Catalase activity is regulated by c-Abl and Arg in the oxidative stress response.

The Abl family of mammalian non-receptor tyrosine kinases includes c-Abl and Arg. Recent studies have demonstrated that c-Abl and Arg are activated in the response of cells to oxidative stress. This work demonstrates that catalase, a major effector of the cellular defense against H2O2, interacts with c-Abl and Arg. The results show that H2O2 induced binding of c-Abl and Arg to catalase. The SH3 domains of c-Abl and Arg bound directly to catalase at a P293FNP site. c-Abl and Arg phosphorylated catalase at Tyr231 and Tyr386 in vitro and in the response of cells to H2O2. The functional significance of the interaction is supported by the demonstration that cells deficient in both c-Abl and Arg exhibit substantial increases in H2O2 levels. In addition, c-abl-/- arg-/- cells exhibited a marked increase in H2O2-induced apoptosis compared with that found in the absence of either kinase. These findings indicate that c-Abl and Arg regulate catalase and that this signaling pathway is of importance to apoptosis in the oxidative stress response.

Normal aerobic metabolism is associated with the production of reactive oxygen species (ROS), 1 including superoxides, hydrogen peroxide (H 2 O 2 ), hydroxyl radicals, and nitric oxide. A substantial amount of oxygen reduced by the mitochondrial respiratory chain is converted to superoxide and then to H 2 O 2 by the mitochondrial superoxide dismutase (1). H 2 O 2 is readily diffusible across cell membranes and functions as a signaling molecule in diverse cellular events. Mitogenic signals induced by certain growth factors and activated Ras are mediated by H 2 O 2 production (2,3). H 2 O 2 activates c-Jun, c-Fos, and NF-B and thereby regulates gene transcription (4,5). In addition, the extracellular signal-regulated protein kinase, c-Jun N-terminal kinases, p70 S6K , and p90 rsk are activated by H 2 O 2 signaling (2,6,7).
The generation of H 2 O 2 by normal cellular metabolism is also associated with damage to DNA, proteins, and lipids (8,9) and the induction of apoptosis (10,11). Although few insights are available regarding the mechanisms responsible for ROSinduced cell death, H 2 O 2 activates topoisomerase II-mediated cleavage of chromosomal DNA and thereby apoptosis (12). The p66 shc adaptor protein (13) and the p85 subunit of phosphatidylinositol 3-kinase (14) have been implicated in the apoptotic response to H 2 O 2 . Other studies have indicated that p53-induced apoptosis is mediated by ROS (12,15,16) and that H 2 O 2 -induced apoptosis is p53-dependent (13,14).
The predominant enzymatic mechanisms that regulate intracellular H 2 O 2 levels are mediated by catalase and glutathione peroxidase. The tetrameric catalase converts H 2 O 2 to H 2 O and O 2 in peroxisomes (17). With the exception of rat myocardial cells, catalase is not detectable in mitochondria (18). Glutathione peroxidase converts H 2 O 2 to H 2 O in a reaction that oxidizes GSH to its disulfide form (GSSG). In turn, GSH is regenerated from GSSG by glutathione reductase. Regulation of H 2 O 2 by the glutathione redox cycle is mediated in the cytosol and mitochondria. Little is known about the regulation of catalase or glutathione peroxidase, particularly the effects of post-translational modifications on their activities.
The widely expressed mammalian c-Abl and Arg non-receptor tyrosine kinases (19 -21) have been implicated in cellular responses to oxidative and other types of stress (22)(23)(24)(25). The N-terminal SH3, SH2, and kinase domains of c-Abl and Arg share ϳ90% identity. The C-terminal regions of c-Abl and Arg share 29% identity and are distinguished from the other nonreceptor tyrosine kinases by the presence of globular and filamentous actin-binding domains (26). c-Abl differs from Arg by the presence of a nuclear localization signal (26) and DNAbinding sequences (27) in the C-terminal region. c-Abl also differs from Arg by localization to the nucleus and extranuclear organelles, whereas Arg is expressed predominantly in the cytoplasm (28).
