Protein Kinase C (cid:1) Regulates Apoptosis via Activation of STAT1*

ProteinkinaseC (cid:1) (PKC (cid:1) )isrequiredformitochondria-dependent apoptosis; however, little is known about downstream effectors of PKC (cid:1) in apoptotic cells. Here we show that activation of STAT1 is an early response to DNA damage and that STAT1 activation requires PKC (cid:1) . Treatment of HeLa cells with etoposide results in phosphorylation of STAT1 on Ser 727 and the association of STAT1 with PKC (cid:1) . Etoposide increases transcription from STAT1-dependent reporter constructs. Increased transcription, as well as STAT1 Ser 727 phosphorylation, can be blocked by inhibition or depletion of PKC (cid:1) . To ask if STAT1 is required for PKC (cid:1) -mediated apoptosis, we utilized U3A STAT1-deficient cells. Induction of apoptosis by PKC (cid:1) is suppressed in U3A cells but can be rescued by co-transfection with STAT1 (cid:2) but not STAT1 mutated at Ser 727 . Nuclear accumulation of STAT1, phospho-Ser 727 STAT1, and PKC (cid:1) are detectable 30–60 min after treatment with etoposide. Nuclear localization is necessary for apoptosis, since a nuclear localization mutant of PKC (cid:1) does not induce apoptosis in U3A cells reconstituted with STAT1 (cid:2) , and a nuclear localization mutant of Immunoblot analysis was done as previously described using 50 or 100 (cid:4) g of protein (5). For immunoprecipitation, 0.5 or 1 mg of protein was mixed with the appropriate antibody overnight, and the antigen-antibody complexes were recovered by incubation with Sepharose-protein A (Sigma) for 1–4 h at 4 °C. The beads were washed three times in MSLB buffer, and the precipitated proteins were resolved on a 10% SDS-polyacrylamide gel. Plasmid Constructs— Generation of mouse pGFPPKC (cid:1) and pGFPNLMFL8 (cid:1) has been previously described (12). The plasmids pSTAT1 (cid:3) (ST1 (cid:3) ) and pSTAT1 (cid:3) S727A (ST1 (cid:3) S727A) were obtained from Dr. James Darnell (Rockefeller (23). The plasmids pGFPST- AT1 (cid:3) (ST1 (cid:3) ) and pGFPST1 (cid:3) L407A (L407A) were obtained from Dr. Nancy Reich (SUNY (22). The interferon stimulatory re- sponse element (ISRE) and GAS cis-reporting systems and the pCIS-CK negative control plasmid were purchased from Stratagene. To generate the PKC (cid:1) -specific siRNA, pRETRO-SUPER (37) was digested with BglII/HindIII and ligated to the double-stranded oligonucleotide

Protein kinase C␦ (PKC␦) is required for mitochondriadependent apoptosis; however, little is known about downstream effectors of PKC␦ in apoptotic cells. Here we show that activation of STAT1 is an early response to DNA damage and that STAT1 activation requires PKC␦. Treatment of HeLa cells with etoposide results in phosphorylation of STAT1 on Ser 727 and the association of STAT1 with PKC␦. Etoposide increases transcription from STAT1-dependent reporter constructs. Increased transcription, as well as STAT1 Ser 727 phosphorylation, can be blocked by inhibition or depletion of PKC␦. To ask if STAT1 is required for PKC␦-mediated apoptosis, we utilized U3A STAT1-deficient cells. Induction of apoptosis by PKC␦ is suppressed in U3A cells but can be rescued by co-transfection with STAT1␣ but not STAT1 mutated at Ser 727 . Nuclear accumulation of STAT1, phospho-Ser 727 STAT1, and PKC␦ are detectable 30 -60 min after treatment with etoposide. Nuclear localization is necessary for apoptosis, since a nuclear localization mutant of PKC␦ does not induce apoptosis in U3A cells reconstituted with STAT1␣, and a nuclear localization mutant of STAT1 does not support PKC␦-induced apoptosis in U3A cells. Our data identify STAT1 as a downstream target of PKC␦ and suggest that PKC␦ may regulate apoptosis by activation of STAT1 target genes.
