The v-Crk Oncogene Enhances Cell Survival and Induces Activation of Protein Kinase B/Akt*

The v-Crk oncogene encodes an adaptor protein containing an SH2 domain and an SH3 domain. v-Crk-transformed fibroblast cells display enhanced tyrosine phosphorylation levels, and the v-Crk protein localizes in focal adhesions, suggesting that transformation may be due to enhanced focal complex signaling. Here we investigated the mechanism of transformation and found that v-Crk-transformed NIH 3T3 cells display growth rates and serum requirements similar to control cells. However, v-Crk enhanced survival in conditions of serum starvation. Both an intact SH2 and SH3 domain are required; moreover, SH2 mutants displayed dominant interfering properties, enhancing cell death. Using other cell death-inducing stimuli, it appeared that v-Crk in general inhibits apoptosis and enhances cell survival. In search of the signaling pathways involved, we found that v-Crk-transformed cells show constitutively higher levels of phospho-protein kinase B (PKB)/Akt and PKB/Akt activity, especially in conditions of serum starvation. These data strongly suggest involvement of the phosphatidylinositol 3-kinase/PKB survival pathway in the v-Crk-induced protection against apoptosis. In accordance, inhibition of this pathway by wortmannin or LY924002 reduced protection against starvation-induced apoptosis. In addition to the phosphatidylinositol 3-kinase/PKB pathway, a MEK-dependent pathway and an unknown additional pathway are also implicated in resistance against apoptosis. Activation of survival pathways may be the most important function of v-Crk in its oncogenic properties.

The v-Crk oncogene has been isolated as the transforming gene of the avian sarcoma retrovirus CT10 (1). The v-Crk gene product is an adaptor protein containing a viral gag portion fused with a SH2 and SH3 domain. The corresponding cellular proto-oncogene was shown to exist in two splice variants, Crk I (28 kDa) and Crk II (42 kDa), of which Crk II encodes an additional N-terminal SH3 domain (2). In addition, another closely related gene has been identified, CRKL, which is similar to Crk II (3). Both c-Crk II and CRKL are ubiquitously expressed (4), suggesting an important cellular function.
How does v-Crk induce oncogenicity? As an adaptor protein the v-Crk protein does not possess catalytic activity. Nevertheless the v-Crk-transformed cells contain enhanced levels of proteins phosphorylated on tyrosine residues, indicating the activation of a tyrosine kinase or the inactivation of a phosphotyrosine phosphatase (for review, see Ref. 5). Since tyrosine kinase activities often have been implicated in signal transduction and oncogenesis, this is likely to be involved. In v-Crktransformed cells the most prominently phosphorylated proteins are Paxillin and p130 cas (6 -8), both adapter proteins described as have signaling functions in focal adhesions (9). Focal adhesions are the sites where cells are bound to the extracellular matrix through integrin receptors that are connected to the intracellular actin cytoskeleton. In addition to their structural function, focal adhesions also serve to transduce extracellular signals via multiprotein complexes. Focal adhesion signaling has important functions in the regulation of cell morphology, cell cycle progression, and cell survival (10,11). v-Crk not only induced the phosphorylation of Paxillin and p130 cas but also bound stably to their phosphorylated tyrosine residues through its SH2 domain (6 -8). In accordance, v-Crk was shown to localize in focal adhesions (12). Through its SH3 domain, v-Crk can bind to several downstream signaling molecules like DOCK180, Sos, and C3G and the kinases Abl, KHS, and HPK1 (for review, see Ref. 5). These data suggest that v-Crk enhances focal complex signaling, but the mechanism resulting in oncogenicity is not known. v-Crk not only induces oncogenicity in fibroblast model systems; various data also suggest the involvement of Crk in human tumors. Particularly, CRKL is a major substrate protein of leukemogenic BCR-ABL proteins (13). These are the gene products of the chimeric BCR/ABL oncogenes that are generated by the well known Philadelphia chromosome translocation and which cause chronic myelogenous leukemia. CRKL-and p130 cas -mediated signaling is thought to play an important role in this disease (for review, see Ref. 14). Furthermore, recently BCAR1, a protein showing 91% amino acid identity with mouse and rat p130 cas , has been identified by a random retroviral mutagenesis approach as being involved in breast cancer (15). Particularly, high levels of BCAR1 were associated with poor survival of patients (16). Since p130 cas contains many binding sites for Crk (8), Crk may be the main downstream effector. In both cases the mechanism of tumorigenesis is not fully clear. Various aspects of focal adhesion signaling are shown to be stimulated by Crk, which might contribute to the oncogenic properties. Cell morphology and regulation of the actin cytoskeleton are affected by Crk-mediated activation of Rho-like GTPases, particularly Rac (17,18). In neuronal cells, v-Crk was shown to delay apoptosis (19), whereas in experiments using Xenopus lysates, Crk appeared to have an apoptosis stimulatory role (20). Furthermore, Crk may play a role in cell cycle progression (21). In addition, however, various data indicate a role for Crk in growth factorstimulation (e.g. Refs. 22 and 23).
