In order to understand protein-protein interactions, researchers face the challenge of parsing out the actual physical interfaces at which they occur — and not all proteins behave according to the rules. Jeffrey Skolnick, director of the Center for the Study of Systems Biology at the Georgia Institute of Technology, and his colleagues are studying "promiscuous" proteins — those that misbehave and bind to many other proteins in addition to their intended targets. In what they say is the first systematic study of the nature of protein-protein interfaces, Skolnick and his team have uncovered more than 1,000 structurally distinct protein-protein interactions that hinge on the underlying physics of the proteins themselves. Their work was published in PNAS in December.

Skolnick says that a number of requirements had to be satisfied before his team could even try to tackle a study of this scope. "Most importantly, there had to be enough quaternary structures in the Protein Data Bank to make such a comparison of interfaces meaningful. A fast and sensitive interface alignment algorithm needed to be developed so that one could compare all the interfaces in the PDB," Skolnick says. In addition, "a library of computer-generated homopolypeptide structures was required so that we could separate out features that are intrinsic to protein structures from those that are generated by evolution."

To study the structural similarity of protein-protein interfaces involving dimers, Skolnick developed iAlign, a computational method that classifies interfaces known to exist among native proteins. What he observed was that even in cases without structural similarity between monomers that form dimeric complexes, almost 90 percent of the interfaces had a close structural neighbor — a finding that could have implications for drugs designed to affect protein-protein interactions.

In the future, Skolnick and his team will focus their attention toward applying these ideas to better predict protein-protein interactions in proteomes. "We are also interested in exploring the conservation of hot spots in interfaces of proteins with similar interfaces, but different tertiary structures," he adds. "These ideas will be adapted to test out some new approaches to inhibiting protein-protein interactions."