Drugs, chemicals, irradiation, and cellular stress induce apoptosis via activation of the mitochondria-dependent, or intrinsic, apoptotic pathway (1)(2)(3). Apoptotic signals act at the mitochondria to induce the release of cytochrome c, activation of the initiator caspase, caspase-9, and the subsequent activation of effector caspases (2,4). PKC␦, a ubiquitously expressed member of the novel subfamily of protein kinase C (PKC) 1 isoforms, is required for activation of this pathway by diverse agents including etoposide (5), ionizing radiation (6), Ara-C (7), FAS ligand (8), brefeldin A, and taxol (9). Studies from our laboratory and others have shown that PKC␦ functions early in the apoptotic pathway and that inhibition of PKC␦ suppresses the release of cytochrome c and caspase activation (7)(8)(9)(10). PKC␦ has been shown to translocate to the nucleus in response to DNA-damaging agents including etoposide and ␥-irradiation (11)(12)(13), and we have previously shown that nuclear localization of PKC␦ is both necessary and sufficient for initiation of the mitochondria-dependent apoptotic pathway (12). Substrates of PKC␦ in apoptotic cells appear to be largely nuclear and include lamin B, the DNA repair protein, DNA-PK, the p53 family member p73, and the cell cycle protein Rad9 (14 -16).
Members of the signal transducer and activator of transcription (STAT) family of transcription factors are induced by interferon and cytokines and regulate the expression of a large group of genes involved in inflammation and antiviral defense (17). Studies in cell lines and transgenic mice suggest that some members of the STAT family also regulate cell death. STAT1 is required for apoptosis induced by ischemia/reperfusion in cardiac myocytes and by tumor necrosis factor-␣, oxysterols, and DNA damage (18 -20). STAT1-deficient U3A human fibroblasts cells are resistant to apoptosis, and this correlates with suppression of caspase and p21WAF1 expression (18,19). Fibroblasts derived from STAT1Ϫ/Ϫ mouse embryos show a reduced p53 response to DNA-damaging agents and increased expression of the p53 inhibitor, Mdm2, suggesting that STAT1 is required for p53-dependent apoptosis (20).
Activation of STAT1, leading to the transcription of STAT1regulated genes, is regulated by phosphorylation at Tyr 701 and Ser 727 . Phosphorylation at Tyr 701 by the Janus family of tyrosine kinases (JAK) leads to STAT1 dimerization via its Src homology 2 domains, exposure of a dimer-specific nuclear localization signal, and subsequent nuclear translocation (21,22). Phosphorylation of STAT1 at Ser 727 is critical for maximal transcriptional activation and for the interaction of STAT1 with transcriptional co-factors (18,23,24). Activated STATs, together with co-factors such as c-Jun, BRCA1, MCM5, IRF-1, IRF-9, and cAMP-response element-binding protein, regulate transcription by binding to GAS or ISRE elements in the promoters of specific target genes (21,25). In the context of apoptosis, phosphorylation of STAT1 at Ser 727 appears to be essential, whereas phosphorylation at Tyr 701 may be dispensable (18, 26 -28). A variety of serine/threonine kinases have been shown to contribute to STAT1 Ser 727 phosphorylation, including members of the extracellular signal-regulated kinase family and Akt (29,30). Recently, STAT1 has been identified as a PKC␦ substrate in cells treated with Type I or Type II interferon (31,32).
The identification of STAT1 Ser 727 as a potential regulatory site for STAT1 proapoptotic functions, and the observation that PKC␦ can phosphorylate STAT1 at Ser 727 in response to some stimuli, suggests a link between these two proapoptotic signaling pathways. Here we show that activation of STAT1 is an early response to DNA-damaging agents and that activation of STAT1 in apoptotic cells requires PKC␦. Our studies indicate that induction of apoptosis requires nuclear translocation of PKC␦ and STAT1 as well as STAT1 Ser 727 phosphorylation. These studies demonstrate a novel and functional interaction between PKC␦ and STAT1 during DNA damage-induced apoptosis.
