Multiscale Simulation of Nucleic Acid Nanostructures

Abstract

Nucleic acids, namely DNA and RNA, are arguably the most studied biological molecules. Utilizing the key properties of nucleic acids, such as their persistence lengths and Watson-Crick base-pairing complementarity, nanometer-scale devices and other functional materials can be created. Numerous experimental studies have been performed on these nanostructures both in vitro as well as in vivo. However, in many cases, it is hard to decipher their in-situ structures, interpret molecular level understanding of their self-assembly and understand many experimental observations, particularly those that originate from structural changes at the single-molecule level. In this thesis, using multiscale molecular dynamics simulations, we have studied the structure, mechanics, and thermodynamics of various self-assembled and artificial nucleic acid nanostructures. In particular, we have developed a de novo computational framework to model atomistic as well as coarse-grained nucleic acid nanotubes. Using various theoretical models, we have estimated their microscopic structure, mechanical properties and found excellent agreement with available values from the experiments. We have also studied a minimal bead-spring coarse-grained model of DNA nanostars which self-assemble into complex polymeric network known as DNA hydrogel at low temperature. The phase transition from an unstructured fluid to a gel-like network structure with the lowering of temperature has been explained by calculations of structural parameters, thermodynamic quantities and dynamics of the systems. Finally, the microscopic origin of liquid crystal ordering of ultra-short nucleic acids that do not satisfy the shape anisotropy criterion of Onsager's cylindrical rods is studied. Employing advanced free-energy calculation techniques, we observe that ultra-short nucleic acids prefer to stack on top of each other while repelling sideways, leading to the formation of supramolecular columns that undergo LC ordering at high volume fraction. The study presented in this thesis has showcased the ability of multiscale molecular simulation and statistical physics to predict the structural, mechanical, and thermodynamic properties of various self-assembled and engineered DNA nanostructures.