Proteolytic Activation of Protein Kinase C-ε by Caspase-mediated Processing and Transduction of Antiapoptotic Signals*

Several novel protein kinase C (PKC) isozymes have been identified as substrates for caspase-3. We have previously shown that novel PKCε is cleaved during apoptosis in MCF-7 cells that lack any functional caspase-3. In the present study, we show thatin vitro-translated PKCε is processed by human recombinant caspase-3, -7, and -9. Tumor necrosis factor-α (TNF) triggered processing of PKCε to a 43-kDa carboxyl-terminal fragment, and cell-permeable caspase inhibitors prevented TNF-induced processing of PKCε in MCF-7 cells. PKCε was cleaved primarily at the SSPD↓G site to generate two fragments with an approximate molecular mass of 43 kDa. It was also cleaved at the DDVD↓C site to generate two fragments with molecular masses of 52 and 35 kDa. Treatment of MCF-7 cells with TNF resulted in the activation of PKCε, and mutation at the SSPD↓G (D383A) site inhibited proteolytic activation of PKCε. Overexpression of wild-type but not dominant-negative PKCε in MCF-7 cells delayed TNF-induced apoptosis, and mutation at the D383A site prevented antiapoptotic activity of PKCε. These results suggest that cleavage of PKCε by caspase-7 at the SSPD↓G site results in the activation of PKCε. Furthermore, activation of PKCε was associated with its antiapoptotic function.

Protein kinase C (PKC) 1 is a family of phospholipid-dependent serine/threonine kinases that play a central role in the growth factor signal transduction pathway and regulate a wide variety of cellular functions, including cell proliferation, differentiation, and cell death (1). PKC represents a family of 11 isozymes that have been categorized into three groups: group A or conventional PKCs (␣, ␤I, ␤II, and ␥), group B or novel PKCs (␦, ⑀, , and ), and group C or atypical PKCs ( and /) (2)(3)(4). In addition, PKC resembles novel PKCs structurally but resembles atypical PKCs functionally (5). Whereas conventional PKCs require Ca 2ϩ and diacylglycerol/phorbol esters for their activities, novel PKCs and atypical PKCs are Ca 2ϩ -independent. Atypical PKCs are also insensitive to diacylglycerol and phorbol esters.
In the native conformation, the regulatory domain of PKC interacts with the catalytic domain through pseudosubstrate sequences and prevents access of substrates to the catalytic site (3). The binding of cofactors, such as phosphatidylserine, Ca 2ϩ , and diacylglycerol/phorbol esters, to the regulatory domain induces a conformational change in the enzyme, thereby exposing the substrate-binding site, and catalysis takes place. The proteolytic cleavage of PKC at the hinge region can also separate the regulatory domain from the catalytic domain. The catalytic fragment (PKM) thus generated does not require any activators or cofactors for activation. In the case of conventional PKCs, proteolytic activation can be achieved by Ca 2ϩactivated proteases, calpains.
It has been shown that novel PKC isozymes, including PKC␦, -, and -, are substrates for caspases (6 -8), a family of cysteine proteases that specifically cleave proteins after Asp residues and play an essential role in the induction of apoptosis (9,10). All caspases exist as inactive proenzymes, which are proteolytically processed to the active heterodimeric form. To date, 14 caspases have been identified. Whereas caspase-8, -9, and -10 participate in the initiation phase of apoptosis, caspase-3, -6, and -7 are involved in the execution phase of apoptosis (10). Activation of these executioner caspases results in the cleavage of critical cellular proteins, including poly(ADP-ribose) polymerase (PARP), DNA-dependent protein kinase, and lamin B.
Cleavage of novel PKC isozymes by caspases also generates catalytically active carboxyl-terminal fragments. Proteolytic activation of these novel PKC isozymes has been directly associated with cell death (6 -8). No detectable cleavage of PKC␣, -␤, -⑀, and -was noted in these studies, and it was claimed that novel PKC␦ and -were selectively involved in apoptosis (7). We and others have recently shown that both novel PKC⑀ and atypical PKC are also cleaved by apoptotic stimuli (11)(12)(13). Atypical PKC serves as a substrate for several caspases, including caspase-3, -6, -7, and -8 (13). Caspase-mediated processing of atypical PKC generates carboxyl-terminal fragments that are catalytically active. Although the functional significance of atypical PKC processing by caspases is not clear, it has been postulated that PKC activation may promote cell survival (13).
