Development of Methods for the Investigation of RNA-Ligand Interactions.

[摘要] Three critical features of RNA make it a unique challenge for drug discovery: a) it is highly negatively charged, increasing non-specific binding, b) it can be highly dynamic, adopting different conformations upon binding varying ligands, and c) it has solvent exposed shallow binding pockets. All these properties represent distinct problems in the advancement of RNA-drug discovery. To address this first problem, MATCH was developed to rapidly, accurately, and universally parameterize small molecules for docking. MATCH accomplishes this by deconstructing a force field into a set of fundamental rules which best replicates existing parameters and permits extension to new molecules. MATCH is not only necessary to study RNA-ligand interactions en masse but will also contribute to understanding the charge-charge consequences of ligand binding. To address RNA flexibility, a method to combine NMR chemical shifts and Molecular Dynamics (MD) was developed to generate dynamic ensembles. To benchmark this technique, a set of 26 RNA structures with experimentally determined chemical shift was selected. An ensemble of structures was optimized to match the chemical shifts of each system. These ensembles were also shown to be consistent with of NMR NOE and RDCs constraints. To further demonstrate the utility of this method a large pool of structures (~350,000) was used to generate an ensemble for a prominent RNA target – the ribosomal decoding site. The conformations within this ensemble were found on favorable areas of the free energy landscape, independently indicating the validity of these structures. Finally to address the solvent exposed binding pocket of RNA and its flexible ligands, a new docking approach for RNA was developed, which performs an enhanced sampling technique by fragmenting the ligand and independently optimizing the conformation of each fragment. To properly benchmark this novel algorithm, a large set of 230 nucleic acid-ligand complexes was compiled. Utilizing this large set of this enhanced sampling technique was compared to ICM – a leading docking program. ICM produced native-like conformations 45% of the time, while our approach yields native-like conformations 55% of the time. Demonstrating the effectiveness of this novel sampling procedure.

[发布机构] University of Michigan

[效力级别] RNA

[关键词] Drug Discovery;RNA;Ligand Parameterization;Chemical Shifts;Computational Docking;Chemical Engineering;Science;Biophysics