Characterization of a cDNA encoding the heparin and collagen binding domains of human thrombospondin.

Thrombospondin (TSP) cDNA clones were isolated from a human fibroblast cDNA library by screening with degenerate synthetic oligodeoxynucleotides. The probes were designed based on the sequences of peptides isolated from tryptic digests of TSP. The largest clone identified contains an open reading frame that encodes the amino-terminal heparin binding domain of TSP and part of the immediately adjacent collagen binding domain, a region of TSP that includes the site of disulfide crosslinking responsible for trimer formation. In addition to the known amino-terminal sequence of mature TSP (which is identical to that of the heparin binding domain) and that of the collagen binding domain, the open reading frame predicts the amino acid sequences of four tryptic peptides including the one from which the probe sequences were derived. The sequence of the mature protein is preceded by an unusual signal peptide sequence of 18 residues. The heparin binding domain contains 240 amino acids including only one intradomain disulfide bond. In contrast, 80 residues beyond the start of the collagen binding domain, a cluster of 10 cysteine residues occurs within 40 amino acids corresponding to the point of interchain disulfide crosslinking in the TSP trimer. RNA blot analysis indicates a TSP mRNA size of approximately equal to 8 kilobases and Southern blots are consistent with a single gene encoding TSP.