Novel PKC⑀ plays a very important role in cell survival and cell death (12, 14 -19). Although novel PKC␦, -, and -are * This work was supported by Grants CA71727 (to A. B.) and CA85682 (to A. B.) from the NCI, National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 1 The abbreviations used are: PKC, protein kinase C; z-DEVD-fmk, benzyloxycarbonyl-Asp-Glu-Val-Asp-7-amino-4-fluoromethylketone; z-IETD-fmk, benzyloxycarbonyl-Ileu-Glu-Thr-Asp-fluoromethylketone; z-LEHD-fmk, benzyloxycarbonyl-Leu-Glu-His-Asp-fluoromethylketone; DN, dominant-negative; MBP, myelin basic protein; PARP, poly(ADPribose) polymerase; PI, propidium iodide; TNF, tumor necrosis factor-␣; substrates for caspase-3, TNF induced cleavage of PKC⑀ in MCF-7 breast cancer cells that lack functional caspase-3 (19). Little is known about the regulation of PKC⑀ activity by caspases and the functional significance of caspase-mediated processing of PKC⑀ on cell death. In the present study, we report for the first time that PKC⑀ is cleaved by caspase-7, proteolytic cleavage of PKC⑀ by TNF results in its activation, and inhibition of PKC⑀ activity abrogates its antiapoptotic activity.

EXPERIMENTAL PROCEDURES
Materials-TNF was purchased from R&D Systems (Minneapolis, MN), and cell-permeable caspase inhibitors were obtained from Kamiya Biomedical Co. (Seattle, WA). Calpeptin was from Calbiochem-Novabiochem. Annexin V conjugated to Alexa Fluor 488 and PI were purchased from Molecular Probes (Eugene, OR). Human recombinant caspase-7 and -8 and monoclonal antibody to PARP were from Pharmingen. Human recombinant caspase-2, -3, -6, and -9 and caspase substrates were from BioVision (Palo Alto, CA). MBP was purchased from Sigma. Horseradish peroxidase-conjugated goat anti-mouse and donkey anti-rabbit antibodies were obtained from Jackson ImmunoRe-search Laboratory Inc. (West Grove, PA). Polyvinylidene difluoride membrane was from Millipore, and enhanced chemiluminescence detection kit was from Amersham Biosciences.
Cell Culture and Transfection-MCF-7 cells were maintained in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum and 2 mM glutamine and kept in a humidified incubator at 37°C with 95% air and 5% CO 2 . Cells were transfected using Fu-GENE 6 transfection reagent (Roche Molecular Biochemicals) according to the manufacturer's protocol and selected using Geneticin (Invitrogen).
Site-directed Mutagenesis, in Vitro Translation, and Caspase Cleavage Assay-Full-length PKC⑀ was cloned from human muscle cDNA library by PCR and subcloned into pcDNA3. Caspase cleavage sites were mutated using the QuikChange site-directed mutagenesis kit (Stratagene) and the manufacturer's protocol. Dominant-negative PKC⑀ was generated by mutation of Lys 437 to Arg (K437R). Sequences were confirmed by DNA sequencing. [ 35 S]Met-labeled wild-type and mutant PKC⑀ were synthesized by in vitro coupled transcription and translation with the T7 Quick TNT kit (Promega). Labeled proteins were incubated with human recombinant caspases in 50 mM Hepes, pH 7.5, 0.1% CHAPS, 5 mM dithiothreitol, 10% glycerol, and 0.1 mM EDTA at 37°C for 1 h. Proteins were separated by SDS-PAGE, and autoradiography was performed with the dried gel.
Immunoblot Analysis-Equivalent amounts of total cellular extracts were electrophoresed by SDS-PAGE and transferred electrophoretically to polyvinylidene difluoride membrane. Immunoblot analyses were performed as described previously (11).
Immunokinase Assay-Cells were treated with or without TNF as indicated in the text. PKC⑀ was immunoprecipitated using a polyclonal antibody to PKC⑀, and immunokinase assay was performed as described previously (20) using MBP as the substrate in the absence of any cofactors. The reaction mixture was separated by SDS-PAGE, and autoradiography was performed with the dried gel.
Annexin V/Propidium Iodide Binding Assay-Cells were treated with or without TNF for 12 h. At the end of the incubation, both detached cells and attached cells were collected and washed with phosphate-buffered saline. Cells were then stained with annexin V-Alexa 488 conjugate and propidium iodide according to the manufacturer's protocol and analyzed using a flow cytometer (Coulter Epics).

