Interaction of protein kinase C with phosphatidylserine. 2. Specificity and regulation.

The roles of specific and nonspecific interactions in the regulation of protein kinase C by lipid have been examined. Binding and activity measurements reveal two mechanisms by which protein kinase C interacts with membranes: (1) a specific binding to the activating lipid phosphatidylserine and (2) a nonspecific binding to nonactivating, acidic lipids. The specific interaction with phosphatidylserine is relatively insensitive to ionic strength, surface charge, and the presence of nonactivating lipids. The two second messengers of the kinase, diacylglycerol and Ca2+, increase markedly the affinity of the kinase for phosphatidylserine. In contrast, the nonspecific interaction is sensitive to ionic strength and surface charge, and is unaffected by diacylglycerol. These results suggest that electrostatic interactions promote the binding of protein kinase C to membranes but the cooperative and selective binding of phosphatidylserine is the dominant driving force in a productive protein-lipid interaction.     

Interaction of protein kinase C with phosphoinositides.

Calcium/phosphatidylserine-dependent protein kinase C (PKC) is activated by phosphatidylinositol 4,5-bisphosphate (PIP2), as well as by diacylglycerol (DG) and phorbol esters. Here we report that PIP2, like DG, increases the affinity of PKC for Ca2+, and causes Ca(2+)-dependent translocation of the enzyme from the soluble to a particulate fraction (liposomes). Phosphatidylinositol 4-phosphate (PIP) also displaces phorbol ester from PKC and causes Ca(2+)-dependent translocation of the enzyme to liposomes, but is much less efficient than PIP2, and a much weaker activator, with a histone phosphorylation v(PIP)/v(PIP2) of approximately 0.15. Scatchard analysis indicates competitive inhibition between PIP and phorbol ester with Ki(PIP) = 0.26 mol% as compared with Ki(PIP2) = 0.043 mol%. No effect of phosphatidylinositol (PI) on phorbol ester binding to PKC, translocation of PKC, or activation of PKC was observed. These results suggest that both PIP and PIP2 can complex with PKC, but full activation of the enzyme takes place only when PIP is converted to PIP2. We suggest that an inositide interconversion shuttle has a role in the regulation of protein phosphorylation.     

Interaction of endotoxic lipid A and lipid X with purified human platelet protein kinase C

Lipid A, the toxic principle of endotoxic lipopolysaccharide, and its precursor, Lipid X, interact with human platelets and modulate protein kinase C therein (Grabarek, J., Timmons, S., and Hawiger, J. (1988) J. Clin. Invest. 82, 964-971). We have now purified protein kinase C from human platelets and studied its interaction with endotoxic Lipids A and X. Protein kinase C-dependent phosphorylation of histone III-S was increased 15 times in the presence of Lipid A and 300 microM Ca2+. The Ca2+ requirement for such activation was lower when 4 beta-phorbol 12-myristate 13-acetate (PMA) or 1,2-diolein were added. Lipid A also induced autophosphorylation of protein kinase C, and its activation was enhanced by phosphatidylserine without reducing the Ca2+ requirement. Kinetic analysis of protein kinase C activation induced by Lipid A, in regard to ATP as a substrate, demonstrated that Lipid A increased the rate of the reaction (Vmax) without modifying the affinity of the enzyme (Km) for the substrate. Lipid X inhibited the activation of the enzyme induced by Lipid A. Lipid X also inhibited protein kinase C activation by phosphatidylserine, 1,2-diolein, and PMA. However, 10 times more of Lipid X was required for 50% inhibition (IC50) when PMA was used as an activator of protein kinase C in the presence of phosphatidylserine than when Lipid A and 1,2-diolein were used. These results support the hypothesis that endotoxic Lipid A and Lipid X exert their biological effect in platelets through direct interactions with protein kinase C.     

