Recent advances in our understanding of biomacromolecular binding processes (in drug design for example) has been aided and guided by computational methods. Computational docking techniques in particular utilize free energy functions and algorithms, to rapidly produce a ranking of likely binding sites based on their binding free energy. The objective of this work is to validate the predicted geometries of a computational docking method using a comprehensive set of fourty-four ligand-DNA physically bound bimolecular systems. The fourty-four ligands differ in their chemical properties (net charges and torsional degrees of freedom) as well as their physical binding modes to DNA (minor groove binding or intercalation). We utilized a physics-based free energy function optimized by a Lamarkian genetic algorithm (LGA) to predict the geometries of small molecules upon physical binding to DNA. The free energy function includes a van der Waals term, an electrostatic term, a hydrogen bond term, a desolvation term, and a torsional term. AutoDock 4.2 was used for the implementation of the free energy function and the LGA. We compared the geometry of each predicted binding site with the experimentally observed geometry for all fourty-four bimolecular systems. Our results show that for most ligands with ten or less torsional degrees of freedom one of the predicted binding sites is within 2.5 Angstroms root mean square deviation from the experimental geometry. The predicted binding site closest to the experimental geometry has a binding free energy within 2 kcal/mol of the lowest binding free energy site. Overall, our results show this docking method to be effective in predicting the physical binding sites of ligands, such as drugs and carcinogens, to DNA, for ligands with up to about ten torsional degrees of freedom.

Comprehensive validation of a computational docking methodology to identify the noncovalent binding site of a ligand to DNA

246th National Meeting (2013)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.