Platelet Homogenate, Cytosol, and Membranes- Platelets in Tyrode's buffer with PGI, were pelleted down at 800 X g for 15 min at room temperature with brakes off and resuspended in homogenizing buffer (20 mm Tris-HC1, ph 7.5, 2 mm EDTA, ph 7.4, 10 mm EGTA, ph 7.4, 250 mm Sucrose, 1 mm phenylmethylsulfonyl fluoride, 0.02% leupeptin) at a concentration of 2.9 X lo9 platelets/ ml. Samples were sonicated on ice for 15 s X 3. Cytosol and membranes were then obtained from homogenate by ultracentrifugation at 100,000 X g with the supernatant being the cytosol. The pellet, consisting of platelet membranes, was resuspended in an equal volume of homogenizing buffer. For reconstituted systems, membranes and cytosol were mixed at equal volumes, vortexed vigorously, and allowed to equilibrate for 10 min. 40-kDa Substrate Phosphorylation in Platelets pl of platelet cytosol, membranes, membranes and cytosol, or homogenate were used as a source of protein kinase C and 40-kDa substrate. The standard reaction was with 20 mol % PS, 2 mol % DiCIk1 in 0.3% Triton X-100 (w/v), 10 mm Me, 100 pm ATP, and 20 mm Tris- HC1, ph 7.5. Calcium was varied from 50 to 500nM, and sodium oleate was used at concentrations of5-150 p~ as indicated in the figure legends. Reactions were performed in microcentrifuge tubes in a 100-pl reaction volume. Reactions were incubated for 10 min at room temperature, then stopped with 100 pl of 2 X sample buffer. The samples were boiled for 5 min and 100 pl was loaded on 12.5% sodium dodecyl sulfate-polyacylamide gels. The gels were stained and fixed, autoradiography was performed, and the 40-kDa substrate band was excised and counted in an LKB scintillation counter in 10 ml of scintillation fluid. Partitioning of Oleate in Platelet Membranes-Various concentra- tions of sodium oleate (5-500 p ~ along ) with [9,10-3H]sodium oleate at a specific activity of 0.01pCi/pMwere incubated with intact platelets, platelet homogenate, or a mix of cytosol and membrane (500 pl) for 10 min at room temperature in a microfuge tube. Samples were then pelleted down in a microfuge at 14,000 X g for 10 min. The supernatant was drawn off and counted, and the pellet was resuspended in an equal volume of homogenizing buffer and counted in an LKB scintillation counter in 10 ml of scintillation fluid. An aliquot of each was extracted by the method of Bligh and Dyer and applied to a thin layer chromatography plate in the solvent system EtOAc/ 2,2,4-trimethylpentane/acetic acid/water (90:50:20:100) and autoradiography was performed.

3 RESULTS Activation of Protein Kinase C by Free But Not by Micellar Oleate-Zn vitro, protein kinase C is activated by cis-unsaturated fatty acids in the absence of membranes or additional phospholipid cofactors (8, 9). The a and B isoenzymes of protein kinase C require Caz+ for this activation which occurs at concentrations of oleate significantly lower than its critical micellar concentration (9). Moreover, we have shown that oleate is unable to activate protein kinase C in the presence of Triton X-1OO/PS mixed micelles (9) (Fig. 1). It was therefore concluded that sodium oleate activates soluble protein kinase C and does not activate micelle-bound enzyme. These studies, however, did not evaluate the role of partitioning of oleate into micelles and the relative contribution of free versus micellar oleate to the activation of protein kinase C. The inability of oleate to activate protein kinase C in the presence of Triton X-1OO/PS mixed micelles could result from protein kinase C associating with membrane and therefore being inaccessible for activation by oleate; for example, due to competition between PS and oleate at a putative lipidbinding site. Alternatively, activation of protein kinase C may have a specific requirement for monomeric oleate. In this case, the presence of detergent micelles serves to sequester oleate in micellar form where it is unable to activate protein kinase C. To evaluate the activation of protein kinase C by micellar oleate, the dependence of protein kinase C activity on oleate concentration was measured in the presence of calcium and in the presence or absence of Triton X-100 micelles (Fig. 1). In the absence of detergent, protein kinase C activity was dependent on the concentration of sodium oleate with activity Na M.am I TX-lW 0 too ,.,. I. I Na Oleate (MM) I C I ZOO Ne Oleate (MM) FIG. 1. Effects of Triton X-100fPS