Apaf-1 Oligomerizes into Biologically Active ∼700-kDa and Inactive ∼1.4-MDa Apoptosome Complexes

Apaf-1, by binding to and activating caspase-9, plays a critical role in apoptosis. Oligomerization of Apaf-1, in the presence of dATP and cytochrome c, is required for the activation of caspase-9 and produces a caspase activating apoptosome complex. Reconstitution studies with recombinant proteins have indicated that the size of this complex is very large in the order of ∼1.4 MDa. We now demonstrate that dATP activation of cell lysates results in the formation of two large Apaf-1-containing apoptosome complexes with M r values of ∼1.4 MDa and ∼700 kDa. Kinetic analysis demonstrates that in vitro the ∼700-kDa complex is produced more rapidly than the ∼1.4 MDa complex and exhibits a much greater ability to activate effector caspases. Significantly, in human tumor monocytic cells undergoing apoptosis after treatment with either etoposide orN-tosyl-l-phenylalanyl chloromethyl ketone (TPCK), the ∼700-kDa Apaf-1 containing apoptosome complex was predominately formed. This complex processed effector caspases. Thus, the ∼700-kDa complex appears to be the correctly formed and biologically active apoptosome complex, which is assembled during apoptosis.

Apoptotic cell death is characterized by a commitment and an execution phase, the latter being distinguished by a series of stereotypic morphological and biochemical changes, which require the activation of a family of caspases (1,2). Caspases are constitutively expressed in cells as inactive proenzymes and require proteolytic cleavage to be active. Presently, fourteen caspases have been identified and these can be divided into two classes. "Initiator" caspases, such as caspases-8 and -9, possess long pro-domains and either directly or indirectly activate "effector" caspases, such as caspase-3, -6, and -7, which contain short pro-domains (3)(4)(5). Activation of initiator caspases involves interaction of their long pro-domains with either caspase recruitment domains (CARD) 1 (6) or the death effector domains (DEDs) of adapter molecules and results in the formation of cellular complexes. Thus with receptor-mediated cell death, ligation of the CD95 (Fas/APO-1) receptor results in the formation of a death-inducing signaling complex (DISC), in which the death effector domains of the adapter molecule FADD (MORT1) interact with the N-terminal death effector domains of caspase-8 (7,8). Many apoptotic stimuli cause cell death by inducing the release of cytochrome c (9), which together with Apaf-1 and ATP/dATP facilitates the processing of caspase-9 and initiates a caspase cascade (10,11). Apaf-1, the first identified mammalian homologue of CED-4, is a 130-kDa protein. It contains an N-terminal CARD domain, followed by a region homologous to CED-4 and a C-terminal domain containing multiple WD-40 repeats, which are involved in protein-protein interactions (11). Both CED-4 and Apaf-1 undergo self-oligomerization, which is required for the activation of CED-3 and caspase-9, respectively (12,13). In the presence of dATP and cytochrome c, the CARD domains of caspase-9 and Apaf-1 interact to produce an apoptosome complex in which procaspase-9 is processed to its active form, which in turn activates caspase-3 (10,14,24).
The apoptosome has been defined as an Apaf-1-containing complex that catalyzes the activation of caspases (15,16). Recently, using purified recombinant Apaf-1 in the presence of cytochrome c and dATP, the formation of an ϳ1.4-MDa apoptosome complex has been described (17,18). This large complex recruits and activates procaspase-9, which dissociates from the apoptosome before activating caspase-3 (17,18). In contrast, we have isolated an ϳ700-kDa Apaf-1 apoptosome (aposome) 2 complex containing activated caspase-9, -3, and -7 from dATPactivated cell lysates (19). The difference in size of the recombinant (ϳ1.4 MDa) and native (ϳ700 kDa) apoptosome complexes led us to question the relative significance of these two large complexes in processing caspases during dATP-dependent activation and apoptotic cell death. In the present study using Superose 6 gel filtration, we demonstrate that both the ϳ700-kDa and ϳ 1.4-MDa apoptosome complexes are formed, following in vitro dATP activation of cell lysates. However, only the ϳ700-kDa complex is capable of processing and activating effector caspases. Furthermore we also show that, during apoptotic cell death, the ϳ700 kDa apoptosome complex predominates, supporting the hypothesis that this is the correctly assembled and biologically active form of the apoptosome.

