Modeling RNA, protein, and synthetic molecules using coarse-grained and all-atom representations

Date

2016-12-05

Authors

Bell, David Russell

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

The aim of computational chemistry is to depict and understand the dynamics and interactions of molecular systems. In addition to increased comprehension in the physical and life sciences, this insight yields important applications to therapeutic design and materials science. In computational chemistry, molecules can be modeled in a number of representations depending on the molecular system and phenomena of interest. In this work, both simplified, coarse-grained representations and all-atom representations are used to model the interactions of RNA, cucurbituril host-guest chemistry, and cadmium selenide quantum dot binding to the Src homology 3 domain. For RNA, a coarse-grained model was developed termed RACER (RnA CoarsE-gRained) to accurately predict RNA structure and folding free energy. After optimization to statistical potentials, RACER accurately predicted the structures of 14 RNAs with an average 4.15Å root mean square deviation (RMSD) to the experimental structure. Further, RACER captured the sequence-specific variation in folding free energy for a set of 6 RNA hairpins and 5 RNA duplexes, with a R² correlation of 0.96 to experiment. The binding free energies of a cucurbituril host with 14 guests were computed using a polarizable force field and the free energy techniques of Bennett acceptance ratio and the orthogonal space random walk. The polarizable force field captured binding accurately, yet unexpectedly, the orthogonal space random walk method converged slowly, albeit at still reduced computational expense to the Bennett acceptance ratio. Lastly, the nanotoxicity effects of trioctylphosphine oxide coated cadmium selenide quantum dots are investigated with the model Src homology 3 protein domain in complex with its native proline rich motif ligand. With increasing quantum dot concentration, there is an increasing preference for the quantum dots to bind to the proline rich motif active site, inhibiting Src homology 3 function.

Description

LCSH Subject Headings

Citation