In this study we set out to investigate which mechanism could be involved in the v-Crk-induced oncogenesis by analyzing stably transformed NIH 3T3 fibroblast cell lines expressing v-Crk or v-Crk mutants that were previously described (12). We and others showed that v-Crk induces in fibroblasts the capacity to grow in soft agar. This property is highly correlated with tumorigenicity and is considered to reflect anchorage independence of growth. In soft agar growth assays, v-Crk is less oncogenic than e.g. oncogenic Ras variants or v-Src. Similarly v-Crk induces an intermediary phenotype with respect to induction of a highly refractile spindle-shaped morphology in fibroblast cells (12). In this study we compared the oncogenic properties of v-Crk-transformed NIH 3T3 cells with those of RasL61-transformed cells and found that v-Crk does not stimulate serum-independent growth. In contrast, our data indicate that the major v-Crk-induced effect is the protection of the fibroblast cells against cell death and particularly against apoptosis, which is also observed in the RasL61-transformed cells. Analysis of the pathways involved in the protection by v-Crk indicates the activation of the PKB 1 /Akt survival pathway as one of the mechanisms.
Cell Culture-The NIH 3T3 cells stably transfected with v-Crk and with v-CrkSH2Ϫ and SH3Ϫ have previously been described (12). NIH 3T3 cells stably transfected with RasL61 were kindly provided by Dr. B. Burgering. COS-1-, NIH 3T3 (ATCC)-, and NIH 3T3-derived cell lines were grown in Dulbecco's modified Eagle's medium supplemented with 7.5% fetal calf serum (Life Technologies, Inc.) at 37°C in a humidified atmosphere with 5% CO 2 .
Cell Growth and Viability Assays, Nuclear Condensation, and UV Treatment-Relative cell number and viability was determined using MTT as a substrate (24). In short, cells were cultured in 12-well plates and were incubated with the MTT solution for 60 min at 37°C. Cells were lysed, and the formazan salt was solubilized in 0.04 N HCl in isopropanol under continuous shaking. The absorbance was measured at 570 nm with a microplate reader (Bio-Rad). The background value without cells (A 570 ϳ 0.010) was subtracted. For the cell lines used, the MTT signal in the presented range was linearly correlated with the cell number, as counted by phase contrast microscopy. The MTT signal per cell was roughly similar for the cell lines used. For serum starvation survival experiments cells were incubated in serum-free medium for the indicated times. For UVC treatment (25) cells were grown in a 12-well dish. Before UVC treatment cells were incubated for 20 h in Dulbecco's modified Eagle's medium containing 0.075% fetal calf serum. Medium was removed, and UV treatment was performed in a Stratalinker (Stratagene) with the indicated doses (standard: 200 J/m 2 ). After treatment with UV light, 1 ml of serum-free medium was added, and cells were incubated at 37°C. At the indicated time points random field photographs were taken, and cell death was morphologically scored as the percentage of cells that are in the process of blebbing, that have a condensed nucleus, or that have lysed. Also, for visualization of condensed nuclei, the cells were grown on glass coverslips and similarly treated. They were fixed in ice-cold methanol (Ϫ20°C) and subsequently incubated with 1 g/ml DAPI in methanol for 5 min (room temperature). Cells were washed once with methanol and once with phosphate-buffered saline (PBS), air-dried, and mounted in 9 l of Mowiol (Hoechst, Frankfurt, FRG) supplemented with 0.1% paraphenylene diamine. Fluorescent DNA-DAPI complexes were visualized in a Leica-Aristoplan microscope equipped with epi-illumination. Alternatively, when phase contrast pictures were taken, cells were fixed in 4% paraformaldehyde in PBS for 15 min, rinsed with PBS, permeabilized with ice-cold 0.1% Triton X-100, 0.1% sodium citrate in PBS for 2 min, rinsed with PBS, stained with 0.4 g/ml DAPI in PBS for 1 min, rinsed twice with PBS, and air-dried (26).
Annexin V Plus PI-For annexin V staining, cells on glass coverslips were rinsed rapidly in Dulbecco's modified Eagle's medium (DMEM), 20 mM Hepes, 0.1% bovine serum albumin (BSA) and incubated in DMEM, 20 mM Hepes, 0.1% BSA containing 0.25 g of propidium iodine and 1 l of concentrated annexin V-FITC (Nexins Research) for 10 min at room temperature in the dark. After washing in Dulbecco's modified Eagle's medium, 20 mM Hepes, 0.1% bovine serum albumin cells were mounted in Mowiol/ paraphenylene diamine.
