DNA-dependent protein kinase is a target for a CPP32-like apoptotic protease.

We demonstrate that the catalytic subunit of the DNA-dependent protein kinase (DNA-PKcs) is specifically, proteolytically cleaved in HL-60 cells treated with staurosporine (STS), a potent inducer of apoptosis. The proteolysis of DNA-PKcs correlated with or preceded apoptotic chromosomal DNA degradation. Cell-free extracts prepared from STS-treated HL-60 cells recapitulated the proteolysis of DNA-PKcs in an in vitro assay using purified DNA-PK as the substrate. Western blot analyses of the apoptotic cell extract showed that the 32-kDa precursor of CPP32 is expressed in HL-60 cells and processed following STS treatment. In addition, whereas the DNA-PKcs protease activity was not inhibitable by many conventional protease inhibitors, it was inhibitable by a highly selective peptide-derived inhibitor of CPP32. These data strongly suggest that CPP32, or a CPP32-like protease, is responsible for DNA-PKcs proteolysis. Finally, our results demonstrated that the cleavage of DNA-PKcs in vitro proceeded in the presence of Bcl-2, indicating that the function provided by Bcl-2 lies upstream the proteolysis of DNA-PKcs.

Programmed cell death is of fundamental importance for the proper differentiation and development of multicellular organisms (reviewed in Refs. 1 and 2). In addition, dysregulated apoptosis appears to play a causative role in many human pathological states, including cancer, AIDS, and Alzheimer's disease (reviewed in Ref. 3). The regulation of apoptosis is thought to be mediated by the interplay of various apoptotic suppressor and inducer molecules (for recent reviews, see Refs. 4 and 5). In particular, one gene that is known to play an important role in suppressing apoptosis is bcl-2. bcl-2 encodes a strong apoptotic suppressor (reviewed in Ref. 5), and bcl-2 knockout mice suffer massive apoptotic cell deaths in the lymphoid system, confirming that the expression of this gene is biologically important (6).
The mechanism by which apoptotic suppressors such as Bcl-2 exert their biological function(s) is unknown, but it is likely that they may regulate one or more proteolytic processing events mediated by the interleukin-1␤ converting enzyme (ICE) 1 family of proteases. The identification of ICE as a putative apoptotic protease (7) was followed by the isolation of a large number of homologues of ICE, including Ich-1/Nedd2, CPP32, Ich-2, and Mch-2 (for reviews, see Refs. 8 and 9). All the ICE homologues are cytoplasmic cysteine proteases, and it has been shown that their overexpression induces apoptosis (reviewed in Refs. 2 and 10). Interestingly, ICE family members are initially synthesized as inactive precursors and require proteolytic processing to generate the active protease, suggesting that the apoptotic machinery is regulated by a protease cascade.
Several targets of the cytoplasmic apoptotic proteases that have been identified so far are, surprisingly, nuclear proteins including but not restricted to lamin B1 (11,12), topoisomerase I (12), the 70-kDa protein component of the U1 small nuclear ribonucleoprotein (13), and the DNA repair enzyme poly(ADPribose) polymerase (PARP) (14,15), although a cytoplasmic protein, fodrin, has also been shown to be proteolytically cleaved during apoptosis (16). With the exception of PARP, which has been shown to be a target of CPP32 (17,18), the proteases that degrade the other proteins remain to be identified. In addition, it is unclear how many more proteolytic target proteins there may be that are specifically related to apoptosis. Therefore, the identification of other biologically relevant target proteins and their corresponding apoptotic proteases is one of the critical areas of future investigation.
