INTRODUCTION

Apoptosis is a process of physiological cell death characterized by a series of stereotyped morphological alterations and endogenous endonuclease activation leading to the generation of large (50 kb¹) and oligonucleosomal DNA fragments (1, 2). Although inhibitors of macromolecular synthesis prevent apoptosis in many circumstances, it is currently thought that the essential enzymatic components of the effector pathway are constitutively expressed in all mammalian cells beyond the blastomere stage of embryonic development (3). The identities of these cell death mediators are currently not known, but a major recent breakthrough toward their identification has come from work on the genes that regulate programmed cell death in the nematode Caenorhabditis elegans (4). This work has revealed that the C. elegans death promoter ced-3 is structurally homologous to a growing family of proteins related to human interleukin 1-converting enzyme (ICE) (5), a cysteine protease. Subsequent research has shown that specific viral and peptide-based inhibitors of the ICE family block apoptotic cell death in many mammalian models (4, 5, 6, 7, 8, 9), indicating that ICE and its homologs play crucial roles in the effector phase of apoptosis.

Recent work has shown that the nuclear lamins are cleaved by one or more members of the ICE family during apoptosis. Early work by Kaufmann (10) demonstrated that the oligonucleosomal DNA fragmentation induced by cancer chemotherapeutic agents in a human myelogenous leukemia line (HL-60) was associated with the degradation of lamin B₁ (and several other nuclear polypeptides). Subsequent work has shown that lamin degradation is associated with DNA fragmentation and cell death in several other models of apoptosis (11, 12, 13, 14) and with chromatin condensation and endonuclease activation in isolated nuclei incubated with apoptosis-promoting cell extracts (14, 15) or Ca²⁺ (11). The identities of the lamin proteases are not known, although preliminary evidence suggests that an ICE family member may be directly involved (15), and a Ca²⁺-dependent serine protease exhibiting ``chymotrypsin-like'' activity that is associated with the nuclear scaffold (NS) has been implicated in another (16). A direct role for lamin degradation in chromatin condensation and DNA fragmentation is attractive, as lamin solubilization is associated with chromatin condensation in both apoptosis and mitosis, the spacing of lamin binding sites in chromatin (periodicity of approximately 20-50 kb) mirrors the sizes of large DNA fragments produced at the commitment phase of apoptosis (17, 18, 19, 20, 21), and the kinetics of lamin degradation and 50-kb DNA fragment formation overlap (11).

Much of our previous work has been concerned with identifying the biochemical mechanisms that regulate endogenous endonuclease activation and cell death in immature thymocytes undergoing apoptosis in response to a variety of stimuli. We have shown that early, sustained increases in the cytosolic Ca²⁺ concentration are commonly involved, and Ca²⁺ buffering agents can interfere with endonuclease activation and cell death (22). It is possible that these Ca²⁺ alterations exert direct effects on components of the apoptotic effector machinery resident in the nucleus, as a number of investigators have demonstrated that typical morphological alterations and chromatin cleavage can be induced in isolated nuclei from untreated cells upon incubation in the presence of exogenous Ca²⁺ (23, 24, 25, 26, 27). However, to date the proteases responsible for Ca²⁺-dependent lamin cleavage in apoptotic cells and isolated nuclei have not been characterized, nor has the possible interrelationship between protease activation and DNA fragmentation been defined. To address these issues was the aim of the present study.

EXPERIMENTAL PROCEDURES

Thymocyte suspensions prepared from 4-8-week-old BALB/c mice were maintained in RPMI 1640 medium supplemented with 10% fetal calf serum, antibiotics, and 50 µM 2-mercaptoethanol. The T-T hybridoma 2B4.11 was maintained in the RPMI 1640 medium described. Transfectants were generated by electroporation at 100 V and 800 microfarads using 1 × 10⁷ 2B4.11 cells and 10 µg of a linearized plasmid containing human bcl-2 under the control of the SV40 promoter/enhancer (Bil2-Neo; generously provided by Dr. Timothy McDonnell, The University of Texas M. D. Anderson Cancer Center). Cells were then selected for 2 weeks in the presence of 0.4 mg/ml G418 (Geneticin, Life Technologies, Inc.), and BCL-2 expression in the resistant cells was verified by immunoblotting with the anti-human BCL-2 monoclonal antibody 6C8. The metastatic murine melanoma line K1735 clone M2X21 was maintained in modified Eagle's medium supplemented with 10% fetal calf serum, vitamins, and sodium pyruvate. Transfectants were generated by calcium phosphate precipitation using the Bil2-Neo plasmid as described previously (28). Cells were then selected for 3 weeks in the presence of 0.4 mg/ml G418, and BCL-2 expression in resistant colonies was determined by immunoblotting. Viability was assessed by trypan blue exclusion.