The cytoplasmic form of c-Abl is activated in the response of cells to H 2 O 2 by a mechanism dependent on protein kinase C␦ (23,29). c-Abl is targeted to mitochondria in H 2 O 2 -treated cells and induces loss of the mitochondrial transmembrane potential (30). Mitochondrial targeting of c-Abl in response to oxidative stress is also associated with cytochrome c release and induction of apoptosis (23). In concert with these findings, c-Abldeficient cells exhibit resistance to H 2 O 2 -induced cell death (23,30). Other studies have shown that Arg is activated by oxidative stress and that this response involves Arg-mediated phosphorylation of the pro-apoptotic Siva-1 protein (24). H 2 O 2 -induced apoptosis is attenuated in Arg-deficient cells, and this defect is corrected by reconstituting Arg expression (24). Moreover, recent findings indicate that ROS induce c-Abl-Arg heterodimers and that both c-Abl and Arg are necessary as effectors in the apoptotic response to oxidative stress (31).
This study demonstrates that c-Abl and Arg interact with catalase in the response of cells to oxidative stress. The functional significance of these findings is supported by the demonstration that cells deficient in both c-Abl and Arg exhibit elevated intracellular H 2 O 2 levels and a pronounced apoptotic response to oxidative stress.
Vectors-FLAG-tagged c-Abl, Arg, and catalase and their mutants were expressed by cloning into the pcDNA3.1-based FLAG vector. His/ Express-tagged constructs were prepared by cloning into pcDNA4HisMAX, which contains N-terminal His and Express tags (Invitrogen). Myc-tagged c-Abl and catalase vectors were prepared by cloning into pCMV-Myc (Clontech). Glutathione S-transferase (GST) fusion proteins were generated by expression of pGEX4T2-based vectors in Escherichia coli BL21(DE3).
Binding Assays-Cell lysates were incubated with 5 g of GST or GST fusion proteins bound to glutathione beads. After incubation for 2 h at 4°C, the adsorbates were washed with lysis buffer and then subjected to immunoblot analysis. An aliquot of the total lysate (2%, v/v) was included as a control. For direct binding assays, purified GST fusion proteins were incubated with in vitro translated 35 S-labeled catalase. The adsorbates were analyzed by SDS-PAGE and autoradiography.
Assessment of Catalase Activity-For the substrate solution, 0.1 ml of 30% H 2 O 2 was added to 50 ml of 0.05 M phosphate buffer (pH 7.0). The A 240 of the solution was adjusted to between 0.550 and 0.520. The reactions were initiated by adding 0.1 ml of the enzyme solution to 2.9 ml of the substrate solution in a silica cuvette (1-cm light path) at 25°C. The time required for the A 240 to decrease from 0.450 to 0.400 corresponds to the decomposition of 3.45 mol of H 2 O 2 in the 3-ml assay (Sigma catalog). FLAG-catalase or the FLAG-catalase(Y-F) mutants were expressed in 293 cells with c-Abl, c-Abl(K-R), Arg, or Arg(K-R). Anti-FLAG immunoprecipitates were assayed for catalase activity and by immunoblotting for catalase protein. Densitometric scanning of the signals was compared with that obtained with dilutions of crystalline catalase (Sigma). Apoptosis Assays-DNA content was assessed by staining ethanolfixed and citrate buffer-permeabilized cells with propidium iodide and monitoring by FACScan (BD Biosciences). The numbers of cells with sub-G 1 DNA were determined with a MODFIT LT program (24).

c-Abl and Arg
Associate with Catalase-To determine whether c-Abl associates with catalase in the cellular response to oxidative stress, lysates from H 2 O 2 -treated MCF-7 cells were subjected to immunoprecipitation with anti-c-Abl antibody. Immunoblot analysis of the precipitates with anti-catalase antibody demonstrated a low level of constitutive c-Abl binding to catalase (Fig. 1A). However, treatment with 0.25 mM H 2 O 2 resulted in increased formation of c-Abl-catalase complexes (Fig. 1A). Similar findings were obtained when cells were treated with 0.5 and 1.0 mM H 2 O 2 (Fig. 1A). However, exposures to 2.0 mM H 2 O 2 were associated with decreased formation of c-Abl-catalase complexes (Fig. 1A). In the reciprocal experiment, analysis of anti-catalase immunoprecipitates by immunoblotting with anti-c-Abl antibody confirmed that H 2 O 2 induced the association of c-Abl and catalase and that the extent of the interaction was dependent on the concentration of H 2 O 2 (Fig. 1B). Similar results were obtained for an H 2 O 2 -dependent association between Arg and catalase (Fig. 1B). The demonstration that H 2 O 2 induced binding of catalase with c-Abl and Arg in U-937 cells indicates that the interaction with catalase occurs in different cell types (Fig. 1C). To confirm binding of catalase to c-Abl and Arg, 293 cells were transfected to express Myc-tagged catalase and FLAG-tagged c-Abl or Arg. Immunoblot analysis of anti-FLAG immunoprecipitates with anti-catalase antibody demonstrated constitutive binding of catalase and c-Abl (Fig. 1D). The finding that kinase-inactive c-Abl(K-R) also associated with catalase demonstrates that binding is independent of the c-Abl kinase function (Fig. 1D). Similar results were obtained for a constitutive interaction between catalase and Arg or Arg(K-R) (Fig. 1D). These findings indicate that c-Abl and Arg associate with catalase and that the interaction is increased in response to oxidative stress.
