The v-rasH oncogene of Harvey murine sarcoma virus encodes a 21,000-dalton p21 protein which has been expressed at a high level as a fusion protein in Escherichia coli. We have purified the p21 to over 90% in purity without the use of any detergent or protein denaturant. The purified p21 possesses full biochemical activities of GTP/GDP binding, autokinase, and GTPase. Scatchard analysis indicates a single class of binding sites with Kd values of 0.83 x 10-8 M for GTP and 1.0 x 10-8 M for GDP. The binding site can be specifically labeled with a [3H]GTP photoaffinity analog, P3-(4-azidoanilido)-5' GTP. To probe for the active center of p21, we used a battery of six monoclonal antibodies to p21 to examine their effects on p21 activities. We found that only one monoclonal antibody, Y13-259, was capable of inhibiting both GTP/GDP binding and autokinase enzymatic activities, suggesting that these p21 activities are related activities conferred by a single active center within the p21 molecule. These observations together with the recent finding that microinjection of the same monoclonal antibody into NIH 3T3 cells specifically blocks p21 in vivo function (Mulcahy et al., Nature [London] 313:241, 1985) strongly suggest that p21 in vitro activities are responsible for its cellular function.

The EJ bladder carcinoma oncogene is activated by a point mutation in the c-rasH proto-oncogene at the 12th amino acid codon. In an attempt to understand the mechanism of oncogenic activation, a comparative study was undertaken to examine the metabolic turnover and subcellular localization of the p21 protein encoded by the EJ oncogene, the viral oncogene, and its normal cellular homolog. Pulse-labeling experiments indicated that both c-ras p21 proteins were synthesized by a very similar pathway, as was observed for the viral p21 protein of Harvey murine sarcoma virus. The pro-p21 proteins were detected in free cytosol, and the processed products were associated with plasma membrane. The intracellular half-life of p21 proteins was determined by pulselabeling and chasing in the presence of excess unlabeled methionine. Although both p21 proteins of EJ and the normal c-ras genes which are not phosphorylated have a half-life of 20 h, the viral p21 protein of Harvey murine sarcoma virus which includes a phosphorylated form is much more stable in cells, having a half-life of 42 h, apparently due to phosphorylation.

The gene for the Harvey murine sarcoma virus (Ha-MuSV) p21ras protein was fused to the amino-terminal portion of the bacteriophage lambda cII gene on the expression vector pJL6. The fusion was such that transcription was controlled by the well-regulated phage lambda pL promoter, and translation initiated in the cII gene continued in frame into the ras gene sequences that code for p21. When the pL promoter was derepressed, the Escherichia coli cells harboring the fusion plasmid synthesized 23,000-dalton protein, which represented more than 10 percent of the total cellular protein. This protein was chimeric and contained 14 residues, which were specified by the vector; these residues were followed by all of the amino acids that make up Ha-MuSV p21ras except for four residues at the amino-terminal end. The protein appears similar to Ha-MuSV p21ras in that it undergoes immunoprecipitation by monoclonal antibodies directed toward that protein, binds guanosine diphosphate, and is capable of autophosphorylation.

The p21 proteins of ras oncogenes are synthesized as precursors in the cytosol. After processing, which involves acylation, the products are associated with the plasma membrane in eucaryotic cells. The p21 overproduced in Escherichia coli, however, is not processed by acylation. A synthetic tetrapeptide of the p21 C terminus is used to identify the acylation site in eucaryotic p21 as cysteine-186. The same peptide of bacterial p21 is not acylated. Although p21 of Harvey murine sarcoma virus-transformed NRK cells can be metabolically labeled with either [3H]palmitate or [3H]myristate, the lipid moiety of the hydrophobic peptide is identified as palmitic acid. We suggest that the enzymatic mechanism for p21 palmitylation may be different from N-terminal myristylation of many other membrane proteins.