mixed micelles and Triton X- 100 micelles on activation of protein kinase C by oleate. Protein kinase C activity was determined in the presence of the indicated concentrations of sodium oleate in the absence of detergent, in the presence of Triton X-100 micelles, or in the presence of Triton X-100 mixed micelles containing 10 mol % PS. These results are representative of three to five similar experiments. A, total purified protein kianse C; B, type I11 (a) protein kinase C; C, type I1 (p) protein kinase C. Activation of Protein Kinase C by Oleic Acid 3607 observed at low pm concentrations (Fig. 1). In the presence of 0.3% (w/v) Triton X-100 (a concentration significantly higher than CMC of %), sodium oleate was unable to activate protein kinase C. Under these conditions, protein kinase C remains unassociated with the micelles (16) while oleate partitions into micelles (see below). These results suggest that formation of Triton X-lOO/oleate mixed micelles may cause surface dilution of monomeric sodium oleate, thus preventing oleate from activating protein kinase C. Since the above studies were determined with total purified protein kinase C it became important to determine the interaction of oleate with individual isoenzymes. Platelet protein kinase C was resolved into two major peaks on hydroxyapatite chromatography, corresponding to types I1 and I11 (or p and a, respectively). Sodium oleate showed similar kinetics in activating types I1 and 111 (Fig. I, B and C). Oleate was unable to activate either isoenzyme in the presence of Triton X-100 (Fig. 1, B and C). Incidently, it was noted that type I11 displayed significant activation by Triton X-lOO/PS mixed micelles alone as compared to type 11; the significance of this observation is unknown. The role of mixed micelle formation was further investigated by evaluating the dose response for sodium oleate activation of protein kinase C at different concentrations of Triton X-100 below and above the CMC of Triton X-100. Fig. 2 shows that concentrations of Triton X-100 above CMC (0.03 and 0.3%) resulted in abrogation of activation of protein kinase C by sodium oleate. However, at a concentration of Triton X-100 well below CMC (0.003%) only minor effects on activation of protein kinase C by sodium oleate were observed with moderate stimulation of enzyme activity at the lower concentrations of oleate ( p ~ Fig., 2 A ) but not at higher concentrations ( PM, Fig. 24). Interestingly, at concentrations slightly below or at CMC3 (0.06 and 0.012% respectively) Triton X-100 showed an inhibitory effect on sodium oleate activation of protein kinase C at low concentrations of sodium oleate (Fig. 2A). Higher concentrations of sodium oleate appeared to activate protein kinase C even in the presence of these low concentrations of Triton X-100 (Fig. 24). This shift to the right in the dose response activation of protein kinase C by sodium oleate in the presence of Triton X-100 at concentrations around CMC may indicate the formation of a limiting number of Triton X-lOO/oleate mixed micelles (see below). Oleate also interacted with types I1 and I11 protein kinase C isoenzymes with similar kinetics in the absence or presence of different concentrations of Triton X-100 (Fig. 2, B and C). The effects of oleate and detergent on other isoenzymes of protein kinase C not present in platelets were not evaluated in this study. In order to establish the existence of Triton X-lOO/oleate mixed micelles, two approaches were undertaken. In the first approach, rapid gel filtration of Triton X-lOO/oleate mixed micelles was effected using Sephadex G-25 gel permeation chromatography. For these experiments, tritium-labeled sodium oleate at 50 pm was added to solutions of Triton X-100 at different concentrations. The mixtures were applied to a Sephadex G-25 column equilibrated at 0.012% Triton X-100; a concentration slightly below CMC used in order to minimize breakdown of micelles. Radioactivity in the void volume (where micelles elute) indicates the percent of the added label that has associated with Triton X-100 mixed micelles. Fig. 3A shows that at 0.3% Triton X-100, the majority of sodium oleate (85%) elutes with micellar Triton X-100. Interestingly, concentrations of Triton X-100 of and 0.012% showed The CMC of Triton X-100 in the absence of lipids is approximately 0.015%, but in the presence of lipids the CMC becomes lower.

