Differential Roles of Akt, Rac, and Ral in R-Ras-Mediated Cellular Transformation, Adhesion, and Survival

Abstract

Multiple biological functions have been ascribed to the Ras-related G protein R-Ras. These include the ability to transform NIH 3T3 fibroblasts, the promotion of cell adhesion, and the regulation of apoptotic responses in hematopoietic cells. To investigate the signaling mechanisms responsible for these biological phenotypes, we compared three R-Ras effector loop mutants (S61, G63, and C66) for their relative biological and biochemical properties. While the S61 mutant retained the ability to cause transformation, both the G63 and the C66 mutants were defective in this biological activity. On the other hand, while both the S61 and the C66 mutants failed to promote cell adhesion and survival in 32D cells, the G63 mutant retained the ability to induce these biological activities. Thus, the ability of R-Ras to transform cells could be dissociated from its propensity to promote cell adhesion and survival. Although the transformation-competent S61 mutant bound preferentially to c-Raf, it only weakly stimulated the mitogen-activated protein kinase (MAPK) activity, and a dominant negative mutant of MEK did not significantly perturb R-Ras oncogenicity. Instead, a dominant negative mutant of phosphatidylinositol 3-kinase (PI3-K) drastically inhibited the oncogenic potential of R-Ras. Interestingly, the ability of the G63 mutant to induce cell adhesion and survival was closely associated with the PI3-K-dependent signaling cascades. To further delineate R-Ras downstream signaling events, we observed that while a dominant negative mutant of Akt/protein kinase inhibited the ability of R-Ras to promote cell survival, both dominant negative mutants of Rac and Ral suppressed cell adhesion stimulated by R-Ras. Thus, the biological actions of R-Ras are mediated by multiple effectors, with PI3-K-dependent signaling cascades being critical to its functions.

The human R-Ras gene was first isolated by low-stringency hybridization with a viral H-ras probe, and its predicted gene product has a molecular mass of approximately 23 kDa (29). Unlike the proteins encoded by the H-, K- or N-ras oncogene, R-Ras protein has a unique 30-amino-acid (aa) sequence in its N terminus, but it otherwise shows extensive sequence similarity in the catalytic domain responsible for effector and guanine nucleotide binding. The C terminus of R-Ras, which displays the highest sequence divergence among Ras family members, shows features not found in other small G proteins except its closest relative, TC21 (5, 10, 16). Based on these C-terminal sequences, both TC21 and R-Ras are predicted to be modified posttranslationally by geranylgeranyltransferase. In addition, R-Ras also diverges biologically from Ras in its failure to induce neurite outgrowth in PC12 cells and does not promote DNA synthesis or membrane ruffling in Swiss 3T3 cells (41).

Previously, members of our laboratory and others have demonstrated that R-Ras could readily transform rodent fibroblasts when mutated at amino acids 12 and 61 (8, 45). In contrast to the ras oncogenes, the ability of R-Ras to transform NIH 3T3 cells is much weaker in terms of focus-forming efficiency, morphological criteria, and the latency of tumor formation in animals inoculated with the R-Ras transfectants. Although mitogen-activated protein kinase (MAPK) activity was found to be up-regulated in NIH 3T3 cells stably transformed by an R-Ras oncogene (8), the signaling cascades responsible for R-Ras transforming potential still remain unidentified.

More recently, a GTPase-deficient mutant of R-Ras, R-Ras V38, has been shown to promote cell adhesion by enhancing the affinity state of integrin receptors (58). The wide spectrum of integrin receptors (α₅β₁, α₄β₁, and α_vβ₃) that are activated and the failure of the wild-type R-Ras to induce cell adhesion led to the speculation that R-Ras plays a central role in the inside-out signaling of integrins (15). This is in contrast to the outside-in mode of integrin signaling, which requires the initial binding of integrins to specific substrates of the extracellular matrix. Thus, R-Ras may be a component of an intracellular signaling pathway that regulates the affinity state of integrins. The mechanisms and signaling events responsible for this novel biological property of R-Ras are thus far unknown. It is also paradoxical that cellular transformation caused by the ras oncogenes is usually associated with the down-regulation of integrins and the loss of focal adhesion complexes (35). These findings give rise to the speculation that R-Ras signaling events are distinct from those of the conventional ras oncogenes.

R-Ras has also been shown to promote cell survival in an interleukin-3 (IL-3)-dependent mouse pro-B-cell line, BaF3 (50). The fact that wortmannin and PD98059 are able to inhibit this novel function of R-Ras suggests that both phosphatidylinositol 3-kinase (PI3-K) and MAPK signaling cascades are involved. A growing body of evidence suggests that one of the downstream mediators of PI3-K actions, Akt, plays a critical role in mediating survival signals propagated from growth factors and cytokines (1, 9, 12, 13, 32, 46). Indeed, R-Ras has been shown to stimulate Akt activity in transient-transfection assays (30), raising the possibility that R-Ras prevents cell death through the activation of Akt.

The region of the Ras protein encompassing amino acids 32 to 40 constitutes the effector-binding site that is present in all members of the Ras subfamily of small GTPases. Single-point mutations introduced into this region have resulted in Ras mutants that show differential capacities to interact with downstream substrates and that display partially transformed phenotypes (19, 23, 42, 53, 57). For example, while an H-Ras S35 mutant binds only c-Raf, activates MAPK, and promotes DNA synthesis, it is defective in inducing membrane ruffling. In contrast, an H-Ras C40 mutant binds specifically to PI3-K and is able to induce membrane ruffling when microinjected into REF-52 cells (19). Also, while an H-Ras G37 mutant is defective in promoting all of these biological phenotypes, it nevertheless binds preferentially to RalGDS (23). Interestingly, coexpression of pairwise combinations of these H-Ras effector loop mutants partially restores the full Ras transforming activity (23, 42). On the basis of these findings, Ras transformation could be conceived as the result of the input of a multitude of signaling cascades. To test the hypothesis that multiple signaling pathways are responsible for R-Ras-induced biological functions, we generated a panel of R-Ras effector loop mutants and examined them for their relative biological and biochemical properties.

MATERIALS AND METHODS

Cell culture.

The NIH 3T3 and COS-7 cell lines were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% calf serum (CS). The 293T and BOSC cell lines were maintained in DMEM supplemented with 10% fetal calf serum (FCS). Both the IL-3-dependent cell lines 32D and BaF3 (kindly provided by Mitchell Goldfarb) were maintained in RPMI medium containing 15% FCS and 10% WEHI-conditioned medium.

