Inhibition of Membrane Lipid-independent Protein Kinase Cα Activity by Phorbol Esters, Diacylglycerols, and Bryostatin-1*

The activity of membrane-associated protein kinase C (PKC) has previously been shown to be regulated by two discrete high and low affinity binding regions for diacylglycerols and phorbol esters (Slater, S. J., Ho, C., Kelly, M. B., Larkin, J. D., Taddeo, F. J., Yeager, M. D., and Stubbs, C. D. (1996) J. Biol. Chem. 271, 4627–4631). PKC is also known to interact with both cytoskeletal and nuclear proteins; however, less is known concerning the mode of activation of this non-membrane form of PKC. By using the fluorescent phorbol ester, sapintoxin D (SAPD), PKCα, alone, was found to possess both low and high affinity phorbol ester-binding sites, showing that interaction with these sites does not require association with the membrane. Importantly, a fusion protein containing the isolated C1A/C1B (C1) domain of PKCα also bound SAPD with low and high affinity, indicating that the sites may be confined to this domain rather than residing elsewhere on the enzyme molecule. Both high and low affinity interactions with native PKCα were enhanced by protamine sulfate, which activates the enzyme without requiring Ca2+ or membrane lipids. However, this “non-membrane” PKC activity was inhibited by the phorbol ester 4β-12-O-tetradecanoylphorbol-13-acetate (TPA) and also by the fluorescent analog, SAPD, opposite to its effect on membrane-associated PKCα. Bryostatin-1 and the soluble diacylglycerol, 1-oleoyl-2-acetylglycerol, both potent activators of membrane-associated PKC, also competed for both low and high affinity SAPD binding and inhibited protamine sulfate-induced activity. Furthermore, the inactive phorbol ester analog 4α-TPA (4α-12-O-tetradecanoylphorbol-13-acetate) also inhibited non-membrane-associated PKC. In keeping with these observations, although TPA could displace high affinity SAPD binding from both forms of the enzyme, 4α-TPA was only effective at displacing high affinity SAPD binding from non-membrane-associated PKC. 4α-TPA also displaced SAPD from the isolated C1 domain. These results show that although high and low affinity phorbol ester-binding sites are found on non-membrane-associated PKC, the phorbol ester binding properties change significantly upon association with membranes.

Protein kinase C (PKC) 1 constitutes a group of isozymes that are central in cellular signaling pathways that regulate numerous cellular processes, including cell growth, differentiation, and metabolism (1). Each isoform can be classified into one of three major classes according to the cofactor and activator requirements. The "conventional" PKC␣, -␤I, -␤II, and -␥ isoforms are Ca 2ϩ -and anionic phospholipid-dependent, whereas the "novel" PKC␦, -⑀, -, and -and "atypical" PKC andisozymes retain a phospholipid dependence but lack a Ca 2ϩ requirement (2). In addition, the activities of all PKC isoforms, except atypical PKC, are potentiated by the lipid second messenger, diacylglycerol, derived from the receptor-G-protein and phospholipase-catalyzed hydrolysis of phosphatidylinositides and phosphatidylcholines (3) and also by the potent tumorpromoting phorbol esters (4).
The Ca 2ϩ and phospholipid requirements for PKC activity differ according to the lysine and arginine content of the substrate (5). Thus, the PKC-catalyzed phosphorylation of the lysine-rich protein, histone H1, requires the presence of both Ca 2ϩ and phospholipid, whereas the phosphorylation of the arginine-rich protein, protamine sulfate, requires neither cofactor (6 -10). The mechanism of activation by protamine sulfate, which also acts as a substrate, has been suggested to involve a binding site(s) for arginine-rich proteins on the PKC molecule, separate from the active center of the enzyme (6,11). Similar to that which occurs upon membrane association induced by Ca 2ϩ and diacylglycerol or phorbol esters, interaction with protamine sulfate has been suggested to mediate in an allosteric activating conformational change resulting in the removal of a pseudosubstrate region from the active site which, in the inactive state, blocks substrate binding (12). Therefore, use of protamine sulfate provides a useful model for the activation of PKC in the absence of lipids by interaction with other proteins. The question of the mechanism by which PKC activity induced by protein-protein interactions is regulated has become an urgent concern, since it has become apparent that there are a large number of non-membrane protein targets for the enzyme, such as, for example, cytoskeletal and nuclear elements (e.g. Refs. [13][14][15][16][17][18][19][20][21]. Although interaction with arginine-rich proteins such as protamine sulfate relieves the Ca 2ϩ and phospholipid requirements for PKC activity, it is not known whether this nonmembrane PKC activity is modulated by phorbol esters and/or diacylglycerols in a similar manner to the membrane-associated enzyme. Evidence supporting this possibility was provided recently by the finding that PKC␦ is capable of binding phorbol esters with low affinity in the absence of membrane lipids (22) and that these compounds can induce binding of PKC⑀ to fila-mentous actin (F-actin) (23). Although F-actin itself was not found to be a substrate for this isoform, this may lead to enhanced phosphorylation of other protein targets. PKC␤II has also been shown to bind F-actin which is reported to be a substrate for this isoform (24). In this respect, F-actin can be considered to be an example of a number of specific intracellular proteins that bind the activated form of PKC, termed "receptors for activated C-kinase" (RACKs) (20,21), which allow precise targeting of PKC isoforms to specific cellular locations. However, the non-membrane, phorbol-induced interaction of PKC with F-actin departs from the original definition of RACKs that were described as proteins that bind PKC that has been activated by membrane association (25).