Binding of c-Abl and Arg to Catalase-To further define the association between c-Abl and catalase, lysates from MCF-7 cells were incubated with GST fusion proteins containing the c-Abl SH2 or SH3 domain. Analysis of the adsorbates by immunoblotting with anti-catalase antibody demonstrated binding of catalase to the c-Abl SH2 and SH3 domains ( Fig. 2A). Similar findings were obtained for binding of catalase to Arg SH2 and Arg SH3 domain ( Fig. 2A). As a control, there was no detectable binding of catalase to GST proteins containing the Grb2 SH2 or SH3 domain ( Fig. 2A). To assess whether binding of c-Abl and catalase is direct, GST-c-Abl and GST-Arg fusion proteins were incubated with purified 35 S-labeled FLAG-catalase. Analysis of the adsorbates by SDS-PAGE and autoradiography demonstrated binding to the c-Abl and Arg SH3 domains (Fig. 2B). By contrast, there was no detectable binding of catalase to the GST proteins containing the c-Abl or Arg SH2 domain (Fig. 2B). These results suggest that catalase is tyrosine-phosphorylated in MCF-7 cells, but not after in vitro translation. Catalase contains a potential proline-rich site ( 293 PFNP 296 ) for SH3 domain binding. Lysates from 293 cells expressing a FLAG-catalase(P293A) mutant were analyzed for binding to the c-Abl SH3 domain. The results demonstrate that, in contrast to wild-type catalase, binding of the c-Abl SH3 domain was abrogated with catalase(P293A) (Fig. 2C). Similar results were obtained with the Arg SH3 domain (Fig. 2C). To show that the PFNP motif is important for the interaction between c-Abl and catalase in vivo, lysates from cells expressing Myc-c-Abl and FLAG-catalase or FLAG-catalase(P293A) were subjected to immunoprecipitation with anti-Myc antibody. Analysis of the immunoprecipitates by immunoblotting with anti-FLAG antibody demonstrated binding of c-Abl to catalase, but not to catalase(P293A) (Fig. 2D). In a similar analysis, there was no detectable binding of Myc-Arg to catalase(P293A) (data not shown). These findings demonstrate that c-Abl and Arg bind directly to catalase through interactions of their SH3 domains and the catalase PFNP site.
Tyrosine Phosphorylation of Catalase in Response to Oxidative Stress-To determine whether c-Abl phosphorylates catalase, recombinant c-Abl was incubated with purified catalase and [␥-32 P]ATP. Analysis of the reaction products by SDS-PAGE and autoradiography demonstrated phosphorylation of catalase (Fig. 3A, left panel). To confirm these findings, similar reactions were performed in the presence of unlabeled ATP. Immunoblot analysis of the reaction products with anti-Tyr(P) antibody demonstrated that catalase was phosphorylated at tyrosine (Fig. 3A, right panel). In concert with the demonstration that ROS-induced activation of c-Abl is mediated by protein kinase C␦ (23,29), the addition of H 2 O 2 to the reaction had no apparent effect on tyrosine phosphorylation of catalase (data not shown). The results also demonstrate that catalase is a substrate for Arg (Fig. 3B). To assess c-Abl-mediated phosphorylation of catalase in vivo, lysates from cells expressing Myc-catalase were immunoprecipitated with anti-Myc antibody. Analysis of the precipitates by immunoblotting with anti-Tyr(P) antibody demonstrated phosphorylation of catalase at tyrosine (Fig. 3C). The finding that treatment of the cells with the c-Abl/Arg kinase inhibitor STI571 (33) decreased tyrosine phosphorylation of catalase provided support for a c-Abl-and/or Arg-dependent mechanism (Fig. 3C). Cells expressing Myccatalase and c-Abl also showed phosphorylation of catalase at tyrosine (Fig. 3D). By contrast, tyrosine phosphorylation of catalase was inhibited in cells expressing c-Abl(K-R) (Fig. 3D). Catalase was also subject to tyrosine phosphorylation in cells expressing kinase-active Arg, but not kinase-inactive Arg(K-R) (Fig. 3D).
