Bcl-xL acts downstream of caspase-8 activation by the CD95 death-inducing signaling complex.

The Bcl-2 family member Bcl-xL has often been correlated with apoptosis resistance. We have shown recently that in peripheral human T cells resistance to CD95-mediated apoptosis is characterized by a lack of caspase-8 recruitment to the CD95 death-inducing signaling complex (DISC) and by increased expression of Bcl-xL (Peter, M. E., Kischkel, F. C., Scheuerpflug, C. G., Medema, J. P., Debatin, K.-M., and Krammer, P. H. (1997) Eur. J. Immunol. 27, 1207-1212). This raises the possibility that Bcl-xL directly prevents caspase-8 activation by the DISC. To test this hypothesis a cell line in which CD95 signaling was inhibited by overexpression of Bcl-xL was used. In these MCF7-Fas-bcl-xL cells Bcl-xL had no effect on the recruitment of caspase-8 to the DISC. It did not affect the activity of the DISC nor the generation of the caspase-8 active subunits p18 and p10. In contrast, cleavage of a typical substrate for caspase-3-like proteases, poly(ADP-ribose) polymerase, was inhibited in comparison with the control-transfected CD95-sensitive MCF7-Fas cells. To test whether Bcl-xL would inhibit active caspase-8 subunits in the cytoplasm, a number of immunoprecipitation experiments were performed. Using monoclonal antibodies directed against different domains of caspase-8, anti-Bcl-xL antibodies, or fusion proteins of glutathione S-transferase with different domains of caspase-8, no evidence for a direct or indirect physical interaction between caspase-8 and Bcl-xL was found. Moreover, overexpression of Bcl-xL did not inhibit the activity of the caspase-8 active subunits p18/p10. Therefore, in this cell line that has become resistant to CD95-induced apoptosis due to overexpression of Bcl-xL, Bcl-xL acts independently and downstream of caspase-8.

Cross-linking of CD95 (APO-1/Fas) results in the recruitment of a set of proteins that include FADD/MORT1 (7,8) and caspase-8 (FLICE/MACH/Mch5; Refs. 9 -11) to the receptor forming the death-inducing signaling complex (DISC) 1 (12,9). Recruitment of caspase-8, which results in its activation (12), requires the presence of FADD and is mediated by an interaction of the death effector domain (DED) of FADD and the first DED of caspase-8 (13,14). Interestingly, this DED was a region found to share weak homology with two areas in CED-4 (15,16). In addition, a recent report demonstrated coimmunoprecipitation of Bcl-x L with caspase-8 and vice versa (17). These data suggest a regulatory mechanism in mammalian cells in which the CED-9 homolog Bcl-x L influences the activity of the CED-3 homolog caspase-8 possibly involving the CED-4 homology region of caspase-8.
We have recently reported that in a CD95-resistant peripheral T cell population, FADD was recruited to the activated CD95 receptor, but recruitment and activation of procaspase-8 was not detected. In addition, this apoptosis resistance correlated with the expression of the anti-apoptotic molecule Bcl-x L (18), suggesting that it might directly render cells CD95 apoptosis resistant by interfering with the DISC. Recently, it was shown that overexpression of Bcl-x L inhibited CD95-mediated apoptosis without blocking initial caspase activation (19). We therefore tested whether Bcl-x L might interfere with recruitment of caspase-8 to the DISC or with the activity of caspase-8 using a breast carcinoma cell line MCF7 expressing high levels of CD95 (MCF7-Fas) (20). Overexpression of Bcl-x L in these cells renders them resistant to CD95-induced apoptosis (20). In addition, no DNA fragmentation, morphological changes, or cleavage of poly(ADP-ribose) polymerase (PARP) could be detected in these cells. However, cleavage of caspase-8 was not affected by overexpression of Bcl-x L . No direct or indirect association between Bcl-x L and caspase-8 could be detected, indicating that in these cells Bcl-x L acted independently and downstream of caspase-8.

EXPERIMENTAL PROCEDURES
Cell Lines-The breast carcinoma cell line MCF7 transfected with CD95 and a control plasmid (MCF7-Fas-vector) or CD95 and Bcl-x L (MCF7-Fas-bcl-x L ) were a kind gift from M. Jää ttelä (Kopenhagen, Denmark). They were cultured as described elsewhere (20). The SKW6.4 cells were cultured in RPMI ϩ 10% fetal calf serum and 0.05 mg/ml gentamycin.
