Isolation, cDNA cloning, and overexpression of a 33-kD cell surface glycoprotein that binds to the globular 'heads' of C1q

Abstract

This work describes the functional characterization, cDNA cloning, and expression of a novel cell surface protein. This protein designated gC1q-R, was first isolated from Raji cells and was found to bind to the globular 'heads' of C1q molecules, at physiological ionic strength, and also to inhibit complement-mediated lysis of sheep erythrocytes by human serum. The NH 2-terminal amino acid sequence of the first 24 residues of the C1q- binding protein was determined and this information allowed the synthesis of two degenerate polymerase chain reaction primers for use in the preparation of a probe in the screening of a B cell cDNA library. The cDNA isolated, using this probe, was found to encode a pre-pro protein of 282 residues. The NH 2 terminus of the protein isolated from Raji cells started at residue 74 of the predicted pre-pro sequence. The cDNA sequence shows that the purified protein has three potential N-glycosylation residues and is a highly charged, acidic molecule. Hence, its binding to C1q may be primarily but not exclusively due to ionic interactions. The 'mature' protein, corresponding to amino acid residues 74-282 of the predicted pre-pro sequence, was overexpressed in Escherichia coli and was purified to homogeneity. This recombinant protein was also able to inhibit the complement-mediated lysis of sheep erythrocytes by human serum and was shown to be a tetramer by gel filtration in nondissociating conditions. Northern blot and RT-PCR studies showed that the C1q-binding protein is expressed at high levels in Raji and Daudi cell lines, at moderate levels in U937, Molt-4, and HepG2 cell lines, and at a very low level in the HL60 cell line. However, it is not expressed in the K562 cell line. Comparison of gC1q-R NH 2-terminal sequence with that of the receptor for the collagen-like domain of C1q (cC1q-R) showed no similarity. Furthermore, antibodies to gC1q-R or an 18-amino acid residue- long NH 2-terminal synthetic gC1q-R peptide did not cross-react with antibodies to cC1q-R. Anti-gC1q-R immunoblotted a 33-kD Raji cell membrane protein, whereas anti cC1q-R recognized a molecule of ~60 kD. The NH 2- terminal sequence of gC1g-R appears to be displayed extracellularly since anti-gC1g-R peptide reacted with surface molecules on lymphocytes, polymorphonuclear leukocytes, and platelets, as assessed by flow cytometric and confocal laser scanning microscopic analyses. In addition, all or part of the gC1q binding domain may reside within the 24 amino acid stretch of the NH 2-terminal sequence of gC1q-R since the 18 amino acid residue long- synthetic peptide corresponding to this region inhibited serum C1q hemolytic activity. The data presented in this report suggest that there are at least two types of C1q-R which appear to be expressed on the same type of cells and these receptors individually or in concert may contribute to the diversity of C1q-mediated responses.