Transformation Potential of Ras Isoforms Correlates with Activation of Phosphatidylinositol 3-Kinase but Not ERK*

The Ras oncoproteins activate the Raf-MEK-ERK kinase pathway, which plays an important role in cellular transformation. We observed that H-RasV12 exhibited a higher transforming potential than either K-RasV12 or N-RasV12 in both NIH3T3 fibroblasts and RIE-1 rat epithelial cell cultures. Surprisingly N-Ras and K-Ras were more potent than H-Ras in activation of mitogen-activated protein (MAP) kinase activity and ternary complex factor-dependent transcription. In contrast, H-Ras was more effective in activation of phosphatidylinositol 3-kinase (PI3K) and AKT. Co-expression of constitutively active AKT, a downstream target of PI3K, cooperated with H-RasV12, K-RasV12, or N-RasV12 in transformation. Furthermore co-expression of the constitutively active MEK and AKT resulted in focus formation, while neither active MEK1 nor active AKT alone transformed NIH3T3 cells. Our data demonstrated that the transforming potential of Ras was not directly correlated with the ability of Ras to activate the MAP kinase cascade. In contrast, the ability to activate PI3K and AKT correlated with the ability of Ras to induce cellular transformation, suggesting an important role of PI3K-AKT in cellular transformation. Our data also demonstrated that, under these assay conditions, activation of the MAP kinase cascade was not sufficient to induce NIH3T3 cell transformation.

The small GTPase Ras plays a critical role in the regulation of numerous cellular functions, including cell proliferation, differentiation, and oncogenic transformation (1). Ras cycles between the inactive GDP-bound form and the active GTP-bound form. Stimulation of tyrosine kinase receptors results in a rapid activation of Ras that is essential for receptor tyrosine kinase-mediated signal transduction (2,3). Oncogenic mutations of Ras, frequently at amino acid residues 12 and 61, result in the stabilization of Ras in GTP-bound conformation, thereby activating the Ras protein. Many downstream targets of Ras have been identified and characterized (3,4). These Ras effectors specifically bind to the GTP form of Ras and mediate the biochemical responses of Ras activity (5). One of the best characterized downstream targets of Ras is the Raf-mitogen-activated protein (MAP) 1 kinase pathway. Ras has been shown to directly bind to and activate Raf, although the molecular mechanism of Raf activation by Ras is currently not fully understood (6,7). The active Raf then phosphorylates and activates MEK, which in turn phosphorylates and activates ERK, also known as the mitogen-activated protein kinase (8). The activated ERK can phosphorylate and regulate functions of numerous cellular proteins (8,9), one of which is the ternary complex factor responsible for induction of many growth factor-inducible genes (10). Studies with dominant negative mutants of Raf, MEK, and ERK clearly have demonstrated that the MAP kinase pathway plays an essential role in Ras-induced cell transformation. Overexpression of constitutively active MEK stimulates cell growth and transformation in NIH3T3 cells, indicating a major role of the MAP kinase cascade in transformation (11,12). The importance of this pathway in tumor formation is also supported by the observation that mutation in B-Raf has been observed in human cancers, especially in melanomas (13,14) However, it is also clear that other downstream effectors of Ras are involved in Ras transformation (3,4). Phosphatidylinositol 3-kinase (PI3K) is another well characterized downstream target of Ras. Ras can directly bind to and activate PI3K (15). Furthermore PI3K has been shown to cooperate with the Raf-MAP kinase pathway in cellular transformation (16,17). AKT has been shown to be activated by the PI3K pathway (18 -20). The products of PI3K bind to the pleckstrin homology domain of AKT. In addition, the PI3K products bind to and activate phospholipid-dependent kinase 1, which directly phosphorylates and activates AKT (21,22). It has been shown PI3K-AKT is involved in Ras-induced cellular transformation (23,24). AKT plays a major role in inhibiting apoptosis and stimulating cell growth (18).
