Changes in glycosylation alter the affinity of the human transferrin receptor for its ligand

When transferrin receptors of human erythroleukemic cells were pulse-labeled with [35S]methionine and then chased in the absence of radioactive precursor, the first detectable immunoprecipitable form of the receptor had a molecular mass of 85 kDa. This form of the receptor was converted to the mature form of 93 kDa with a half-time of about 40-60 min. Both the immature (85 kDa) and mature (93 kDa) receptors associated as dimers, the native form of the receptor. The 85-kDa, as well as the 93-kDa, receptors bound to a monoclonal antibody raised against the transferrin receptor or to transferrin-Sepharose. In order to determine whether glycosylation was necessary for ligand binding, purified receptors were isolated from cells grown in the presence of tunicamycin. When K562 cells were grown in the presence of tunicamycin, an 80-kDa nonglycosylated form of the receptor was synthesized. This nonglycosylated receptor was also capable of dimer formation; however, much less of it reached the cell surface than the fully glycosylated form, although both untreated and tunicamycin-grown cells appeared to synthesize transferrin receptors at similar rates. Although the number of receptor molecules/cell was similar in control and tunicamycin-treated cells, the nonglycosylated receptors exhibited a much lower affinity for transferrin than those of untreated cells; in contrast, when receptors were purified by immunoprecipitation and digested with bacterial alkaline phosphatase, no difference was observed between the affinity of these receptors and undigested immunoprecipitated receptors. These results suggest that glycosylation is not necessary for specific binding of transferrin to its receptor, but the affinity of this binding can be influenced greatly by the presence or absence of carbohydrate residues.