The Protein Kinase Cδ Catalytic Fragment Is Critical for Maintenance of the G2/M DNA Damage Checkpoint*

Protein kinase Cδ (PKCδ) is an essential component of the intrinsic apoptotic program. Following DNA damage, such as exposure to UV radiation, PKCδ is cleaved in a caspase-dependent manner, generating a constitutively active catalytic fragment (PKCδ-cat), which is necessary and sufficient for keratinocyte apoptosis. We found that in addition to inducing apoptosis, expression of PKCδ-cat caused a pronounced G2/M cell cycle arrest in both primary human keratinocytes and immortalized HaCaT cells. Consistent with a G2/M arrest, PKCδ-cat induced phosphorylation of Cdk1 (Tyr15), a critical event in the G2/M checkpoint. Treatment with the ATM/ATR inhibitor caffeine was unable to prevent PKCδ-cat-induced G2/M arrest, suggesting that PKCδ-cat is functioning downstream of ATM/ATR in the G2/M checkpoint. To better understand the role of PKCδ and PKCδ-cat in the cell cycle response to DNA damage, we exposed wild-type and PKCδ null mouse embryonic fibroblasts (MEFs) to UV radiation. Wild-type MEFs underwent a pronounced G2/M arrest, Cdk1 phosphorylation, and induction of apoptosis following UV exposure, whereas PKCδ null MEFs were resistant to these effects. Expression of PKCδ-green fluorescent protein, but not caspase-resistant or kinase-inactive PKCδ, was able to restore G2/M checkpoint integrity in PKCδ null MEFs. The function of PKCδ in the DNA damage-induced G2/M cell cycle checkpoint may be a critical component of its tumor suppressor function.

Mcl-1 to accelerate apoptosis (16,19). This cleavage event is vital to the apoptotic cascade because the overexpression of a mutant PKC␦ that cannot be cleaved by caspase 3 suppresses UV radiation-induced apoptosis in human keratinocytes (KCs) (20). It has also been reported that PKC␦ contains a nuclear localization sequence and that nuclear localization of PKC␦ is of critical importance to successful completion of apoptosis (21,22).
PKC␦ expression is lost in both human squamous cell carcinomas and chemically induced mouse skin tumors, supporting its function as a cutaneous squamous cell carcinoma tumor suppressor gene (23,24). The vital role of PKC␦-mediated apoptosis in tumor suppression is bolstered by work demonstrating that transgenic mice overexpressing PKC␦ are resistant to chemically induced squamous cell carcinomas and have elevated 12-O-tetradecanoylphorbol-13-acetate-induced apoptosis (24,25). Furthermore, re-expression of PKC␦ in human squamous carcinoma cells induces spontaneous apoptosis and inhibits tumorigenesis (23).
In addition to the well studied pro-apoptotic effects of PKC␦, several studies have reported effects of PKC␦ on the cell cycle. For example, PKC␦ stimulates apoptosis by initiating an S phase arrest in rat thyroid cells (26). Other studies have tied PKC␦ to the G 2 /M phases of the cell cycle by demonstrating that PKC␦ can localize to the nucleus, where it is associated with chromatin condensation as well as inhibition of cytokinesis (27,28). Furthermore, research on the effects of PKC␦ overexpression in HCT116 cells revealed that PKC␦ induced many of the morphological features of mitotic catastrophe, including multimicronucleation and centrosomal amplification (29). In this study, we report that PKC␦-cat plays a critically important role in enforcing the G 2 /M checkpoint in response to UV radiation.
Cell Culture and UV Treatment-Primary human KCs were isolated from neonatal foreskins with Loyola Institutional Review Board approval as described previously (31,32). KCs and HaCaT cells were cultured in Medium 154CF with 0.07 mM calcium and human keratinocyte growth supplement (Cascade Biologics). Spontaneously immortalized MEFs from wild-type or PKC␦ null mice were cultured in Dulbecco's modified Eagle's medium (Invitrogen) containing 10% fetal bovine serum and were kindly provided by Dr. Anning Lin (University of Chicago). PKC␦-cat-ER was activated by treatment with 10 M 4-hydroxytamoxifen (Alexis Biochemicals) (18). In all experiments using PKC␦-cat-ER, control (Linker) transduced cells were also treated with 4-hydroxytamoxifen. ATM/ATR inhibition was achieved by the addition of 2 mM caffeine in water to the culture medium (33). UV irradiation was done using a UV Panelite unit with ϳ65% emission in the UVB spectrum as previously described (23).