For membrane-associated PKC, activation by diacylglycerol proceeds by two parallel mechanisms. The first involves an increased affinity for the membrane, which is revealed as a decrease in the concentrations of both Ca 2ϩ and anionic phospholipid required for maximal activity (4,26,27). Second, interaction with diacylglycerol and phosphatidylserine induces an activating conformational change that results in the folding out of an N-terminal pseudosubstrate region (12, 28 -31). Phorbol esters have been suggested to potentiate PKC activity in a similar manner to diacylglycerol so that the two activator types initially appeared to share a single site or two identical sites of interaction on the enzyme (4,32,33). However, in recent studies of activator binding to PKC␣, using the fluorescent phorbol ester sapintoxin D (SAPD), we showed that diacylglycerols and phorbol esters interact with differing affinities with two activator binding sites on the enzyme molecule (34 -36). Furthermore, it was shown that the site with low affinity for phorbol esters binds diacylglycerol with a relatively higher affinity, suggesting that the specificity of this site may differ from the high affinity phorbol ester-binding site (35). Interaction of diacylglycerol with this low affinity phorbol ester-binding site was found to lead to an enhancement in the level of high affinity phorbol ester binding and consequently to a potentiated level of PKC activity. By contrast to diacylglycerol, the potent PKC activator and anti-tumor agent, bryostatin-1, was found to compete more effectively for high affinity phorbol ester binding and did not potentiate the level of phorbol ester-induced PKC activity. Based on these results, a model for PKC activation was proposed in which interaction of an activator with the low affinity for phorbol ester-binding site leads to an enhanced level of binding of either the same or a second activator to the high affinity phorbol ester-binding site and consequently to an elevated level of PKC␣ activity (35).
The activator binding site with relatively high affinity for phorbol esters likely resides within the C1 domain, which consists of two cystine-rich zinc finger motifs, termed C1A and C1B (37). Interaction of phorbol esters with these subdomains has been extensively characterized by truncation/deletion and mutagenesis studies (38 -42) along with both crystal (43) and solution-state structural determinations (44,45). By using glutathione S-transferase (GST) fusion proteins containing either the C1A or C1B subdomains, it has been shown that both are capable of binding phorbol esters (39,42,46). However, the phorbol ester binding affinities of the two subdomains may differ for each PKC isoform. For example, whereas PKC␥ C1A and C1B appear to bind phorbol esters with similar affinities, in the case of the novel PKCs, PKC and PKC␦ C1B bind phorbol esters with higher affinity than C1B (47,48). Recent studies have indicated that these isoform-specific differences in phorbol ester binding to C1A and C1B may be carried over into the native enzyme. While for PKC␦ C1B has been identified as being a high affinity phorbol ester-binding site, due to its role in phorbol ester-induced intracellular translocation (49), for PKC␥ the high affinity site may correspond to C1A (46).
The first aim of this study was to determine if the high and low affinity activator binding sites, previously identified on membrane-associated PKC␣ (35), similarly exist on the nonmembrane form of this isoform. Second, the effects of phorbol esters and other activators of membrane-associated PKC on lipid-independent activity induced by protein-protein interactions with protamine sulfate were determined. Finally, while the high affinity phorbol ester-binding site clearly resides with the C1 domain, whether the low affinity binding site is similarly located was also investigated. The results indicate that both high and low affinity phorbol ester-binding sites on nonmembrane PKC␣ pre-exist in the absence of membrane lipids and that both sites are confined within the C1 domain of this isoform. However, phorbol esters, diacylglycerol, and bryostatin-1, although clearly described as being potent activators of membrane-associated PKC␣, are shown in the present study to be potent inhibitors of non-membrane PKC␣ activity induced by protamine sulfate.
Preparation of Intact PKC␣-The recombinant conventional PKC isoform, PKC␣ (rat brain), was prepared using the baculovirus Spodoptera frugiperda (Sf9) insect cell expression system (50) and purified to homogeneity, as described previously (51). The specific activity of the PKC␣ preparation was typically ϳ1 nmol min Ϫ1 g Ϫ1 .