Three highly related Ras genes, H-, K-, and N-ras are found in the mammalian genome. The three ras gene products share extremely high sequence identity, suggesting they may have similar functions (2). In fact, most of the biochemical and functional characterizations of Ras proteins indicated that the three Ras proteins indeed have similar properties. However, gene knock-out experiments in mice have demonstrated that the physiological functions of the three ras genes are clearly different. Inactivation of N-ras caused little change in phenotype in transgenic mice, while inactivation of K-ras resulted in embryonic lethal phenotypes (25,26). Furthermore in vitro studies also indicate that different Ras proteins may have different functions (5,27). K-Ras is more potent than H-Ras in Raf activation, while H-Ras is more potent than K-Ras in PI3K activation (5,27).
In this study, we compared the three Ras proteins in cell transformation and activation of ERK and PI3K. H-Ras was the most potent of the three in inducing transformation of both NIH3T3 fibroblasts and rat epithelial RIE-1 cells. While showing low activity in transformation, N-Ras and K-Ras were the stronger activators of the MAP kinase and Elk-1-dependent transcription. Interestingly the transformation potentials of Ras isoforms correlated with their abilities to activate PI3K in contrast to the widely accepted model that activation of the MAP kinase pathway plays the major role in Ras-induced transformation. Our results demonstrated the differential activities of the three Ras proteins in transformation and activation of downstream targets and also showed that activation of PI3K, but not MAP kinase, was a limiting factor in Ras-induced cellular transformation.

MATERIALS AND METHODS
Cell Culture and Transfection Conditions-NIH3T3 cells were cultured in DMEM supplemented with 10% calf serum in a humidified atmosphere of 10% CO 2 . Human embryonic kidney (HEK) 293 and rat intestinal epithelial (RIE-1) cells were cultured in DMEM supplemented with 10% fetal bovine serum in a humidified atmosphere of 5% CO 2 . PC12 cells were cultured in DMEM supplemented with 5% fetal bovine serum and 10% horse serum. Transfections of NIH3T3, HEK293, RIE-1, and PC12 cells were performed by LipofectAMINE method as suggested by the manufacturer (Invitrogen).
Plasmids-cDNAs for human H-ras, K-ras, and N-ras were obtained from Drs. Quilliam and Vojtek (28,29). The H-ras cDNA and N-ras were amplified by PCR and subcloned into the EcoRI and BamHI sites of pCDNA3 (Invitrogen). K-ras cDNA was amplified by PCR and subcloned into the BamHI site of pCDNA3. Mutation of ras was created by the Quik Change site-directed mutagenesis (Stratagene). ras plasmids were confirmed by DNA sequencing. The plasmids of HA-p110 and FLAG-p85 were from Dr. Skolnik (New York University Medical Center) (30). Myc-ERK (31) and Elk (32) are laboratory stocks.
Transformation Assay-Transformation of NIH3T3 cells was performed following standard protocols (33). Low passage NIH3T3 cells were plated in 6-well plates (30 mm) at a density of ϳ2 ϫ 10 5 cells/well. pCMV-␤-galactosidase (100 ng) was co-transfected as a control for transfection efficiency. 48 h after transfection, cells were trypsinized and plated onto a 100-mm plate. Cells were kept in DMEM with 6% calf serum. The culture medium was changed every 3-4 days. Approximately 14 days after transfection, cells were stained by crystal violet as described previously (33), and foci were then counted. RIE-1 cells were obtained from Dr. C. J. Der (University of North Carolina at Chapel Hill) (34). Transformation of RIE-1 cells was performed similarly to that for NIH3T3 cells except that the RIE-1 cells were cultured in DMEM with 10% fetal bovine serum. Foci were stained 21 days after transfection.
For stable lines, NIH3T3 stable clones were obtained by selection for resistance to G418 (300 g/ml, Invitrogen). NIH3T3 cells were transfected with Ras or vector (the total amount of DNA was always normalized to 1 g with vector). 48 h after transfection, cells were trypsinized and plated onto 100-mm plates (10 and 20% of the original transfected cells were used for each 100-mm plate). Cells were cultured in DMEM with 10% bovine serum and 300 g/ml G418 for 2 weeks. Morphology of G418-resistant colonies was examined by a light microscope. The G418resistant colonies were isolated and cloned. The purified Ras-expressing clones were plated at low density for morphology analysis.
