INTRODUCTION

Ras proteins (N-Ras, Ha-Ras, Ki-Ras) are three closely related members of the Ras family which act as a molecular switch for signal transduction pathways to control cell growth and differentiation (1). Like other guanine nucleotide-binding proteins, Ras cycles between an active GTP-bound form and an inactive GDP-bound form. The GDP-bound form is converted to the GTP-bound form through a GDP/GTP exchange reaction that is facilitated by guanine nucleotide-releasing factors (2). On the other hand the GTP-bound form is converted to the GDP-bound form by the intrinsic GTPase activity, which is accelerated by GTPase-activating proteins (GAPs)¹ (3).

R-Ras is a member of the Ras family proteins and is highly homologous to Ras (4). Despite a high sequence similarity between Ras and R-Ras, R-Ras does not transform Rat1 fibroblastic cells (5). However, recent results have demonstrated that the activated form of R-Ras transforms NIH 3T3 cells, and the transformant forms tumors in athymic nude mice (6, 7). Since R-Ras has an effector binding domain the amino acid sequence of which is very similar to that of Ras, R-Ras binds to and activates the c-Raf-1/mitogen-activated protein kinase cascade (6, 8). Besides its roles in the stimulation of cell proliferation, R-Ras may play other roles in different cell biological processes. It is reported that R-Ras binds to Bcl-2, which is a key molecule controlling the process of apoptosis; however, the binding to Bcl-2 is not GTP-dependent (9). It was also reported that R-Ras promotes apoptosis induced by growth factor deprivation by a mechanism that is suppressed by overexpression of Bcl-2 (10). Recently it has been described that R-Ras enhances cell adhesion to extracellular matrix substrates through the activation of integrins (11). However, the biochemical mechanisms by which R-Ras activity is regulated are still to be clarified.

Three mammalian GAPs for Ras have been identified so far. p120GAP, which was first described, is a prototype of this class of proteins (12). Besides a catalytic domain that stimulates Ras GTPase, p120GAP has two SH2 (Src homology 2) domains, one SH3 domain, one PH (plekstrin homology) domain, and one phospholipid binding domain (13). The second is neurofibromin (NF1), a product of the neurofibromatosis type I gene (14). Neurofibromin has a region that shows a sequence similarity to the catalytic domain of p120GAP and Ira proteins of Saccharomyces cerevisiae, and the domain was termed as a GAP-related domain. Indeed this region was shown to possess GAP activity for Ras and to suppress ira2 mutation (15). We have isolated the third Ras GAP (Gap1^m) which is a mammalian homolog of the Drosophila Gap1 gene (16). In addition to the GAP-related domain, Gap1^m has two putative phospholipid binding domains and a region similar to the domain unique to Btk tyrosine kinase (16).

Recently R-Ras GAP, the entire structure of which is closely related to Gap1^m, was isolated (17). The identity of the amino acid sequences of R-Ras GAP and Gap1^m is 60%. Despite the high sequence similarity to Gap1^m, R-Ras GAP stimulates the GTPase of R-Ras better than that of Ras (17). A previous study described that p120GAP stimulates the GTPase of R-Ras as efficiently as Ha-Ras (18). We showed that Gap1^m stimulates the GTPase of the wild type of Ras but not that of the activated form of Ras, Rap1, or GTP-binding proteins of other families (19). Hence we investigated in this study the enzymatic properties of Gap1^m and compared them with those of p120GAP using Ras and R-Ras proteins as substrates. The results indicate that both Gap1^m and p120GAP promote the GTPase of R-Ras and that Gap1^m stimulates the GTPase of Ras better than that of R-Ras, in contrast to p120GAP the activity of which is higher with R-Ras as the substrate.