Peptide inhibitors of ICE (YVADfmk, zVADfmk, based on a cleavage site in interleukin-1) and CPP32 (DEVDfmk, based on a cleavage site in PARP) were purchased from Enzyme Systems Products, Inc. (Dublin, CA). Another peptide inhibitor of ICE (YVADcmk) and an inhibitor of the NS protease (APFcmk) were from Bachem (King of Prussia, PA). All other inhibitors were purchased from Sigma.

Cells were lysed in 0.1% Nonidet P-40 and nuclei isolated by centrifugation at 200 × g through a glycerol cushion as described previously (11). Nuclei were washed twice in incubation buffer (150 mM KCl, 5 mM MgCl₂, 25 mM Tris, pH 7.5) and resuspended in the same buffer for incubation. The purity of representative preparations from each cell type was verified by electron microscopy.

Quantification of DNA fragmentation was accomplished using the diphenylamine reagent as described previously (29). Briefly, cells were harvested by centrifugation, lysed in 0.5 ml of a buffer containing 0.5% Triton X-100, 25 mM Tris, pH 8.0, 10 mM EGTA, and 10 mM EDTA for 15 min on ice, and samples were centrifuged for 20 min at 12,000 × g to separate DNA fragments (supernatants) from intact chromatin (pellets). The DNA content in each fraction was determined using the diphenylamine reagent (30); results are expressed as the percentage of total DNA in each sample which resisted sedimentation at 12,000 × g. Alternatively, the DNA present in the centrifuged supernatants was precipitated overnight at 20 °C following the addition of 2 volumes of isopropyl alcohol and NaCl (to 0.5 M final concentration). After centrifugation at 12,000 × g for 10 min, precipitates were harvested, dried, and incubated for 1 h in TE buffer (10 mM Tris, pH 8.0, 1 mM EDTA) containing 0.2 mg/ml proteinase K and 1 mg/ml RNase A. The DNA fragments were resolved by electrophoresis for 1 h at 75 V on 1.5% agarose gels preimpregnated with ethidium bromide. The DNA fragments were then photographed.

Field-inversion gel electrophoresis (FIGE) (FIGE mapper, Bio-Rad) was used to detect large DNA fragments. Aliquots of nuclei were mixed with equal volumes of a melted plug solution containing 150 mM NaCl, 10 mM Tris, pH 7.5, 10 mM EDTA, and 1.2% low melting point agarose, transferred to wells of a FIGE plug mold, and incubated for 15 min at 4 °C. Plugs were then ejected and incubated for 16 h at 37 °C in a lysis buffer containing 50 mM Tris, pH 8.0, 50 mM EDTA, 1% sarcosyl, and 0.2 mg/ml proteinase K. Plugs were then washed three times for 1 h at 4 °C in a buffer containing 10 mM Tris and 1 mM EDTA, and the plugs were inserted into the wells of a 1% agarose gel. DNA fragments were then resolved by FIGE at a forward voltage of 180 V and a reverse voltage of 120 V for 16 h using a switch time ramp of 0.1-3.5 s (program 4 provided by the manufacturer). Gels were then stained with ethidium bromide and photographed.

