Comparative analysis of p21 proteins from various cell types by two-dimensional gel electrophoresis

The products of the ras gene family are related proteins at a molecular weight of 21 kDa, designated p21. In the present study we used two-dimensional gel electrophoresis to compare p21 proteins from five different normal and malignant cell lines. Using a known protein (3H-labeled translation initiation factor [eIF-4D]) as a standard internal marker for isoelectric point (pI), we show that p21 proteins from various cells differ only slightly in molecular weight (21-24 kDa) but express a wide variety in charge (pI 4.8 to 7) that could only be detected by the use of two-dimensional gel electrophoresis. p21 in NIH/3T3 cells was expressed as a single protein, which migrated at 21 kDa and pI 5.1. This peptide, which is probably the product of the normal cellular ras gene, was also detected in normal human lymphocytes. The synthesis of this peptide was not elevated in the transformed cells. However, transformation of NIH/3T3 fibroblasts and of human leukocytes was found to be associated with expression of qualitatively different forms of p21 peptides. Four additional p21-associated peptides of identical molecular weight (23 kDa), but multiple charge forms, were detected selectively in Kirsten murine sarcoma virus-transformed NIH/3T3 cells. Transformation of cells with Harvey murine sarcoma virus was found to be associated with prominent expression of two major pairs of p21-associated proteins, one at 21 kDa (pI, 5.2 and 5.3) and the other at 23 kDa (pI, 5.1 and 5.2). In HL-60 leukemic cells there was an additional, more acidic form (pI 5.0) of p21, which appeared to be absent or reduced in normal human lymphocytes. These results indicate that p21 from viral origin or cellular origin might be expressed in the cells in multiple charge forms. The capability to distinguish multiple forms of p21 and slight charge modifications associated with malignancy should call for the use of 2-D gel electrophoresis as an important tool in future studies involving p21 proteins.     

Expression of ras-like proteins in embryonic and adult cells of Xenopus laevis

A polyclonal antibody raised against v-Ha-ras p21 was purified and its specificity was checked on Ha-ras transformed cell lines. It was used to immunoprecipitate p21 from different Xenopus laevis cell types: brain cells, blood cells, and embryonic material. By one-dimensional Western blot analysis, we show that ras p21 is synthesized very early in oogenesis and accumulates throughout vitellogenesis. The ras p21 content, estimated to be 1.1 ng in the full-grown oocyte, remains constant during oocyte maturation and egg cleavage. Increase in the amount of ras p21 occurs at the beginning of neurulation. Two-dimensional Western blot patterns reveal the presence of multiple molecular forms of p21 in all Xenopus cell types studied. The numerous resolved polypeptides were ascribed to the expression of at least two different ras genes. Furthermore, specific charge modifications of the ras polypeptides are observed in brain, blood, and embryonic cells. During oogenesis and early embryonic development, differences in two-dimensional patterns mainly concern variations in the relative amounts of the different polypeptides. The results are discussed in relation to the well documented synthesis activities of the growing oocyte and of the early developing embryo.     

Association of p21ras with cellular polypeptides

p21ras specific antiserum was used to immunoprecipitate p21ras polypeptides from human A431 cells. In addition to p21ras, this antiserum precipitated a series of polypeptides with relative molecular weights of 150,000, 120,000, 105,000, and 50,000. The precipitation of these polypeptides was prevented by preincubation of the antiserum with an excess of purified Ras protein. These polypeptides do not share an epitope with p21ras, and two of them (120 and 150 kDa) copurify with a fraction of p21ras. The co-precipitation of p21ras with these polypeptides was detected in a variety of cell types. The pattern of the immunoprecipitates was consistently different in normal and ras-transformed cells. The 120- and 150-kDa polypeptides are phosphorylated on serine and threonine in A431 cells. Serum treatment resulted in a 2-fold increase in the phosphoserine content of the 120-kDa polypeptides.     