NIH 3T3 cells carrying different plasmid constructs were derived by transfecting 1.5 × 10⁵ cells with 1 μg of DNA by a standard calcium phosphate precipitation method (55). Transfectants were selected in Geneticin (750 μg/ml) and were passaged once prior to the characterization of growth properties in vitro and in vivo. For the generation of 32D transfectants, around 5 × 10⁶ cells in 0.4 ml were electroporated with 10 μg of plasmid DNA by a modified electroporation procedure described previously (34). After electroporation, serial dilutions were performed in 24-well plates to obtain clonal subpopulations of cells in selection medium containing Geneticin (400 μg/ml). To generate BaF3 transfectants, individual cDNA was first cloned into the retroviral vector pBabepuro, and 5 μg of each construct was transfected into the packaging cell line BOSC. Retroviral particles were collected and their titers were determined on NIH 3T3 cells by selecting in medium containing 2 μg of puromycin per ml, with titers of ∼10⁵ to 10⁶ CFU/ml being routinely obtained. For standard infection, ∼10⁶ viral particles were incubated with 5 × 10⁵ BaF3 cells in a final volume of 2 ml in the presence of Polybrene (4 μg/ml) for 5 h. Infected cells were then subjected to selection in 1.5 μg of puromycin per ml, and a mass population of selected cells was used for subsequent analysis. To generate 32D cells expressing different dominant negative mutants, a similar retroviral infection protocol was utilized. In this case, biological assays were performed only with transfectants that had been in selection medium for no more than 10 days.

Plasmids.

The R-Ras S61, G63, and C66 mutant plasmids were generated by PCR with an oncogenic R-Ras mutant, R-Ras L87, as a template. A BssHI (GCGCGC)-tagged forward primer, 5′ GTGGGCAAGAGCGCGCTGACCATC, and individual PstI (CTGCAG)-tagged reversed primers which contain single base substitutions in the codons encoding amino acids 61, 63, and 66 were used in PCR mutagenesis reactions. Oligonucleotides used were as follows: for S61, 5′ ATCCACACTGCAGATCTTCGTGTAGGAGTCCTCAATAcTGGGGTC; for G63, 5′ ATCCACACTGCAGATCTTCGTGTAGGAGTCCcCAATAGT; and for C66, 5′ ATCCACACTGCAGATCTTCGTGcAGGAGTC, where lowercase letters indicate the substitutions and underlining indicates the tag. The resulting amplified products were cloned into the BssHII-PstI sites of the parental R-Ras L87 cDNA, and the authenticity of all constructs was confirmed by sequencing analysis. These R-Ras effector loop mutants, together with the parental R-Ras L87 cDNA, were cloned into the BamHI-SalI sites of the eukaryotic expression vector pCEV29, the NheI-SalI sites of the mouse mammary tumor virus (MMTV) promoter-containing vector pMAMneo, the BglII-EcoRI sites of the AU5 epitope-containing vector pCEFLKZAU5 (kindly provided by Silvio Gutkind), the BamHI-SalI sites of the retroviral vector pBabepuro, and the BamHI-SalI sites of the yeast expression vector pEG202. For the yeast two-hybrid interaction experiments, the Ras-binding domains (RBD) of human c-Raf (aa 48 to 176) and mouse RalGDS (aa 726 to 853) were amplified by PCR and cloned into the EcoRI-XhoI sites of the yeast expression plasmid pJG4-5. Expression plasmids used in immune-complex kinase transient-transfection assays included pLTR-HA-erk2 (25), CMV5-HA-Akt, and CMV6-myr-Akt (1, 9) and have all been described previously. For experiments using various dominant inhibitory molecules in blocking R-Ras transformation, all cDNAs were cloned into the expression vector pCEV29-CAT (kindly provided by Makoto Igarashi). These include dominant negative mutants of MEK (MEKA) (kindly provided by Silvio Gutkind); PI3-K (p85 ΔiSH2-N) (a generous gift from Julian Downward); Rac (Rac N17), RhoA (RhoA N19), and Cdc42 (Cdc42 N17) (kindly provided by Toru Miki); Akt (Akt K179M); and RalA N28 (a generous gift from Jacques Camonis). The v-Src cDNA in the expression vector pBabepuro was kindly provided by Toru Ouchi.

Antibodies.

An anti-R-Ras rabbit polyclonal antibody, R732-4, was generated with a glutathione S-transferase (GST)–R-Ras wild-type fusion protein as an immunogen. The antihemagglutinin (anti-HA) antibody was obtained from the Monoclonal Core Facility, The Mount Sinai School of Medicine. All other antibody reagents were purchased from commercial sources: anti-AU5 (Covance), anti-Rac (Upstate Biotechnology), and anti-RalA (Transduction Laboratories).

Immunoprecipitation and Western blot analysis.

Unless otherwise stated, all cell solubilization steps were performed with the standard radioimmunoprecipitation assay buffer. For immunoprecipitation, usually 1 μg of a monoclonal antibody was added to ∼500 μg of cleared lysates, incubated at 4°C for 1 h, and then absorbed onto 30 μl of Gamma Bind G Sepharose beads (Pharmacia) for an additional 1 h at 4°C. Immune complexes were then washed three times with lysis buffer, and bound proteins were eluted by boiling in Laemmli buffer and resolved on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels. Following transfer onto nitrocellulose membranes (Schleicher & Schuell), proteins were detected by sequential incubation with the primary antibody (1:500 dilution) and then either a rabbit anti-mouse secondary antibody (Jackson ImmunoResearch Laboratories) in the case of a mouse monoclonal antibody or ¹²⁵I-labeled protein A (0.2 μCi/ml; ICN) in the case of a rabbit polyclonal antibody. Dried filters were then subjected to autoradiography and the relative intensities of individual bands were quantified by an imaging densitometer (Bio-Rad).

Focus-forming assay.

Approximately 1.5 × 10⁵ NIH 3T3 cells were plated onto a 100-mm-diameter culture dish, and DNA transfection was performed by the standard calcium phosphate precipitation method (55). In all cases, the amount of DNA added was normalized by the addition of a control expression vector, pCEV29-CAT. After transfection, the medium was changed twice a week with DMEM containing 5% CS, and the number and quality of foci were scored every week for up to 3 weeks. All plates were then fixed in 70% methanol and stained with Giemsa solution for further quantification.

Cell adhesion assay.

Equal numbers of 32D cells (10⁶ per well) were plated in triplicate onto 6-well plates precoated with human plasma fibronectin (1 μg/ml; Life Technologies). Transfectants were then either left untreated or treated with the inducer dexamethasone (5 μM) and then were allowed to attach at 37°C for 12 h. Individual wells were washed three times with 2 ml of phosphate-buffered saline (PBS) to remove nonadherent cells, and firmly attached cells were scraped off the plates and counted on a hematocytometer.

Cell survival assay.

