We have observed that the approximately 90-kDa non-steroid-binding component of nonactivated glucocorticoid receptors purified from WEHI-7 mouse thymoma cells (which has been identified as the approximately 90-kDa heat shock protein) consistently migrates as a doublet during polyacrylamide gel electrophoresis under denaturing and reducing conditions. It has recently been reported that murine Meth A cells contain a tumor-specific transplantation antigen (TSTA) which is related or identical to the approximately 90-kDa heat shock protein (Ullrich, S.J., Robinson, E.A., Law, L.W., Willingham, M., and Appella, E. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 3121-3125). The observation that TSTA and the approximately 90-kDa heat shock protein isolated from these cells exists as two isoforms of similar molecular mass and charge has suggested to us that the doublet we observed is also due to the existence of two isoforms. However, unlike TSTA, which appears to contain the two isoforms in similar relative abundance, nonactivated glucocorticoid-receptor complexes seem to contain predominantly the lower molecular mass isoform. We have therefore conducted this study to determine whether TSTA and the approximately 90-kDa component of glucocorticoid receptors are indeed related, to establish whether the receptor preferentially binds one isoform of the approximately 90-kDa heat shock protein, and to investigate the stoichiometry of the nonactivated receptor complex. By comparing Meth A TSTA and the approximately 90-kDa component of the receptor in their reactions with the AC88 monoclonal antibody (specific for the approximately 90-kDa heat shock protein) and a polyclonal antibody directed against Meth A TSTA, we found that these two proteins are indistinguishable and probably identical. We then used the BuGR1 (directed against the steroid-binding subunit of glucocorticoid receptors) and AC88 monoclonal antibodies to purify, respectively, receptor-associated and free approximately 90-kDa heat shock protein from WEHI-7 cells grown for 48 h with [35S]methionine to metabolically label proteins to steady state. Following analysis of the proteins by polyacrylamide gel electrophoresis under denaturing and reducing conditions, the relative amounts of the two isoforms in each sample were determined from the 35S counts and the known methionine content of each isoform. We found that approximately three-quarters of both the receptor-associated and the free approximately 90-kDa heat shock protein is present as the lower molecular weight isoform, indicating no preferential binding of either isoform in the receptor. The long-term metabolic labeling approach has also enabled us to direc