Energy embedding of trypsin inhibitor.

Abstract

AbstractEnergy embedding has been shown recently to be a useful extension of the distance geometry approach to conformational calculations in the case of very small molecules and simple energy functions. This paper tests the ability of energy embedding to locate low energy conformations satisfying both weak and strong geometric constraints when the molecule is the small protein, bovine pancreatic trypsin inhibitor, and the energy function is the complicated Oobatake‐Crippen residue–residue potential. Using the potential function alone, the algorithm reaches a structure with energy lower than that of the native conformation, but with little resemblance to it. Aided by numerous geometric constraints, such as preformed secondary structure segments, the algorithm again finds a local minimum with energy better than that of the native, and with only 3.3 Å rms deviation from it. This is significantly closer to the native value than can be obtained using standard distance geometry and the geometric constraints alone. Thus, energy embedding using the Oobatake‐Crippen potential function is a significant help in finding native conformations of proteins. However, additional trials on a hairpin bend fragment of trypsin inhibitor demonstrate the potential's shortcomings in encouraging proper secondary structure.