Soft agar assays were performed according to published methods (33). Stable clones of NIH3T3 cells were plated onto DMEM soft agar plates at a density of 1, 2, and 6 ϫ 10 4 cells/plate (35-mm plate). Plates were incubated for 3 weeks and examined by microscopy.
ERK Kinase Assay-NIH3T3 cells were transfected with a Myc-ERK plasmid together with varying amounts of Ras plasmids (the total amount of DNA in transfection was kept constant by addition of vector DNA). 24 h after transfection, cells were serum-starved for 12 h. Cells were washed in phosphate-buffered saline and then lysed in buffer (10 mM Tris-HCl, pH 7.5, 2 mM EDTA, 150 mM NaCl, 1% Nonidet P-40, 50 mM NaF, 1 mM dithiothreitol, 1 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, 2 g/ml leupeptin). Cell lysates were cleared by centrifugation in a microcentrifuge for 5 min and incubated with 1 g of 9E10 antibody (Babco) for 2 h at 4°C followed by addition of protein G-Sepharose (Amersham Biosciences) for 1 h. Immunoprecipitated kinases were washed three times with lysis buffer followed by one wash with 25 mM HEPES, pH 7.5, 1 mM dithiothreitol, and 0.5 mM EDTA. Kinase activities for ERK were measured as described previously using 2 g of GST-Elk as a substrate (35).
Phosphatidylinositol 3-Kinase and AKT Assay-NIH3T3 or HEK293 cells were transfected with 100 ng of p110 and p85 plasmids. Where indicated, varying amounts of H-ras12V, K-ras12V, and N-ras12V were co-transfected. 24 h after transfection, cells were starved in serum-free medium overnight and washed with ice-cold phosphate-buffered saline. PI3K assays were performed following a recently described procedure with some modifications (36). Cells were lysed in buffer containing 50 mM HEPES, pH 7.5, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl 2 , 1 mM EGTA, 100 mM NaF, 10 mM sodium phosphate, 1 mM phenylmethylsulfonyl fluoride, 1 mM NaVO 4 , 10 g/ml aprotinin, and 10 g/ml leupeptin. Cleared lysates were normalized for protein content and incubated with 1 g of HA antibody at 4°C for 2 h followed by addition of protein A-agarose beads (Pierce) for 1 h. The immunoprecipitates were washed three times with lysis buffer; twice in 0.5 M LiCl, 100 mM Tris-Cl, pH 7.6; twice with 10 mM Tris-Cl, pH 7.6, 100 mM NaCl, 1 mM EDTA; and twice with 20 mM HEPES, pH 7.5, 50 mM NaCl, 5 mM EDTA, 0.03% Nonidet P-40, 30 mM Na 4 P 2 O 7 , 0.2 mM NaVO 4 , 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, and 10 g/ml leupeptin. Kinase reactions were carried out in 30 l of reaction buffer for 10 min at 30°C. Reactions contained 20 mM Tris-Cl, pH 7.6, 75 mM NaCl, 10 mM MgCl 2 , 2. pCDNA3-HA-AKT (100 ng) was transfected into NIH3T3 cells together with varying amounts of Ras plasmids using the LipofectAMINE method. One and a half days after transfection, cells were starved in serum-free medium overnight. Cells were then lysed in lysis buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 10% glycerol, 1% Nonidet P-40, 10 mM NaF, 1 mM sodium vanadate, 1 mM Na 4 O 7 P 2 , 2 M leupeptin, 2 M aprotinin, 1 mM phenylmethylsulfonyl fluoride). The lysates were cleared by centrifugation. Immunoprecipitation was performed by addition of monoclonal anti-HA antibody (1-2 g). The immune complexes were collected with protein A-agarose and washed three times with lysis buffer and once with kinase buffer without ATP and substrate (20 mM HEPES, pH 7.4, 10 mM MgCl 2 , 1 mM dithiothreitol). The kinase reaction was performed in 20 l of kinase reaction buffer containing 50 M ATP (with 10 Ci of [ 32 P]ATP) and 2 g of histone 2B. The reaction was incubated at 30°C for 20 min and followed by analysis by SDS-PAGE and autoradiography. Phosphorylation of histone 2B was quantified by PhosphorImager.
