Apoptosis Induced by the Nuclear Death Domain Protein p84N5 Is Associated with Caspase-6 and NF-κB Activation*

Although the mechanisms involved in responses to extracellular or mitochondrial apoptotic signals have received considerable attention, the mechanisms utilized within the nucleus to transduce apoptotic signals are not well understood. We have characterized apoptosis induced by the nuclear death domain-containing protein p84N5. Adenovirus-mediated N5 gene transfer or transfection of p84N5 expression vectors induces apoptosis in tumor cell lines with nearly 100% efficiency as indicated by cellular morphology, DNA fragmentation, and annexin V staining. Using peptide substrates and Western blotting, we have determined that N5-induced apoptosis is initially accompanied by activation of caspase-6. Activation of caspases-3 and -9 does not peak until 3 days after the peak of caspase-6 activity. Expression of p84N5 also leads to activation of NF-κB as indicated by nuclear translocation of p65RelA and transcriptional activation of a NF-κB-dependent reporter promoter. Changes in the relative expression level of Bcl-2family proteins, including Bak and Bcl-Xs, are also observed during p84N5-induced apoptosis. Finally, we demonstrate that p84N5-induced apoptosis does not require p53 and is not inhibited by p53 coexpression. We propose that p84N5 is involved in an apoptotic pathway distinct from those triggered by death domain-containing receptors or by p53.

Programmed cell death (PCD) 1 is essential for normal development, tissue homeostasis, and host defense mechanisms. PCD is recognized by a collection of distinctive morphological and biochemical characteristics termed apoptosis (1). An intrinsic genetic program that is conserved through evolution controls PCD in mammals (2). For example, the ced-3 gene that is required for PCD in Caenorhabditis elegans encodes a protein that is a functional homologue of the family of mammalian cysteine proteases termed caspases (3). Activation of these caspases by proteolytic processing is required for the execution of apoptosis. Members of the mammalian Bcl-2 family of apoptotic regulatory proteins also have a homologue in C. elegans, the ced-9 gene (4). Some members of this family, like Bcl-2 and Bcl-XL, inhibit apoptosis, whereas others like Bax, Bak, or Bcl-Xs promote apoptosis (5).
The pathways leading to apoptosis in response to extracellular stimuli, like tumor necrosis factor, or in response to mitochondrial apoptotic signals, such as cytochrome c release, are well characterized (6,7). Initiator caspases are typically recruited to protein complexes composed of death receptors and/or adapter molecules. These proteins contain signature protein interaction motifs like the death domain, the death effector domain, or the CARD domain. The locally high concentration of recruited caspase proenzyme presumably leads to proteolytic processing to more active forms, thereby triggering a caspase proteolytic cascade. Apoptotic signals can also originate from within the nucleus. For example, DNA damage caused by radiation triggers a stress response that can result in apoptotic cell death (8). The mechanisms utilized by nuclear apoptotic signals to initiate caspase proteolytic cascades are not well understood.
A number of nuclear transcription factors can induce apoptosis, including p53. Although p53 has a well-documented role in the cell's response to DNA damage, the mechanism used by p53 to initiate apoptosis is controversial. Although it is presumed to trigger apoptosis from within the nucleus by altering the expression of genes more directly involved in the execution of apoptosis, non-nuclear mechanisms unrelated to transcriptional regulation have also been proposed (12). Few proteins other than transcription factors have been documented to initiate apoptosis from within the nucleus. Nuclear localization of expanded polyglutamine repeat proteins is required for their ability to induce apoptosis that causes progressive neurodegenerative diseases like Huntington's disease or spinocerebellar ataxia (9,10). Activation of caspase-8 is a required step in this process (11). Activation of caspase-8 apparently occurs by a novel mechanism involving recruitment of the proenzyme to characteristic protein aggregates that are associated with these diseases.
