PKCdelta is required for mitochondrial-dependent apoptosis in salivary epithelial cells.

We report here that the novel protein kinase C isoform, PKCdelta, is required at or prior to the level of the mitochondria for apoptosis induced by a diverse group of cell toxins. We have used adenoviral expression of a kinase-dead (KD) mutant of PKCdelta to explore the requirement for PKCdelta in the mitochondrial-dependent apoptotic pathway. Expression of PKCdeltaKD, but not PKCalphaKD, in salivary epithelial cells resulted in a dose-dependent inhibition of apoptosis induced by etoposide, UV-irradiation, brefeldin A, and paclitaxel. DNA fragmentation was blocked up to 71% in parotid C5 cells infected with the PKCdeltaKD adenovirus, whereas caspase-3 activity was inhibited up to 65%. The activation of caspase-9-like proteases by all agents was also inhibited in parotid C5 cells expressing PKCdeltaKD. The ability of PKCdeltaKD to block the loss of mitochondrial membrane potential was similarly determined. Expression of PKCdeltaKD blocked the decrease in mitochondrial membrane potential observed in cells treated with etoposide, UV, brefeldin A, or paclitaxel in a dose-dependent manner. In contrast to the protective function of PKCdeltaKD, expression of PKCdeltaWT resulted in a potent induction of apoptosis, which could be inhibited by co-infection with PKCdeltaKD. These results suggest that PKCdelta is a common intermediate in mitochondrial-dependent apoptosis in salivary epithelial cells.

Apoptosis is a genetically regulated form of cell death that plays a critical role in the destruction of unwanted cells such as tumor cells and cells damaged by viral infection, drugs, chemical radiation, and aging. Critical genes in the apoptotic process have been identified and include the Bcl-2 family of proteins (1), as well as a family of cysteine proteases known as caspases, which specifically cleave their substrates at aspartic acid residues (2). Apoptosis is initiated either by ligand binding to cell surface receptors or by cell toxins via a pathway that targets the mitochondria (3). In the receptor-mediated pathway, ligand binding to the tumor necrosis factor (TNF)␣/FAS family of receptors results in activation of the initiator caspase, caspase 8. Active caspase-8 then activates downstream "effector" caspases such as 3, 6, and 7, which cleave cell proteins ultimately resulting in DNA fragmentation and cell death.
The mitochondrial-dependent apoptotic pathway is induced by agents such as drugs, chemicals, irradiation, and some types of cell stress. Although the specific cell signals delivered by these agents differ, all appear to converge at the mitochondria resulting in the release of cytochrome c, activation of the initiator caspase, caspase-9, and the subsequent activation of downstream caspases. How many diverse signals initiate this common pathway is not known. One possibility is that specific protein kinase cascades may function to integrate these signals upstream of the mitochondria. In line with this, recent evidence implicates specific serine/threonine protein kinases as regulators of apoptosis. These include the phosphoinositide 3-kinase/ AKT pathway (4,5), members of the mitogen-activated protein kinase/extracellular-regulated kinase family (6 -8), and the protein kinase C (PKC) 1 pathway (9 -12).
The PKC family consists of 12 serine/threonine kinases, of which specific isoforms have been shown to be either proapoptotic or anti-apoptotic, depending on the stimulus and cell type (11)(12)(13)(14)(15). For instance, PKC␣ suppresses apoptosis in several cell types, and work from Ruvolo et al. (13) suggests that this occurs through phosphorylation and activation of Bcl-2 (12). In contrast, PKC␦ is required for apoptosis induced by genotoxins (11), phorbol ester (16), and Fas ligand (17). Recent reports from two laboratories suggest that PKC␦ may function at the mitochondria to facilitate apoptosis (16,18). Specific isoforms of PKC are cleaved by caspases to release a catalytically active fragment in cells induced to undergo apoptosis. These include PKC (19), PKC (20), PKC (21,22), and PKC␦ (23)(24)(25). Expression of the PKC␦ or PKC catalytic domain induces apoptosis (19,23). Interestingly, although PKC␦ has been shown to contribute to phorbol ester-induced apoptosis, caspase cleavage of PKC␦ does not occur under these conditions, suggesting a role for the uncleaved form of PKC␦ in apoptosis (9).
