Influence of glycosylation on allelic and cell-specific M(r) variation, receptor processing, and ligand binding of the human complement C3b/C4b receptor

The human complement C3b/C4b receptor (CR1), a single chain membrane glycoprotein of M(r) ~ 200,000, has several cell-specific M(r) variations as well as an allelic variation involving four phenotypes whose M(r) values span a range of 90,000. We investigated the role of glycosylation in these structural variations and in receptor metabolism. In the human HL-60 promyelocytic leukemic cell line differentiated toward granulocytes or monocytes and in Epstein-Barr virus-transformed B lymphoblastoid cell lines of the common phenotype, CR1 is synthesized as a precursor of M(r) = 222,000 that is converted into mature CR1s with differing M(r) values. Endoglycosidase F treatment of mature CR1 from all these cell types produced the same lower M(r) band. Additionally, the previously noted 5,000 higher M(r) of CR1 from human peripheral granulocytes versus erythrocytes was abrogated by endoglycosidase F. Hence these cell-specific variations are due to differences in N-glycosylation. Lectin affinity chromatography shows that these N-linked oligosaccharides are mostly tri- and tetraantennary complex-type species with specific differences in fractionation patterns that correlate with their differing M(r)s. In lymphoblastoid cell lines, the four allelic variants each have a precursor 6,000 lower in M(r) than the respective mature CR1. In the presence of tunicamycin, each of the CR1 allelic products is 25,000 lower in M(r) than the glycosylated receptor. The failure to radiolabel CR1 with [³H]glucosamine in the presence of tunicamycin indicates the lack of O-linked oligosaccharide on CR1. These data, taken together, strongly suggest that the CR1 polymorphism resides at the polypeptide level. Nonglycosylated CR1 synthesized in the presence of tunicamycin had twice the turnover rate of glycosylated CR1. The efficiency of surface membrane insertion and of ligand binding to hemolytically inactive C3 were markedly reduced for nonglycosylated CR1, suggesting that glycosylation is important for the proper expression of CR1 function.