The Protein Kinase Cδ Catalytic Fragment Targets Mcl-1 for Degradation to Trigger Apoptosis*

Proteolytic cleavage and subsequent activation of protein kinase C (PKC) δ is required for apoptosis induced by a variety of genotoxic agent, including UV radiation. In addition, overexpression of the constitutively active PKCδ catalytic fragment (PKCδ-cat) is sufficient to trigger Bax activation, cytochrome c release, and apoptosis. While PKCδ is a key apoptotic effector, the downstream target(s) responsible for the mitochondrial apoptotic cascade are not known. We found that expression of the active PKCδ-cat in HaCaT cells triggers a reduction in the anti-apoptotic protein Mcl-1, similar to UV radiation. The down-regulation of Mcl-1 induced by PKCδ-cat was not at the mRNA level but was due to decreased protein half-life. Overexpression of Mcl-1 protected HaCaT cells from both UV and PKCδ-cat-induced apoptosis and blocked the release of cytochrome c from the mitochondria, indicating that Mcl-1 down-regulation was required for apoptosis signaling. Indeed, down-regulation of Mcl-1 with siRNA slightly increased the basal apoptotic rate of HaCaT cells and dramatically sensitized them to UV or PKCδ-cat-induced apoptosis. HaCaT cells with down-regulated Mcl-1 had higher activated Bax protein, as measured by Bax cross-linking, indicating that Mcl-1 down-regulation is sufficient for Bax activation. Finally, recombinant PKCδ could phosphorylate Mcl-1 in vitro, identifying Mcl-1 as a direct target for PKCδ. Overall our results identify Mcl-1 as an important target for PKCδ-cat that can mediate its pro-apoptotic effects on mitochondria to amplify the apoptotic signaling induced by a wide range of apoptotic stimuli.

The induction of apoptosis in keratinocytes by UV radiation is an important protective mechanism from sunlight-induced skin tumors (1). UV induces keratinocyte apoptosis primarily via the intrinsic or mitochondrial death effector pathway (2,3). The UV mitochondrial apoptotic signaling pathway requires the early loss of Mcl-1, which allows the release of cytochrome c and other pro-apoptotic factors from the mitochondria. Release of these pro-apoptotic mediators triggers the activation of caspase-9 and ultimately the activation of effector caspases such as caspase-3 (2,4). While caspase-3 cleaves a wide range of protein substrates, the widely expressed death substrate PKC␦ 2 is critical as inhibition of PKC␦ cleavage and/or activation protects from progressive caspase activation and apoptosis induced by a variety of stimuli, indicating positive feedback regulation between PKC␦ cleavage and caspase activation (5)(6)(7)(8)(9)(10). Furthermore, ectopic expression of the constitutively active PKC␦-cat is sufficient to induce apoptosis in several cell types (8,(11)(12)(13). PKC␦-cat has been localized to the mitochondria and the nucleus (6,8) and induces apoptosis involving Bax activation, cytochrome c release, and caspase activation (8,13). Although several PKC␦ targets have been identified, including p73␤ (14), DNA-PK (12), Rad9 (15), phospholipid scramblases 1 and 3 (16 -18), p38␦-ERK1/2 (19), and nuclear lamin B (20), it is unclear how these substrates can directly initiate the mitochondrial apoptotic cascade.
During UV-induced apoptosis, the loss of Mcl-1 is required for the mitochondrial translocation of Bax, release of cytochrome c, and caspase activation (4). The anti-apoptotic activity of Mcl-1 is due to its ability to bind and sequester pro-apoptotic Bcl-2 family proteins such as Bak and Bax (21)(22)(23). Upon Mcl-1 loss, Bak and/or Bax are free to oligomerize and mediate pore formation in the mitochondria, allowing the release of cytochrome c (24). In the cell types examined to date, Mcl-1 has a short half-life, and its levels are rapidly modulated during apoptosis at the transcriptional and translational level (25). In normal human keratinocytes, the UV-induced loss of Mcl-1 is at the mRNA level, and no increase in Mcl-1 protein turnover was noted in HeLa cells exposed to UVC (4,26).
