Multiple amino acids within I-EGF1 can contribute to the binding of type II anti–HPA-1a antibodies. (A) Comparison of human vs murine PSI and I-EGF1 domain sequences, with differences highlighted in red. Note especially the APLD sequences in the PSI domain and the Q470M, H446P, G463D, and P464Q differences within EGF1. (B) Structural model of the variable region of antibody B2G1 bound with the β3 PSI and I-EGF1 domains. The antibody is shown as a tan surface with the CDR loops indicated, and the side chains of integrin β3 residues at the antigen-antibody interface are shown as sticks and dots. Note that interface interacting residues include not only polymorphic amino acid 33 but also P32 in the PSI domain and H₄₄₆ and Q₄₇₀ of I-EGF1. Also note that G₄₆₃ and P₄₆₄ are nowhere near the interface. (C) Top: HEK 293 cells transiently transfected with plasmids expressing human GPIIb and a muGPIIIa isoform that had been mutated to express the indicated humanized amino acid substitution were incubated with the indicated antibodies and subjected to flow cytometric analysis. The PSI domain–specific mAb PSIB1 was used as a control for transfection efficiency. Note that mAb 26.4 requires Q₄₇₀ for its binding, whereas B2G1 requires both Q₄₇₀ and H₄₄₆, as predicted in the docking model in panel B. Bottom: HEK 293 cells transfected with plasmids expressing human GPIIb and a human GPIIIa isoform that had been mutated to express the indicated mouse amino acids were subjected to flow cytometric analysis using the indicated antibodies. Note that the Q₄₇₀→M mutation results in loss of binding of both 26.4 and B2G1, whereas the H₄₄₆→P amino acid substitution affects only B2G1.