Regulation of Protein Kinase C (cid:109) by Basic Peptides and Heparin PUTATIVE ROLE OF AN ACIDIC DOMAIN IN THE ACTIVATION OF THE KINASE*

Protein kinase C (cid:109) is a novel member of the protein kinase C (PKC) family that differs from the other isoen- zymes in structural and enzymatic properties. No substrate proteins of PKC (cid:109) have been identified as yet. Moreover, the regulation of PKC (cid:109) activity remains ob-scure, since a structural region corresponding to the pseudosubstrate domains of other PKC isoenzymes has not been found for PKC (cid:109) . Here we show that aldolase is phosphorylated by PKC (cid:109) in vitro . Phosphorylation of aldolase and of two substrate peptides by PKC (cid:109) is inhibited by various proteins and peptides, including typical PKC substrates such as histone H1, myelin basic protein, and p53. This inhibitory activity seems to depend on clusters of basic amino acids in the protein/peptide structures. More-over, in contrast to other PKC isoenzymes PKC (cid:109) is activated by heparin and dextran sulfate. Maximal activa- tion by heparin is about twice and that by dextran sulfate four times as effective as maximal activation by phosphatidylserine plus 12- O -tetradecanoylphorbol-13-acetate, the conventional activators of c- and nPKC isoforms.Wepostulate that PKC (cid:109) contains an acidic domain, which is involved in the formation and stabilization of an active state and which, in the inactive enzyme, is blocked by

Protein kinase C (PKC) 1 is a serine/threonine protein kinase that is phospholipid-dependent and activated by diacylglycerol and the phorbol ester TPA (1)(2)(3). In this respect, PKC behaves similarly as most PKC isoenzymes (cPKCs and nPKCs, for reviews see Refs. 4 and 5) of the PKC family. However, PKC differs in some structural and enzymatic features from the other PKC isotypes known so far (1)(2)(3)6), indicating that it may represent a novel subfamily of PKC. Thus, the two cysteine-rich domains, which serve as binding sites for cofactors and activators, are much further apart in PKC than in all the other PKCs and in contrast to the other PKC isoenzymes, PKC contains a pleckstrin homology domain and lacks a region corresponding to pseudosubstrate regions of the PKC family members. Moreover, PKC is not inhibited by a PKCspecific inhibitor (7) and is not down-regulated upon prolonged TPA treatment of murine keratinocytes and epidermis (8). As yet, no substrate proteins of PKC have been found (3,9), neither in vivo nor in vitro, even though numerous proteins are known that are phosphorylated by the other PKC isoenzymes. Recently, Sidorenko et al. (10) claimed that the tyrosine kinase Syk and the phospholipase C␥1 were substrates of PKC. However, incorporation of phosphate into these proteins was lower than into myelin basic protein, which was shown by us and others to be an extremely poor substrate of PKC (3,9). As PKC substrates are not known so far, the function of PKC in cellular signaling is for the most part obscure. Very recently, data were presented suggesting that PKC is located at the Golgi apparatus and is involved in basal transport processes (11). Moreover, it was suggested that PKC regulates lymphocyte signaling (10).
Here, we report on a possibly differential regulation of the kinase activities of PKC and other PKC isoenzymes. In contrast to other PKCs, PKC is inhibited by various proteins and peptides, most likely due to clusters of basic residues in their structure, and is activated by heparin and dextran sulfate. Based on these results, the putative role of an acidic domain in the activation of PKC is discussed. Other materials were bought from companies as indicated: [␥-32 P]ATP (specific activity, 5000 Ci/mmol), Hartman Analytic (Braunschweig, Germany); aldolase, Boehringer (Mannheim, Germany); heparin, phosphatidylserine (PS), protamine sulfate, dextran sulfate, histone H1 (III-S), myelin basic protein, poly-L-lysine (M r 15,000 -60,000), poly-L-lysine (M r 1,000 -4,000) poly-L-arginine (M r 15,000 -60,000), Llysine, histamine, quinine, Sigma (Munich, Germany).

Materials
Recombinant PKC-Sf 158 cells were infected with recombinant PKC baculovirus, and cell extracts were prepared and used as source for PKC as described previously (3,7).