Caspase Inhibitors Prevent Proteolytic Processing of PKC⑀
Induced by TNF-We have previously shown that both DNAdamaging agents and TNF could induce cleavage of PKC⑀ (11,19). To examine whether proteolytic activation of PKC⑀ is mediated by caspases, we examined the effects of several cellpermeable caspase inhibitors on TNF-induced processing of PKC⑀. Fig. 1 shows that treatment of MCF-7 cells with TNF resulted in a decrease in full-length PKC⑀ with a concomitant increase in a fragment with an approximate molecular mass of 43 kDa. Because binding of TNF to its receptors induces activation of apical caspase-8, we examined the effect of z-IETDfmk, a cell-permeable inhibitor of caspase-8, on TNF-induced processing of PKC⑀. As shown in Fig. 1, caspase-8 inhibitor blocked TNF-induced cleavage of PKC⑀. z-DEVD-fmk, an inhibitor of caspase-3 and -7, was as effective as z-IETD-fmk in preventing TNF-induced processing of PKC⑀, whereas z-LEHD-fmk, an inhibitor of caspase-9, exhibited a partial inhibition. In contrast, calpeptin, a cell-permeable inhibitor of calpains, had little effect on the proteolytic cleavage of PKC⑀, even at 50 M concentrations (data not shown). These results suggest that TNF-induced activation of caspases results in the processing of PKC⑀.
In Vitro-translated PKC⑀ Is Cleaved by Recombinant Caspases-To determine whether PKC⑀ is in fact a substrate for caspases, we examined the effect of human recombinant caspases on the cleavage of 35 S-labeled in vitro-translated PKC⑀. Fig. 2 shows that PKC⑀ was cleaved by caspase-3, -7, and -9. Human recombinant caspase-8 had only a slight effect, but caspase-2 and -6 were unable to process PKC⑀ (data not MCF-7 cells transfected with an empty vector, wild-type PKC⑀, or dominant-negative PKC⑀ were treated with 1 nM TNF for the indicated periods of time. PKC⑀ was immunoprecipitated, and an immunokinase assay was performed using MBP as the substrate in the absence of any cofactors. We have used twice as much protein during immunoprecipitation in A (500 g) as compared with B (250 g). Proteins were separated by SDS-PAGE, and autoradiography was performed with the dried gel. Equal amounts of total cellular extracts were subjected to Western blot analysis with PKC⑀ antibody.
FIG. 9. Effect of D451A and D383A mutation on TNF-induced activation of PKC⑀. MCF-7 cells expressing empty vector, wild-type, or mutant PKC⑀ were treated with 1 nM TNF for 10 h, and immunokinase assay was performed as described in the Fig. 8 legend, except that the same amount of protein (250 g) was used for immunoprecipitation in all four cell lines. Equal amounts of total cellular extracts were subjected to Western blot analysis with PKC⑀ antibody.
shown). Cleavage of PKC⑀ by human recombinant caspases generated fragments with apparent molecular masses of 52, 43, and 35 kDa. Caspase inhibitors z-DEVD-fmk and z-LEHD-fmk were able to inhibit processing of PKC⑀ by caspase-3/-7 and caspase-9, respectively (Fig. 2). The cleavage pattern of PKC⑀ generated by caspase-3, -7, and -9 was similar, suggesting that there may be overlapping specificity in an in vitro cleavage assay. However, the 52-kDa band was not detectable in the caspase-7-treated lane, presumably because it was further cleaved to 43-kDa and 9-kDa fragments. Because caspase-7 was most effective in processing PKC⑀, we examined the time course of PKC⑀ cleavage by caspase-7. As shown in Fig. 3, significant processing of PKC⑀ took place by 30 min, and little full-length PKC⑀ remained after incubation of PKC⑀ with caspase-7 for 2 h.
PKC⑀ Is Cleaved at SSPD2G and DDVD2C Sites-To determine the potential caspase cleavage sites in PKC⑀, we mutated several Asp residues to Ala in PKC⑀ by site-directed mutagenesis. As shown in Fig. 4, the intensity of the band near 43 kDa produced by the treatment of in vitro-translated PKC⑀ with caspase-7 was much greater compared with that of the 52-and 35-kDa fragments. Mutation at the DDVD2C (D451A) site abolished cleavage of PKC⑀ to 52-and 35-kDa fragments, whereas mutation at the SSPD2G (D383A) site prevented generation of fragments near 43 kDa. Cleavage of PKC⑀ at the SSPD2G site presumably produced two fragments of identical molecular mass around 43 kDa. During the TNT assay, sometimes we could detect some faint bands, including one near the 43-kDa region, even without any caspase treatment. This protein results from tRNA-dependent, ribosome-independent addition of [ 35 S]methionine on proteins (21). The important point is that the intensity of the 43-kDa band was similar in control and caspase-treated D383A mutant PKC⑀. Thus, it was not generated by the cleavage of PKC⑀ by caspases.