Preparation of GST-C1-(His) 6 Fusion Protein-Based on evidence that GST stabilizes the isolated C1 domain and to ensure the production of only full-length peptides, the fusion protein, GST-C1-(His) 6 , was constructed. The nucleotide consensus sequence assigned to the C1 domain of PKC␣ was that described previously (52). This was amplified by PCR with Pfu polymerase (Stratagene, La Jolla, CA) using fulllength rat cDNA as the template. Primers were designed so that EcoRI and HpaI restriction sites were inserted at the 5Ј and 3Ј ends of the domain, respectively, to facilitate insertion into pGEX-5X-2 (Amersham Pharmacia Biotech). The reverse primer also encoded a blunt restriction site after the last amino acid of the domain which was utilized to insert a (His) 6 sequence (5Ј-CAT CAC CAT CAC CAT CAC TGA-3Ј). The domain fragment was subcloned into pGEX-5x-2 yielding a plasmid, the sequence of which was confirmed by dideoxy sequencing (Nucleic Acid Facility, Thomas Jefferson University), that encoded GST-C1-(His) 6 , a protein-tagged N terminus with GST and C terminus with (His) 6 . Along with the GST-C1-(His) 6 , a fusion protein lacking the C1 domain was prepared by insertion of the (His) 6 sequence alone into pGEX-5x-2 (GST-(His) 6 ).
Escherichia coli BL21 cells harboring the expression plasmids for GST-C1-(His) 6 or GST-(His) 6 were grown in LB medium containing 100 g/ml ampicillin until the absorbance at 600 nm (A 600 ) was ϳ1. Expression of the fusion proteins was induced at room temperature with 0.1 mM isopropylthiogalactoside for 4 h, after which the cells were pelleted, washed once with phosphate-buffered saline, re-pelleted, and stored at Ϫ80°C. The fusion proteins were purified from the frozen E. coli pellets as described previously (52). Briefly, frozen cell pellets were homogenized at 4°C in buffer A (50 mM Hepes, pH 8.0, 10% ethylene glycol, 1% v/v Triton X-100, 0.5 mg/ml lysozyme, 320 units of benzonase, 1 mM phenylmethylsulfonyl fluoride, 10 mM benzamidine, 4 g/ml pepstatin A, 4 g/ml aprotinin, 10 g/ml leupeptin). The homogenate was placed on ice for 20 min and then centrifuged at 30,000 ϫ g for 30 min at 4°C. Sufficient dithiothreitol was added to the cleared lysate to a yield a final concentration of 1 mM which was then loaded slowly onto a 1-ml column containing glutathione agarose (Sigma) previously equilibrated with buffer A. The eluate was re-applied to the column two times followed by extensive washing with 1 mM sodium phosphate buffer, pH 7.3, containing 15 mM NaCl, 0.5% v/v Triton X-100. The fusion proteins were eluted in a pH 8.0 buffer containing 50 mM Hepes, 10% v/v ethylene glycol, 15 mM reduced glutathione, and 0.4% w/v sucrose monolaurate. In order to isolate full-length peptides the crude preparation was further purified by metal affinity chromatography using TALON resin (CLONTECH, Palo Alto, CA) according to the manufacturer's procedures. Fractions containing the purified fusion protein (detected by SDS-polyacrylamide gel electrophoresis followed by Coomassie Blue staining) were pooled and dialyzed extensively against 50 mM Hepes, pH 8.0, containing 10% v/v ethylene glycol. The resultant product was finally concentrated by dialysis against the same buffer saturated with polyethylene glycol and stored at Ϫ80°C in the presence of 20% v/v glycerol.
Preparation of PKC-(His) 6 -To facilitate isolation and purification, a (His) 6 affinity tag was added to the C terminus of PKC. Briefly, the last 1100 base pairs were amplified by PCR using Pfu polymerase (Stratagene, La Jolla, CA) and using PKC cDNA as a template, which was cloned in this laboratory. 2 The reverse primer was designed so that the stop codon was eliminated, and the (His) 6 sequence (CATCACCAT-CACCATCACTGA) followed by a stop codon and an HpaI restriction site was inserted in frame with the coding sequence. The PCR product was gel-purified on 1% agarose, A-tailed using TAQ polymerase and dATP, and subcloned into pCR 2.1 (Invitrogen, Carlsbad, CA). The nucleotide sequence was confirmed by dideoxy sequencing (Nucleic Acid Facility, Thomas Jefferson University). The first 1203 nucleotides of PKC were excised from PKC/pCR2.1 using EcoRI/BlnI, and the last 576 nucleotides including the (His) 6 tag were excised from PKC-3Ј(His) 6 /pCR2.1 using BlnI/HpaI. The fragments were gel-purified, ligated into the EcoRI/StuI site of pFastBac1 (Life Technologies, Inc.), and transformed into DH5␣ E. coli. A clone containing the full-length PKC coding sequence including the (His) 6 tag was selected by restriction analysis of plasmid DNA. A recombinant baculovirus containing the PKC-(His) 6 sequence was generated using the Bac-to-Bac Baculovirus Expression System (Life Technologies, Inc.). Sf9 cells were infected with the recombinant baculovirus at a multiplicity of infection of 5, incubated for 3 days at 27°C, harvested by centrifugation, and pellets stored at Ϫ80°C.
