Maternal alloantibodies against the human platelet Ag (HPA)-1a allotype of the platelet beta(3) integrin GpIIb/IIIa can cause severe fetal or neonatal hemorrhage. Almost all anti-HPA-1a-immune mothers are homozygous for HPA-1b and carry HLA-DR52a (DRB3*0101). The single Pro(33) -->Leu substitution (HPA-1b-->HPA-1a) was previously predicted to create a binding motif for HLA-DR52a that can lead to alloimmunization. We have isolated six CD4(+) T cell clones from three such mothers, which all respond to intact HPA-1a(+), but not HPA-1b(+), platelets. We used them to define the "core" and "anchor" residues of this natural T cell epitope. Molecular modeling based on a recently published crystal structure can explain the preferential presentation of the Leu(33) (but not Pro(33) variant) by HLA-DR52a rather than the linked HLA-DR3 or the allelic DR52b. The modeling also predicts efficient anchoring at position 33 by several alternative hydrophobic alpha-amino acids; indeed, a recently identified variant with Val(33) is presented well to two clones, and is therefore potentially alloimmunogenic. Finally, these HPA-1a-specific T cell clones use a variety of T cell receptors, but all have a "Th1" (IFN-gamma-producing) profile and are suitable for testing selective immunotherapies that might be applicable in vivo.