Protein Kinase A Regulates Caspase-9 Activation by Apaf-1 Downstream of Cytochrome c*

The cyclic AMP signal transduction pathway modulates apoptosis in diverse cell types, although the mechanism is poorly understood. A critical component of the intrinsic apoptotic pathway is caspase-9, which is activated by Apaf-1 in the apoptosome, a large complex assembled in response to release of cytochrome c from mitochondria. Caspase-9 cleaves and activates effector caspases, predominantly caspase-3, resulting in the demise of the cell. Here we identified a distinct mechanism by which cyclic AMP regulates this apoptotic pathway through activation of protein kinase A. We show that protein kinase A inhibits activation of caspase-9 and caspase-3 downstream of cytochrome c in Xenopus egg extracts and in a human cell-free system. Protein kinase A directly phosphorylates human caspase-9 at serines 99, 183, and 195. However, mutational analysis demonstrated that phosphorylation at these sites is not required for the inhibitory effect of protein kinase A on caspase-9 activation. Importantly, protein kinase A inhibits cytochrome c-dependent recruitment of procaspase-9 to Apaf-1 but not activation of caspase-9 by a constitutively activated form of Apaf-1. These data indicate that extracellular signals that elevate cyclic AMP and activate protein kinase A may suppress apoptosis by inhibiting apoptosome formation downstream of cytochrome c release from mitochondria.

Apoptosis is a physiological form of cell death that is fundamental for cell selection during development and tissue homeostasis in multicellular organisms (1). Apoptosis is important for the removal of cells damaged by cellular stresses or those that have undergone oncogenic transformation (2). The apoptotic program involves activation of a group of cysteine proteases termed caspases, which are present in cells as inactive or low activity proenzymes (3). Initiation of an intrinsic apoptotic pathway by a wide variety of stimuli causes the release of cytochrome c from mitochondria (4 -6). In the cytosol, cytochrome c binds the apoptotic factor Apaf-1 and induces Apaf-1 oligomerization, leading to recruitment and activation of procaspase-9 in a large complex termed the apoptosome (7,8). As a result, procaspase-9 autoprocesses and cleaves the effector procaspases 3 and 7, which are activated to cleave key struc-tural and regulatory proteins and thereby bring about the biochemical and morphological changes associated with apoptotic cell death (3). Signal transduction pathways activated by extracellular and intracellular stimuli can impinge on this apoptotic pathway to control cell fate.
The cyclic AMP signal transduction pathway regulates a diverse array of cellular processes, including proliferation, differentiation, and secretion. Activation of this signaling pathway occurs through specific ligation of G-protein-coupled receptors, initiating formation of cAMP from ATP through the action of adenylate cyclase. Cyclic AMP binds to the regulatory subunits of the heterotetrameric cAMP-dependent protein kinase or protein kinase A (PKA), 1 dissociating the holoenzyme and releasing the free catalytic (C) subunits. The active C subunits subsequently phosphorylate target proteins on serine or threonine residues within a relatively well defined consensus sequence (9,10). In addition, cAMP may act independently of PKA through Epac, a guanine-nucleotide exchange factor for the Rap1 GTPase that is directly activated by cAMP (11). Recently, it has emerged that one important facet of cAMP signal transduction is to regulate apoptosis. While elevation of cAMP is associated with induction of apoptosis in lymphoid cells by glucocorticoids (12), in many cell types elevated cAMP protects against apoptosis (13)(14)(15)(16)(17)(18)(19)(20)(21). In contrast to the well characterized molecular mechanisms by which cAMP regulates other biological processes, little is known about the mechanism by which this ubiquitous second messenger modulates apoptosis.
The mitochondrial apoptotic pathway can be regulated at multiple stages to promote or suppress cell death (22). One important point of control is the release of cytochrome c from mitochondria. This event is regulated by the pro-and antiapoptotic proteins of the Bcl-2 family (4,6). In addition, this apoptotic pathway can be regulated downstream of cytochrome c release by caspase inhibitor proteins such as XIAP (23). The activity of XIAP may be controlled by release of other factors such as Smac/Diablo from mitochondria (24). Components of the pathway such as caspase-9 (25,26) and XIAP (27) are also regulated post-translationally through phosphorylation by protein kinases activated by signaling pathways. Abnormal or constitutive activation of these signaling pathways may contribute to the survival of cancer cells despite initiation of upstream apoptotic responses. Nevertheless, the mechanisms by which signaling pathways reg-ulate caspase activation downstream of cytochrome c remain to be fully characterized.