To identify PKC␦ targets that can initiate the mitochondrial apoptotic pathway, we have focused on the Bcl-2 family since many of these proteins are regulated by phosphorylation. We previously found that PKC␦-cat triggers the up-regulation and activation of Bax (13). In this study, we find that expression of the active PKC␦-cat causes increased turnover and reduced levels of Mcl-1, and we have identified Mcl-1 as a direct target phosphorylated by PKC␦-cat. These findings can explain the ability of PKC␦-cat to trigger mitochondrial apoptosis and identify Mcl-1 as a key link in the positive feedback loop amplifying the apoptotic cascade of a wide variety of genotoxic agents. * This work was supported by a grant from the Potts Foundation and National

EXPERIMENTAL PROCEDURES
Cell Culture and Retrovirus Infections-The immortalized human keratinocyte cell line HaCaT (27), kindly provided by Dr. Norbert Fusenig (German Cancer Research Center, Heidelberg, Germany), was cultured in Media 154 (Cascade Biologics, Inc., Portland, OR) with 0.07 mM calcium added. Cells were irradiated with 30 mJ/cm 2 UV from a bank of four UVB bulbs (FS36T12/UVB-VHO) with the dish lids removed. Cells were infected with retroviruses by spinning at 300 ϫ g for 1 h at 32°C as described previously (2,8). For siRNA experiments, cells were infected with pSUPER.retro.puro viruses and selected with 1 g/ml puromycin for at least 3 days until plated for the experiment.
Retrovirus Production-The catalytic fragment of PKC␦ was routinely expressed from an LZRS-based retroviral vector as an estrogen receptor ligand binding domain fusion protein (PKC␦-ER) that is activated by treating the cells with 100 nM 4-hydroxytamoxifen (Tam) (13). The PKC␦-ER fusion protein has some basal PKC␦ catalytic fragment activity, and Tam treatment is used to activate it further. In some experiments, the catalytic fragment of PKC␦ (PKC␦-cat) in LZRS was used instead of the PKC␦-ER fusion protein (8). The PKC␦-ER virus permitted higher expression because the virus can be produced in the absence of Tam and packaging cell apoptosis is minimized (13). FLAG-tagged Mcl-1 was kindly provided by Dr. W. Douglas Cress (H. Lee Moffitt Cancer Center, Tampa, FL) (28) and was cloned into the Hind III/XhoI sites of the retroviral vector LZRS-Linker (8,29). Construction of the Bcl-2 retrovirus was described previously (13). As a negative control, the empty retroviral vector LZRS-Linker (Linker) was used. All retroviruses were prepared in Phoenix-Ampho packaging cells (ATCC with permission from Garry P. Nolan, Stanford University Medi-cal Center) by calcium phosphate transfection and selection with 1 g/ml puromycin (29).
Retroviruses encoding hairpin siRNA were constructed using the pSUPER.retro.puro vector system (OligoEngine, Seattle, WA). A double-stranded hairpin oligonucleotide designed to target the human Mcl-1 cDNA nucleotides 13-33 (AAGAAACGCGGTAAT-CGGA) (30) was cloned into the BglII/HindIII sites of pSUPER.retro. puro. A control siRNA virus was constructed, which encoded a 19-bp target sequence (GCGCGCTTTGTAGG-ATTCG) with no significant homology to any gene in the human genome. Correct insertion of the oligonucleotides was confirmed by DNA sequencing. pSUPER viruses were produced in Phoenix-Ampho packaging cells as described for LZRS viruses.
Apoptosis Assays-Apoptosis was assayed by flow cytometry using Annexin V-FITC kit (Beckman Coulter, Inc.) as recommended by the manufacturer. The cells were run on a Coulter Epics XL-MCL flow cytometer to determine the percentage Annexin V-positive cells.
Cell Fractionation and Immunoblotting-Whole cell lysates were prepared by lysing floating and attached cells in 20 mM Tris-HCl, pH 7.5, 1% CHAPS, 5 mM EDTA, and 1ϫ Complete protease inhibitor mixture (Roche Applied Science). Cytoplasmic extracts for cytochrome c release were prepared by trypsinizing the cells, washing once with phosphate-buffered saline, and suspending them in isotonic sucrose buffer: 250 mM sucrose, 10 mM HEPES, pH 7.4, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EGTA, 1 mM dithiothreitol, 1ϫ Complete protease inhibitor mixture. Digitonin was added to 0.05%, and the cells mixed gently for 2 min at room temperature. The permeabolized cells were pelleted by spinning at 15,000 ϫ g for 10 min at 4°C and the cytoplasmic extracts analyzed by SDS-PAGE. Proteins were transferred to nitrocellulose, stained with antibodies, and visualized with ECL (Amersham Biosciences) or on an Odyssey Infrared Imaging System (LI-COR Biosciences, Inc., Lincoln, NE). Antibodies used for immunoblotting were Mcl-1 (sc-819, Santa Cruz Biotechnology, Santa Cruz, CA) at 1:500, ␤-actin  fected Phoenix-Ampho cells using anti-FLAG-agarose and incubated in the presence of [␥-32 P]ATP and for 10 min at 30°C. To generate the active PKC␦-cat, recombinant PKC␦ was incubated with recombinant active caspase-3 (Alexis Biochemicals, San Diego, CA) as described previously (8). After the kinase reaction, the agarose beads were washed with immunoprecipitation buffer and boiled in SDS sample buffer. Phosphorylated proteins were run on SDS-PAGE and transferred to nitrocellulose for detection by autoradiography and Western blotting.
