The "Action-at-a-distance calculator" provides an easy interface for computing likely candidate mutations to harness action-at-a-distance electrostatic interactions.
Charged residues at the periphery of a protein–protein binding interface can make non-trivial contributions to the binding free energy, even when the residue remains fully exposed to solvent in the bound state, and at distances as far as 10 Å from the binding partner. We have termed these moderate-range, through-solvent electrostatic contributions to binding affinity "action-at-a-distance" interactions.
The greatest attraction of action-at-a-distance interaction is the location of the residues involved. Mutations to residues on the surface of a protein are likely to have comparatively small effects on protein stability. When the residue remains exposed to solvent even in the bound state, the details of the three-dimensional packing of amino-acid chains are also likely to play a minor role. As a result, relatively simple procedures can be used to identify likely candidate mutations to harness action-at-a-distance interactions.
Many design methods are based on a careful balance of opposing interactions. Accurate design of residues making intimate contact with the binding partner requires balancing favorable energetic contributions of electrostatic and dispersion forces between the binding partners with unfavorable contributions from side-chain entropy, loss of interactions with solvent, and possible steric clashes. The necessity to account for this balance can require a high-level of structural accuracy for successful application in design, and thus many approaches are poorly suited to applications where the resolution of the available structural data is lower (as may be the case for homology models and membrane proteins). As action-at-a-distance interactions take place over moderately long distances, they may be less dependent on local variations in structure and thus applicable even when only moderate-resolution structural data is available.