PKC␣ Activity-PKC␣ activity was determined in the presence of protamine sulfate, which acts both as a phosphate acceptor and an activator of the enzyme, as described previously (34). Briefly, the assay consisted of 50 mM Tris/HCl, pH 7.4, PKC␣ (0.04 ng/l or 0.52 nM final), and protamine sulfate (0.4 mg ml Ϫ1 ). To this was added either TPA or 4␣-TPA from 0.5 mM dimethyl sulfoxide (Me 2 SO) stock solutions, OAG from a 10 mM methanol stock, or bryostatin-1 from a 50 M methanol stock, so that the total assay volume was 75 l. The assay was allowed to thermally equilibrate to 30°C and initiated by the addition of ATP (15 M), [␥-32 P]ATP (ϳ1 Ci), and MgCl 2 (15 mM) in 50 mM Tris/HCl, pH 7.4. After 30 min, the assay was quenched by the addition of 100 l of phosphoric acid (175 mM), 100 l of which was transferred to Whatman P81 anion exchange papers. These were washed three times with 75 mM phosphoric acid, air-dried, and placed into scintillation mixture. The incorporation of 32 P into protamine sulfate was then measured by scintillation counting. Results are expressed as specific activities. The linearity of the assay was confirmed in separate experiments (results not shown).
The concentration of phorbol ester, diacylglycerol, or bryostatin-1 required to reduce protamine sulfate-induced activity by 50% (IC 50 ) and the corresponding Hill coefficient (n) were determined by fitting activity data to a linear Hill equation by linear regression analysis (53) as shown in Equation 1.
where, v 0 and v I are the reaction velocities in the absence and presence of inhibitor.
Phorbol Ester Binding to Non-membrane PKC␣-Binding of the fluorescent phorbol ester, SAPD, to PKC␣, PKC-(His) 6 , or GST-C1-(His) 6 in the absence of membranes was quantified as described previously by measuring the resonance energy transfer (RET) from tryptophans to the 2-(N-methylamino)benzoyl fluorophore attached at the 12-position of the phorbol moiety (35). The fluorescence intensities, obtained upon excitation of the tryptophan fluorophore at 290 nm, were determined using a PTI Alphascan spectrofluorimeter (Photon Technology International, Inc., South Brunswick, NJ) at 333 and 425 nm, corresponding to the emission maxima of tryptophan and SAPD, respectively. The binding assay system (2 ml) consisted of 50 mM Tris/HCl, pH 7.4, 5 g/ml (or 0.12 M) PKC␣, PKC-(His) 6 , or GST-C1-(His) 6 and protamine sulfate at the indicated concentrations. To this was added TPA, 4␣-TPA, OAG, or bryostatin-1 at the required concentration. After incubation for 10 min at 30°C in a stirred quartz cuvette, SAPD was titrated from stock solutions of the required concentration prepared from a Me 2 SO solution of the phorbol ester. The fluorescence intensities were recorded for each phorbol ester addition after allowing equilibration. To isolate the fluorescence signal resulting from RET, the observed fluorescence intensities at each SAPD concentration were corrected for volume changes incurred during the titration procedure and normalized for the contribution from the direct excitation of the SAPD fluorophore according to the following: ϪPKC are the fluorescence intensities measured after each SAPD addition, in the presence and absence of PKC␣, respectively, and F 0,ϩPKC and F 0,ϪPKC are the fluorescence intensities measured in the absence of SAPD, in the presence and absence of PKC␣, respectively. The resultant data were fitted by nonlinear least squares analysis to a modified Hill equation, assuming a model involving two independent SAPD-binding sites (35) as shown in Equation 2.
where F min, H , F max, H , F min, L , and F max, L are the minimum and maximum fluorescence intensities (i.e. the corrected RET signal for saturation of the high and low affinity phorbol ester-binding sites by SAPD, addition of further SAPD not further increasing the signal); K H and K L are binding constants (defined as the SAPD concentration corresponding a half-maximal fluorescence intensity increase); and n H and n L are the Hill coefficients for high and low affinity binding, respectively. Upon activation by protamine sulfate there was small decrease in the (nonmembrane) PKC tryptophan fluorescence intensity, possibly due to a conformational change as the enzyme became active; however, this does not affect the validity of Equation 2 for non-membrane-associated PKC.
Fluorescence anisotropy measurements of SAPD were made using an SLM 48000 in T format anisotropy mode, excitation being at 355 nm and emission at Ͼ390 nm (using a red pass filter). The fluorescence anisotropy was determined as described elsewhere (54).

RESULTS
Previous studies from this laboratory demonstrated the existence of high and low affinity phorbol ester-binding sites on membrane-associated PKC␣ (34,35). In order to determine if these binding site(s) "pre-exist" on lipid-independent PKC␣ or whether they are exposed upon membrane association, phorbol ester binding to non-membrane PKC␣ was studied. The effects of TPA, SAPD, OAG, and bryostatin-1 on non-membrane PKC␣ activity induced by protamine sulfate was also determined. To control for potential nonspecific interactions of these compounds with non-membrane PKC␣, phorbol ester binding to the isolated C1 domain of this isoform, and also to the atypical isoform, PKC, which is incapable of binding phorbol esters (55,56), was determined. The binding assay used was based on the increase in fluorescence intensity due to RET from tryptophans to the fluorophore of the phorbol ester, SAPD, as recently described (35).