Immunohistochemical analysis of normal and mutated ras oncogene p21 expression in human pulmonary and pleural neoplasms

In this study we examined 214 cases of primary human pulmonary neoplasms for the expression of a mutated form of the ras oncogene p21 product, recognized by the monoclonal antibody (MCA) DWP. Adjacent serial sections from these same cases had previously been used to demonstrate the frequency of ras p21 expression using the broadly reactive anti-ras p21 MCA RAP-5. Confirmation of the increased expression of p21 was accomplished using MCA Y13-259. The use of adjacent tissue sections from these cases allows the direct comparison of the expression of the mutated and non-mutated forms of ras p21. If reactivity with DWP would prove to be significantly more restrictive than that of the "pan" ras MCAs, RAP-5 and Y13-259, it would lend support to the possibility that DWP (and similar MCAs which detect other specific mutations) could be used to define subsets of these neoplasms based on their specific ras p21 phenotype. Since one would anticipate that the valine/cysteine substitution at position 12 of the ras p21 would occur at only low frequencies in human tumors, our results with DWP are consistent with this hypothesis. As previously reported, RAP-5 reacted with a high proportion of lung tumors (100/214 or 47%). In this report, we demonstrate the selective expression of the mutation recognized by the MCA DWP in only 5% of these same tumors (13/214), and that the expression of this mutated form is not restricted to any of the conventional histological subclasses of pulmonary neoplasms.     

Studies on the structure of sphingomyelinase. Amino acid composition and heterogeneity on isoelectric focusing.

Sphingomyelinase, purified to apparent homogeneity from human placenta, is an acidic protein, as judged from its amino acid composition and by isoelectric focusing of the carboxymethylated protein. The amino acid composition is characterized by an approximately equal content of hydrophobic and polar amino acid residues. The reduced-alkylated polypeptides were separated into two groups. Most of the polypeptides were heterogeneous with pI values of 4.4-5.0, but an additional more minor component was observed at pI 5.4. Liquid isoelectric focusing resolved the purified enzyme into a single major component (pI 4.7-4.8), a minor component (pI 5.0-5.4) and a plateau region of activity (pI 6-7). On thin-layer isoelectric focusing, the protein profile obtained from each of these regions was the same. In addition, the substrate specificity, Km values and effect of inhibitory substances were identical. We conclude that sphingomyelinase is an acidic, microheterogeneous protein that likely exists as a holopolymer of a single major polypeptide chain. the heterogeneity of the intact protein on isoelectric focusing appears to reflect this microheterogeneity, which is influenced by a tendency to associate with itself and with detergents such as Triton X-100.     

Activation of cellular p21ras by myristoylation

p21ras is palmitoylated on a cysteine residue near the C-terminus. Changing Cys-186 to Ser in oncogenic forms produces a non-palmitoylated protein that fails to associate with membranes and does not transform NIH 3T3 cells. To examine whether palmitate acts in a general way to increase ras protein hydrophobicity, or is involved in more specific interactions between p21ras and membranes, we constructed genes that encode non-palmitoylated ras proteins containing myristic acid at their N-termini. Myristoylated, activated ras, without palmitate (61Leu/186Ser) exhibited both efficient membrane association and full transforming activity. Unexpectedly, we found that myristoylated forms of normal cellular ras were also potently transforming. Myristoylated c-ras retained the high GTP binding and GTPase characteristic of the cellular protein and, moreover, bound predominantly GDP in vivo. This implied that it continued to interact with GAP (GTPase-activating protein). While the membrane binding induced by myristate permitted transformation, only palmitate produced a normal (non-transforming) association of ras with membranes and must therefore regulate ras function by some unique property that myristate does not mimic. Myristoylation thus represents a novel mechanism by which the ras proto-oncogene protein can become transforming.     

Deletion mutants of Harvey ras p21 protein reveal the absolute requirement of at least two distant regions for GTP-binding and transforming activities.