Cytotoxicity Assay-10 6 cells were incubated in 24-well plates (Costar, Cambridge, MA) with 1 g/ml anti-APO-1 in medium (5 ϫ 10 5 cells/ml) for 16 h at 37°C. After incubation cells were photographed at a 400-fold magnification. Quantification of DNA fragmentation as a specific measure of apoptosis was carried out essentially as described elsewhere (23,24).
GST Fusion Proteins and Immunoprecipitation-Using standard polymerase chain reaction and cloning techniques the following GST fusion proteins were generated: GST-FADD, GST-C-FADD (amino acids 89 -207), GST-caspase-8, GST-N-caspase-8 (amino acids 1-180), and GST-C-caspase-8 (amino acids 181-478). GST fusion proteins were purified as described previously (10), and 20 g of each GST fusion protein precoupled to glutathione beads was used per precipitation. For immunoprecipitation 10 g mAbs (25 g in the case of C5) were coupled to either anti-IgG1 Agarose beads (C1, N2) (Sigma) or to protein A/Gplus agarose (C5) (Santa Cruz Biotechnology), and postnuclear supernatants of 10 7 cells were added and incubated for more than 1 h at 4°C. Beads were then washed three times with lysis buffer (30 mM Tris/HCl, pH 7.5, 150 mM NaCl, 1 mM phenylmethylsulfonyl fluoride and small peptide inhibitors (12), 1% Triton X-100 (Serva) and 10% glycerol). The amount of DISC-associated FADD was determined as follows: 10 7 MCF7 cells were either first treated with 2 g/ml anti-APO-1 for 5 min at 37°C and then lysed (stimulated condition) or first lysed and then supplemented with 2 g/ml anti-APO-1 (unstimulated condition) as described previously (12). Triton X-100 solubilization of cellular proteins using the above lysis buffer and immunoprecipitation of CD95 was done as described previously (12). All immunoprecipitates were separated on 12% SDS-PAGE.
In Vitro Caspase-8 Activation Assay-CD95 DISC was immunoprecipitated from 5 ϫ 10 7 activated or nonactivated cells as described above, and immunoprecipitates were incubated with in vitro translated 35 S-labeled caspase-8 in caspase-8 cleavage buffer (50 mM Hepes, pH 7.4, 100 mM NaCl, 0.1% 3-(cyclohexylamino)-1-propanesulfonic acid, 10 mM dithiothreitol, and 10% sucrose) for 24 h at 4°C. The cleavage reactions were stopped by addition of standard reducing sample buffer. After boiling for 3 min samples were subjected to 15% SDS-PAGE, amplified (Amplify, Amersham), dried, and subjected to autoradiography. For in vitro PARP cleavage of active caspase-8 subunits, 5 ϫ 10 7 MCF7-Fas cells were left untreated or stimulated with 2 g/ml anti-APO-1. After 2 h cells were lysed, and lysates were made as described above. Lysates were precleared with mouse IgG, and the active caspase-8 subunits were subsequently precipitated with the C5 mAb as described above. After intensive washing precipitates were incubated in caspase-8 cleavage buffer together with 300 ng of recombinant PARP for 4 h at 37°C. Reactions were stopped by the addition of standard reducing sample buffer, separated on 10% SDS-PAGE, and immunoblotted with C-II-10.

MCF7-Fas-bcl-x L Cells
Are Resistant to CD95-mediated Apoptosis-Since the CD95 DISC seemed to determine apoptosis sensitivity in peripheral T cells (18), we investigated whether expression of Bcl-x L would directly affect the composition of the DISC. However, normal T cells can hardly be transfected and overexpression of Bcl-2 or Bcl-x L in lymphoid cells does not necessarily block the CD95 pathway (25)(26)(27)(28)(29). Hence, we tested the effect of overexpression of Bcl-x L in a breast carcinoma cell MCF7-Fas, one of the few cell lines that have been reported to become resistant to CD95-mediated apoptosis upon overexpression of Bcl-x L (20). Indeed overexpression of Bcl-x L blocked the morphological changes (i.e. rounding up and detaching) induced by cross-linking of CD95 (Fig. 1A). In addition, fragmentation of chromosomal DNA as compared with control-transfected MCF7-Fas cells was almost completely inhibited in the Bcl-x L transfectants (Fig. 1B). This difference in sensitivity toward cross-linking of the CD95 receptor was not due to a difference in expression of the receptor, since both transfectants express similar amounts (data not shown), but due to the expression of Bcl-x L (Fig. 1B).
