Given the remarkable advance in computational power over the past decade, why has molecular simulation not had a more significant impact on the drug discovery process? While there are certainly noteworthy successes, the impact of virtual screening is limited by approximate treatment of ligand-protein interactions along two orthogonal dimensions: (1) Incorporation of backbone flexibility of the receptor, and (2) The accuracy with which molecular interactions are computed at the atomic level.
The first Aim seeks a solution to issue (1) by proposing a novel approach to virtual screening by targeting ensembles of receptor conformations, as sampled in native like environments during microsecond timescale simulation. In contrast to previous efforts, the present proposal suggests a computationally expedient solution to the problem of estimating the entropy of binding.
The second Aim seeks a solution to problem (2) by developing a new class of intermolecular potential based on recent advances in the quantum mechanical treatment of weak nonbonded interactions. Previously published results indicate at least a factor of five improvement in accuracy over standard empirical approaches. By bringing these advances to the field of protein-ligand interactions, dramatic improvement in the accuracy of these calculations is expected.
Both Aims will be pursued in the context of the A2A adenosine receptor, a member of the G-protein coupled receptor family and a target for several disorders of the central nervous system, including Parkinson's disease. Hits identified from small molecule libraries will be experimentally validated via a collaboration with a lab with extensive expertise in A2A biochemistry. We will also apply our methods to the optimization of a series of androgen receptor antagonists developed at UD, with the long term goal of treating prostate cancer. Overall, success in either Aim will have a profound and widespread, positive impact on the predictive validity of calculations of small molecule-protein interactions. This will in turn improve the value of hits identified in virtual screens, and help to realize the predictive promise of virtual screening.
This project seeks to significantly advance the field of virtual screening for drug discovery, a potentially powerful tool in the search for new therapies for known molecular targets. The proposed methods will be developed in the context of two protein targets: One that is under investigation for treatment of Parkinson's and Huntington's diseases, and one for the treatment of prostate cancer.
Showing the most recent 10 out of 93 publications