Deletions of small sequences from the viral Harvey ras gene have been generated, and resulting ras p21 mutants have been expressed in Escherichia coli. Purification of each deleted protein allowed the in vitro characterization of GTP-binding, GTPase and autokinase activity of the proteins. Microinjection of the highly purified proteins into quiescent NIH/3T3 cells, as well as transfection experiments utilizing a long terminal repeat (LTR)-containing vector, were utilized to analyze the biological activity of the deleted proteins. Two small regions located at 6-23 and 152-165 residues are shown to be absolutely required for in vitro and in vivo activities of the ras product. By contrast, the variable region comprising amino acids 165-184 was shown not to be necessary for either in vitro or in vivo activities. Thus, we demonstrate that: (i) amino acid sequences at positions 5-23 and 152-165 of ras p21 protein are probably directly involved in the GTP-binding activity; (ii) GTP-binding is required for the transforming activity of ras p21 and by extension for the normal function of the proto-oncogene product; and (iii) the variable region at the C-terminal end of the ras p21 molecule from amino acids 165 to 184 is not required for transformation.     

Molecular basis of the isozymes of bovine glucose-6-phosphate isomerase

Glucose-6-phosphate isomerase occurs in different bovine tissues as multiple, catalytically active isozymes which can be resolved by polyacrylamide gel electrophoresis (PAGE) and isoelectric focusing (IEF). Most differentiated tissues have five distinct forms with apparent pI values of 7.2, 7.0, 6.8, 6.6, and 6.4. Young, mitotically active, cells of the intestinal mucosa and the epithelium of the eye lens show only the two more basic isozymes, while old cells in the cortex and nucleus of the eye lens accumulate the more acidic isozymes. All of the isozymes exhibit equal separation based on charge-to-mass ratio (PAGE) and charge (IEF), thus indicating only charge changes. The isozyme patterns are unchanged in the presence of reducing agents or protease inhibitors. Each isozyme was purified to homogeneity and shown to exhibit identical subunit molecular weights (59,000) on SDS-gel electrophoresis. Each of the isolated isozymes, when subjected to PAGE or IEF, exhibited a single band, indicating that the isozymes are not generated as a result of electrophoresis. When the most basic isozyme was incubated in vitro under mild alkaline conditions, there was a spontaneous generation of the more acidic isozymes with properties identical to those found in vivo. The isozymes, thus, appear to be the result of spontaneous, postsynthetic modifications involving the addition of equal numbers of negative charges and are consistent with the deamidation of specific asparagine and/or glutamine residues.     

Abnormal regulation of mammalian p21ras contributes to malignant tumor growth in von Recklinghausen (type 1) neurofibromatosis.

Tumor cell lines derived from malignant schwannomas removed from patients with neurofibromatosis type 1 (NF1) have been examined for the level of expression of NF1 protein. All three NF1 lines examined expressed lower levels of NF1 protein than control cells, and the level in one line was barely detectable. The tumor lines expressed normal levels of p120GAP and p21ras. Although the p21ras proteins isolated from the tumor cells had normal (nonmutant) biochemical properties in vitro, they displayed elevated levels of bound GTP in vivo. The level of total cellular GAP-like activity was reduced in extracts from the tumor line that expresses very little NF1 protein. Introduction of the catalytic region of GAP into this line resulted in morphological reversion and lower in vivo GTP binding by endogenous p21ras. These data implicate NF1 protein as a tumor suppressor gene product that negatively regulates p21ras and define a "positive" growth role for ras activity in NF1 malignancies.     

Post-translational modifications of p21rho proteins.