Cells were lysed for 5 min at 37 °C and rotated for 1 h at 4 °C in a stringent lysis buffer containing 150 mM NaCl, 1% Triton X-100, 0.5% sodium dodecyl sulfate (SDS), 0.5% sodium deoxycholate, 1 mM phenylmethylsulfonyl fluoride, and 25 mM Tris, pH 7.5 (lamin B₁ and histone H1 experiments) or for 1 h at 4 °C in the same buffer lacking SDS and sodium deoxycholate (BCL-2 experiments). Debris was pelletted by centrifugation for 5 min at 12,000 × g, and pellets and/or supernatants were solubilized for 5 min at 100 °C in Laemmli's SDS-polyacrylamide gel electrophoresis sample buffer containing 100 mM dithiothreitol. Polypeptides were resolved at 100 V on 12% gels and electrophoretically transferred to nitrocellulose (0.2 µm, Schleicher & Schuell) for 1 h at 100 V. Membranes were blocked overnight (lamin experiments) or for 2 h (histone H1 and BCL-2 experiments) in a Tris-buffered saline solution containing Tween 20 (TBS-T) (25 mM Tris, pH 8.0, 150 mM NaCl, and 0.05% Tween 20) containing 3% (w/v) nonfat dried milk. To detect lamin B₁, membranes were probed for 2 h with the chicken anti-lamin B₁ antibody (1/5000 in TBS-T), 1 h with rabbit anti-chicken Ig (1/2000 in TBS-T), and 1 h with donkey anti-rabbit Ig (1/2000 in TBS-T) and then developed by enhanced chemiluminescence (ECL) (Amersham Corp.) according to the manufacturer's instructions. To detect histone H1, membranes were probed overnight with a mouse monoclonal antibody to histone H1 (Biogenex Inc., 1/1000 in block) and 1 h with sheep anti-mouse Ig (1/2000 in TBS-T) and developed by ECL. To detect human BCL-2, membranes were blocked overnight, probed for 1 h with 6C8 (1 µg/ml in TBS-T), 1 h with rabbit anti-hamster Ig (1/2000 in TBS-T), and 1 h with donkey anti-rabbit Ig (1/2000 in TBS-T) and developed by ECL.

RESULTS

Previous work using specific peptide-based protease inhibitors has implicated two distinct proteases (ICE family members and the Ca²⁺-dependent NS protease) in the cleavage of lamin B₁ in apoptotic cells and in a cell-free system. We therefore tested the effects of these peptide inhibitors on lamin B₁ cleavage and DNA fragmentation in thymocytes induced to undergo apoptosis upon exposure to glucocorticoid, the Ca²⁺-mobilizing agent thapsigargin, and antibodies to the T cell receptor complex. The peptide inhibitor of ICE, YVADfmk (7, 8, 9, 31), effectively inhibited lamin B₁ cleavage in response to all three stimuli (Fig. 1A) and partially prevented both spontaneous and induced DNA fragmentation in the cells (Fig. 1, B and C). However, a peptide inhibitor of the NS protease, APFcmk (16), was even more potent, completely blocking lamin cleavage (Fig. 1A) and DNA fragmentation (Fig. 1, B and C) in all circumstances. Both inhibitors also exerted corresponding effects on the decrease in cell size normally associated with apoptosis in thymocytes (measured by forward scatter by fluorescence-activated cell sorter, not shown) and cell death (Fig. 1D). Together, these results confirm other recent findings indicating that peptide inhibitors of the ICE family (31, 32) or the NS protease (32) interfere with all of the biochemical events of apoptosis in whole thymocytes that are exposed to certain apoptotic stimuli.

Recently, several groups have used different cell-free systems to study the regulation of various components of the apoptotic effector machinery (7, 14, 33, 34). However, we and others have shown that glucocorticoids, thapsigargin, and T cell receptor triggering all induce apoptosis in thymocytes via mechanisms that involve alterations in intracellular Ca²⁺ homeostasis (22) and that exogenous Ca²⁺ can directly promote the nuclear changes of apoptosis in isolated nuclei in the absence of cytoplasm (11, 18, 24, 27, 35). Moreover, it has been suggested that Ca²⁺-mediated DNA fragmentation involves unidentified serine proteases constitutively present in isolated rat liver nuclei (36). We therefore investigated the involvement of various proteases in DNA fragmentation observed in nuclei incubated with Ca²⁺. Preliminary work demonstrated that exogenous Ca²⁺ promoted DNA fragmentation in isolated thymocyte nuclei in a dose-dependent fashion (Fig. 2A). The response was characterized by the formation of both large 30-50-kb (Fig. 2B) and oligonucleosomal (not shown) DNA fragments. Similarly, Ca²⁺ promoted the cleavage of lamin B₁ to a distinct fragment of roughly 21 kDa (Fig. 2C) and the solubilization and cleavage of histone H1 (Fig. 2D) over a similar dose range.