Phorbol Ester Binding to Non-membrane PKC␣ in Isolation and in the Presence of Protamine Sulfate-Interaction of phorbol esters with PKC␣ was determined using a binding assay based on the increase in fluorescence intensity resulting from RET from tryptophans to the fluorophore of the phorbol ester, SAPD, as recently described (35). The binding isotherm obtained for the interaction of SAPD with PKC␣ in the absence of protamine sulfate (or membrane lipids), shown in Fig. 1A, was found to be "dual sigmoidal," indicating that even alone, PKC␣ contains two SAPD-binding sites of low and high affinity, respectively. Also shown in Fig. 1A is a comparison of the binding isotherm obtained for SAPD binding to this non-membrane form of PKC␣ with that determined previously for the enzyme associated with membrane lipid vesicles consisting of POPC and BPS (35). The binding constant for low affinity SAPD binding to non-membrane PKC␣ in the absence of lipids, K H , obtained by fitting the binding data shown in Fig. 1A to Equation 2, was similar to that determined for low affinity binding to the membrane-associated enzyme (Table I). Low affinity SAPD-binding site is therefore suggested to be lipid-independ-ent (i.e. it does not require membrane association). By contrast, the strength of the high affinity interaction of SAPD with non-membrane PKC␣ was ϳ5-fold less than that for high affinity binding to the enzyme associated with membranes containing PS (Table I). In control experiments TPA was found to cause a small blue shift in the tryptophan emission maxima and a small increase in the fluorescence intensity, reflecting a conformational change in the enzyme. Whether this occurs with SAPD and thus influences the distance-dependent RET signal is difficult to determine, since addition of SAPD reduces the tryptophan signal due to fluorescence quenching as part of the RET process. However, in separate measurements of the tryptophan emission tail at 425 nm, where the RET signal is measured, obtained by adding the same TPA concentration as used in the SAPD binding data, it was possible to approximately correct for this. The resulting dual sigmoidal SAPD binding curve was not significantly influenced by this small correction.
Binding isotherms obtained for the interaction of SAPD with non-membrane PKC␣ in the presence of increasing concentrations of protamine sulfate were again dual sigmoidal (Fig. 1B), further indicating the existence of low and high affinity phorbol ester-binding sites on the non-membrane-associated enzyme. The effect of increasing protamine sulfate concentration, shown in Table I, was to increase the resonance energy transfer signal proportionately with increasing SAPD concentration. However, neither the affinities (K H and K L ) nor the Hill coefficients (n H and n L ) for high and low affinity SAPD binding were found to be affected by interaction with protamine sulfate (Table I).
Effects of the Phorbol Ester, TPA, and Its "Inactive" Epimer, 4␣-TPA, the Diglyceride OAG, and Bryostatin-1 on SAPD Binding to Non-membrane PKC␣ in the Presence of Protamine Sulfate-In order to determine the specificity of low and high affinity phorbol ester binding to non-membrane PKC␣ associated with protamine sulfate, the effects of TPA and the inactive 4␣-OH epimer, 4␣-TPA, on SAPD binding was first determined. This was accomplished by utilizing the increase in fluorescence anisotropy of SAPD as it binds to the phorbol ester-binding site, reflecting a more restricted motion compared with membraneassociated or free SAPD. Addition of SAPD (1 M, sufficient to saturate the high affinity site), in the absence of lipids, led to a marked increase in r s consistent with binding to the non-membrane enzyme (Fig. 2A). Addition of TPA (1 M) resulted in a decrease in r s , due to the displacement of bound SAPD. The  a Binding constants and Hill coefficients were calculated using Equation 2 for PKC␣ in the absence of protamine sulfate and in the presence of BPS:POPC membrane lipid vesicles (1:4 molar, 150 M total lipid concentration) and Ca 2ϩ (0.1 mM), as previously described (35).
b Binding constants and Hill coefficients for SAPD binding to the isolated C1 domain of PKC␣ (GST-C1-(His) 6 ).
inactive epimer 4␣-TPA (1 M) also displaced SAPD but at a slower rate indicating that the affinity of 4␣-TPA for nonmembrane PKC␣ is reduced compared with TPA. Addition of SAPD (1 M) to PKC␣, in the presence of lipids (150 M) as PS:PC vesicles (1:4, molar), again resulted in an increase in r s , consistent with binding (Fig. 2B). This was enhanced by addition of 0.1 mM Ca 2ϩ , sufficient to induce complete membrane association. Addition of TPA (1 M) again displaced high affinity SAPD binding, as found for non-membrane-associated PKC␣. However, high affinity SAPD binding was not inhibited by 4␣-TPA (1 M). These results clearly show the competition of TPA for SAPD binding and show that the phorbol ester-binding site properties of PKC change markedly upon membrane association.