Protein Kinase C Assay-Phosphorylation reactions were carried out in a total volume of 100 l containing buffer I (50 mM Tris-HCl, pH 7.5, 10 mM ␤-mercaptoethanol), 4 mM MgCl 2 , 5 l of a Sf 158 cell extract containing recombinant PKC, 35 M ATP containing 1 Ci of [␥-32 P]ATP and 5 g of syntide 2 or -peptide 1 as substrates. PS, TPA, heparin, Gö6976, Gö6983, or various other compounds (see Table I) were added at concentrations indicated in the legends of the figures and Table I. After incubation for 7 min at 30°C, the reaction was terminated by transferring 50 l of the assay mixture onto a 20-mm square piece of phosphocellulose paper (Whatman p81), which was washed three times in deionized water and twice in acetone. The radioactivity on each paper was determined by liquid scintillation counting. Phosphate incorporated into the substrate peptide was obtained by subtracting values determined in the absence of kinase.
Autophosphorylation and Phosphorylation of Aldolase or Histone H1-These phosphorylations were carried out essentially as described for the protein kinase C assay. However, no substrate was added for the autophosphorylation, and for the phosphorylation of aldolase or histone H1, these proteins instead of the substrate peptides were added, at the concentrations indicated in the text. The assay contained 7 Ci of [␥-32 P]ATP. Proteins of the reaction mixture were separated by SDSpolyacrylamide gel electrophoresis and visualized by autoradiography.

RESULTS AND DISCUSSION
No substrate protein of PKC, neither in vitro nor in vivo, has been found so far. In accordance with previous reports (3,9), we observed that typical PKC substrates, such as histone H1, myelin basic protein, and protamine sulfate, were phosphorylated very weakly by PKC and, therefore, cannot be considered as PKC substrates. Recently, the tyrosine kinase Syk and the phospholipase C␥1 were claimed to be substrates of PKC. However, phosphorylation by PKC in vitro of these proteins was even weaker than that of myelin basic protein (10). Here we show that aldolase can serve as a substrate for PKC in vitro. Aldolase was phosphorylated by PKC much more effectively than histone H1 (Fig. 1). To our surprise, when aldolase (5 g) and histone (5 or 10 g) were both present in the kinase assay, histone H1 suppressed the phosphorylation of aldolase almost completely and also autophosphorylation of PKC was inhibited ( Fig. 1).
To determine the dose dependence of the inhibitory effect of histone H1 on PKC, we used syntide 2 as well as a novel synthetic peptide that we termed -peptide 1 as substrates for PKC. The -peptide 1 with the amino acid sequence RKRYS-VDKTLSHPWL, corresponding to the sequence 825-839 of human PKC, proved to be a potent PKC substrate. It incorporated around 30% more phosphate than syntide 2 on phosphorylation with PKC. 2 Phosphorylation of syntide 2 and -peptide 1 by PKC was inhibited by histone H1 depending on the concentration of the inhibitor (Fig. 2). Half-maximal inhibition was reached with 0.2-0.3 g of histone H1/100 l (IC 50 ).
To prove the hypothesis that the basic properties of histone H1 were responsible for the suppression of PKC activity, we tested various proteins and peptides containing basic regions and other basic compounds (5 g/100 l each) for their inhibitory capacity (Table I). All proteins tested suppressed the PKC-catalyzed phosphorylation, with the exception of the elongation factor EF-1␣. Protamine sulfate and histone H1 were most effective in this respect, followed by myelin basic protein and the human tumor suppressor protein p53. The p53-peptide with the amino acid sequence SHLKSKKGQS-TSRHKK, corresponding to sequence 367-382 of human p53, was similarly active as the p53 protein. Peptides derived from  proteins, basic peptides, and other basic compounds Incorporation of phosphate into syntide 2 by PKC was determined by applying the kinase assay as described under "Experimental Procedures." Mutated amino acids in the Ser-pseudosubstrates 1 and 2 are in bold. The elongation factor EF-1␣ was purified from porcine spleen as described previously (12).