EXPERIMENTAL PROCEDURES
Reagents and Western Blot Analysis-Most reagents, including antibodies to caspase-3, -7 and -9, were obtained from indicated sources (15,21). Protein samples (ϳ20 g) were resolved on 10 or 12.5% SDSpolyacrylamide gels (PAGE) and transferred onto nitrocellulose membranes (Hybond-C Extra, Amersham Pharmacia Biotech), and antibody binding was detected as described previously (22). To assess the kinetics of apoptosome assembly, Western blots of Apaf-1 from each time point were probed with identical antibody and enhanced chemiluminescence solutions and simultaneously exposed to the same photographic film. The exposed films were scanned with a Molecular Dynamics Densitometer, and the images were analyzed with a Multi-Analyst Image analysis program (Bio-Rad Laboratories, CA).
Cell Culture and Preparation of Control and Caspase-activated Cell Lysates-Human monocytic tumor cells (THP.1) were grown in RPMI 1640 media with 10% heat-inactivated fetal bovine serum in 5% CO 2 at 37°C. Apoptosis was induced with etoposide (10 M), a DNA topoisomerase II inhibitor, or TPCK (75 M), a chymotrypsin-like serine protease inhibitor, as described previously (22,23). Lysates (100,000 ϫ g supernatants) from THP.1 cells were prepared as described previously and activated by incubation (10 mg/ml) for 1 h at 37°C with 2 mM dATP and 2 mM MgCl 2 (19). These cell lysates contained low but detectable levels of cytochrome c and did not require the addition of exogenous cytochrome c for full caspase activation.
Chromatographic Methods-Lysates were fractionated by size-exclusion chromatography, using a fast protein liquid chromatography protein purification system on analytical (16/60), preparative (26/60) Hi-Prep S300 Sephacryl high-resolution (19) and Superose 6 HR 10/30 columns (Amersham Pharmacia Biotech, Herts, UK). The Sephacryl columns were eluted at 4°C with 5% (w/v) sucrose, 0.1% (w/v) CHAPS, 20 mM HEPES/NaOH, 5 mM dithiothreitol, pH 7.0, and fractions collected. The Superose 6 column was eluted with the same buffer supplemented with 50 mM NaCl, at 0.4 ml/min, and 0.5-or 1.0-ml fractions were collected. The columns were calibrated with protein standards (Amersham Pharmacia Biotech), including blue dextran, thyroglobulin, ferritin, catalase, bovine serum albumin, ovalbumin, and bovine heart cytochrome c (Sigma) with some of their sizes (kDa) indicated in Fig. 1. Some differences (see Results) were noted in the elution profiles of the effector caspases on S300 and Superose 6 columns. This was most probably because of addition of 50 mM NaCl in the buffer, which ensured good separation of large protein complexes on Superose 6 columns.
Assay of Apoptosome Effector Caspase Activating Activity-To assess the biological activities of the ϳ1.4-MDa and ϳ700-kDa apoptosome complexes, we developed a simple in vitro assay. Briefly, apoptosome complexes were isolated from apoptotic or dATP-treated lysates by Superose 6 gel filtration as described above. The appropriate column fractions (100 l aliquots) were then incubated with procaspases (25 g of protein) for 30 min before assaying for DEVDase activity. The procaspases were obtained following fractionation of control lysates (fractions 18 -20) by gel filtration on a preparative (26/60) Hi-Prep S300 Sephacryl high-resolution column. These fractions, which contained procaspase-3, -7, and -9, were concentrated in Vivaspin 4 (Vivascience, Lincoln, UK) concentrators (10 kDa cut-off) to a final concentration of 8 -14 mg/ml and stored at Ϫ80°C prior to use. In preliminary experiments, it was determined that the isolated active apoptosome complexes required neither dATP/ATP nor cytochrome c to process added procaspases, presumably because these cofactors are only required for the proper oligomerization of Apaf-1. Thus, this assay measures the ability of an assembled Apaf-1 complex to both process and activate effector caspase-3 and -7.