32D transfectants were deprived of IL-3 by washing 5 × 10⁵ cells three times with RPMI (15% FCS) medium. The cells were resuspended in medium without or with IL-3 in a final volume of 1 ml. Cells were allowed to incubate for various times, and the percent viable cells was determined by staining with a 0.2% solution of trypan blue (Sigma). For the isolation of apoptotic DNA, 5 × 10⁶ cells were rinsed twice with PBS and solubilized in 500 μl of hypotonic buffer composed of 10 mM Tris-HCl (pH 7.5), 1 mM EDTA, and 0.2% Triton X-100. Following incubation of cell extracts for 30 min on ice, high-molecular-weight DNA was pelleted at 15,000 × g for 10 min. Apoptotic DNA in the supernatant was then precipitated with an equal volume of isopropanol in a 0.1 M NaCl solution. After centrifugation at 15,000 × g for 30 min, DNA samples were resuspended in 10 mM Tris-HCl (pH 7.5)–1 mM EDTA–1 μg of RNase per ml, incubated for 30 min at 37°C, and resolved on a 2% agarose gel.

Yeast two-hybrid system.

All putative R-Ras effectors were cloned into a “prey” expression vector, pJG4-5, and various R-Ras mutants were cloned into a “bait” expression vector, pEG202. By using the yeast strain EGY48, which harbors a lacZ reporter plasmid, pSH18-34, transformation was carried out with 0.5 μg of each expression plasmid in a solution containing 40% polyethylene glycol, 0.1 M lithium acetate (pH 7.5), 10 mM Tris-HCl (pH 7.5), 1 mM EDTA (pH 8.0), and 50 μg of denatured salmon sperm DNA (Bio 101). Yeast cells harboring the appropriate plasmids were selected as viable colonies on agar plates lacking the amino acids uracil, histidine, and tryptophan. At least four independent colonies were then streaked onto plates containing 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) (80 μg/ml) to test for the transactivation of the reporter gene lacZ.

In vitro binding assay.

The RBD of p110α PI3-K (aa 127 to 314) was cloned in frame in the prokaryotic expression vector pGEX-KG. The ∼50-kDa GST-p110 RBD fusion protein was purified on glutathione agarose beads (Molecular Probes). To generate the R-Ras proteins, 10⁶ 293T cells were transfected with 5 μg of various AU5-tagged R-Ras mutants in the expression vector pCEFLKZ-AU5. Approximately 36 h after transfection, cells were harvested in 600 μl of binding buffer composed of 20 mM Tris-HCl (pH 7.5), 50 mM NaCl, 6 mM MgCl₂, 10% glycerol, and 1% NP-40. In general, the lysate prepared from a 100-mm-diameter plate was used in a single binding reaction. For a typical reaction, 50-μl aliquots of the beads containing ∼10 μg of GST-p110 RBD were incubated with 500 μl of 293T cell lysates. Following incubation for 2 h at 4°C, reaction mixtures were washed four times in the same binding buffer, and bound proteins were eluted by boiling in Laemmli buffer and resolved on an SDS-polyacrylamide gel. The extent of binding between different AU5-tagged R-Ras mutants and GST-p110 RBD was then assessed by Western blot analysis with an anti-AU5 monoclonal antibody.

MAPK transient-transfection assay.

For transient-transfection assays, 10⁶ NIH 3T3 cells were transfected by a standard calcium phosphate precipitation method with 2 μg of an HA-epitope-tagged erk2 expression plasmid and 5 μg of different R-Ras mutants in the expression vector pCEV29 (6). Approximately 12 h after transfection, cells were placed in serum-depleted medium for 18 h and then solubilized in 600 μl of a solution containing 20 mM HEPES (pH 7.5), 2.5 mM MgCl₂, 10 mM EGTA (pH 8.0), 1% NP-40, 40 mM β-glycerophosphate, 2 mM Na₃VO₄, 2 mM leupeptin, 2 mM aprotinin, and 1 mM phenylmethylsulfonyl fluoride. Exogenously expressed HA-MAPK/erk2 was immunoprecipitated from total cell lysates with 1 μg of an anti-HA monoclonal antibody coupled to 30 μl of Gamma Bind G Sepharose beads (Pharmacia). Immune complexes were washed three times with 1% NP-40–2 mM Na₃VO₄ in PBS, once with 0.5 mM LiCl–100 mM Tris-HCl (pH 7.5), and once in kinase buffer (see below). Protein G-Sepharose beads were then resuspended in 30 μl of kinase reaction buffer containing 12.5 mM morpholinepropanesulfonic acid (MOPS) (pH 7.5), 12.5 mM β-glycerophosphate, 7.5 mM MgCl₂, 0.5 mM EGTA, 0.5 mM NaF, 0.5 mM Na₃VO₄, 1 μCi of [γ-³²P]ATP, 20 μM ATP, 3.3 mM dithiothreitol, and 60 μg of myelin basic protein (Sigma). Following incubation at 30°C for 30 min, 60 μl of 2× Laemmli buffer was added to stop the reactions. Approximately 20 μl of the solubilized materials was then resolved by SDS–14% PAGE, dried, and exposed to X-ray films at −70°C.

Akt assay.

Approximately 5 × 10⁵ COS-7 cells were transfected with different R-Ras effector loop mutants in the expression vector pCEFLKZAU5 (5 μg) together with an HA-Akt plasmid (1 μg) (13). About 12 h after transfection, cultures were serum starved for 18 h and lysed in a buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 10% glycerol, 1% NP-40, 10 mM NaF, 1 mM Na₃VO₄, 1 mM Na₃P₃O₄, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 2 mM leupeptin, and 2 mM aprotinin. Exogenously expressed HA-Akt was immunoprecipitated with 1 μg of an anti-HA monoclonal antibody coupled to 30 μl of Gamma Bind G Sepharose beads (Pharmacia) at 4°C for 3 h. Immune complexes were washed three times with lysis buffer, once with ice cold water, and once with kinase buffer (20 mM HEPES [pH 7.4], 10 mM MnCl₂, and 10 mM MgCl₂). Kinase reactions were started by adding 30 μl of a mixture containing 1 mM dithiothreitol, 5 μM ATP, 20 μCi of [γ-³²P]ATP, and 3 μg of histone 2B (H2B; Boehringer Mannheim) in kinase reaction buffer. Following incubation at 30°C for 30 min, 60 μl of 2× Laemmli buffer was added to stop the reactions. Approximately 20 μl of the solubilized materials was then resolved on an SDS–14% PAGE gel, dried, and exposed to X-ray films at −70°C.

RESULTS

Differential abilities of R-Ras effector loop mutants to transform NIH 3T3 cells.