Bcl-x L Does Not Associate with Caspase-8 or Block Its Activity-Recently, Chinnaiyan et al. (17) demonstrated in transient cotransfection experiments that Bcl-x L could be coimmunoprecipitated together with caspase-8 and vice versa. This association was indirect as it required an additional factor believed to be a mammalian CED-4 homolog (17). Most importantly the data suggested that the CED-9 homolog Bcl-x L might act upstream of the CED-3 homolog caspase-8. Since in the MCF7-Fas cells apoptosis through CD95 involved activation of caspase-8 and CD95-mediated apoptosis in these cells was blocked by overexpression of Bcl-x L , we could now test this hypothesis in vivo in stably transfected cells. To this end we generated a number of reagents to use in immunoprecipitation experiments. All following experiments were done both with untreated and anti-APO-1-treated MCF7-Fas and MCF7-Fasbcl-x L cells. First we tested whether Bcl-x L could be immunoprecipitated from MCF7 cells using fusion proteins of GST with different functional domains of caspase-8 or FADD (Fig. 2). Using GST-caspase-8 and GST-N-caspase-8 (comprising the first 188 amino acids of caspase-8 containing both DED), we could precipitate FADD from lysates of both transfectants (Fig.  2, B and D, lanes 7-10). Therefore, both DED containing GSTcaspase-8 constructs were functionally active in binding the physiologically important signaling molecule FADD through their N termini. FADD was not precipitated when GST-Ccaspase-8 was used (lacking the two N-terminal DED) (Fig. 2,  B and D, lanes 11 and 12). We then tested whether cytoplasmic Bcl-x L from anti-APO-1-treated and untreated cells would coprecipitate with GST-caspase-8 or GST-C-caspase-8. As a control for specificity, we used GST, GST-FADD, or GST-C-FADD (lacking the DED). Bcl-x L was clearly detectable by Western blot analysis in Triton X-100 lysates from 5 ϫ 10 5 MCF7-Fasbcl-x L cells (Fig. 2C, lanes 13 and 14). However, no associated Bcl-x L could be detected with any of the GST fusion proteins tested, although 20-fold more lysate (10 7 cells) was used for precipitation ( Fig. 2C, lanes 1-12).
Since the GST domain might have prevented interaction of caspase-8 with Bcl-x L , specific anti-caspase-8 and anti-Bcl-x L antibodies were used for immunoprecipitation experiments (Figs. 3 and 4). The two expressed caspase-8 isoforms (21) and their cleavage products (p43, p41, p26, p24, p18) (21) were clearly detected in similar amounts in cellular lysates from unstimulated and anti-APO-1-stimulated cells (Ϫ/ϩ in Fig. 3). Immunoprecipitation with a rabbit anti-Bcl-x L serum detected Bcl-x L (Fig. 3C). However, it completely failed to coimmunoprecipitate either full sized caspase-8 or any of its active subunits (Fig. 3, A and B).
To exclude that the antibodies against Bcl-x L interfered with the binding to caspase-8, we performed a reverse immunoprecipitation with different mAbs against either the C terminus or the N terminus of caspase-8 and determined whether coimmunoprecipitation of Bcl-x L could be detected. Using the C1 mAb, which recognizes the active caspase-8 subunit p18, we could immunoprecipitate full sized caspase-8 (Fig. 4A, lanes 5, 6, 13,  and 14) and both the p18 and the p10 subunits (Fig. 4, B and C,  lanes 6 and 14), indicating that the mAb did not interfere with the complex formation between p10 and p18. However, no coimmunoprecipitation of Bcl-x L could be detected using this antibody (Fig. 4D, lanes 5, 6, 13, and 14). The N2 antibody, which reacted with the caspase-8 prodomain, immunoprecipitated full sized caspase-8 (Fig. 4A, lane 7, 8, 15, and 16), yet did not precipitate any Bcl-x L (Fig. 4D, lanes 7, 8, 15, and 16).