Cell-free systems that faithfully reproduce the regulation of apoptosis have proved to be useful for dissecting the biochemical mechanisms controlling caspase activation, including regulation by signaling pathways (28). In this study, we have investigated the ability of the cAMP signaling pathway to regulate caspase activation using cell-free systems derived from Xenopus eggs and mammalian cultured cells. We show that activation of PKA by cAMP blocks caspase-9 activation downstream of cytochrome c. Caspase-9 is phosphorylated directly by PKA, although mutational analysis shows that this phosphorylation is not required to inhibit caspase-9 activation. We demonstrate that PKA inhibits the recruitment of procaspase-9 to Apaf-1 in response to cytochrome c, indicating that PKA controls apoptosome formation by a novel mechanism.

Plasmid Constructs and Site-directed Mutagenesis-Human
Caspase-9 and Apaf-1-(1-541) cDNAs were amplified from U2OS cells and cloned into pcDNA3 (Invitrogen, Carlsbad, CA). Codons for specific residues in caspase-9 were replaced with those encoding alanine or glutamate using the QuikChange TM site-directed mutagenesis kit (Stratagene, Cedar Creek, TX), and the mutations were verified by sequence analysis. For subsequent expression of caspase-9 proteins in bacteria, cDNAs were used that also encoded alanine at residue 287 instead of the cysteine required for catalytic activity. These cDNAs were subcloned into pGEX-4T-1 (Amersham Biosciences), and expression of recombinant proteins was induced in Escherichia coli BLR(DE3) at 30°C for 2 h by the addition of 1 mM isopropyl 1-thio-␤-D-galactopyranoside. Glutathione S-transferase (GST)-tagged proteins were affinity-purified with glutathione-Sepharose 4B (Amersham Biosciences), eluting with 15 mM glutathione in buffer A (10 mM Hepes-KOH at pH 7.5, 150 mM NaCl, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, and 1 g ml Ϫ1 each of aprotinin, leupeptin, and pepstatin A). Glutathione was removed by filtration through a PD-10 column (Amersham Biosciences). The caspase recruitment domain (CARD) of mouse Apaf-1 (cDNA generously provided by Dr. J. Silke, Parkville, Australia) was cloned into pET28a and expressed as a His 6 -tagged protein which was purified on a nickel-agarose column (Qiagen, Hilden, Germany). The protein was eluted with 250 mM imidazole in buffer B (10 mM Hepes-KOH at pH 7.5, 150 mM NaCl, 0.07% 2-mercaptoethanol), and then imidazole was removed using a PD-10 column.
Antibodies-Polyclonal rabbit anti-caspase-9 (Pharmingen, San Diego, CA), monoclonal mouse anti-caspase-9 (Santa Cruz Biotechnology, Santa Cruz, CA), and monoclonal mouse anti-Apaf-1 antibody (Pharmingen) were purchased from commercial suppliers. Secondary antibodies for Western blots were goat anti-rabbit or anti-mouse IgG coupled to horseradish peroxidase (Bio-Rad). A sheep polyclonal antibody was generated against His 6 -tagged caspase-9 C287A protein and affinity-purified on a GST-caspase-9 C287A column. A rabbit polyclonal phosphospecific antibody, pS183-Casp-9, was raised against a phosphopeptide derived from caspase-9, LRTRTGS*NIDCEK, where S* represents phosphoserine (Moravian Biotech, Brno, Czech Republic). This antibody was affinity-purified by two rounds of negative selections against nonphosphorylated peptide followed by one round of positive selection with phosphorylated peptide.
Measurement of Caspase Activation-Caspase activation was induced by incubating 200 g of HeLa S100 extract together with 1 mM ATP, 10 g ml Ϫ1 creatine kinase, 5 mM creatine phosphate (Roche Diagnostics), and 10 M bovine heart cytochrome c (Sigma) at 30°C for the times indicated. HEK293 extracts were incubated with the same additions except that 200 nM cytochrome c was used. Caspase activation in Xenopus egg extracts was performed as described previously (29). Other reagents were used at the following concentrations: 0.48 unit l Ϫ1 PKA catalytic subunit (Promega, Southampton, UK), 10 M H89, 300 M dibutyryl cAMP, and 300 M 8-Br-cAMP (Calbiochem). For analysis of caspase-9 processing, 25-g samples were analyzed by SDS-PAGE and Western blotting using chemiluminescence detection. For measurement of caspase-3 activity, samples were incubated with Ac-DEVD-AMC (Calbiochem), and the release of free AMC was measured using a fluorescence microtiter plate reader as described (26).