Reporter Analysis-The reporter plasmid, c-fos luciferase (37), was co-transfected with varying amounts of H-, K-, and N-ras DNA. Cells were starved in serum-free DMEM overnight 24 h after transfection. Luciferase activity was measured as described previously (38). Activity of co-transfected ␤-galactosidase was also determined and used to normalize the luciferase activity.

Different Transforming Activities of H-, K-, and N-Ras in
NIH3T3 and RIE-1 Cells-Functional differences between the three Ras isoforms, H-, K-, and N-Ras, have not been appreciated until recently (5,25,26,40). Gene knock-out experiments clearly demonstrate that the three Ras genes have different biological functions. In addition, it has been shown that H-Ras is more potent than K-Ras in PI3K activation (5,27). In contrast, K-Ras is more potent than H-Ras in Raf activation. We observed that expression of the GTP-bound V12 mutant of the three Ras proteins under control of the human CMV promoter produced dramatically different results with respect to cellular transformation in NIH3T3 cells. The focus formation assay in NIH3T3 cells indicated that H-RasV12 was a very potent inducer of transformation (Fig. 1, A and B), while K-RasV12 and N-RasV12 showed low transforming activity. The foci formed by H-RasV12-transformed cells were typical of Ras transformation: highly refractive with irregular edges (Fig. 2A). To exclude the possibility that the inability of K-Ras or N-Ras to transform NIH3T3 cells was due to different levels of expres-sion, immunoblots with antibodies recognizing all isoforms of Ras were performed. The three Ras constructs showed similar levels of protein expression (Fig. 1C).
It has been shown that Ras can transform rat intestinal epithelial cells (34). We tested the three active RasV12 mutants in transformation of RIE-1 cells. Similar results were obtained showing that H-Ras is more potent in transformation than K-Ras and N-Ras in RIE-1 cells (data not shown). These results confirmed that the different transformation activities of different Ras isoforms were not restricted in fibroblasts but were also observed in epithelial cells.
Transforming Phenotypes of H-, K-, and N-Ras Stable Cell Lines-The focus formation assay described above was based on the ability of transiently transfected cells to grow on a confluent monolayer. To extend the results of transient transfections, we examined the transformation phenotype of stable Ras cell lines. Stable cell lines of NIH3T3 cells transfected with one of the three Ras constructs were selected by resistance to G418. The cells in H-RasV12-transfected colonies were highly refractive, typical of transformation (Fig. 2, A and B). In contrast, the majority of cells in K-RasV12-transfected colonies were flatter and did not show features of transformation ( Fig.  2A). Cells in N-RasV12-transfected colonies showed morphology significantly different from that of vector-transfected cells although not as refractive as the H-Ras-transfected cells ( Fig.  2A). The N-Ras-transfected cells showed a very regular pattern and higher saturation density than the vector control. However, the majority of N-Ras-transfected cells did not form foci unlike the H-Ras-transfected cells.
Four G418-resistant clones from each Ras-transfected plate were selected randomly, purified, and characterized along with vector controls. Vector-transfected cells were flat and did not grow in soft agar (Fig. 2, B and C). All four clones of H-RasV12transfected cells showed transformed morphology: highly refractive, elongated cell shape, and decreased adhesion to culture plates (Fig. 2B). All four H-RasV12-transfected clones formed colonies in soft agar assays, although the efficiency varied among the four clones (Fig. 2C). In contrast, the K-Rasor N-Ras-transfected clones exhibited less transformed phenotype and did not grow in soft agar under similar conditions (Fig. 2, B and C). The expression levels of Ras were similar in these stable cell lines (Fig. 2D). These results further confirmed that H-Ras was more potent in transformation than K-RasV12 or N-RasV12 under our assay conditions.
