Activation of the Cytoplasmic c-Abl Tyrosine Kinase by Reactive Oxygen Species*

The ubiquitously expressed c-Abl protein tyrosine kinase localizes to both the nucleus and cytoplasm. The nuclear form of c-Abl is activated in the cellular response to genotoxic stress. Here we show that cytoplasmic c-Abl is activated by oxidative stress. The results also demonstrate that mitochondrial cytochrome c is released in the cellular response to H2O2 and that this effect is mediated by a c-Abl-dependent mechanism. In concert with these results, we show that H2O2-induced apoptosis is attenuated in c-Abl-deficient cells. These findings demonstrate that cytoplasmic c-Abl is involved in the apoptotic response of cells to oxidative stress.

Normal cellular metabolism is associated with the production of reactive oxygen species (ROS) 1 and, as a consequence, damage to DNA and proteins (1,2). The generation of ROS is also known to induce apoptosis; however, the molecular mechanisms responsible for ROS-induced apoptosis are unclear. Studies have indicated that ROS induce activation of topoisomerase II-mediated cleavage of chromosomal DNA and thereby apoptosis (3). Other work has suggested that ROSinduced apoptosis is p53-dependent (4,5) and that p53-induced apoptosis is mediated by ROS (6 -8). In addition, the p66 shc adaptor protein (5) and the p85 subunit of phosphatidylinositol 3-kinase (PI3K) (4) have been implicated in the apoptotic response to oxidative stress.
The nuclear form of the c-Abl tyrosine kinase is activated in the cellular response to genotoxic stress (9). Nuclear c-Abl has been implicated in the apoptotic response to DNA damage by mechanisms in part dependent on p53 and its homolog, p73 (10 -14). c-Abl also functions as an upstream effector of the proapoptotic SAPK/JNK and p38 mitogen activated protein kinase (MAPK) pathways in the genotoxic stress response (9,15,16). Other studies have demonstrated that c-Abl phosphorylates p85 and thereby inhibits PI3K activity in the apoptotic response to DNA damage (17). Additional evidence supporting a role for c-Abl in apoptosis has been provided by the findings that cells deficient in c-Abl or expressing a dominant-negative c-Abl mutant exhibit an attenuated apoptotic response to genotoxic agents (18,19).
Recent work has shown that c-Abl phosphorylates protein kinase C (PKC) ␦ in cells treated with H 2 O 2 (20). The present results demonstrate that the cytoplasmic, and not the nuclear, form of c-Abl is activated in the cellular response to H 2 O 2 . We also show that H 2 O 2 induces mitochondrial cytochrome c release and apoptosis by a c-Abl-dependent mechanism.
Isolation of Cytoplasmic and Nuclear Fractions-Cells were disrupted in lysis buffer containing 0.05% Nonidet P-40. The cytoplasmic and nuclear fractions were prepared as described (24).
Preparation of Cytoplasts-Enucleated cells were prepared by density centrifugation as described (25). Cells were incubated in 21 M cytochalasin B for 1 h at 37°C, layered over a discontinuous Ficoll gradient, and centrifuged at 80,000 ϫ g for 1 h. Cytoplasts were collected at the 12.5-15% Ficoll interface. Cytoplast purity was assessed by staining with 0.5 g/ml 4Ј,6-diamino-2-phenylindole (DAPI) and was greater than 95% free of whole cells.
Apoptosis Assays-DNA content was assessed by staining ethanolfixed cells with propidium iodide and monitoring by FACScan (Becton-Dickinson).