Although Ras and R-Ras have identical core effector-binding site sequences, these G proteins display disparate biological and biochemical properties. In addition, it was unclear if the cellular transformation, adhesion, and survival functions of R-Ras were mutually exclusive or were mediated by a common signaling pathway. We generated a panel of three R-Ras effector loop mutants very similar to those described for H-Ras in order to address these uncertainties. Three mutations were introduced in the core effector loop of an R-Ras activated mutant, R-Ras L87, by altering amino acid 61 from threonine to serine (R-Ras S61; amino acid 35 in H-Ras), amino acid 63 from glutamic acid to glycine (R-Ras G63; amino acid 37 in H-Ras), and amino acid 66 from tyrosine to cysteine (R-Ras C66; amino acid 40 in H-Ras) (Fig. 1A). Next, we separately introduced plasmids encoding these mutants into NIH 3T3, 32D, and BaF3 cells in order to determine their relative biological and biochemical properties.

FIG. 1.

Transforming potential of R-Ras effector loop mutants. (A) Schematic representation of R-Ras effector loop mutants and the corresponding region of H-Ras. Amino acid changes are indicated by italics. (B) The transforming ability of different R-Ras effector loop mutants was assayed with NIH 3T3 cells by transfecting 100 ng of R-Ras S61, G63, and C66 mutants. Parallel cultures were transfected with the same amount of the parental R-Ras L87 plasmid (WT) or an empty vector (Neo) as a negative control. Two representative sets of results from four independent experiments are shown. (C) Focus morphology of NIH 3T3 cells transfected with R-Ras effector loop mutants was examined after 3 weeks in culture by Giemsa stain. (D) The expression levels of different R-Ras effector loop mutants were assessed by Western blot (WB) analysis with cell lysates derived from marker-selected mass cultures.

As shown in Fig. 1B, R-Ras effector loop mutants differed significantly in their relative abilities to transform NIH 3T3 cells. While the G63 mutant was completely devoid of transforming activity, the C66 mutant produced only a few foci when similar amounts of DNA were transfected. In contrast, the S61 mutant retained almost 60% of the transforming potential, although the foci produced were considerably smaller (Fig. 1C). Moreover, the mutants’ ability to induce focus formation correlated with the efficiency of the corresponding transfectants in proliferating in semisolid agar (data not shown). Finally, the levels of R-Ras effector loop mutant proteins in marker-selected mass cultures correlated closely with their relative transforming potential (Fig. 1D). Such preferential selection for transformed cells with a higher ectopic expression of an oncogene product has frequently been observed.

Differential abilities of R-Ras effector loop mutants to promote cell adhesion.

To examine the relative abilities of different R-Ras effector loop mutants to induce cell adhesion, different R-Ras mutant plasmids were electroporated into 32D cells. In this case, the expression of R-Ras was under the control of the MMTV long terminal repeat-inducible promoter. As shown in Fig. 2A, when 32D cells harboring the parental R-Ras L87 mutant were exposed to the inducer dexamethasone, a dose-dependent increase in cell adhesion that peaked at ∼24 h was observed (Fig. 2B).

FIG. 2.

Biological effects of ectopic expression of R-Ras L87 in 32D cells. Murine myeloid progenitor stem cells, 32D cells, were electroporated with an R-Ras L87 expression plasmid with expression under the control of an MMTV-inducible promoter. (A) An increase in cell adhesion and morphological conversion of 32D cells from nonadherent rounded cells to adherent spindle-shaped cells after induction with dexamethasone (Dex) for 20 h is shown. (B) Time course analysis of cell adhesion after the addition of 5 μM dexamethasone. The nonadherent and adherent cells at the indicated time points were counted, and they are expressed as percentages of cells attached to the culture plate. The expression of R-Ras protein was assessed by Western blot analysis with an anti-R-Ras polyclonal antibody, R732-4. We speculate that the higher-molecular-weight bands represent a posttranslationally modified form of R-Ras.

Next, we subjected 32D transfectants expressing different R-Ras effector loop mutants to similar adhesion assays. As shown in Fig. 3A, whereas 32D cells expressing the transformation-defective G63 mutant retained ∼70% of the adhesion function, both S61 and C66 transfectants failed to attach to fibronectin-coated plates. Western blot analysis of total cell lysates revealed similar inducibilities of R-Ras protein in different 32D transfectants (Fig. 3B). When similar adhesion assays were performed with the corresponding NIH 3T3 transfectants, the G63 mutant-transfected cells reproducibly displayed a stronger affinity for fibronectin matrix than did the S61 and C66 mutant-transfected cells (data not shown). Our data, therefore, provide evidence that the abilities of R-Ras to induce focus formation and cell adhesion can be dissociated and may therefore be mediated by different downstream signaling molecules.

FIG. 3.

Effects of R-Ras effector loop mutants on cell adhesion. Expression plasmids harboring R-Ras L87 (WT), S61, G63, and C66 cDNAs, as well as a vector control (Neo), were introduced into 32D cells. Transfectants were either left untreated (−) or treated with dexamethasone (+) for 20 h, and the percentages of adherent cells were determined. Cell adhesion was expressed as fold increase in cell attachment with the addition of dexamethasone. Each data point is the mean ± standard deviation of triplicate samples. The results are representative of five independent experiments. Similar results were also obtained in an independent electroporation experiment. (B) The expression level of different R-Ras mutants in 32D transfectants was measured by Western blot (WB) analysis with an anti-R-Ras polyclonal antibody.

Differential abilities of R-Ras effector loop mutants to confer cell survival.

An oncogenic mutant of R-Ras was previously shown to protect a pro-B-cell line, BaF3, from apoptosis upon the removal of IL-3 (50). Since 32D cells also depend on IL-3 for viability, we compared the abilities of different R-Ras effector loop mutants to confer survival on this cell line. Like BaF3 cells, 32D cells deprived of IL-3 lost viability, and nucleosomal DNA fragments appeared after 16 h. However, 32D cells expressing the R-Ras L87 oncogenic mutant efficiently blocked these processes, with ∼80% of the cells remaining viable 16 h after the withdrawal of IL-3 (Fig. 4A and B).

FIG. 4.

Effects of R-Ras effector loop mutants on cell survival. (A) 32D cells transfected with either R-Ras L87 or control vector (Neo) were cultured in the presence (+) or absence (−) of IL-3. Cells were harvested after 16 h and small-molecular-size DNA was extracted and analyzed on an ethidium bromide-stained agarose gel. (B) Cell survival assays were performed by plating 5 × 10⁵ individual 32D transfectants per well in 24-well plates. The same clones of 32D cells used in cell adhesion assays were tested in cell survival assays. Cells were cultured in IL-3-free medium for 24 h in the presence of dexamethasone. Cell survival was measured by trypan blue exclusion. Data are the means ± standard deviations of triplicate samples and are representative of five independent experiments. (C) Similar experiments were performed with BaF3 cells infected with R-Ras L87 as well as S61, G63, C66, mutants. Mass-selected cultures were generated and cell survival assays were performed as described for panel B. The levels of expression of different R-Ras mutants are shown in the lower panel. WB, Western blot.