One of the terminal steps of apoptosis is nucleolytic degradation of chromosomes into double-stranded, nucleosome-sized fragments. A complex, which has DNA-dependent serine-threonine protein kinase activity (DNA-PK) and which consists of at least three components, the 460-kDa catalytic subunit (DNA-PK cs ) and the heterodimeric Ku protein, has been shown to be intimately involved in DNA double-strand break (DSB) repair (reviewed in Refs. 19 -21). DNA-PK cs is thought to be the product of the severe combined immune deficiency (scid) gene, and it has long been known that animals homozygously mutated at this locus were defective in DNA DSB repair (22,23). Ku is a heterodimeric protein of 70-and 86-kDa subunits, which binds tightly to a variety of double-stranded DNA ends (reviewed in Refs. 20 and 21). Recently, it was shown that members of the fifth x-ray cross-complementation group (XRCC5), which are DNA DSB repair-defective and lack Ku DNA end-binding activity, are mutated in the Ku86 gene (24).
Thus, DNA-PK has been unequivocally identified as an important mammalian DNA repair complex, and mutations in either the DNA-PK cs or the 86-kDa subunit of Ku result in DNA DSB repair defects.
In this study, we demonstrate that DNA-PK cs is selectively proteolytically cleaved in the human HL-60 promyelocytic leukemia cell line following apoptotic induction. In addition, we demonstrate that this proteolytic step can be blocked in vivo by the apoptotic protease inhibitors, iodoacetamide and N-ethylmaleimide. Recapitulation of DNA-PK cs proteolysis was achieved in vitro, and inhibitors highly selective for the CPP32 apoptotic protease prevented cleavage. These studies demonstrate that DNA-PK is an apoptotic target and provide important insights into the terminal steps of apoptotic induction.
Electrophoretic Analysis of Apoptotic DNA Degradation-The isolation of degraded DNA and the electrophoretic analysis of DNA degradation during apoptosis have been described (30,32).
Purification of DNA-PK-DNA-PK was purified from HeLa cell nuclear extracts as described (26), except that the Sepharose column step was replaced by Superose fast protein liquid chromatography (Superdex 200, Prep grade, 16 mm ϫ 60 cm, Pharmacia Biotech Inc.) as the third and final step.
Western Blot Analysis of DNA-PK Proteolysis during Apoptosis-Cells were treated with 5 M STS for various lengths of time. At the end of each treatment, cells were washed twice in phosphate-buffered saline. Cells were then resuspended in an aliquot of 20 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , and immediately heated at 100°C for 10 min. DNase I (10 g/10 7 cells) was added to the lysate, and the sample was incubated at 37°C for 5 min. The sample was then mixed with an equal volume of a 2 ϫ urea/glycerol/SDS solution (7 M urea, 30% glycerol, 3% SDS), and heated at 100°C for 10 min. Protein concentration of the sample was subsequently determined using detergent-compatible protein assay reagents (Bio-Rad). The protein extracts were subjected to electrophoresis in a SDS-polyacrylamide gel (6% for DNA-PK cs , 10% for Ku, and Ich-1 L , and 12% for Bcl-2, Bax, Bak, and CPP32 analyses). The proteins were transferred onto nitrocellulose paper by electrophoresis in a Trans-Blot chamber (Bio-Rad), and proteins were identified using antibodies and ECL-Western reagents (Amersham Corp.).
Preparation of Cell Extracts Containing DNA-PK cs Protease Activity-HL-60 cells (5 ϫ 10 8 ) were either left untreated or were exposed to 5 M STS for 2 h. Cells were then washed in 50 ml of phosphate-buffered saline three times and resuspended in extraction buffer (10 mM HEPES, pH 7.5; 250 mM NaCl; 5 mM MgCl 2 ; 5 mM CaCl 2 ; 1 mM dithiothreitol; 0.5 mM PMSF; 50 g/ml leupeptin, aprotinin, and antipain; 20% glycerol). The cell suspension was subjected to four or five cycles of freezing and thawing in a isopropyl alcohol/dry ice bath until Ͼ95% of the cells were lysed as assessed by microscopic observation. The lysate was centrifuged at 100,000 ϫ g for 2 h at 4°C. The 100,000 ϫ g supernatant was subsequently centrifuged at 150,000 ϫ g for 1 h at 4°C to remove any remaining insoluble material. The clear supernatant was kept at Ϫ85°C and assayed for the presence of DNA-PK cs protease activity.