Western Blotting-Cell lysates and immunoprecipitates were solubilized in Laemmli sample buffer containing 0.1 M dithiothreitol and separated by SDS-polyacrylamide gel electrophoresis (8%). Proteins were blotted onto polyvinylidene difluoride filters that were blocked with 3% milk proteins (w/v, Protifar) in TBST (20 mM Tris-HCl, pH 7.6, 140 mM NaCl, and 0.1% Tween 20) for at least 1 h. The filters were probed with primary antibody in 0.3% protifar in TBST for at least 1 h (or overnight), washed four times in TBST, incubated with the appropriate secondary antibody (horseradish peroxidase-coupled goat antimouse or goat anti-rabbit) in 0.3% protifar in TBST buffer for 30 min and washed again four times with TBST. Proteins were visualized using enhanced chemiluminescence (Renaissance). For quantification of protein amounts, a densitometer (Molecular Dynamics) and Image-Quant software were used. Antibodies against PKB/Akt (1:2000) were from Transduction Laboratories (Lexington, KY). Where indicated, the phospho-specific Akt/PKB antibodies against phospho-Ser-473 (1:1500) was obtained from New England Biolabs (Beverly, MA), whereas in addition, phospho-Ser-473-PKB antibodies (1:1500) were used from BIOSOURCE International (Camarillo, CA) since the Biolabs antibodies gave significantly more background signal with nonphosphorylated PKB/Akt. The secondary antibodies (1:10000) were from Jackson Im-munoResearch (Pennsylvania, PA).
Transfections and Immunoprecipitations-Transient transfections in COS-1 cells were performed at 40% confluency with 4 g of DNA/ 6-cm dish using the Fugene (Roche Molecular Biochemicals) transfection protocol. For co-transfection 2 g of each plasmid/6-cm dish were transfected. Twenty-four hours after transfection, cells were serumstarved for 16 h. Stimulated and unstimulated cells were washed once with ice-cold phosphate-buffered saline and lysed in lysis buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 5 mM EDTA, 1% Triton X-100) with freshly added 1 mM phenylmethylsulfonyl fluoride, 40 mM ␤-glycerophosphate, 1 mM sodium vanadate, 50 mM sodium fluoride, and 10 g/ml aprotinin. Lysates were incubated on ice for 5 min and centrifuged for 5 min at 15,000 ϫ g, and supernatants were precleared by incubating with protein A-Sepharose beads (Amersham Pharmacia Biotech) for 1 h at 4°C and centrifuging for 5 min at 15,000 ϫ g. HA-PKB was immune-precipitated from aliquots (200 g of protein) of the precleared extracts using 2 g of the monoclonal anti-HA antibody (12CA5) coupled to a (1:1) mixture of protein G-agarose beads and protein A-Sepharose beads (Sigma). Immunoprecipitates were washed twice with lysis buffer and twice with low salt buffer (50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 ) before Western blot analysis or twice with high salt buffer (50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , and 0.5 M LiCl) and twice with low salt buffer before activity measurements. One-third of the immunoprecipitates was used for Western blot analyses. Immune precipitation of endogenously expressed PKB/Akt from NIH 3T3-derived cell lines was performed similarly, except that cells were grown in 10-cm dishes and immune precipitation was performed with Immobilized Akt 1G1 monoclonal antibody (New England Biolabs).
In Vitro Kinase Assays for PKB/Akt-PKB activity was assayed with the Crosstide peptide (GRPRTSSFAEG) as a substrate as previously described (27). Immunoprecipitates were incubated with 45 l of kinase assay mixture (50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 1 mM dithiothreitol, 30 M Crosstide peptide, 1 M specific peptide inhibitor of cyclic AMP-dependent protein kinase (PKI) (Bachem, Bubendorf, Switzerland), 50 M unlabeled ATP, and 3 Ci of [␥-32 P]ATP. After incubation for 20 min at 30°C under continuous shaking, reactions were stopped by the addition of 5 l 200 mM EDTA. Proteins were precipitated by the addition of 12.5 l 25% trichloroacetic acid and centrifuged for 1 min at 14,000 rpm. Supernatants containing the phosphorylated peptide were spotted onto p81 phosphocellulose filters (Whatman), washed three times with 1% (v/v) orthophosphoric acid, and analyzed by Cerenkov counting. Under the conditions used, the kinase assays are linear for at least 60 min.