Interestingly, when similar assays were performed with 32D cells transfected with different R-Ras effector loop mutants, over 50% of the G63 transfectants remained viable in medium lacking IL-3 (Fig. 4B). In contrast, only 10% of the S61 and C66 transfectants survived under similar experimental conditions. To confirm this finding, all three R-Ras effector loop mutants were introduced into BaF3 cells by retroviral transduction, and similar survival assays were then performed. As expected, BaF3 cells expressing different R-Ras effector loop mutants displayed survival properties very similar to those of their 32D counterparts. In this case, while the G63 transfectants remained mostly viable in medium lacking IL-3, cells transfected with the S61 and C66 mutants survived to a much lesser extent (Fig. 4C). These observations led us to conclude that the ability of R-Ras to confer cell survival correlates with its propensity to promote cell adhesion.

R-Ras effector loop mutants display differential capacities for binding to known downstream substrates.

In previous studies using effector loop mutants of both H-Ras and Rac, it was possible to associate a defined biological function with binding to a specific downstream effector (19, 20, 23, 28, 42, 52, 53, 57). Several putative R-Ras effectors were described, namely, c-Raf (48), p110α PI3-K (30), and RalGDS (2, 47). To explore the potential roles of these putative substrates in mediating the biological functions induced by R-Ras, we tested their interactions with different R-Ras effector loop mutants by the yeast two-hybrid interaction trap system. Accordingly, the RBD of p74 c-Raf and RalGDS were cloned into the yeast expression vector pJG4-5. To optimize the probability of nuclear localization of various R-Ras mutants, an arginine residue was substituted for the conserved C-terminal cysteine residue that is important for membrane targeting. As shown in Fig. 5A and summarized in Table 1, the relative abilities of different R-Ras mutants to transform NIH 3T3 cells correlated well with the binding to c-Raf since only the S61 mutant showed significant interaction with the c-Raf RBD. Interestingly, the RalGDS RBD displayed a strong interaction with the G63 mutant, which was devoid of transforming activity but retained significant cell adhesion and survival properties. Considerable binding between the RalGDS RBD and the S61 mutant was also observed. However, we failed to observe any detectable interaction between the C66 mutant and the two RBD examined.

FIG. 5.

Interaction of R-Ras effector loop mutants with known effectors. (A) The interaction of different R-Ras effector loop mutants with the RBD of both c-Raf and RalGDS was assessed by the yeast two-hybrid system. Yeast cells harboring the indicated combination were patched onto nutrient plates supplemented with X-Gal. As positive controls, the parental R-Ras L87 and H-Ras R12 mutants were included. Strong interaction was manifested by the appearance of dark blue color. Similar results were obtained in two additional experiments. (B) Binding between p110α PI3-K and different R-Ras effector loop mutants was evaluated by an in vitro pull-down assay. AU5-tagged R-Ras wild-type (Ctr) as well as R-Ras L87 (WT), S61, G63, and C66 mutant cDNAs were transfected into 293T cells. Approximately 36 h after transfection, cells were solubilized in NP-40 lysis buffer. An equal amount of cell extract was added to GST-p110 RBD-bound glutathione beads and binding proceeded for 2 h at 4°C. Bound proteins were eluted with sample buffer and subjected to SDS-PAGE analysis. The extent of binding of R-Ras effector loop mutants to GST-p110 RBD was determined by blotting with an anti-AU5 antibody. Results are expressed as fold differences relative to the control following densitometric analysis. Equivalent amounts of 293T lysates and eluted materials were also blotted separately with anti-AU5 and anti-GST antibodies to confirm equal loading. Data are the means ± standard deviations of results derived from three independent experiments. The slight variations among different AU5-R-Ras mutant proteins in each lane have been normalized for the calculation of relative binding capacity. WB, Western blot.

TABLE 1.

Differential abilities of R-Ras effector loop mutants to induce various biological phenotypes

R-Ras L87 protein	Capacity for binding toa:			Transformationb	Cell adhesionc	Cell survivalc
	Raf	RalGDS	PI3-K
Wild type	+++	+++	+++	++	+++	+++
Mutants
S61	++	+	+	+	+	+
G63	−	++	++	−	++	++
C66	−	−	+	±	+	+

^a+++, very strong binding; ++, strong binding; +, weak binding; −, no binding. 

^b++, >100 foci; +, <100 foci; ±, <10 foci; −, no foci. 

^c+++, very strong phenotype; ++, strong phenotype; +, weak phenotype. 

Next, we investigated if R-Ras effector loop mutants would show differential degrees of binding to the p110α catalytic subunit of PI3-K. Initial attempts to perform binding studies using the yeast two-hybrid system failed to reveal detectable interactions between the RBD of p110α PI3-K with either H-Ras or R-Ras. We attribute this to potential deleterious conformational changes resulting from the fusion of the p110α PI3-K RBD to the LexA transactivation domain. For this reason, we sought to perform this particular interaction study by expressing the RBD of p110α PI3-K as a GST fusion protein, GST-p110 RBD. A similar fusion protein has previously been reported to interact with p21 Ras in vitro in a GTP-dependent manner (44). The GST-p110 RBD protein was first immobilized onto glutathione beads and then incubated with total cell lysates derived from cells expressing different AU5-tagged R-Ras effector loop mutants. As expected, R-Ras L87 showed a fivefold-higher capacity for binding to GST-p110 RBD than the wild-type counterpart (Fig. 5B). In contrast, all three R-Ras effector loop mutants showed a considerably weaker binding when experiments were performed under similar conditions. Nevertheless, the G63 mutant displayed a 2.8-fold-higher capacity for binding to GST-p110 RBD than did the R-Ras wild-type protein. On the other hand, the S61 and C66 mutants elicited increases of only 1.9- and 1.6-fold, respectively. As expected, when similar binding experiments were performed with the H-Ras G37 and C40 mutants, we observed a preferential binding of the C40 mutant to the GST-p110 RBD (data not shown). We conclude from these binding studies that the ability of R-Ras to transform cells seems to be correlated with its ability to bind to the c-Raf RBD and that cell adhesion and survival phenotypes elicited by R-Ras may be mediated through the interaction with the RBD of RalGDS and PI3-K.

R-Ras effector loop mutants display differential abilities to activate known downstream signaling cascades.