To assess tyrosine phosphorylation of catalase in response to oxidative stress, anti-catalase immunoprecipitates from lysates of H 2 O 2 -treated MCF-7 cells were subjected to immunoblotting with anti-Tyr(P) antibody. The results demonstrate that tyrosine phosphorylation of catalase was increased in response to 1.0 (but not 0.25 or 2.0) mM H 2 O 2 (Fig. 4A). By contrast, H 2 O 2 -induced increases in tyrosine phosphorylation of catalase were not found in MCF-7 cells stably expressing c-Abl(K-R) or Arg(K-R) (Fig. 4B). Moreover, pretreatment with 1 and 10 M STI571 blocked H 2 O 2 -induced tyrosine phospho- Lysates not subjected to immunoprecipitation were used as controls for catalase and c-Abl expression. B, lysates from MCF-7 cells treated with the indicated concentrations of H 2 O 2 were subjected to immunoprecipitation with anti-catalase antibody or IgG. The immunoprecipitates were analyzed by immunoblotting with anti-c-Abl, anti-Arg, or anti-catalase antibody. Lysates not subjected to immunoprecipitation were used as controls. C, anti-catalase immunoprecipitates from U-937 cells treated with H 2 O 2 for 2 h were analyzed by immunoblotting with anti-c-Abl, anti-Arg, or anti-catalase antibody. Immunoprecipitates prepared with IgG and lysates not subjected to immunoprecipitation were used as controls. D, 293 cells were cotransfected with 4 g of Myc-catalase and 1 g of FLAG-Arg, FLAG-Arg(K-R), FLAG-c-Abl, FLAG-c-Abl(K-R), or FLAG vector. Anti-FLAG immunoprecipitates prepared at 24 h after transfection were analyzed by immunoblotting with anti-catalase or horseradish peroxidase-conjugated anti-FLAG antibody. rylation of catalase (Fig. 4C). These findings collectively demonstrate that catalase is phosphorylated by a c-Abl-and/or Arg-dependent mechanism in the oxidative stress response.
Catalase Is Regulated by Phosphorylation at Tyr 231 and Tyr 386 -To define the phosphorylation sites, catalase was incubated with c-Abl and then subjected to tryptic digestion. Analysis of the fragments by high pressure liquid chromatography separation and Edman sequencing demonstrated phosphorylation at Tyr 231 and Tyr 386 . Compared with wild-type catalase, mutation of Tyr 231 to Phe was associated with a decrease in c-Abl-mediated phosphorylation (Fig. 5A). Mutation of Tyr 386 to Phe was also associated with a decrease in c-Abl phosphorylation (Fig. 5A). The catalase(Y231F/Y386F) double mutant exhibited comparable decreases in c-Abl phosphorylation, but not complete abrogation (Fig. 5A). Similar results were obtained for Arg-mediated phosphorylation of the catalase mutants (data not shown). Because these findings indicate that phosphorylation of catalase by c-Abl and Arg is not restricted to Tyr 231 and Tyr 386 , we generated mutants at the other 14 tyrosine sites. The absence of a detectable effect of these mutants on c-Abl and Arg phosphorylation indicated that Tyr 231 and Tyr 386 are the predominant sites (data not shown).
