Stability of macromolecular complexes

Macromolecular interactions are crucial in numerous biologic processes, yet few general principles are available that establish firm expectations for the strength of these interactions or the expected contribution of specific forces. The simplest principle would be a monotonic increase in interactions as the size of the interface grows. The exact relationship might be linear or nonlinear depending on the nature of the forces involved. Simple “linear–free energy” relationships based on atomic properties have been well documented, for example, additivity for the interaction of small molecules with solvent, and, recently, have been explored for ligand–receptor interactions. Horton and Lewis propose such additivity based on buried surface area for protein–protein complexes. We investigated macromolecular interactions and found that the highest‐affinity complexes do not fulfill this simple expectation. Instead, binding free energies of the tightest macromolecular complexes are roughly constant, independent of interface size, with the notable exception of DNA duplexes. By comparing these results to an earlier study of protein–ligand interactions we find that: (1) The maximum affinity is approximately 1.5 kcal/mol per nonhydrogen atom or 120 cal/mol Å2 of buried surface area, comparable to results of our earlier work; (2) the lack of an increase in affinity with interface size is likely due to nonthermodynamic factors, such as functional and evolutionary constraints rather than some fundamental physical limitation. The implication of these results have some importance for molecular design because they suggest that: (1) The stability of any given complex can be increased significantly if desired; (2) small molecule inhibitors of macromolecular interactions are feasible; and (3) different functional classes of protein–protein complexes exhibit differences in maximal stability, perhaps in response to differing evolutionary pressures. These results are consistent with the widespread observation that proteins have not evolved to maximize thermodynamic stability, but are only marginally stable. Proteins 2002;48:645–653. © 2002 Wiley‐Liss, Inc.