Growth Characteristics of v-Crk-transformed Cells-To in-
vestigate how v-Crk induces oncogenicity we first analyzed serum dependence and growth rate of stably transfected NIH 3T3 fibroblast cell lines expressing v-Crk (3T3v-Crk) or v-Crk mutants, mutated either in the SH2 (3T3SH2Ϫ) or the SH3 (3T3SH3Ϫ) domain of v-Crk. These previously described cell lines (12) were compared with control NIH 3T3 cells (3T3) and RasL61-transformed cells (3T3Ras). Growth of these cell lines in the presence of 0.5, 2.0, or 7.5% fetal calf serum was monitored daily using a MTT assay. Fig. 1 shows a representative time point, showing the relative cell densities after 3 days.
Clearly no profound differences in growth were observed between 3T3v-Crk and NIH 3T3 control cells that could explain oncogenicity. It appeared that their doubling time is similar and their growth is similarly correlated with the serum concentration. At 0.5% serum, the cells complete their cell cycle, but in contrast to RasL61-transformed cells, none of the other cell lines continued to grow, showing a strict dependence on serum. The only difference observed is that 3T3v-Crk cells grow to higher saturation density than the control cells (12), as was apparent with the MTT assay after 4 days of growth (not shown) and was confirmed by phase contrast microscopy. For comparison, highly oncogenic 3T3Ras-transformed cells require less serum, grow more rapidly, and grow to even higher saturation densities (Fig. 1). They grow even in serum-free conditions (see next paragraph). The 3T3SH3Ϫ cell line that expresses the SH3Ϫ mutant v-Crk protein has growth characteristics similar to NIH 3T3. In contrast, the 3T3SH2Ϫ cells grow more slowly, reach lower cell densities, and die in the presence of 0.5% fetal calf serum. Apparently the SH2 mutant of v-Crk has a negative effect, both with respect to cell growth and survival. As compared with RasL61, v-Crk clearly induces more restricted oncogenic properties.

3T3v-Crk Cells Survive in Serum Starvation Conditions-To
investigate a possible role of v-Crk in cell survival, the effect of serum deprivation was further investigated. Serum deprivation results in rapid cell death of control NIH 3T3 cells; within 48 h, ϳ95% of the cells die (Fig. 2). The 3T3v-Crk cells on the other hand are more resistant against serum withdrawal; ϳ40 -50% of the cells survive for even longer than a week, as is also seen by phase contrast microscopy. RasL61-transformed cells continue growing in serum-free conditions. The surviving v-Crk-transformed cells were fully capable of growing when serum was added again (not shown). To determine the role of the SH2 and SH3 domains of v-Crk in mediating protection against cell death, the mutant cell lines were assayed. The 3T3SH3Ϫ cell line dies with similar kinetics to NIH 3T3, whereas the 3T3SH2Ϫ cell line is significantly more sensitive to serum starvation-induced cell death (Fig. 2). Thus, v-Crk protects against cell death, and both the SH2 and SH3 domain of v-Crk are required for this protection. In addition, the v-Crk mutant protein mutated in the SH2 domain (v-CrkSH2Ϫ) enhances cell death, having a dominant negative function with respect to survival.
v-Crk Protects against Apoptosis-Figs. 1 and 2 indicate that the main growth difference between NIH 3T3-and v-Crk-transformed NIH 3T3 cells is the ability of 3T3v-Crk to survive in serum-free conditions. Inhibition of apoptosis by v-Crk would provide a physiological selective advantage in vivo for v-Crktransformed cells. The MTT assay (Fig. 2) measures the number of surviving cells but does not distinguish between death by necrosis or apoptosis. Actually, an important characteristic of apoptosis is that during the programmed cell death the cell membrane remains impermeable, and the cell content is not released. Using phase contrast microscopy we noticed that blebbing apoptotic cells are still capable of forming dark blue formazan crystals in the MTT assay. A decrease in MTT signal due to apoptosis is therefore only observed when (apoptotic) cells detach from the substratum or when they have lysed due to secondary necrosis. Using nuclear DAPI staining, we found that cell death upon serum starvation (48 h) results in nuclear condensation, characteristic of apoptosis (Fig. 3A). To further prove that cell death occurred due to apoptosis, annexin V binding assays were performed. Annexin V binds to phosphatidylserine, a phospholipid that normally is present at the inner leaflet of the cell membrane but during apoptosis is exposed at the outer leaflet. In the presence of serum, neither NIH 3T3 cells nor 3T3v-Crk cells bound annexin V. However, after 24 h of serum depletion the NIH 3T3 cells heavily stained with annexin V-FITC, and 3T3v-Crk cultures stained significantly less (Fig. 3B, panels B and E). Annexin V-FITC staining of starved NIH 3T3 cells revealed that many cell fragments (presumably derived from blebs: see next paragraph) are shedded off. In addition, some NIH 3T3 cells are in necrosis, as is revealed by their plasma membrane permeability for propidium iodine (Fig. 3B, panel C, red arrow). In conclusion, both DAPI and annexin V staining demonstrate the protective effects of v-Crk expression against apoptosis induced by serum withdrawal.