Immunoprecipitation of Apoptosomes-Apoptosome formation was stimulated by incubating 1.5 mg of HeLa S100 extract together with 1 mM ATP, 10 g ml Ϫ1 creatine kinase, 5 mM creatine phosphate, and 10 M cytochrome c for 4 h at 30°C. Where indicated, 10 M H89 and 0.24 unit l Ϫ1 PKA catalytic subunit were added. The sample was incubated with 5 l of polyclonal sheep anti-caspase-9 antibody for 1 h at 4°C. 20 l of washed protein A beads (Sigma) were then added and incubated for 1.5 h at 4°C with rotation. Beads were then pelleted by centrifugation and washed three times in buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 2 mM EDTA, 1 mM Na 3 VO 4 , 50 mM NaF, and 5 mM ␤-glycerophosphate) before boiling in reducing SDS-PAGE sample buffer. Proteins were separated on SDS-PAGE and analyzed by Western blotting.
Phosphorylation of Recombinant GST-Caspase-9 and GST-Apaf-1-(1-541)-GST-caspase-9 proteins (all C287A) were incubated for the times indicated at 30°C with 0.48 unit l Ϫ1 PKA catalytic subunit in kinase buffer (50 mM Tris-HCl at pH 7.5, 10 mM MgCl 2 , and 1 mM dithiothreitol) containing 100 M [␥-32 P]-ATP (specific activity 6 ϫ 10 6 cpm nmol Ϫ1 ). GST-Apaf-1-(1-541) was phosphorylated under the same conditions, except that the incubation was for 90 min. The reaction was terminated by the addition of reducing SDS-PAGE loading buffer. Samples were then subjected to SDS-PAGE followed by autoradiography. For identification of phosphorylation sites, 10 g of GST-caspase-9 was incubated in a total volume of 30 l for 90 min. Tryptic digestion, purification of the phosphopeptides by HPLC, mass spectrometry, and sequencing were carried out as described previously (26). For analysis by Western blotting, reactions contained 0.25 g of GST-caspase-9 and nonradioactive ATP, and blots were probed with a polyclonal rabbit anti-caspase-9 phosphospecific antibody, pS183-Casp-9.
In Vitro Translation and Assay of Caspase-9 Activity-In vitro translation of caspase-9 cDNAs (which encoded the active site cysteine at residue 287) was performed according to the TNT Quick-coupled transcription/translation protocol (Promega). For induction of caspase-9 processing, 1 l of in vitro translated caspase-9 labeled with [ 35 S]methionine was incubated at 30°C for 90 min with 200 nM cytochrome c together with 2 l of TNT lysate in a 10-l buffered reaction (20 mM Hepes-KOH, pH 7.5, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA, and 1 mM dithiothreitol). Alternatively, caspase-9 processing was induced using 1 l of in vitro translated caspase-9 incubated for 90 min at 30°C with 2-8 ng of Apaf-1-(1-541) or 300 ng of Apaf-1 CARD in the same reaction buffer. Caspase-9 D330A was used for the incubation with Apaf-1-(1-541). Samples were subjected to SDS-PAGE and visualized by autoradiography.
Immunodepletion of Caspase-9 -For immunodepletion of caspase-9, 40 l of washed protein A beads (Sigma) were incubated with 5 l of sheep polyclonal antibody against caspase-9 in 200 l of buffer (20 mM Hepes-KOH, pH 7.5, 10 mM KCl, and 2 mM MgCl 2 ) at 4°C overnight. Unbound antibody was removed by washing in the same buffer. Beads were then added to 100 l of HeLa S100 extract (10 mg/ml) and incubated for 2 h at 4°C with rotation. Beads were pelleted by centrifugation, and the recovered supernatant was incubated with a second aliquot of anti-caspase-9 antibody bound to protein A beads. After centrifugation, the supernatant was recovered and used as caspase-9depleted extract. For analysis of caspase-3 activation by caspase-9, HeLa cytosolic extract depleted of endogenous caspase-9 was reconstituted with in vitro translated caspase-9 and cytochrome c.

