A Novel GTPase-activating Protein for R-Ras*

R-Ras, belonging to the Ras small GTP-binding pro- tein superfamily, has been implicated in regulation of various cell functions such as gene expression, cell pro- liferation, and apoptotic cell death. In the present study, we purified an R-Ras-interacting protein with molecu- lar mass of about 98 kDa (p98) from bovine brain cytosol by glutathione S -transferase (GST)-R-Ras affinity col- umn chromatography. This protein bound to GTP (cid:103) S (guanosine 5 (cid:42) -(3- O -thio)triphosphate, a nonhydrolyz-able GTP analog) (cid:122) R-Ras but not to GDP (cid:122) R-Ras, GTP (cid:103) S (cid:122) R-Ras with a mutation in the effector domain (R-Ras A64 ), GTP (cid:103) S (cid:122) Ha-Ras, or GTP (cid:103) S (cid:122) RalA. We obtained a cDNA encoding p98 on the basis of its partial amino acid se- quences. The predicted protein consists of 834 amino acids whose calculated mass, 95,384 Da, is close to the apparent molecular mass of p98. The amino acid sequence shows a high degree of sequence similarity to the entire sequence of Gap1 m , one of the GTPase-acti-vating proteins (GAP) for Ha-Ras. A recombinant pro- tein consisting of the GAP-related domain of p98 fused to maltose-binding protein stimulated GTPase -ATAAGGTACCATGAGCTCTGGTGCTGC-3 (cid:57) and 5 (cid:57) -TATAGGTAC-CAAGGCACACAGTGGCAG-3 (cid:57) , and cloned into the Kpn I site of pGEX2T- Kpn I, which was made by introducing the Kpn I site to the Bam HI site of pGEX-2T. The R-Ras A64 mutant was produced by PCR of pGEX-R-Ras with the primers 5 (cid:57) -TCCTACACGAAGATCTGC-3 (cid:57) and 5 (cid:57) -TCGTGTAGGAGGCCTCAATAGTG-3 (cid:57) , and cloned into pGEX-2T- Kpn I. Human RalA cDNA was amplified by PCR from human brain cerebral cortex Quick Clone cDNA (Clontech) with the primers 5 (cid:57) -* Bovine into small pieces with and in ml of homogenization m M m M m M m M g/ml peptin, Teflon-glass homogenizer four homogenate centrifuged GTPase Assay— GTPase activity of decreased radioactivity incubated GST-NF1-GRD rapid filtration using nitrocellulose filters. washed with and bound radioactivity was determined liquid (cid:57) which were designed from rat R-Ras GAP partial cDNA se- quences. The PCR products were subjected to 6% polyacrylamide gel electrophoresis (21). Other Procedures— SDS-PAGE was performed as described previously (22). Protein concentrations were determined with bovine serum albumin as a reference protein as described (23). The BLAST program was used for protein homology search (24).

R-Ras was originally identified as the gene product of a Ras homologue (10). R-Ras has been reported to physically associate with Bcl-2, which is known to be a blocker of apoptotic cell death (11,12). Recently, the activated R-Ras V38 that presumably remains mostly in the GTP-bound form due to impaired GTPase activity has been reported to enhance the apoptotic cell death in cytokine-deprived 32D.3 cells and serum-deprived NIH/3T3 cells (13). Bcl-2 abrogates most of the effects of R-Ras V38 , indicating that R-Ras promotes apoptosis caused by growth factor deprivation via a Bcl-2-suppressible mechanism. In NIH/3T3 cells, R-Ras V38 confers the ability to proliferate under low serum conditions, to form colonies in soft agar, and to form tumors in nude mice, although its ability is weaker than that of the activated Ha-Ras (14). R-Ras as well as Ha-Ras have been shown to interact with c-Raf-1, p120 GAP, and NF1, to induce MAP kinase activation, and to stimulate Ras response elements in certain cells (15,16). On the other hand, R-Ras does not induce maturation of Xenopus oocytes or differentiation of PC12 cells. Taken together, although Ras and R-Ras share some biochemical and cellular functions, these proteins seem to play different biological roles.