We have previously shown that activation of PKC by phorbol ester is sufficient to induce an apoptotic program in parotid salivary acinar cells (10) and that PKC␦ is essential for genotoxin-induced cell death (11). Here we demonstrate that expression of a dominant inhibitory form of PKC␦ from an adeno-viral vector is sufficient to suppress apoptosis in parotid cells in response to diverse stimuli and that PKC␦ is required for loss of mitochondrial membrane potential in apoptotic cells. This argues that PKC␦ is important for the integration of diverse death signals at or prior to their convergence at the mitochondria.

EXPERIMENTAL PROCEDURES
Cells and Cell Culture-The isolation of the salivary parotid C5 cell line has been described elsewhere (26). Cells were cultured on Primaria 60-mm culture dishes (Falcon Plastics, Franklin Lakes, NJ) in a 1:1 mixture of Dulbecco's modified Eagle's medium/nutrient mixture F-12 supplemented with 2.5% fetal calf serum, 5 g/ml transferrin, 1.1 M hydrocortisone, 0.1 M retinoic acid, 2.0 nM T3, 5 g/ml insulin, 80 ng/ml epidermal growth factor (Collaborative Biomedical Products, Bedford, MA), 5 mM L-glutamine, 50 g/ml gentamicin sulfate, and a trace element mixture (Biofluids, Rockville, MD). Tissue culture reagents were obtained from Life Technologies, Inc. (Manassas, VA) unless otherwise indicated. Etoposide, paclitaxel (taxol), and brefeldin A were purchased from Sigma-Aldrich. UVC irradiation was done using a UV cross-linker (Fisher Scientific model UVXL-1000) at a wavelength of 254 nm.
Construction of Adenoviral Vectors and Infection of Parotid C5 Cells-Generation of the wild-type and kinase-dead recombinant rat PKC␦ adenoviruses has been previously described (27). The kinasedead mutant (K376R) has been shown to function as an isoform-specific dominant inhibitory kinase (28). To make the kinase-dead mutant of PKC␣ (PKC␣KD), mutagenesis was performed using a primer that targeted the conserved lysine in the ATP binding domain (TACGCCAT-CAGGATCCTGAAG) of the mouse PKC␣ cDNA. The resulting mutant (K368R) cDNA was subcloned into the adenoviral shuttle vector, pX-CMV, and recombinant adenovirus was prepared essentially as described (29). An adenoviral vector expressing the ␤-galactosidase gene (AdLacZ) was a generous gift from J. Schaack, University of Colorado Health Sciences Center (30). Adenoviruses were titered on 293HEK cells using a focus-forming assay that detects expression of the adenoviral protein E2 (31).
Subconfluent parotid C5 cells were infected with AdLacZ, PKC␦KD, PKC␦WT, or PKC␣KD for one hour at different multiplicities of infection (MOIs) (12-300 focus-forming units/cell) in a 1:1 mixture of Dulbecco's modified Eagle's medium/nutrient mixture F-12 supplemented as described above but without the addition of serum. Following the infection period, the virus-containing medium was replaced with medium containing 2.5% fetal bovine serum, and cells were incubated for an additional 24 h prior to the addition of apoptosis-inducing agents. Unless indicated otherwise, doses of the apoptotic agents were as follows: etoposide, 50 M; UV, 128 J/m 2 ; brefeldin A, 3 M; taxol, 5 M.