4 3608 Activation of Protein Kinase 6 X E c - > - > 40-0 * a a Na Oleate (FM) C by Oleic Acid B 1001 TritonX-100 (% w/v) 04.,.,. I. I.,., c Na Oleata Na Oleata FIG. 2. Effects of concentrations of Triton X-100 on activation of protein kinase C by oleate. The dose dependence of protein kinase C activation by sodium oleate was determined in the absence or in the presence of the indicated concentrations of Triton X-100. These results are representative of three similar experiments. A, total purified protein kinase C. 0, 0% Triton X-100; D, 0.003% Triton X-100; 0, 0.006% Triton X-100; A, 0.012% Triton X-100; A, 0.03% Triton X-100; 0, 0.3% Triton X-100. B, type I11 (a) protein kinase C. 0, 0% Triton X-100; T, 0,006% Triton X-100; A, 0.012% Triton X-100; +, 0.03% Triton X-100; C, type I1 (0) protein kinase C. 0, 0% Triton X-100; 0, 0.006% Triton X-100; A, 0.012% Triton X-100; e, 0.03% Triton X-100. a definite ability to enrich the void volume with sodium oleate (Fig. 3A). This indicates the presence of sodium oleate in Triton X-100 mixed micelles even though the absolute number of micelles under these conditions is significantly less than those present at higher concentrations of Triton X It is important to note that the conditions for gel permeation chromatography do not allow for equilibrium considerations: as the micelles chromatograph through the column, dilution and breakdown of Triton X-100 micelles occurs. Therefore, the percent oleate in micelles determined under these conditions should be considered as an underestimate of micellar oleate. These considerations are of greater importance for concentrations of Triton X-100 at CMC since any dilution would result in significant loss of micelles. In an equilibrium approach, the ability of tritium-labeled sodium oleate to partition with Triton X-100 was investigated using equilibrium dialysis. For these experiments, tritiumlabeled oleate was added at 50 FM into a dialysis bag in the presence or absence of Triton X-100, and the mixture was equilibrated against 20 mm Tris-HC1, ph 7.5, buffer for 72 h. Similar results were obtained when oleate was added to the From A,, values, it is estimated that the number of Triton X- 100 micelles in the void volume when and 0.012% was applied to the column is 5.9% and 18.4% of the number of micelles with 0.3% Triton X I. I I 0.30 Triton X-100 (% wlv) FIG. 3. Partitioning of oleate into Triton X-100 micelles. A, coelution of oleate with Triton X-100 micelles on gel permeation chromatography. Tritium-labeled sodium oleate at 50 PM was applied to Sephadex G-25 column chromatography in the presence of the indicated concentrations of Triton X-100. The percent of oleate eluting in the void volume with Triton X-100 is plotted versus the applied concentration of Triton X-100. Similar results were obtained in three different experiments. B, Determination of micelle-bound oleate by equilibrium dialysis. Equilibrium dialysis was performed as described under Experimental Procedures. The oleate bound to micelles (as percent of total oleate) is plotted against the concentration of Triton X-100. These are representative of four different experiments. TABLE I Effects of detergents on protein kinase C activation by oleate Protein kinase C activation by 0.5 Conc. mm Na-oleate (% V, with PSf DiCld Detergent <CMC XMC <CMC XMC None 63.8 Triton X % 0.03 % P-Octyl glucoside 5 mm 50 mm Deoxycholate 0.5 mm 5 mm Zwitteraent mm 5 mm Protein kinase C activity was determined in the presence of 0.5 mm oleate in the absence or presence of the above detergents. outside buffer thus assuring equilibrium and allowing the calculation of micelle-associated oleate as described under Experimental Procedures. The majority of oleate (>go%) was found in association with Triton X-100 micelles when Triton X-100 was present at concentrations above CMC (Fig. 3B). Concentrations of Triton X-100 around CMC caused significant (40-60%) partitioning of oleate (Fig. 3B). As compared with gel filtration, equilibrium dialysis studies showed that a greater fraction of oleate was associated with micelles at any given concentration of Triton X-100 (compare Fig. 3, A and B). These studies establish the existence of Triton X- 100/oleate mixed micelles even at concentrations of Triton X-100 at CMC. Moreover, significant, and nearly total, partitioning of oleate into micelles was observed. The effects of micellization of oleate on protein kinase C activity were not restricted to Triton X-100. Other detergents had similar effects, as illustrated in Table I. Concentrations below CMC did not abrogate the effects of oleate while con-