Post-translational modifications of the ras proteins, which are required for plasma membrane localization and biological function of the proteins, have been shown to include prenylation and carboxymethylation at the carboxyl terminal cysteine residue of the cysteine-aliphatic amino acid-aliphatic amino acid-any amino acid (CAAX) box. In addition, p21Ha-ras and p21N-ras, but not p21K-ras (B), are palmitoylated. The three mammalian rho proteins (A, B, and C) are also members of the ras superfamily but have distinct biological activities and different intracellular distributions from p21ras. Analysis showed all three rho proteins are modified by a COOH-terminal carboxymethylation similar to p21ras, whereas p21rhoC labeled with [3H]mevalonic acid in vivo revealed the presence of a C20 prenoid, similar to that already described for p21rhoA. However, in vivo and in vitro studies of p21rhoB showed this protein to be modified by both C15 and C20 prenoids. Mutation of C193 in the CAAX box abolished prenylation, whereas mutation of the adjacent C192 resulted in a significant reduction in the amount of the C20, but not C15 prenoid, recovered from p21rhoB. In vivo labeling studies with [3H]palmitic acid and mutational analysis showed that both cysteine residues at 189 and 192 upstream of the CAAX box in p21rhoB are sites for palmitoylation. We conclude that there are different populations of post-translationally modified p21rhoB in the cell and that the sequence specificity for geranylgeranyl- and farnesyltransferases may be more complicated than previously proposed.     

An affinity labeling of ras p21 protein and its use in the identification of ras p21 in cellular and tissue extracts

We have carried out photoaffinity labeling of the ras p21 protein, a ras oncogene product, with [alpha-32P]GTP. Based on our studies, a sensitive, rapid, and specific assay for the detection of multiple forms of ras p21 has been developed. The specificity of this protocol is shown by (a) sensitivity of affinity labeling of ras p21 to known inhibitors of GTP binding and (b) immunoprecipitation of affinity labeled protein with anti-ras p21 serum. Detection and semiquantitation of ras p21 by this method is accomplished in less than 24 h and requires as little as 100,000 cells or about 5 mg of tissue sample from skin tumor, liver, and mammary tumor tissues. Furthermore, using this approach, we were able to detect the selective loss of one species of ras p21 in transplanted Morris hepatoma cells.     

Post-translational modification of low molecular mass GTP-binding proteins by isoprenoid

Several proteins in mammalian cells are modified post-translationally by the isoprenoid, farnesol. Incubation of cultured cells with [3H]mevalonate, an isoprenoid precursor, results in the labeling of multiple polypeptides, the most prominent of which migrate in the range of 21-26 kDa on sodium dodecyl sulfate-polyacrylamide gels. In Rat-6 fibroblasts transformed by H-ras, one of the farnesylated proteins was identified as p21ras by two-dimensional immunoblotting. However, this protein accounted for only a small proportion of the [3H]mevalonate-derived radioactivity incorporated into 21-26-kDa proteins. Murine erythroleukemia cells, which did not express immunodetectable quantities of p21ras, contained several 21-26-kDa farnesylated proteins distributed in both the cytosolic and particulate fractions. At least eight of these proteins were capable of binding [alpha-32P]GTP on nitrocellulose membranes. Pulse-chase studies showed that the isoprenoid modification did not necessarily result in the translocation of the cytosolic proteins to the cell membrane. A prominent group of carboxyl-methylated proteins in murine erythroleukemia cells overlapped with the 21-26-kDa farnesylated proteins on one-dimensional sodium dodecyl sulfate gels. Methylation of this group of proteins was selectively abolished when cells were treated with lovastatin, an inhibitor of isoprenoid synthesis. Addition of exogenous mevalonate to the lovastatin-treated cells fully restored carboxyl methylation. These studies suggest that the 21-26-kDa farnesylated proteins in mammalian cells are members of a recently discovered family of low molecular mass GTP-binding proteins which, although ras-related, appear to be distinct structurally and possibly functionally from the products of the ras genes. The observed isoprenoid-dependent carboxyl methylation of a group of 21-26-kDa proteins suggests that the low molecular mass GTP-binding proteins may undergo a series of post-translational C-terminal cysteine modifications (i.e. farnesylation, carboxyl methylation) analogous to those recently elucidated for p21ras.     