Antibodies and Western Blotting-Cell lysates were collected by scraping cells in radioimmune precipitation assay buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, 1% sodium deoxycholate). Lysates were briefly sonicated and centrifuged at 14,000 rpm for 5 min to remove cellular debris. Protein concentrations were determined using standard Bradford reagent methodology.
Flow Cytometry-Cells were collected by trypsinization, pelleted, and washed in fluorescence-activated cell sorter (FACS) buffer (phosphate-buffered saline, 5% fetal bovine serum). Cells were then pelleted and resuspended in 100 l of fetal bovine serum. Cells were fixed by addition of ice-cold 100% ethanol for at least 30 min. RNA was digested by treatment with 10 g/ml RNase for 15 min at 37°C. Propidium iodide was added to a final concentration of 50 g/ml, and samples were incubated on ice for at least 1 h. Cell cycle profiles were analyzed using a Beckman Coulter EPICS XL-MCL flow cytometer. Histogram overlays were generated using FlowJo software.
Mitotic Index-For mitotic index measurements in MEFs, the cells were treated with or without UV radiation, followed by 10 ng/ml nocodazole to trap any cells that had overcome UV radiation-induced G 2 /M arrest and entered mitosis. Mitotic index measurements in HaCaT cells were performed 3 days after PKC␦-cat-ER transduction and thus were not treated with nocodazole. Cells were collected and washed in FACS buffer before being fixed in 3.7% formaldehyde for 10 min. After fixation, cells were permeabilized in 70% EtOH at 4°C for 30 min. After washing in FACS buffer, cells were incubated for 2 h in 100 l of phosphorylated histone H3 (Ser 10 ) antibody. Cells were washed and incubated on ice in 100 l of diluted Alexa Fluor 488-conjugated anti-rabbit secondary antibody for 30 min. Following secondary antibody incubation, cells were washed twice and incubated in 500 l of solution containing 10 g/ml RNase and 50 g/ml propidium iodide for 30 min prior to FACS analysis.
Nuclear/Cytoplasmic Fractionation-Cells were collected in phosphate-buffered saline and incubated in hypotonic buffer (10 mM KCl, 50 mM HEPES, pH 7.9, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, Complete protease inhibitor (Roche Applied Science), 40 mM glycerophosphate, 2 mM sodium fluoride, and 1 mM sodium orthovanadate) for 15 min on ice. Triton X-100 was added to a final concentration of 0.5% to complete lysis. Samples were centrifuged, and the resulting supernatant contained enriched cytoplasmic proteins. The pellet (containing nuclei) was washed in hypotonic buffer plus 1 M sucrose and lysed in high-salt buffer (400 mM KCl, 50 mM HEPES, pH 7.0, 1 mM EDTA, 1 mM EGTA, 10% glycerol, 1 mM dithiothreitol, Complete Protease Inhibitor, and phosphatase inhibitors). Samples were centrifuged, and the resulting supernatant contained enriched nuclear proteins.

RESULTS
PKC␦-cat Induces G 2 /M Cell Cycle Arrest-To determine the effect of PKC␦-cat on the cell cycle, we retrovirally expressed PKC␦-cat or an inducible PKC␦-cat-ER fusion protein in cultured primary human KCs, HaCaT cells (Fig. 1A), and MEFs (supplemental Fig. 1). Propidium iodide staining revealed that PKC␦-cat expression increased G 2 /M KCs ϳ2-fold (p Ͻ 0.005) compared with the control Linker virus population (Fig. 1, B and C). PKC␦-cat expression had a similar effect on the immortalized HaCaT cell line, which harbors mutant p53, although the effect was slightly diminished (34). To distinguish between arrest in the G 2 or M phase, mitotic index measurements were performed on PKC␦-cat-ER-transduced HaCaT cells. Fig. 1, D and E, shows that PKC␦-cat-ER caused a significant (p Ͻ 0.001) reduction in mitotic cells, signifying that the cells arrested with 4N DNA were in the G 2 phase. Because PKC␦-cat can also induce apoptosis, we inhibited apoptosis with the caspase 3 inhibitor benzyloxycarbonyl-VAD to determine whether the cell cycle effects were independent of apoptosis. Benzyloxycarbonyl-VAD slightly accentuated the G 2 /M accumulation induced by PKC␦-cat, although the effect was not statistically significant, suggesting that G 2 /M arrest may precede apoptosis (data not shown).