To test if v-Crk also protects against apoptosis induced by other methods, UVC treatment was applied. After overnight serum starvation, cells were treated with various doses of UVC light and analyzed for apoptosis by morphological criteria and by analyzing nuclear condensation by DAPI staining. After treatment with 200 J/m 2 UV light for 5 h, more than 80% of the NIH 3T3 cells are apoptotic or in (secondary) necrosis. Fig. 4A illustrates that UVC irradiation induces apoptosis, as shown both by nuclear condensation and cell membrane blebbing. A higher magnification of a typical blebbing cell shows many cell protrusions that are connected to the cell body only through a thin stalk, as is nicely illustrated in cells expressing green fluorescent protein (Fig. 4B). Later in the apoptotic process the protrusions are pinched off while still containing green fluorescent protein, implying that the plasma membrane is still intact, which is characteristic of apoptosis (results not shown). The  Fig. 4C. It shows that the v-Crk-transformed cells are protected against UV-induced apoptosis, as compared with NIH 3T3 cells. Similar conclusions were reached using other UV doses (50, 100, 400, 600 J/m 2 ) and at different time points after UV treatment (results not shown). Again, survival of the v-CrkSH3 mutant-expressing cells is similar to the NIH 3T3 cells. The SH2 mutant-expressing cells are slightly more sensitive to UV-induced apoptosis. RasL61-transformed cells do not undergo apoptosis at all in these conditions. Other apoptotic stimuli including H 2 O 2 and osmotic stress were also tested (not shown). In all cases tested, v-Crk provided protection against apoptosis.
Which Signaling Pathways Are Involved in Crk-mediated Protection against Apoptosis?-As a first approach in determining which signaling pathways might be activated by v-Crk to inhibit apoptosis, we used pharmacological inhibitors. The best-described anti-apoptotic pathway is the PI 3-kinase-PKB/ Akt pathway, which can be inhibited by the PI 3-kinase inhibitors LY924002 and wortmannin. Because v-Crk has been implicated in activation of the Ras-Raf-MEK-Erk pathway (28,29), which plays a protective role in various systems (30,31), the inhibitor PD98059 was used to inhibit activation of MEK. Another MAPK implicated in apoptosis is the p38 MAPK, which is specifically inhibited by SB202190 (32). Fig. 5A shows that upon serum starvation, the v-Crk-induced protection against apoptosis is strongly decreased by LY294002 (10 M). The same was found when PI 3-kinase was inhibited with wortmannin (100 nM; not shown), indicating that v-Crk-induced protection may be mediated by the PI 3-kinase/PKB pathway. Survival of the 3T3Ras cells was not affected by LY294002, whereas PD98059 (10 M) largely abrogated survival. Inhibition of MEK modestly decreased protection of 3T3v-Crk cells, implying that activation of the MAPK pathway, although to a lesser extent, is also required for full protection. In contrast, inhibition of the p38 MAPK did not influence survival. Induction of apoptosis by UV-induced cell damage yielded a slightly different result. Again p38 MAPK does not play a role of importance. However, Fig. 5B shows that both PI 3-kinase-and MEK-mediated survival contribute only weakly to v-Crk-induced protection. Both in serum starvation-induced apoptosis and UV-induced apoptosis, inhibition of both pathways simultaneously hardly showed an additive effect. Even in the presence of both inhibitors 3T3v-Crk was more resistant than NIH 3T3, indicating that v-Crk induces yet additional protective effects, the nature of which still has to be determined.

Protection against Apoptosis Is Correlated with Enhanced PKB/Akt Activation and Enhanced PKB/Akt Phosphorylation on Ser-473-The inhibitory effect of PI 3-kinase inhibitors on
v-Crk-induced protection against apoptosis caused us to investigate the activity of PKB/Akt in v-Crk-expressing cells. PKB/ Akt was immune-precipitated and assayed for kinase activity using a specific peptide (Crosstide) as a substrate (27). Whereas serum-starved v-Crk-transformed cells and control NIH 3T3 cells contain similar amounts of PKB/Akt (Fig. 6B), it appeared that the PKB/Akt activity was at least 3-fold more in v-Crk-transformed cells than in control NIH 3T3 cells (Fig. 6A). As a positive control, PKB/Akt activation by PDGF was also assayed. For both, the addition of PDGF (20 ng/ml for 10 min) resulted in a further induction of PKB/Akt activity. Thus, v-Crk expression results in a 3-fold activation of PKB/Akt even in the absence of growth factors and induces an enhancement of PDGF stimulation of PKB/Akt activity.
