Computational methods to design biophysical experiments for the study of protein dynamics

"In recent years, new software and automated instruments have enabled us to imagine autonomous or ""self-driving"" laboratories of the future. However, ways to design new scientific studies remain unexplored due to challenges such as minimizing associated time, labor, and expense of sample preparation and data acquisition. In the field of protein biophysics, computational simulations such as molecular dynamics and spectroscopy-based experiments such as double electron-electron resonance and Fluorescence resonance energy transfer techniques have emerged as critical experimental tools to capture protein dynamic behavior, a change in protein structure as a function of time which is important for their cellular functions. These techniques can lead to the characterization of key protein conformations and can capture protein motions over a diverse range of timescales. This work addresses the problem of the choice of probe positions in a protein, which residue-pairs should experimentalists choose for spectroscopy experiments. For this purpose, molecular dynamics simulations and Markov state models of protein conformational dynamics are utilized to rank sets of labeled residue-pairs in terms of their ability to capture the conformational dynamics of the protein. The applications of our experimental study design methodology called OptimalProbes on different types of proteins and experimental techniques are examined. In order to utilize this method for a previously uncharacterized protein, atomistic molecular dynamics simulations are performed to study a bacterial di/tri-peptide transporter a typical representative of the Major Facilitator Superfamily of membrane proteins. This was followed by ideal double electron-electron resonance experimental choice predictions based on the simulation data. The predicted choices are superior to the residue-pair choices made by experimentalists which failed to capture the slowest dynamical processes in the conformational ensemble obtained from our long timescale simulations. For molecular dynamics simulations based design of experimental studies to succeed both ensembles need to be comparable. Since this has not been the case for double electron-electron resonance distance distributions and molecular simulations, we explore possible reasons that can lead to mismatches between experiments and simulations in order to reconcile simulated ensembles with experimentally obtained distance traces. This work is one of the first studies towards integrating spectroscopy experiment design into a computational method systematically based on molecular simulations."

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Biophysics and Quantitative Biology

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2020

Author dc:creator

• Mittal, Shriyaa

Contributors dc:contributor

• Shukla, Diwakar
• Gruebele, Martin
• Luthey-Schulten, Zaida
• Sinha, Saurabh

• molecular dynamics
• computational tools
• spectroscopy

Statement dc:rights

• Copyright 2020 Shriyaa Mittal

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/108420

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/108420