Recently, we demonstrated that the nuclear protein encoded by the N5 gene (p84N5) was capable of initiating apoptosis (13). The N5 gene was originally isolated based on the ability of p84N5 to bind an amino-terminal domain of the retinoblastoma tumor suppressor protein (pRb) (14). Association of p84N5 with pRb inhibits p84N5-induced apoptosis, identifying p84N5 as a potential mediator of pRb's inhibitory effects on apoptosis (15). Because p84N5 is unique among nuclear proteins in containing a death domain that is required for its ability to induce apoptosis (13), it may participate directly in a nuclear apoptotic pathway. In the current study, we characterize this pathway by analyzing caspase activation, Bcl-2 family member expression, the effects of p53, and NF-B activation during p84N5-induced apoptosis.

MATERIALS AND METHODS
Cell Culture-SAOS-2, 293, MCF-7, and C-33A cell lines were obtained from the American Type Culture Collection and maintained in Dulbecco's modified Eagle's medium with 10% heat-inactivated fetal bovine serum and antibiotics (100 units/ml penicillin, 100 g/ml streptomycin), in a 5% CO 2 incubator at 37°C. Growth curves with these cell lines were assayed by plating 500,000 cells per well of a 60-mm tissue culture dish, treating with indicated virus at an multiplicity of infection (m.o.i.) of 10, harvesting cells by trypsinization at the indicated time after infection, and counting the number of cells that exclude trypan blue with a hemacytometer.
The Colo357 X cell line was obtained from Dr. Keping Xie (M.D. Anderson Cancer Center) and was grown as above. Growth assays of virally infected Colo357 X cells were performed by plating at a density of 20,000 cells/well in 24-well plates in triplicate. One day later, the cells were infected with the indicated adenoviruses at a total m.o.i. for all viruses of 50. An equal number of cells were treated with PBS as a control. Cells were harvested at different time intervals and viable cells, as indicated by trypan blue exclusion, were counted using a hemacytometer.
Plasmids and Adenovirus-The full-length p84N5 cDNA was subcloned into pCEP4 (Invitrogen, Carlsbad, CA) expression vector to create the expression vector pCMVN5. The Bcl-2 (16), p53 (17), and pRb (18) expression vectors were previously described. The NF-kB luciferase reporter plasmids (19) and RelA expression plasmid (20) were described previously. To generate the recombinant p84N5 expressing adenovirus, the p84N5 cDNA was subcloned into the pAdCMV (AS)-BGHpa vector (Dr. T. J. Liu, M.D. Anderson Cancer Center). The resulting plasmid, was coelectroporated with pJM17, the adenovirus backbone plasmid, into 293 cells and recombinant N5 adenovirus (AdN5) identified by polymerase chain reaction using primers specific for N5 essentially as described (21). Recombinant adenovirus was purified by CsCl equilibrium density gradient centrifugation, and viral particle numbers were estimated by A 260 in the presence of SDS as described (22). Infectious titer was determined by an end point assay using an adenovirus direct immunofluorescence kit (Chemicon International, Inc.). The GFP-and p53-expressing adenoviruses were gifts of Dr. T. J. Liu and were prepared as above.
Transfection and Apoptosis Assays-For transfections, cells were seeded on 100-mm dishes and transfected the following day by calcium phosphate precipitation (23) using 6 -30 g of total DNA. To analyze transfected or infected cells for DNA fragmentation, cells were collected 24 h after treatment and free DNA ends were stained with terminal deoxytransferase (TUNEL) using the APO-DIRECT kit (Phenix Flow Systems, San Diego, CA) according to manufacture's instructions. Annexin V staining was assayed 2 days following infection by harvesting cells by trypsinization, washing with PBS, and staining with annexin V Alexa 568 following the manufacturer's instructions (Roche Molecular Biochemicals). FACS analysis was performed on a FACSCalibur instrument (Becton Dickinson, San Jose, CA). Hoechst 33342 (Roche Molecular Biochemicals) staining was performed by adding the dye at a final concentration of 1 g/ml to the culture media. Cells were incubated with the dye for 10 min and immediately photographed at 100ϫ magnification under a fluorescence microscope with a UV filter. To assay activation of caspase activity, cells were infected by addition of AdGFP or AdN5 to the culture media. Cells were incubated at 37°C for the desired length of time and then collected, washed twice with PBS, and extracted, and the extracts were analyzed for caspase activity using colorimetric peptide substrates as directed by the manufacturer (Bio-Vision, Palo Alto, CA). For analysis of luciferase reporter gene activity, transfected cells were washed twice with PBS and lysed, and aliquots of the lysate corresponding to 5 ϫ 10 4 successfully transfected cells were assayed for luciferase activity using reagents from the Analytical Luminescence Laboratory (Ann Arbor, MI). The pEGFP-C1 vector was included in the transfections to determine the transfection efficiency. Cells were examined for GFP by fluorescence microscopy prior to extraction.