Transient Transfection-For terminal deoxynucleotidyl transferasemediated dUTP nick end labeling (TUNEL) analysis, subconfluent 2fTGH and U3A cells were transiently transfected using a 6:1 lipid/ DNA ratio of FuGene 6 transfection reagent (Roche Applied Science) according to the manufacturer's protocol. Etoposide (Sigma) was dissolved in Me 2 SO and used at a final concentration of 50 or 100 M as indicated in the figures. Human recombinant interferon-␥ (IFN␥) was purchased from Sigma and added at a final concentration of 5 ng/ml 30 min prior to harvesting. HeLa cells were transfected with the pISRE and pGAS cis-reporter plasmids or the pCIS-CK control plasmid using the calcium phosphate method. Cells were co-transfected with pCMV ␤-galactosidase as a control for transfection efficiency. Luciferase and ␤-galactosidase were assayed as previously described (38). The siRNA vector, pRETRO-SUPER, was introduced into HeLa cells by square wave electroporation.
TUNEL Analysis and Cell Counts-TUNEL analysis was performed using the in situ cell death detection kit TMR Red (Roche Applied Science) according to the manufacturer's protocol. GFP-positive cells were visualized by immunofluorescence microscopy, and TUNEL-positive cells containing GFP were quantified as the percentage of the total GFP-positive cells per field. Greater than 250 cells were counted for each variable per experiment.

Etoposide Regulates STAT1-dependent Transcription-STAT1
is required for apoptosis induced by doxorubicin and cisplatin and for the induction of apoptotic genes such as Bax, Fas, and Noxa, suggesting that induction of STAT1-dependent gene transcription is an important component of apoptosis induced by DNA damage (20). STAT1 drives transcription through binding to GAS or ISRE elements in STAT1-responsive genes (21). To determine whether the DNA-damaging agent, etoposide, can regulate STAT1-dependent transcription, HeLa cells were transfected with promoter-reporter plasmids containing either a GAS or ISRE element, and transfected cells were treated with etoposide for 6 h. As shown in Fig. 1A, both promoters are expressed basally, and the addition of etoposide increases transcription from each promoter Ͼ2-fold. We have previously shown that PKC␦ is required for etoposide induced apoptosis (12). To determine the contribution of PKC␦ to the induction of STAT1-dependent transcription by etoposide, we have used the PKC␦-specific inhibitor, rottlerin, as well as depletion of PKC␦ by expression of a PKC␦-specific siRNA. In the experiment shown in Fig. 1B, HeLa cells were transfected with the pGAS or pISRE reporter plasmids and then pretreated with the PKC␦ inhibitor, rottlerin, prior to the addition of etoposide. Rottlerin has previously been shown to block etoposide-induced apoptosis (5). As shown in Fig. 1B, rottlerin completely suppresses etoposide induction of the pGAS reporter and suppresses induction of the pISRE reporter by about 80%, indicating that PKC␦ activity is required for the induction of STAT1-dependent transcription. In the experiment shown in Fig. 1C, HeLa cells were transfected by electroporation with a plasmid that expresses an siRNA directed against PKC␦ or the vector control. After 48 h, PKC␦ protein is depleted Ͼ90% using this strategy (see Fig. 5C). Cells were then transfected a second time with pGAS luc or pISRE luc using the calcium phosphate method, followed by treatment with etoposide. As shown in Fig.  1C, inhibition of PKC␦ expression suppressed the etoposidemediated induction of both promoters while only slightly suppressing basal expression of these promoters. The etoposide induction of pGAS luc was suppressed by 50%, whereas the induction of pISREluc was reduced by 30%. Taken together, these data suggest that etoposide induces transcriptional activation of STAT1-responsive genes in apoptotic cells and that this activation requires PKC␦ activity. In the case of both rottlerin and the siRNA to PKC␦, etoposide induction of the GAS promoter was suppressed to a greater extent than induction of the ISRE promoter. This differential sensitivity may reflect the fact that the GAS element preferentially binds STAT1 homodimers, whereas ISRE element preferentially binds STAT1/STAT2 heterodimers (21).
