c-Abl tyrosine kinase regulates caspase-9 autocleavage in the apoptotic response to DNA damage.

Activation of the initiator caspase-9 is essential for induction of apoptosis by developmental signals, oncogenic transformation, and genotoxic stress. The c-Abl tyrosine kinase is also involved in the apoptotic response to DNA damage. The present results demonstrate that c-Abl binds directly to caspase-9. We show that c-Abl phosphorylates caspase-9 on Tyr-153 in vitro and in cells treated with DNA damaging agents. Moreover, inhibition of c-Abl with STI571 blocked DNA damage-induced autoprocessing of caspase-9 to the p35 subunit and activation of caspase-3. Caspase-9(Y153F) also attenuated DNA damage-induced processing of caspase-9 to p35, activation of caspase-3, and apoptosis. These findings indicate that caspase-9 autoprocessing is regulated by c-Abl in the apoptotic response to genotoxic stress.

Caspase-9 is the initiator caspase of the apoptosome, an oligomeric complex that controls the intrinsic apoptotic pathway. Formation of the apoptosome is induced by release of mitochondrial cytochrome c into the cytosol. Cytochrome c associates with Apaf-1 and thereby promotes its oligomerization and recruitment of caspase-9 (1-3). Binding to Apaf-1 increases activity of the caspase-9 protease and autocleavage of the p46 pro-caspase-9 at Asp-315 to yield p35 and p12 subunits (4 -6). Caspase-9 activation requires interaction with the Apaf-1 caspase recruitment domain, an increase in local concentrations of caspase-9, and the formation of caspase-9 dimers (7,8). Following autoprocessing in the apoptosome, caspase-9 cleaves and activates caspase-3. In turn, caspase-3 directs feedback cleavage of caspase-9 at Asp-330 to generate p37 and p10 subunits (4,9). The caspase-9 p12, and not the p10, subunit contains four N-terminal amino acids that bind to the third baculoviral repeat of the X-linked inhibitor of apoptosis (10), which maintains caspase-9 in the inactive monomer conformation (11,12) and functions as a tether for caspase-3 (13,14). Other studies have demonstrated that caspase-9 activity is inhibited by Akt-mediated phosphorylation on  and by extracellular signal-regulated kinase-mediated phosphorylation on Thr-125 (16). However, it is not known whether phosphorylation of caspase-9 contributes to autocleavage of this important apoptotic initiator.
The c-Abl tyrosine kinase is activated in the response of cells to genotoxic stress (17). The product of the gene mutated in ataxia telangiectasia is responsible in part for c-Abl activation (18,19). Other work has demonstrated that nuclear c-Abl interacts with the DNA-dependent protein kinase (DNA-PK)-Ku complex (20,21). Phosphorylation of c-Abl by the catalytic subunit DNA-PKcs stimulates c-Abl activity (20). Activation of c-Abl by DNA damage or inhibition of DNA replication contributes to the induction of apoptosis by mechanisms in part dependent on the p53 tumor suppressor and its homolog p73 (22)(23)(24)(25)(26). c-Abl also contributes to DNA damage-induced activation of the mitogen-activated protein kinase/extracellular signal-regulated kinase kinase (MEK) kinase-1, c-Jun N-terminal kinase, and p38 mitogen-activated protein kinase pathways (17,(27)(28)(29). Moreover, c-Abl interactions with Rad51, Rad9, and the hTERT telomerase catalytic subunit have been implicated in the apoptotic response to DNA damage (30 -32). In concert with these studies, DNA damage-induced apoptosis is attenuated in cells that i) express a kinase-inactive, dominantnegative c-Abl(K-R) mutant, ii) are null for c-Abl (c-abl Ϫ/Ϫ ), or iii) are treated with the c-Abl kinase inhibitor STI571 (23,33,34). Notably, however, there are no known interactions between c-Abl and the initiator or effector caspases.
The present studies demonstrate that c-Abl phosphorylates caspase-9 on Tyr-153 in vitro and in the response to DNA damage. We also show that c-Abl-mediated phosphorylation of caspase-9 contributes to DNA damage-induced autoprocessing of caspase-9, activation of caspase-3, and apoptosis.