Immunoblotting-Adherent and floating cells were scraped into the culture medium, collected by centrifugation (3,000 ϫ g for 10 min), washed once with phosphate-buffered saline, and resuspended in 1 ml of JNK lysis buffer (25 mM HEPES, pH 7.5, 20 mM ␤-glycerophosphate, 0.1 mM sodium orthovanadate, 0.1% Triton X-100, 0.3 M NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, 0.5 mM dithiothreitol, 10 mM NaF, 1 mM phenylmethylsulfonyl fluoride, and 4 g/ml each aprotinin and leupeptin). The lysate was allowed to sit on ice for 30 min and then clarified by spinning at 12,500 rpm for 5 min in a refrigerated Savant SRF13K microcentrifuge. Protein concentration was determined using a Bradford assay kit purchased from Bio-Rad. Cell lysates (25-50 g) were resolved on a 10% gel, transferred to an Immobilon membrane (Millipore), and immunoblotted with the desired antibody as described previously (11). Enhanced chemiluminescence (PerkinElmer Life Sciences) followed by autoradiography was used to detect the signal. Antibodies to all PKC isoforms were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). All anti-PKC antibodies recognize epitopes in the carboxyl-terminal portion of the protein. For immunoblots from cells infected with PKC␦KD, the PKC␣ and PKC antibodies were preabsorbed with four times excess of PKC␦-blocking peptide (Santa Cruz Biotechnologies) by shaking for 1 h at room temperature before use.

Preparation of Cells for Fluorescence-activated Cell Sorting-
The medium containing floating cells was removed and saved. Cells were removed from a P60 dish by the addition of 3 ml of cell dispersion solution (0.25% trypsin-EDTA, 68 M EGTA, pH 7.4, 100 g/ml Pronase) and incubated at 37°C for 15 min. An equal volume of trypsin inhibitor mixture (86 mM NaCl, 30 mM KCl, 1 mM NaH 2 PO 4 , 3 mM MgSO 4 , 0.5 mM CaCl 2 , 15 mM glucose, 18 mM NaCO 3 , 0.5 mM adenosine, 20 mM taurine, 2 mM DL-carnitine, 0.5% bovine serum albumin, and 80 g/ml trypsin inhibitor (Sigma)) was added and a single cell suspension was made by pushing cells 4 times through a 20-gauge needle, 2 times through a 23-gauge needle, and 2 times through a 26-gauge needle. Suspended cells were combined with the medium containing floating cells and centrifuged at 1000 ϫ g for 3 min. The pellet was washed once with phosphate-buffered saline, and the cells were then resuspended in the desired reagent.
Analysis of DNA Content-Cells were prepared for fluorescenceactivated cell sorting (FACS) as described above and the cell pellet was resuspended in 0.5 ml of saponin/propidium iodide solution (0.3 mg/ml saponin, 25 g/ml propidium iodide, 10 mM EDTA, and 5 g/ml RNase). Cells were stained for 5-24 h at 4°C in the dark prior to FACS analysis using a Beckman Coulter XL.
Analysis for Mitochondrial Membrane Potential-Cells were prepared for FACS as described above and stained using the Mitocapture Apoptosis Detection kit obtained from Biovision (Palo Alto, CA). This assay uses an intramitochondrial dye, which forms aggregates in healthy cells leading to increased 575 fluorescence and indicating a normal mitochondrial membrane potential (MMP). In apoptotic cells however, the dye is excluded from the mitochondria leading to loss of 575 fluorescence (aggregate) and an increase in 525 fluorescence (monomer). Data were converted to density plots using System 2 software. In some experiments, FACS data were confirmed by fluorescent microscopy.

Expression of PKC␦KD Inhibits Apoptosis in Response to
Diverse Agents-We have previously demonstrated that etoposide-induced apoptosis in parotid C5 cells is accompanied by specific changes in PKC signaling pathways, which include caspase-dependent cleavage and activation of PKC␦ (11). The studies reported here utilize expression of a dominant inhibitory form of PKC␦ from an adenoviral vector (PKC␦KD) to explore the contribution of PKC␦ to specific events in the apoptotic pathway. In Fig. 1, expression of PKC␦ protein was analyzed by immunoblotting in parotid C5 cells infected with increasing amounts of PKC␦KD. As seen here, the abundance of PKC␦ protein increased with increasing MOI and was 15-20-fold over endogenous levels at an MOI of 150 as determined by densitometery analysis of immunoblots. The abundance of the other major isoforms of PKC expressed in parotid C5 cells, PKC␣ and PKC (11), was unchanged ( Fig. 1). Control studies in parotid C5 cells infected with AdLacZ and stained for ␤-ga- lactosidase expression demonstrated that Ͼ98% of cells were infected at an MOI of 100 (data not shown).