Compound (5 g Poly-L-arginine, M r 15,000-60,000 90 Poly-L-lysine, M r 15,000-60,000 76 Poly-L-lysine, M r 1,000-4,000 64 L-Lysine, histamine, quinine (free base) 0 the pseudosubstrate domain of PKC (IYRRGSRRWRKL) and (RKRQRSMRRRVH), which contain serine instead of alanine and therefore serve as substrates for several PKC isoenzymes, were also found to effectively inhibit PKC. However, the respective peptide derived from the pseudosubstrate of PKC␦ (MNRRGSIKQAKI) as well as the -peptide 2 (GVRRRRL), corresponding to the amino acid sequence 198 -204 of human PKC, did not show such an inhibitory effect. A major difference between the two peptides and the inhibitory peptides (-peptide, -peptide, myristoylated alanine-rich protein kinase C substrate-peptide, and p53-peptide) exists in the total number and clustering of basic amino acids (Arg/Lys). The ␦-peptide and the -peptide 2 contain four basic amino acids and one cluster of two or four basic amino acids, respectively, whereas the inhibitory peptides have at least six basic amino acids arranged in two or three clusters. The myristoylated alanine-rich protein kinase C substrate-peptide (KKKKKRFS-FKKSFKLSGFSFKKSK) with 12 basic residues and three clusters was the most effective inhibitor peptide. Thus, a peptide might require a minimal positive net charge and/or specific clusters of basic amino acids to be able to inhibit PKC. In fact, an exchange of one or two basic amino acids in the -peptide for neutral residues resulting in -peptide-1 (IYRRGSIRWRKL) and -peptide-2 (IYRRGSIRWAKL) caused a gradual loss of inhibitory activity (Table I). -peptide-2 has a similar arrangement of basic amino acids as the ␦-peptide. EF-1␣ protein, which does not contain any cluster of basic amino acids even though it is basic (pI ϭ 9), did not inhibit PKC thus further supporting our notion. The strongly basic polypeptides poly-Larginine and poly-L-lysine (molecular weights of 15-60 kDa), and even the smaller poly-L-lysine (molecular mass of 1-4 kDa) inhibited PKC effectively. L-Lysine and other basic low mo-lecular weight compounds, such as histamine and quinine, were on the other hand unable to inhibit PKC activity. This indicates that structural features, such as the above mentioned clusters of basic amino acids, rather than a positive net charge, determine the suitability of a compound to act as PKC inhibitor, thus pointing to some specificity of the interaction with the kinase. Most of the proteins and peptides inhibiting PKC are substrates rather than inhibitors of the other PKC isoenzymes, and some of them, such as protamine and poly-L-arginine, have been found to activate other PKC isoenzymes (13). On the other hand, none of the inhibitory proteins and peptides was significantly phosphorylated by PKC. Thus, inhibition of PKC was not likely to be due to a competition of the inhibitory compound with the substrate syntide 2 for ATP. Inhibition was not reduced by increasing substrate concentrations, as demonstrated in Fig. 3 for the inhibition by protamine sulfate of syntide 2 phosphorylation by PKC. For comparative purposes, Fig. 3 shows also the inhibition of PKC␦ by the pseudosubstrate peptide. In this case, inhibition decreased upon increasing the concentration of the substrate syntide 2. This clearly demonstrates that the PKC-inhibiting peptides do not act like the well known pseudosubstrate peptides that inhibit other PKC isoenzymes by competing with the PKC substrate for its binding site (14). Therefore, we postulate that PKC contains an acidic domain, different from the acidic substrate-binding motif of other PKCs (see Ref. 15), which is involved in enzyme activation or stabilization of the active state of the kinase. In the active state PKC is inhibited by proteins and peptides containing clusters of basic residues probably due to an interaction with this "activating" domain. In the inactive state the acidic domain might not be accessible, due to an interaction with an autoregulatory basic domain of the enzyme. Indeed, PKC exhibits a highly acidic domain (amino acid sequence 336 -391 of human PKC) in the regulatory part close to the C terminus of the cysteine-rich regions. This domain contains 40% acidic and just 2% basic residues and, in a smaller region (342-362), even 48% acidic residues. It is intriguing that the other PKC isoenzymes lack a comparable domain.