PKC␦-cat Induces G 2 /M Checkpoint Pathway-To have a direct role in cell cycle regulation, it is likely that PKC␦-cat would need to have at least some nuclear localization. To examine this, we performed nuclear/cytoplasmic fractionation on HaCaT cells before and 18 h after exposure to 30 mJ/cm 2 UV radiation. Fractionation results revealed that both the fulllength PKC␦ and PKC␦-cat were present in the nucleus, with PKC␦-cat detected only after UV exposure ( Fig. 2A). UV radiation caused a significant fraction of histone H3 to be detected in the cytoplasmic extract, suggesting that apoptotic degradation of the nuclear membrane occurred in the cells exposed to UV radiation. Confocal microscopy using a carboxyl-terminal GFP fusion of PKC␦ (PKC␦-GFP) and 4Ј,6-diamidino-2-phenylindole staining was used to confirm the subcellular fractionation data. PKC␦-GFP was localized to both the cytoplasm and nucleus before and after UV irradiation (Fig. 2B). This localization provides a means by which PKC␦ may be interacting with important components of the G 2 /M cell cycle checkpoint pathway. We next tested whether PKC␦-cat was sufficient to induce the G 2 /M DNA damage checkpoint. One key aspect of this checkpoint involves the phosphorylation of Cdk1 on Tyr 15 by Wee1 and Myt1 kinases, an event that has been shown to inhibit Cdk1 activity and to prevent entry into mitosis (8,9). We measured relative levels of p-Cdk1 (Tyr 15 ) in PKC␦-cat-expressing KCs and found that PKC␦-cat expression induced elevated levels of p-Cdk1 (Tyr 15 ) compared with the control Linker-transduced cells. The kinase-dead mutant (K378A) did not induce p-Cdk1 (Tyr 15 ) (Fig. 2C). As a positive control, UV radiation also induced p-Cdk1 (Tyr 15 ).
UV Radiation-induced G 2 /M Cell Cycle Arrest Requires PKC␦-Because expression of PKC␦-cat was capable of inducing a pronounced G 2 /M arrest and checkpoint activation in KCs and HaCaT cells, we next determined the requirement for PKC␦ in DNA damage-induced cell cycle arrest. To address this issue, wild-type and PKC␦ null MEFs were exposed to UV radiation (Fig. 3A). We found that exposure of the wild-type MEFs to 30 mJ/cm 2 UV radiation induced a pronounced G 2 /M arrest, which persisted up to 24 h after exposure (Fig. 3, B-D). Strikingly, UV exposure failed to induce G 2 /M arrest in PKC␦ null MEFs, although there was a slight increase in the percentage of cells in S phase. PKC␦ null MEFs did undergo a G 2 /M arrest when transduced with PKC␦-cat-ER virus (supplemental Fig.  1). Mitotic index measurements revealed that UV caused a significant (p Ͻ 0.05) decrease in mitotic wild-type MEFs, but not PKC␦ null MEFs, at 6 h, indicating that the UV radiation-induced arrest was in the G 2 phase (Fig. 3, E and F). These results suggest that PKC␦ is critical for G 2 /M checkpoint integrity after DNA damage.
Interestingly, G 2 /M checkpoint override by the ATM/ATR inhibitor caffeine drove UV radiation-irradiated wild-type MEFs into apoptosis but did not induce apoptosis in PKC␦ null MEFs (supplemental Fig. 2). The lack of apoptosis after caffeine-mediated inhibition of checkpoint activation in PKC␦ null MEFs is likely a reflection of the crucial pro-apoptotic functions of PKC␦. Interestingly, PKC␦ null cells treated with UV radiation plus caffeine accumulated in S phase (p Ͻ 10 Ϫ4 ), indicating a possible S phase checkpoint independent of PKC␦ and ATM/ATR function. G 2 /M Checkpoint Integrity Requires PKC␦ Cleavage and Kinase Activity-We next attempted to rescue G 2 /M checkpoint integrity in PKC␦ null MEFs by retrovirally transducing them with the catalytically competent PKC␦-GFP (Fig. 4A). To address whether PKC␦ kinase activity or caspase cleavage is required to restore the G 2 /M checkpoint response, we also transduced a kinase-dead mutant, PKC␦(K376R)-GFP, or a kinase competent mutant that possesses a mutated caspase cleavage site, PKC␦(D327A)-GFP. All PKC␦-GFP constructs were expressed at similar levels (Fig. 4A). Expression of PKC␦-GFP in PKC␦ null MEFs had little effect on the cell cycle distribution of non-damaged cells (Fig. 4, B and C). Expression of PKC␦-GFP in PKC␦ null MEFs restored the G 2 /M checkpoint response to UV radiation (Fig. 4, B and C). The kinase-dead PKC␦(K376R)-GFP mutant was not able to restore the G 2 /M checkpoint in the PKC␦ null MEF background, indicating that kinase activity is necessary for this PKC␦ checkpoint function. Strikingly, expression of the cleavage site mutant PKC␦ was also unable to restore UV radiation-induced G 2 /M arrest in the PKC␦ null MEFs (Fig. 4, B and C). The dose of UV radiation (30 mJ/cm 2 ) used in these experiments was sufficient to induce caspase 2 and 3 activities, both of which are capable of cleaving PKC␦ (data not shown). This suggests that PKC␦ cleavage is a critically important event in the enforcement of the G 2 /M checkpoint after exposure to UV radiation.