The observation that Ca²⁺ promoted the solubilization and cleavage of both lamin B₁ and histone H1 suggested that activation of endogenous nuclear proteases was associated with Ca²⁺-stimulated DNA fragmentation in isolated thymocyte nuclei. We therefore employed a panel of protease inhibitors to characterize further the activities involved. Two different calpain inhibitors, the serine protease inhibitor TLCK and YVADfmk both failed to affect DNA fragmentation (Fig. 3, A and B), lamin B₁ cleavage (Fig. 3C), or histone H1 solubilization and cleavage (Fig. 3D). Similarly, another peptide-based ICE family antagonist, DEVDfmk (37), had no effect on Ca²⁺-mediated proteolysis or DNA fragmentation in isolated thymocyte nuclei over a wide dose range (1-200 µM, not shown). The only two inhibitors that were effective in this system were the serine protease antagonist TPCK and the NS protease antagonist APFcmk, which blocked all of the Ca²⁺-induced biochemical events (Fig. 3) in a dose-dependent fashion (Fig. 4). Thus, a protease sensitive to TPCK and APFcmk, but insensitive to ICE family and calpain antagonists, is required for Ca²⁺-mediated lamin B₁ and histone H1 solubilization and cleavage and for endonuclease activation in isolated thymocyte nuclei.

Effects of various protease inhibitors on DNA fragmentation and lamin cleavage in isolated nuclei. Panel A, quantitative comparison. Isolated nuclei were incubated for 2 h with or without 5 mM Ca²⁺ in the absence or presence of 100 µM calpain inhibitor I (CP I), TLCK, TPCK, YVADfmk, or APFcmk, and DNA fragmentation was quantified using the diphenylamine reagent. Mean ± S.D., n = 3. Panel B, effects on oligonucleosomal DNA fragmentation. Isolated nuclei were incubated as described in panel A, and the presence of oligonucleosomal DNA fragmmentation was determined by agarose gel electrophoresis. Panel C, effects on cleavage of lamin B₁. Isolated nuclei were incubated for 1 h with or without 5 mM Ca²⁺ in the absence or presence of 100 µM calpain inhibitor I, 100 µM TLCK, 100 µM YVADfmk, 100 µM TPCK, 100 µM APFcmk, 25 µM calmidizolium, 1 mM ATA, or 0.5 mM Zn²⁺, and lamin proteolysis was measured by immunoblotting. Panel D, effects on histone H1 solubilization and cleavage. Isolated nuclei were incubated with or without 5 mM Ca²⁺ in the absence or presence of 100 µM calpain inhibitor I, 100 µM TLCK, 100 µM TPCK, or 100 µM YVADfmk, and histone H1 was detected by immunoblotting.

Previous work has shown that the nuclease inhibitors aurintricarboxylic acid (ATA) and Zn²⁺, are capable of inhibiting DNA fragmentation and cell death in thymocytes undergoing apoptosis (38) and can interfere with Ca²⁺-induced DNA fragmentation in isolated nuclei (18, 26, 36). However, whether these agents directly inhibit endonuclease activation or whether they interfere with the protease activation required for endonuclease activation has not been established. Incubation with a concentration of Zn²⁺ which blocks apoptosis in whole thymocytes (26) did not prevent the cleavage of lamin B₁, whereas ATA partially prevented the response (Fig. 3C). The nuclease inhibitor ATA also partially inhibited large DNA fragment formation (Fig. 5A) and largely inhibited oligonucleosomal DNA fragmentation (Fig. 5B), consistent with previous observations. Interestingly, submillimolar concentrations of Zn²⁺ totally abrogated formation of both the large and oligonucleosomal DNA fragments (Fig. 5, A and B), whereas at no dose did Zn²⁺ inhibit lamin B₁ cleavage (Fig. 5C). Together, these observations suggest that the effects of ATA on large fragment formation may be caused in part by inhibition of the Ca²⁺-stimulated NS protease, but the effects of Zn²⁺ can be dissociated from NS protease inhibition and in this system may be linked solely to effects on nuclease activation.