Western Blotting-For Western analysis, transfected cells were extracted in a buffer containing 50 mM Tris, pH 7.4, 250 mM NaCl, 5 mM EDTA, 0.1% Nonidet P-40, 50 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin on ice for 10 min. The total protein concentration of the soluble extract was determined by Bradford assay according to the manufacturer's instructions (Bio-Rad, Hercules, CA). 70 g of total protein for each sample was resolved by 10% SDS-polyacrylamide gel electrophoresis, blotted, and stained as described previously (13). Antibodies directed against p84N5 (14) and pRb (24) were described previously. All other primary antibodies were used as directed by the manufacturer. Antibodies directed against caspases were obtained from Oncogene Research Products (Cambridge, MA); all other antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Primary antibodies were detected using a peroxidase-conjugated secondary antibody and enhanced chemiluminescence according to the manufacturer's recommendations (Amersham Pharmacia Biotech).

Expression of p84N5 Induces Apoptosis through Activation of
Caspase-6 -The full-length, wild type N5 cDNA was used to generate a recombinant, E1-deleted, replication-defective adenovirus (AdN5) that expressed wild-type p84N5 under control of the cytomegalovirus early promoter and the bovine growth hormone polyadenylation signal. This adenovirus was used to drive expression of p84N5 in infected cells. The proliferation of cells infected with AdN5 at a multiplicity of infection (m.o.i.) of 10 slowed significantly by the third day after infection (Fig.  1A). Nearly all AdN5-infected cells died by 6 days after infection. The proliferation of cells infected with a similarly constructed GFP-expressing recombinant adenovirus (AdGFP) was similar to uninfected cells. Both AdGFP-infected and untreated cells reached confluency by day 4, explaining the decrease in proliferation rate observed at the later stages of the experiment. To test whether the observed decrease in cell viability of AdN5-infected cells was due to apoptosis, cells were assayed for the presence of fragmented DNA by TUNEL 3 days after infection. Nearly 80% of cells infected with AdN5 contained fragmented DNA (Fig. 1B). Infection with AdGFP did not alter the percentage of TUNEL-positive cells relative to uninfected cells. Consistent with this observation, AdN5-infected cells also demonstrated an increase in phosphatidylserine exposure on the outer leaflet of the plasma membrane as determined by annexin V staining (Fig. 1C). Finally, AdN5infected cells exhibited an increased permeability to Hoechst 33342 and condensed chromatin (Fig. 1D). AdGFP-infected cells did not show increases in annexin V or Hoechst 33342 staining. To ensure that apoptosis induced by AdN5 coincided with increased expression of p84N5, protein extracts prepared from infected cells were analyzed by Western blotting using a mouse anti-N5 monoclonal antibody. 24 h after infection, the level of p84N5 expression had increased in AdN5-treated cells at least 5-fold relative to AdGFP-treated cells (Fig. 1E). The efficiency of adenoviral-mediated gene transduction under the conditions used in these experiments was 98% as determined by fluorescent microscopy of AdGFP-infected cells.