Inhibition of protein kinase C zeta subspecies blocks the activation of an NF-kappa B-like activity in Xenopus laevis oocytes.

Nuclear factor kappa B (NF-kappa B) plays a critical role in the regulation of a large variety of cellular genes. However, the mechanism whereby this nuclear factor is activated remains to be determined. In this report, we present evidence that in oocytes from Xenopus laevis, (i) ras p21- and phospholipase C (PLC)-mediated phosphatidylcholine (PC) hydrolysis activates NF-kappa B and (ii) protein kinase C zeta subspecies is involved in the activation of NF-kappa B in response to insulin/ras p21/PC-PLC. Thus, the microinjection of either ras p21 or PC-PLC, or the exposure of oocytes to insulin, promotes a significant translocation to the nucleus of an NF-kappa B-like activity. This effect is not observed when oocytes are incubated with phorbol myristate acetate or progesterone, both of which utilize a ras p21-independent pathway for oocyte activation. These data strongly suggest a critical role of the insulin/ras p21/PC-PLC/protein kinase C zeta pathway in the control of NF-kappa B activation.     

The human Rab genes encode a family of GTP-binding proteins related to yeast YPT1 and SEC4 products involved in secretion

Seven cDNA clones corresponding to the rab1, rab2, rab3A, rab3B, rab4, rab5, and rab6 genes were isolated from a human pheochromocytoma cDNA library. They encode 23-25 kDa polypeptides which share approximately 30-50% homology and belong to the ras superfamily. The rab1, rab2, rab3A, and rab4 proteins are the human counterparts of the rat rab gene products that we have previously characterized. Comparison of the seven human rab proteins with the yeast YPT1 (YPT1p) and SEC4 (SEC4p) proteins reveals highly significant sequence similarities. H-rab1p shows 75% amino acid identity with YPT1p and may be therefore considered as its human counterpart. The other proteins share approximately 40% homology with YPT1p and SEC4p. The homology (approximately 30%) between these rab proteins and p21ras is restricted to the four conserved domains involved in the GTP/GDP binding. Human rab proteins were produced in Escherichia coli. Large amounts of rab proteins in soluble form can be extracted and purified without the use of detergents. All six proteins bind GTP and exhibit GTPase activities. A possible involvement of the rab proteins in secretion is discussed.     

Purification and characterization from rat liver cytosol of a GDP dissociation inhibitor (GDI) for liver 24K G, a ras p21-like GTP-binding protein, with properties similar to those of smg p25A GDI

A regulatory protein for a liver GTP-binding protein (G protein) with a molecular weight value of 24,000 (24K G), which we have recently purified, was purified to near-homogeneity from rat liver cytosol and characterized. This regulatory protein, designated here as GDP dissociation inhibitor for 24K G (24K G GDI), inhibited the dissociation of GDP from and the subsequent binding of GTP to 24K G. 24K G GDI was inactive for other ras p21/ras p21-like small G proteins including c-Ha-ras p21, rhoB p20, smg p21B, and smg p25A. 24K G was, however, recognized by bovine brain smg p25A GDI which regulated the GDP/GTP exchange reaction of smg p25A. By analyses of sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE), immunoblotting with anti-smg p25A GDI antibody, two-dimensional PAGE, and C4 column chromatography, 24K G GDI showed physical properties very similar to those of smg p25A GDI. The peptide map and the partial amino acid sequences of 24K G GDI were not identical with those of smg p25A GDI. Among the 83 residues, 2 amino acids were different between rat liver 24K G GDI and bovine brain smg p25A GDI. These results indicate that there is a specific regulatory protein for 24K G, 24K G GDI, in rat liver cytosol and that 24K G GDI has close similarity to smg p25A GDI.     