To determine whether R-Ras effector loop mutants differ in their biochemical properties, transient-transfection assays were performed to test two signaling cascades known to be activated by R-Ras, namely, MAPK and PI3-K (7, 30, 49). When MAPK transient-transfection assays were carried out with NIH 3T3 cells, we observed a modest stimulation of MAPK activity by all R-Ras mutants. The overall magnitude was substantially lower than that of an H-Ras R12 mutant, which routinely displayed a 20-fold stimulatory effect. However, the relative ability of different R-Ras effector loop mutants to stimulate MAPK appeared to correlate with their respective transforming potential (Fig. 6A). In this case, the maximal activation was achieved by the S61 mutant, which exhibited an ∼4-fold stimulation of MAPK activity. To investigate whether the observed R-Ras-induced MAPK activity was propagated through the conventional Raf-MEK-MAPK linear pathway, transient-transfection assays were performed in the presence of the anti-MEK drug PD98059. As a negative control, an anti-PI3-K inhibitor, wortmannin, was tested in parallel cultures. As shown in Fig. 6B, PD98059 effectively abolished the ability of R-Ras to activate MAPK, while the addition of wortmannin did not have a drastic inhibitory effect.

FIG. 6.

Differential effects of R-Ras effector loop mutants in the activation of MAPK. (A) A kinase transient-transfection assay was performed with NIH 3T3 cells by transfecting 5 μg of the indicated plasmids together with 2 μg of an HA-tagged erk2 plasmid. MAPK activity was measured by using myelin basic protein (MBP) as a substrate in the presence of [γ-³²P]ATP. Kinase reaction mixtures were resolved by SDS–14% PAGE, dried, and exposed to the X-ray films. The extent of phosphorylation of MBP (kinase) was determined by densitometric analysis and is expressed as fold activation relative to the control (Neo). Equal aliquots of beads after immunoprecipitation were loaded onto an SDS–12.5% polyacrylamide gel to ascertain the expression of HA-erk2. The expression levels of p23 R-Ras in the transfected cells were examined with a polyclonal antibody. Data are the means ± standard deviations of triplicate samples and are representative of five independent experiments. WT, wild type. (B) The ability of R-Ras to stimulate MAPK activity was measured in cultures pretreated for 30 min prior to lysis either with dimethyl sulfoxide (Ctr) or in the presence of PD98059 (20 μM) and wortmannin (100 nM).

Next, we examined the relative abilities of R-Ras effector loop mutants to activate PI3-K-dependent signaling cascades. For this, we utilized one of the PI3-K downstream substrates, Akt, as a readout for its lipid kinase activity. This is of particular relevance since Akt has been extensively implicated in mediating cell survival under numerous apoptotic conditions (12, 22, 24, 46). Transient-transfection assays were performed in COS-7 cells by cotransfecting an HA-tagged wild-type Akt (HA-Akt) and individual AU5-tagged R-Ras effector loop mutant plasmids. As shown in Fig. 7A, the R-Ras L87 mutant routinely induced an ∼10-fold stimulation of Akt activity as measured by the phosphorylation of the substrate, H2B. This level of activation was comparable to those reported by Marte et al. for similar transient-transfection assays performed with A14 cells (30). As for the effector loop mutants, whereas the S61 and C66 mutants activated Akt kinase activity 3- and 1.9-fold, respectively, the G63 mutant was able to produce a 7-fold stimulatory effect. As expected, a constitutively active myristoylated form of Akt, myr-Akt, phosphorylated H2B ∼30-fold. To examine the dependency on PI3-K of these kinase reactions, we performed similar experiments in the presence of increasing concentrations of wortmannin. As shown in Fig. 7B, wortmannin prevented R-Ras L87 from activating Akt in a dose-dependent manner, with a half-maximal inhibition achieved at ∼3 nM. We conclude from these data that the observed cell adhesion and survival phenotypes exhibited by the G63 mutant correlate with its ability to stimulate PI3-K-dependent signaling events.

FIG. 7.

Differential effects of R-Ras effector loop mutants on the activation of Akt. (A) A kinase transient-transfection assay was performed by transfecting COS-7 cells with 5 μg of the indicated plasmids together with 1 μg of an HA-tagged Akt plasmid. Akt activity was measured with H2B as a substrate in the presence of [γ-³²P]ATP. Kinase reaction mixtures were resolved by SDS–14% PAGE, dried, and exposed to the X-ray films. The extent of phosphorylation of H2B was determined by densitometric analysis and is expressed as fold activation relative to the control (Ctr). Equal aliquots of beads after immunoprecipitation were loaded onto an SDS–12.5% polyacrylamide gel to ascertain the expression of HA-Akt. The expression levels of the AU5-tagged R-Ras mutants in the transfected cells were examined with an anti-AU5 monoclonal antibody. Data are the means ± standard deviations of triplicate samples and are representative of five independent experiments. WT, wild type; WB, Western blot. (B) The effect of wortmannin on R-Ras-induced Akt activity was assessed by similar kinase assays except that wortmannin at the indicated concentrations was added to the culture medium 30 min prior to the kinase assay. DMSO, dimethyl sulfoxide.

Dependence of R-Ras-induced biological activities on intracellular signaling cascades.

To firmly establish the role of intracellular signaling cascades in R-Ras-induced biological functions, pharmacological compounds and dominant negative mutants were tested in different biological assays. For cellular transformation, we examined a repertoire of dominant negative mutants previously used successfully in dissecting Ras downstream signaling. They were MEKA (7), Rac N17 (37), RhoA N19 (36, 39), Cdc42 N17 (38), p85 ΔiSH2-N (42), Akt K179M (9), and RalA N28 (54). Focus-forming assays were performed with NIH 3T3 cells by cotransfecting 50 ng of the oncogenic R-Ras L87 mutant together with an ∼20-fold molar excess (1.0 μg) of different dominant inhibitory mutants. To account for the potential existence of nonspecific inhibitory effects due to the overexpression of these dominant negative mutants, we also included the v-src oncogene as a control. As shown in Fig. 8, a dominant negative mutant of MEK, MEKA, which represents an inhibitor of the MAPK signaling cascade, did not show significant inhibitory effect on the transforming activity of R-Ras. The same mutant, however, suppressed H-Ras R12 transformation by 60% (data not shown). Interestingly, the p85 ΔiSH2-N mutant, which has been shown to inhibit PI3-K-dependent signaling events, drastically dampened the ability of R-Ras to induce foci, by ∼80%. Among the downstream signaling intermediates of PI3-K, dominant negative mutants of neither Akt nor Cdc42 had any significant effects on R-Ras transformation. However, Rac N17, RhoA N19, and RalA N28 inhibited R-Ras transformation by 10, 30, and 15%, respectively. In most cases, the transforming activity of v-src was relatively not perturbed by the coexpression of these dominant negative mutants. In addition, we did not observe a drastic reduction in the number of marker selectable colonies in NIH 3T3 cells transfected with similar amounts of these dominant negative mutants (data not shown).

FIG. 8.

Effects of dominant negative mutants on R-Ras-induced transformation. Approximately 50 ng of R-Ras L87 and 60 ng of v-src were transfected into NIH 3T3 cells along with 1.0 μg of the indicated dominant negative mutants. The effects of dominant negative mutants on the ability of R-Ras to induce focus formation are expressed as the percentage of focus-forming units compared to the vector control (Ctr) plate. Results are the means ± standard deviations of triplicate plates from two independent experiments which were repeated at least twice.