5 - E sp 100 C -3 & 8o k 0 c 60 $ 40 c PI Cytosol Cytosol & Activation of Protein Kinase C by Oleic Acid 3609 Membrane FIG. 4. Activation of platelet protein kinase C by oleate in different platelet compartments. 40-kDa phosphorylation in different platelet compartments was determined as described under Experimental Procedures. Degree of phosphorylation in the reconstituted cytosol and membrane is expressed as percent of phosphorylation in cytosol. These results are representative of six different experiments. 0, calcium/ps/dag; El, sodium oleate. centrations above CMC caused total inhibition of activation of protein kinase C by oleate (Table I). These studies clearly demonstrate that micellar sodium oleate is unable to activate protein kinase C. Therefore, activation of protein kinase C by fatty acids shows a definite requirement for both free enzyme and monomeric fatty acid? Effects of Membrane Partitioning of Oleate on Activation of Protein Kinase C-The above in vitro studies suggest that physiologic regulation of protein kinase C by cis-unsaturated fatty acids requires the presence of soluble protein kinase C as well as membrane-free fatty acid. Because the presence of physiologic membranes may result in significant partitioning of fatty acids it became important to investigate the effects of cell membranes on protein kinase C activation by oleate. As a model system, the ability of oleate to activate endogenous platelet protein kinase C in the presence of platelet protein substrates was established and compared with the effects of DAG. These studies were conducted using platelet cytosol, membranes, total homogenates, and reconstituted mixtures of cytosol and membranes. In platelet cytosol, activation of protein kinase C by DiClel in the presence of PS and 100 p~ calcium resulted in significant phosphorylation of the well-characterized 40-kDa substrate (18). Sodium oleate at a concentration of 50 pm also caused phosphorylation of the same 40-kDa substrate in the presence of 50 nm calcium. In platelet membranes, DiC,,:. caused minimal phosphorylation of 40-kDa substrate6 while sodium oleate had no detectable effect on 40-kDa phosphorylation (data not shown). When platelet cytosol and membranes were reconstituted, DiClel caused significant phosphorylation of the 40-kDa substrate whereas sodium oleate lost its ability to cause protein kinase C-induced phosphorylation of 40-kDa substrate. When the amount of phosphorylation was quantitated, it was found that, in the reconstituted system, DiCIeI was able to achieve 80% phosphorylation of the 40-kDa substrate as compared with cytosol alone, while sodium oleate resulted in less than 30% phosphorylation of 40-kDa substrate (Fig. 4). Nearly identical results were obtained when the experiments were performed in total platelet homogenates (data not shown). The dose dependence of protein kinase C activation by sodium oleate in the different compartments was evaluated next. Fig. 5A shows that sodium oleate activates soluble protein kinase C in a dose-dependent manner in the presence of 50 nm calcium. Sodium oleate had little effect on membrane-bound protein kinase C. In total platelet homogenates, sodium oleate also activated protein kinase C in a dosedependent manner. However, the dose dependence showed a definite shift to the right (Fig. 5A) suggesting a significant decrease in the potency of sodium oleate in activating protein kinase C. Because the physiologic range of intracellular calcium in platelets varies from a resting level of 50 nm to an activated level of 500 nm (19), these studies were also done at the higher calcium concentration, and very similar results were obtained (Fig. 5B). These studies strongly suggest that the presence of platelet membranes modulates the ability of sodium oleate to activate protein kinase C. To rule out metabolism of oleate as the reason for the diminished potency, platelet homogenates were incubated with tritium-labeled sodium oleate at 50 pm for variable durations, and then platelet lipids were extracted and separated on thin layer chromatography. Following extraction, radioactivity was retained in the organic phase (greater than 97%). Autoradiography of the thin layer chromatography plates demonstrated that sodium oleate remained intact for over 60 min of incubation with platelet homogenates (data not shown). Similarly, no significant metabolism or incorporation of oleate could be detected when examined in either Oleate (@I) 6There remains aformal possibility where protein kinase C is micelle bound and oleate is free. These conditions could not be established in vitro since the requirements for membrane-association of protein kinase C (micelles or liposomes) result in predominant surface partitioning of oleate. Platelet membranes contain approximately 30% of total protein kinase C activity, therefore, the low level of phosphorylation appears to indicate the selective localization of 40-kDa substrate in platelet cytosol Oleate (pm) FIG. 5. Concentration dependence of IO-kDa phosphorylation on sodium oleate in cytosol, membranes, and homogenate in the presence of 50 nm calcium (A) or 500 nm calcium (B). These results are representative of three experiments. 0, cytosol; 0, membrane; A, homogenate.