Protein Half-life Determination-The half-life of Mcl-1 was determined by treating cells with 10 g/ml cyclohexamide for 0 -6 h to inhibit protein synthesis and then preparing cell lysates to determine Mcl-1 levels by Western blotting. Mcl-1 protein levels were quantified by densitometry scanning and analysis using Scion Image. Mcl-1 band intensities for each treatment condition were normalized so that the percent of Mcl-1 in the no cyclohexamide group equaled 100%, and semi-log plots were generated in Microsoft Excel. Linear regression was performed and the half-life calculated from the fitted line equation.
Ribonuclease Protection Assays-RNA was isolated from HaCaT cells using TRIzol reagent (Invitrogen). The RNA was used in the Ribo-Quant multi-probe RNase protection assays system (Pharmingen) with the hAPO-2c probe set as described in the manufacturer's instructions.

Expression of PKC␦-cat Reduces Steady-state Mcl-1 Protein Levels-
Mcl-1 is an anti-apoptotic Bcl-2 family protein that can bind to and sequester pro-apoptotic proteins such as Bax and Bak (21)(22)(23). Since apoptosis induced by both PKC␦cat and UV involves the activation of Bax (13), we examined Mcl-1 protein levels in HaCaT cells induced to undergo apoptosis by either PKC␦cat expression or UV irradiation. To induce the activity of PKC␦-cat, we infected HaCaT cells with a retrovirus encoding a PKC␦-cat/estrogen receptor ligand binding domain fusion protein (PKC␦-ER). This PKC␦-ER fusion protein has some basal PKC␦ enzymatic activity that can be activated further by treating the cells with Tam (13). As a negative control, cells were infected with an empty virus (Linker) and also treated with Tam. The Western blot in Fig. 1A shows that total Mcl-1 levels were reduced to 57% of control in cells infected with the PKC␦-ER virus, and Mcl-1 was reduced further (41% of control) when the PKC␦-ER was activated with Tam. Consistent with previous reports, UV caused a dramatic reduction in Mcl-1 levels, which was almost undetectable 18 h after UV irradiation.  The asterisks indicate significant differences between the treated groups and the Linker ϩ Tam group ( p Ͻ 0.05). C, the half-life of Mcl-1 protein in HeLa cells was determined as described for A. Cells were treated with CHX alone, exposed to UV, or exposed to UV and treated with CHX and the half-life of Mcl-1 calculated from Mcl-1 Western blots. Note that in HeLa cells Mcl-1 has a short half-life (ϳ40 min) that is not affected by UV.
(Linker virus plus Tam), and activation of PKC␦-ER with Tam reduced the half-life to 2 h 30 min. UV also reduced the half-life of Mcl-1 protein to 2 h, but the kinase inactive PKC␦(K378A)-ER mutant did not significantly affect Mcl-1 turnover. As a control, we also measured the half-life of Mcl-1 in HeLa cells which are reported to have rapidly turned over Mcl-1 (half-life of ϳ40 min) that is not influenced by UV radiation (4). Fig. 2C confirms that the half-life of Mcl-1 in HeLa cells is ϳ40 min and is not decreased further by UV, indicating that cell type differences exist for the regulation and turnover of Mcl-1.
We also examined Mcl-1 mRNA levels to determine whether PKC␦-cat or UV caused a decrease which could contribute to the reduced Mcl-1 protein levels. The ribonuclease protection assay in Fig. 3 shows that PKC␦-ER plus Tam did not cause an appreciable decrease in Mcl-1 mRNA levels. The mRNA levels of other Bcl-2 family members (Bcl-w, Bcl-x L , Bid, Bik, Bak, and Bax) were also not affected. UV caused a generalized decrease in mRNA levels for all Bcl-2 family members examined. Together, these results indicate that PKC␦-cat triggers loss of Mcl-1 by destabilizing the protein.
