Geometric Similarities of Protein–Protein Interfaces at Atomic Resolution Are Only Observed within Homologous Families: An Exhaustive Structural Classification Study

To elucidate the structural basis of the diversity and universality in protein–protein interactions, an exhaustive all-against-all structural comparison of all known protein interfaces in the Protein Data Bank was performed at atomic resolution. After similar interfaces were clustered, approximately 20,000 structural motifs with at least two members were identified, out of which 3678 motifs consisted of at least 10 interfaces. Except for some trivial interfaces involving single α helices, almost all motifs were found to be confined within single protein families. Furthermore, the interaction partners of each motif were found to be very limited, and, accordingly, the interaction networks of the motifs tend to be small and are much more restricted than the binding sites for small ligand molecules. These findings suggest that, at the level of atomic structures, protein–protein interactions are precisely designed; hence, protein interfaces with multiple interacting partners should involve incompletely overlapping multiple interfaces and/or accommodate structural changes upon binding to their targets.