Assessment of binding using RET from PKC tryptophans to SAPD to measure binding is shown in Fig. 3. The binding curves shown in Fig. 3A suggest that TPA competed for both low and high affinity SAPD binding to the non-membrane form of the enzyme, as observed for membrane-associated PKC␣. Similar to TPA, 4␣-TPA also competed for both high and low affinity SAPD binding to PKC␣ associated with protamine sulfate, although less effectively (Fig. 3A). These effects contrast with those on SAPD binding to the membrane-associated enzyme, as previously observed in studies from this laboratory, where it was found that 4␣-TPA competed only for low affinity SAPD binding, whereas high affinity binding was unaffected (35) in keeping with the observation shown in Fig. 2 (above).
Similar to the effects of the phorbol esters, TPA and 4␣-TPA, Fig. 3B shows that the presence of OAG resulted in an inhibition of both high and low affinity SAPD binding to protamine sulfate-activated PKC␣. This result again contrasts with previously observed effects of diacylglycerol on SAPD binding to membrane-associated PKC␣, where it was found that while low affinity SAPD binding was inhibited, the level of high affinity binding was enhanced (35). Along with phorbol esters and diacylglycerols, bryostatin-1 has also been commonly described as a potent "activator" of PKC (57). However, despite this similarity, bryostatin and phorbol esters have been shown to have dramatically different effects on PKC-regulated cellular processes. Indeed, it appears that bryostatin may antagonize many of the cellular effects induced by phorbol esters, such as tumor promotion (58 -62). It was therefore of interest to determine whether the effects of bryostatin on non-membrane PKC␣ activity induced by protamine sulfate in the absence of lipids also differed from those on the membrane-associated enzyme. Fig. 3C shows that protamine sulfate-associated PKC␣ was also inhibited by bryostatin-1. These effects contrast with those on binding to membrane-associated PKC␣ observed previously, where high affinity SAPD binding was found to be inhibited by bryostatin-1, whereas the low affinity SAPD interaction was relatively unaffected (35).
Interaction of SAPD with the Isolated C1 Domain of PKC␣ (GST-C1-(His) 6 ) and with PKC-To address the possibility that the compounds studied may interact "nonspecifically" with a site(s) outside of the C1 domain, binding of SAPD to the isolated C1 domain of PKC␣ was determined. The fusion pro- tein GST-C1-(His) 6 contains a single tryptophan within the C1B region, as well as four tryptophan residues within the GST moiety, thus allowing for determination of binding from measurements of RET to the SAPD fluorophore. The SAPD binding isotherm obtained for the isolated C1 domain was found to be dual sigmoidal, as shown in Fig. 4A, providing evidence that both the high and low affinity SAPD-binding sites, observed to exist on non-membrane PKC␣, may be confined within the C1 domain of this isoform. The strengths of both high and low affinity SAPD interactions with the isolated C1 domain were marginally reduced compared with the native intact enzyme, observed as an increase in K H and K L (Table I). However, both high and low affinity SAPD-binding sites appeared to have similar specificities to native PKC␣, in that both interactions were inhibited by TPA, 4␣-TPA, OAG, and bryostatin-1 (Fig.  4A). The possibility that SAPD may bind in a nonspecific manner to the GST or (His) 6 moieties was ruled out by the finding that the level of RET from the tryptophans of GST-(His) 6 was insignificant within the phorbol ester concentration range used. Furthermore, evidence against the low and high affinity SAPD interactions being of a nonspecific nature outside the C1 domain was provided by the finding that SAPD failed to bind to PKC, as shown in Fig. 4B, which is in keeping with the inability of this isoform to bind phorbol esters (56).
Effects of Phorbol Esters on Non-membrane, Protamine Sulfate-induced PKC␣ Activity-PKC is effectively activated by protamine sulfate without the need for membrane association or Ca 2ϩ ; however, whether this non-membrane activity is af-fected by phorbol esters (or other activators) is currently not known. Therefore, it was important in the present study to determine the effects of phorbol esters on protamine sulfateinduced PKC␣ activity. Membrane lipid-independent PKC␣ activity induced by protamine sulfate was potently inhibited in a concentration-dependent manner both by the fluorescent phorbol ester, SAPD, and also by TPA, as shown in Fig. 5, A and B, respectively. This surprising observation contrasts completely with the fact that phorbol esters activate the membrane-associated enzyme (35). Interestingly, the "non-active" phorbol ester, 4␣-TPA, also inhibited this activity, although the potency of the inhibition by this phorbol ester was less than with TPA ( Fig. 5A and see Table II). The Hill coefficients, shown in Table II, for the inhibition of activity by TPA and 4␣-TPA and SAPD were ϳ1, again indicating that the effects of both phorbol esters may be mediated by at least one inhibitory site. Furthermore, the IC 50 for the inhibition of protamine sulfate-induced activity by SAPD, obtained by fitting the doseresponse curve shown in Fig. 5B to Equation 1 by linear regression (Table II), was similar to the SAPD concentration required for half-maximal binding to the high affinity site, determined from nonlinear regression analysis of the data shown in Fig. 1A using Equation 2 (Table I). This suggests that the inhibitory effect of phorbol esters may, at least in part, be mediated by the high affinity phorbol ester-binding site. Finally, the inhibition of protamine sulfate-induced activity by SAPD was ϳ5-fold greater than the concentration of phorbol ester required for the half-maximal activation of membraneassociated PKC␣ (EC 50 ), in keeping with the increase in the strength of high affinity binding observed upon membrane association (see Table II).