To understand the specific functions of R-Ras, we attempted to identify proteins that specifically interact with R-Ras in the present study, and have purified an R-Ras-interacting protein with molecular mass of about 98 kDa by GST-R-Ras affinity column chromatography, cloned its cDNA, determined its primary structure, and identified it as a novel R-Ras GAP.

EXPERIMENTAL PROCEDURES
Materials and Chemicals-Polyvinylidene difluoride membranes (Problott, 0.45 m pore size) were purchased from Applied Biosystems. Achromobacter protease I was from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). All materials used in the nucleic acid study were purchased from Takara Shuzo Co., Ltd. (Kyoto, Japan). Prokaryotic expression plasmid pGEX-2T and pMal-c2 were obtained from Pharmacia Biotech Inc. and New England Biolabs, respectively. In vitro transcription plasmid pGEM-3zf(ϩ) was obtained from Promega Corp. Other materials and chemicals were obtained from commercial sources.
Plasmids and Protein Preparation-pGEX-R-Ras was constructed as follows. Human R-Ras cDNA was amplified by PCR from the pTRGA (a kind gift from Drs. David G. Lowe and Alan Hall) with the primers 5Ј-ATAAGGTACCATGAGCTCTGGTGCTGC-3Ј and 5Ј-TATAGGTAC-CAAGGCACACAGTGGCAG-3Ј, and cloned into the KpnI site of pGEX2T-KpnI, which was made by introducing the KpnI site to the BamHI site of pGEX-2T. The R-Ras A64 mutant was produced by PCR of pGEX-R-Ras with the primers 5Ј-TCCTACACGAAGATCTGC-3Ј and 5Ј-TCGTGTAGGAGGCCTCAATAGTG-3Ј, and cloned into pGEX-2T-KpnI. Human RalA cDNA was amplified by PCR from human brain cerebral cortex Quick Clone cDNA (Clontech) with the primers 5Ј-* This investigation was supported by grants-in-aid for scientific research and for cancer research from the Ministry of Education, Science, and Culture, Japan (1994), and by a grant for research on metabolic disease from the Yamanouchi Foundation (1994). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EMBL Data Bank with accession number(s) U30857.
Cytosol Preparation-Bovine brain gray matter (100 g) was cut into small pieces with scissors and suspended in 300 ml of homogenization buffer (25 mM Tris/HCl at pH 7.5, 1 mM DTT, 5 mM EGTA, 10 mM MgCl 2 , 10 M (p-amidinophenyl)methanesulfonyl fluoride, 1 g/ml leupeptin, and 10% sucrose). The suspension was homogenized with a Potter-Elvehjem Teflon-glass homogenizer and filtered through four layers of gauze. The homogenate was centrifuged at 20,000 ϫ g for 30 min at 4°C and then at 100,000 ϫ g for 60 min at 4°C. The supernatant was stored at Ϫ80°C as the crude cytosolic fraction.
GST-R-Ras Affinity Column Chromatography-The guanine nucleotide-bound forms of GST-small G proteins were made by incubating small G proteins (1.5 nmol) for 1 h at 30°C with 15 M GDP or GTP␥S in 1 ml of a reaction mixture (20 mM Tris/HCl at pH 7.5, 10 mM EDTA, 1 mM DTT, 5 mM MgCl 2 , 1 mM L-␣-dimyristoylphosphatidylcholine, and 0.3% CHAPS (17). GST-small G proteins (each 30 nmol) were immobilized on 1.25 ml of glutathione-Sepharose 4B, which packed into columns. Then, 300 ml of brain cytosolic fraction was preabsorbed to remove the native GST with 1 ml of glutathione-Sepharose 4B and was loaded onto the GST-small G protein affinity columns. The columns were washed with 12.5 ml (10 volumes) of buffer A (20 mM Tris/HCl at pH 7.5, 1 mM EDTA, 1 mM DTT, and 5 mM MgCl 2 ), followed by washing with 12.5 ml (10 volumes) of buffer A containing 50 mM NaCl. The proteins bound to the affinity columns were eluted four times by addition of 1 ml (0.8 volumes) of buffer A containing 200 mM NaCl.