Interaction of phosphoglycerate kinase with phosphatidylserine liposomes

The interaction of 3-phosphoglycerate kinase from bovine heart with natural phosphatidylserine (I) and synthetic dipalmitoyl phosphatidylserine (II) in form of liposomes was investigated by measuring fluorescence and activity of the enzyme. The addition of increasing amounts of I resulted in progressive quenching of protein fluorescence with no shift in the emission maximum. In contrast, II did not cause any change in the fluorescence. In the presence of low amounts of I and II (lipid/protein molar ratio 10-40) full enzymatic activity of 3-phosphoglycerate kinase was observed even after 80 min of incubation, whereas without phospholipids the activity considerably decreased. At higher lipid concentrations I strongly inactivated the enzyme and the inactivation by II was only insignificant. It was concluded that the phospholipid membrane protects the enzyme against thermal denaturation, whereas the inactivation is mainly due to phospholipid impurities.     

Substitution of phosphatidylserine by lipid A in the activation of purified rabbit brain protein kinase C

Three lipid A derivatives (hexaacyl monophosphoryl lipid A, hexaacyl diphosphoryl lipid A, and disaccharide precursor IVA) were shown to activate protein kinase C from rabbit brain. These derivatives substituted for phosphatidylserine in a concentration-dependent manner and did not compete for binding of [3H]phorbol dibutyrate to its receptor site. Instead, phorbol dibutyrate binding was increased on raising the concentration of the derivatives in a similar manner to phosphatidylserine. The phorbol ester 12-0-tetra-decanol 13-acetate augmented the activation of protein kinase C by the lipid A derivatives.     

Substitution by fatty acids for phosphatidylserine in a reconstitution of phorbol ester binding to protein kinase C.

1. Fatty acids can be substituted or phosphatidylserine in a reconstitution of phorbol ester binding to protein kinase C. 2. Phorbol ester, however, does not seem to be effectively utilized for the activation of the enzyme. 3. It is suggested that fatty acids play a role on the activation of protein kinase C in the abnormal conditions such as ischemia, while the phospholipid-dependent activation has a physiological significance in normal conditions.     

Interaction of protein kinase C isozymes with phosphatidylinositol 4,5-bisphosphate.

Interaction of protein kinase C (PKC) isozymes with phosphatidylinositol 4,5-bisphosphate (PIP2) was investigated by monitoring the changes in the intrinsic fluorescence of the enzyme, the kinase activity, and phorbol ester binding. Incubation of PKC I, II, and III with PIP2 resulted in different rates of quenching of PKC fluorescence and different degrees of inactivation of these enzymes. Other inositol-containing phospholipids such as phosphatidylinositol and phosphatidylinositol 4-phosphate also caused differential rates of quenching of the intrinsic fluorescence of these enzymes. These latter two phospholipids were, however, less potent in the inactivation of PKCs than PIP2. The IC50 of PIP2 were 2, 4, and 11 microM for PKC I, II, and III, respectively. Inactivation of PKCs by PIP2 cannot be reversed by extensive dilution of PIP2 with Nonidet P-40 nor by digestion of PIP2 with phospholipase C. Interaction of PIP2 with the various PKC isozymes was greatly facilitated in the presence of Mg2+ or Ca2+ as evidenced by the accelerated quenching of the PKC fluorescence, however, these divalent metal ions protected PKC from the PIP2-induced inactivation. Binding of PIP2 to PKC in the absence of divalent metal ion also caused a reduction of [3H]phorbol 12,13-dibutyrate binding as a result of reducing the affinity of the enzyme for phorbol ester. Based on gel filtration chromatography, it was estimated that one molecule of PKC interacted with one PIP2 micelle with an aggregation number of 80-90. The PIP2-bound PKC could further interact with phosphatidylserine in the presence of Ca2+ to form a larger complex. Binding of PKC to both PIP2 and phosphatidylserine in the presence of Ca2+ was also evident by changes in the intrinsic fluorescence of PKC. As the interaction of PKC with PIP2, but not with phosphatidylserine, could be enhanced by millimolar concentrations of Mg2+, we propose that PIP2 may be a component of the membrane anchor for PKC under basal physiological conditions when [Ca2+]i is low and Mg2+ is plentiful. Under the in vitro assay conditions, PIP2 could stimulate PKC activity to a level approximately 10-20% of that by diacylglycerol. The stimulatory effect of PIP2 on PKC apparently is not due to binding to the same site recognized by diacylglycerol or phorbol ester, because PIP2 cannot effectively compete with phorbol 12,13-dibutyrate in the binding assay.     

Calcium and phosphatidylserine stimulate the self-association of conventional protein kinase C isoforms.