STAT1 Is Phosphorylated at Ser 727 in Response to Etoposide-Activation of STAT1, leading to the transcription of STAT1-regulated genes, is regulated by phosphorylation of STAT1 at Tyr 701 and Ser 727 . Phosphorylation at Tyr 701 is important for nuclear translocation and retention (22,39,40), whereas phosphorylation at Ser 727 is required for maximal transcriptional activation of STAT1-dependent genes, including genes required for the apoptotic response (23,28). To ask whether changes in phosphorylation of STAT1 occur in response to etoposide, we utilized antibodies that specifically recognize STAT1 phosphorylated at Tyr 701 or Ser 727 . As demonstrated in Fig. 2A (top), STAT1 is phosphorylated at Ser 727 in untreated HeLa cells, and treatment with etoposide results in an increase in phosphorylation at this site within 60 min, which is sustained for at least 4 h. In contrast, no basal phosphorylation at STAT1 Tyr 701 is detectable in HeLa cells ( Fig.  2A, middle). Likewise, phospho-STAT1 Tyr 701 was not detectable in cells treated with etoposide; however, Tyr 701 phosphorylation was observed in response to IFN␥ treatment as has been previously reported. Previous studies from other labora-tories suggest that STAT1 phosphorylated at Tyr 701 is dephosphorylated in the nucleus and subsequently exported into the cytoplasm (39 -41). To investigate whether the lack of detectable STAT1 phospho-Tyr 701 is due to its rapid dephosphorylation, HeLa cells were pretreated with pervanadate prior to the addition of etoposide. Whereas some accumulation of phospho-Tyr 701 STAT1 was observed in pervanadate pretreated cells, etoposide did not further increase the amount of phospho-Tyr 701 STAT1 (data not shown). Tyrosine phosphorylation of STAT1 is mediated by the JAK tyrosine kinases, JAK1, JAK2, and TYK2. To further address the contribution of Tyr 701 phosphorylation to etoposide-induced STAT1 activation, we have utilized three human fibroblast cells lines, each deficient in one of these tyrosine kinases. As shown in Fig. 2B, treatment of the control cell lines, 2fTGH and 2C4, with etoposide results in the rapid phosphorylation of STAT1 on Ser 727 . A similar time course of STAT1 Ser 727 phosphorylation was observed in U1A (TYK2-deficient) and U4A (JAK1-deficient) cells. Etoposideinduced STAT1 S727 phosphorylation was also observed in the ␥2A (JAK2-deficient) cells; however, in this cell line, the time course was slightly delayed, and phosphorylation appeared to be more robust, suggesting that a JAK2-regulated pathway may suppress the etoposide-induced phosphorylation of STAT1. Taken together, these data indicate that etoposideinduced activation of STAT1 does not require the JAK family of tyrosine kinases and suggests that Ser 727 phosphorylation of STAT1 is sufficient for activation of STAT1 in apoptotic cells.
Association of STAT1 and PKC␦ in Etoposide-treated Cells-Recent studies suggest that PKC␦ is required for phosphorylation of STAT1 at Ser 727 in response to IFN␣, IFN␥, and interleukin-6 (31,42). To determine the contribution of PKC␦ to etoposide-induced phosphorylation of STAT1, cell lysates were immunoprecipitated with an antibody to STAT1 and immunoblotted for PKC␦ or phospho-Ser 727 STAT1 (Fig. 3A). An association of STAT1 with full-length PKC␦ was observed as early as 30 min after etoposide treatment and was concomitant with phosphorylation of STAT1 at Ser 727 (Fig. 3A, top and  bottom). In the experiment shown in Fig. 3B, lysates from etoposide-treated cells were immunoprecipitated with anti-PKC␦ and immunoblotted for STAT1. Immunoprecipitation of PKC␦ likewise demonstrated an etoposide-dependent association with STAT1, which was detectable within 30 min of etoposide stimulation. Since in HeLa cells the cleavage product of PKC␦ does not accumulate until 12-24 h after etoposide treatment (data not shown), whereas PKC␦ and STAT1 associate within 30 min, our data suggest that STAT1 interacts with PKC␦ prior to its caspase cleavage. Furthermore, as shown in Fig. 3C, pretreatment of HeLa cells with the broad spectrum caspase inhibitor benzyloxycarbonyl-VAD-fluoromethyl ketone does not inhibit the association of PKC␦ and STAT1 or the phosphorylation of STAT1 Ser 727 in etoposide-treated cells. Thus, caspase activation is not required for interaction of PKC␦ with STAT1 or for phosphorylation of STAT1 at Ser 727 .
