Superposition of protein structures using electrostatic isopotentials

Algorithms for comparing protein structures are widely used to identify proteins with similar functions and to examine the mechanisms of binding specificity. In order to make accurate comparisons, two structures must first be superposed, so that differences in position and orientation do not create misleading dissimilarities. Most algorithms generate these superpositions by aligning atoms of the peptide backbone. This approach is rapid, but it may not reflect similarities or differences in all mechanisms that proteins use to bind other molecules. Electric fields, for example, play a large role in recognition and their substantial range can interact with other molecules long before backbone contacts occur. To compare proteins based on their electric fields, we have developed the first algorithm designed to superpose protein structures using electric fields alone. Our method works by searching rotational and translational space for a superposition that maximizes the overlapping volume between electrostatic isopotentials. Applying this method to compare the serine protease and enolase superfamilies, our results demonstrate that our electrostatic superposition algorithm can distinguish very similar proteins with different binding preferences.