Conflicting evidence exists as to whether "conventional" protein kinase C isoforms (cPKCs) function as monomers or oligomers. In this report, we demonstrate that purified cPKC isoforms can be rapidly cross-linked by the sulfhydryl-selective cross-linker bis(maleimido)hexane, but only in the presence of both Ca(2+) and phosphatidylserine; cross-linking was minimal in the presence of either of these activators alone. In addition, cross-linking of these cPKCs did not require Mg(2+) or ATP. Among the various phospholipids tested, phosphatidylserine was found to be the most effective in the promotion of cPKC self-association and for the stimulation of protein kinase activity toward the exogenous substrate histone. Phosphatidic acid and phosphatidylinositol were less effective in this regard, whereas phosphatidylcholine exhibited little ability to induce cPKC self-association or to stimulate kinase activity. An examination of the mechanism by which the cPKC isoforms self-associate in the presence of phospholipid/Ca(2+) revealed that this process occurred independently of phospholipid aggregation. Moreover, self-association was not inhibited by saturating the enzyme active site with a peptide substrate, suggesting that self-association is distinct from an enzyme-substrate interaction. Isoform-specific antibodies revealed that all cPKC isoforms (alpha, beta, and gamma) self-associate and that, in a mixture of cPKC isoforms, PKC-alpha forms primarily alpha-alpha homodimers. Besides cPKC interactions detected with purified enzyme, PKC-alpha also appeared capable of self-association in murine B82L fibroblasts that were treated with calcium ionophore, phorbol ester, or epidermal growth factor but not in untreated cells. Collectively, these data indicate that self-association occurs in parallel with cPKC activation, that self-association is not mediated by the substrate binding site, and, at least in the case of PKC-alpha, that the formation of isoform homodimers predominates.     

Phorbol ester binding and activation of protein kinase C on triton X-100 mixed micelles containing phosphatidylserine

A mixed micellar assay for the binding of phorbol-esters to protein kinase C was developed to investigate the specificity and stoichiometry of phospholipid cofactor dependence and oligomeric state of protein kinase C (Ca2+/phospholipid-dependent enzyme) required for phorbol ester binding. [3H]Phorbol dibutyrate was bound to protein kinase C in the presence of Triton X-100 mixed micelles containing 20 mol % phosphatidylserine (PS) in a calcium-dependent manner with a Kd of 5 X 10(-9) M. The [3H]phorbol dibutyrate X protein kinase C . Triton X-100 . PS mixed micellar complex eluted on a Sephacryl S-200 molecular sieve at an Mr of approximately 200,000; this demonstrates that monomeric protein kinase C binds phorbol dibutyrate. This conclusion was supported by molecular sieve chromatography of a similar complex where Triton X-100 was replaced with beta-octylglucoside. Phorbol dibutyrate activation of protein kinase C in Triton X-100/PS mixed micelles occurred and was dependent on calcium. The PS dependence of both phorbol ester activation and binding to protein kinase C lagged initially and then was highly cooperative. The minimal mole per cent PS required was strongly dependent on the concentration of phorbol dibutyrate or phorbol myristic acetate employed. Even at the highest concentration of phorbol ester tested, a minimum of 3 mol % PS was required; this indicates that approximately four molecules of PS are required. [3H]Phorbol dibutyrate binding was independent of micelle number at 20 mol % PS. The phospholipid dependencies of phorbol ester binding and activation were similar, with PS being the most effective; anionic phospholipids (cardiolipin, phosphatidic acid, and phosphatidylglycerol were less effective, whereas phosphatidylcholine, phosphatidylethanolamine, and sphingomyelin did not support binding or activation. sn-1,2-Dioleoylglycerol displaced [3H]phorbol dibutyrate quantitatively and competitively. The data are discussed in relation to a molecular model of protein kinase C activation.     