Proteolysis of DNA-PK cs in Vitro-Purified DNA-PK (50 -200 ng) was incubated with various amounts of cell extract in a final volume of 15 l and a defined buffer condition (10 mM HEPES, pH 7.5; 50 mM NaCl; 2 mM MgCl 2 ; 2 mM CaCl; 50 g/ml leupeptin, aprotinin, and antipain; 0.5 mM PMSF; 1 mM dithiothreitol; 5% glycerol) for various times at 37°C as indicated in the text (see Fig. 4). When protease inhibitors were studied, the extract was first incubated with an inhibitor for 15 min at room temperature and then mixed with DNA-PK and incubated at 37°C for 45 min. At the end of the incubation, 10 l of SDS-glycerol (5% SDS and 50% glycerol) was added to the sample, and the reaction was terminated by heating the sample at 100°C for 5 min. Proteolysis of DNA-PK cs was subsequently analyzed by electrophoresis and Western blot analysis.

Apoptotic Responses of HL-60 and a Clonal Variant, HCW-2-
We have recently demonstrated (30,31) and others have also shown (33)(34)(35)(36) that the human promyelocytic leukemic HL-60 cell line is hypersensitive to a wide variety of apoptotic inducers including differentiation agents, protein synthesis inhibitors, protein kinase inhibitors, and DNA damage. In addition, we have described the isolation of a stable clonal variant of the HL-60 cell line, HCW-2, which is resistant to many of these apoptotic inducers (30,31). Since one of the terminal steps of apoptosis is the induction of chromosomal DNA DSBs and much, if not all, of mammalian DNA DSB repair is mediated by the DNA-PK complex (reviewed in Refs. 20 and 21), we investigated the status of this activity in HL-60 and HCW-2 cells. These results are presented below.
DNA-PK cs Is Proteolytically Cleaved following Apoptotic Induction by Staurosporine-The protein kinase inhibitor, STS, is an extremely potent inducer of apoptosis in HL-60 cells (31,36). Therefore, HL-60 and HCW-2 cells were either left untreated or were exposed to 5 M STS for 5 h, and then whole cell extracts were prepared and subjected to Western blot analysis. When a monoclonal antibody Ref. 26) that recognizes DNA-PK cs was used, a significant fraction of the DNA-PK cs in HL-60 cells was detected that exhibited a greatly increased gel mobility (Fig. 1A). The rapidly migrating protein fragment (hereafter termed ⌬DNA-PK cs ) migrated as an approximately 150-kDa polypeptide instead of the intact 460-kDa protein.
Two lines of evidence suggested that the presumed proteolysis of DNA-PK cs was specific and related to apoptosis. First, no cleavage of Ku86 or Ku70 proteins was detected when either cell line was exposed to STS (Fig. 1B). Secondly, no cleavage of DNA-PK cs was detected following STS treatment of the apoptotic-resistant HCW-2 cell line (Fig. 1A). From these experiments we conclude that DNA-PK cs is specifically cleaved in HL-60 cells undergoing apoptosis.
Proteolysis of DNA-PK cs Correlates with the Onset of Apoptosis-To investigate the cleavage of DNA-PK cs in more detail we next carried out a time course experiment. HL-60 cells were exposed to 5 M STS, and at hourly intervals samples were analyzed for the appearance of ⌬DNA-PK cs and degradation of DNA into nucleosome-sized fragments, which is a hallmark of cells undergoing apoptosis (1). ⌬DNA-PK cs could be detected as early as 1 h after STS treatment but increased dramatically at 2 h (Fig. 2, top). Interestingly, detectable DNA fragmentation was observed in HL-60 cells only after 2 h post-STS exposure, and this increased at later times (Fig. 2, bottom). From this experiment we conclude that the appearance of ⌬DNA-PK cs either coincides with or precedes the apoptotic DNA degradation process in HL-60 cells following STS treatment.