To determine the cleavage pattern of PKC⑀ generated by TNF in intact cells, we overexpressed PKC⑀ in MCF-7 cells and treated them with TNF for varying periods of time. We have used an antibody that recognizes the carboxyl-terminal domain of PKC⑀. Fig. 5 shows that a 6-h exposure to TNF resulted in the generation of a fragment near 43 kDa and that the intensity of the 43-kDa band increased with time. Another band appeared above the 43-kDa fragment after treatment with TNF. The intensity of the band was maximum at 12 h and then rapidly decreased. This band likely represents a phosphoryl-ated form of the 43-kDa fragment because mutation of Asp to Ala of nearby LSFD2N (D370A) and LGLD2E (D406A) sites did not prevent caspase-mediated processing of PKC⑀ (data not shown). TNF also induced cleavage of PKC⑀ to a 35-kDa fragment, and the intensity of the 35-kDa band increased when cells were treated with TNF for Ͼ12 h. In addition, the intensity of the 35-kDa fragment was much less compared with that of the 43-kDa fragment. A faint band with an approximate molecular mass of 50 kDa was also apparent after treatment of the cells with TNF for 24 h. This could represent the carboxylterminal fragment of PKC⑀ processed by caspase-9 (Fig. 2). Thus, PKC⑀ was cleaved primarily at the SSPD2G site in intact cells in response to TNF.
To determine whether mutation of PKC⑀ at the DDVD2C (D451A) or the SSPD2G (D383A) site blocks corresponding cleavage of PKC⑀ in intact cells, we introduced PKC⑀ harboring a mutation at the D451A or D383A site in MCF-7 cells. As shown in Fig. 6, the intensity of the 43-kDa band increased dramatically after treatment of PKC⑀-overexpressing MCF-7 cells with TNF. TNF also induced cleavage of PKC⑀ to 43-kDa fragments in cells expressing the D451A mutant, but mutation at the D383A site prevented TNF-induced processing of PKC⑀ to 43-kDa fragments. These results corroborate that PKC⑀ is processed primarily at the SSPD2G (D383A) site in intact cells. The same blot was probed with tubulin to control for loading differences.
TNF Induces Proteolytic Activation of PKC⑀-To determine whether TNF-induced processing of PKC⑀ results in the activation of PKC⑀, we performed immunokinase assay using MBP as the substrate. As depicted in Fig. 7, proteolytic separation of the autoinhibitory regulatory domain from the catalytic domain is expected to result in the activation of PKC⑀ in the absence of any cofactors. Fig. 8A shows that TNF caused a slight increase in MBP phosphorylation in vector-transfected MCF-7 cells. In contrast, MBP phosphorylation was almost undetectable in cells expressing DN-PKC⑀, even when cells were treated with TNF. TNF had a dramatic effect on MBP phosphorylation in PKC⑀-overexpressing cells (Fig. 8B). We have used twice as much protein during immunoprecipitation of PKC⑀ from vector-transfected and DN-PKC⑀-expressing cells (Fig. 8A) compared with PKC⑀-overexpressing cells (Fig. 8B) to allow detection of PKC⑀ activation in vector-transfected cells and to permit comparison between vector-transfected cells and DN-PKC⑀-expressing cells. Because the SSPD2G site is localized at the hinge region of PKC⑀ (Fig. 7), we examined whether processing at the SSPD2G site was necessary for TNF-induced activation of PKC⑀. We compared the ability of TNF to activate PKC⑀ in cells expressing wild-type or mutant PKC⑀ using an immunokinase assay. Fig. 9 shows that mutation at the D451A site that retains the ability of TNF to generate the 43-kDa catalytic fragment of PKC⑀ was not sufficient to prevent activation of PKC⑀ by TNF.
In contrast, mutation at the D383A site that prevents generation of the 43-kDa fragment abolished TNF-induced activation of PKC⑀. Thus, processing of PKC⑀ by TNF at the D383A site results in its activation.