EXPERIMENTAL PROCEDURES

The pGEX expression vectors for the Ras family proteins (17) were generously provided by Dr. K. Kaibuchi. The N-Ras expression system was a kind gift of Dr. A. Wittinghofer. Glutathione S-transferase (GST)-Ras fusion proteins were induced in Escherichia coli and purified using glutathione-Sepharose 4B (Pharmacia Biotech Inc.) as described by Smith and Johnson (20). Fusion proteins were dialyzed overnight against 100 volumes of buffer A (20 mM Tris-HCl (pH 8.0), 150 mM NaCl, 10% (v/v) glycerol, and 2 mM MgCl₂) and then concentrated to approximately 20 mg/ml using Centricon-10 (Amicon Inc.). Each fusion protein (500 µg) was digested with 1 µg of thrombin (Sigma) in buffer A containing 1 mM CaCl₂. The digested samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The gel was stained by a SYPRO^TM Orange stain kit (Bio-Rad), and the protein concentration was determined by quantitating the intensity of each band by Fluor Imager (Molecular Dynamics) using bovine serum albumin as a standard. Concentrations of active proteins were determined by measuring their [³H]GDP binding activities (0.5-0.6 pmol of [³H]GDP bound per pmol of protein). These concentrations were employed in all experiments.

To produce the GST-Gap1^m fusion protein, the NcoI-EcoRI fragment of the cDNA of Gap1^m (nucleotides 54 to 3 end, accession number D30734) which covers the entire coding sequence was ligated to the BamHI site of pGEX2T by using synthetic oligonucleotides (5-GTGGATC-3 and 5-GATCCAC-3). The SacII-EcoRI fragment of rat p120GAP cDNA (21) (a generous gift of Dr. Y. Kaziro, nucleotides 214 to 3 end, accession number L13151) was cloned into the SmaI site of pGEX2T after blunting of the fragment which directed the expression of GST-p120GAP (amino acids 39 to carboxyl terminus) fusion protein. Each GST-fusion protein was cleaved by procedures similar to those described above and then applied to a column of heparin-Sepharose CL-6B (Pharmacia) which had been equilibrated with buffer B (10 mM Tris-HCl (pH 8.0), 0.5 mM EDTA, 5% glycerol). The adsorbed materials were eluted by a 0-1 M NaCl gradient in buffer B. Fractions containing Gap1^m (0.7-0.8 M NaCl) and p120GAP (0.3-0.4 M NaCl) were pooled and stored at 80 °C until use. The purity of Gap1^m and p120GAP was more than 85% as revealed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Fig. 1).

GAP activity was measured as described previously (19) except that the reaction was carried out at 30 °C. Typically, 2 pmol of each Ras·[-³²P]GTP (30 Ci/mmol, ICN) was incubated at 30 °C with the indicated amounts of Gap1^m or p120GAP for the times specified in the figure in 20 µl of a solution containing 50 mM Tris-HCl (pH 8.0), 5 mM MgCl₂, 0.25 mM GTP, 5 mg/ml bovine serum albumin, and 5% glycerol. The reactions were quenched by keeping the sample on ice. After the addition of 0.5 ml of ice-cold washing buffer (20 mM Tris-HCl (pH 8.0), 100 mM NaCl, and 5 mM MgCl₂), the sample was passed through a nitrocellulose filter (0.45 µm, Schleicher & Schuell). The filter was washed three times with 1 ml of ice-cold washing buffer, and the radioactivity trapped on the filter was determined using a liquid scintillation counter. For determination of the enzymatic properties (K_m and k_cat) of Gap1^m and p120GAP, a constant amount of either Gap1^m (1.2 nM for Ha-Ras and 1.5 nM for R-Ras) or p120GAP (20.0 nM for Ha-Ras and 3.0 nM for R-Ras) was incubated with various concentrations of Ha-Ras·[-³²P]GTP or R-Ras·[-³²P]GTP at 30 °C for 10 min. The amount of GTP hydrolyzed during the reaction was determined as a decrease in the radioactivity trapped on the filter. K_m and k_cat values were calculated using the Michaelis-Menten equation. The data were obtained from at least three independent experiments.

Competitive Inhibition of GAP Activity by Ha-Ras or R-Ras Bound to GTPS or GDP

Ha-Ras·[-³²P]GTP (2 nM) or R-Ras·[-³²P]GTP (2 nM) was incubated with a constant amount of either Gap1^m (0.3 nM for Ha-Ras and 1.2 nM for R-Ras) or p120GAP (20 nM for Ha-Ras and 2 nM for R-Ras) in the presence of various concentrations of competitors, Ha-Ras·GTPS, Ha-Ras·GDP, R-Ras·GTPS, or R-Ras·GDP at 30 °C for 10 min. The GAP activity observed with a control sample without added competitor was taken as 100%, and the GAP activity of each sample was expressed as a percentage of the control sample.