Cysteine-alkylating Agents Block the Proteolysis of DNA-PK cs -It has been demonstrated that the cysteine-alkylating agents IA and NEM are potent inhibitors of apoptotic proteolysis (13,14,17). This inhibitory profile is consistent with the apoptotic cleavage enzymes being members of the ICE-like family of cysteine proteases (for a recent review, see Ref. 9).
Importantly, most serine-, metallo-, and aspartate-protease inhibitors and E-64, which can inhibit certain cysteine proteases, in general do not appear to affect apoptotic proteolytic cleavages (13,17). Thus, we co-treated HL-60 cells with 5 M STS and a variety of protease inhibitors for 5 h. Cellular extracts were then prepared and subjected to Western analysis. Impressively, IA and NEM completely blocked the appearance of ⌬DNA-PK cs , whereas none of the other inhibitors had any effect on the proteolysis (Fig. 3). This result is consistent with the hypothesis that the appearance of ⌬DNA-PK cs is most likely attributable to an apoptotic cysteine protease.
In Vitro Recapitulation of STS-induced DNA-PK cs Proteolysis-To study the process of STS-induced proteolysis of DNA-PK cs , cytosolic extract was prepared from HL-60 cells, which were treated with 5 M STS for 2 h, as this appeared to be a minimum treatment required for the induction of proteolysis of DNA-PK cs (see Fig. 2). The presence of a DNA-PK cs protease in the extract was investigated in vitro using purified DNA-PK as the substrate. Extract prepared from STS-treated cells (hereafter referred to as "activated extract") already contained a basal level of the cleaved endogenous ⌬DNA-PK cs (Fig. 4A, lane  2). However, this did not obscure the striking degradation of exogenous DNA-PK in the extract. Typically, incubation of an additional 100 ng of purified DNA-PK with 40 g of activated extract for 45-60 min induced complete proteolysis of the purified DNA-PK cs and resulted in a cleaved product with an approximate molecular mass of 150 kDa (Fig. 4A, lanes 3-8). In contrast, incubation of purified DNA-PK with the same amount of extract prepared from untreated HL-60 cells had no proteolytic effect on DNA-PK cs (Fig. 4A, lane 1), nor was proteolysis observed upon incubation of purified DNA-PK with buffer alone for 90 min (Fig. 4A, lane 9). These results suggested that the proteolysis of DNA-PK cs during apoptotic induction of HL-60 cells could be faithfully reconstituted in vitro.
In vivo, DNA-PK cs , but not the Ku subunits, appeared to be an apoptotic proteolytic target (see Fig. 1). To assess whether this specificity was maintained in our in vitro system we assayed the stability of Ku86 and Ku70 by Western blot analysis following incubation of 40 g of activated extract with 100 ng of purified DNA-PK at 37°C for 45 min. No evidence of proteolysis of either Ku subunit was observed (Fig. 4B). Thus, the protease activated by STS treatment of HL-60 cells showed the same specificity on exogenous DNA-PK in vitro as had been observed for endogenous DNA-PK in vivo.
In Vitro Proteolysis of DNA-PK cs Is Inhibited by IA and NEM-Since the in vitro proteolysis assay buffer contained 0.5 mM PMSF and 50 g/ml of leupeptin, aprotinin, and antipain, we inferred that proteases sensitive to these inhibitors were not responsible for cleaving DNA-PK cs . Therefore, the effect of , and then whole cell extracts were prepared. 50 g of each extract was subsequently subjected to gel electrophoresis on either a 6% (for DNA-PK cs ) or a 10% (for Ku70 and Ku86) polyacrylamide gel and then subjected to Western analyses. A, specific proteolysis of DNA-PK cs in cells undergoing apoptosis. DNA-PK cs was identified using monoclonal antibody 42-27. The position of the predominant cleavage product is designated as ⌬DNA-PK cs . B, proteolysis of the Ku subunits does not occur in cells undergoing apoptosis. Ku86 and Ku70 were identified using a mixture of monoclonal antibodies 111 and N3H10, respectively. Only the relevant portion of each Western blot is shown.