UV Radiation-induced G 2 /M Checkpoint Activation in Wildtype and PKC␦ Null MEFs-We next examined components of the G 2 /M checkpoint pathway before and after UV exposure in wild-type and PKC␦ null MEFs. In agreement with the cell cycle analysis in Figs. 3 and 4, we found that UV exposure induced G 2 /M checkpoint activation as assessed by increased levels of p-Cdk1 (Tyr 15 ) in wild-type but not PKC␦ null MEFs (Fig. 5A). Despite lacking an intact G 2 /M checkpoint response, PKC␦ null MEFs still exhibited UV radiation-induced ␥H2A.X and p-p53 (Ser 15 ), both substrates of ATM and ATR kinases (35)(36)(37). The induction of ␥H2A.X and p-p53 (Ser 15 ), but not p-Cdk1, suggests that PKC␦ may be acting downstream of ATM/ATR in the DNA damage. To examine this, we treated primary human KCs and HaCaT cells ectopically expressing the constitutively active PKC␦-cat with 2 mM caffeine (Fig. 5B). We found that inhibition of ATM/ATR activity by caffeine treatment did not block PKC␦-cat-induced G 2 /M arrest in primary KCs or HaCaT cells, indicating a role for PKC␦-cat in the G 2 /M checkpoint downstream of ATM/ATR activation.
In agreement with the cell cycle data presented in Fig. 4, re-expression of PKC␦-GFP, but not either kinase-dead PKC␦ or caspase cleavage mutant PKC␦, restored phosphorylation of p-Cdk1 (Tyr 15 ) after UV exposure (Fig. 5A). This supports the idea that both kinase activity and caspase cleavage of PKC␦ are important events in the activation of the G 2 /M checkpoint after DNA damage.

DISCUSSION
The majority of studies on PKC␦ function have focused on the pro-apoptotic role that the kinase plays both in vivo and in vitro. We have now demonstrated that in addition to these proapoptotic effects, PKC␦ also regulates cell cycle progression by participating in the G 2 /M DNA damage checkpoint. We have shown that retroviral expression of PKC␦-cat in primary KCs, immortalized KCs with mutant p53, and MEFs is sufficient to induce G 2 /M checkpoint activation and cell cycle arrest (34). PKC␦-cat-induced G 2 /M checkpoint activation requires kinase competent PKC␦ because a kinase-dead PKC␦ mutant failed to activate the G 2 /M checkpoint (Fig. 2C).
The cleavage of PKC␦ appears to be critical to checkpoint integrity because the expression of the caspase-resistant PKC␦(D327A)-GFP mutant was unable to restore the G 2 /M checkpoint in PKC␦ null MEFs (Figs. 4 and 5). PKC␦ is cleavable by numerous proteases including caspases 2 and 3, but the protease responsible for cleaving PKC␦ in this circumstance remains unclear (19,38). It is possible that PKC␦ may be a substrate for the caspase 2 containing DNA-PKcs-PIDDosome complex, which was recently demonstrated to be integral to proper G 2 /M checkpoint maintenance after ionizing radiation exposure (39). PKC␦ appears to be required for G 2 /M checkpoint maintenance rather than induction because a decrease in mitotic index was observed in both wild-type and PKC␦ null MEFs 1 h after UV exposure (data not shown) but not after 6 h (Fig. 3, E and F). Proteins such as p53, p21, and the DNA-PKcs-PIDDosome complex have been similarly implicated in G 2 /M checkpoint maintenance rather than initial checkpoint activation (39,40).