Purification of p98 and Peptide Sequence-To purify p98, 3 liters of brain cytosolic fraction was used for GTP␥S⅐GST-R-Ras affinity column chromatography as described above. The second and third fractions of the 200 mM NaCl-eluates were dialyzed three times with distilled water and concentrated by freeze-drying. The concentrated samples were subjected to SDS-PAGE and transferred onto polyvinylidene difluoride membranes (18). The immobilized p98 was reduced and S-carboxymethylated, followed by in situ digestion with Achromobacter protease I and Asp-N as described previously (18). The digested peptides were fractionated by C18 column chromatography and subjected to amino acid sequencing (18).
Molecular Cloning and Determination of Nucleic Acid Sequence of Bovine R-Ras GAP cDNA-To amplify a partial fragment of R-Ras GAP cDNA, we performed PCR from bovine quick clone cDNA (Clontech) using degenerate oligonucleotide primers corresponding to the peptide sequences indicated by double underlines in Fig. 2. The amplified fragment was labeled with [␣-32 P]dCTP using a Random Primer DNA labeling kit (Takara Shuzo Co.) and used to screen a bovine brain cDNA library (1.2 ϫ 10 6 independent plaques in total) (19). The cDNA inserted into gt10 phage DNA was cloned into pUC18 for the nucleotide sequencing with an Applied Biosystems model 373S DNA sequencer.
Interaction of in Vitro Translated Recombinant R-Ras GAP with GST Small G Proteins-In vitro translation of pGEM-R-Ras GAP was performed using the TNT T7-coupled reticulocyte lysate system (Promega) under the conditions described in the instruction manual. GST-small G proteins (0.75 nmol each) were immobilized onto 31 l of glutathione-Sepharose 4B beads and washed with 310 l (10 volumes) of buffer A. The immobilized beads were added to 40 l of in vitro translated mixture and incubated for 1 h at 4°C with gentle mixing. The beads were washed three times with 102 l (3.3 volumes) of buffer A, and the bound proteins were eluted with GST-small G proteins by addition of 102 l (3.3 volumes) of buffer A containing 10 mM glutathione.
GTPase Assay-GTPase activity of Ha-or R-Ras was assayed by measuring the decreased radioactivity of [␥-32 P]GTP⅐GST-Ha-or R-Ras (20). [␥-32 P]GTP⅐GST-Ha-or R-Ras (10 pmol) was incubated at 30°C for appropriate periods in the presence or absence of MBP-R-Ras GAP-GRD or GST-NF1-GRD in 100 l of reaction mixture (20 mM Tris/HCl at pH 7.5, 5.35 mM EDTA, 1 mM DTT, 10 mM MgCl 2 , 1 mM GTP, and 0.15% CHAPS). The reaction was stopped by adding 3 ml of ice-cold stopping buffer (20 mM Tris/HCl at pH 8.0, 100 mM NaCl, and 25 mM MgCl 2 ), followed by rapid filtration using nitrocellulose filters. The filters were washed three times with the same ice-cold buffer, and bound radioactivity was determined using a liquid scintillation counter.
