Three recent reports have suggested that rat liver microsomal glucose-6 phosphatase (Glc6Pase) should be a 62-64 kDa polypeptide. In this work, we examine the possibility that the 62-64 kDa could represent a functional dimeric form of the 36.5 kDa glucose-6 phosphate (Glc6P)-phosphohydrolase, previously identified [Countaway et al. (1988) J. Biol. Chem. 263, 2672-2678]. From 32P-labeling experiments with 32P-Glc6P and analysis of 32P-labeled protein by SDS-PAGE and autoradiography, we show that three different rat liver microsomal polypeptides, the apparent molecular masses of which are 62, 54, and 37 kDa, may be specifically labeled with 32P-Glc6P. We demonstrate that the 62 kDa polypeptide is a microsome-bound form of cytosolic phosphoglucomutase, by combining labeling competition experiments and enzymatic assay. It should likely not account for a putative dimeric form of Glc6P-phosphohydrolase. The 37 kDa polypeptide fulfills the criteria of Glc6P-phosphohydrolase. We have not obtained any definitive evidence for its assembly as a dimer under functioning conditions. The 32P-Glc6P-labeling characteristics of the 54 kDa polypeptide are those expected for a protein displaying affinity in the millimolar range of concentration and a high binding capacity for Glc6P. They are consistent with those of a 54 kDa microsomal polypeptide, previously suggested to be involved in Glc6Pase activity.