FIG. 2. Proteolysis of DNA-PK cs correlates with the onset of STS-induced apoptosis in HL-60 cells.
HL-60 cells were either left untreated (0 h) or were treated for up to 5 h with 5 M STS. At hourly intervals (1, 2, 3, 4, or 5 h), samples of cells were removed and divided into two aliquots. One was used to make whole cell extracts and subjected to a Western analysis, while the other was processed for quantitation of DNA fragmentation. Top, proteolysis of DNA-PK cs occurs after apoptotic induction by STS. ⌬DNA-PK cs was identified by Western analysis using monoclonal antibody 42-27. Only the portion of the blot corresponding to the cleavage product is shown for the sake of presentation. Bottom, significant internucleosomal DNA fragmentation in HL-60 cells occurs after apoptotic induction by STS. Degraded chromosomal DNA was isolated from cells, subjected to electrophoresis in a 2% agarose gel, and then stained with ethidium bromide.  (13,17). Whole cell extracts were subsequently prepared and analyzed by Western analysis for the appearance of ⌬DNA-PK cs using monoclonal antibody 42-27. other protease inhibitors on the DNA-PK cs proteolysis was investigated. The results showed that among the conventional inhibitors tested, only IA and NEM were able to block the DNA-PK cs protease activity in vitro (Fig. 5). This result recapitulated our in vivo observations (for comparison, see Fig. 3) and again suggested that the DNA-PK cs protease is a cysteine protease.
The Expression of CPP32 Correlates with DNA-PK cs Proteolysis-A partial composition of the activated extract was determined to help identify potential critical apoptotic components. Western blot analysis of the cell extracts demonstrated that the 32-kDa precursor peptide of the CPP32 apoptotic protease was expressed in HL-60 cells under normal (nonactivated) conditions (Fig. 6A, lane 1). Upon activation, the amount of precursor CPP32 was reduced (Fig. 6A, lane 2). This most likely represents the conversion of this precursor protein into the two smaller subunits that make up the active CPP32 protease (17,37,38). Unfortunately, our antibody, the only commercial CPP32 monoclonal antibody available, does not recognize either of the processed subunits, so we could not demonstrate this directly. Nevertheless, this result suggested that CPP32 might catalyze the proteolysis of DNA-PK cs . This hypothesis was supported by the observation that the level of another apoptotic protease, Ich-1 L (39), which also requires processing for activation, was not altered by STS treatment (Fig. 6B,  compare lanes 1 and 2) and by the demonstration, using Northern blot analysis, that HL-60 cells do not even express ICE (data not shown). Thus, the expression pattern of CPP32 was most consistent with it being responsible for DNA-PK cs cleavage.
A Peptide Inhibitor Highly Selective for CPP32 Blocks DNA-PK cs Protease Activity-To extend the above results, the effect of Ac-DEVD-CHO peptide, a highly selective inhibitor of CPP32 (17), on DNA-PK cs cleavage in vitro was investigated. The results showed that 1 M Ac-DEVD-CHO completely blocked proteolysis of DNA-PK cs by the activated extract (Fig.  7). As a control, the effect of Ac-YVAD-CHO peptide, a highly selective inhibitor of the ICE protease (25,40), was examined, and as expected, the presence of up to 50 M Ac-YVAD-CHO had no effect on the proteolysis of DNA-PK cs (Fig. 7). These results, combined with those from Western analyses (Fig. 6) strongly suggested that CPP32, or a CPP32-like enzyme, is most likely the DNA-PK cs protease.