Michaelis plots of reaction velocity as a function of protamine sulfate concentration for the inhibitory effects of TPA are shown in Fig. 6. Fitting the reaction velocity data, obtained in the absence of TPA, to the Hill equation by nonlinear regression analysis, yielded a Hill coefficient of 1. reciprocal of specific activity against the reciprocal of protamine sulfate concentration yielded a curve that deviated positively from linearity, diagnostic of positive cooperativity with respect to protamine sulfate concentration (result not shown). In keeping with the results of a previous study (6), these data suggest that the stimulation of PKC␣ activity induced by protamine sulfate involves interaction with at least one allosteric binding site(s) on the enzyme molecule, the inhibition would then be attributable to a reduction in the maximal level of activity attainable (i.e. V max ).
Effects of OAG, and Bryostatin-1 on Non-membrane, Protamine Sulfate-induced PKC␣ Activity-The diglyceride, OAG, also inhibited protamine sulfate-induced PKC␣ activity in a concentration-dependent manner (Fig. 7A), although the potency of this inhibitory effect was less than for TPA, 4␣-TPA, or SAPD (Table II). This inhibitory effect again contrasts with the well described potentiating effect of diacylglycerols on membrane-associated PKC. The Hill coefficient, shown in Table II, for the inhibitory effect of OAG was ϳ1, again indicating at least one site of action. Interestingly, for OAG, the IC 50 for the inhibition of non-membrane, protamine sulfate-induced PKC␣ activity was similar to the EC 50 for the activation of the membrane-associated enzyme.
Finally, bryostatin-1 also inhibited protamine sulfate-induced PKC␣ activity ( Fig. 7B and see Table II). The potency of this inhibitory effect was ϳ100-fold greater than that observed for TPA. The Hill coefficient for this process was ϳ1 again indicating at least one inhibitory site, as found for SAPD, TPA, 4␣-TPA, and OAG (see Table II). The potency of inhibition of non-membrane, protamine sulfate-induced PKC␣ activity (IC 50 ) by bryostatin-1 was similar to the EC 50 for the activation of the membrane-associated enzyme (see Table II).

DISCUSSION
In the present study, it is shown that the low and high affinity phorbol ester-binding sites, previously shown to exist on membrane-associated PKC␣ (35), also exist on soluble PKC␣. The results show that although membrane-association is not required to expose these sites, the affinities and specificities of the sites are modified by this interaction. Furthermore, evidence is also presented that both high and low affinity phorbol ester-binding sites may be confined within the C1 domain of the enzyme. However, binding of phorbol esters, which are commonly classed as activators of membrane-associated PKC, resulted in a potent inhibition of lipid-independent enzyme activity induced by protein-protein interactions with the arginine-rich protein, protamine sulfate. Furthermore, two other important activators of membrane-associated PKC, diacylglycerol and bryostatin-1, competed for both high and low affinity phorbol ester binding and also inhibited protamine sulfate-induced activity.
The observation that PKC␣ alone bound SAPD in the absence of protamine sulfate (or membrane lipids) is consistent with membrane association not being an absolute requirement for phorbol ester binding, as observed in several other studies. For example, it has been previously observed that phorbol esters are capable of inducing a low level of PKC activity in the absence of membranes or protein elements, providing evidence for a lipid-independent interaction of phorbol esters with PKC (63). Finally, it has been shown that the novel Ca 2ϩ -independent isoform, PKC␦, and also a peptide corresponding to the C1B region, bound phorbol ester in the absence of lipids (22). In the present study, binding of phorbol ester to non-membrane PKC␣ utilized SAPD as a probe for low and high affinity binding sites, as previously observed to exist on the membrane-associated  50 and n for non-membrane, protamine sulfate-induced PKC␣ activity were calculated by fitting concentration-response data for each compound by linear regression to Equation 2.
b EC 50 and n for the activation of membrane-associated PKC␣ were obtained from activator concentration-response curves determined in the presence of POPC:BPS lipid vesicles, as previously described (35).
c Membrane-associated PKC␣ activity was previously shown to be unaffected by 4␣-TPA under similar assay conditions (35).  (35). Phorbol ester binding to PKC is commonly determined using phorbol 12,13-dibutyrate (PDBu) which, being relatively hydrophilic compared with TPA (or SAPD), minimizes nonspecific interactions and enables the physical separation of bound from free ligand. However, this precludes the detection of low affinity binding. By contrast, the SAPD binding assay used in the present study does not require physical separation of bound from free ligand, and the affinity of SAPD for binding to PKC␣ is ϳ10-fold greater than that of PDBu (64), which allows for the detection low affinity phorbol ester binding. The results show that both the high and low affinity phorbol ester-binding sites, previously shown to be present on the membrane-associated enzyme (35), pre-exist on non-membrane PKC␣. Furthermore, the observation that the isolated C1 domain of PKC␣ also bound SAPD with high and low affinities provides evidence that, as for the high affinity site, the low affinity phorbol ester-binding sites may also be contained within this domain, rather than residing elsewhere on the PKC molecule. This is also apparent from the observation that PKC failed to bind SAPD, in keeping with the inability of this isoform to bind other phorbol esters. However, further experiments are required to determine the nature of the low affinity phorbol ester-binding site within the C1 domain and whether this site corresponds to either the C1A or B subdomains.