Activation of PKC⑀ Is Necessary for Its Antiapoptotic Action-We have previously shown that PKC activators protect cells against TNF-induced cytotoxicity (22). Fig. 10 shows that when cells transfected with a vector containing neomycin re- FIG. 11. Effect of wild-type, D383A mutant, and dominant-negative PKC⑀ on TNF-induced apoptosis. Cells were treated with or without 1 nM TNF for 12 h and then stained with annexin V-Alexa 488 conjugate and PI and analyzed by flow cytometry as described under "Experimental Procedures." sistance gene but without PKC⑀ construct (Fig. 10, Neo) were treated with TNF, cells rounded up and detached from the tissue culture dish. Overexpression of PKC⑀ prevented the morphological changes indicative of apoptosis when cells were treated with TNF for 12 h, although cells started to round up after exposure to TNF for 18 h. Introduction of either D383A mutant or DN-PKC⑀ failed to prevent cell death by TNF, suggesting that activation of PKC⑀ was necessary for the antiapoptotic activity of PKC⑀.
To quantify cell death by apoptosis, we performed an annexin V binding assay (Fig. 11). When cells undergo apoptosis, phosphatidylserine is flipped from the inner to the outer leaflet of plasma membrane. During an early stage of apoptosis, annexin V binds to phosphatidylserine on the cell surface. Cells were co-stained with the cell-impermeable dye PI to distinguish apoptotic cells from necrotic cells. However, during late-stage apoptosis, annexin V can enter through the membrane, and therefore late-stage apoptosis cannot be distinguished from necrosis. Fig. 11 shows a representative dot blot analysis of annexin V conjugate-PI-stained cells. Cells stained with annexin V conjugate alone (bottom right quadrant) represent apoptotic cells, whereas cells co-stained with annexin V conjugate and PI (top right quadrant) represent late apoptotic and necrotic cells. Viable cells are shown at the bottom left quadrant (negative for both annexin V and PI). We have shown the mean Ϯ S.E. of total percentage of cell death from two to three independent experiments at the top of each panel. A small percentage of cells undergo apoptosis even in the absence of any TNF treatment. TNF increased cell death to ϳ15% in vector-transfected MCF-7 cells (Neo), but the effect of TNF was attenuated in cells overexpressing PKC⑀ such that ϳ5% cells underwent apoptosis. In contrast, introduction of mutant PKC⑀ (D383A) caused an increase in apoptotic cell death to 25%. At present, it is not clear why blockage of PKC⑀ cleavage not only inhibited the antiapoptotic effect of PKC⑀ but also potentiated cell death by TNF. Overexpression of dominant-negative PKC⑀ by itself increased cell death to ϳ8%, and TNF further enhanced cell death to 16%. We have consistently found that when DN-PKC⑀-expressing cells are cultured for several passages, their ability to potentiate TNF-induced cell death decreases, presumably because cells that are resistant to apoptosis survive selectively over cells that are prone to apoptosis.
Because MCF-7 cells lack functional caspase-3, they do not undergo DNA fragmentation (23). Therefore, we also assessed cleavage of PARP or DNA-dependent protein kinase to monitor cells undergoing apoptosis. Fig. 12 shows that treatment of MCF-7 cells with TNF resulted in the cleavage of PARP, as evidenced by the appearance of the 85-kDa cleavage product of 116-kDa full-length PARP. Although PKC␦ is cleaved during apoptosis, TNF had no effect on the processing of PKC␦ in MCF-7 cells, which lack caspase-3. Overexpression of PKC⑀ in MCF-7 cells prevented TNF-induced cleavage of PARP and DNA-dependent protein kinase. By contrast, introduction of D383A mutant PKC⑀ in MCF-7 cells failed to prevent TNFinduced cleavage of PARP. These results suggest that activation of PKC⑀ inhibited TNF-induced apoptosis.

DISCUSSION
Protein kinase C plays an important role in regulating cell survival and cell death. Different PKC isozymes may exhibit distinct functions in cell death regulation. Proteolytic activation of novel PKC␦ and -has been shown to be important for the induction of apoptosis (6 -8). Results of our present study demonstrate that PKC⑀ was also cleaved by caspases during apoptosis, but proteolytic cleavage of PKC⑀ prevented its antiapoptotic signal.