Activation of ERK1 and PI3K by RasV12 Isoforms-One of the major downstream effectors of activated Ras is the Raf-MEK-ERK kinase pathway (3,4,8). To test whether the low transforming activity of K-RasV12 and N-RasV12 correlated with their inability to activate the ERK pathway, we determined ERK activation by Ras. Myc-tagged ERK1 was co-transfected with increasing amounts of a Ras plasmid, and the ERK kinase activity was determined. H-RasV12 stimulated ERK in a dose-dependent manner (Fig. 3A). Interestingly both K-RasV12 and N-RasV12 were very potent in activation of ERK (Fig. 3A). While maximal ERK activation levels achieved by the three Ras isoforms were similar at high concentrations of DNA, at low concentrations of DNA, K-rasV12 and N-rasV12 were more effective than H-rasV12 in ERK activation (Fig. 3A). These data suggested that although K-RasV12 and N-RasV12 effectively activate ERK1, they showed low transforming activity compared with H-RasV12. The low transforming ability of K-RasV12 and N-RasV12 was not due to a defect in ERK activation. These results also suggested that activation of ERK by oncogenic Ras was not sufficient to induce NIH3T3 cell transformation at least under these experimental conditions.
Ras has been shown to bind and activate PI3K (15, 41). The effect of Ras expression on PI3K activity was determined. Whereas all three Ras proteins activated PI3K, H-RasV12 was the most potent PI3K activator (Fig. 3, B and C). It is noteworthy that activation of PI3K required the transfection of Ras DNA at higher doses than was required for ERK activation. In addition, the maximal level of activation of PI3K by K-RasV12 or N-RasV12 was still lower than that achieved by H-RasV12, while the maximal activation of ERK1 by the three RasV12 was comparable.
The lipid products of PI3K are known to bind pleckstrin homology domains existing in numerous signaling molecules (18). Binding of the PI3K products to the pleckstrin homology domain of phospholipid-dependent kinase 1 increases the kinase activity of phospholipid-dependent kinase 1, which can subsequently phosphorylate and activate AKT, which also contains a pleckstrin homology domain (21). We determined the activation of AKT by Ras (Fig. 3D). HA-AKT was co-transfected with Ras in NIH3T3 cells. The transfected AKT was immunoprecipitated and assayed for kinase activity using histone 2B as a substrate. H-RasV12 was the most potent activator of AKT, while K-and N-RasV12 displayed a much lower AKT activation. Our data showed that activation of AKT by different Ras isoforms correlated with the relative activation of PI3K.
We extended our studies to include different cell lines. Both ERK and PI3K activation were examined in HEK293 cells. HEK293 cells can be transfected with higher efficiency than NIH3T3 cells; therefore, less RasV12 DNA was needed to stimulate ERK activity. Again N-Ras was most potent in ERK activation, while H-Ras was the weakest ERK activator in HEK293 cells (Fig. 4A). However, high concentrations of H-RasV12 (200 ng) activated ERK1 to a degree (ϳ500-fold) similar to that activated by K-or N-RasV12 (data not shown). As a control, EGF effectively activated ERK1 in HEK293 cells. PI3K activity was also determined in 293 cells in response to coexpression of the different Ras isoforms. Although all three Ras proteins were able to activate PI3K, H-RasV12 was the most potent PI3K activator (Fig. 4, B and C), similar to the results obtained with NIH3T3 cells (Fig. 3, B and C). These observations further confirmed that H-RasV12, K-RasV12, and N-RasV12 displayed different activities in stimulating ERK and PI3K. Our results also indicated a correlation between PI3K activation and transformation by Ras. Activation of Elk-1-dependent Transcription and Elk-1 Phosphorylation-One of the downstream effects of the Ras-MAP kinase pathway is the phosphorylation and activation of the Elk-1 transcription factor (10). Expression of active Ras results in phosphorylation of Elk-1, which can be easily detected by