We have identified the subunits of the insulin receptor using immunoprecipitation by antibodies to the insulin receptor after either biosynthetic or surface labeling of cultured human lymphocytes (IM-9). With this approach, we have found there are two major, Mr = 135,000 (alpha), Mr = 95,000 (beta) and one minor, Mr = 210,000 (gamma) subunit. Peptide mapping clearly demonstrates that the major peptides of the alpha and beta subunits are different, whereas similarities exist in the peptide fragments of the gamma subunit and the alpha and beta subunits after limited proteolysis. The gamma subunit, however, is not simply a disulfide heterodimer of alpha and beta subunits, since this subunit was not reduced by 100 mM dithiothreitol plus 5% 2-mercaptoethanol, or even under more potent denaturing conditions, such as 8 M guanidine-HCL and mercaptoethanol at pH 10.5. In nonreduced gels, free insulin receptor subunits are observed, as well as two higher molecular weight bands of Mr = 520,000 and 350,000. On reduction, the 520,000 band was composed primarily of Mr = 210,000 and 95,000 subunits, whereas the 350,000 band was composed primarily of Mr = 135,000 and 95,000 subunits. These data suggest that the two major subunits of the insulin receptor (alpha and beta) are distinct. In addition, there is a third component of the receptor identifiable of 210,000 which may be a proreceptor or some closely associated effector protein. Furthermore, it appears that in the native state several kinds of disulfide oligomers of these subunits exist. These findings suggest a complex model for insulin receptor synthesis and insertion into the membrane.

Rat hepatoma cells were labeled with [32P]orthophosphate and the insulin receptor subunits were identified by immunoprecipitation and sodium dodecyl sulfate-acrylamide gel electrophoresis. In the basal state, only the Mr = 95,000 (beta) subunit of the insulin receptor was phosphorylated. The covalent labeling with 32P of this subunit was stimulated about 3-fold by insulin (10(-6) M). This stimulation was due to an increase in the content of phosphoserine, the appearance of phosphotyrosine, and a possible increase in phosphothreonine as well. These results suggest phosphorylation of the insulin receptor at multiple sites is an early event in insulin action.

An analogue of porcine insulin which differs from the native molecule in that the amino-acid residue B22-L-arginine is replaced by its D-enantiomer has been synthesized. The [D ArgB22]B-chain was synthesized by the segment condensation method and purified as the di-S-sulfonate by ion exchange chromatoggraphy on SP-Sephadex at pH 3.5. Combination with native porcine sulfhydryl A-chain gave [DArgB22]insulin which was purified by ion exchange chromatography on SP-Sephadex at pH 4.5 with a linear NaCl gradient. The biological activity of this analogue as measured by glucose oxidation in rat epididymal adipocytes was 2%. Thymidine incorporation into DNA of human fibroblast was 16%. The immunoreactivity using antipork insulin antibody in a double antibody immunoassay was 4%. The receptor-binding affinity as measured by radioreceptor assays was 2% with cultured human fibroblasts and 1% with rat adipocytes. These results suggest that the L-configuration at B22-arginine is essential for retaining the biological, immunological and receptor-binding properties of the hormone.

In the present study, we attempted to define possible subpopulations of antibodies which theoretically could be directed against evolutionarily conserved regions of the insulin molecule in sera from insulin-treated diabetic patients using a variety of labelled and unlabelled insulins which differ widely in structure but are very similar in functional properties. Ten high titre human insulin antisera from patients treated with mixed beef-pork insulin were examined. All sera were found to bind 125I-pork insulin better than labelled chicken insulin which bound better than labelled fish insulin. Detailed studies were conducted with four of the antisera using the pork and fish tracers. With two of the antisera, a subpopulation of antibody could be detected with 125I-fish insulin which had similar affinity for both fish and pork insulin, but reacted much less well with guinea pig insulin and the desoctapeptide derivative of porcine insulin. Based on the known properties of these four insulins, the data provide suggestive evidence consistent with the hypothesis that there are subpopulations of antibodies recognizing regions on the insulin molecule that are well conserved, possibly the region involved in the formation of insulin dimers or receptor binding.