CD2 antigen mediated activation of the guanine nucleotide binding proteins p21ras in human T lymphocytes.

T cell stimulation via the TCR complex (TCR/CD3 complex) results in activation of the guanine nucleotide binding proteins encoded by the ras protooncogenes (p21ras). In the present study we show that the activation state of p21ras in T lymphocytes can also be controlled by triggering of the CD2 Ag. The activation state of p21ras is controlled by GTP levels on p21ras. In T cells stimulation of protein kinase C is able to induce an accumulation of "active" p21ras-GTP complexes due to an inhibitory effect of protein kinase C stimulation on the intrinsic GTPase activity of p21ras. The regulatory effect of protein kinase C on p21ras GTPase activity appears to be mediated via regulation of GAP, the GTPase activating protein of p21ras. In the present report, we demonstrate that the TCR/CD3 complex and the CD2 Ag control the accumulation of p21ras-GTP complexes via a regulatory effect on p21ras GTPase activity. The TCR/CD3 complex and CD2 Ag are also able to control the cellular activity of GAP. These data demonstrate that p21ras is part of the signal transduction responses controlled by the CD2 Ag, and reveal that the TCR/CD3 complex and CD2 Ag control the activation state of p21ras via a similar mechanism.     

Expression and characterization of ras mRNAs from Saccharomyces cerevisiae.

The cellular homologs of the Harvey and Kirsten murine sarcoma virus oncogenes comprise a multigene family, ras, that displays striking evolutionary conservation. We recently reported [DeFeo-Jones et al., Nature (London) 306:707-709, 1983] the cloning of two ras homologs from the yeast Saccharomyces cerevisiae. The nucleotide sequences of these genes predict polypeptides that show remarkable homology to p21, the mammalian ras gene product. We have also found proteins in yeast lysates with serological cross-reactivity to p21 (Papageorge et al., Mol. Cell. Biol. 4:23-29, 1984). In this work, we explored the relationship between the immunoprecipitated proteins and the yeast ras genes. We show that both ras genes are expressed in the wild-type cell. Furthermore, we demonstrate by in vitro translation of hybrid-selected RASsc1 mRNA and immunoprecipitation of the translation products that the cloned RASsc1 gene encodes the proteins immunoprecipitated from yeast lysates by anti-p21 monoclonal antibody. Finally, we used anti-p21 monoclonal antibodies to detect a guanine nucleotide binding activity in yeast lysates. The structural and biochemical homologies between ras gene products of S. cerevisiae and mammalian cells suggest that information obtained by genetic analysis of ras function in a lower eucaryote should be applicable to higher organisms as well.     

Proteolytic conversion of insulin-like growth factors to an acidic form(s).