To determine whether in vivo phosphorylation of catalase is affected by the tyrosine mutations, 293 cells were transfected to express Myc-c-Abl and FLAG-catalase or FLAG-catalase mutated at Tyr 231 , Tyr 386 , or Tyr 231 /Tyr 386 . Immunoblot analysis of anti-FLAG immunoprecipitates with anti-Tyr(P) antibody showed a substantial decrease in tyrosine phosphorylation of the catalase(Y231F) mutant (Fig. 5B). The extent of tyrosine phosphorylation was also decreased with catalase(Y386F) and the double mutant (Fig. 5B). Similar findings were obtained when Express-Arg was expressed with wildtype catalase and the Tyr-to-Phe mutants (Fig. 5C). To determine whether catalase is regulated by phosphorylation at Tyr 231 and Tyr 386 , anti-FLAG immunoprecipitates from 293 cells expressing FLAG-catalase and c-Abl or c-Abl(K-R) were analyzed for catalase activity. The results demonstrate that, compared with wild-type c-Abl, cotransfection of c-Abl(K-R) substantially decreased catalase activity (Fig. 5D). Compared with wild-type FLAG-catalase, the activity was decreased when FLAG-catalase(Y231F) or FLAG-catalase(Y386F) was expressed with c-Abl (Fig. 5D). The demonstration that the activity of the FLAG-catalase(Y231F/Y386F) double mutant was comparable to that of the single Tyr-to-Phe mutants indicates that phosphorylation at both Tyr 231 and Tyr 386 is needed for stimulation of catalase activity (Fig. 5D). Moreover, the inhibitory effects of c-Abl(K-R) on catalase activity were much less pronounced with these mutants than with the wild-type enzyme (Fig. 5D). Similar results were obtained when the wild-type and mutant FLAG-catalase proteins were expressed with Arg or Arg(K-R) (data not shown). These findings indicate that (i) c-Abl and Arg activate catalase by phosphorylation at both Tyr 231 and Tyr 386 , and (ii) expression of c-Abl(K-R) or Arg(K-R) attenuates catalase activity.
c-Abl and Arg Regulate Intracellular H 2 O 2 Levels-Intracellular H 2 O 2 levels are controlled predominantly by catalase and the glutathione redox cycle. To assess the protective effects of catalase, BSO was used to inhibit the glutathione redox cycle by depleting intracellular levels of glutathione (7). MEFs were treated with BSO and monitored for intracellular H 2 O 2 levels by fluorescence spectrophotometry using the oxidant-sensitive dye DCF-DA. The results demonstrate that, compared with wild-type MEFs, arg Ϫ/Ϫ cells exhibited somewhat higher intracellular H 2 O 2 levels (Fig. 6A, upper panel). Increased H 2 O 2 levels were more apparent in c-abl Ϫ/Ϫ cells (Fig. 6A, upper  panel). Importantly, even higher levels of H 2 O 2 were found in c-abl Ϫ/Ϫ arg Ϫ/Ϫ cells (Fig. 6A, upper panel). Treatment of the wild-type MEFs with H 2 O 2 was associated with an increase in H 2 O 2 levels (Fig. 6A, lower panel). H 2 O 2 treatment of c-abl Ϫ/Ϫ and arg Ϫ/Ϫ cells was also associated with higher intracellular H 2 O 2 levels (Fig. 6A, lower panel). Moreover, the finding that the c-abl Ϫ/Ϫ arg Ϫ/Ϫ cells exhibited a striking increase in H 2 O 2 levels supports more pronounced dysregulation of catalase than that observed in c-abl Ϫ/Ϫ or arg Ϫ/Ϫ cells (Fig. 6A, lower  panel). Analysis of three experiments confirmed significant differences for H 2 O 2 levels in control and H 2 O 2 -treated c-abl Ϫ/Ϫ arg Ϫ/Ϫ cells compared with wild-type, c-abl Ϫ/Ϫ , and arg Ϫ/Ϫ cells (Fig. 6B). In addition, analysis of two independent clones of c-abl Ϫ/Ϫ arg Ϫ/Ϫ cells indicated that the substantial differences were not due to clonal variation (Fig. 6B). To confirm the basis for dysregulation of H 2 O 2 homeostasis, c-abl Ϫ/Ϫ arg Ϫ/Ϫ cells were transduced with a retroviral vector expressing Arg (24). The results demonstrate that Arg expression was associated with decreases in H 2 O 2 levels (Fig. 6C)  cells exhibit resistance to H 2 O 2 -induced apoptosis (23). Arg Ϫ/Ϫ cells also exhibit an attenuated apoptotic response to oxidative stress (24). In concert with these findings, treatment of wildtype MEFs with H 2 O 2 was associated with induction of apoptosis, whereas c-abl Ϫ/Ϫ and arg Ϫ/Ϫ cells showed loss of this response (Fig. 7A). By contrast to c-abl Ϫ/Ϫ and arg Ϫ/Ϫ cells, c-abl Ϫ/Ϫ arg Ϫ/Ϫ cells exhibited an apoptotic response to H 2 O 2 that was more pronounced than that found with wild-type MEFs (Fig. 7A). Analysis of separate experiments confirmed the increased sensitivity of c-abl Ϫ/Ϫ arg Ϫ/Ϫ cells compared with c-abl Ϫ/Ϫ and arg Ϫ/Ϫ cells (Fig. 7B). Studies of two separate c-abl Ϫ/Ϫ arg Ϫ/Ϫ clones demonstrated similar results (Fig. 7B). These findings demonstrate that cells deficient in both c-Abl and Arg, but not either alone, exhibit a hypersensitive apoptotic response to oxidative stress.