The ϳ1.4-MDa and ϳ700-kDa apoptosome complexes were also assayed for their ability to activate exogenous effector caspases. Only the ϳ700-kDa apoptosome complex (Fig. 1C, fractions 10 -15) actively processed the effector caspases as assayed by the induced DEVDase activity. Thus, following dATP activation, two large apoptosome complexes of ϳ700 kDa and ϳ1.4 MDa were formed, but only the ϳ700-kDa complex Lysates (10 mg/ml) from THP.1 cells were incubated with or without dATP (2 mM) for 1 h, and the lysate (2 mg) was separated on a Superose 6 column and immunoblotted. A, Apaf-1 from control lysates eluted as a monomer and caspase-3 and -9 as unprocessed zymogens. B, after dATP activation, Apaf-1 eluted both as an ϳ1.4-MDa and an ϳ700-kDa apoptosome complex. Almost, all of the caspase-9 was processed to its p35 and 37 forms, which eluted with the ϳ1.4-MDa and the ϳ700-kDa apoptosome complexes, and as free heterotetramers. Caspase-3 was processed to its active large subunit. C, the ϳ700-kDa but not the ϳ1.4-MDa apoptosome possessed the ability to process and activate most of the effector caspases. The Apaf-1-containing fractions (fractions 5-21, Fig. 1B) were assayed for caspase processing activity by measuring DEVDase activity (q-q) produced when the fractions were incubated with "free caspases" as described under "Experimental Procedures." Thus, the ϳ700-kDa apoptosome complex (fractions 10 -15) initially had very little or no DEVDase activity but had the inherent ability to activate exogenous effector caspases. All the fractions from panel B were also initially assayed for DEVDase activity (E-E). The NaCl in the elution buffer resulted in the DEVDase (i.e. free active effector caspase-3 and -7) activity eluting in fractions 22-28. The elution positions of M r markers are indicated with arrows.
Biologically Active and Inactive Apaf-1 Complexes 6068 appeared to constitute the correctly assembled and functionally active apoptosome.
Kinetics of Formation of the ϳ700-kDa and ϳ1.4-MDa Apoptosome Complexes-To understand the relationship between the ϳ700-kDa and ϳ1.4-MDa apoptosome complexes, we studied their time-dependent formation and activities following dATP activation. The ϳ700-kDa complex was more rapidly formed than the ϳ1.4-MDa complex ( Fig. 2A). Production of the ϳ700-kDa complex was maximal within 5 min of initiating dATP activation, whereas assembly of the ϳ1.4-MDa complex did not reach a maximum until after 20 min (Fig. 2A). However, only the ϳ700-kDa apoptosome complex exhibited significant biological activity as assessed by its ability to process caspases to yield active effector caspase activity (DEVDase, Fig. 2B). The biological activity of the assembled ϳ700-kDa complex was almost maximal 5 min after activation was initiated, whereas the ϳ1.4-MDa complex was virtually inactive at all time points measured (Fig. 2B).
During Apoptosis, the ϳ700-kDa Apoptosome Complex Predominates-To determine whether the ϳ700-kDa and/or ϳ1.4-MDa apoptosome complexes were formed during apoptosis, lysates were prepared from control and apoptotic cells and fractionated using a Superose 6 column. Apoptosis was induced in THP.1 cells by two different stimuli, i.e. etoposide and TPCK (22,23), which gave 30 and 38% apoptosis, respectively, as determined by Annexin V binding. All of the Apaf-1 in lysates isolated from control cells eluted as the free ϳ130-kDa monomeric form (Fig. 3A). In marked contrast, Apaf-1, in lysates from etoposide or TPCK-treated THP.1 cells, eluted mainly as the free monomeric form or as the ϳ700-kDa complex, and little if any Apaf-1 eluted as an ϳ1.4-MDa complex (Fig. 3, B and C). Furthermore, only the ϳ700-kDa apoptosome complex processed and activated effector caspases (Fig. 3, B and C), suggesting that this was the biologically active complex which was formed during etoposide-or TPCK-induced apoptosis. Thus, in THP.1 cells, induction of apoptosis by either etoposide or TPCK resulted in the formation of an ϳ700-kDa caspase processing complex. DISCUSSION Reconstitution studies with recombinant proteins have shown that Apaf-1 and caspase-9 in the presence of cytochrome c and dATP forms a very large (ϳ1.4 MDa) apoptosome