| Abstract: | Methods for measuring the attachment of viable cells to immobilized peptides to identify sequences possessing cell binding activity are described. The 212-residue C-terminal cell binding domain (CBD) of human thrombospondin-1 (TS1 or TS) was thought to contain a site or sites for the attachment of a number of cell types based on inhibition of cell attachment to TS1 with monoclonal antibodies. The presence of such a site was confirmed by expression of this region of TS1 in Escherichia coli and demonstration of its cell binding activity independent of other known cell attachment sites in TS1. To further define active sequences within the CBD, a set of eight overlapping peptides that spanned the 212-residue domain was synthesized with standard techniques. Several of these long (30-mer) peptides were found to be insoluble in aqueous buffers; thus, they could not be tested as soluble inhibitors of the attachment of cells to TS1, a standard approach to identification of active cell binding sequences. Using chemically derivatized plates prepared by Costar Corp., we devised assays to test the direct attachment of cells to the peptides immobilized on plastic wells. We also found that the peptides were adsorbed to underivatized plastic when solubilized in 6 M guanidine hydrochloride. Even rather short peptides (5 residues) were efficiently immobilized with these methods. Peptides immobilized directly on plastic gave higher levels of cell attachment than the same peptides conjugated to a carrier protein and then immobilized. Using these methods, two of the 30-mers were found to support cell attachment. The active sequences within the 30-mers were then narrowed down to a hexamer in one case and a pentamer, both containing the tripeptide VVM. Shorter, soluble peptides containing these active sequences act synergistically to inhibit the attachment of cells to TS1 and its recombinant CBD. |