Our hypothesis would imply that polyanions are able to break up the autoinhibitory interaction between the acidic and the basic domain. In fact, the highly sulfated polysaccharides heparin and dextran sulfate were found to function as potent activators of PKC. Maximal activation of PKC by heparin alone, i.e. in the absence of any other cofactor, was about twice and that by dextran sulfate four times as effective as maximal activation by PS/TPA (Fig. 4). Dextran sulfate contains more sulfate groups than heparin and is, therefore probably, more active than heparin in stimulating PKC. Application of PS/ TPA together with heparin or dextran sulfate did not further increase the activity of PKC. Activation of PKC by heparin or dextran sulfate was saturable at low concentrations (Fig. 5). The K a values for heparin and dextran sulfate, as determined by a Lineweaver-Burk plot, were approximately 0.36 and 0.05 M (based on an average molecular weight of 20,000 and 500,000, respectively, as given by the supplier). Thus, heparin and dextran sulfate are very effective activators as compared for instance with diacylglycerol (e.g. the K a value of dioctanoylglycerol for PKC␦ is 10 M, see Ref. 16). As shown in Fig. 6, A and B, the maximal velocities (V max ) of the heparin-and dextran sulfate-activated syntide 2 phosphorylations (23.3 pmol/ min and 41.7 pmol/min, respectively) were much higher than that of the PS/TPA-activated phosphorylation (9.5 pmol/min). On the other hand, the affinity of the enzyme for the substrate was rather lower upon activation with the two polyanions (same K m for both: 9.5) than with PS/TPA (K m : 4.8 M). Thus, the much more effective incorporation of phosphate into syntide 2 by the polyanion-activated PKC than by the PS/TPAactivated kinase is due to an increase in the maximal velocity of the phosphorylation reaction. As the mechanisms of action of heparin and dextran sulfate are likely to be identical, we will in the following just deal with heparin. Autophosphorylation of PKC was also more efficiently stimulated by heparin alone than by PS/TPA (Fig. 7). Both, heparin-and PS/TPA-activated (7) autophosphorylation could be strongly suppressed by 1 M of the PKC inhibitor Gö6976, but not at all by 1 M of Gö6983, an effective inhibitor of the other PKC isoenzymes. These inhibitors are known to interact with the ATP binding site of PKC.
TPA is generally thought to activate PKC by a conformational change that results from its binding to the zinc finger regions of the enzymes (4). As a consequence, an inhibitory pseudosubstrate domain is removed from the substrate binding site. Whether this mechanism can explain the activation by TPA/PS of PKC remains an open question, since a domain corresponding to the pseudosubstrate regions of other PKC isoenzymes has not been found in the PKC structure (7). This does not exclude, however, that upon identification of bona fide in vivo substrates of PKC, a specific pseudosubstrate sequence will be identified in the future. Within the PKC family the activation by heparin may, on the other hand, turn out to be a specific feature of PKC, since PKC␦ activity was not affected by heparin (data not shown) and a PKC preparation from rat brain (containing mainly PKC ␣, ␤, ␥) was even inhibited by heparin (17). The latter result is in agreement with the finding that heparin might block smooth muscle cell proliferation by inhibition of PKC␣ (18). Several other protein kinases, such as casein kinase 1 and 2, nuclear kinases, the tyrosine kinase Syk, and G-protein-coupled receptor kinases are inhibited by heparin (17,19,20, and references in Ref. 17). On the other hand, activation by heparin was reported for instance for a RNAactivated protein kinase (21) and a Lyn-related tyrosine protein kinase (22). Activation of each of the two kinases by heparin was shown to occur through mechanisms different from those of other known activators of these kinases, thus resembling the activation of PKC by heparin. Moreover, many growth factors are known to bear specific heparin-binding sites that contain a cluster of basic amino acid residues (23,24).
The activation by heparin or dextran sulfate of PKC appears to be rather specific, as other acidic compounds, such as chondroitin sulfate, cholesterol sulfate, double-stranded polyinosinic-polycytidylic acid, DNA (calf thymus), poly-L-aspartic acid, and poly-L-glutamic acid, did not or just very weakly activate PKC (data not shown). This supports the notion that heparin and dextran sulfate specifically break up the intramolecular interaction of basic residues with an acidic domain of PKC. The apparent specificity of the stimulatory effect may be taken as an indication for a physiological function of heparin or heparin-like compounds in the control of PKC activity. Heparin has been shown to affect various intracellular signaling pathways, including PKC-dependent pathways, and to be a potent proliferation inhibitor of several cell types (Refs. 25 and 26 and references therein). However, these effects are thought to be mediated by binding of heparin to cell surface binding sites or growth factors (23,24,27). Little is known about possibly direct actions of heparin on signaling pathways inside the cell. As heparin is synthesized in the Golgi complex and PKC has recently been shown to be localized there (11), a direct action of heparin on PKC in this cellular compartment is conceivable. Alternatively, heparin might mimic the effects of heparin-like factors in vivo. Such factors are produced, for instance, by endothelial and smooth muscle cells and are growth-inhibitory for these cells (28,29).