We hypothesize that PKC␦ functions as an integrator of cell toxins that converge at the mitochondria to induce apoptosis. To address this directly, we have analyzed the effect of PKC␦KD expression on apoptosis induced by etoposide, as well as three additional cell toxins that have distinct mechanisms of action. Etoposide inhibits topoisomerase type II, resulting in the accumulation of double-stranded breaks (33). UV irradiation results in direct DNA damage as well as free radicalmediated oxidative damage (34,35). Brefeldin A is a potent inhibitor of vesicular transport and protein secretion and has been proposed to function by inhibiting the membrane binding of COP coat proteins (36,37). Taxol inhibits micotubule dissociation and blocks cells in G 2 /M (38). In the experiment shown in Fig. 2, parotid C5 cells were infected with PKC␦KD or the control virus, AdLacZ, for 24 h and then challenged with etoposide, brefeldin A, or UV irradiation. Apoptosis in response to these agents is indicated morphologically by rounding, blebbing, and detachment of cells from the monolayer (Fig. 2, top row). Infection with PKC␦KD prior to the addition of the stimuli significantly inhibited appearance of the apoptotic morphology (Fig. 2, middle row). However, infection with the control vector, AdLacZ, had no effect on appearance of the apoptotic morphology (Fig. 2, bottom row). PKC␦KD expression similarly protected against apoptosis induced by taxol (data not shown).
To determine whether the suppression of the morphologic characteristics of apoptosis is associated with suppression of biochemical indicators of apoptosis, the ability of PKC␦KD to inhibit DNA fragmentation was determined. Parotid C5 cells were infected with PKC␦KD or AdLacZ, treated with etoposide, stained with propidium iodide, and the percent sub-G 1 DNA determined by FACS analysis. As seen in Fig. 3A, in untreated cells (panel A) or cells infected with PKC␦KD alone (panel D), there is little detectable sub-G 1 DNA; however, nearly all the DNA in etoposide-treated cells is found in the sub-G 1 fraction (panel B). Infection of cells with PKC␦KD prior to the addition of etoposide resulted in a dose-dependent decrease in sub-G 1 DNA and an increase in 2N to 4N DNA (panels C, E, F). Fig. 3B shows the quantification of sub-G 1 DNA in a similar experiment. PKC␦KD suppressed etoposide-induced DNA fragmentation, as indicated by sub-G 1 DNA, by 64% when cells were infected at an MOI of 150 and 72% at an MOI of 300. Expression of PKC␦KD at an MOI of 75 and 150 also slightly suppressed basal DNA fragmentation in infected but untreated cells, an observation that was consistent in multiple experiments. In contrast, infection with AdLacZ offered no protection against etoposide-induced DNA fragmentation. As a control, parotid C5 cells were infected with an adenovirus which expresses a kinase-dead PKC␣ protein (PKC␣KD). As seen in Fig.  3C, in contrast to PKC␦KD, expression of PKC␣KD offered no protection against etoposide-induced apoptosis. Fig. 4 shows a series of similar experiments using UV irradiation, brefeldin A, or taxol to induce apoptosis. As seen here, expression of PKC␦KD resulted in a dose-dependent decrease in sub-G 1 DNA in cells treated with UV (panel A), brefeldin A (panel B), or taxol (panel C). The analysis of multiple experiments using these agents is shown in Table I and shows that suppression of DNA fragmentation by PKC␦KD was 67-71% for cells treated with brefeldin A or UV and 54 -57% for cells treated with etoposide or taxol. Suppression of brefeldin A and UVinduced apoptosis was slightly but significantly greater than suppression of etoposide-induced apoptosis (p ϭ 0.04), and suppression of UV-induced DNA fragmentation was slightly but significantly greater than suppression of taxol-induced apoptosis (p ϭ 0.05).