RESULTS

We identified and isolated Gap1^m as the third GAP for Ras (16, 19). Since the entire structure of Gap1^m is not similar to that of p120GAP or neurofibromin, we deduced that Gap1^m may have distinct enzymatic properties. To investigate the substrate specificities of Gap1^m and p120GAP, various concentrations of either Gap1^m or p120GAP were incubated with each of the four Ras family proteins, and the enzyme concentration at which 50% of the maximal stimulation was achieved (EC₅₀) was determined (Fig. 2). Gap1^m showed a similar effect on the GTPase activities of three Ras proteins (EC₅₀ values for N-Ras, Ha-Ras, and Ki-Ras were 0.56 ± 0.09, 0.48 ± 0.02, and 0.46 ± 0.04 nM, respectively). However, Gap1^m was less effective in stimulation of the GTPase of R-Ras (EC₅₀ 1.13 ± 0.12 nM). In contrast, p120GAP exhibited higher activity in stimulation of the GTPase of R-Ras (EC₅₀ 3.86 ± 0.38 nM) than those of three Ras proteins (EC₅₀ values for Ki-Ras, N-Ras, and Ha-Ras were 12.5 ± 0.9, 23.1 ± 1.9, and 22.7 ± 1.5 nM, respectively). A similar tendency of substrate specificities of Gap1^m and p120GAP was also observed when time course experiments were carried out using the four Ras family proteins as substrates (Fig. 3). These data demonstrated that Gap1^m stimulated GTPase activities of both Ras and R-Ras where Ras was activated higher than R-Ras. In contrast, p120GAP showed much higher activity toward R-Ras than Ras proteins under the same experimental conditions.

Inhibition of GAP Activity by Ras Family Proteins Bound to GTPS or GDP

We next investigated whether there may be a relationship between the substrate preferences of Gap1^m and p120GAP and their affinities to the substrates (Figs. 4 and 5). We measured the ability of Ha-Ras or R-Ras bound to either GTPS (a nonhydrolyzable analog of GTP) or GDP to inhibit competitively the GAP activity of Gap1^m (Fig. 4). The inhibition constant (K_i) is obtained as a concentration of the inhibitor at which 50% of GAP activity is inhibited (19). Ha-Ras·GTPS was much more inhibitory on the activity of Gap1^m (K_i = 0.83 ± 0.12 µM) than Ha-Ras·GDP (K_i = 3.27 ± 0.67 µM). R-Ras·GTPS was also more effective in the inhibition of Gap1^m activity (K_i = 2.00 ± 0.21 µM) than R-Ras·GDP (K_i = 4.31 ± 1.02 µM). In this analysis, the K_i of R-Ras·GTPS for Gap1^m activity was 2.5 times higher than that of Ha-Ras·GTPS. This higher affinity of Gap1^m to Ha-Ras·GTPS was in good agreement with the substrate preference of Gap1^m; the EC₅₀ of Gap1^m for Ha-Ras was 2.5 times lower than that for R-Ras. The K_i value of Ha-Ras·GDP was also lower than that of R-Ras·GDP.

A similar experiment was carried out using p120GAP (Fig. 5). The concentrations of the inhibitors were not enough for full inhibition of the activity. R-Ras·GTPS was most effective in the inhibition of p120GAP activity; however, the K_i of R-Ras·GTPS was still over 10 µM. The effect of Ha-Ras·GTPS was weaker than R-Ras·GTPS, indicating that R-Ras·GTPS bound to p120GAP more tightly than Ha-Ras·GTPS did. Neither Ha-Ras·GDP nor R-Ras·GDP was inhibitory on p120GAP activity at the concentrations tested. Although we could not determine the K_i of Ha-Ras·GTPS, Vogel et al. (22) reported that the K_i of Ha-Ras·GTP is 110 µM. Thus, the relative affinities of p120GAP to R-Ras and Ha-Ras again correlated well with the substrate preference of p120GAP. These results suggested that the substrate specificities of both Gap1^m and p120GAP may be determined by their affinities to the substrates.