Ectopic  HaCaT cells were infected with Linker, PKC␦-ER, or PKC␦(K378A)-ER viruses and treated with Tam to activate PKC␦-ER or exposed to UV (30 mJ/cm 2 ) as indicated. RNA was isolated 3 days after viral infection or 18 h after UV exposure and an RNase protection assay performed. Hybridization to L32 and glyceraldehyde-3phosphate dehydrogenase (GAPDH) was used as control to ensure equal levels of input RNA. Note that Mcl-1 mRNA levels were not decreased by PKC␦cat (PKC␦-ERϩTam). apoptosis, confirming that kinase activity is required for PKC␦cat to induce apoptosis.
The release of cytochrome c from the mitochondria is a key event in the activation of caspase-9 and is under negative regulation by anti-apoptotic Bcl-2 family proteins such as Mcl-1 (4). To determine whether Mcl-1 expression can inhibit the release of cytochrome c caused by PKC␦-cat (13), we assayed cytoplasmic cytochrome c levels by Western blot. As shown in Fig. 5, PKC␦-ER plus Tam triggered the release of cytochrome c into the cytoplasm, and this could be almost completely blocked by Mcl-1 expression. Mcl-1 expression also blocked UV-induced release of cytochrome c from the mitochondria. Note that the caspase inhibitor Z-VAD did not inhibit cytochrome c release triggered by PKC␦-cat, indicating that cytochrome c release is an early, caspase-independent event. Together these results suggest that the loss of Mcl-1 triggered by PKC␦-cat may facilitate the release of cytochrome c and subsequent apoptotic signaling.
Suppression of Mcl-1 with siRNA Sensitizes HaCaT Cells to Apoptosis-To determine whether the reduction in Mcl-1 caused by PKC␦-cat is sufficient to trigger apoptosis, we infected HaCaT cells with a retrovirus encoding a siRNA targeting Mcl-1 and assessed apoptosis. Fig. 6A shows that the Mcl-1 siRNA reduced Mcl-1 protein levels to 42-67% of control, with an average reduction to 55% of control (n ϭ 3 experiments). Note that this level of inhibition is similar to the reduction in Mcl-1 triggered by PKC␦-ER (41% of control, Fig. 1). In Fig. 6B, quantitation of apoptosis by Annexin V binding shows that Mcl-1 siRNA alone cause a small but statistically significant ( p Ͻ 0.05) increase in apoptosis over a 2-day time period. However in HaCaT cells expressing the catalytic fragment of PKC␦ or exposed to UV radiation, the Mcl-1 siRNA caused a substantial increase in apoptosis that was highly significant ( p Ͻ 0.001).
Mcl-1 can bind and sequester the pro-apoptotic protein Bax (21), and the induction of apoptosis by the catalytic fragment of PKC␦ or UV involves the up-regulation and activation/oligomerization of Bax (13). We determined whether partial down-regulation of Mcl-1 with siRNA was sufficient to activate Bax. The Bax cross-linking experiment in Fig. 6C shows that the Mcl-1 siRNA induced the appearance of Bax dimers (2ϫ) and increased the total amount of Bax 2.4-fold. Activation of the PKC␦-cat and UV irradiation also increased the Bax dimer and total Bax levels, and the Mcl-1 siRNA further enhanced the oligomerization and up-regulation of Bax. Taken together, these results indicate that even a partial reduction in Mcl-1 levels can dramatically sensitize cells to apoptotic stimuli. Tam to activate PKC␦-ER, or exposed to UV as indicated. The general caspase inhibitor Z-VAD was added at 10 g/ml as indicated. Cytosolic fractions were analyzed by Western blotting for release of cytochrome c from the mitochondria. Note that PKC␦-cat and UV triggered cytochrome c release that was inhibited by Mcl-1. Z-VAD did not prevent cytochrome c release. A Western blot for ␤-actin was done as a control of protein loading. siRNA viruses were selected with puromycin and then re-infected with PKC␦cat virus or exposed to UV as indicated. Three days after the PKC␦-cat infection or 18 h after UV irradiation, apoptosis was assayed by Annexin V/propidium iodide staining. Each condition in the experiment was performed in triplicate, with error bars denoting standard deviation. C, HaCaT cells infected with control or Mcl-1 siRNA viruses were selected with puromycin and re-infected with PKC␦-ER virus as indicated. PKC␦ activity was induced by addition of Tam to the media of PKC␦-ER-infected cells. Three days after PKC␦-ER infection or 18 h after UV irradiation, the heavy membrane fraction was isolated and a Western blot for Bax performed. Total Bax levels were quantified by densitometry and are indicated at the bottom of the Western blot. Note the increase in total Bax levels, especially the dimer (2ϫ) and trimer (3ϫ) forms, indicative of Bax activation.