Several novel PKC isozymes, including PKC␦, -, and -, have been identified as substrates for caspase-3 (6 -8). Although PKC⑀ serves as a substrate for caspase-3 in an in vitro cleavage assay, it is processed in MCF-7 cells lacking functional caspase-3, suggesting that some caspase other than caspase-3 is responsible for the cleavage of PKC⑀. Caspase-3 and -7 share similar substrate specificities: they both recognize DEVD sequence and cleave PARP (24 -26). However, PKC␦ is not processed by apoptotic stimuli in MCF-7 cells, suggesting that there may be some specificity in intact cells (19). Because both caspase-3 and -7 generated similar cleavage products, it is likely that caspase-7 is responsible for the cleavage of PKC⑀ in MCF-7 cells. Human recombinant caspase-9 also generated cleavage products similar to those of caspase-3 and -7, in addition to two faint bands with approximate molecular masses of 50 and 38 kDa. Thus, caspase-9 may cleave PKC⑀ at an additional site, and there may be overlapping specificity in an in vitro cleavage assay. It has been shown that PKC is processed at the same site by several caspases, including caspase-8, -3, -6, and -7 (13). PKC⑀ contains a DDVD2C site that is recognized by group II caspases, including caspase-3, -7, and -2 (27). Both caspase-3 and -7, but not caspase-2, cleaved PKC⑀ at the DDVD2C site, although it was predominantly processed at the SSPD2G site both in vitro and in intact cells. Because a cell-permeable inhibitor of caspase-3/-7 was more effective than the inhibitor of caspase-9 in preventing generation of the 43-kDa fragment, it is likely that caspase-7 also cleaves PKC⑀ at the SSPD2G site. This site was recognized irrespective of the cell type or apoptotic stimuli. For example, TNF induced processing of PKC⑀ to 43-kDa fragments in several breast cancer cells (19). In addition, treatment of HeLa cells with cisplatin, a DNAdamaging agent, also cleaved PKC⑀ to generate a 43-kDa carboxyl-terminal fragment (11). Whereas caspase-3 recognized conventional caspase cleavage sites in PKC␦ and -, it cleaved PKC at an unconventional CQND2S site, even though PKC possessed DDND2S and DHED2S sites that could be recognized by caspase-3-like proteases (6 -8). PKC was also cleaved primarily at an unconventional EETD2G site by several caspases, whereas caspase-3 and -7 processed PKC at a con- ventional DGMD2G site (13). It is conceivable that the hinge region of PKCs is more accessible to caspases. However, mutation or cleavage of the SSPD2G site at the hinge region may expose other caspase cleavage sites. This may explain why cleavage of PKC⑀ by caspase-7 was more pronounced in the D383A mutant compared with wild-type PKC⑀ (Fig. 4).
The primary cleavage site for all of the PKC isozymes was at the hinge region between the regulatory domain and the catalytic domain. Because the pseudosubstrate sequence at the regulatory domain interacts with the catalytic domain, caspase-mediated cleavage of PKCs separates the catalytic fragment from the autoinhibitory regulatory domain, resulting in activation of PKCs. Thus, in response to apoptotic stimuli, PKCs may be activated in the absence of any cofactors. Cleavage of PKC⑀ by caspases at the hinge region (Ser-Ser-Pro-Asp 383 -Gly) would generate a carboxyl-terminal fragment containing the catalytic domain. In contrast, cleavage of PKC⑀ at the Asp-Asp-Val-Asp 451 -Cys site would generate a carboxylterminal fragment lacking the ATP binding site. We have shown that treatment of MCF-7 cells with TNF results in PKC⑀ activation and that mutation at the D383A site prevents TNFinduced activation of PKC⑀. In contrast, mutation at the D451A site did not affect TNF-induced activation of PKC⑀. Therefore, caspase-mediated processing of PKC⑀ at the Ser-Ser-Pro-Asp 383 -Gly site results in its activation.
PKC activators have been shown to block TNF-induced cell death. In the present study, we have shown that overexpression of PKC⑀ in MCF-7 cells inhibited TNF-induced apoptosis, suggesting that PKC⑀ acts as an antiapoptotic protein. This is consistent with our previous report that overexpression of PKC⑀ in rat embryo fibroblasts delayed apoptosis induced by the DNAdamaging agent cisplatin (17). We have also examined the functional significance of PKC⑀ activation on TNF-induced cell death. In contrast to wild-type PKC⑀, introduction of DN-PKC⑀ in MCF-7 cells by itself caused some cell death and failed to inhibit TNF-induced apoptosis. Blockage of the caspase cleavage site (D383A) of PKC⑀ prevented antiapoptotic activity of PKC⑀, suggesting that activation of PKC⑀ was necessary for the antiapoptotic function of PKC⑀. Thus, our results provide a direct link between activation of PKC⑀ and its antiapoptotic function.