To investigate further the enzymatic properties of Gap1^m and p120GAP, we determined the kinetics of the GTPase of Ha-Ras and R-Ras stimulated by GAPs. A fixed amount of Gap1^m or p120GAP was incubated with various concentrations of either Ha-Ras·[-³²P]GTP or R-Ras·[-³²P]GTP, and the reaction rate of GTPase was determined (Fig. 6). Gap1^m was much more active than p120GAP in the stimulation of the GTPase activity of Ha-Ras. In contrast, activation of the R-Ras GTPase by Gap1^m was similar to that by p120GAP at low concentrations of R-Ras, and p120GAP was more active at high concentrations of R-Ras. Since the activity of p120GAP was not saturable under the experimental conditions, the activity of p120GAP for Ha-Ras might also be higher than Gap1^m at very high concentrations.

The rate of GAP-stimulated GTPase was plotted as a function of substrate concentration according to the Michaelis-Menten equation which gave k_cat (V_max/E₀) and K_m for the reactions (Fig. 7). The K_m values of Gap1^m for Ha-Ras and R-Ras were 1.53 ± 0.27 and 3.38 ± 0.53 µM, respectively (Fig. 7A). This result indicated that Gap1^m had a higher binding affinity to Ha-Ras than R-Ras which agreed well with the results presented in Fig. 4. At saturating concentrations of the substrates, Gap1^m showed somewhat higher activity for R-Ras than for Ha-Ras (k_cat values for Ha-Ras and R-Ras were 3.96 ± 0.31 and 4.96 ± 0.74 s¹, respectively).

p120GAP shows 7.5 times higher activity for R-Ras than Ha-Ras at any of the substrate concentrations examined (Fig. 6). By fitting the data to a double reciprocal plot, the K_m and the k_cat values of p120GAP for Ha-Ras were determined to be 145 ± 11 µM and 23.0 ± 3.4 s¹, respectively (Fig. 7B). Accurate values of the K_m and k_cat of p120GAP for R-Ras could not be determined because the specific activity was completely proportional to the substrate concentration. These results are summarized in Table I.

Table I. Enzymatic properties of Gap1m and p120GAP GAPs and Substrates kcat Km EC50 Ki s1 µM nM µM Gap1m   Ha-Ras 3.96  ± 0.31 1.53  ± 0.27 0.48  ± 0.02 0.83  ± 0.12a   R-Ras 4.96  ± 0.74 3.38  ± 0.53 1.13  ± 0.12 2.00  ± 0.21a p120GAP   Ha-Ras 23.0  ± 3.4 145  ± 11 23.1  ± 1.9 110b   R-Ras NDc ND 3.86  ± 0.38 >10a a Determined by using Ha-Ras · GTPS or R-Ras · GTPS as a competitor. b Taken from Ref. 22. c Not determined.

DISCUSSION

In this study we examined the enzymatic properties and substrate specificity of Gap1^m and compared them with those of p120GAP. Previously we demonstrated that Gap1^m does not stimulate the GTPase of Rap1, Rho, and Ram25K (19). Therefore, in this study we used three Ras proteins and R-Ras as the substrates.

Both Gap1^m and p120GAP stimulated the GTPase of Ras and R-Ras. EC₅₀ values of Gap1^m for Ras proteins were two times lower than that for R-Ras (Fig. 1 and Table I), thus Gap1^m stimulates Ras GTPase better than that of R-Ras (p < 0.01). In contrast, the EC₅₀ of p120GAP for R-Ras was much lower than those for Ras proteins. Thus p120GAP activates the GTPase of R-Ras better than that of Ras. A previous report described that the catalytic domain of p120GAP (GAPette) stimulates the GTPase of both Ras and R-Ras with almost equal EC₅₀ values (18). However, it was described that domains of p120GAP outside of the catalytic domain are necessary for the full activity (23, 24). Since the p120GAP used in this study lacks only 38 amino acid residues at the amino terminus, it seems to be an intrinsic property of p120GAP that it is more active for R-Ras than for Ras.