Generally PKB/Akt activity is correlated with phosphorylation of Ser-473 and Thr-308, both contributing to maximal PKB/Akt activity (for reviews, see Refs. [33][34][35]. Using phosphospecific antibodies, we found that upon starvation, the Ser(P)-473 PKB levels in 3T3v-Crk are increased as compared with NIH 3T3 (Fig. 6B), which correlates well with the PKB/Akt activity in these cells (as determined in Fig. 6A). The phosphorylation of PKB/Akt on Ser-473 is abrogated by inhibition of PI 3-kinase with the chemical inhibitor LY924002 (Fig. 6C), also indicating the involvement of PI 3-kinase-mediated signaling in v-Crk-induced PKB/Akt activation. Again, both the SH2 and SH3 domains of v-Crk are required; Ser-473 PKB phosphorylation in the SH3 mutant is the same as in NIH 3T3 cells, whereas the SH2 mutant displays decreased levels of Ser-473 PKB phosphorylation. Thus, the PKB/Akt phosphorylation levels in these cell lines correspond well to their resistance against serum starvation-induced apoptosis (Fig. 2). The RasL61transformed cells display low levels of PKB/Akt phosphorylation, indicating that other signaling pathways may be involved in protection against apoptosis in these cells, as was also clear in Fig. 5A. In conclusion, v-Crk expression induces a PI 3-kinasedependent activation of PKB/Akt and stimulates the phosphorylation of Ser-473.
v-Crk Also Enhances PKB/Akt Phosphorylation upon Transient Transfection in COS-1 Cells-To further investigate the activation of PKB/Akt by v-Crk, transient transfection experiments were performed. COS-1 cells were transfected with HAtagged PKB/Akt or co-transfected with HA-tagged PKB/Akt and v-Crk together. PKB/Akt was immune-precipitated and analyzed for activity (Fig. 7A) and for phosphorylation state using the phosphospecific antibody against Ser-473 (Fig. 7B). Fig. 7A shows that immune-precipitated PKB/Akt displays a basal level of PKB activity even in serum starvation conditions (bars 1 and 2), which is stimulated by PDGF treatment (bar 4) and is PI 3-kinase-dependent (bars 3 and 5). Co-transfection with v-Crk results in clearly enhanced PKB/Akt activity, both in serum starvation conditions (bars 7 versus 2) and after PDGF induction (bars 9 versus 4), which is also mostly PI 3-kinase-dependent (bars 8 and 10). Thus, these results demonstrate that transient expression of v-Crk also activates a pathway resulting in PKB/Akt activation. Fig. 7B shows the parallel analysis of the PKB/Akt phosphorylation state. It appears that in this transient assay system, v-Crk also induces the phosphorylation of Ser-473, which is nicely correlated with the observed PKB/Akt activity. The phosphorylation of tran-siently expressed PKB/Akt as well as the enhanced phosphorylation due to v-Crk expression is completely PI 3-kinase-dependent (lanes 3, 5, and 8, 10, respectively). DISCUSSION

v-Crk Protects against Apoptosis-
The v-Crk oncogene product induces in fibroblasts the capacity to form tumors (1), which is correlated with a transformed phenotype and the capacity to grow in soft agar. Here we further investigated which mecha-FIG. 6. PKB/Akt is constitutively activated in 3T3v-Crk cells. A, relative in vitro PKB/Akt kinase activity (using the Crosstide peptide as a substrate) measured after immune-precipitation of PKB/Akt from NIH 3T3 and 3T3v-Crk cells. The PKB/Akt activity of serum-starved 3T3v-Crk is (arbitrarily) taken as 100%. SF, serum starvation for 20 h with 0.075% fetal calf serum; ϩPDGF, induction with 20 ng/ml PDGF for 10 min. B, Western blot analysis of lysates of stably transfected NIH 3T3 cell lines obtained after a 24-h serum starvation (SF) using ␣-phospho-Ser-473-PKB/Akt antibody (Biolabs). The control panel, treated with ␣PKB, shows the total PKB/Akt independent of the phosphorylation state. ϩPDGF, induction with 20 ng/ml PDGF for 10 min. C. Western blot analysis of lysates of NIH 3T3 and 3T3v-Crk, using ␣-phospho-Ser-473-PKB/Akt antibody (BIOSOURCE International). Lysates were obtained after starvation for 24 h (SF) and after starvation plus, in addition, treatment with LY294002 (10 M) for 30 min (SFϩLY). nisms are involved in the transformation. Compared with Rastransformed cells, v-Crk appears mildly transforming. Rastransformed cells are more aggressive in soft agar growth and when injected in animals. In correlation with this, we found that v-Crk does not induce obvious growth advantage as compared with NIH 3T3. In contrast to the activated Ras mutant RasL61, v-Crk does not induce growth in serum-free conditions, and the serum requirement and growth rate are similar to NIH 3T3 control cells. However, similar to RasL61-transformed cells upon serum starvation, v-Crk-expressing cells were largely protected against cell death. In view of the physiological relevance of apoptosis, we further investigated whether the protection against cell death by v-Crk could be due to the inhibition of apoptosis.