Several PKC␦ substrates involved in the DNA damage response have been identified. These include both p53 and the FIGURE 3. PKC␦ null MEFs fail to arrest in G 2 /M phase following UV irradiation. A, Western blot displaying PKC␦ protein levels in wild-type (WT) and PKC␦ null MEF whole cell lysates. ␣-Tubulin levels are shown as a loading control. B, representative DNA content histograms of wild-type and PKC␦ null MEFs before and 18 h after exposure to 30 mJ/cm 2 UV radiation. C, the percentage of wild-type and PKC␦ null MEFs in G 2 /M phase of the cell cycle before and after exposure to 30 mJ/cm 2 UV radiation is displayed. Error bars denote the S.D. from experiments performed in triplicate. *, Student's t test value of p Ͻ 0.005. D, the percentage of wild-type and PKC␦ null MEFs with G 2 /M DNA content at various times following exposure to 30 mJ/cm 2 UV radiation is displayed. Error bars denote the S.D. from experiments performed in triplicate. #, Student's t test value of p Ͻ 0.005. E, wild-type and PKC␦ null MEFs were exposed to 10 mJ/cm 2 UV radiation, treated with 100 ng/ml nocodazole, and stained for phosphorylated histone H3 (Ser 10 ) and propidium iodide 6 h after UV exposure. Flow cytometry analysis is shown, with the phosphorylated histone H3 (Ser 10 )-positive G 2 /M DNA content cells circled as the mitotic cells. F, quantitation of mitotic indices from HaCaT cells treated as described for D. **, Student's t test value of p Ͻ 0.05. p53 family member p73␤ (15,41,42). PKC␦ has been demonstrated to phosphorylate and activate the DNA repair protein Rad9 after exposure to the DNA-damaging agent 5-azacytidine in an ATM-dependent manner (43). It is possible that ATR, a major kinase activated by UV radiation, is also capable of activating PKC␦ in a similar fashion. Other components of the DNA damage response that have been identified as PKC␦ substrates include DNA-PKcs and topoisomerase II␣ (44,45). Together, these studies reveal the importance of PKC␦ in the DNA damage repair pathway and indicate that it may be functioning at several levels. Nuclear localization of PKC␦ is important for DNA damage-induced apoptosis and may be required during multiple processes following DNA damage, including checkpoint activation (21,22).
The novel cell cycle regulatory role for PKC␦ described in this study has important implications for the role of this protein as a tumor suppressor. Based on these results, the loss of PKC␦ during tumor development would allow for the cell to progress through the cell cycle, even in the presence of DNA damage, while at the same time evading apoptosis. This is illustrated by wild-type MEFs being driven into apoptosis by caffeine-induced G 2 /M checkpoint override after UV irradiation, whereas PKC␦ null MEFs do not arrest or undergo apoptosis even after treatment with caffeine (supplemental Fig. 2). Similar to wildtype MEFs, KCs are highly sensitive to caffeine treatment after UV irradiation (46,47). Thus, KCs lacking PKC␦, such as in squamous papillomas and carcinomas, are also likely to have this dual defect in the response to DNA damage, resulting in both reduced cell cycle arrest and reduced apoptosis (23,24,48).
Furthermore, concomitant treatment of LZRS-PKC␦-cat-ER-transduced cells with the ATM/ATR inhibitor caffeine was unable to prevent PKC␦-cat-ER-induced G 2 /M arrest. This suggests that PKC␦-cat is inducing G 2 /M checkpoint activation independent of the ATM/ATR kinases and may actually function internally to the G 2 /M checkpoint signal cascade. Consistent with this, UV induced ␥H2A.X and p-p53 (Ser 15 ) in PKC␦ null MEFs but did not induce p-Cdk1 (Tyr 15 ) (Fig. 5A). It will be important in the future to identify substrates for PKC␦ within the G 2 /M checkpoint signaling hierarchy. Possible targets include, but are not limited to, the Cdc25 phosphatases, the checkpoint kinases Chk1 and Chk2, and the Cdk1 kinase Wee1 (8,49). The importance of cell cycle checkpoint function to cancer cells has prompted intensive study of small molecule inhibitors targeted toward components of the G 2 /M checkpoint such as ATM/ATR, Chk1, Chk2, and the Aurora family for potential use as cancer therapeutics (50 -52). Better understanding of how PKC␦ participates in this cell cycle checkpoint is important for developing better cancer treatments in the future.