6 3610 Activation of Protein Kinase C by Oleic Acid cytosol or membrane fractions separately (data not shown). The lack of significant metabolism of oleate and the in vitro studies strongly suggest that the diminished potency of sodium oleate in platelet homogenates is due to membrane partitioning of oleate. If that is the case, then sodium oleate should have a finite solubility in the presence of platelet membranes, thus allowing sufficient monomeric concentrations to cause activation of protein kinase C. The soluble oleate would be a fraction of the total oleate, and this would account for the shift to the right in the dose response. To test this hypothesis, the partitioning of sodium oleate into membranes uersus soluble fractions was determined first. Tritium-labeled oleate at 50 PM was added to either platelet homogenates or reconstituted platelet membrane and cytosolic fractions and incubated for 5 min followed by separation of cytosolic and membrane fractions. Approximately 70% of sodium oleate was detected in the membrane fraction and 30% was found in the cytosolic fraction (Fig. 6). Similar results were obtained when oleate was added to intact platelets before homogenization. Again 30% of oleate was recovered in cytosol and 70% in the membrane fraction (Fig. 6). When the dose response of protein kinase C activation by sodium oleate in homogenate was corrected for the free sodium oleate (30% of total), nearly identical activation curves were observed for platelet cytosol and homogenate at 50 nm Ca2+ (Fig. 7A) and 500 nm Ca2' (Fig. 7B). These experiments suggest that partitioning of oleate into membranes may be the primary determinant of free oleate concentration and hence the effective oleate concentration in activating protein kinase C. That is, protein kinase C activity is primarily dependent on free oleate concentrations, and the presence of membranes serves to decrease the effective free oleate. The ability of different parameters to modulate the partitioning of oleate between cytosolic and membrane fractions of platelets were evaluated next. There was no change in membrane-associated oleate with increasing durations of preincubation of platelet homogenates with tritium-labeled oleate (data not shown) suggesting rapid attainment of equilibrium between the two compartments. Also, there was little difference in partitioning at different ph over the physiologic range or with different temperatures of preincubation (data not shown). At higher temperatures (50-70 "C), oleate showed greater membrane partitioning probably due to increased fluidity of membranes. To rule out denaturation of potential oleate-binding proteins from cytosol as a cause for decreased free oleate at high temperatures, platelet homogenates were heated to 70 "C then cooled down prior to incubation with Membrane &Cytosol Cytosol Membrane FIG. 6. Partitioning of sodium oleate between cytosol and membrane fractions of platelets. Tritium-labeled sodium oleate at 50 GM was incubated with either intact platelets, platelet homogenates, or reconstituted membrane and cytosolic fractions. Platelet membranes were separated from cytosolic fractions as described under "Experimental Procedures." The amount of oleate associated with the pelleted membrane fractions or with the supernatant cytosolic fraction are expressed as percent of total oleate Cytosol - Homogenate W 150 Free Oleate (pm) l Z 0 V loo A A Cytoml 20 - Homogenate W 150 Free Oleate (pm) FIG. 7. Induction of 40-kDa phosphorylation by oleate. Results shown in Fig. 5 are replotted to show 40-kDa phosphorylation (as percent of maximum) as a function of free oleate concentration. For the cytosolic fraction, this is taken as the total concentration of oleate, and for the homogenate fraction, this is derived from the total oleate multiplied by the percent oleate that is not associated with membranes (approximately 30%). These studies were performed at 50 nm calcium (A) and 500 nm calcium (B)., < /,,,, Oleate tpm) FIG. 8. Effect of oleate concentration on degree of partitioning of oleate between cytosol and membrane. The total mass of oleate present in cytosol or membrane is plotted against total oleate concentration. tritium-labeled sodium oleate. Under these conditions, there was no change in the partitioning of sodium oleate between membrane and cytosolic compartments. There was a slight dependence of partitioning on the concentration of oleate with a modest drop in cytosolic oleate at increasing concentrations, but this did not show saturation kinetics for the total amount of solubilized sodium oleate (Fig. 8). Therefore, it appears that sodium oleate partitions rapidly between membrane and cytosolic fractions with approximately 30% of sodium oleate remaining in a free form that is capable of activating protein kinase C. DISCUSSION The inability of oleate and other cis-unsaturated fatty acids to activate membrane-bound protein kinase C raises two important issues. First, what are the roles of membrane and

7 Activation of Protein Kinase C by Oleic Acid 3611 FIG. 9. Schematic illustration of compartmentalization of protein kinase C activation by lipid second messengers. Agonist interaction with cell membrane receptors may result in activation of phospholipases C (PL-C) or phospholipases A2 (PL-A,), thus causing the hydrolysis of membrane phospholipids (PL) and the subsequent generation of DAG and free fatty acids (FFA), respectively. Free fatty acids undergo a rapid distribution between membrane and cytosol (1) so that in the case of arachidonic acid 30% is in the membrane and 70% is soluble (Table 11). The soluble free fatty acid is then able to activate soluble protein kinase C (stippled) (2). Inactive soluble protein kinase C (shown with the psuedosubstrate site interacting with the protein substrate site) is a target for activation by free fatty acid (3) and is presumed to be in equilibrium with membrane-bound protein