To identify the caspases activated during N5-induced apoptosis, extracts prepared from infected cells have been assayed for caspase activity using seven different colorimetric peptide substrates. Extracts prepared 48 h after infection contain significant VEID peptide cleavage activity ( Fig. 2A). VEID is an efficient substrate for caspase-6 (25). Interestingly, the level AdN5-induced DEVD cleavage activity does not increase above background until 4 days after infection, and the relative extent of activation is lower than VEID cleavage activity. DEVD is an efficient substrate for caspases-3 and -7. A small increase in LEHD cleavage activity is also observed 4 -5 days after infection. LEHD is an efficient substrate for caspase-9. Increases in cleavage activity of the other peptides tested, which include efficient substrates for capases-1, -2, -4, -5, and -8, were not detected upon AdN5 infection. AdGFP-infected cell extracts did not contain cleavage activity above background for any of the peptides tested. 2 To confirm that caspase-6 was involved in the increase in VEID cleavage activity upon AdN5 infection, extracts were analyzed for proteolytic processing of caspases-6 and -9 by Western blotting. Proteolytically processed caspase-6 was detected in extracts prepared 2 or 3 days after infection with AdN5, but not in extracts prepared from AdGFP-infected cells (Fig. 2B). Processing of caspase-9 was also detected in extracts prepared 4 days after infection. These observations indicated that the activation of VEID cleavage activity detected was likely due, at least in part, to activation of caspase-6.
Activation of NF-B during p84N5-induced Apoptosis-Activation of NF-B serves as a survival signal in response to varied apoptotic stimuli, including genotoxic agents like ionizing radiation (26). Activation of NF-kB involves translocation of the protein from the cytoplasm to the nucleus where it functions as a transcription factor to regulate the expression of target genes. To examine possible activation of NF-B by p84N5, nuclear or cytoplasmic extracts prepared from 293 cells transfected with pCMVN5 or empty vector have been analyzed by Western blotting using an anti-p65RelA monoclonal antibody. In comparison to cells transfected with empty vector, the level of p65RelA in the cytoplasm of pCMVN5-transfected cells decreases while the relative level of p65RelA in the nucleus increases (Fig. 3A). To test whether the nuclear translocation of p65RelA coincides with increased transcriptional activity, a luciferase reporter gene driven by an artificial NF-B-responsive promoter was cotransfected with pCMVN5 or empty vector, and the levels of luciferase activity in the extracts of transfected cells were assayed. Expression of p84N5 potently increases luciferase activity generated from the NF-B-responsive promoter construct but not from a reporter construct containing a mutant promoter that is insensitive to NF-B (Fig. 3B). The level of reporter gene induction by p84N5 expression was similar to that achieved by expression of exogenous p65RelA.
Changes in Bcl-2 Family Member Gene Expression during p84N5-induced Apoptosis-Bcl-2 and its homologues are critical regulators of cell death. Changes in the expression of either pro-apoptotic or anti-apoptotic family members can facilitate apoptosis. We have examined the expression of Bak, Bax, Bad, Bcl-Xs, and Bcl-2 during p84N5-induced apoptosis by Western blot analysis. Transfection of pCMVN5 into 293 cells or SAOS-2 cells causes detectable apoptosis within 48 h (13). An increase in relative protein expression of Bak and Bcl-Xs was detected in similarly transfected 293 cells relative to cells transfected with empty vector (Fig. 4A). No change in the relative level of expression of Bad, Bax, or Bcl-2 was detected in N5-transfected cells. Nor were their changes in MAD protein expression.
Bak and Bcl-Xs are apoptotic agonists that heterodimerize with Bcl-2 and inhibit its ability to protect cells from cell death (27)(28)(29). The increase in expression of these genes may facilitate p84N5-induced apoptosis. If true, a compensating increase in Bcl-2 expression would be predicted to inhibit p84N5-induced apoptosis. To test this possibility, we have coexpressed Bcl-2 with p84N5 and measured the effect on apoptosis using the TUNEL assay. More than 90% of transfected cells express-2 J. Doostzadeh-Cizeron, unpublished observation.

FIG. 2. AdN5-induced apoptosis involves activation of caspase-6.