PKC␦ Functions Prior to the Activation of Caspase-3 and Caspase-9-like Proteases-To address where in the apoptotic pathway PKC␦ functions, we have sequentially analyzed specific events for their dependence on PKC␦. Activation of effector caspases such as caspase-3 is required for dismantling of structural proteins, the activation or inactivation of protein kinases, and DNA fragmentation. In the experiment shown in Fig. 5, parotid C5 cells were infected with PKC␦KD and Ac-DEVD-pNA cleavage, an indicator of caspase-3 activity, was assayed after the addition of etoposide, brefeldin A, taxol, or treatment with UV. Although the extent of caspase-3 activation varied between agents, Ac-DEVD-pNA cleavage was suppressed in all cases by expression of PKC␦KD, indicating that PKC␦ is required for, or at a point prior to, activation of caspase-3 protease. Interestingly, expression of PKC␦KD did not suppress caspase-3 cleavage entirely, and about the same residual level of caspase-3 activity was seen regardless of the apoptotic agent. This suggests that a PKC␦-independent pathway leading to caspase-3 activation is also induced by this group of agents. The suppression of Ac-DEVD-pNA cleavage by PKC␦KD in multiple experiments is shown in Table I. Inhibition of caspase-3 cleavage is similar for all agents; however, PKC␦KD appears to be a slightly, although not significantly, less potent inhibitor of etoposide-induced Ac-DEVD-pNA cleavage than of cleavage induced by brefeldin A or UV irradiation.
Because caspase-3 is cleaved and activated by caspase-9 in the mitochondrial-dependent pathway, the ability of PKC␦KD to inhibit cleavage of Ac-LEHD-pNA, a caspase-9 substrate, was also assayed. As seen in Fig. 6, etoposide (panel A) all induced Ac-LEHD-pNA cleavage, which can be inhibited in a dose-dependent manner by expression of PKC␦KD. In contrast, expression of AdLacZ had no effect on Ac-LEHD-pNA cleavage. Inhibition of Ac-LEHD-pNA cleavage by PKC␦KD was about 80% for brefeldin A and UV about 45% for etoposide and about 65% for taxol at an MOI of 200, reflecting the relative potencies observed previously for inhibition of sub-G 1 DNA and caspase-3 activity.
PKC␦ Is Required for Loss of Mitochondrial Membrane Potential in Response to Agents That Induce Apoptosis-Loss of the transmembrane potential of mitochondria (MMP) has been shown to occur prior to nuclear condensation and caspase activation and is linked to cytochrome c release in many but not all apoptotic cells (39 -41). To determine whether loss of MMP in response to cell toxins requires PKC␦, we assayed MMP during apoptosis in parotid C5 cells infected with PKC␦KD or the control vector, AdLacZ. Intracellular fluorescence was assayed by FACS after loading cells with an intramitochondrial dye. High fluorescence at 575 nm corresponds to the aggregated form of the dye and is proportional to an intact MMP, whereas loss of MMP leads to a loss of 575 fluorescence and an increase in fluorescence at 525 nm (monomeric form of the dye). indicating a loss of MMP. However, in cells infected with PKC␦KD, loss of MMP in response to etoposide, brefeldin A, or UV is significantly inhibited. This is quantified in Fig. 8 in which the MMP is expressed as the ratio of 575 (aggregate) to 525 (monomer) fluorescence. As seen here, loss of MMP in response to each of these agents is inhibited in a dose-dependent manner by expression of PKC␦KD. With the exception of etoposide, MMP is nearly restored to that seen in cells infected with PKC␦KD alone. Expression of AdLacZ did not inhibit loss of MMP (data not shown).