anti-phospho-Elk antibody. Our results indicated that Elk-1 phosphorylation was stimulated by co-expression of H-RasV12, K-RasV12, or N-RasV12 (Fig. 5A). The abilities of the three Ras isoforms to stimulate Elk-1 phosphorylation generally correlated with their abilities to stimulate ERK activation. K-RasV12 and N-RasV12 were much more effective in inducing Elk-1 phosphorylation than H-RasV12 (Fig. 5A). Activation of immunoprecipitated with anti-HA, and activity was determined. Lane 1 is a negative control without p85 and p110 DNA. Lane 2 represents the basal PI3K activity control without any ras DNA. Lanes 3, 4, and 5 were co-transfected with 10, 50, and 200 ng of H-rasV12, respectively. Corresponding amounts of rasV12 DNA were used for lanes 6, 7, and 8 for K-rasV12, or lanes 9, 10, and 11 for N-rasV12 DNA, respectively. Lane 12 is a sample treated with insulin. The amount of p110 in immunoprecipitates was detected by Western blot and is shown below the corresponding PI3K reactions. The relative expression of Ras in cell lysates is shown at the bottom. PIP denotes phosphatidylinositol 3-phosphate, the product of the PI3K reaction. C, quantification of PI3K activity. The radioactivity of phosphatidylinositol 3-phosphate (relative activity) was quantified by phosphorimage analysis. All lanes are the same as corresponding lanes in B. D, differential activation of AKT by Ras isoforms. HA-AKT (100 ng) was co-transfected with varying quantities of ras. The amount of Ras DNA used in each transfection is indicated. AKT was immunoprecipitated with anti-HA antibody. Kinase assays were determined using histone 2B as a substrate (top panel). AKT in the immunoprecipitates was detected by anti-HA Western blot (middle panel). Expression of co-transfected Ras in cell lysates was detected by anti- Ras Western blot (bottom panel). Results shown are a representative of three independent experiments. The amount of Ras protein in cell lysates was detected by anti-Ras antibodies that recognize all three isoforms of Ras (bottom panel). Lane 2 is a control for basal ERK activity in the absence of Ras. Lanes 3, 4, and 5 were co-transfected with 1, 5, and 20 ng of H-rasV12, respectively. Lanes 6, 7, and 8 were co-transfected with 1, 5, and 20 ng of K-rasV12, respectively. Lanes 9, 10, and 11 were co-transfected with 1, 5, and 20 ng of N-rasV12, respectively. Lane 12 is a positive control of cells stimulated with epidermal growth factor (50 ng/ml) for 5 min. B, H-RasV12 is more potent than K-RasV12 and N-RasV12 in activation of PI3K in HEK293 cells. The experiments were performed similarly to that in Fig. 2C. Lane 1 is a negative control with no PI3K in the transfection. Lane 2 is a control for basal PI3K activity. Lanes 3, 4, and 5 were co-transfected with 10, 50, and 200 ng of H-rasV12, respectively. Lanes 6, 7, and 8 were co-transfected with 10, 50, and 200 ng of K-rasV12, respectively. Lanes 9, 10, and 11 were co-transfected with 10, 50, and 200 ng of N-rasV12, respectively. PI3K activity is shown in the top panel. The immunoprecipitated catalytic subunit of PI3K (p110) was detected by anti-HA antibody (middle panel), while the levels of transfected Ras proteins were detected by Western blot (bottom panel). C, quantification of PI3K activity. Radioactivity of phosphatidylinositol 3-phosphate (PIP) (relative activity) was determined by phosphorimage analysis. Lanes are the same as corresponding lanes in B.
ternary complex factors (Elk-1 is a member of ternary complex factor) plays a critical role in the induction of the c-fos protooncogene (10). We therefore tested the effect of Ras on a c-Fos reporter, which expresses luciferase under the control of the c-fos promoter (37). Both K-RasV12 and N-RasV12 stimulated the c-Fos reporter more efficiently than H-RasV12 (Fig. 5B), although all three Ras isoforms stimulated the c-Fos reporter to a similar extent at high DNA doses (data not shown). Induction of c-Fos promoter qualitatively correlated with the phosphorylation of Elk-1.
