Amarda Shehu, Molecules in Motion: Computing Structural Flexibility

The biological functionality of flexible molecules such as proteins often depends on their ability to change shape as needed to fit in molecular complexes. A thorough understanding of protein function requires characterizing the entire conformational space available to a protein at equilibrium (room temperature). Two factors make the search for conformations populated at equilibrium especially challenging: (i) the space that needs to be searched is high-dimensional; and (ii) desired conformations lie on the very basin of a rugged funnel-like energy landscape associated with protein conformational space.

This talk presents our efforts in developing a general framework to explore conformational space for low-energy conformations that sit on the energy basin. The framework exploits analogies between robot kinematic chains and protein polypeptide chains to efficiently sample conformational space. Topological constraints that characterize conformations of this basin are employed to narrow down the search to relevant lower-dimensional subspaces.

Two classes of developed methods will be presented. In the first one, knowledge of an average experimentally-determined protein structure is used to obtain other low-energy conformations in the vicinity. Conformations obtained in this fashion agree very well with available experimental data on equilibrium macroscopic properties of a protein. In the second, a priori information is reduced to a general topological feature such as cyclization. Results obtained on cyclic cysteine-rich polypeptides confirm the significant contribution of the developed method in obtaining a detailed conformational view of energy minima populated at equilibrium.