Plasmid Construction-The vector expressing pCDNA3-caspase-9 has been described (36). The caspase-9(Y153F) mutant was generated by site-directed mutagenesis and confirmed by DNA sequencing. Caspase-9 and caspase-9(Y153F) were subcloned into pGEX4T-1 (Amersham Biosciences) at the BamH1 and EcoR1 sites and into pLXIN (Clontech) at the BamH1 site. Retroviral Transduction-PT67 cells were transiently transfected with pLXIN or pLXIN-caspase-9(Y153F) in the presence of Lipofectamine. At 48 h after transfection, cells were selected in medium containing G418. Retroviral titers of the supernatants were determined using NIH3T3 cells. Filtered retroviral supernatants containing polybrene (Sigma) were used to infect U-937 cells. At 24 h after infection, cells were seeded into methylcellulose medium (Stem Cell Technologies) and single cell clones were selected in G418.
In Vitro Binding Assays-Cell lysates were incubated with purified GST or GST-caspase-9 in lysis buffer for 2 h at 4°C. In other experiments, GST, GST-c-Abl SH3, GST-c-Abl, or GST-c-Abl(K-R) were incubated with purified His-caspase-9. Adsorbates to glutathione beads were analyzed by immunoblotting.
Apoptosis Assays-DNA content was assessed by staining ethanolfixed cells with propidium iodide and monitoring by FACScan (Becton Dickinson). Percentage of cells with sub-G 1 DNA were determined by the MODFIT LT Program (Verity Software). DNA laddering was performed as described (38).

RESULTS
c-Abl Associates with Caspase-9 -To determine whether c-Abl associates with caspase-9, lysates from U-937 cells were immunoprecipitated with anti-caspase-9. Immunoblot analysis of the precipitates with anti-c-Abl demonstrated little if any signal (Fig. 1A). By contrast, complexes of c-Abl and caspase-9 were detectable when similar experiments were performed on U-937 cells treated with araC, an inhibitor of DNA replication (39) and inducer of c-Abl activation (33,34) (Fig. 1A). The association of c-Abl and caspase-9 was found at 30 min of araC treatment and was maximal at 60 -120 min (Fig. 1A). To confirm the association, U-937 cell lysates were incubated with GST or a GST-caspase-9 fusion protein. Immunoblot analysis of the adsorbates showed binding of c-Abl to GST-caspase-9 and not GST (Fig. 1B). To determine whether the interaction is direct, purified His-tagged caspase-9 was incubated with GST fusion proteins containing kinase-active GST-c-Abl or kinaseinactive GST-c-Abl(K-R). Immunoblotting of the adsorbates with anti-caspase-9 demonstrated binding to both c-Abl and c-Abl(K-R) (Fig. 1C). To further define the c-Abl sequences responsible for binding to caspase-9, we incubated Hiscaspase-9 with GST or a GST fusion protein that contains the c-Abl SH3 domain (GST-c-Abl SH3). The results show that c-Abl SH3 binds to His-caspase-9 (Fig. 1D). These findings indicate that c-Abl associates with caspase-9 in the response to genotoxic stress and that the interaction is mediated by direct binding through the c-Abl SH3 domain.
c-Abl Phosphorylates Caspase-9 on Tyr-153-To determine whether caspase-9 functions as a c-Abl substrate, we incubated His-caspase-9 with purified GST-c-Abl or GST-c-Abl(K-R) and [␥-32 P]ATP. Analysis of the products by SDS-PAGE and autoradiography showed that caspase-9 is phosphorylated by c-Abl and not c-Abl(K-R) ( Fig. 2A, left). To define the phosphorylation site, individual tyrosines in His-caspase-9 were mutated to phenylalanine (Tyr 3 Phe), and the mutants were analyzed as substrates for c-Abl. The results demonstrated that, in contrast to other mutants (Y251F, Y345F, Y363F, Y397F), there was no detectable phosphorylation of His-caspase-9 with the Y153F mutation ( Fig. 2A, right). To determine whether c-Abl phosphorylates caspase-9 in vivo, studies were performed with wildtype and c-abl Ϫ/Ϫ mouse fibroblasts. Constitutive and araCinduced tyrosine phosphorylation of caspase-9 was detectable in wild-type, but not c-abl Ϫ/Ϫ , cells (Fig. 2B). Importantly, expression of c-Abl in the c-abl Ϫ/Ϫ cells (c-abl ϩ/Ϫ ) was also associated with constitutive and araC-induced tyrosine phosphorylation of caspase-9 (Fig. 2B). To determine whether caspase-9 is phosphorylated on Tyr-153 in cells, we generated U-937 cells stably expressing an empty vector or caspase-9(Y153F) (Fig. 2C, left). U-937 cells transfected with wild-type caspase-9 were not viable. As found for endogenous caspase-9, araC treatment was associated with coprecipitation of c-Abl and caspase-9(Y153F) (Fig. 2C, right). In addition, immunoblot analysis of anti-caspase-9 immunoprecipitates with anti-phospho-Tyr demonstrated araC-induced tyrosine phosphorylation of caspase-9 in the U-937/vector cells (Fig. 2D, left). By contrast, there was little detectable tyrosine phosphorylation of caspase-9 in the araC-treated U-937/caspase-9(Y153F) cells (Fig. 2D, right). These findings indicate that c-Abl phosphorylates caspase-9 on Tyr-153 in vitro and in the response of cells to DNA damage.