AKT Cooperates with Ras in Transformation-To further characterize the correlation between the low transformation activity of K-and N-Ras and their reduced ability to activate PI3K and AKT, we examined the possibility that enhanced AKT activity might cooperate with Ras in cellular transformation. AKT is an important downstream target of PI3K and has been implicated in cellular transformation (16,23,24). The constitutively active AKT, when co-transfected with ras plasmids, enhanced the transforming activity of Ras (Fig. 6). The low quantity of AKT, K-RasV12, or N-RasV12 alone did not induce focus formation in NIH3T3 cells under our assay conditions. However, the combination of constitutively active AKT FIG. 5. K-RasV12 and N-RasV12 are functional in transcription induction. A, K-RasV12 and N-RasV12 are more potent than H-RasV12 in stimulation of Elk-1 phosphorylation. NIH3T3 cells were transfected with 300 ng of Elk-1 plasmid except lane 1, which was not transfected with Elk-1. Various Ras plasmids were co-transfected as indicated. The transfected cells were starved in serum-free medium for 12 h before lysis and Western blot with anti-Elk-1 antibody (bottom panel) or anti-phospho-Elk-1 (pElk-1) antibody, which recognizes the phosphorylated and active Elk-1 (top panel). Lane 2, control with no ras. Lanes 3,4, and 5 were co-transfected with 10, 50, and 200 ng of H-rasV12, respectively. Lanes 6, 7, and 8 were co-transfected with 10, 50, and 200 ng of K-rasV12, respectively. Lanes 9, 10, and 11 were co-transfected with 10, 50, and 200 ng of N-rasV12, respectively. Results shown are a representative of three similar experiments. B, activation of c-Fos reporter by Ras. NIH3T3 cells were transfected with a c-fos reporter, ␤-galactosidase control, and varying amounts of ras. 24 h after transfection, cells were starved in serum-free medium for 12 h. Cells were lysed, and luciferase activity was determined and normalized against the co-transfected ␤-galactosidase activity. Lane 1 is a control without Ras co-transfection. Lanes 2, 3, 4, and 5 were cotransfected with 1, 3, 10, and 50 ng of H-rasV12, respectively. Lanes 6, 7, 8, and 9 were co-transfected with 1, 3, 10, and 50 ng of K-rasV12, respectively. Lanes 10, 11, 12, and 13 were co-transfected with 1, 3, 10, and 50 ng of N-rasV12, respectively. C, expression levels of co-transfected Ras correspond to the bar graph of luciferase assay in B. with either K-RasV12 or N-RasV12 resulted in efficient transformation (Fig. 6). AKT also enhanced H-RasV12 in transformation (Fig. 6). Co-transfection of AKT with the constitutively active MEK1 also enhanced the transformation potential albeit the colonies were smaller. These data demonstrated that AKT had a critical role in cellular transformation and that the low transforming activities of K-or N-Ras were likely due to a low activation of PI3K. DISCUSSION Although it is generally accepted that H-, K-, and N-Ras have similar biochemical and biological functions, gene inactivation studies and in vitro characterization indicate that the three Ras proteins may have different functions or activities. In this study, we showed that H-RasV12 had the highest transformation potential in NIH3T3 cells. We were surprised to find that the three RasV12 mutants showed dramatic differences in their abilities to transform cells in culture given their high degree of sequence identity. These results were confirmed in RIE-1 cells. The different transformation potentials of the three Ras proteins might be due to a difference in interaction/ activation of downstream targets. We tested the interaction of several known downstream Ras targets. Our data showed that the V12 mutants of all three Ras proteins interacted similarly with C-Raf, PI3K, and Ral guanine nucleotide dissociation stimulator in a GTP-dependent manner in the yeast two-hybrid system (data not shown).
It has been shown that H-Ras activates PI3K better than K-Ras, while K-Ras activates Raf better than H-Ras (5,27). Our data are completely consistent with the previously reported observations. We further extended previous observations in the following aspects. We investigated the transformation activities of the three Ras isoforms. We demonstrated that activation of MAP kinase cascade was not the limiting factor in Ras transformation. We correlated that transformation potential of Ras to the activation of PI3K and AKT. We also showed that activation of AKT could cooperate with Ras, especially Kand N-Ras, in transformation. These data indicated a critical role of PI3K/AKT in Ras transformation that was consistent with published results (23).
Ras is known to directly interact with downstream targets to transmit its signal. Paradoxically the N-terminal 86 residues including the effector domain (residues 30 -42) are absolutely conserved among the three Ras proteins. This effector domain of Ras is believed to mediate the interaction between Ras and its downstream targets. Therefore, it is reasonable to assume that the differential activities of Ras isoforms are due to the C-terminal variable domain. Previous studies indicate that the C-terminal domain of Ras plays a role in transformation (42,43). Interestingly the C-terminal domains of Drosophila and Caenorhabditis elegans Ras are more similar to K-Ras than H-Ras. It should be noted that all the experiments described in this report used mutationally activated Ras. The functions of endogenous Ras genes will certainly depend on the relative expression of Ras proteins in different tissues.