In Vitro Proteolysis of DNA-PK cs Occurs in the Presence of Bcl-2-The expression of apoptotic suppressors and effectors is known to profoundly affect the susceptibility of cells to apoptotic signals. Thus, we examined by Western blot analyses the levels of apoptotic suppressor and effector proteins in our in vitro DNA-PK proteolysis assay samples. The activated extract contained the apoptosis suppressor protein Bcl-2, and the amount of Bcl-2 remained unchanged (Fig. 8B) at a time when  the majority of DNA-PK cs was cleaved (Fig. 8A). Interestingly, the activated extracts did not contain detectable levels of the cell death effectors Bax and Bak (Fig. 8, C and D), although these proteins are expressed in HL-60 cells (Refs. 30, 31, and 49, and data not shown). From these results we conclude two things. First, Bcl-2 in the in vitro assay is not itself an apoptotic protease target, and secondly, and most importantly, neither the presence of Bcl-2 nor the absence of Bax and Bak in the extract prevented the proteolysis of DNA-PK cs by the activated CPP32-like activity. DISCUSSION In this study, we have shown that the catalytic subunit of the DNA-PK complex is a proteolytic target in HL-60 cells undergoing apoptosis. In vivo inhibitor studies suggested that the DNA-PK cs protease is a member of the ICE family of apoptotic proteases. Reconstitution of the DNA-PK cs proteolysis reaction in vitro allowed the identification of the ICE homologue, CPP32, or a CPP32-like enzyme, as the protease most likely responsible for the DNA-PK cs proteolysis. The significance of these results is discussed below.
DNA-PK cs Is a Target for an Apoptotic Protease-The observation that proteolysis of DNA-PK cs was specific to cells undergoing apoptosis and that this cleavage could be abrogated by treating cells with the protease inhibitors IA and NEM in vivo (Fig. 3) and with the CPP32-selective tetrapeptide inhibitor Ac-DEVD-CHO in vitro (Fig. 7) demonstrated that DNA-PK cs is an apoptotic target. These data confirm and significantly extend previous studies that had suggested that the DNA-PK complex was altered in cells undergoing apoptosis. Ajmani et al. (41) initially demonstrated that DNA-PK cs protein levels were greatly diminished in mature human neutrophils destined to undergo apoptosis and in apoptotic HL-60 cells. We have shown here that the reduced DNA-PK cs protein levels are probably due directly to proteolysis (Fig. 1A). Second, Casciola-Rosen et al. (42) recently reported that DNA-PK cs was proteolytically cleaved in apoptotic HeLa cells after ultraviolet light exposure. Our studies, utilizing a different cell type and a different apoptotic inducer, extend the observations of Rosen et al. and suggest that the proteolysis of DNA-PK cs may be a common event in many cells undergoing apoptosis.
The biological implication of the above results is that proteolysis of DNA-PK cs facilitates apoptosis. This seems likely. Proteolytic cleavage of DNA-PK cs almost certainly inactivates DNA-PK activity. Work from a large number of laboratories has shown that mutations that deleteriously affect DNA-PK result in greatly diminished DNA DSB repair activity (reviewed in Refs. 19 -21). Since one of the terminal steps of apoptosis is the activation of nucleases that induce DSBs in chromosomal DNA, it clearly would be counterproductive for the cell to simultaneously try to repair the damaged chromosomes. Thus, apoptosis should be facilitated by inactivating the major DNA DSB repair system. A correlate of this would be that apoptosis should be inhibited or abrogated by blocking the proteolysis of DNA-PK cs . Thus, DNA-PK cs may prove to be an important therapeutic target for antiapoptotic drugs.
CPP32 Is Likely Responsible for DNA-PK cs Proteolysis-The second important observation that this study provides is that the apoptotic protease, CPP32, is most likely responsible for DNA-PK cs proteolysis. While the identification of apoptotic target proteins (for a review, see Ref. 8) and apoptotic proteases (for a review, see Ref. 9) is rapidly progressing, the identification of which proteases are responsible for cleaving which target proteins has lagged behind. Recently, several studies have shown that CPP32 is also the apoptotic protease responsible for the proteolysis of PARP (15,17,18). Thus, CPP32 is currently the most biologically relevant mammalian apoptotic protease. This is not to suggest that the other ICE family homologues are irrelevant to apoptosis, but rather that their substrates, tissue specificity, and/or developmental timing may not yet have been identified.