complex (17,18). This complex was active as judged by its ability to process 35 S-labeled procaspase-9 and -3. However, in the current study we have demonstrated that, during dATP-dependent activation of cell lysates, two large apoptosome complexes with M r values of ϳ700 kDa and ϳ1.4 MDa were formed (Figs. 1 and 2). The rate of formation of the ϳ700-kDa complex was very rapid (within 5-10 min) compared with the formation of the ϳ1.4-MDa complex (ϳ20 min). Moreover, the ϳ700-kDa complex efficiently processed exogenous caspases, whereas the ϳ1.4-MDa complex was relatively inactive (Figs. 1 and 2). Why the ϳ1.4-MDa apoptosome complex was incapable of caspase processing is unclear because both complexes contained oli-FIG. 2. The ϳ700-kDa but not the ϳ1.4-MDa apoptosome complex possesses caspase processing activity. Lysates were activated with dATP as described in the legend to Fig. 1 and at the indicated times fractionated by Superose 6 chromatography, essentially as described under "Experimental Procedures" except that 1-ml fractions were taken. Each fraction was assayed for its caspase processing activity and then subjected to SDS-PAGE. A, the fractions were then analyzed by Western blotting with an antibody to Apaf-1 and quantified by densitometry as described under "Experimental Procedures." The timedependent formation of the ϳ1.4-MDa (-q-) and ϳ700-kDa (-f-) apoptosome complexes is shown. The imaged Apaf-1 data is shown for one experiment (similar results were obtained in a second separate experiment). B, caspase processing activity for the ϳ700-kDa and ϳ1.4-MDa complexes is shown. This is the average of two separate time course experiments.  (21). Lysates were then freshly prepared and fractionated by Superose 6 gel filtration as described in the legend to Fig. 1. The fractions were run on SDS-PAGE and immunoblotted for Apaf-1 and assayed for caspase processing activity. In both etoposide (B) and TPCK-treated cells (C), Apaf-1 eluted predominantly either as the monomeric form or as a ϳ700-kDa apoptosome complex. Almost all the caspase processing activity was associated with the ϳ700-kDa apoptosome complex. Control cell lysates did not have any detectable caspase processing activity (data not shown).
gomerized Apaf-1 and processed caspase-9 (Fig. 1). We cannot rule out the possibility that an inhibitor is associated with the larger complex or alternatively that the ϳ1.4-MDa complex may be inappropriately oligomerized and structurally different from the ϳ1.4-MDa complex previously described (17,18). Nevertheless, the ϳ700-kDa complex does not appear to be formed when reconstitution of the apoptosome is carried out with recombinant Apaf-1, procaspase-9, cytochrome c, and dATP (17,18).
The biological significance of two large Apaf-1 apoptosome complexes is as yet unclear. However, our experiments with apoptotic lysates provide important evidence which indicates that the ϳ700-kDa complex is the biologically significant complex that is formed during apoptosis. Thus, in THP.1 cells using two unrelated stimuli, i.e. a DNA topoisomerase II inhibitor (etoposide) and a serine chymotrypsin-like protease inhibitor (TPCK), we have demonstrated that the predominant complex formed is an ϳ700-kDa apoptosome complex (Fig. 3). Taken together our results suggest that the ϳ700-kDa complex is the major form of the apoptosome formed in vivo and that the correct assembly of the apoptosome may require other factors present in cell lysates, in addition to Apaf-1, cytochrome c, dATP, and caspase-9. Thus, clear differences exist in the apoptosome complexes formed in vitro from recombinant proteins when compared with those assembled in cell lysates and during apoptosis.
In summary, following dATP activation of lysates, two Apaf-1-containing complexes are formed, a biologically inactive ϳ1.4-MDa and a biologically active ϳ700-kDa apoptosome. Kinetic analysis demonstrates a more rapid formation of the smaller ϳ700-kDa complex, which exhibits significant caspaseactivating activity. In apoptotic cells, the ϳ700-kDa complex predominates, and we therefore propose that the ϳ700-kDa complex is the functionally active form of the apoptosome in apoptosis.