Interaction of protein kinase C isozymes with Rho GTPases

Evidence is provided for direct protein-protein interactions between protein kinase C (PKC) alpha, betaI, betaII, gamma, delta, epsilon, and zeta and members of the Rho family of small GTPases. Previous investigations, based on the immunoprecipitation approach, have provided evidence consistent with a direct interaction, but this remained to be proven. In the study presented here, an in vitro assay, consisting only of purified proteins and the requisite PKC activators and cofactors, was used to determine the effects of Rho GTPases on the activities of the different PKC isoforms. It was found that the activity of PKCalpha was potently enhanced by RhoA and Cdc42 and to a lesser extent by Rac1, whereas the effects on the activities of PKCbetaI, -betaII, -gamma, -delta, -epsilon, and -zeta were much reduced. These results indicate a direct interaction between PKCalpha and each of the Rho GTPases. However, the Rho GTPase concentration dependencies for the potentiating effects on PKCalpha activity differed for each Rho GTPase and were in the following order: RhoA > Cdc42 > Rac1. PKCalpha was activated in a phorbol ester- and Ca(2+)-dependent manner. This was reflected by a substantial decrease in the phorbol ester concentration requirements for activity in the presence of Ca(2+), which for each Rho GTPase was induced within a low nanomolar phorbol ester concentration range. The activity of PKCalpha also was found to be dependent on the nature of the GTP- or GDP-bound state of the Rho GTPases, suggesting that the interaction may be regulated by conformational changes in both PKCalpha and Rho GTPases. Such an interaction could result in significant cross-talk between the distinct pathways regulated by these two signaling elements.     

Activation of protein kinase C in lipid monolayers

The potential of lipid monolayers spread at an air-water interface was investigated as a well defined membrane model able to support protein kinase C (PKC) association and activation. PKC association to a mixed phospholipid film (phosphatidylcholine, phosphatidylserine) could be detected by an increase of the monolayer surface pressure. This association was strikingly dependent upon the presence of submicromolar concentrations of Ca2+. The effect of Ca2+ resulted in an increase of the PKC penetration into the lipid core at a given permissive surface pressure as well as in a marked increase of the critical surface pressure (29-38 dynes/cm) above which the enzyme was excluded from the membrane. Inclusion of diacylglycerol or tetradecanoate phorbol acetate (TPA) did not modify the PKC-monolayer association in a detectable manner. PKC associated to the lipid layer exhibited the expected catalytic property and was fully activated when diacylglycerol or TPA was included in the membrane. PKC activity was highly dependent upon the surface pressure of the lipid monolayer, being optimal between 30 and 35 dynes/cm. Study of the compression isotherm of various diacylglycerol structures revealed that all potent PKC agonists exhibited an expanded liquid phase behavior with collapse pressure below 40 dynes/cm, in contrast to weak activators which showed condensed isotherms with high collapse pressure (approximately equal to 60 dynes/cm). These observations showed that the lipid monolayer system is well adapted to the study of the molecular mechanisms involved in the regulation of PKC activity at a model membrane interface. They are in line with the suggestion of a major role of Ca2+ in the association (translocation) of PKC to membrane in living cell and suggest that diacylglycerol (and TPA) might activate membrane-associated PKC through local change in the surrounding lipid phase organization.     

Phosphorylation of CCAAT-enhancer binding protein by protein kinase C attenuates site-selective DNA binding.

Four DNA-recombinant proteins, corresponding to the DNA-binding domain of CCAAT/enhancer binding protein (C/EBP), were phosphorylated in vitro by protein kinase C (PKC). High-performance liquid chromatography-peptide mapping of 32P-labeled C/EBP indicated the presence of three major 32P-labeled peptides: S299 (P)RDK, AKKS277 (P)VDK, and GAAGLPGPGGS248 (P)LK. Phosphorylation of C/EBP by PKC or M-kinase resulted in an attenuation of binding to a 32P-labeled CCAAT oligodeoxynucleotide. Three other truncated forms of C/EBP, C/EBP87, C/EBP87S-C, and C/EBP60, were studied to define the sites of phosphorylation affecting DNA binding. Phosphorylation of the C/EBP87, containing sites Ser299 and Ser277, and C/EBP60, containing only site Ser299, by PKC also resulted in attenuation of DNA binding. In contrast, phosphorylation of C/EBP87S-C, which retained Ser277 but had a Cys in place of Ser299, had no effect on DNA binding. Ser299 could not be phosphorylated by PKC if the protein is already bound to specific DNA. Phosphorylation of intact C/EBP from liver nuclear extract by PKC or M-kinase occurred at Ser299 and Ser277 and at an additional site, as demonstrated by immunoprecipitation and peptide mapping.     