RESULTS AND DISCUSSION
To determine whether c-Abl is activated by ROS, lysates from COS7 cells exposed to H 2 O 2 were subjected to immunoprecipitation with mouse IgG, as a control, or anti-c-Abl antibody. The precipitates were assayed for phosphorylation of a GST-Crk (120 -225) fusion protein (27,28). There was no detectable phosphorylation of GST-Crk (120 -225) with the control immunoprecipitates (Fig. 1a, left). A low level of GST-Crk (120 -225) phosphorylation was detectable when assaying antic-Abl immunoprecipitates from control cells, whereas exposure to H 2 O 2 resulted in stimulation (4 -5-fold) of the Crk kinase activity (Fig. 1a, right, first 2 lanes). By contrast, there was no detectable H 2 O 2 -induced phosphorylation of a GST-Crk (120 -212) fusion protein that lacks the c-Abl phosphorylation site at Tyr-221 (Fig. 1a, right, last 2 lanes). The results also show that H 2 O 2 treatment is not associated with increases in the level of c-Abl protein (Fig. 1a, right). To confirm involvement of ROS in c-Abl activation, cells were treated with NAC, a scavenger of reactive oxygen intermediates and precursor of glutathione (29,30). NAC treatment inhibited H 2 O 2 -induced phosphorylation of GST-Crk (120 -225) by c-Abl (Fig. 1b). The induction of c-Abl activity was dependent on H 2 O 2 concentration, with 5-fold increases upon exposure to 1 mM H 2 O 2 (Fig. 1c, left). In addition, maximal induction of c-Abl activity was observed at 30 -60 min (Fig. 1c, right). The finding that human DLD1 cells respond to H 2 O 2 with activation of c-Abl further indicated that the results are not restricted to certain cell types (Fig. 1d).
To extend the analysis of H 2 O 2 -induced activation of c-Abl to other pathways involved in the ROS response, we studied mouse embryo fibroblasts that are null for c-Abl expression (c-Abl Ϫ/Ϫ MEFs) (21). There was no detectable c-Abl activity in control or H 2 O 2 -treated c-Abl Ϫ/Ϫ cells (Fig. 2a). By contrast, wild-type MEFs responded to H 2 O 2 with induction of c-Abl activity. Recent studies have demonstrated that c-Abl interacts with PKC␦ in the response to oxidative stress (20). To determine whether c-Abl is required for activation of PKC␦, we assayed anti-PKC␦ immunoprecipitates from c-Abl Ϫ/Ϫ and wild-type MEFs. The results demonstrate that, whereas PKC␦ is required for activation of c-Abl (20), c-Abl is dispensable for activation of PKC␦ in the ROS response (Fig. 2b). Other studies have demonstrated that ERK1 is activated in cells exposed to H 2 O 2 (31). Analysis of anti-ERK1 immunoprecipitates from H 2 O 2 -treated c-Abl Ϫ/Ϫ and wild-type MEFs demonstrated activation of ERK1 by a c-Abl-independent mechanism (Fig. 2c). These findings demonstrate that activation of c-Abl in the ROS response is not functional in the induction of PKC␦ or ERK1 activities.
As nuclear c-Abl is activated in the stress response to DNA  damage (9), studies were performed to define the subcellular localization of ROS-induced c-Abl activation. Cells were treated with H 2 O 2 before preparation of nuclear and cytoplasmic fractions. Analysis of cytoplasmic anti-c-Abl immunoprecipitates demonstrated increased phosphorylation of GST-Crk (120 -225) (Fig. 3a). By contrast, there was no detectable activation of c-Abl in the nuclear fraction (Fig. 3a). Oxygen radicals induce lesions in nuclear DNA (32,33), and nuclear c-Abl is activated by DNA damage (9). To determine whether a nuclear signal is required for H 2 O 2 -induced activation of cytoplasmic c-Abl, we assayed cytoplasts devoid of nuclei. Treatment of the cytoplasts with H 2 O 2 was associated with induction of c-Abl activity (Fig.  2b). By contrast, cisplatin treatment, which activates nuclear c-Abl (9), had no detectable effect on c-Abl activity in cytoplasts (data not shown). These findings indicate that cytoplasmic c-Abl is activated in the response to oxidative stress by a mechanism independent of nuclear signals.
The cellular response to genotoxic stress includes release of mitochondrial cytochrome c and the induction of apoptosis (34). To determine whether oxidative stress induces cytochrome c release, cytoplasmic lysates from wild-type and c-Abl Ϫ/Ϫ cells treated with H 2 O 2 were subjected to immunoblotting with anticytochrome c. The results demonstrate that H 2 O 2 treatment of wild-type MEFs is associated with increased levels of cytochrome c (Fig. 4a). By contrast, cytochrome c release was not detectable in c-Abl Ϫ/Ϫ MEFs treated with H 2 O 2 (Fig. 4a). To determine whether c-Abl contributes to the induction of apoptosis by oxidative stress, H 2 O 2 -treated MEFs were assayed for the appearance of sub-G 1 DNA. The results demonstrate that, compared with wild-type MEFs, the c-Abl Ϫ/Ϫ MEFs exhibit an attenuated apoptotic response to H 2 O 2 exposure (Fig. 4b).
Analysis at 3 to 24 h of H 2 O 2 exposure confirmed that cells deficient in c-Abl expression exhibit a defective apoptotic response (Fig. 4c). The finding that H 2 O 2 -induced release of cytochrome c is completely abrogated in c-Abl Ϫ/Ϫ cells indicates that the attenuated induction of apoptosis in response to H 2 O 2 is mediated by a cytochrome c-independent pathway (Fig. 4b).
These results collectively demonstrate that cytoplasmic H 2 O 2 induces cytochrome c release and apoptosis by a c-Abl-dependent mechanism.
Oxidative cellular damage contributes to aging (5) and, in the presence of acute ROS exposure, the induction of apoptosis (35). Previous work has shown that the nuclear c-Abl kinase is activated in the apoptotic response of cells to genotoxic stress (12)(13)(14)18). Conversely, the present studies demonstrate that cytoplasmic, and not nuclear, c-Abl is activated in the apoptotic response to oxidative stress. Whereas DNA damage-induced apoptosis is mediated by activation of c-Abl and the release of mitochondrial cytochrome c (34), less is known about involvement of mitochondrial signals in H 2 O 2 -induced cell death. The present results demonstrate that cytochrome c release is also induced in response to oxidative stress and that this event is c-Abl-dependent. These findings support a model in which c-Abl functions in determining cell fate by conferring stressinduced signals to the release of cytochrome c and thereby apoptosis. The findings further indicate that the subcellular distribution of c-Abl determines localization of the specific response to apparently diverse environmental stresses. However, ROS induce damage to DNA (1), as well as other cellular components (2), and thus it is conceivable that cells have evolved with conservation of a similar response to both genotoxic and oxidative stress. The present findings demonstrate that, analogous to activation of nuclear c-Abl by DNA-damaging agents (9), cytoplasmic c-Abl is activated by ROS-induced stress.