A, SAOS-2 were infected with AdN5 or AdGFP at an m.o.i. of 10. At the indicated times after infection, cell extracts were prepared and assayed for caspase activity using the colorimetric peptide substrates shown. Cleavage of the peptide causes an increase in OD405. The data presented show the mean OD405 of assays done in triplicate The data for AdGFP infection are not shown, because none of the extracts demonstrated OD405 readings above background. Only three peptides were cleaved by the AdN5-infected extracts, VEID, DEVD, and LEHD (shown as white symbols). B, equal quantities of total protein from extracts prepared as above were analyzed for caspases-6 or -9 by Western blotting. The caspase-6 antibody recognizes the zymogen and the p18 processed form. The caspase-9 antibody recognizes the zymogen and the p35 processed form. Processed caspase-6 can be readily detected by day 2 following infection, whereas caspase-9 is not readily detected until day 4. ing p84N5 alone contain fragmented DNA. Coexpression of Bcl-2 with p84N5 decreases the percentage of cells containing fragmented DNA to near background levels (Fig. 4B).
N5-induced Apoptosis Does Not Require and Is Not Inhibited by p53-We have demonstrated that expression of p84N5 in p53 null SAOS-2 cells (30) induces apoptosis (Fig. 1A), indicating that p53 was not required for p84N5-induced apoptosis. We wished to explore the possibility that p53 may inhibit p84N5induced apoptosis. To this aim, p84N5 and wild-type p53 were expressed in SAOS-2 cells and the cells assayed for apoptosis by TUNEL. Coexpression of wild-type p53 did not affect the percentage of cells containing fragmented DNA (Fig. 5A). Consistent with our previous results (13), expression of a constitutively active pRb mutant (Rb⌬CDK) decreased the percentage of cells containing fragmented DNA. Expression of p53 and Rb⌬CDK upon transfection of the appropriate expression vectors was confirmed by Western blotting. 2 We also compared the effects of pCMVN5 transfection in wild-type p53 expressing MCF-7 cells or SAOS-2 cells. MCF-7 and SAOS-2 cells were equally sensitive to p84N5-induced apoptosis as indicated by the similar percentage of transfected cells containing fragmented DNA (Fig. 5B). MCF-7 and SAOS-2 cells were also equally sensitive to apoptosis induced by AdN5 (32). Finally, we checked the effect of p84N5 expression on the relative level of endogenous p53, p110Rb, p16, or p27 expression by Western blotting. The expression level of each of these proteins was similar in CMVN5-transfected cells and cells transfected with empty vector (Fig. 5C).
These observations suggested that N5 and p53 function in different apoptotic pathways. If true, the effects of each gene on the inhibition of cell growth would be expected to be at least additive. To test this prediction, the effect of AdN5 infection, Adp53 infection, or a combination of both on the growth of Colo357 X cells was determined. These cells are relatively resistant to the effects of AdN5, 3 allowing any increase in the effect of infection on cell growth to be determined. Both AdN5 infection and Adp53 infection decreased the growth rate of these cells. Treatment with either virus caused the number of cells to decrease by day 6. The kinetics of cell accumulation upon AdN5 or Adp53 infection appeared similar; AdN5 had a slightly greater effect on cell growth than did Adp53. Interestingly, infection with a combination of both viruses had an even greater effect on cell growth (Fig. 6). The number of cells infected with AdN5 and Adp53 at any one time were typically 2-to 4-fold less than the number of cells infected with either AdN5 or Adp53 alone. This difference is likely an underestimate, because the m.o.i. of each virus in the combination infection is half that of the m.o.i. in the single infections. DISCUSSION Our data indicate that forced expression of p84N5, either by transfection or adenovirus-mediated gene transfer, renders cells inviable. Loss of viability is due to induction of apoptotic cell death, because p84N5-expressing cells exhibit many of the typical characteristics of this process, including fragmentation of DNA, exposure of phosphatidylserine on the outer leaflet of the plasma membrane, changes in membrane permeability, and changes in cellular and nuclear morphology. The p84N5 protein has significant sequence similarity to the death domains of other proteins involved in apoptosis (31), and characteristic point mutations known to inactivate other death do-3 S. Yin, unpublished observation.

FIG. 3. Activation of NF-B during p84N5-induced apoptosis.