Activation of Caspase-9 and Caspase-3 Is Inhibited by cAMP
Analogues through PKA-To investigate the role of the cAMP signal transduction cascade in apoptotic regulation, we sought to determine whether this signaling pathway modulated activation of the caspase-9/caspase-3 pathway. We made use of cell-free systems that allow dissection of the biochemical molecular mechanisms by which signal transduction cascades may impinge on this pathway. Xenopus egg extracts supplemented with an ATP-regenerating system undergo caspase activation upon prolonged incubation as cytochrome c is released from mitochondria (6,29). The addition of a stable cAMP analogue, 8-Br-cAMP, to egg extracts at the start of the incubation completely blocked caspase-3 activation detected by cleavage of a fluorogenic substrate, AcDEVD-AMC (Fig. 1A). When egg extracts were fractionated to yield a post-mitochon-drial supernatant and a pellet containing mitochondria, 8-Br-cAMP inhibited the activation of caspase-3 in the supernatant in response to addition of the mitochondrial pellet or exogenous cytochrome c (Fig. 1A). 8-Br-cAMP also blocked upstream caspase-9 processing induced by cytochrome c in the supernatant (Fig. 1B).
Similarly, in human somatic cell (HeLa) extracts lacking mitochondria, caspase-3 activation (Fig. 2, A and B) and caspase-9 processing (Fig. 2C) induced by the addition of exogenous cytochrome c was inhibited by co-incubation with 8-Br-cAMP or another cAMP analogue, dibutryl cAMP (Bt 2 cAMP). Fig. 2C shows that Bt 2 cAMP inhibited both autoprocessing of caspase-9 from the proenzyme to a p35 form (a direct measure of caspase-9 activation) and processing to a p37 form that is produced by caspase-3, which is activated downstream of caspase-9 (30). Inhibition of both caspase-3 activation and caspase-9 activation by cAMP analogues was reversed by the pharmacological PKA inhibitor, H89. Consistent with a role for PKA in the regulation of this apoptotic pathway, we found incubation of HEK293 (Fig. 3A) or HeLa (Fig. 3, B and C) cytosolic extracts with active PKA catalytic subunit strongly inhibited activation of both caspase-3 (Fig. 3, A and B) and caspase-9 (Fig. 3C) in response to cytochrome c. These inhibitory effects were substantially reversed by H89. These results show that cAMP signaling inhibits the caspase-9/caspase-3 apoptotic pathway at the stage of caspase-9 activation and that this inhibitory effect is mediated through PKA.
Purified Caspase-9 Is Phosphorylated by PKA at Three Sites-Phosphorylation of caspase-9 is a mechanism by which survival pathways may inhibit caspase-9 activation (25,26). To test whether PKA could target caspase-9 directly, we incubated recombinant caspase-9 expressed as a GST fusion protein with PKA and [␥-32 P]ATP. We found that GST-caspase-9 was phosphorylated with a stoichiometry of 0.52 mol phosphate/mol protein in this reaction (Fig. 4A). To identify the phosphorylation sites, 32 P-labeled GST-caspase-9 was digested with trypsin, and the resulting phosphopeptides were purified by HPLC (Fig. 4B). Four major phosphopeptides (P1-P4) were obtained; these were analyzed by mass spectrometry and sequencing. This analysis revealed that P1 (M r 2653.50) corresponded to amino acids 94 -116 of caspase-9 (Fig. 4C)  To confirm that Ser-99, Ser-183, and Ser-195 were the major phosphorylation sites in caspase-9 targeted by PKA, we used site-directed mutagenesis to generate recombinant GSTcaspase-9 in which these serines were mutated to nonphosphorylatable alanines. Phosphorylation was abolished in GSTcaspase-9 S99A/S183A/S195A (Fig. 4D), confirming that these were indeed the only sites targeted directly by PKA.