Interaction of SNF1 protein kinase with its activating kinase Sak1

The Saccharomyces cerevisiae SNF1 protein kinase, a member of the SNF1/AMP-activated protein kinase (AMPK) family, is activated by three kinases, Sak1, Tos3, and Elm1, which phosphorylate the Snf1 catalytic subunit on Thr-210 in response to glucose limitation and other stresses. Sak1 is the primary Snf1-activating kinase and is associated with Snf1 in a complex. Here we examine the interaction of Sak1 with SNF1. We report that Sak1 coimmunopurifies with the Snf1 catalytic subunit from extracts of both glucose-replete and glucose-limited cultures and that interaction occurs independently of the phosphorylation state of Snf1 Thr-210, Snf1 catalytic activity, and other SNF1 subunits. Sak1 interacts with the Snf1 kinase domain, and nonconserved sequences C terminal to the Sak1 kinase domain mediate interaction with Snf1 and augment the phosphorylation and activation of Snf1. The Sak1 C terminus is modified in response to glucose depletion, dependent on SNF1 activity. Replacement of the C terminus of Elm1 (or Tos3) with that of Sak1 enhanced the ability of the Elm1 kinase domain to interact with and phosphorylate Snf1. These findings indicate that the C terminus of Sak1 confers its function as the primary Snf1-activating kinase and suggest that the physical association of Sak1 with SNF1 facilitates responses to environmental change.     

Interaction of Wnt-1 proteins with the binding protein BiP.

The mouse Wnt-1 gene, a target for insertional activation in mouse mammary tumor virus-induced mammary tumors, encodes poorly secreted, cysteine-rich glycoproteins required for proper central nervous system development. We have been analyzing the biosynthesis of Wnt-1 proteins in several cell lines that express Wnt-1 cDNA from heterologous promoters. A protein of 78 kDa was found to be associated with the intracellular forms of Wnt-1 proteins in mammalian and avian cells by using multiple antisera against Wnt-1 proteins. We have identified p78 as the binding protein BiP with anti-BiP antisera and by its release from Wnt-1 immunoprecipitates upon incubation with MgCl2 and ATP. Experiments with a Wnt-1 mutant that lacks the sequence encoding the signal peptide indicates that Wnt-1 proteins must enter the secretory pathway in order to interact with BiP. We demonstrate that Wnt-1 proteins are associated with BiP in cells in which active Wnt-1 proteins are produced, such as a cultured mammary epithelial cell line and Wnt-1 transgenic mouse mammary tumor cells. The association of Wnt-1 proteins with BiP may be a factor in determining the efficiency of secretion of Wnt-1 gene products.     

Reversible exposure of the pseudosubstrate domain of protein kinase C by phosphatidylserine and diacylglycerol.

The lipid activators of protein kinase C, phosphatidylserine and diacylglycerol, induce a reversible conformational change that exposes the auto-inhibitory pseudosubstrate domain of the enzyme. The pseudosubstrate domain of beta-II protein kinase C is cleaved after the first residue, arginine 19, by the endoproteinase Arg-C only when the kinase is bound to the activating lipid phosphatidylserine. Exposure of this residue is markedly enhanced by diacylglycerol. In contrast, the pseudosubstrate domain is not cleaved in the absence of lipids, when protein kinase C is bound to non-activating acidic lipids, when the kinase has autophosphorylated on the amino terminus, or after dilution of the activating lipids. This work reveals specificity in the interaction of protein kinase C with phosphatidylserine since only this phospholipid causes the specific conformational change detected in the regulatory domain of the enzyme, and demonstrates that allosteric regulators expose the intramolecular auto-inhibitory domain of a kinase.     

RACK1, a protein kinase C anchoring protein, coordinates the binding of activated protein kinase C and select pleckstrin homology domains in vitro.