RT-PCR Analysis-To design rat R-Ras GAP primers for RT-PCR, we determined partial sequence of the rat R-Ras GAP cDNA as follows. A partial fragment of rat R-Ras GAP cDNA was amplified by PCR from rat brain Quick Clone cDNA (Clontech) with the same degenerate oligonucleotide primers as used for amplification of bovine R-Ras GAP partial cDNA. The amplified fragment of about 400 base pairs was cloned into pCRII vector (Invitrogen) and subjected to nucleotide sequencing as described above. Total RNA (1 g) from various adult rat tissues was subjected to RT-PCR with the following primers, 5Ј-ACA-CAGAAGACCACGTCTTCTC-3Ј and 5Ј-ATGGCTGCCTCTTGCTTATC-TC-3Ј which were designed from rat R-Ras GAP partial cDNA sequences. The PCR products were subjected to 6% polyacrylamide gel electrophoresis (21).
Other Procedures-SDS-PAGE was performed as described previously (22). Protein concentrations were determined with bovine serum albumin as a reference protein as described (23). The BLAST program was used for protein homology search (24).

RESULTS
Purification of GTP␥S⅐R-Ras-interacting Molecule-To detect R-Ras-interacting molecules, bovine brain cytosolic fraction was loaded onto a GST-R-Ras affinity column. The proteins bound to the affinity column were eluted by addition of 200 mM NaCl. A protein with mass of about 98 kDa (p98) was detected in the 200 mM NaCl-eluate from GTP␥S⅐GST-R-Ras affinity column but not from GST or GDP⅐GST-R-Ras affinity column (Fig. 1A). p98 was not detected in the eluate of GTP␥S⅐GST-R-Ras A64 affinity column. R-Ras A64 is structurally equivalent to Ha-Ras A38 , which has a mutation in the effectorinteracting domain (1,2). To further confirm the specificity of the interaction, affinity column chromatography using GST-Ha-Ras and GST-RalA was performed. p98 was eluted from neither GST-Ha-Ras nor GST-RalA affinity columns (Fig. 1B). Amino Acid Sequence Analysis of p98 -To identify the GTP␥S⅐R-Ras-interacting molecule, p98 was subjected to amino acid sequencing as described under "Experimental Procedures." Thirteen peptide sequences derived from p98 were determined. Two sequences of the peptides were used to design the degenerate oligonucleotide primers for amplification of the specific DNA fragments derived from the p98 cDNA. A fragment of about 400 base pairs was obtained and used as a probe for library screening. Of 1.2 ϫ 10 6 recombinant phage plaques from a bovine brain cDNA library, two clones hybridized with the probe. The nucleotide sequence of one of the cloned cDNAs of about 3.9 kilobase pairs was determined. The cDNA contained an open reading frame encoding a protein consisting of 834 amino acids. The calculated molecular mass was 95,384 Da, which is close to the apparent molecular mass of p98 estimated by SDS-PAGE. The deduced amino acid sequence is shown in Fig. 2. All of the 13 peptide sequences obtained were found within the deduced amino acid sequence. The neighboring sequence around the initiation codon was consistent with the translation initiation start site proposed by Kozak (25) but we found no termination codon in the preceding region. To confirm whether the first ATG is the real initiation codon, in vitro translation was performed using p98 cDNA cloned downstream of the T7 promoter of pGEM-3zf(ϩ). In vitro translated protein migrated with an apparent size of about 98 kDa, which was the same size as native p98, and co-migrated with native p98 purified from GTP␥S⅐GST-R-Ras affinity column (data not shown).
Structural Characteristics of p98 -As a result of homology search in GenBank protein data base, p98 showed a high degree of sequence similarity with rat Gap1 m , which is thought to be the mammalian counterpart of Drosophila Gap1 (9, 26). The identity of nucleic acid and amino acid sequences was 62.5% and 59.7%, respectively. To examine relationship between p98 and rat Gap1 m , we determined partial sequence of rat p98 using of a cDNA fragment amplified by PCR (data not shown). The partial amino acid sequences of rat p98 showed identities of 95.2% with bovine p98 but only 55.4% with Gap1 m . Thus, we concluded that p98 is a homologue of rat Gap1 m rather than its counterpart. Similarly to rat Gap1 m , p98 contained two C2 domains (27), GRD (28), and PH domain (29) (Fig. 3). The alignment of amino acid sequences in each domain are shown in Fig. 3. Since p98 shows a high degree of amino acid sequence similarly with Gap1 m and exhibits GAP activity toward R-Ras (see below), we designated it as R-Ras GAP.