Phosphorylation of Caspase-9 at Ser-183 in Cell Extracts-To test whether caspase-9 was phosphorylated in response to PKA activation in cell extracts, we raised an antibody against one phosphorylated site, Ser-183. The purified antibody recognized caspase-9 only when phosphorylated at this site, since GSTcaspase-9 S183A was not recognized even after incubation with PKA (Fig. 5A). Using this phosphospecific antibody, we showed that caspase-9 was indeed phosphorylated at Ser-183 in HeLa cytosolic extracts in response to the addition of the active catalytic subunit of PKA or the cAMP analogues 8-Br-cAMP and Bt 2 cAMP (Fig. 5B). However, phosphorylation of Ser-183 in response to cAMP analogues was rather weak compared with the addition of the active catalytic subunit, indicating a relatively low stoichiometry of phosphorylation. Using the pSer-183 antibody, we were unable to detect significant phosphorylation of this site in cells in response to elevation of cAMP (data not shown), possibly because the antibody was insufficiently sensitive or because this site was not phosphorylated strongly by PKA in vivo.
PKA Inhibits Caspase-9 Activation Independently of Direct Phosphorylation-To test whether phosphorylation of caspase-9 is the principal mechanism by which PKA controls activation of the protease in cell extracts, we examined the ability of PKA to regulate caspase-9 activation and processing when Ser-183, Ser-99, and Ser-195 were mutated to nonphosphorylatable alanine residues. As found previously, PKA strongly inhibited cytochrome c-induced processing of in vitro translated wild-type caspase-9. Strikingly, however, mutation of all three sites to alanine did not prevent the inhibition of caspase-9 processing by PKA in reticulocyte lysate (Fig. 6A). Using HeLa cytosolic extracts, we immunodepleted endogenous caspase-9 and abrogated activation of caspase-3 by cytochrome c (Fig. 6, B and C). Caspase-3 activation was restored in the depleted extract by the addition of in vitro translated wildtype caspase-9 or mutated caspase-9 S99A/S183A/S195A , showing that both proteins were functional. As expected, caspase-3 activation by wild-type caspase-9 was strongly inhibited by PKA. Importantly, PKA also suppressed caspase-3 activation by caspase-9 S99A/S183A/S195A to a similar extent (Fig. 6B). Analysis of caspase-9 processing confirmed that activation of caspase-9 S99A/S183A/S195A , like wild-type caspase-9, was inhibited by PKA (Fig. 6C). These results demonstrate that PKA can inhibit caspase-9 activation by a mechanism that is independent of direct caspase-9 phosphorylation.
PKA Inhibits Apoptosome Formation-Cytochrome c induces the ATP/dATP-dependent assembly of Apaf-1 and the recruitment of procaspase-9 into a large oligomeric complex, the apoptosome (7,8). PKA could therefore potentially inhibit caspase-9 activation by regulating the interaction of caspase-9 with Apaf-1. To study the formation of native apoptosome complexes, we used immunoprecipitation of endogenous caspase-9 with detection of caspase-9 and Apaf-1 by Western blotting of the immunoprecipitates (31). We confirmed that incubation of HeLa cytosolic extracts with cytochrome c induced the association of Apaf-1 with caspase-9 in the presence of ATP (Fig. 7A). Pre-incubation with active PKA or Bt 2 cAMP strongly reduced the association between caspase-9 and Apaf-1 and also blocked the activation of caspase-9 detected by its processing. These inhibitory effects were reversed by the PKA inhibitor H89 (Fig. 7B). This demonstrates that PKA inhibits the assembly of the apopto- some in response to cytochrome c, thereby preventing the subsequent activation of caspase-9 and downstream activation of caspase-3.