The data presented above show that PKC␦ activity is required for the loss of MMP in apoptotic cells. In most cases, loss of MMP had been shown to occur prior to, and independent of, caspase activation (42,43); however, recently it has been reported that caspase activity may be required for changes in MMP (42). To determine whether loss of MMP in response to cell toxins requires activated caspase, parotid C5 cells were pretreated with the broad spectrum caspase inhibitor, ZVAD-.FMK, prior to induction of apoptosis. As seen in Table II, pretreatment with ZVAD.FMK did not block the loss of MMP seen in cells treated with etoposide, brefeldin A, UV, or taxol, indicating that caspase activation is not required for loss of MMP. ZVAD.FMK however was a potent inhibitor of caspase-3 activity in parotid C5 cells. Taken together, these data indicate that loss of MMP occurs independent of activation of caspase-3. Moreover, whereas PKC␦ activity is required for loss of MMP, activation of PKC␦ by caspase-directed cleavage per se is not required.

Expression of PKC␦WT Is a Potent Inducer of Apoptosis in
Parotid C5 Cells-The studies shown above demonstrate an essential role for PKC␦ in apoptosis induced by cell toxins and DNA damage. To determine whether overexpression of PKC␦WT alone is sufficient to induce apoptosis, parotid C5 cells were infected with an adenoviral vector, which drives expression of PKC␦WT. As seen in Fig. 9A, expression of PKC␦WT resulted in a robust induction of apoptosis as indicated by DNA ladder formation. Similar results were seen when DNA fragmentation was measured by assaying sub-G 1 DNA accumulation (data not shown). DNA fragmentation was accompanied by a dose-dependent activation of caspase-3 as seen in Fig. 9B. Induction of apoptosis by PKC␦WT did not require the addition of a PKC activator such as phorbol ester, indicating that at least some of the overexpressed PKC␦ is activated intracellularly. To verify that adenovirus-expressed PKC␦KD blocks apoptosis via inhibition of PKC␦WT, parotid C5 cells were infected with PKC␦WT or co-infected with PKC␦WT together with PKC␦KD. As seen in Fig. 9C, co-infection with PKC␦KD results in inhibition of the caspase-3 activation induced by PKC␦WT. DNA fragmentation induced by infection with PKC␦WT was also inhibited by co-infection of PKC␦KD (data not shown).
PKC␦ is cleaved by caspase-3 to generate a 40-kDa carboxylterminal fragment in most but not all cells induced to undergo apoptosis by genotoxins and other cell toxins (11,24,25). Our studies indicate that treatment of parotid C5 cells with etoposide, UV, brefeldin A, and taxol all result in the caspase-dependent cleavage of endogenous PKC␦ to generate a 40-kDa fragment (data not shown). To determine whether induction of apoptosis by PKC␦ correlates with generation of the 40-kDa catalytic fragment, parotid C5 cells were infected with PKC␦WT or PKC␦KD, and PKC␦ expression was analyzed by immunoblot. As seen in Fig. 9D, induction of apoptosis by PKC␦WT is accompanied by cleavage of PKC␦ to generate a 40-kDa fragment. In contrast, the 40-kDa PKC␦ fragment is not seen in etoposide-treated cells, which express PKC␦KD. Because activation of caspase-3 is required for PKC␦ cleavage (23), we conclude that PKC␦ regulates an event in the apoptotic pathway upstream of caspase-3.

DISCUSSION
Many chemicals and drugs induce apoptosis through a mitochondrial-dependent mechanism. However, how these diverse signals are integrated is not known. The data presented here indicate that PKC␦ is a required intermediate in the apoptotic pathway induced by agents that cause DNA damage (etoposide; UV) as well as toxins that target the Golgi (brefeldin A) and microtubule network (taxol). Although all agents required PKC␦ for a maximal apoptotic response, the dependence on PKC␦ for apoptosis induced by UV and brefeldin A was slightly greater than for taxol and etoposide for all apoptotic parameters assayed. This suggests that whereas PKC␦ functions as a common integrator of these apoptotic signals, alternative pathways also exist to integrate apoptotic signals initiated by specific types of cell damage. In this regard, etoposide has been reported to induce two distinct pathways of apoptosis: a caspasedependent pathway at low doses (10 M) and a caspase-independent pathway at higher doses (Ͼ25 M) (42). The less efficient inhibition of etoposide-induced apoptosis by PKC␦KD may indicate that PKC␦ is required for one but not both of these pathways. For all agents except etoposide, upstream events in the apoptotic pathway, such as loss of MMP and activation of caspase-9-like protease, were inhibited to a greater extent than downstream events such as caspase-3 activation and DNA fragmentation. This suggests that PKC␦ is required for regulation of a key early component(s) in the apoptotic pathway and functions at, or prior to, the mitochondria.