Since the reported interaction between Bcl-x L and caspase-8 Bcl-x L was immunoprecipitated from 10 7 untreated (Ϫ) or 2-h anti-APO-1-treated (ϩ) MCF7-Fas or MCF7-Fas-bcl-x L . As a control, cellular lysates were incubated with rabbit IgG. Immunoprecipitates were analyzed for the presence of Bcl-x L (C) and for coimmunoprecipitation of endogenous full sized caspase-8 (A) or the active caspase-8 subunit p18 (B). Full-length caspase-8 and the cleaved subunits were detected in cellular lysates from 5 ϫ 10 5 cells (lanes 7-10). Immunoblots were done with the mAb N2 directed against the N terminus (A), with the p18 mAb C15 (B), and with a polyclonal anti-Bcl-x L antibody (C). Positions of rabbit (rb)IgG H (detected due to some cross-reactivity of the secondary anti-mIgG1 Ab with the heavy chain of the anti-Bcl-x L anti-serum), the two expressed caspase-8 isoforms, their cleavage intermediates p43/ p41, their prodomains p26/p24, and the caspase-8 active subunit p18 and Bcl-x L are indicated. was shown to involve both the N terminus and the p18 subunit of caspase-8 (17), both mAbs against caspase-8 used could still interfere with the interaction of Bcl-x L with caspase-8. Therefore, we made use of a third caspase-8 mAb, C5, that is directed against the p10 subunit. This region was reported by Chinnaiyan et al. (17) not to be involved in the binding of Bcl-x L . The C5 mAb immunoprecipitated both the full-length caspase-8 isoforms (Fig. 4A, lanes 3, 4, 11, and 12) and the active subunits p18 and p10 (Fig. 4, B and C, lanes 4 and 12), indicating that the mAb did not interfere with complex formation between p10 and p18. However, also C5 did not coimmunoprecipitate any Bcl-x L (Fig. 4D, lanes 3, 4, 11, and 12). Taken together in all experiments, no caspase-8/Bcl-x L association was detected even when the Western blots were highly overexposed (data not shown).
To test whether Bcl-x L overexpression would inhibit the activity of caspase-8, we immunoprecipitated the active subunits p18/p10 and determined the in vitro PARP cleavage activity of the immunoprecipitated complex. The immunoprecipitate of the active subunits from anti-APO-1-treated cells using the C5 mAb displayed clear PARP cleavage in vitro (Fig. 4E, lanes 4  and 12), while no cleavage of PARP was observed when caspase-8 was immunoprecipitated from untreated cells (Fig.  4E, lanes 3 and 11) or when mIgG control immunoprecipitates were used (Fig. 6E, lanes 1, 2, 9, and 10). Importantly, again no difference could be detected in in vitro PARP cleavage when the active subunits were immunoprecipitated from MCF7-Fas or MCF7-Fas-bcl-x L cells, indicating that overexpression of Bcl-x L did not affect the enzymatic activity of caspase-8.
Formation of the DISC and Activation of Caspase-8 Are Unaffected in MCF7-Fas-bcl-x L Cells-The above data indicate that Bcl-x L likely does not exert its inhibitory activity by physically interacting with procaspase-8 or active caspase-8 sub-units in the cytoplasm, regardless whether anti-APO-1-treated or untreated cells were tested. We have shown recently that in vivo caspase-8 is activated by recruitment to the CD95 DISC (13). If Bcl-x L functioned upstream of caspase-8, it should interfere with caspase-8 cleavage by the DISC. The first step in DISC formation is recruitment of FADD (12). This can be visualized by immunoprecipitation of CD95 and subsequent immunoblotting with a FADD-specific mAb. Treatment of SKW6.4 B cell lymphomas with anti-APO-1 clearly resulted in an association of FADD with CD95 (Fig. 5A, lanes 7 and 8). Similarly, both MCF7-Fas-vector and MCF7-Fas-bcl-x L cells, which express comparable amounts of FADD (Fig. 5A, lanes 1  and 2), efficiently recruited FADD to CD95 after treatment with anti-APO-1 (Fig. 5A, lanes 3-6). We next tested whether the activity of the DISC to process caspase-8 was affected by Bcl-x L overexpression. We have recently shown that this activity can be determined in vitro by incubation of immunoprecipitated DISC with in vitro translated 35 S-labeled caspase-8 (13). Testing the MCF7-Fas transfectants in this assay demonstrated that Bcl-x L overexpression did also not alter the activity of the DISC to activate caspase-8 in vitro (Fig. 5B).