Interaction of Recombinant R-Ras GAP with GTP␥S⅐GST-R-Ras-To address whether recombinant R-Ras GAP interacts with GTP␥S⅐R-Ras, immobilized GST-small G proteins were mixed with in vitro translated R-Ras GAP and interacting proteins were eluted with GST small G proteins by addition of glutathione. In vitro translated R-Ras GAP was co-eluted strongly with GTP␥S⅐GST-R-Ras but weakly with GST, GDP⅐GST-R-Ras, GTP␥S⅐GST-R-Ras A64 , GST-Ha-Ras, and GST-RalA (Fig. 4). The weak bands detected in the eluates other than that from GTP␥S⅐GST-R-Ras may result from nonspecific interaction. The slightly faster migrated band may be a degraded product of R-Ras GAP.
GAP Activity of R-Ras GAP-We examined whether R-Ras GAP-GRD stimulates intrinsic GTPase activity of R-Ras. As described previously (15), GST-NF1-GRD stimulated GTPase activity of R-Ras in both a time-and dose-dependent manner (Figs. 5 and 6). The rate constant of GST-NF1-GRD for Ha-Ras was about 5-fold higher than that for R-Ras. MBP-R-Ras GAP-GRD also stimulated GTPase activity of R-Ras in a time-and Small GTP-binding Protein and R-Ras-interacting Protein dose-dependent manner (Figs. 5 and 6). In contrast with the GAP activity of GST-NF1-GRD, the rate constant of MBP-R-Ras GAP-GRD for R-Ras was about 5-fold higher than that for Ha-Ras. MBP-R-Ras GAP-GRD did not stimulate the GTPase activities of Rap1 or Rho (data not shown). It has been reported that p120 Ras GAP and NF1 do not stimulate GTPase activity of Ha-Ras, which has a mutation in the effector-interacting domain. The GTPase activity of R-Ras A64 was not stimulated by GST-NF1-GRD as described for Ha-Ras A38 (Fig. 5A). We examined the GAP activity of MBP-R-Ras GAP-GRD toward the effector mutant of R-Ras. MBP-R-Ras GAP-GRD weakly stimulated the GTPase activity of R-Ras A64 but the activity was much lower than that toward wild-type R-Ras (Fig. 5B). Under the similar conditions, MBP-R-Ras GAP-GRD did not affect guanine nucleotide release from R-Ras (data not shown).
Tissue Distribution of R-Ras GAP-To examine the tissue distribution of R-Ras GAP, RT-PCR was performed using mRNA prepared from various rat tissues (Fig. 7). Primers were designed from partial nucleotide sequences of rat R-Ras GAP. R-Ras GAP mRNA was expressed highly in cerebrum and cerebellum; moderately in heart, spleen, thymus, lung, liver, kidney, and pancreas; and hardly in skeletal muscle, small intestine, adrenal grand, and testis.

DISCUSSION
In the present study, we purified an R-Ras-interacting protein, p98, by GST-R-Ras affinity column chromatography. p98 interacts with GTP␥S⅐R-Ras but not with GDP⅐R-Ras, GTP␥S⅐R-Ras A64 , GTP␥S⅐Ha-Ras, or GTP␥S⅐RalA. We determined partial amino acid sequences of peptides derived from p98, cloned its cDNA, and determined its primary structure. p98 shows a high degree of amino acid sequence similarity to Gap1 m , and recombinant GRD of p98 showed GAP activity toward R-Ras higher than that toward Ha-Ras. Taken together, these results clearly indicate that p98 serves as GAP for R-Ras. Since GAP specific for R-Ras was identified here for the first time, we designated p98 as R-Ras GAP.
