The DNA sequences encoding for two proteins of the cell-binding domain (CBD) of human fibronectin (FN), namely a 33-kDa protein (aa 1329-1722) and a 40-kDa (aa 1380-1851) protein, were cloned and expressed in Escherichia coli. The interactions of the resulting rCBD proteins, refolded and purified to homogeneity, with human platelets were studied in comparison with those of the pentapeptide GRGDS. The binding of both the 33-kDa and the 40-kDa proteins to washed platelets appeared to be dependent upon platelet activation. In the case of the 33-kDa protein, binding to stimulated platelets was shown to be saturable, with Kd = 2 microM (thrombin as agonist). Moreover, both the 33-kDa and the 40-kDa proteins inhibited fibrinogen binding (at 0.1 microM) to ADP- or thrombin-stimulated platelets with IC50 values in the same concentration range. Binding seemed therefore to occur mainly at the GPIIb/IIIa receptor, and accordingly monoclonal antibodies against this receptor prevented up to 85% of the binding of the 33-kDa protein to platelets. With most stimuli the 33-kDa and the 40-kDa proteins inhibited platelet aggregation at concentrations 15- to 25-fold lower than those required by GRGDS and, in the case of the 33-kDa protein, this was shown to occur in either platelet-rich plasma, washed platelets, or whole blood. The 33-kDa protein also inhibited platelet aggregation and thromboxane A2 (TXA2) generation on the subendothelial extracellular matrix, whereas the GRGDS peptide inhibited only matrix-induced platelet aggregation, but not TXA2 formation. Furthermore, the 33-kDa protein, which is derived from the human FN CBD, seemed to be highly selective, since it inhibited the aggregation of platelets from primates only, and not from other animals tested. Finally, the 33-kDa protein did not promote fibroblast cell attachment, as was observed for both whole FN and the 40-kDa protein, thus displaying a selectivity toward platelets. In conclusion, the unique properties of the 33-kDa protein, and, in particular, its special affinity directed only toward activated primate platelets, seem to hold a promising potential for the further development of an antithrombotic agent.