Comparison of isotherms corresponding to SAPD binding to membrane-associated PKC␣ with that obtained for binding to the soluble isozyme (Fig. 1A) revealed that interaction with membrane lipid vesicles containing PS results in an ϳ5-fold increase in the strength of high affinity SAPD binding. This is in keeping with the results of studies by Kazanietz and coworkers (22), who showed that the structure-activity relationship for phorbol ester binding to soluble, non-membrane PKC␦, differed markedly compared with that for binding to the membrane-associated isoform. For example, although it was shown that the affinity of PDBu for binding to PKC␦ in the presence of phosphatidylserine (PS) was enhanced 80-fold compared with the soluble isozyme, the binding affinity of phorbol-12-decanoate was found to be only enhanced 6-fold.
In the present study it was shown that 4␣-TPA, TPA, OAG, and bryostatin-1 each displaced both high and low affinity SAPD binding to native PKC␣ and to the isolated C1 domain, suggesting that both sites on the non-membrane enzyme may have similar binding specificities. By contrast, the low and high affinity phorbol ester-binding sites on membrane-associated PKC␣ have previously been shown to have differing specificities (35). In particular, the observation that 4␣-TPA inhibited high affinity SAPD binding to non-membrane PKC␣ while having negligible effects on high affinity SAPD binding to the membrane-associated enzyme (Fig. 2) suggests that this site on non-membrane PKC␣ may have a reduced specificity for the orientation of the 4-OH and/or the A and B rings of SAPD, compared with that on the membrane-associated enzyme. Furthermore, whereas bryostatin-1 appears to compete for low affinity SAPD binding to non-membrane PKC␣ and to the isolated C1 domain, this compound failed to compete for low affinity SAPD binding to the membrane-associated enzyme, suggesting that the specificities of the low affinity sites on non-membrane and membrane-associated PKC␣ may also differ. In keeping with this was the finding that the ratios of the values of IC 50 for the inhibition of non-membrane PKC␣ to the corresponding values of EC 50 for the activation of the membrane-associated enzyme differed for the compounds studied (see Table II).
The mechanism of activation of PKC by arginine-and lysinerich proteins has been suggested to involve the formation of high order aggregates (6, 7). Also consistent with the formation of an aggregate was the observation that the specific activity of protamine sulfate phosphorylation displayed positive cooperativity with respect to protamine sulfate concentration, as observed previously (6). Importantly, the observation that neither the Hill coefficient nor the mid-point of the specific activity against protamine sulfate curve was affected by the presence of SAPD argues against the possibility that the inhibition of protamine sulfate activity resulted from an effect on PKC␣-protamine sulfate interactions. Rather, the apparent decrease in the maximal reaction velocity suggests that the inhibitory effect involves an attenuation of the activating conformational change induced by interaction with the arginine-rich protein. This conformational change could provide a basis for the observed increase in RET induced by protamine sulfate, since the molecular rearrangement would result in a change in the average distance between participating PKC␣ tryptophan and SAPD fluorophores.
The finding that the value of the IC 50 for the inhibition of protamine sulfate activity by SAPD was close to the phorbol ester concentration required for half-maximal binding to the high affinity site on protamine sulfate-associated PKC␣ strongly suggests that the inhibitory effect may be mediated by interaction with this site. However, based on the present data, it is not possible to determine whether, similar to SAPD, the inhibitory effect of these compounds is mediated specifically by interaction with the non-membrane high affinity phorbol esterbinding site per se. Only direct measurements of binding of these compounds to the PKC␣ would resolve this issue, which is technically difficult to achieve at the sensitivity required. Interaction of phorbol esters, diacylglycerol, and bryostatin with the SAPD-binding sites on PKC␣ appears to inhibit the protamine sulfate-induced activating conformational change, based on the observation that the inhibitory effect of SAPD on protamine sulfate-induced activity corresponded to a reduction in the maximal reaction velocity (V max ), rather than an effect on the PKC␣-protamine sulfate interactions. Conversely, the protamine sulfate-induced activating conformational change does not appear to impact on phorbol ester binding, based on the observation that neither the binding constants nor the Hill coefficients for high and low affinity SAPD binding changed significantly in the presence of increasing protamine sulfate concentrations. These observations suggest that the inhibition of activity by phorbol esters (and other activators) may proceed by an "uncompetitive" type mechanism in which the effect of SAPD binding is to prevent rather than reverse the activating conformational change.
Finally, the observed differences in the specificities of the low and high affinity phorbol ester-binding sites on non-membrane and membrane-associated PKC␣ may partially contribute to the distinct effects of bryostatin, compared with phorbol esters, on PKC-regulated cellular processes adding to differences in down-regulation, isoform-selectivity, and cellular location.