Degree Type

Dissertation

Date of Award

2019

Degree Name

Doctor of Philosophy

Department

Biochemistry, Biophysics and Molecular Biology

Major

Bioinformatics and Computational Biology

First Advisor

Robert . Jernigan

Second Advisor

Guang . Song

Abstract

It is widely accepted that the structure of a protein and its motions are critical for a protein’s function, and that protein functions are usually accompanied by highly specific conformational changes. However, in many cases it is still unclear how the details of motion relate to a protein’s functionality and especially what causes conformational changes, despite having a significant number of proteins with multiple experimentally determined conformations. Here we investigate the conformational changes in proteins by collecting ensembles of different conformations of the same protein structure and simulate the application of external forces originating from exothermic chemical reactions such as ATP hydrolysis or the forces arising upon physical impact of ligands when they bind. External forces are applied to a structure in a novel approach of introducing directed forces at single residues as well as the Metropolis Monte Carlo simulation where more randomness is introduced. Both of these types of simulations are conducted within the framework of elastic network models.

By applying single iterative forces to single residues, our approach shows that the forces able to drive the conformation to the known final structure are usually highly directional in nature. Our simulations also reveal that external forces can push a conformation to the known target form by pushing on only a few residues, and that such residues are sequentially conserved, indicating their functional importance. During the Metropolis Monte Carlo simulations, we observe the that forces enable a protein structure to overcome energy barriers in moving towards the known final form, and for all structures studied, the final state reached is within less than 3.8Å from the known final conformation, in terms of root-mean-square-deviation, and usually substantially closer. We also generate energy landscapes to investigate conformational transition pathways. The landscapes are generated by computing the free energies interpolated from known experimental structures and extracting the dominant motions in terms of their principal component. The generated energy landscapes agree with the concept that native structures usually fall within low energy basins. We project the conformational transition pathways generated by Metropolis Monte Carlo simulations and observe that the pathways generally follow low energy pathways and are overall energetically favorable.

Copyright Owner

Yuan Wang

Language

en

File Format

application/pdf

File Size

157 pages

Share

COinS