Previous studies showing a requirement for PKC␦ for apoptosis have relied heavily, but not exclusively, on the use of rottlerin to inhibit PKC␦. A recent publication by Davies et al. (44) however indicates that rotterlin does not inhibit PKC␦ in vitro, bringing into question reports by Gschwendt et al. (45) and Keenan et al. (46), which demonstrate specific inhibition of PKC␦ by rottlerin. Thus alternative inhibitory strategies are required to verify the role of PKC␦ in apoptosis. Our current studies, which use a kinase-dead mutant of PKC␦ to inhibit endogenous PKC␦ activity, provide independent verification of the importance of PKC␦ as a regulator of apoptosis. This approach has been used extensively to inhibit the activity of FIG. 7. PKC␦ is required for mitochondrial membrane depolarization in response to apoptotic agents. Parotid C5 cells were untreated or infected with PKC␦KD at the indicated MOI for 24 h prior to the addition of the apoptotic agent for an additional 18 h. MMP was assayed by measuring fluorescence at 575 and 525 nm as described under "Experimental Procedures." Healthy cells have high 575 fluorescence, indicating a normal MMP. In apoptotic cells however the dye is excluded from the mitochondria leading to loss of 575 fluorescence (aggregate) and an increase in 525 fluorescence (monomer). Data were converted to density plots using System 2 software. protein kinases; however, pitfalls include a potential lack of specificity, as well as the possibility that the subcellular localization of the kinase may be altered because of the mutation and/or overexpression. To address the issue of specificity, we have shown that inhibition is specific for PKC␦, as expression of PKC␣KD does not inhibit etoposide-induced apoptosis. Likewise, preliminary experiments using PKCKD also show no inhibition of apoptosis. 3 Furthermore, as expected, expression of PKC␦WT has the opposite effect of PKC␦KD, i.e. it is an inducer of apoptosis. Expression of PKC␦KD can block apoptosis induced by overexpression of PKC␦WT, indicating that PKC␦WT is a target of PKC␦KD. Inhibition of apoptosis with PKC␦KD increased in a dose dependent manner when the kinase-dead mutant was expressed in the range of 5-30-fold over endogenous PKC␦ (MOI of 25-250). Whereas it is not clear why such high ratios of kinase-dead to endogenous PKC␦ are required, one possibility, not addressed in the current manuscript, is that some/much of the overexpressed protein does not localize to the correct subcellular localization, and hence is not available to suppress the biological effect being assayed.
PKC␦ is proteolytically activated by caspase-3 in response to many agents that induce mitochondrial-dependent apoptosis, including the agents utilized in these studies. Whereas caspase cleavage and activation of PKC␦ clearly contributes to apoptosis in many cells, our data are also consistent with a role for PKC␦ in apoptosis prior to caspase cleavage and activation.