Cytochrome c is thought to induce the exposure of the caspase recruitment domain of Apaf-1 that interacts with caspase-9 (3). We found that processing of caspase-9 or caspase-9 S99A/S183A/S195A induced by Apaf-1 CARD (32) was not inhibited by PKA, in striking contrast to the effect of PKA on cytochrome c-induced processing (Fig. 8A). These data point to a mechanism by which PKA inhibits Apaf-1 activation and exposure of the CARD, thereby preventing the subsequent recruit-ment of caspase-9 to the apoptosome independently of caspase-9 phosphorylation. To test further whether PKA acts at the stage of Apaf-1 oligomerization and activation, we examined the ability of PKA to affect caspase-9 processing induced by Apaf-1-(1-541), a constitutively activated form. Apaf-1-(1-541) induces caspase-9 processing to the p35 form, whereas cytochrome c results in formation of both p35 and p37 forms because of downstream caspase-3 activity (30,33). To ensure that comparable activation of caspase-9 was being induced by these two stimuli, we used caspase-9 D330A , which cannot be cleaved by caspase-3 to the p37 form, and titrated the amount of Apaf-1-(1-541) that was added. We found that PKA had no effect on caspase-9 processing induced by Apaf-1-(1-541), even when the extent of caspase-9 processing was similar to that induced by cytochrome c. This indicates that PKA cannot inhibit Apaf-1 function once Apaf-1 has been oligomerized and activated in response to cytochrome c. One possible mechanism is through phosphorylation of Apaf-1 by PKA; indeed, we found that purified GST-Apaf-1-(1-541) was phosphorylated directly when incubated with PKA (Fig. 8C). DISCUSSION The ubiquitous second messenger cyclic AMP regulates apoptosis in a diverse range of cells, but the mechanism by which this occurs has remained elusive. In many cell types, elevation of cAMP is associated with prevention of apoptosis. This report provides evidence that cAMP acts through PKA to inhibit a major apoptotic pathway at a point downstream of cytochrome c release from mitochondria. We have shown that PKA inhibits caspase-9 activation by Apaf-1 in response to cytochrome c. PKA phosphorylates purified caspase-9 at three sites, although PKA can strongly inhibit caspase-9 activation independently of caspase-9 phosphorylation.
Much of the previous focus of research on the regulation of apoptosis has been on the expression or activity of proteins of the Bcl-2 family, which can control the release of cytochrome c from mitochondria into the cytosol. The cAMP signal transduction pathway has been suggested to regulate apoptosis at this by an acetonitrile gradient up to 100% on HPLC. C, identity of phosphopeptides corresponding to P1-P4 (underlined) and the position of phosphorylated serine residues (bold, numbered) in caspase-9. D, phosphorylation of GST-caspase-9 (WT, wild type) or mutated GST-caspase-9 lacking the three sites of phosphorylation (S99A/S183A/S195A) by PKA catalytic subunit and [␥-32 P]ATP, analyzed by SDS-PAGE and autoradiography.
FIG. 5. Phosphorylation of caspase-9 at serine 183 by PKA. A, recognition of GST-caspase-9 (WT, wild type) but not GST-caspase-9 S183A (S183A) after incubation with PKA by a rabbit polyclonal antibody raised against a caspase-9 phosphopeptide containing phospho-Ser-183 (pS183-Casp-9). Total caspase-9 detected by a sheep anticaspase-9 antibody is also shown. B, phosphorylation of caspase-9 at Ser-183 in HeLa S100 extract supplemented with active PKA, dibutryl cAMP (db-cAMP), or 8-Br-cAMP. Detection of GST-caspase-9 (WT) and caspase-9 S183A (S183A) is compared. stage by targeting the pro-apoptotic Bcl-2 family protein Bad, which is phosphorylated on certain serine residues by PKA. These phosphorylated residues cause sequestration of Bad by 14-3-3, with the resultant dissociation of Bad from its antiapoptotic partner, Bcl-x L (34 -38). However, expression of Bad is restricted to certain cell types, and Bad-deficient mice appear to display no developmental abnormality. In most cell types, cell death by growth factor withdrawal proceeds normally in the absence of Bad. Thus, Bad cannot account exclusively for the induction of apoptosis in response to loss of extracellular signals such as those operating through elevation of cAMP (39,40).