Effects of nuclease inhibitors on DNA fragmentation and lamin proteolysis. Panel A, effects on 50-kb DNA fragmentation. Isolated nuclei were incubated for 1 h with or without 5 mM Ca²⁺ in the absence or presence of the indicated concentrations of ATA or Zn²⁺, and DNA fragmentation was assessed by FIGE. Panel B, quantitative analysis. Isolated nuclei were incubated for 2 h with or without 5 mM Ca²⁺ and the indicated concentrations of ATA () or Zn²⁺ (), and DNA fragmentation was quantified using the diphenylamine reagent. Mean ± S.D., n = 3. Panel C, dose-dependent effects of Zn²⁺ on lamin cleavage. Isolated nuclei were incubated for 1 h with or without 5 mM Ca²⁺ in the absence or presence of the indicated concentrations of Zn²⁺, and proteolysis of lamin B₁ was measured by immunoblotting.

Previous work has shown that calmodulin antagonists are capable of inhibiting DNA fragmentation in thymocytes exposed to apoptotic triggers (29, 39), and it has been suggested that the NS protease may be regulated by calmodulin (40). We therefore tested the effects of the calmodulin antagonist calmidizolium on DNA fragmentation and lamin B₁ cleavage in isolated nuclei. Calmidizolium blocked Ca²⁺-induced oligonucleosomal DNA fragmentation in a dose-dependent fashion (Fig. 6A); however, it failed to prevent cleavage of lamin B₁ (Fig. 6B) or the accumulation of 30-50-kb DNA fragments (Fig. 6C).

Members of the BCL-2 family play evolutionarily conserved roles in regulating apoptotic cell death, although the points at which they act within the pathway remain unclear, as effects of BCL-2 upstream and downstream (33, 34) of ICE family protease activation have been suggested. Recent efforts from our laboratory and others have demonstrated that a significant fraction of total cellular BCL-2 protein is localized to the nuclear envelope (28, 41, 42), and we have demonstrated that overexpression of BCL-2 can block Ca²⁺-stimulated oligonucleosomal DNA fragmentation in isolated nuclei from a Ras-transformed murine fibroblast line (28). To investigate further the effects of BCL-2 on the nuclear events of apoptosis, we transfected the T cell hybridoma 2B4.11, an extensively characterized cell line that responds to apoptotic triggers in fashion similar to thymocytes (43), with human BCL-2. Our choice of this line was reinforced by a recent report confirming that isolated nuclei from the parental line contain up to three polypeptides (49, 47, and 45 kDa) termed inducible Ca²⁺/Mg²⁺ lymphocyte endonuclease, which exhibit Ca²⁺-dependent nuclease activity inhibitable by Zn²⁺ (44). Expression of BCL-2 in whole cells was confirmed by immunoblotting (not shown), and preliminary experiments confirmed that the cells were resistant to induction of apoptosis in response to glucocorticoid or thapsigargin (Fig. 7, A and B). Consistent with previous studies, isolated nuclei from these cells contained substantial levels of human BCL-2 (Fig. 7C). Strikingly, incubation of nuclei from the BCL-2 transfectants in the presence of Ca²⁺ failed to stimulate lamin cleavage (Fig. 7D) or DNA fragmentation (Fig. 7E) for up to 2 h, although both responses were observed after longer times of incubation (not shown). To test whether the effects of BCL-2 were cell type-specific, we conducted analogous experiments with isolated nuclei from pooled BCL-2 transfectants of a metastatic murine melanoma line (K1735 clone M2X21) (45). Again, Ca²⁺-mediated cleavage of lamin B₁ and DNA fragmentation were dramatically reduced in nuclei from the BCL-2 transfectants compared with the levels observed in nuclei from the parental cells (Fig. 7, D and E). Thus, suppression of Ca²⁺-dependent protease activation appears to represent one important aspect of BCL-2 function in the nucleus that may explain, at least in part, its effects on the Ca²⁺-sensitive pathway of endonuclease activation.