Previous studies have alluded to the importance of both PI3-K- and MAPK-dependent pathways in the regulation of cell adhesion and survival (13, 18). Since the measurement of these biological properties was of short-term nature, they were amenable to analysis through the use specific pharmacological inhibitors. Accordingly, we used two well-characterized inhibitors, wortmannin and PD98059, as specific blockers of the PI3-K- and MAPK-dependent signaling cascades, respectively. When the aforementioned cell adhesion and survival assays were performed with R-Ras L87-expressing 32D (R-Ras L87/32D) cells in the presence of either drug, we observed differential effects on these biological processes. While wortmannin at the concentration of 100 nM significantly disrupted the attachment of R-Ras L87/32D cells to fibronectin, by ∼40%, PD98059 at 20 μM failed to inhibit cell adhesion markedly (Fig. 9A). As for cell survival, both wortmannin and PD98059 significantly reduced the viability of R-Ras L87/32D cells, from 65% to 12 and 38%, respectively (Fig. 9B).

FIG. 9.

Effects of PI3-K- and MAPK-specific inhibitors on R-Ras-induced cell adhesion and survival. 32D cells transfected with R-Ras L87 were treated with wortmannin (100 nM), PD98059 (20 μM), or dimethyl sulfoxide (Ctr) for 20 h, and cell adhesion (A) and survival (B) assays were performed as described in the legends to Fig. 3 and 4. Data are the means ± standard deviations of triplicate samples and are representative of three independent experiments. dex, dexamethasone.

The preferential sensitivity of R-Ras-induced biological functions to wortmannin prompted us to investigate the role of known PI3-K downstream substrates in cell adhesion and survival. Several signaling molecules have been shown to be the targets of PI3-K: they are Rac, Akt, protein kinase C, and p⁷⁰S6K (reviewed in reference 14). Among these effectors, Rac and Akt have previously been demonstrated to play vital roles in cell adhesion and survival, respectively (11, 13). In addition, since the ability of the G63 mutant to promote cell adhesion and survival in 32D cells correlated with RalGDS binding, we also examined the importance of the Ral pathway in these two R-Ras biological functions. For this, dominant negative mutants of Akt (Akt K179M), Rac (Rac N17), and RalA (RalA N28) were introduced into R-Ras L87/32D cells by retroviral transduction. As shown in Fig. 10, we observed that while the ectopic expression of the Akt K179M mutant had no detectable effect on cell adhesion, it reduced the ability of R-Ras L87/32D cells to survive in IL-3-depleted medium by ∼65%. When similar experiments were performed with Rac N17 and RalA N28 mutant-infected cells, we observed ∼65 and ∼40% reductions in their adhesion properties, respectively. In contrast, the expression of these two mutants failed to alter the viability of R-Ras L87/32D cells. In all cases, the expression of these inhibitory mutants was confirmed and the magnitude of inducibility of R-Ras L87 protein was not significantly altered in various transfectants (Fig. 10B). Furthermore, these inhibitory mutants did not have any detectable effect on the proliferation rate of the parental 32D cells (data not shown). We conclude from all these studies that PI3-K-dependent signaling pathways play a critical role in R-Ras-induced biological functions. Specifically, the transforming potential of R-Ras is highly dependent on the PI3-K pathway, with the potential involvement of small GTPases, Rac, RhoA, and RalA. On the other hand, the ability of R-Ras to promote cell survival is mediated primarily by Akt, while its positive effect on cell adhesion requires both functional Rac and Ral GTPases.

FIG. 10.

Effects of dominant negative mutants on R-Ras-induced cell adhesion and survival. Dominant negative mutants of Akt, Rac, and RalA were introduced into 32D cells harboring an R-Ras L87 mutant by retroviral transduction. As a negative control, a parallel culture was infected with a control vector (Ctr) containing the chloramphenicol acetyltransferase gene. Following coselection in Geneticin and puromycin for 7 days, four populations of each cotransfectant were analyzed for expression of the dominant negative mutants and R-Ras expression. Cell adhesion (A) and survival (B) assays were performed as described in the legends to Fig. 3 and 4. Data are the means ± standard deviations of triplicate samples and are representative of three independent experiments. To ascertain the expression of R-Ras L87, Akt K179M, Rac N17, and RalA N28, Western blot (WB) analysis was performed with their respective antibodies (lower panels).

DISCUSSION

Signaling mechanisms responsible for the transforming potential of the R-Ras oncogene.

Transformed foci induced by R-Ras display features that are distinct from those exhibited by the ras oncogenes. They are consist mostly of compact foci of less refractile cells (17, 45). The transformation-competent R-Ras S61 mutant used in this study is similar to the corresponding H-Ras S35 mutant in binding preferentially to c-Raf. However, unlike the latter, R-Ras S61 mutant only weakly activates MAPK. It is therefore possible that novel Raf gene family members or unknown R-Ras substrates are responsible for the transforming activity observed for the S61 mutant. Moreover, our observations are consistent with previous findings that although R-Ras binds c-Raf in vitro (41, 48), it inevitably fails to elicit robust stimulation of c-Raf, B-Raf, and MAPK in vivo (17, 30). Indeed, the highest MAPK activation detected among different R-Ras mutants was only 4-fold, whereas 20- to 30-fold increases were routinely observed for the H-Ras R12 mutant. Although our results obtained with PD98059 tend to argue for the role of MEK in R-Ras downstream signaling, we do not exclude the possibility that the observed MAPK activation is due to the creation of an autocrine stimulatory loop. Taking all these observations together in addition to our finding that a dominant inhibitory mutant of MEK fails to block R-Ras transformation, we conclude that the MAPK pathway is unlikely to play a substantial role in R-Ras transformation.

Paradoxically, while a dominant negative mutant of PI3-K markedly inhibits R-Ras transformation, the G63 mutant, which activates PI3-K-dependent signaling, is completely devoid of focus-forming activity. This discrepancy could be explained by the fact that PI3-K pathway is essential but not sufficient to elicit morphological transformation (27). Alternatively, signaling events activated by the G63 mutant may be below the threshold required for morphological transformation to be registered. Indeed, cotransfection of the G63 mutant with either the S61 or C66 mutant significantly enhances the overall number of foci being detected (unpublished observations). These findings imply that R-Ras utilizes multiple signaling pathways in mediating its oncogenic effects. A detailed analysis of different parameters of transformation, such as DNA synthesis, cytoskeletal organization, and low-serum and soft-agar proliferation for each R-Ras effector loop mutant may help to define the role of relevant signaling molecules.

Signaling mechanisms responsible for the ability of R-Ras to induce cell adhesion.