Our studies indicate that in HeLa cells the full-length PKC␦ protein associates with STAT1 prior to caspase cleavage of PKC␦. To address this further, we have utilized parotid C5 cells, which undergo early and robust cleavage of PKC␦ in response to etoposide (5). Fig. 3D shows that in parotid C5 cells, etoposide induces the association of PKC␦ and STAT1 by 2 h, and this association is greatly diminished by 4 h. Loss of the association of PKC␦ and STAT1 parallels the initiation of cleavage of PKC␦ at 4 h as shown in Fig. 3E. Taken together, these studies demonstrate that the etoposide-induced association of PKC␦ and STAT1 is an early event in the apoptotic pathway and that loss of this association is co-incident with caspase cleavage of PKC␦. Additionally, the observed co-association between PKC␦ and STAT1 in response to etoposide in both HeLa and salivary epithelial cells suggests that this interaction is not cell type-dependent and probably represents a general occurrence in response to DNA-damaging agents.
Previous studies from our laboratory and others suggest a role for PKC␦ and STAT1 in the nucleus of apoptotic cells. To determine whether nuclear translocation of STAT1 and PKC␦ accompanies etoposide-induced apoptosis in HeLa cells, we prepared nuclear extracts from etoposide-treated cells and analyzed PKC␦, total STAT1, and phospho-Ser 727 STAT1 by immunoblot. As shown in Fig. 4, the accumulation of total STAT1 and phospho-Ser 727 STAT1 can be seen in the nucleus 15-30 min after the addition of etoposide, with maximal nuclear accumulation by 30 -60 min. Nuclear accumulation of PKC␦ follows a similar time course, with peak accumulation seen 30 -60 min after the addition of etoposide. These results sug-gest that PKC␦-dependent activation of STAT1 occurs in the nucleus of apoptotic cells.
PKC␦ Is Required for Etoposide-induced Phosphorylation of STAT1 at Ser 727 -To determine whether PKC␦ is required for Ser 727 phosphorylation of STAT1 in response to etoposide, HeLa cells were pretreated with the PKC inhibitor Go6976, which inhibits PKC␣, -␤, and -␥; Go6983, which inhibits PKC␣, -␤, -␥, -␦, and -; or rottlerin, which inhibits PKC␦. As demonstrated in Fig. 5A, top, pretreatment with Go6983 or rottlerin, both of which inhibit PKC␦, suppresses etoposide-induced Ser 727 phosphorylation of STAT1, whereas pretreatment with Go6976, which inhibits only the conventional PKC isoforms, does not. To further investigate the contribution of PKC␦ to STAT1 Ser 727 phosphorylation, we utilized siRNA to deplete HeLa cells of endogenous PKC␦. As shown in Fig. 5C, expression of siRNA to PKC␦ results in depletion of Ͼ90% of the PKC␦ protein. The addition of etoposide to HeLa cells transfected with the vector control results in phosphorylation of FIG. 3. Etoposide induces the association of PKC␦ with STAT1. A, HeLa cells were treated for the indicated times with 100 M etoposide, and whole cell lysates were prepared as described under "Experimental Procedures." Lysates were immunoprecipitated (IP) with an antibody to STAT1 and immunoblotted for PKC␦ (top), total STAT1 protein (middle), or phospho-Ser 727 STAT1 (P-S727; bottom). B, cell lysates were immunoprecipitated with an antibody to PKC␦ and immunoblotted for STAT1 (top) or PKC␦ (bottom). C, HeLa cells were incubated with or without 100 M benzyloxycarbonyl-VAD-fluoromethyl ketone (ZVAD) for 1 h, followed by treatment for an additional 1 h with 100 M etoposide (E). Lysates were immunoprecipitated with an antibody to STAT1 or Protein A beads alone. Proteins were separated by SDS-PAGE and immunoblotted for phospho-Ser 727 STAT1 (top), total STAT1 protein (middle), or PKC␦ (bottom). WCL, whole cell lysates. UT, untreated. D, parotid C5 cells were treated with 50 M etoposide for the indicated times, and whole cell lysates were immunoprecipitated with an antibody to the C terminus of PKC␦ and immunoblotted for STAT1. E, whole cell lysates from the experiment shown in B were immunoblotted with an antibody directed to the C terminus of human PKC␦. Representative experiments are shown; each experiment was repeated three or more times. STAT1 at Ser 727 , whereas this is suppressed in cells transfected with the siRNA to PKC␦ (Fig. 5B). When the abundance of STAT1 phospho-Ser 727 is normalized to total STAT1, the average of three experiments shows that etoposide increases the abundance of phospho-Ser 727 by Ͼ3-fold in vector-transfected cells, whereas in cells depleted of PKC␦, the increase in phospho-Ser 727 is reduced to about 1.6-fold. These studies indicate that etoposide-induced phosphorylation of STAT1 at Ser 727 is a PKC␦-dependent event. The experiments in Figs. 1 and 5 show a more robust suppression of both STAT1-dependent transcription and STAT1 Ser 727 phosphorylation by rottlerin as compared with the PKC␦ siRNA. This may be due to incomplete depletion of PKC␦ by the PKC␦ siRNA or may indicate contributions from additional signaling pathways.