kinase C (4). Membrane binding of protein kinase C requires PS and calcium. Under these conditions the enzyme is able to interact with substrate although with very low catalytic efficiency (26) and is, therefore, shown with the psuedosubstrate site detached from the protein substrate-binding site on the catalytic domain but in an inactive form (white-filled). Membrane-bound protein kinase C is then catalytically activated (stippled) by membrane DAG (5). TABLE I1 Partitioning of arachidonate in platelet membranes Partitioning of oleate and arachidonate between platelet membranes and cytosol was determined as described under Experimental Procedures. Cvtosol Membrane % of total f S.D. Oleate 30.2 & 2.9 % 69.7 t (50 Arachidonate (50 KM f 1.4 % % surface interactions in modulating activation of protein kinase C by fatty acids? Second, in order to discern physiologic regulation of protein kinase C by fatty acids, it becomes important to determine the membrane partitioning of fatty acids and the ability of fatty acids to activate protein kinase C in different cellular compartments (soluble versus membrane-bound enzyme). Little data is available on the partitioning of oleate in membranes. Recently, the equilibrium distribution of oleic acid between emulsions and phospholipid bilayers was examined. It was found that the affinity of oleic acid for emulsions or bilayers was similar (20). The partitioning of lower concentrations of fatty acids between membranes and cytosolic fractions and the role of these processes on physiologic activities of fatty acids have not been addressed. The results from this study clearly demonstrate the formation of Triton X-lOO/oleate mixed micelles and that at concentrations above CMC, most of the oleate partitions into micelles with small amounts of free oleate. Micellar oleate appears to be unable to activate protein kinase C. Only when the capacity of micelles to retain oleate is exceeded does free oleate accumulate sufficiently to activate soluble enzyme. Therefore, in vitro, the data appear to favor activation of protein kinase C by oleate only when both oleate and enzyme are monomeric. The current results also show that these considerations extend to physiologic systems. In human platelets, 70% of added oleate appears to partition in platelet membranes while 30% remains free. The fraction of oleate that partitions into membranes does not appear to be modulated by physiologic ranges of temperature and ph. Also, the free oleate fraction remains unchanged even at high concentrations of oleate and does not appear to require proteins that may enhance oleate solubility. Therefore, oleate appears to follow similar equilibrium partitioning in physiologic membranes as was demonstrated in detergent micelles. Membrane partitioning of oleate appears to have a significant effect on oleate s ability to activate protein kinase C so that the concentration dependence of protein kinase C activation by oleate is shifted to the right in the presence of platelet membranes. Our results indicate that the non-membrane-bound oleate is the form of oleate that interacts with protein kinase C since enzyme activity shows identical dependence on free (non-membrane bound) oleate whether membranes are present or not (Fig. 7). These studies have important implications for mechanistic studies on protein kinase C regulation and on the physiologic role of cis-unsaturated fatty acids as activators of protein kinase C. The major implication for mechanisms of activation of protein kinase C by lipid cofactors is the elucidation of two modes of activation. On the one hand, protein kinase C may be activated by diacylglycerol (or phorbol esters) only when the enzyme is bound to a surface (cell membranes (21), liposomes (22, 23), or detergent/phospholipid mixed micelles (16)). On the other hand, sodium oleate and other cis-unsaturated fatty acids activate protein kinase C when both fatty acid and enzyme are not membrane-associated. The distinct physical requirements for activation of protein kinase C by DAG and cis-unsaturated fatty acids appears to extend to the cellular environment. It has been proposed that physiologic regulation of protein kinase C may occur by two pathways: DAG generated by phospholipase C activation and arachidonic acid generated by phospholipase AS activity (7, 8). The current studies suggest that this dual activation of protein kinase C may occur in two distinct compartments. Membrane-associated protein kinase C is activated upon the generation of diacylglycerol second messengers while soluble protein kinase C is activated by cis-unsaturated fatty acids (Fig. 9). Although it is difficult to determine the fraction of protein kinase C that is membrane-associated under physiologic conditions, in our studies we detect approximately 60-70% of protein kinase C in the cytosolic fraction in the presence of nm calcium. The generation of DAG results in an increase in the membrane-bound fraction of protein kinase C, so called translocation (24). On the other hand, very little information is available on the concentra- tions of free arachidonate that are generated by the action of phospholipase A2, although it has been suggested that the total concentration may reach the micromolar range in activated platelets (25). In our studies, we find that only 30% of arachidonate partitions into platelet membranes (Table 11) and arachidonate has similar in vitro activity as oleate in activating protein kinase C (data not shown). Therefore, arachidonate appears to be better suited to activate soluble protein kinase C under physiological conditions where up to 70% of arachidonate remains membrane-free (Table 11). Preliminary studie suggest the ability of endogenously generated arachidonate to activate protein kinase C independently of DAG in collagen-stimulated platelet^.^ Such evidence is re- W. A. Khan, M. Werner, and Y. A. Hannun, manuscript in preparation.

