The platelet fibrinogen receptor α IIbβ 3 exists in an equilibrium between inactive and active conformations. In its inactive conformation, the transmembrane (TM) domains of α IIb and β 3 interact, but they separate when α IIbβ 3 assumes its active conformation. Peptides corresponding to the α IIb TM domain form homodimers in vitro and in bacterial membranes and the interface that mediates this interaction overlaps with the interface that mediates the heteromeric association of the α IIb TM domain with that of β 3. Because the homomeric association of α IIb TM domain is relatively strong, we expected that peptides spanning the α IIb TM domain might activate α IIbβ 3 by binding to the α IIb TM domain, thereby disrupting the α IIb/β 3 TM domain heterodimer. We synthesized a 26 residue peptide corresponding to the α IIb TM domain, flanked by pairs of lysine residues to increase its solubility in water, and assessed its ability to activate washed or gel-filtered platelets in an aggregometer. Addition of 3 μM peptide induced platelet aggregation after a short lag, but unlike aggregation stimulated by ADP, peptide-induced aggregation was not preceded by platelet shape change. Nonetheless, like ADP-induced aggregation, peptide-induced aggregation was inhibited by EDTA and the α IIbβ 3-specific monoclonal antibody A2A9. On the other hand, pre-incubating platelets with PGE1 or apyrase caused only a small decrease in the rate and extent of peptide-induced aggregation, suggesting that the peptide induced aggregation by binding directly to α IIbβ 3. A peptide corresponding the β 3 TM domain behaved in an identical manner, whereas an unrelated peptide, the model TM domain MS-1, had no effect. Similarly, there was little response to an α IIb TM peptide in which glycines 972 and 976, the first and last residues of a GxxxG motif, were changed to leucine. Confirming that the α IIb TM peptide binds to α IIbβ 3, we found that a FITC-labeled α IIb TM peptide co-eluted with purified α IIbβ 3 during size exclusion gel filtration. To determine whether the latter interaction involved the α IIb TM domain, we used the TOXCAT assay. In TOXCAT, the α IIb TM domain-mediated dimerization of a fusion protein containing the α IIb TM domain interposed between maltose binding protein and the ToxR' transcriptional activation domain in the inner membrane of E. coli drives the activation of the chloramphenicol acetyl transferase (CAT) gene. We found that adding the α IIb TM peptide to the bacterial growth medium consistently decreased CAT synthesis, implying that its presence impaired α IIb-mediated fusion protein dimerization. These results provide evidence for the presence of an α IIb/β 3 TM domain heterodimer in unstimulated human platelets and suggest that the GxxxG motif in the α IIb TM domain is involved in its formation. Because the homomeric interaction of α IIb and β 3 TM domains is substantially stronger than their heteromeric interaction, our data are also consistent with the hypothesis that the TM peptides disrupt the heterodimer which functions to maintain α IIbβ 3 in an inactive state. Lastly, the data suggest that stabilizing the α IIb/β 3 TM heterodimer may represent a new approach for the development of novel anti-platelet agents.