A, 293 cells were transfected with pCMV or pCMVN5. Cytoplasmic or nuclear extracts were prepared from these cells 24 h later. Equal quantities of total protein were analyzed by Western blotting using an anti-p65RelA antibody. Staining with anti-␤-actin served as a protein loading control. B, 293 cells were cotransfected with a luciferase reporter gene construct whose expression was driven by a wild-type or mutant NF-B-dependent promoter, pEGFPC-1, and either pCMVN5, a p65RelA expression vector, or the empty vector (CMV). Luciferase activity was determined 24 h after transfection and normalized on the basis of transfection efficiency as determined by the percentage of GFP fluorescent cells. Values shown are the mean of the relative light units for three experiments. main proteins also compromise p84N5-induced cell death (13). Interestingly, p84N5 is unique among death domain-containing proteins in its nuclear localization (14) and nuclear localization is required for p84N5-induced cell death. 4 There is a relatively small number of proteins that are known to initiate apoptosis from within the nucleus, including the expanded polyglutamine repeat proteins and possibly transcription factors like p53. To test whether the apoptotic pathway triggered by p84N5 expression is similar to that triggered by these other nuclear proteins, we have examined the activation of NF-B, changes in Bcl-2 family protein expression, the effects of p53, and caspase activation during p84N5-induced apoptosis.
NF-B activation during p84N5-induced apoptosis is indicated by the change in localization of p65RelA from the cytoplasm to the nucleus in AdN5-infected cells but not in AdGFPinfected cells. The change in subcellular localization is accompanied by an increase in NF-B-dependent transcriptional activation as determined by increased levels of a luciferase reporter gene whose transcription is dependent on NF-B transactivation. In contrast to p84N5, expression of p53 inhibits NF-B activation (33,34). Activation of NF-B is observed during apoptosis initiated by DNA-damaging agents like ionizing radiation (35,36), among others (37,38). Although similar mechanisms may be utilized to activate NF-B, p84N5induced apoptosis is unlikely to be mediated by NF-Bmediated Fas ligand expression, and CD95 ligation as has been proposed for some DNA-damaging agents (37). Fas-mediated apoptosis requires activation of caspase-8 and is CrmA-sensitive, whereas p84N5-induced apoptosis does not require caspase-8 activation and is insensitive to Crm A (see below).
Our results also demonstrate that apoptosis induced by p84N5 is accompanied by an increase in the relative expression of pro-apoptotic Bcl-2 family members Bak and Bcl-Xs. The expression of Bcl-2, Bax, and Bad are unchanged during p84N5-induced apoptosis. In contrast, p53-induced apoptosis involves an increase in the relative expression of Bax (39). Alteration in the levels of Bcl-2 family protein expression can change mitochondrial membrane permeability allowing release of cytochrome c. Cytoplasmic cytochrome c participates in a protein complex composed of caspase-9 and Apaf-1 that leads to caspase activation (40). For example, Bak can accelerate release of cytochrome c through mitochondrial porin channels (41). Bak can also disrupt association of Apaf-1 and the anti-apoptotic Bcl-2 family protein Boo (42). Hence, p84N5-induced apoptosis may be facilitated by the increase in Bak expression observed. Consistent with this hypothesis, we demonstrate that coexpression of Bcl-2 can inhibit p84N5-induced apoptosis. Furthermore, like p84N5-induced apoptosis, apoptosis induced by forced expression of Bak is insensitive to CrmA (43). However, because p84N5-induced apoptosis is initiated by caspase-6 activation rather than caspase-9, Bak either facilitates apoptosis by a different mechanism or it participates in the later execution stages of the process. Whether p84N5 can directly influence transcription is unknown. However, NF-B can regulate the transcription of some Bcl-2 family genes (44,45), so the effect of p84N5 on the expression of Bcl-2 family proteins may be indirect.