The pleckstrin homology (PH) domain, identified in numerous signaling proteins including the beta-adrenergic receptor kinase (betaARK), was found to bind to various phospholipids as well as the beta subunit of heterotrimeric G proteins (Gbeta) [Touhara, K., et al. (1994) J. Biol. Chem. 269, 10217-10220]. Several PH domain-containing proteins are also substrates of protein kinase C (PKC). Because RACK1, an anchoring protein for activated PKC, is homologous to Gbeta (both contain seven repeats of the WD-40 motif), we determined (i) whether a direct interaction between various PH domains and RACK1 occurs and (ii) the effect of PKC on this interaction. We found that recombinant PH domains of several proteins exhibited differential binding to RACK1. Activated PKC and the PH domain of beta-spectrin or dynamin-1 concomitantly bound to RACK1. Although PH domains bind acidic phospholipids, the interaction between various PH domains and RACK1 was not dependent on the phospholipid activators of PKC, phosphatidylserine and 1, 2-diacylglycerol. Binding of these PH domains to RACK1 was also not affected by either inositol 1,4,5-triphosphate (IP(3)) or phosphatidylinositol 4,5-bisphosphate (PIP(2)). Our in vitro data suggest that RACK1 binds selective PH domains, and that PKC regulates this interaction. We propose that, in vivo, RACK1 may colocalize the kinase with its PH domain-containing substrates.     

Importance of C1B domain for lipid messenger-induced targeting of protein kinase C

The molecular mechanisms by which arachidonic acid (AA) and ceramide elicit translocation of protein kinase C (PKC) were investigated. Ceramide translocated epsilonPKC from the cytoplasm to the Golgi complex, but with a mechanism distinct from that utilized by AA. Using fluorescence recovery after photobleaching, we showed that, upon treatment with AA, epsilonPKC was tightly associated with the Golgi complex; ceramide elicited an accumulation of epsilonPKC which was exchangeable with the cytoplasm. Stimulation with ceramide after AA converted the AA-induced Golgi complex staining to one elicited by ceramide alone; AA had no effect on the ceramide-stimulated localization. Using point mutants and deletions of epsilonPKC, we determined that the epsilonC1B domain was responsible for the ceramide- and AA-induced translocation. Switch chimeras, containing the C1B from epsilonPKC in the context of deltaPKC (delta(epsilonC1B)) and vice versa (epsilon(deltaC1B)), were generated and tested for their translocation in response to ceramide and AA. delta(epsilonC1B) translocated upon treatment with both ceramide and AA; epsilon(deltaC1B) responded only to ceramide. Thus, through the C1B domain, AA and ceramide induce different patterns of epsilonPKC translocation and the C1B domain defines the subtype specific sensitivity of PKCs to lipid second messengers.     

The C1 and C2 domains of protein kinase C are independent membrane targeting modules, with specificity for phosphatidylserine conferred by the C1 domain

Protein kinase C is specifically activated by binding two membrane lipids: the second messenger, diacylglycerol, and the amino phospholipid, phosphatidylserine. This binding provides the energy to release an autoinhibitory pseudosubstrate from the active site. Interaction with these lipids recruits the enzyme to the membrane by engaging two membrane-targeting modules: the C1 domain (present as a tandem repeat in most protein kinase Cs) and the C2 domain. Here we dissect the contribution of each domain in recruiting protein kinase C betaII to membranes. Binding analyses of recombinant domains reveal that the C2 domain binds anionic lipids in a Ca(2+)-dependent, but diacylglycerol-independent, manner, with little selectivity for phospholipid headgroup beyond the requirement for negative charge. The C1B domain binds membranes in a diacylglycerol/phorbol ester-dependent, but Ca(2+)-independent manner. Like the C2 domain, the C1B domain preferentially binds anionic lipids. However, in striking contrast to the C2 domain, the C1B domain binds phosphatidylserine with an order of magnitude higher affinity than other anionic lipids. This preference for phosphatidylserine is, like that of the full-length protein, stereoselective for sn-1, 2-phosphatidyl-L-serine. Quantitative analysis of binding constants of individual domains and that of full-length protein reveals that the full-length protein binds membranes with lower affinity than expected based on the binding affinity of isolated domains. In addition to entropic and steric considerations, the difference in binding energy may reflect the energy required to expel the pseudosubstrate from the substrate binding cavity. This study establishes that each module is an independent membrane-targeting module with each, independently of the other, containing determinants for membrane recognition. The presence of each of these modules, separately, in a number of other signaling proteins epitomizes the use of these modules as discreet membrane targets.