First, PKC␦ is required for mitochondrial membrane depolarization, a process that we show here to be caspase-independent in parotid C5 cells. Second, expression of PKC␦KD inhibits activation of both initiator (caspase-9) and effector (caspase-3) caspases. Third, whereas induction of apoptosis by PKC␦WT is accompanied by cleavage of PKC␦ in cells which express PKC␦KD no cleavage of PKC␦ is seen. This suggests that the kinase function of PKC␦ is required for activation of a pathway, which subsequently mediates caspase activation. However, expression of the caspase-cleavage fragment of PKC␦ is sufficient to induce apoptosis in many cells, including parotid C5 cells, 2 suggesting that both full-length and caspase-cleaved PKC␦ contribute to the apoptotic program. One explanation is that the function of the activated full-length and caspase-cleaved forms of PKC␦ are the same, with caspase cleavage simply being a more efficient method of activating the kinase. In this paradigm, caspase cleavage would serve to amplify the PKC␦ signal in apoptotic cells. In support of this is the demonstration from our laboratory and Fujii et al. that caspase cleavage of PKC␦ is not required for phorbol ester-induced apoptosis, and that phorbol ester is a much weaker inducer of apoptosis than agents that induce PKC␦ cleavage (9,11). Alternatively, it is tempting to speculate that either changes in conformation or subcellular localization may result in the regulation of targets specific for the caspase-cleaved form of PKC␦. Thus the functions of full-length and caspase-cleaved PKC␦ may be distinct, with full-length PKC␦ regulating an early event(s) and the caspase-cleaved form of PKC␦ regulating a later event(s) in the apoptotic pathway. Consistent with this possibility is the observation that PKC␦ translocates to specific sites following induction of apoptosis (16,18,47).
It should be noted that expression of PKC␦KD did not totally inhibit apoptosis in response to any agent tested. Because PKC␦KD is thought to function as a competitive inhibitor of endogenous PKC␦, it is possible that we cannot achieve the level of expression of the inhibitory protein required to completely suppress the function of the endogenous kinase (28). In addition it is likely that other regulatory signals contribute to the induction and/or amplification of apoptosis by these agents. In particular, members of the mitogen-activated protein kinase/extracellular-regulated kinase family have been shown to regulate apoptosis (6 -8). In addition, the caspase cleavage and constitutive activation of other protein kinases such as MEKK-1, MST1, PAK-2, FAK have also been shown to contribute to apoptosis as does caspase cleavage and inactivation of enzymes required for DNA repair (48,49).
Our studies suggest that PKC␦ functions at, or prior to, the mitochondria in apoptotic cells, and that overexpression of 3 R. Bell and M. E. Reyland, unpublished data.
FIG. 8. Inhibition of mitochondrial membrane depolarization by PKC-␦KD is dose dependent. Parotid C5 cells were untreated or infected with PKC␦KD at the indicated MOI for 24 h prior to the addition of the apoptotic agent. MMP is expressed as the ratio of 575 (aggregate) to 525 (monomer) fluorescence. The striped bars indicate the aggregate/monomer ratio in untreated cells or in cells treated with the apoptotic agent alone. Cells infected with PKC␦KD but not treated with an apoptotic agent are shown in black, whereas the gray bars show cells infected with PKC␦KD and then induced to undergo apoptosis with the indicated agent. A high ratio indicates healthy cells; a decrease in the ratio indicates apoptosis. This experiment has been repeated three times with similar results. PKC␦WT is sufficient to induce apoptosis. This is consistent with a recent report from Li et al. (18) in which phorbol ester treatment of keratinocytes which overexpress PKC␦WT induced apoptosis. In this study apoptosis correlated with translocation of PKC␦WT, but not PKC␦KD to the mitochondria, suggesting that mitochondrial localization of PKC␦ facilitates apoptosis (16,18). This conclusion is supported by data from Majumder et al. (16), which demonstrates that in phorbol estertreated cells, translocation of PKC␦ to the mitochondria facilitates cytochrome c release. Whereas our studies are consistent with the hypothesis that PKC␦ regulates mitochondrial events in apoptosis, they do not address whether this is a direct effect at the level of the mitochondria or an indirect effect perhaps via regulation of another component of the apoptotic pathway.
Notably, the nuclear proteins DNA-PK (50) and lamin B (51) are reported to be substrates for PKC␦ in apoptotic cells and Scheel-Toellner et al. (47) show translocation of PKC␦ to the nucleus in T cells treated with IFN-␤ to induce apoptosis (47). Future studies directed at identifying substrates of PKC␦ in apoptotic cells will greatly enhance our understanding of the function of this essential signaling pathway.