DISCUSSION

Previous work has shown that lamin degradation is associated with apoptosis in several different model systems. In this study, we investigated the nature of the proteases responsible for this event in whole thymocytes exposed to several different apoptotic stimuli and in isolated nuclei incubated with exogenous Ca²⁺. The results indicate that specific peptide-based inhibitors of the ICE family (YVADfmk) and of the NS protease (APFcmk) blocked lamin cleavage, DNA fragmentation, and cell death in whole cells following T cell receptor triggering or treatment with glucocorticoid or the intracellular Ca²⁺-mobilizing agent thapsigargin. Because the NS protease inhibitor was significantly more potent than the ICE antagonist in blocking both lamin cleavage and DNA fragmentation, it is possible that APFcmk affects a process downstream of the ICE family in the pathway of apoptosis in thymocytes. Consistent with this notion, of a panel of inhibitors specific for several different proteases which were implicated previously in apoptosis, only the serine protease inhibitor TPCK and the NS protease-specific peptide APFcmk blocked lamin degradation and DNA fragmentation in isolated thymocyte nuclei that were incubated with Ca²⁺. Because both compounds possess a phenylalanine residue proximal to the chloromethyl ketone derivative (the position that determines inhibitor selectivity), it appears likely that both compounds affect the same protease in nuclei. Identifying this protease will be the subject of our future investigation.

Our results confirm other work indicating that a TPCK-sensitive protease is required for Ca²⁺-mediated endonuclease activation in whole cells and isolated nuclei (36, 46). Although the nature of this interaction between proteases and endonucleases in the apoptotic pathway is still not known, the present results support a hypothetical scenario (18) in which Ca²⁺-dependent activation of the NS protease promotes the degradation of the chromatin-associated polypeptides lamin B₁ and histone H1, resulting in the sequential exposure of the lamin binding sites and internucleosomal regions to a Ca²⁺-dependent endonuclease. Because the lamin binding sites are spaced every 20-50 kb and the histone H1 attachment sites every 180 bases in chromatin, this model would account for the stepwise pattern of DNA cleavage observed in apoptotic thymocytes and other cell types. From these data it might be postulated that chromatin cleavage is mediated by a constitutively active endonuclease whose action is restricted solely by DNA accessibility. However, this idea does not appear to be correct, as Zn²⁺ preferentially blocks the production of both large and oligonucleosomal DNA fragments without interfering detectably with proteolysis of lamin B₁. Conversely, Lazebnik and colleagues (15) have shown that it is possible to inhibit lamin degradation without blocking oligonucleosomal DNA fragmentation with ICE antagonists in nuclei incubated with cytoplasmic extracts from cells arrested at the S and M phases of the cell cycle (S/M extracts). Moreover, our results support the idea suggested by Sun and Cohen (20) that independent endonuclease activities are responsible for the 50-kb and oligonucleosomal DNA fragments, as we have found that formation of the latter but not the former is blocked by calmodulin antagonists in isolated nuclei. In addition to removing chromatin-bound proteins, it is possible that proteolytic processing directly affects endonuclease activation, an idea that is supported by preliminary work suggesting that two candidate Ca²⁺-dependent thymocyte endonucleases appear to exist as high molecular weight precursors in untreated cells (47, 48), and it may account for the appearance of the 47- and 45-kDa inducible Ca²⁺/Mg²⁺ lymphocyte endonuclease polypeptides in isolated nuclei from 2B4.11 cells exposed to glucocorticoid or anti-T cell receptor antibodies prior to the onset of apoptosis (44).