H-RasV12 showed a weaker ability to activate co-transfected ERK1 than K-RasV12 or N-RasV12 in both NIH3T3 and HEK293 cells at low DNA concentrations. However, H-Ras activated ERK to the same extent as K-RasV12 or N-RasV12 at higher concentrations. The same did not hold for PI3K activation. Even at high concentration, neither K-RasV12 nor N-RasV12 activated PI3K to the same extent as H-RasV12. Furthermore activation of PI3K required a much higher concentration of Ras DNA than was required for ERK activation. Another noticeable difference between ERK and PI3K activation by Ras was the magnitude of activation. The maximal PI3K activation by Ras was ϳ10-fold above basal, yet ERK was activated as much as 500-fold. This difference could be attributed to a higher basal activity of PI3K or the amplification effect of the Raf-MEK-ERK kinase cascade. Alternatively Ras might be a more potent activator of ERK than PI3K. Nevertheless the above observations were consistent with the model that activation of PI3K, but not ERK, was the limiting factor in Ras transformation.
Constitutive activation of MEK-ERK has been shown to induce transformation in NIH3T3 cells (11,12), and therefore, activation of ERK is generally thought to mediate the transforming potential of Ras. It is worthy to note that focus formation by constitutively active MEK was achieved with stable cell lines expressing the constitutively active MEK (12). The constitutively active MEK failed to induce foci under the conventional focus formation assay conditions used for Ras and other oncogenes. Point mutants of Ras that fail to activate ERK are also defective in inducing transformation (12). These observations suggest that ERK plays an important role in NIH3T3 transformation. Results in this study showed that both K-RasV12 and N-RasV12 were more effective than H-RasV12 in ERK activation, yet they were less potent in transformation. Therefore, activation of ERK was likely only one of the factors required for Ras transformation. Activation of ERK alone might not be sufficient to induce cellular transformation. Rather activation of other downstream effectors of Ras could be required for transformation.
The qualitative difference of the three Ras proteins in transformation could be due to a quantitative difference in their ability to activate downstream targets such as PI3K. Alternatively it was possible that both K-Ras and N-Ras were insufficient to activate a downstream component, which was effectively activated by H-Ras, essential for cellular transformation. We proposed that the low transforming activities of K-and N-Ras were at least in part due to their reduced abilities to activate PI3K based on the following reasons. Both K-RasV12 and N-RasV12 showed a significantly lower ability to stimulate PI3K. Interestingly both K-and N-RasV12 failed to fully activate AKT to the same extent as H-RasV12. Expression of active AKT could enhance the transformation potential of K-and N-RasV12. Consistent with works from other laboratories (16,17,23,44), our data demonstrate an important role of PI3K-AKT in Ras-mediated transformation.
It is worth noting that mutation of K-and N-Ras is more frequent than mutation of H-Ras in human cancers. How can our results be reconciled with existing data for an important role of K-and N-Ras in human cancer? Another puzzling question is how mutation in B-Raf, which is frequently found in human cancers and only activates the ERK pathway (14), contributes to human cancer. The NIH3T3 and RIE-1 transformation assay is based on expression of a single transforming gene. However, human cancers arise from multiple genetic alterations. For example, mutation in the PTEN tumor suppressor gene occurs in ϳ50% of human cancers (45)(46)(47). PTEN mutation results in constitutive activation of the AKT pathway, which provides a major signal to stimulate cell growth and inhibit apoptosis (48 -50). Furthermore genetic alterations resulting in suppression of apoptosis, such as mutations in Bcl-2 family members (some are targets of AKT) (51,52), may also substitute for the requirement of PI3K activation in cancer growth. In PTEN mutant cells, no additional genetic change is needed to activate AKT; therefore, activation of the ERK pathway by mutations, such as K-Ras, N-Ras, and B-Raf, may be sufficient to cause tumorigenesis. This notion is consistent with the high potential of K-and N-Ras to activate the ERK path-way and the high frequency of K-and N-Ras mutations found in human cancers.