The simian virus 40 large tumor antigen (T-ag) is found in both the nuclei (nT-ag) and plasma membranes (mT-ag) of simian virus 40-infected or -transformed cells. It is not known how newly synthesized T-ag molecules are recognized, sorted, and transported to their ultimate subcellular destinations. One possibility is that these events depend upon structural differences between nT-ag and mT-ag. To test this possibility, we compared the structures of nT-ag and mT-ag from simian virus 40-infected cells. No differences between the two forms of T-ag were detected by migration in polyacrylamide gels, by Staphylococcus aureus V8 partial proteolytic mapping of methionine- or proline-containing peptides, or by two-dimensional tryptic peptide mapping of methionine-containing peptides. The carboxy-terminal, methionine-containing tryptic peptide was identified in the two-dimensional maps and was shown to be identical in nT-ag and mT-ag. Thus, a structural basis for the recognition and differential localization of T-ags could not be demonstrated. The carboxy terminus of the T-ag encoded by mutant dlA2413 is derived from the alternate open reading frame of the simian virus 40 early region, in analogy with the theoretical early gene product, T*-ag. We used this mutant to identify peptides unique to T*-ag. None of these peptides were detected in maps of mT-ag; only wild-type T-ag-specific peptides were found. These findings suggest that T*-ag does not represent the membrane-associated form of T-ag, but that mT-ag is encoded within the same reading frame used for nT-ag.