8 3612 Activation of Protein Kinase quired to demonstrate that arachidonate acts as a second messenger in the regulation of protein kinase C. Acknowledgments-We wish to thank Samia El Touny for expert technical assistance, Supriya Jayadev for careful review of the manuscript, and Marsha Haigood for expert secretarial assistance. REFERENCES 1. Nishizuka, Y. (1989) Cancer 63, Nishizuka, Y. (1988) Nature 334, Rhee, S. G., Suh, P. G., Ryu, S. H., and Lee, S. Y. (1989) Science 244, Loffelholz, K. (1989) Biochem. Pharnacol. 38, Castagna, M., Takai, Y., Kaibuchi, K., Sano, K., Kikkawa, U., and Nishizuka, Y. (1982) J. Biol. Chem. 257, Niedel, J. E., Kuhn, L. J., and Vandenbark, G. R. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, McPhail, L. C., Clayton, C. C., and Snyderman, R. (1984) Science 224, Murakami, K., Chan, S. Y., and Routtenberg, A. (1986) J. Biol. Chem. 261, El Touny, S., Khan, W., and Hannun, Y. (1990) J. Biol. Chem. 265, Sekiguchi, K., Tsukuda, M., Ogita, K., Kikkawa, U., and Nishizuka, Y. (1987) Biochem. Biophys. Res. Commun. 145, Naor, Z., Shearman, M. S., Kishimoto, A., and Nishizuka, Y. (1988) Mol. Endocrinol. 2, C by Oleic Acid 12. Shearman, M. S., Naor, Z., Sekiguchi, K., Kishimoto, A., and Nishizuka, Y. (1989) FEBS Lett. 243, Sharkey, N. A., and Blumberg, P. M. (1985) Biochem. Biophys. Res. Commun. 133, Kitano, T., Go, M., Kikkawa, U., and Nishizuka, Y. (1986) Methods Enzymol. 124, Huang, K. P., Nakabayashi, H., and Huang, F.L. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, Hannun, Y. A., Loomis, C. R., and Bell, R. M. (1985) J. Biol. Chem. 260, Hannun, Y. A,, Loomis, C. R., and Bell, R. M. (1986) J. Bid. Chem. 261, Sano, K., Takai, Y., Yamanishi, J., and Nishizuka, Y. (1983) J. Biol. Chem. 258, Rink, T. J., Smith, S. W., and Tsien, R. Y. (1982) FEBS Lett. 148, Spooner, P. J. R., Gantz, D. L., Hamilton, J. A., and Small, D. M. (1990) J. Biol. Chem. 265, Kishimoto, A,, Takai, Y., Mori, T., Kikkawa, U., and Nishizuka, Y. (1980) J. Biol. Chem. 255, Boni, L. T., and Rando, R. R. (1985) J. Biol. Chem. 260, Bazzi, M. D., and Nelsestuen, G. L. (1987) Biochemistry 26, Kraft, A. S., and Anderson, W. B. (1983) Nature 301, Nishikawa, M., Hidaka, H., and Shirakawa, S. (1988) Biochem. Pharmacol. 37, Hannun, Y. A., and Bell, R. M. (1990) J. Biol. Chem. 265,