The ability of R-Ras to induce cell adhesion in 32D cells closely resembles that described by Zhang et al. (58). In fact, we also observed similar increases in adhesion to fibronectin substratum by BaF3 and NIH 3T3 cells transfected with the R-Ras L87 mutant (unpublished observations). Interestingly, increase in cell adhesion is one of the hallmarks of differentiation of myeloid stem cells (31). In fact, we have observed an increase in the surface expression of the myeloid differentiation marker mac1 integrin in 32D cells harboring an active R-Ras gene (unpublished observations). Whether R-Ras could be activated in response to myeloid-cell-specific differentiation factors remains to be determined.

Interestingly, the Ras-Raf-MAPK pathway has been demonstrated to suppress the ability of integrin receptors to bind to their respective ligands (18). Paradoxically, Ras has also been shown to activate PI3-K (43), which, according to our results, would be predicted to promote cell adhesion. However, its potent stimulatory effect on MAPK pathway may have a dominant effect, giving rise to an overall decrease in cell adhesion. It is, therefore, tempting to speculate that under physiological conditions, cell adhesion is finely controlled by both Ras and R-Ras GTPases.

The fact that the G63 mutant retains significant capacity to induce cell adhesion in 32D cells and stimulates PI3-K-dependent signaling cascades suggests that R-Ras promotes cell adhesion via PI3-K. This conclusion is supported by the observation that cell adhesion induced by R-Ras is more sensitive to the inhibitory effect of wortmannin than of PD98059. More importantly, inhibition of one of the PI3-K downstream substrates, Rac, markedly reduces the ability of R-Ras to induce cell adhesion. This observation is consistent with the finding that an activated Rac promotes cell spreading and adhesion of T lymphocytes (11). In contrast to R-Ras, the modulation of cell adhesion by Rac does not involve an increase in the affinity state of the surface-bound integrin receptors. These findings imply that Rac may not be a direct downstream mediator of the R-Ras cell adhesion function. It is possible that additional novel effectors of R-Ras are present or that Rac may play an indirect role in the post-integrin receptor signaling events. Indeed, R-Ras has recently been implicated in antagonizing Ras effects on integrins through a small death effector domain-containing protein, PEA-15 (40).

Unexpectedly, the R-Ras G63 mutant used in this study interacted with the RBD of both p110 PI3-K and RalGDS. This particular mutant therefore behaves differently from the corresponding H-Ras G37 mutant, which binds only to RalGDS (23). This discrepancy may reflect the nature of the activating mutants being used (H-Ras V12 versus R-Ras L87), as well as other subtle differences in the regions flanking the core effector loops of Ras and R-Ras. Nevertheless, the expression of a dominant negative RalA N28 mutant blocks R-Ras-induced cell adhesion by 40%, suggesting that Ral-related GTPases may be involved in cell adhesion. Recent evidence has implicated Ral in platelet activation (56), and a Ral binding protein, RalBP1/RLIP1, has also been isolated (4, 21, 33). Interestingly, this novel substrate of Ral possesses GAP activity towards Rac and Cdc42 (21), which may in turn modulate cytoskeletal organization and cell adhesion behavior.

Signaling mechanisms that are responsible for the ability of R-Ras to confer cell survival.

The ability of the G63 mutant to promote survival of both BaF3 and 32D cells correlates well with its ability to activate the survival kinase, Akt. The fact that wortmannin effectively blocks R-Ras-induced cell survival suggests a role for PI3-K in this event. More importantly, a dominant negative mutant of Akt, but not those of Rac and RalA, significantly reduces the ability of R-Ras to confer cell survival. Therefore, signaling pathways responsible for R-Ras-induced cell adhesion and survival bifurcate at the level of PI3-K. Indeed, when cell survival assays were performed on poly-2-hydroxyethyl methacrylate-coated plates, we did not observe a reduction in the ability of R-Ras to promote survival of 32D cells (unpublished observations). However, in adherent cultures, R-Ras may activate Akt and promote survival indirectly through its intrinsic ability to activate integrins. In support of this hypothesis, King et al. have shown that integrin-stimulated Akt is mediated by PI3-K (26). In addition, we also observed a substantial suppression of the ability of R-Ras to promote cell survival in the presence of the anti-MEK drug PD98059. This inhibition can be explained by the fact that the treatment of parental 32D cells with PD9809 potentiates cell death in the absence of IL-3 (unpublished observations). On the other hand, Sutor et al. have recently reported that the activation of MEK in a 32D-like cell line, FDC-P1, is dependent on both PI3-K and A-Raf (49). These data raise the possibility that cell survival phenotype promoted by R-Ras in 32D cells may indeed be dependent on both PI3-K and MEK. However, R-Ras activates MAPK activity in 32D cells only 4-fold, while H-Ras R12 stimulates MAPK >50-fold (unpublished observations). It will be interesting to test if A-Raf is activated by R-Ras and whether dominant negative A-Raf can block R-Ras in promoting survival of 32D cells.

Unexpectedly, our data are in conflict with a previous study in which an R-Ras V38 mutant was shown to be proapoptotic in 32D cells upon IL-3 withdrawal (51). We attributed this discrepancy to a potential clonal variation in 32D cells being used and to the fact that different R-Ras oncogenic mutants (V38 versus L87) were being tested. In support of our observations, Suzuki et al. have reported that the ectopic expression of an R-Ras L87 mutant in BaF3 cells prevents cell death under IL-3-deprived conditions (50). In addition, Ahmed et al. have also demonstrated that the ectopic expression of a catalytically active Akt in BaF3 cells inhibits apoptosis (1). Future studies utilizing dominant negative mutants to inhibit endogenous R-Ras protein may shed light on the role of R-Ras in cell survival.

Implications of R-Ras signaling events for its physiological function.

The normal physiological function of R-Ras still remains elusive. On the basis of our data as well as those of others, it is clear that R-Ras, in sharp contrast to the products of prototypic ras oncogenes, is a weak activator of the MAPK pathway. However, its preferential activation of PI3-K-dependent signaling cascades may play a crucial role in physiological states whereby adhesion and survival are preferred over proliferation. For example, R-Ras may be a critical factor in the maintenance of cells in their quiescent state during differentiation and senescence. The finding of a single R-Ras effector loop mutant having both cell adhesion and survival functions is not unexpected. Indeed, the loss of cell adhesion in cultured epithelial cells and the associated cell death caused by anoikis are well documented (24). In addition, cell attachment to substratum is a prerequisite for cell cycle progression from G₁ to S phase (3). In contrast, cells undergoing mitosis are usually characterized by rounding and a reduction in adhesion. It is, therefore, reasonable to assume that R-Ras together with Ras may regulate cell cycle progression in a coordinate fashion. It will be of great interest to examine the relative GTP-bound states of both R-Ras and Ras during cell division. Findings derived from studies of this type will have important implications for the understanding of such important physiological processes as cell division and differentiation.