The relative amounts of the various forms of bioassayable insulin-like growth factors (IGF) isolated from human serum or serum fraction Cohn IV-1 depend on the purification procedure. With acid gel filtration or acid/ethanol extraction as the initial step, IGF-II (pI approximately 6.5) was the most abundant (40-70%) followed by somatomedin A (pI approximately 7.4; 15-23%), an acidic form of insulin-like activity (ILA pI 4.8) (13-21%) and IGF-I (pI approximately 8.5; 5-27%). If, however, pH 5.5 ion-exchange chromatography on SP-Sephadex was used prior to acid gel filtration, the acidic pI 4.8 form was the major (greater than 90%) species recovered and was accompanied by a quantitative loss of the other IGF species. This suggested a possible conversion of IGF-I, somatomedin A and/or IGF-II to the acidic ILA pI 4.8 form(s) during the SP-Sephadex procedure. Further experiments indicated that differences in the yields of ILA pI 4.8 were not due simply to differences in the initial pH conditions of the various methods (i.e. acid versus neutral), although exposure to pH 9.7 (a pH experienced during elution of IGF activity from the SP-Sephadex) did appear to play a role. The involvement of the carrier protein in the conversion process was tested by subjecting carrier-free IGF-I and IGF-II to the SP-Sephadex procedure. No conversion of the free forms to ILA pI 4.8 occurred. To examine the possible role of proteinase in the conversion of IGFs to ILA pI 4.8, SP-Sephadex chromatography was performed in the presence of a broad spectrum proteinase inhibitor. The IGF distribution pattern obtained closely resembled the 'normal' pattern seen with acid gel filtration, indicating that proteinase inactivation had prevented conversion to ILA pI 4.8. These data suggest that proteolytic conversion of IGF-I, somatomedin A and IGF-II to more acidic ILA pI 4.8 form(s) (i) occurs during SP-Sephadex chromatography, (ii) is not prevented simply by prior acid exposure, and (iii) takes place only when IGF-I and -II are in their high-Mr carrier-bound forms. Since IGF-I and IGF-II, although homologous, have unique amino acid sequences, the conversion of both IGFs implies that at least two acidic ILA forms exist. Nevertheless, because ILA pI 4.8 retains the full spectrum of IGF bioactivities in vitro, and significant quantities are present in normal human serum (21%), it would suggest that proteolytic conversion of IGF-I, somatomedin A and IGF-II to ILA pI 4.8 in vivo may be a physiologically significant event.     

Stoichiometric interaction of smg p21 with its GDP/GTP exchange protein and its novel action to regulate the translocation of smg p21 between membrane and cytoplasm

We have previously purified a GDP/GTP exchange protein for smg p21A and -B, members of a ras p21/ras p21-like small GTP-binding protein superfamily. This regulatory protein, named smg p21 GDP dissociation stimulator (GDS), stimulates the dissociation of both GDP and GTP from and the subsequent binding of both GDP and GTP to smg p21s. We show here that smg p21 GDS forms a complex with both the GDP- and GTP-bound forms of smg p21B at a molar ratio of about 1:1. Both the GDP- and GTP-bound forms of smg p21B bound to membranes. smg p21 GDS inhibited this binding and moreover induced the dissociation of the prebound smg p21B from the membranes. These results indicate that smg p21 GDS stoichiometrically interacts with smg p21B and thereby regulates its GDP/GTP exchange reaction and its translocation between membranes and cytoplasm.     

Inhibition of the ras p21 GTPase-activating protein-stimulated GTPase activity of c-Ha-ras p21 by smg p21 having the same putative effector domain as ras p21s

ras p21 GTPase-activating protein (GAP) has been proposed to interact with the putative effector domain of ras p21s, and smg p21, a ras p21-like guanine nucleotide binding protein (G protein), has been shown to have the same amino acid sequence as ras p21s in this region. In the present studies, we examined the effects of ras p21 GAP on the GTPase activity of smg p21 purified from human platelets, of smg p21 on the ras p21 GAP-stimulated GTPase activity of c-Ha-ras p21 purified from Escherichia coli, and of c-Ha-ras p21 on the smg p21 GAP1- or -2-stimulated GTPase activity of smg p21. ras p21 GAP stimulated the GTPase activity of c-Ha-ras p21 but not that of smg p21. The GTP-bound form of smg p21, however, inhibited the ras p21 GAP-stimulated GTPase activity of c-Ha-ras p21 in a dose-dependent manner. The half-maximum inhibition by smg p21 was obtained at 0.4 microM which was more potent than previously observed for ras p21 (2-200 microM). The GDP-bound form also inhibited the ras p21 GAP-stimulated GTPase activity of c-Ha-ras p21, but the efficiency was 40-50% that of the GTP-bound form. smg p21 GAP1 and -2 stimulated the GTPase activity of smg p21 but not that of c-Ha-ras p21. c-Ha-ras p21 did not inhibit the smg p21 GAP1- or -2-stimulated GTPase activity of smg p21. These results indicate that ras p21 GAP interacts with smg p21 without the subsequent stimulation of its GTPase activity.