Whereas 3T3v-Crk cells were clearly protected against cell death by serum starvation, the propensity of fibroblasts to undergo apoptosis is not very well defined. The hallmark of apoptosis is that upon a death stimulus, the cell content is largely degraded within the boundary of the intact cell membranes, and therefore, no leakage of cell content and degrading enzymes occurs. This is generally accompanied by the typical apoptotic changes in morphology of cells called blebbing. We found that the majority of dead fibroblast cells generated by serum starvation did not meet this criterion since their membranes were permeable for ethidium bromide and propidium iodine. They did, however, display condensed nuclei, which is another feature taken as a criterion for apoptosis. The explanation is that in tissue cultures, apoptosis is followed by secondary necrosis. Annexin V staining of serum-starved cells provided further evidence that the NIH 3T3 cells indeed undergo apoptosis before lysis. Annexin V staining showed that many large cell fragments (most likely derived from blebs) are shedded off from the dying cells, which made this method less useful for quantitative purposes. Nevertheless, the 3T3v-Crk cells were protected also by this criterion.
Other methods of apoptosis induction were also tested. A very suitable method appeared to be UVC treatment, which induced clear blebbing of the cells within hours. During blebbing the fibroblasts displayed cellular protrusions typically attached to the cell only by a stalk, which were subsequently pinched off. The blebbing preceded nuclear condensation, and both were significantly delayed in 3T3v-Crk. Other apoptosisinducing agents like hydrogen peroxide and osmotic stress also showed that v-Crk protects against apoptosis, but these displayed less the blebbing phenotype. Whereas serum starvation most likely induces apoptosis by deregulating cellular metabolism, UVC will do so by radiation damage. v-Crk was shown to protect in both cases.
v-Crk Activates Multiple Protection Pathways-In search of signaling pathways involved in protection against apoptosis and considering that focal adhesion signaling may be stimulated by v-Crk, we investigated the possible involvement of the well known anti-apoptotic PI 3-kinase/PKB pathway. Furthermore, among the signaling pathways that have been reported to be activated by v-Crk (5,13,28,29), the Ras-Erk pathway has been ascribed anti-apoptotic properties (30,31) and was, therefore, also investigated. Using pharmacological inhibitors, we found that v-Crk-mediated protection against serum starvation-induced apoptosis was largely abrogated by inhibiting the PI 3-kinase/PKB pathway, whereas the MEK-Erk pathway played a smaller role. Indeed, v-Crk induced in a PI 3-kinasedependent manner the phosphorylation and activation of PKB/ Akt, as is outlined below. Remarkably, the protection against UV-induced apoptosis appears to involve other mechanisms, since inactivation of PKB/Akt by LY924002 treatment only poorly decreased resistance against UV-induced apoptosis. This indicates that v-Crk activates, in addition to the PI 3-kinase/PKB pathway, another pathway(s) that plays a role in protection against UV-induced apoptosis. Comparably, the Ras oncogene was recently also reported to inhibit different apoptotic induction mechanisms independently of each other, which in that case was attributed to independent activation of both the Ras/PKB and the Ras/Erk pathway (36). Inhibition of the Ras-Erk pathway by MEK inhibitor PD98059 showed that in our cells the MAPK pathway does provide some protection. However, whereas our data indicate that in 3T3RasL61 cells this is the major protection pathway upon serum starvation, this is not the case in 3T3v-Crk. For UV-treated cells, the effect of the MEK inhibitor PD98059 is even stronger in NIH 3T3 cells than in 3T3v-Crk, indicating that v-Crk does not play a role in the MEK-Erk-mediated protection. In accordance with this, it was recently reported that Erk is not activated by v-Crk (37) and, therefore, cannot be causally correlated with protection against cell death. Others have shown that the functions of Erk and c-Crk II represent two distinct pathways involved in protection from apoptosis in a three-dimensional collagen matrix (38). Our data indicate that yet another pathway is activated by v-Crk that protects against apoptosis. The nature of this pathway still has to be determined.