Our data indicate that p84N5-induced apoptosis does not require p53. Cells with wild-type p53 (MCF-7) or null for p53 (SAOS-2) are equally sensitive to p84N5-induced cell death. Furthermore, we show that p84N5-induced apoptosis does not alter the level of p53 expression and that coexpression of p53 does not inhibit p84N5-induced apoptosis. For a number of reasons we believe it likely that N5 and p53 function in different apoptotic pathways. As described above, p53 and p84N5 have different effects on NF-B activation and Bcl-2 family protein expression. Furthermore, p53-induced apoptosis is dependent on caspase-3 and/or caspase-9 in many cases (46 -48). These caspases are not required for p84N5-induced apoptosis. Finally, we show that infection with both AdN5 and Adp53 causes a more dramatic reduction of Colo357 X cell growth than infection with either one alone. Genes that participate in or activate the same apoptotic pathway would not be expected to have such an additive or synergistic effect. Consistent with our hypothesis that N5 and p53 function in different pathways, N5-induced apoptosis is preceded by a G 2 /M cell cycle arrest while expression of p53 typically induces a G1 cell cycle arrest. 2 The pattern of caspase activation during p84N5-induced 4 R. Evans and D. W. Goodrich, unpublished observation . apoptosis is atypical. Caspase-6 is the first caspase activated, and it makes up the majority of caspase activity observed in extracts of AdN5-infected cells. This conclusion is based upon cleavage of a peptide substrate, VEID, that is preferred by caspase-6 as well as by detection of proteolytically processed caspase-6. The pattern of caspase activation is unlikely to be influenced by adenoviral infection, because caspases are not activated in AdGFP-infected cells. Because activation of caspase-6 precedes or is coincident with DNA fragmentation, annexin V staining, and loss of cell viability, we conclude that it may mediate p84N5-induced apoptosis. Activation of caspase-9 and caspase-3 activity is also detected in AdN5infected cells. However, activation of these caspases is unlikely to mediate p84N5-induced apoptosis. First, activation of caspase-9 or -3 occurs after execution of apoptosis has already begun. The activation of a caspase-3, for example, is not detected until 4 days after infection, well after the time when DNA fragmentation can be detected. This observation suggests activation of caspases-9 and -3 is a consequence of earlier caspase-6 activation. Second, MCF-7 cells, which lack caspase-3 (49), are sensitive to AdN5. Hence, caspase-3 is not required for p84N5-induced apoptosis. This pattern of caspase activation suggests that p84N5 triggers an apoptotic pathway distinct from those triggered by expanded polyglutamine repeat proteins or death domain-containing receptors. Apoptosis induced by these proteins require caspases-3, -8, or -9, whereas p84N5-induced apoptosis involves caspase-6.
Short prodomain caspases are presumed responsible for the execution phase of apoptosis and typically rely on long prodomain caspases for initial proteolytic activation. Caspase-6 is a short prodomain caspase yet is activated before activation of any other caspase, including long prodomain caspases-8 or -9, during p84N5-induced apoptosis. The activation of caspase-6 is not likely to depend on caspase-9, because caspase-6 is a poor substrate for caspase-9 in vitro (50). Caspase-8 is also unlikely to be involved, because an efficient caspase-8 peptide substrate is not cleaved by extracts of AdN5-treated cells and because p84N5induced apoptosis is insensitive to CrmA (13), a potent inhibitor of caspase-8. Finally, caspase-3 is unlikely to be involved, because it is not required for p84N5-induced apoptosis. Caspase-6 is either activated by a protease whose activity has not been detected in our experiments or is activated by a novel mechanism during p84N5-induced apoptosis. By analogy to other death domain-containing proteins, p84N5 may play an important part in caspase-6 activation.
In this study, we have characterized the apoptotic pathway triggered by the expression of the nuclear death domain-containing p84N5. Our data indicate that p84N5-induced apoptosis is characterized by activation of caspase-6, increases in the expression of Bcl-2 pro-apoptotic family members Bak and Bcl-Xs, and activation of NF-kB. Furthermore, this pathway does not require p53, nor does the presence of p53 affect the efficiency of cell killing. Expression of Rb, however, does inhibit p84N5-induced apoptosis, presumably mediated by physical association between the two proteins (13,14). These observations, coupled with the fact that p84N5 is a death domaincontaining protein localized within the nucleus, suggest that p84N5 participates in an apoptotic pathway that is different from those triggered by p53, death domain-containing receptors, or expanded polyglutamine repeat proteins.