Induction of Apoptosis by Overexpression of PKC␦ Requires both Nuclear Localization and Ser 727
Phosphorylation of STAT1-The data shown above suggest that PKC␦ may regulate apoptosis via STAT1-dependent transcription and that nuclear accumulation of these signaling proteins may be required for their interaction. Transient overexpression of PKC␦ is a potent inducer of apoptosis in a variety of cell types (5-7, 10, 43-47). To ask if PKC␦ induces apoptosis via a STAT1-dependent mechanism, we have utilized the STAT1-deficient human fibroblast cell line, U3A (18,33). U3A cells or the parental 2fTGH cell line were transiently transfected with pGFPPKC␦ or pGFPNLMFL8␦, a PKC␦ nuclear localization mutant that does not induce apoptosis (12). As seen in Fig. 6A, pGFPPKC␦ induced apoptosis in 12% of transfected 2fTGH cells as determined by TUNEL staining, whereas only 4% of cells transfected with pGFPNLMFL8␦ were TUNEL-positive, verifying our previous findings that nuclear localization of PKC␦ is required for induction of apoptosis (12). In contrast, when U3A cells were transfected with pGFPPKC␦, there was no increase in GFP/TUNEL-positive cells over that seen in cells transfected with pGFPNLMFL8␦, demonstrating that STAT1 is required for PKC␦-induced apoptosis. A requirement for STAT1 was verified by reconstituting U3A cells by transfection with pSTAT1␣ together with PKC␦. As seen in Fig. 6A, in STAT1reconstituted U3A cells, pGFPPKC␦-induced apoptosis is rescued and actually enhanced. However, no induction of apoptosis was seen when U3A cells were reconstituted with STAT1 and transfected with pGFPNLMFL8␦, indicating that PKC␦ must translocate to the nucleus to induce apoptosis via STAT1.
The data shown in Fig. 5 suggest that PKC␦ is required for STAT1 Ser 727 phosphorylation in etoposide-treated cells. To determine the contribution of STAT1 phosphorylation to PKC␦induced apoptosis, U3A cells were reconstituted with STAT1 in which the serine at position 727 was mutated to alanine (pST1␣ S727A). Unlike wild type STAT1, expression of pST1␣ S727A was unable to restore apoptosis in pGFPPKC␦ co-transfected cells, indicating that Ser 727 phosphorylation of STAT1 is required for apoptosis induced by PKC␦.
