PAC1 is an IgM kappa murine monoclonal antibody that, like the Arg-Gly-Asp-containing ligand fibrinogen, binds to integrin alpha IIb beta 3 only on activated platelets. The unique binding properties of PAC1 may be determined by its large size, its multivalency, and by variable region sequences, including an Arg-Tyr-Asp at residues 100A-C in H-CDR3. To study the molecular determinants of PAC1 function, baculoviruses containing cloned cDNA for the Fd heavy and kappa light chains of PAC1 were used to co-infect Sf9 insect cells. Infected cells secreted a soluble, monovalent, 50-kDa Fab fragment that bound saturably to agonist-stimulated platelets but not to resting cells. Fab binding was inhibited > 85% by 10 mM EDTA, 1 mM RGDS, 1 mM fibrinogen gamma 397-411, or 12 microM fibrinogen, but not by 1 mM RGES. Compared to PAC1 IgM, a 60-fold higher molar concentration of PAC1 Fab was required for half-maximal binding to platelets or for half-maximal inhibition of fibrinogen binding. PAC1 Fab bound to an activated form of alpha IIb beta 3 expressed in Chinese hamster ovary cells, but not to the resting form of the receptor in these cells or to alpha v beta 3 in human endothelial cells. Conversion of Asp100C to Glu by site-directed mutagenesis rendered the antibody inactive, indicating that the Arg-Tyr-Asp sequence in H-CDR3 is essential for PAC1 recognition of alpha IIb beta 3. Binding of fibrinogen or PAC1 IgM to platelets induced tyrosine phosphorylation of a 140-kDa platelet protein, but binding of PAC1 Fab did not. These studies demonstrate that the specificity of PAC1 for activated alpha IIb beta 3 is determined by an integrin recognition sequence within H-CDR3. However, the strength of this binding interaction and the ability of PAC1 to trigger signaling in platelets also depend on antibody valency.