Caspase-9(Y153F) or STI571 Treatment Attenuates DNA Damage-induced Autoprocessing of Caspase-9 -Treatment of U-937 cells with araC is associated with release of mitochondrial cytochrome c to the cytosol and activation of caspase-3 at 4 and 6 h (34). araC treatment of U-937/vector cells was also associated with processing of caspase-9 to the p37 and p35 subunits (Fig. 3A). Importantly, however, processing of caspase-9 to the p35 fragment was attenuated in the araC-FIG. 1. c-Abl associates with caspase-9. A, U-937 cells were left untreated or treated with 10 M araC for the indicated times. Cell lysates were immunoprecipitated with anti-caspase-9. Immune complexes were subjected to immunoblotting (IB) with anti-c-Abl and anti-caspase-9. B, U-937 cell lysates were incubated with GST or GST-caspase-9. The adsorbates were analyzed by immunoblotting with anti-c-Abl. C, columnpurified His-caspase-9 was incubated with glutathione beads containing GST, GST-c-Abl, or GST-c-Abl(K-R). The adsorbates were subjected to immunoblot analysis with anti-caspase-9. D, His-caspase-9 was incubated with GST or GST-c-Abl SH3. The adsorbates were immunoblotted with anti-caspase-9. treated U-937/caspase-9(Y153F) cells (Fig. 3A). Moreover, araC-induced activation of caspase-3 was attenuated in U-937/ caspase-9(Y153F) cells compared with that in U-937/vector cells (Fig. 3B). Caspase-3 induces cleavage of caspase-9 to the p37 subunit (4,9). Thus, cleavage of caspase-9(Y153F) to p37 in the absence of caspase-3 activation indicates that low levels of caspase-3 activity or other proteases, perhaps even caspase-9 itself, may induce cleavage at Asp-330 in the presence of the Y153F mutation. To determine whether inhibition of c-Abl activity has similar effects on caspase-9 and caspase-3, U-937 cells were treated with the c-Abl inhibitor, STI571. Exposure to STI571 had no apparent effect on araC-induced release of cytochrome c (Fig. 3C). However, STI571 treatment was associated with attenuation of araC-induced cleavage of caspase-9 to the p35 subunit (Fig. 3C). STI571 also attenuated araC-induced activation of caspase-3 (data not shown). Consistent with  ). B, wild-type, c-abl Ϫ/Ϫ or c-abl ϩ/Ϫ cells were treated with 10 M araC for indicated times. Anti-caspase-9 immunoprecipitates were analyzed by immunoblotting with anti-phospho-Tyr or anti-caspase-9. C, lysates from U-937/vector or U-937/caspase-9(Y153F) cells (lanes A and B represent two separately isolated clones) were subjected to immunoblot analysis with anti-caspase-9 and anti-␤-actin (left). U-937/caspase-9(Y153F) cells were treated with 10 M araC for 1 h. Cell lysates were immunoprecipitated with anti-caspase-9. Immune complexes were subjected to immunoblotting with anti-c-Abl and anti-caspase-9 (right). D, U-937/vector (left) and U-937/caspase-9(Y153F) (right) cells were treated with 10 M araC for indicated times. Anti-caspase-9 immunoprecipitates were analyzed by immunoblotting with anti-phospho-Tyr and anti-caspase-9.
FIG. 3. Caspase-9(Y153F) and STI571 attenuate DNA damage-induced autocleavage of caspase-9. A and B, U-937/ vector or U-937/caspase-9(Y153F) cells were treated with 10 M araC for the indicated times. Cytoplasmic fractions were subjected to immunoblotting with anticaspase-9 and ␤-actin (left) or with anticaspase-3 and anti-␤-actin (right). C and D, U-937 cells were pretreated with 10 M STI571 for 24 h and then exposed to 10 M araC for 4 h. Cytoplasmic fractions were subjected to immunoblot analysis with the indicated antibodies (C). Anti-caspase-9 immunoprecipitates were analyzed by immunoblotting with the indicated antibodies (D).