The data in Fig. 6A demonstrate that PKC␦ must have access to the nucleus to induce apoptosis via a STAT1-dependent pathway, suggesting that STAT1 is a downstream target of nuclear PKC␦ in apoptotic cells. To determine whether nuclear translocation of STAT1 is also required for apoptosis induced by PKC␦, 2fTGH and U3A cells were transiently transfected with pGFP PKC␦ together with pGFPSTAT1␣ or pGFPST1L407A␣, which is mutated in a nuclear localization sequence and does not localize to the nucleus (22). As shown previously, expression of pGFPPKC␦ alone does not induce apoptosis in U3A cells, whereas co-transfection of pGFPPKC␦ together with pGFPSTAT1␣ results in a complete rescue of pGFPPKC␦-induced apoptosis (Fig. 6, A and B). In contrast, reconstitution of U3A cells with the STAT1 nuclear localization mutant, pGFPL407A␣, does not rescue PKC␦-induced apoptosis in the U3A cells. Analysis of the localization of STAT1 in GFP-transfected cells is shown in Fig. 6C. In U3A cells reconstituted with pGFPSTAT1 alone, STAT1 is localized to the nucleus in Ͼ40% of the transfected cells. However, cotransfection of pGFPSTAT1 together with PKC␦ results in accumulation of STAT1 in the nucleus of 85% of the transfected cells, correlating with the induction of apoptosis seen in Fig.  6B. In contrast, nuclear translocation of STAT1 was not observed in U3A cells co-transfected with pGFPPKC␦ together with pST1L407A. These studies suggest that activation of an apoptotic program by PKC␦ induces the nuclear translocation of STAT1 and implies a functional role for nuclear STAT1 in the apoptotic pathway downstream of nuclear PKC␦ signaling. DISCUSSION PKC␦ is required for apoptosis induced by a wide variety of stimuli in many cell types. Whereas a limited number of substrates for PKC␦ have been described in apoptotic cells, the functional consequences of phosphorylation of these substrates is in large part unknown. Here we have explored the hypothesis that PKC␦ communicates with the cellular apoptotic machinery via activation of the transcription factor STAT1. Our studies show that etoposide induces a STAT1 transcriptional program that requires PKC␦ activity and demonstrate that these two important proapoptotic signaling pathways are functionally linked in cells undergoing apoptosis. We propose that in response to DNA-damaging agents, an apoptotic program is induced, which includes PKC␦-mediated activation of STAT1regulated gene expression.
We demonstrate that STAT1 is phosphorylated on Ser 727 in HeLa cells treated with etoposide and that this correlates with the co-association of STAT1 with full-length PKC␦. Using two inhibitor strategies, we show that STAT1 Ser 727 phosphorylation in apoptotic cells is dependent at least in part on PKC␦. This is in agreement with published studies that show that in cells treated with Type I and Type II interferons and interleukin-6, PKC␦ is required for the induction of STAT1 transcriptional activity and for Ser 727 STAT1 phosphorylation. PKC␦ has also been shown to directly phosphorylate STAT1 at Ser 727 in vitro (31,32). We find no evidence that etoposide induces phosphorylation of PKC␦ at Tyr 701 or that etoposide activation of STAT1 requires the JAK family of tyrosine kinases. The obligatory role for Tyr 701 phosphorylation of STAT1 in the context of apoptosis is somewhat controversial. Induction of apoptosis by tumor necrosis factor-␣ and actinomycin D or oxysterol requires Ser 727 phosphorylation but not Tyr 701 phosphorylation of STAT1 (18,19). However, Tyr 701 phosphorylation is required to enhance cell death during induction of apo- ptosis by heat shock or ischemia (48). Additionally, studies have found that STAT1 can regulate transcription of genes in both tyrosine phosphorylation-dependent and -independent contexts (28,49).
Our current studies demonstrate that both PKC␦ and STAT1 are translocated to the nucleus within 30 min of etoposide treatment and that this is concurrent with the accumulation of phospho-Ser 727 STAT1 in the nucleus. Whereas etoposide does not increase phosphorylation of STAT1 at Tyr 701 , we were able to detect a small amount of Tyr 701 -phosphorylated STAT1 in both untreated and etoposide-treated cells (data not shown), suggesting that basal phosphorylation of STAT1 at tyrosine 701 may contribute to its nuclear accumulation in apoptotic cells. However, it should be noted that even in etoposidetreated cells, the majority of total STAT1 and phospho-Ser 727 STAT1 is found in the cytosol (data not shown). Since Tyr 701 phosphorylation of STAT1 has recently been shown to be necessary for its retention in the nucleus (40), STAT1 may be exported more readily from the nucleus in etoposide-treated cells where Tyr 701 phosphorylation is minimal. Several recent studies suggest that in resting cells, unphosphorylated STAT1 shuttles between the nucleus and cytosol (27,40,50). We propose that the addition of etoposide induces an increase in Ser 727 phosphorylation, and this may increase the abundance of STAT1 in the nucleus by increasing the binding affinity of STAT1 for DNA.