The nuclear localization, Ca²⁺-dependence, and peptide inhibitor sensitivity of the protease implicated in lamin cleavage and DNA fragmentation in this study are all specific characteristics of the NS protease. Although most of the molecular characteristics of this protease are still not known, preliminary work indicates that it is structurally related to the multicatalytic proteinase complex, otherwise known as the proteosome (16). The observation that elevations in the cytosolic Ca²⁺ concentration trigger chromatin condensation and nuclear envelope breakdown during mitosis (49, 50) coupled with the ability of cytosolic Ca²⁺ increases to activate the NS protease directly (16) raises the possibility that the NS protease is not only involved in certain examples of apoptosis but also in some aspects of mitosis. Circumstantial support for this hypothesis is provided by the observation that disregulated proliferation in hyperplastic nodules and hepatocellular carcinomas is associated with elevated NS protease activity (51, 52), and a peptide inhibitor of the NS protease blocks chemical transformation in a murine fibroblast line (53). Coordinate regulation of catabolic enzymes in apoptosis and mitosis is an attractive possibility given the observed similarities in cellular morphology as well as the known capacity of oncogenes (most notably Myc) (54) and cyclin-dependent kinases (55) to promote both responses. It is also possible that the proteosome itself may play some role in the cytoplasmic events of apoptosis, as increases in ubiquitin expression (56) and in enhanced ubiquitination of cellular proteins (57) are linked to programmed cell death in the moth Manduca sexta and in irradiated human peripheral blood lymphocytes, respectively.

Recently, a number of investigators have used cell-free systems in an attempt to identify the biochemical machinery necessary for the nuclear features of apoptosis. These studies were initiated by Lazebnik and co-workers (14), who demonstrated that S/M extracts were capable of inducing characteristic chromatin condensation, nuclear lamina disruption, and DNA fragmentation in isolated nuclei from control cells. Importantly, lamin degradation and endonuclease activation in nuclei incubated with S/M extracts (and other cell-free systems) occurred in the absence of exogenous Ca²⁺ (14), and lamin degradation and DNA fragmentation were insensitive to TPCK and the peptide inhibitor of the NS protease (15). Together with the results reported here, these data indicate that there are at least two pathways to lamin degradation and DNA fragmentation in isolated nuclei, one mediated by activated ICE family members and the other by Ca²⁺-induced activation of the NS protease. It is still formally possible that the ICE pathway activates one of the Ca²⁺-independent enzymatic activities of the NS protease, an issue that requires further investigation.

Despite intensive investigation, the mechanisms underlying the inhibition of apoptosis by BCL-2 are still not clear. In this study, we demonstrated that the amount of BCL-2 constitutively present in isolated nuclei from two divergent cell lines transfected with BCL-2 is sufficient to block Ca²⁺-mediated lamin degradation and DNA fragmentation. Precisely how it does so is not known, although we have recently obtained evidence that BCL-2 may exert its effects, at least in part, via its action on an energy-dependent Ca²⁺ uptake system localized to the nuclear envelope (42). Because the NS protease is apparently localized to the nuclear envelope as well, and we presented evidence previously that nuclear Ca²⁺ uptake promotes endonuclease activation (27), this activity of BCL-2 may account for the inhibition of protease and nuclease activation observed. Alternatively, because recent work demonstrated that the nuclear Ca²⁺ pool regulates transport of large (>10 kDa) polypeptides through nuclear pores (58), it is also possible that BCL-2 exerts its effects on nuclei by blocking translocation of certain cytosolic proteins (i.e. p53, Myc, ICE homologs, DNase I) that may cooperate with the NS protease to promote lamin/histone cleavage and DNA fragmentation. These possibilities are not mutually exclusive, and both are under investigation at present.

An important question raised by this study concerns why specific inhibitors of either the ICE family or the NS protease are capable of inhibiting lamin B₁ degradation and DNA fragmentation in whole thymocytes, whereas only the latter prevent these changes in isolated nuclei supplemented with Ca²⁺. It is possible that the ICE family is involved in promoting the Ca²⁺ alterations that trigger nuclear NS protease activation or that the pathways of apoptosis triggered in thymocytes by glucocorticoids, thapsigargin, and anti-T cell receptor antibodies in thymocytes are actually initiated within the nucleus. The latter would have important implications for current thinking concerning the location of the apoptotic effector machinery in general, as the results obtained with cytoplasmic apoptosis-promoting extracts and enucleated cells (59, 60) imply that all of the important biochemical alterations tied to commitment to cell death occur in the cytoplasm. Both of these explanations for the likely interrelationship between the ICE family and the NS protease are testable hypotheses that are under investigation at present.