DISCUSSION
Involvement of c-Abl and Arg in the ROS Response-The c-Abl and Arg tyrosine kinases are activated in the response of diverse types of cells to oxidative stress (23,24). Activation of c-Abl is associated with its translocation to mitochondria and the release of cytochrome c (23,30). The finding that ROSinduced apoptosis is attenuated in c-Abl-deficient cells has supported a role for c-Abl in the apoptotic response to oxidative stress (23). Moreover, the demonstration that ROS-induced cell death is attenuated in cells expressing a kinase-inactive, dominant-negative c-Abl(K-R) mutant has indicated that this response is, at least in part, dependent on the c-Abl kinase function (30). Other studies indicate that, like c-Abl, activation of Arg contributes to ROS-induced apoptosis (24). In H 2 O 2treated cells, Arg interacts with the pro-apoptotic Siva-1 protein (24) and translocates to mitochondria. 2 The apoptotic response to oxidative stress is attenuated in Arg-deficient cells, and this defect is corrected by reconstituting Arg expression (24). In addition, ROS-induced apoptosis is attenuated in cells 2 C. Cao, Y. Leng, and D. Kufe, unpublished data. c-Abl and Arg Regulate Catalase Activity expressing Arg(K-R) compared with wild-type cells (24). Conversely, cells overexpressing Arg exhibit a marked increase in ROS-induced apoptosis (24). These findings and the demonstration that ROS induce c-Abl-Arg heterodimers (31) indicate that both c-Abl and Arg are functionally involved in the apoptotic response to oxidative stress.
Interaction of c-Abl and Arg with Catalase-The results of this study extend the involvement of c-Abl and Arg in the ROS response by demonstrating their interaction with catalase. Catalase has been implicated as an important effector in the prevention of apoptosis (34 -36). The demonstration that mutations in cytosolic catalase shorten the life span in Caenorhabditis elegans has also supported a role for catalase and ROS in aging of eukaryotic cells (37). Notably, little is known about proteins that interact with catalase or the regulation of catalase activity in response to oxidative stress. Our findings demonstrate that exposure of cells to H 2 O 2 is associated with binding of c-Abl and Arg to catalase. Complexes of c-Abl and catalase were induced at 0.25-1.0 (but not at 2.0) mM H 2 O 2 . A similar biphasic effect was observed for the formation of Arg-catalase complexes. When ectopically expressed, c-Abl and Arg associated with catalase in the absence of H 2 O 2 -induced stress, presumably as a result of their overexpression. The results further show that the SH3 domains of c-Abl and Arg bind directly to a PFNP site (amino acids 293-296) that resides on the surface of catalase (38). Studies in cells confirmed that the PFNP site is necessary for binding of c-Abl and Arg to catalase. Moreover, ectopic expression of c-Abl(K-R) and Arg(K-R) mutants demonstrated that binding to catalase is not dependent on the c-Abl or Arg kinase functions.