Our results demonstrate that the cAMP signaling pathway acting through PKA can directly inhibit the mitochondrial apoptotic pathway downstream of cytochrome c. This role appears to be conserved in vertebrates. In Xenopus, PKA also plays roles in the cell division cycle during early embryonic development when it is periodically activated (41) and may play a role in the timing of M-phase through inhibition of Cdc25 phosphatase (42). Direct suppression of caspase-9 activation by FIG. 6. Inhibition of caspase-9 activation by PKA does not require direct phosphorylation of caspase-9. A, caspase-9 (WT, wild type) or caspase-9 S99A/S183A/S195A (AAA) translated in vitro and incubated with active PKA catalytic subunit before addition of cytochrome c. Samples were analyzed for caspase-9 processing by SDS-PAGE and autoradiography. B, HeLa cytosolic extracts depleted of caspase-9 incubated with in vitro translated wild-type caspase-9 (WT) or caspase-9 S99A/S183A/S195A (AAA) and PKA catalytic subunit as shown. Control incubations lacked cytochrome c or were without in vitro translated caspase-9 (w/o Casp-9). Caspase-3 activity induced by cytochrome c addition was assayed by measuring cleavage of DEVD-AMC. C, shows caspase-9 detected by Western blot in each of the samples assayed in B compared with S100 HeLa cytosolic extract prior to depletion. FIG. 7. PKA inhibits apoptosome assembly. A, HeLa cytosolic extract was incubated with or without cytochrome c. Samples were removed for Western blotting using rabbit polyclonal antibodies against Apaf-1 and caspase-9 (input). Caspase-9 was immunoprecipitated from the remainder of the incubation using a sheep polyclonal antibody to caspase-9, and immunoprecipitates (caspase-9 IP) were immunoblotted with antibodies to Apaf-1 or caspase-9 as indicated. A nonspecific band in the immunoprecipitates detected by the caspase-9 antibody migrates above processed caspase-9 at about 40 kDa. B, caspase-9 immunoprecipitates from incubations containing cytochrome c, H89, PKA, and dibutyryl cAMP (db-cAMP) as indicated, analyzed by Western blotting with antibodies to Apaf-1 or caspase-9.  -1-(1-541) or Apaf-1 CARD. A, caspase-9 (WT, wild type) or caspase-9 S99A/S183A/S195A (AAA) translated in vitro and incubated with active PKA catalytic subunit and cytochrome c or Apaf-1 CARD to induce caspase activation, analyzed by SDS-PAGE and autoradiography. B, caspase-9 D330A translated in vitro and incubated with active PKA catalytic subunit and cytochrome c or differing amounts of GST-Apaf-1-(1-541) to induce caspase activation, analyzed by SDS-PAGE and autoradiography. C, phosphorylation of GST-Apaf-1-(1-541) by PKA using [␥-32 P]ATP, detected by autoradiography. PKA could therefore couple control of apoptosis with cell cycle progression at this stage of development. More generally, activation of PKA by cAMP elevation in response to extracellular signals may play an important role in the regulation of caspase-9 activation in somatic cells.
We have suggested previously that phosphorylation of caspase-9 at Thr-125 by ERK MAPK inhibits activation of the proenzyme through inhibition of a conformational change induced by binding to Apaf-1 that activates the catalytic activity of procaspase-9 (26). Inhibition of caspase-9 activation by PKA might therefore operate through a similar mechanism in which PKA would directly phosphorylate the proenzyme and prevent its activation. However, although we found that PKA does indeed phosphorylate purified caspase-9 at three serine residues, we also found that mutation of these residues to nonphosphorylatable alanine did not abolish the inhibition of caspase-9 activation by PKA in cell extracts. By contrast, mutation of Thr-125 significantly abrogates the inhibitory effect of ERK MAPK on caspase-9 activation (26). Although it remains possible that phosphorylation of one or more of the sites in caspase-9 that are targeted by PKA could play a role in its regulation, it is clear that PKA can operate through a distinct mechanism to prevent caspase-9 activation.
Importantly, we have found that PKA strongly inhibits the recruitment of procaspase-9 to Apaf-1 in response to cytochrome c. This step involves the oligomerization of Apaf-1 to form a large complex and then recruitment of procaspase-9 through interactions between the CARD domain of Apaf-1 and the structurally related prodomain of caspase-9. The inability of PKA to inhibit caspase-9 processing by constitutively activated Apaf-1-(1-541) or Apaf-1 CARD suggests that PKA works prior to Apaf-1 oligomerization and activation, which exposes the CARD domain in native Apaf-1.
One obvious candidate substrate for PKA is Apaf-1 itself, and we have found that Apaf-1 is phosphorylated by PKA in vitro.
To determine the possible function of Apaf-1 phosphorylation in the response to PKA activation, it will be necessary to identify the phosphorylation sites and to test their roles. Interestingly, Kornbluth and colleagues (43) have recently also provided evidence for the control of apoptosome formation and the phosphorylation of Apaf-1 in cell extracts in response to the Bcr-Abl fusion tyrosine kinase, although the mechanism is likely to be indirect, since no tyrosine phosphorylation of Apaf-1 was detected. Alternatively, phosphorylation of an as yet unidentified regulatory factor by PKA could control the activation of Apaf-1 and the recruitment of caspase-9. Elucidation of this novel level of control of the apoptosome is likely to be important in understanding the control of apoptosis by protein kinase signaling pathways.