To examine more directly the relationship between PKC␦ and STAT1 in apoptosis, we have utilized a PKC␦ overexpres-sion model for apoptosis. Our studies clearly indicate that in the absence of STAT1, overexpression of PKC␦ does not induce apoptosis. Furthermore, we demonstrate that both PKC␦ and STAT1 must have access to the nucleus for apoptosis to occur. Nuclear accumulation of STAT1 and PKC␦ occurs both in etoposide-treated HeLa cells and in STAT1 reconstituted U3A cells transiently transfected with PKC␦. Our studies demonstrate that the nuclear localization of both STAT1 and PKC␦ are required for apoptosis induced by full-length PKC␦. Since the nuclear localization signal mutant of full-length PKC␦ did not induce apoptosis, this finding supports our previous results that nuclear localization of PKC␦ is required for the induction of apoptosis (12). Recently, it has been shown that STAT1 Leu 407 is essential for the interaction of STAT1 with importin ␣ and subsequent nuclear translocation (22). The STAT1 Leu 407 mutant, however, retains the ability to become Tyrphosphorylated, to dimerize, and bind DNA in vitro (22). Transiently transfected STAT1 L407A did not localize to the nucleus and was not able to rescue PKC␦-induced apoptosis in U3A cells, indicating that nuclear localization of STAT1 is required for induction of apoptosis by PKC␦. In accordance with these results, recent data from another laboratory has demonstrated that a constitutively active dimerized form of STAT1 that is retained in the nucleus sensitizes cells to IFN␥-induced apoptosis by inducing activation of caspase-2, -3, and -7 (51).
A model of apoptosis has been proposed in which apoptotic agents activate PKC␦, resulting in the activation of caspase-3 and the subsequent cleavage of PKC␦ to generate a constitu- tively active fragment of this kinase. Our laboratory and others have proposed that this constitutively active fragment of PKC␦ is rapidly translocated into the nucleus and functions to amplify the apoptotic pathway (12,52,53). In this regard, PKC␦ has been shown to have a functional role in the nucleus of apoptotic cells, and several nuclear binding partners for the caspase cleavage fragment of PKC␦ have been identified such as p73␤, DNA-PK, lamin B, c-Abl, SHPTP1, and Rad9 (13, 15, 54 -56). We have previously demonstrated that nuclear translocation of PKC␦ in etoposide-treated cells also occurs independently of caspase cleavage, albeit less efficiently, and have suggested a role for full-length PKC␦ in the apoptotic pathway prior to caspase cleavage (12). Our current data show that STAT1 and PKC␦ associate prior to caspase cleavage of PKC␦, providing further support for a role for the full-length form of PKC␦ in apoptotic cells.
How full-length PKC␦ is activated by DNA-damaging agents and how this activation leads to caspase-3 activation is largely unknown. Tyrosine phosphorylation of PKC␦ occurs very rapidly upon exposure of cells to DNA-damaging agents and is likely to be one mechanism by which PKC␦ is "activated" in apoptotic cells. Although numerous studies have attempted to demonstrate a direct effect of tyrosine phosphorylation on the catalytic activity of PKC␦ with variable results, mutation of specific tyrosine residues has been shown to suppress the ability of PKC␦ to induce apoptosis in transfected cells (11). 2 We propose that tyrosine phosphorylation links PKC␦ to downstream signaling pathways by creating docking sites for Src homology 2 domain proteins such as STAT1. PKC␦-mediated activation of these signaling pathways may be required for transcriptional reprogramming of the cell to facilitate apoptosis. This may include induction of STAT1-dependent proapoptotic genes, such as caspases and Bcl-2 family members, and STAT1-dependent regulation of p53-dependent genes such as Bax and NOXA (51). In addition, PKC␦-dependent activation of other signal transduction pathways such as the c-Jun N-terminal kinase pathway has been reported and may lead to transcriptional regulation of a distinct subset of pro-or antiapoptotic genes (57). The summation of these multiple transcriptional end points may serve to activate the apoptotic pathway directly or to "prime" the pathway for activation by additional apoptotic signals.