PKC␦ Directly Phosphorylates
Mcl-1-Since our results demonstrated that the catalytic fragment of PKC␦ causes enhanced turnover of Mcl-1, and phosphorylation of Mcl-1 by a variety of stimuli has been reported to regulate its stability and activity (32)(33)(34), we determined whether PKC␦ can directly phosphorylate Mcl-1. Mcl-1, or Bcl-2 as a control, was immunoprecipitated from Phoenix-Ampho packaging cells, incubated with recombinant PKC␦ in a kinase assay, and the phosphorylated proteins analyzed by SDS-PAGE and autoradiography. Fig. 7A shows that PKC␦ phosphorylated a protein at ϳ45 kDa in the FLAG immunoprecipitate from Mcl-1-transfected cell lysate, and this band corresponds in size to the Mcl-1 in Fig. 7B. No phosphorylation of Bcl-2 (ϳ30 kDa) or Mcl-1 in the absence of exogenous PKC␦ was detected, indicating that the phosphorylation of Mcl-1 was specific and due to PKC␦. Similar results were obtained with PKC␦-cat generated by caspase-3 cleavage (data not shown). A nonspecific band at ϳ85 kDa reacting with the Mcl-1 antibody was not phosphorylated by PKC␦. These results indicated that Mcl-1 is a direct substrate for PKC␦-cat.

DISCUSSION
PKC␦ is a widely expressed death substrate activated by caspase-3 cleavage in cells triggered to undergo apoptosis by a variety of agents. In this study, we have identified Mcl-1, an anti-apoptotic Bcl-2 family member, as an important target for apoptosis mediated by PKC␦-cat. Ectopic expression of PKC␦cat caused the down-regulation of Mcl-1 by increasing Mcl-1 protein turnover (Fig. 2), and the loss of Mcl-1 was required for PKC␦-cat-induced cytochrome c release (Fig. 5) and apoptosis (Fig. 4). The loss of Mcl-1 is also involved in UV-induced apoptosis (4), and we previously demonstrated that the cleavage and activation of PKC␦ were also partially required for the down-regulation of Mcl-1 by UV radiation (10). This study extends these observations by demonstrating that PKC␦-cat is sufficient to cause Mcl-1 down-regulation and provides mechanistic insight into the down-regulation of Mcl-1 by PKC␦-cat.
The expression of Mcl-1 is under multiple types of regulation at both the transcriptional and translation levels and responds rapidly to environmental clues and apoptotic stimuli (25,35,36). Expression of PKC␦-cat caused ϳ50% decrease in steady-state Mcl-1 levels, while UV caused almost complete loss of Mcl-1 (Fig.  1), consistent with the much stronger apoptotic response of HaCaT cells to UV radiation than PKC␦cat. Both PKC␦-cat and UV decreased the half-life of Mcl-1 protein (Fig. 2), and UV, but not PKC␦cat, reduced Mcl-1 mRNA levels. In contrast, UV has been reported to trigger the loss of Mcl-1 in HeLa cells by inhibiting its synthesis and had no effect of Mcl-1 protein turnover (4). To validate our Mcl-1 halflife measurement method, we confirmed that UV did not increase Mcl-1 turnover in HeLa cells (Fig.  2C), indicating cell type-specific regulation for Mcl-1 down-regulation. Understanding the apoptotic regulation of Mcl-1 in HaCaT cells, a human keratinocyte cell line, is highly relevant to UV skin carcinogenesis as these non-tumorigenic cells harbors UV signature mutations in p53 and resemble the premalignant cells frequently targeted by UV radiation (27,37). In normal human keratinocytes, UV caused a reduction in Mcl-1 mRNA and protein levels, consistent with our results in HaCaT cells shown in Figs. 1 and 3 (26).