In Vivo Bcl-x L Acts Downstream of Caspase-8 and Upstream of a PARP Cleaving Caspase-To test whether Bcl-x L would inhibit cleavage of caspase-8 in vivo, the C15 mAb recognizing the p18 subunit of caspase-8 was used. This antibody allowed to follow caspase-8 activation in vivo by detecting both the full-length caspase-8 isoforms as well as the active caspase-8 subunit p18 (Fig. 6A). Importantly, in both control transfected and Bcl-x L overexpressing MCF7-Fas cells the p18 subunit was formed to a similar extent and with identical kinetics (Fig. 6A,  lower panel). Furthermore, as has been shown previously for FIG. 4. Bcl-x L does not coimmunoprecipitate with caspase-8. Cellular lysates from 10 7 untreated or anti-APO-1 treated MCF7-Fas or MCF7-Fas-bcl-x L were incubated with Sepharose-coupled mAb against the caspase-8 p10 subunit (C5), the p18 subunit (C1), the N-terminal part (N2), or as a control mouse IgG. Immunoprecipitates were analyzed for the precipitated caspase-8 subunits using the C1 mAb to detect full-length caspase-8 and the p18 subunit (A and B) or the C5 mAb to detect the p10 subunit (C). Immunoprecipitates were analyzed for coimmunoprecipitation of Bcl-x L by Western blotting with a polyclonal anti-Bcl-x L antibody. As a control, cellular lysates of 5 ϫ 10 5 MCF7-Fas or MCF7-Fas-bcl-x L cells were analyzed (lanes [17][18][19][20]. E, the active caspase-8 subunits were immunoprecipitated using the C5 mAb, and after immunoprecipitation complexes were incubated for 4 h at 37°C with recombinant PARP. Resulting PARP cleavage was determined by immunoblotting with the C-II-10 mAb. Cells were stimulated as in Fig.  3. Immunoprecipitations were performed with 10 g for C1, N2, and mIgG and with 25 g for C5. Indicated are the positions of heavy and light chain (mainly visible in C5 immunoprecipitations due to aspecific cross-reaction of the secondary antibodies with the high amount of heavy and light chain present), caspase-8, p18, p10, Bcl-x L , and the p116 and p85 PARP fragments.  5, and 7) or treated with anti-APO-1 for 5 min (lanes 4, 6, and 8). Subsequently CD95 was immunoprecipitated, and the amount of associated FADD was determined by Western blot analysis. B, MCF7-Fas-vector (lanes 1 and 2) and MCF7-Fas-bcl-x L (lanes 3 and 4) cells were left untreated (lanes 1 and 3) or treated with anti-APO-1 for 5 min (lanes 2 and 4). Subsequently the immunoprecipitated DISC from these cells was analyzed for its in vitro caspase-8 cleavage activity using in vitro translated 35 S-caspase-8. The resulting products were separated on 15% SDS-PAGE and detected by autoradiography. Positions of the products as described recently (13) are indicated on the right. the SKW6.4 cell line (21), all full-length caspase-8 was processed after cross-linking of CD95 in both cell lines (Fig. 6A,  upper panel). In contrast to caspase-8 cleavage, cleavage of PARP, a typical substrate of caspase-3-like proteases (9,30,31), was blocked in the Bcl-x L -overexpressing MCF7-Fas cells, while PARP was cleaved in the control-transfected MCF7-Fas cells upon anti-APO-1 treatment (Fig. 6B). Taken together our data suggest that Bcl-x L inhibits the CD95 signaling pathway downstream of caspase-8 cleavage and upstream of the activity of caspase-3-like proteases.
We have established that Bcl-x L does not block CD95-mediated apoptosis in MCF-Fas cells by directly interfering with either caspase-8 association with the DISC or caspase-8 activation in the DISC. Our data indicate that Bcl-x L does not associate with caspase-8 nor does it inhibit its activity. Therefore, Bcl-x L does not act at the level of caspase-8 activation, it rather acts downstream of caspase-8 and upstream of a PARP cleaving activity in these cells. DISCUSSION Regulation of apoptosis in the nematode C. elegans serves as a paradigm for apoptosis regulation in mammalian cells. In the worm three genes dictate apoptosis during development (2). Mutations in ced-3 or ced-4 abolish apoptosis, while mutation of ced-9 results in aberrant cell death. CED-9 is homologous to the mammalian Bcl-2 family and inhibits apoptosis induced by CED-3, which is homolog to the mammalian family of caspases and therefore responsible for execution of apoptosis. CED-4, another apoptosis-promoting gene product, physically interacts with CED-9 (17,32,33) and is required for CED-9 to inhibit the function of CED-3. No mammalian homolog for CED-4 has yet been discovered. CED-3 is most homologous to caspase-3-like proteases. It has been shown that Bcl-2 functions upstream of these caspases (27, 34 -38). Since caspase-8 has a caspase-3like activity, the question arose whether Bcl-2 family members could act upstream of caspase-8. Recently, it was suggested that Bcl-x L inhibits caspase-8 activity by physically interacting with caspase-8 (17). This association required CED-4 and was postulated to occur at the level of the mitochondria (39). We have demonstrated previously that during CD95-mediated apoptosis caspase-8 is activated by recruitment to the CD95 DISC. After completion of this process all cytosolic caspase-8 can be converted to active caspase-8 subunits detectable in the cytoplasm (13). If Bcl-x L acts upstream of caspase-8, it would have to bind to either procaspase-8 in the DISC or to procaspase-8 or the active caspase-8 subunits in the cytoplasm. We now show that CD95-induced caspase-8 activation or activity is not affected by overexpression of Bcl-x L in the CD95-resistant cell line MCF7-Fas-bcl-x L . Bcl-x L did not affect the activity of the DISC. In addition, Bcl-x L did not interfere with PARP cleavage activity of the caspase-8 active subunits p10/p18 when immunoprecipitated using mAbs against the caspase-8 active subunits p10 (Fig. 6E) or p18 (data not shown).