Among the small G protein-interacting proteins, both target proteins and GAP appear to interact with small G proteins in a GTP-dependent fashion, and not to interact with their effector mutants. Since R-Ras GAP is the first molecule that specifically interacts with R-Ras in a GTP-dependent fashion, we speculate that R-Ras GAP may serve as a downstream target for R-Ras rather than GAP. However, this possibility seems unlikely, because genetic evidence indicates that Gap1, which shows a high degree of amino acid sequence similarity with R-Ras GAP, functions as a GAP rather than a downstream target for Ras in Drosophila (26).
The GAP activity for R-Ras was first detected in human spleen (30). This protein has been partially purified and shown to be the same as p120 Ras GAP. NF1 was also reported to exhibit GAP activity toward R-Ras (15). We have shown here that R-Ras GAP exhibits higher GAP activity toward R-Ras than toward Ha-Ras. Although we cannot rule out the possibility that p120 Ras GAP and NF1 serve as GAP for R-Ras as well as for Ha-Ras, it is more likely that p120 Ras GAP and NF1 primarily serve as GAP for Ha-Ras, and that R-Ras GAP primarily serves as GAP for R-Ras in vivo. Further studies are necessary to estimate how much R-Ras GAP contributes to the regulation of R-Ras in vivo.
RT-PCR experiments indicate that R-Ras GAP is highly expressed in cerebrum and cerebellum, moderately in heart, spleen, thymus, lung, liver, kidney, and pancreas and hardly in skeletal muscle, small intestine, adrenal grand, and testis, suggesting that R-Ras GAP plays important roles in brain. On the other hand, R-Ras is expressed in most tissues including skeletal muscle, small intestine, adrenal grand, and testis (data not shown). From these observations, it is conceivable that isoforms or different types of R-Ras GAP are expressed in the tissues where R-Ras GAP is hardly expressed.
R-Ras GAP has unique structural features such as C2 domains and PH domain which are also observed in Gap1 and Gap1 m . This suggests that Gap1 m and R-Ras GAP may share some functions or be regulated in a similar way in vivo. The C2 domain, which is observed in protein kinase C, synaptotagmin, and Rabphilin-3A, is believed to be involved in the binding to Ca 2ϩ and phospholipid (27,31,32). It is possible that R-Ras GAP is recruited to membranes via the C2 domains upon influx of Ca 2ϩ into cells. The PH domain is assumed to be involved in the binding to phosphatidylinositol-4,5-bisphosphate or ␤␥ subunits of trimeric G proteins (33,34). It is speculated that R-Ras GAP associates with these molecules through the PH domain in vivo. Further studies may provide insight into roles of the C2 and PH domains of R-Ras GAP, leading to better understanding of modes of action and activation of R-Ras GAP.
R-Ras has been reported to interact with Bcl-2 (12). Furthermore, it has been shown that R-Ras V38 increases the rate of apoptotic cell death in the setting of growth factor withdrawal, and that Bcl-2 completely abrogates this effect of R-Ras (13). R-Ras V38 has been shown to interact with c-Raf-1, to activate MAP kinase cascade, and to induce transformation of NIH/3T3 cells (14 -16). However, it is not yet clear how R-Ras V38 accelerates apoptotic cell death in growth factor-deprived cells and promotes transformation of some types of cells such as NIH/ 3T3 cells, and how R-Ras is activated presumably downstream of receptors for some extracellular signals. Several groups have demonstrated that Ras is involved in regulation of a variety of cell functions including cell transformation, proliferation, and differentiation by utilizing p120 Ras GAP as a probe (35,36). Overexpression of p120 Ras GAP suppresses growth factorand Ras-mediated responses leading to cell transformation, proliferation, and differentiation. Similarly to p120 Ras GAP, R-Ras GAP will enable us to dissect how R-Ras regulates various cell functions and how R-Ras is regulated during actions of certain extracellular signals.