The down-regulation of an anti-apoptotic protein such as Mcl-1 is not necessarily sufficient to induce apoptosis as most cells, including HaCaT cells, express multiple pro-survival Bcl-2 family members (4,39,40). Mcl-1 knock-out mice were embryonic lethal, and Mcl-1 has a specific survival role for hematopoietic stem cells (41,42). Using a retrovirus expressing a hairpin siRNA against Mcl-1, we were able to down-regulate Mcl-1 levels ϳ50%, a level comparable with that triggered by PKC␦-cat ( Figs. 1 and 6). We found a small but statistically significant increase in basal apoptosis and total Bax levels ( Fig.  6). However, the down-regulation of Mcl-1 with siRNA dramatically sensitized HaCaT cells to PKC␦-cat and UV-induced Bax activation and apoptosis (Fig. 6). This synergy may be due to either more complete Mcl-1 loss by the combination of PKC␦-cat and Mcl-1 siRNA or the involvement of additional apoptosis inducing signals initiated by PKC␦-cat. Since ϳ50% reduction of Mcl-1 by siRNA did not trigger apoptosis to the same extent as the active PKC␦-ER virus, which also reduced Mcl-1 levels ϳ50%, our results suggest that additional signals initiated by PKC␦-cat cooperate with Mcl-1 loss to trigger apoptosis. In addition, more complete reduction of Mcl-1 by siRNA in other cell types was also not sufficient to induce high levels of apoptosis or cell death (4,22).
We also demonstrated that Mcl-1 is directly phosphorylated by PKC␦-cat in vitro (Fig. 7). Since kinase activity was required for increased turnover of Mcl-1 (Fig. 2) and apoptosis (Fig. 4), phosphorylation is a likely mechanism for targeting Mcl-1 for degradation. Furthermore, both PKC␦-cat and Mcl-1 are localized to the mitochondria, putting the active kinase and its substrate at the same subcellular localization (4,8). Mcl-1 can be phosphorylated by several protein kinases. 12-O-Tetradecanoylphorbol-13-acetate treatment induced phosphorylation at a conserved extracellular signal-regulated kinase (ERK) site in the PEST region of Mcl-1, and G 2 /M arresting drugs or a protein phosphatase 1/2A inhibitor also induced an ERK-independent phosphorylation of Mcl-1 (32,34,43). Mcl-1 was phosphorylated at Ser-121 and Thr-163 by c-Jun NH 2 -terminal kinase in response to H 2 O 2 , and this phosphorylation was associated with inactivation of Mcl-1 and enhanced apoptosis (33).
Mcl-1 was also phosphorylated by glycogen synthase kinase-3 at Ser-159 in response to interleukin-3 withdrawal or AKT inhibition leading to increased ubiquitinylation and degradation (44). Mcl-1 can be degraded by the proteasome and is also subject to cleavage by caspases and granzyme B (4,(43)(44)(45). The BH3 domain-containing Mule/ARF-BP1 E3 ubiquitin ligase is necessary and sufficient for Mcl-1 polyubiquitination and proteasome targeting (46). We found that the proteasome inhibitor MG132 significantly increased Mcl-1 protein levels (data not shown), suggesting a role for the ubiquitin proteasome pathway in Mcl-1 degradation in HaCaT cells.
During apoptosis induced by UV and other stimuli, PKC␦ cleavage and activation are caspase-dependent and occur only after caspase activation, while Mcl-1 loss is an early event required for caspase activation (4,5,8,11,47,48). These kinetics make it impossible for PKC␦-cat to mediate the initial Mcl-1 down-regulation that permits Bax/Bak activation, cytochrome c release, and caspase activation. However, the ability of PKC␦cat to phosphorylate and increase the turnover of Mcl-1 can explain why PKC␦ cleavage and activation are required for a robust apoptotic response, as this would establish a positive feedback loop between PKC␦ cleavage and activation, Mcl-1 loss, and further caspase activation.
Mcl-1 is unique among Bcl-2 family members in that it is responsive to a variety of growth factors or phorbol ester treatment (49) and is required for early pre-implantation and development in the mouse and survival of hematopoietic stem cells (41,42). Mcl-1 also has functions in cell cycle control, indicating it has non-apoptotic functions (50). The multiple regulatory features and cell type-specific regulatory mechanisms of Mcl-1 are consistent with this multifaceted functions and may be required to integrate inputs from diverse signaling pathways, including the PKC␦ apoptotic pathway discussed here. Taken together, our results identify Mcl-1 as a key link in the positive feedback loop targeted by PKC␦-cat in cells destined to undergo apoptosis.