Accurate Prediction of Drug Activity by Computational Methods: Importance of Thermal Capacity.

Heat capacity is one of the most important thermodynamic quantities in protein biochemistry. Upon the binding of small molecules, a change in the heat capacity of proteins is generally observed, and this is often used in drug discovery. However, few computational works dedicated to the study of these phenomena are available in the literature. Here, a simple computational method for determining the change in heat capacity upon the binding of small ligands has been evaluated. The method is based on the accurate calibration of the solvent's thermal properties in the simulation conditions used in order to simply subtract its contribution to calculate the variations in the heat capacity of the system of interest. Using HIV protease as a model system, for which numerous experimental thermodynamic data are available, estimates of the change in heat capacity upon binding were obtained, which were similar to those observed experimentally. Furthermore, the predicted variations in heat capacity appear to be able to discriminate between molecules that behave as effective inhibitors of the enzyme and molecules that are able to bind the enzyme but not inhibit it. The results obtained suggest that this computational approach could be a useful aid in the in silico screening of new ligands for targets of interest.