We also measured the inhibition constant (K_i) of Ras and R-Ras bound to GTPS. The K_i of Ras·GTPS for Gap1^m was 2.5 times lower than that of R-Ras·GTPS, which agrees well with the ratio of K_m values of Gap1^m for Ras and R-Ras and also with the substrate specificity in that Gap1^m stimulates GTPase of Ras better than that of R-Ras. Similarly, p120GAP promotes the GTPase of R-Ras better than that of Ras, and p120GAP binds more tightly to R-Ras than Ras. Hence, there may be a good correlation between the affinity of the GAPs to the substrates and their substrate preference.

The K_m values of p120GAP and neurofibromin for Ras have been reported to be 9.7 and 0.3 µM, respectively (24, 25). Under our experimental conditions the K_m of p120GAP for Ras was 145 µM. This difference in the K_m values of these two p120GAP may be the result of different experimental conditions, i.e. temperatures, substrates, and the sources of p120GAP. The k_cat values of the two preparations are similar (19 s¹ (24) and 23.4 s¹ (this study)). A similar high K_m value of full-length p120GAP for Ras has also been reported (23). The K_m of Gap1^m for Ras was 1.53 µM (Table I). Vogel et al. (22) described that Ras in complex with nonradioactive GTP competes in the p120GAP-catalyzed reaction with the K_i value of 110 µM. The K_i value of Ras·GTPS for recombinant Gap1^m is 0.83 µM (Table I), which is similar to the value obtained using authentic Gap1^m (19). The difference between these K_i values of the two GAPs agrees well with that between the K_m values of p120GAP and Gap1^m. Thus the affinity of p120GAP to the substrates is much weaker than that of Gap1^m or neurofibromin by almost 2 orders of magnitude.

Recently another Gap1^m family termed R-Ras GAP, whose entire domain structure is very similar to Gap1^m, was described (17). The overall identity of the amino acid sequence is 60%. R-Ras GAP also stimulates the GTPase of Ras, but the stimulation is lower than that observed with R-Ras. In contrast, a GAP-related domain of neurofibromin stimulates GTPase of Ras stronger than R-Ras GTPase (17). Thus, all the known Ras GAPs and R-Ras GAP activate the GTPase of both Ras and R-Ras, with some different substrate preferences. GAP1^IP4BP (26) and GAPIII (27) seem to be human and mouse homologs of R-Ras GAP, respectively, since both of them show higher sequence similarity to R-Ras GAP than to Gap1^m.

Residues in the switch I region of Ras which are critical to the stimulation by p120GAP have been extensively characterized (12). Within this region, the amino acid sequence of Ras from residue 32 to 40 is completely preserved in R-Ras and Rap1. However, although Rap1 binds to p120GAP, the GTPase of Rap1 is not stimulated by p120GAP (28). Whereas both position 31 of Ras (glutamic acid) and the equivalent position of R-Ras (aspartic acid) are acidic residues, Rap1 has lysine at the corresponding position. Substitution of the 31st glutamic acid of Ras by lysine renders the mutated Ras a phenotype like Rap1 with concomitant loss of susceptibiltity to p120GAP (29). These findings suggest that residue 31 of Ras or residues at the corresponding positions of R-Ras and Rap1 may be critical to their susceptibility to GAPs.

What may be the biological implication deduced from the enzymatic properties of Ras GAPs? The weak binding affinity of p120GAP to the substrates suggests that factors that help the association of p120GAP with the substrate may increase the activity of p120GAP toward the substrates. Yao and Cooper (30) reported that p120GAP with a membrane targeting signal showed higher specific activity in intact cells than p120GAP without the targeting signal. Since p120GAP binds to autophosphorylated receptors (31), such binding may help the local contact of p120GAP to its substrates located on the membranous structure. As pointed out by Bernards (32), this implies that p120GAP may act as a quencher of growth factor-activated signals rather than a regulator of basal activity of Ras and R-Ras. In contrast, the affinity of Gap1^m or neurofibromin with their substrates is rather high such that they may be able to stimulate the GTPase of the substrates efficiently by themselves. Some portion of Gap1^m and neurofibromin resides in the membranous fraction (19, 33). Such membrane localization of neurofibromin and Gap1^m may facilitate the contact of these GAPs with their substrates. Thus Gap1^m may regulate the basal activity of Ras and R-Ras as does neurofibromin (32).