Recently, the existence of a caspase activity upstream of Bcl-x L was suggested studying Bcl-x L -transfected Jurkat cells using PARP cleavage as a read-out (19). In this study Bcl-x Ltransfected Jurkat cells were resistant to CD95-mediated apoptosis, but PARP was still cleaved. It was suggested that caspase-8 was responsible for the observed PARP cleavage in these cells. However, the identity of the PARP cleaving caspase was not determined. We now show that in MCF7-Fas-bcl-x L cells PARP cleavage was completely blocked by overexpression of Bcl-x L , whereas caspase-8 activity was unchanged. The discrepancy between the two reports is likely due to different cell lines used. Our data demonstrate that first, in vivo PARP was not directly cleaved by caspase-8, and second, another caspase-3-like protease that cleaves PARP was blocked by Bcl-x L .
Our data suggest that Bcl-x L acts downstream of caspase-8. The indirect association between caspase-8 and Bcl-x L that has been reported previously was based on experiments that used a transient transfection assay (17). We now have tested this interaction in vivo in stably transfected MCF7 cells by performing immunoprecipitation experiments using highly sensitive anti-Bcl-x L and anti-caspase-8 antibodies. To prevent competition of the anti-caspase-8 mAbs and Bcl-x L for the same binding site, we used mAbs against three different regions of caspase-8 (the prodomain and the p18 and p10 subunits). Consistent with the lack of functional inhibition of caspase-8 in the DISC, no coimmunoprecipitation of Bcl-x L with caspase-8 or vice versa was detectable.
We have recently shown that CD95 resistance in short term activated peripheral T cells is characterized by a lack of caspase-8 recruitment to the DISC despite high expression of caspase-8 in the cytoplasm (18). In addition, we found that the expression of Bcl-x L correlated with CD95 apoptosis sensitivity in these T cells, making it possible that Bcl-x L directly prevented caspase-8 recruitment to the DISC. However, for caspase-8 being the main target for Bcl-x L is in contradiction to the reported effects of Bcl-2 or Bcl-x L on the CD95 signaling pathway in lymphoid cells. Reports range from no inhibition (27)(28)(29)40) to partial inhibition (25) to substantial inhibition of CD95-mediated apoptosis (19,34,38). One of the few cell lines that have been reported to be completely resistant to CD95 apoptosis after overexpression of Bcl-x L is the breast carcinoma cell line MCF7-Fas (20). Using these cells we did not find any association of Bcl-x L with procaspase-8 or caspase-8 active subunits or any effect of Bcl-x L on caspase-8 activation. However, our experiments do not exclude that Bcl-x L could influence the action of caspase-8 by preventing a possible translocation of the caspase-8 active subunits within the cell from one location to their physiological target(s) without blocking their enzymatic activity.
The presence of homologous functionally interchangeable proteins in C. elegans and mammals suggests a universal principle in apoptosis regulation conserved throughout evolution. The data presented here demonstrate that the CED-9 homolog Bcl-x L does not act by interfering with the activation of the CED-3 homolog caspase-8. Therefore, apoptosis sensitivity is at least regulated at two separate levels: at the level of the DISC and independently further downstream at the level of Bcl-x L likely regulating the activity of other CED-3 homologs. In contrast to C. elegans mammalian cells seem to have developed at least two levels at which caspases act. First level caspases, such as caspase-8, may couple the intracellular death machinery to death receptors and may not be regulated by CED-9-like molecules. Second level caspases likely representing caspase-3-like proteases located downstream of mitochondria may represent the CED-3 homologs likely to be regulated by the CED-9 homolog Bcl-x L (Bcl-2).