Caspase-9(Y153F) Attenuates DNA Damage-induced Apoptosis-The effects of the caspase-9(Y153F) mutant on araCinduced apoptosis were first assessed by monitoring the percentage of cells with sub-G 1 DNA. araC treatment of U-937/ vector cells was associated with a substantial increase in sub-G 1 DNA content, and this response was attenuated in U-937/caspase-9(Y153F) cells (supplemental Fig. S1A). Similar results were obtained at different times of araC exposure and in repetitive experiments (Fig. 4A). U-937 cells also respond to 20 grays of ionizing radiation (IR) with the induction of apoptosis (40). As found with araC, IR-induced apoptosis of U-937/ caspase-9(Y153F) cells was decreased compared with that obtained with U-937/vector cells (supplemental Fig. S1B). Confirmation of these results in additional experiments (Fig.  4B) indicated that U-937/caspase-9(Y153F) cells exhibit an attenuated apoptotic response to diverse DNA damaging agents. As an additional assay for apoptotic cells, araC-induced DNA laddering was attenuated in U-937 cells expressing caspase-9(Y153F) (Fig. 4C). Similar results were obtained with IRtreated U-937/vector and U-937/caspase-9(Y153F) cells (Fig.  4C). These findings indicate that caspase-9(Y153F) attenuates DNA damage-induced apoptosis. DISCUSSION c-Abl Phosphorylates Caspase-9 on Tyr-153-Previous studies have demonstrated that Akt and extracellular signal-regulated kinase phosphorylate and thereby inactivate caspase-9 processing (15,16). The present work demonstrates that c-Abl associates with caspase-9 in the response of cells to genotoxic stress. Low levels of constitutive binding of c-Abl and caspase-9 were also observed when cells were stressed by being overgrown (data not shown). In vitro studies further demonstrated that the c-Abl SH3 domain binds directly to caspase-9. The available evidence indicates that the c-Abl SH3 domain interacts with a proline-rich motif in the caspase-9 caspase recruitment domain region (PEVLRPETP; amino acids 118 -126) (data not shown). The results also demonstrate that c-Abl phosphorylates caspase-9 on Tyr-153. The Tyr-153 site resides in the caspase-9 large subunit near the caspase recruitment domain region. Based on structural studies, Tyr-153 would be exposed on the inactive caspase-9 conformation and possibly buried in the active conformation near the dimer interface (7). Consistent with these predictions, caspase-9 and not caspase-9(Y153F) was subject to tyrosine phosphorylation in the response to DNA damage. DNA damage-induced phosphorylation of caspase-9 by a c-Abl-dependent mechanism was confirmed by the absence of this response in cells null for c-Abl expression. Moreover, expression of c-Abl in c-abl ϩ/Ϫ cells was associated with tyrosine phosphorylation of caspase-9. These findings indicate that caspase-9 is phosphorylated by c-Abl on Tyr-153 in vitro and in the cellular response to DNA damage.
c-Abl Regulates Caspase-9 Autoprocessing in the Apoptotic Response to DNA Damage-c-Abl is activated in the response of cells to genotoxic stress (17). Other studies on wild-type, c-abl Ϫ/Ϫ , and c-abl ϩ/Ϫ cells have shown that c-Abl is essential for DNA damage-induced apoptosis (23,33). Moreover, inhibition of c-Abl with STI571 treatment has been shown to block araC-induced apoptosis (34). c-Abl interacts with diverse signaling molecules that are associated with the apoptotic response (17,(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32); however, there has been no known interaction between c-Abl and the initiator or effector caspases. In the present studies, the functional significance of the c-Ablcaspase-9 interaction in cells is supported by the demonstration that inhibition of c-Abl with STI571 blocks autoprocessing of caspase-9 to the p35 subunit. Caspase-9(Y153F) also blocked processing to the p35 subunit and caspase-3 activation. Our results further demonstrate that caspase-9(Y153F) blocks araC-induced apoptosis. araC misincorporates into DNA and functions as an inhibitor of DNA replication (39). The demonstration that caspase-9(Y153F) similarly attenuates IR-induced apoptosis indicates that the c-Abl 3 caspase-9 pathway is of importance in the response to other genotoxic agents. Of potential significance, caspase-9 Tyr-153 is conserved in sites that are proximal to L1 loops in other caspases, including caspase-3 (Tyr-37) and caspase-7 (Tyr-60). The present findings thus support a model in which caspase-9 autocleavage is regulated by a c-Abl-dependent mechanism in the apoptotic response to genotoxic stress. c-Abl Regulates Caspase-9 Autocleavage