ICE family proteases such as CPP32 have an absolute requirement for aspartic acid at the cleavage site, and this residue is frequently adjacent to a small hydrophobic residue (D-2-X). For example, ICE cleaves the interleukin-1␤ precursor at FEAD-2-G and YVHD-2-A sites (40), and CPP32 cleaves PARP at a DEVD-2-G site (15,17). Interestingly, however, in vitro ICE does not seem to cleave PARP at this site (15). Therefore, each ICE family member protease may recognize and cleave its target(s) using a very specific amino acid sequence. Based on the size of ⌬DNA-PK cs (ϳ150 kDa) the putative proteolytic cleavage site in DNA-PK cs can be approximated, although there are many DX motifs in the DNA-PK cs coding sequence (43,44) including two PARP-like motifs at amino acids 2979 -2983 (DWVDG) and amino acids 2709 -2713 (DEVDN). Cleavage at the DWVD-2-G or DEVD-2-N site would generate C-terminal polypeptides of ϳ125 and ϳ152 kDa, respectively. Since the latter fragment size is very close to the estimated size (ϳ150 kDa) of ⌬DNA-PK cs seen in HL-60 cells undergoing apoptosis it is tempting to speculate that DEVD-2-N may be the CPP32 cleavage site in DNA-PK cs .
Role of Apoptotic Suppressors and Effectors in STS-induced Apoptosis in HL-60 Cells-Bcl-2 can suppress apoptotic induction by many stimuli in a variety of cell lines (reviewed in Ref. 45). Conversely, expression of Bax (46,47) and Bak (48 -50) can enhance apoptosis. Unfortunately, the biochemical mechanism(s) by which these apoptotic suppressors and effectors regulate apoptosis remains unknown. The only protein involved in apoptosis for which a mechanism of action is known is the baculoviral apoptotic suppressor protein, p35, which acts as a competitive inhibitor of the ICE family proteases (51,52). In our cell-free system, the proteolysis of DNA-PK cs took place in the presence of significant amounts of Bcl-2 protein (Fig. 8, A  and B). These results suggest that the apoptotic suppressor protein Bcl-2 does not (or could not) directly inhibit the activated CPP32-like protease and thus is likely to be functioning in a manner distinct from p35. Last, while HL-60 cells express Bax (30,31) and Bak (Ref. 49, data not shown), our cell-free system lacked detectable amounts of these proteins (Fig. 8, C and D, respectively) and yet was clearly very apoptotic-competent. Thus, Bax and Bak are not directly involved in the proteolysis of DNA-PK cs . A simple interpretation of these data is that Bcl-2, Bax, and Bak are not needed for the proximal events of apoptosis but instead are more likely to be involved in upstream events such as apoptotic signal transduction.
Development of an in Vitro System for Studying Apoptosis-A number of recent reports have suggested that cell-free systems can at least carry out the proteolysis of the apoptotic target, PARP (15,17). We have extensively characterized the in vivo response of HL-60 and HCW-2 cells to a variety of apoptotic inducers (30,31). In this report, we demonstrate that many of our previous observations, including, for example, the stability of Bcl-2 during apoptosis (31), can be recapitulated in vitro. In addition, we also demonstrate that the proteolysis of DNA-PK cs that occurs in vivo (Figs. 1 and 2) can be reproduced in vitro (Figs. 4,5,and 8). Furthermore, using isolated nuclei from nonactivated HL-60 cells rather than purified DNA-PK as a substrate, we have recently shown that we can still detect specific proteolysis of DNA-PK cs (data not shown). Thus, these results suggest that our activated cell-free extracts faithfully mimic many of the events that occur in vivo. Further purification and characterization of these extracts should allow for a more complete biochemical description of the process of apoptosis.