c-Abl and Arg Phosphorylate Catalase-Our findings also provide evidence that catalase functions as a substrate for c-Abl and Arg phosphorylation. To our knowledge, there have been no reports that catalase is subject to tyrosine or serine/ threonine phosphorylation. The in vitro results demonstrate that c-Abl phosphorylates catalase predominantly at Tyr 231 and Tyr 386 . The finding that mutation of these sites failed to completely abrogate phosphorylation indicates the potential involvement of other tyrosines. In concert with these findings, catalase was phosphorylated at tyrosine in the response of cells to oxidative stress. Tyrosine phosphorylation of catalase was decreased in H 2 O 2 -treated cells stably expressing c-Abl(K-R) or Arg(K-R). Moreover, tyrosine phosphorylation of catalase in response to H 2 O 2 was attenuated by treating MCF-7 cells with the c-Abl inhibitor STI571. Involvement of the Tyr 231 and Tyr 386 sites is further supported by the demonstration that phosphorylation of the Y231F and Y386F mutants in vivo was substantially decreased compared with that of wild-type catalase. The SH2 domains of c-Abl and Arg precipitated catalase from cells, but failed to bind to in vitro labeled catalase. In this regard, in vitro phosphorylation of catalase by c-Abl or Arg conferred binding of catalase to the c-Abl and Arg SH2 domains (data not shown). These findings support a model in which binding of the c-Abl and Arg SH3 domains facilitates tyrosine phosphorylation of catalase and then interaction through the c-Abl and Arg SH2 domains.  (42). Importantly, catalase blocks loss of the mitochondrial transmembrane potential associated with H 2 O 2 accumulation (7). In concert with previous findings (23,24), our results show that cells deficient in c-Abl or Arg exhibit an attenuated apoptotic response to oxidative stress. In this context, the available evidence indicates that activation of both c-Abl and Arg by H 2 O 2 is necessary for induction of apoptosis (24,30). By contrast, the results also show that cells deficient in both c-Abl and Arg are hypersensitive to H 2 O 2 -induced apoptosis. These findings can be explained, at least in part, by a model in which the functions of c-Abl and Arg differ at low (0.25-1.0 mM H 2 O 2 ) and high (2.0 mM H 2 O 2 ) levels of ROS-induced stress (Fig. 8). Thus, at lower ROS levels, c-Abl and Arg phosphorylate catalase and thereby stimulate catalase activity. In the event that ROS levels continue to increase, c-Abl and Arg dissociate from catalase and confer an apoptotic response. In the absence of both c-Abl and Arg, marked increases in H 2 O 2 levels can occur as a result of catalase dysfunction. Under these conditions, ROS-induced disruption of the mitochondrial transmembrane potential is probably sufficient to mediate the apoptotic response (Fig. 8).
Findings in c-Abl-and Arg-deficient Cells May Explain the Phenotype of Mice with Disruption of Both Genes-Mice with targeted disruption of the c-abl gene are born runted and exhibit head and eye abnormalities (43). Mice deficient in Arg develop normally, but the presence of multiple behavioral abnormalities has suggested that arg Ϫ/Ϫ brains suffer from neurologic dysfunction (44). Embryos deficient in both c-Abl and Arg die before 11 days post coitus with defects in neurulation (44). Notably, c-abl Ϫ/Ϫ arg Ϫ/Ϫ embryos also exhibit massive numbers of apoptotic cells in all tissues (44). Our findings indicate that extensive apoptosis in c-abl Ϫ/Ϫ arg Ϫ/Ϫ mice could be a consequence of increased sensitivity of these cells to ROSinduced apoptosis. The demonstration that cells deficient in either c-Abl or Arg exhibited increased H 2 O 2 levels, but less than those found in the double mutants, could also explain why the extensive apoptosis is not observed in c-abl Ϫ/Ϫ or arg Ϫ/Ϫ mice. Alternatively, the absence of extensive apoptosis in the c-abl Ϫ/Ϫ or arg Ϫ/Ϫ mice could be explained by the attenuated apoptotic response to oxidative stress in c-Abl-or Arg-deficient cells. Our findings in MEFs from the deficient mice suggest that c-Abl and Arg function in protecting cells from H 2 O 2induced damage by regulating catalase-mediated degradation of H 2 O 2 . The results also suggest that c-Abl and Arg represent a potential switch that regulates an anti-apoptotic function of catalase to one that, in the presence of uncontrollable H 2 O 2 levels, confers the apoptotic response to oxidative stress.