v-Crk Regulation of PKB-Further analysis of the activation state of PKB mediated by v-Crk showed a clear correlation FIG. 7. v-Crk also enhances PKB/Akt phosphorylation upon transient transfection in COS-1 cells. COS-1 cells were transfected with HA-PKB/Akt with or without co-transfected v-Crk. HA-PKB was immune-precipitated, and immune precipitates were split and subjected to PKB/Akt kinase activity assays (A) and Western blot analysis (B). A, PKB/Akt activity assays, as described under "Experimental Procedures." The PKB/Akt activity of cells transfected with PKB-HA only is (arbitrarily) taken as 100%. B, Western blot analysis with phosphorylation-specific antibodies using ␣-phospho-Ser-473-PKB/Akt antibody (BIOSOURCE International) and control with antibodies recognizing phosphorylation-independent total PKB/Akt (PKB). SF, serum depletion for 24 h; ϩPDGF, induction with 20 ng/ml PDGF for 10 min; ϩLY, pretreatment with LY924002 (10 M) for 30 min. between enhanced phosphorylation of Ser-473, PKB activity, and survival in serum starvation conditions in the cells stably transformed with various v-Crk constructs. Activation of PKB/ Akt by v-Crk also occurred upon transient co-expression of both in COS-1 cells, indicating that it is not the result of the transformed phenotype of the v-Crk transformed cells but directly due to v-Crk-induced signaling. Both the v-Crk-induced PKB/ Akt activation in stably transformed cells and in transient assays are largely PI 3-kinase-dependent. Furthermore, whereas v-Crk activates PKB/Akt, the cell line stably expressing the v-Crk SH2 mutant showed a reduced amount of Ser(P)-473 even in serum-containing conditions (not shown). A lack of phosphorylation correlated with enhanced sensitivity to cell death inducers. In a recent study, it was shown that soft agar growth of v-Crk-transformed cells was abrogated by inhibition of PI 3-kinase, confirming the correlation with oncogenic properties (37). Using the Ser(P)-473-specific antibody, this study showed that v-Crk also induces PKB/Akt phosphorylation in chicken fibroblasts (37).
How v-Crk activates PKB/Akt still has to be established. Ser(P)-473 phosphorylation was abrogated by LY924002, indicating PI 3-kinase dependence. That PI 3-kinase can localize to focal adhesion complexes (39) suggests that PKB/Akt activation might be due to enhanced PI 3-kinase activity in v-Crktransformed cells, correlating with enhanced focal complex signaling. In this respect both focal adhesion kinase and c-Src and p130 cas have been implicated in signaling toward PI 3-kinase (39 -41). PI 3-kinase binds with the SH2 domains of its regulatory subunit, p85, to focal adhesion kinase, Src, and p130 cas . Interestingly, Src has also been shown to stimulate PI 3-kinase, and moreover, Src is activated in v-Crk-transformed cells (6). Furthermore, both focal adhesion kinase and p130 cas are phosphorylated in v-Crk-transformed cells, creating binding sites for PI 3-kinase. This might result in activation of this lipid kinase.
The results obtained with the v-CrkSH2Ϫ mutant indicate that the regulation of PKB/Akt by Crk isoforms is not restricted to the viral v-Crk only. The v-CrkSH2Ϫ mutant protein is considered to act as a dominant negative protein disturbing normal c-Crk function. The v-CrkSH2Ϫ mutant-expressing cell line grew more slowly than control cells, showed down-regulated phosphorylation of PKB/Akt, and displayed enhanced sensitivity for apoptosis. This implies that the v-Crk SH2 protein titrates-out proteins that are required for protection against cell death and suggests that endogenous c-Crk may also have a function in regulating PKB/Akt and protection against cell death. Further investigations are required to determine if the decreased growth rate of the v-Crk SH2 mutant NIH 3T3 cell line is also correlated with PKB/Akt activity or is due to other dominant negative effects by this v-Crk mutant (e.g. see Ref. 21). As indicated in the Introduction, various data suggest that v-Crk and c-Crk play an important role in focal adhesion signaling. Our results are in agreement with a model in which v-Crk enhances focal adhesion complex signaling; previously, it was shown that PKB/Akt can be activated by focal adhesion signaling (42). Similarly, in general the effect of dominant negative v-Crk SH2 may be the disturbance of focal adhesion complex signaling.
In conclusion we show that v-Crk expression in fibroblast cells results in protection against apoptosis induced by various methods. Protection is mediated by activation of several protection pathways, among them the PI 3-kinase/PKB pathway. We show that in fibroblasts, v-Crk activates PKB/Akt and induces phosphorylation of Ser-473. Furthermore, our results with dominant negative forms of v-Crk suggest that endogenously expressed Crk also plays a role in PKB/Akt activation and regulation of apoptosis. Since, normally, cells that are deregulated will undergo apoptosis, inhibition of apoptosis is thought to be a very important requirement in the process of tumorigenesis. Our results suggest that inhibition of apoptosis is the main mechanism involved in the oncogenesis induced by v-Crk.