Identification and Characterization of Intrinsically Disordered Protein Regions

Graduation Year

2024

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Chemistry

Major Professor

Vladimir Uversky, Ph.D., Sc.D.

Co-Major Professor

Wayne Guida, Ph.D.

Committee Member

Sameer Varma, Ph.D.

Committee Member

Arjan van der Vaart, Ph.D.

Committee Member

H. Lee Woodcock, Ph.D.

Keywords

protein intrinsic disorder, molecular recognition features, coupled folding and binding, structural bioinformatics, molecular dynamics, machine learning

Abstract

This dissertation investigates protein intrinsic disorder and intrinsically disordered protein regions (IDPRs) through the development and application of advanced computational and experimental techniques. Chapter 1 provides an introduction to protein intrinsic disorder, outlining the historical context and fundamental concepts that highlight the importance of intrinsically disordered proteins (IDPs) and IDPRs in various biological processes. Chapter 2 focuses on the rapid prediction and analysis of protein intrinsic disorder. We introduce RIDAO (Rapid Intrinsic Disorder Analysis Online), a high-efficiency web-based tool that integrates multiple disorder predictors. RIDAO significantly outperforms existing predictors in computational efficiency, making it suitable for large-scale proteomic studies. We highlight its potential for genome-scale structural bioinformatics and comparative genomics using a test set of over one million sequences. Chapter 3 investigates the coupled folding and binding of an IDPR contained in the tumor suppressor p53’s transactivation domain. Using isothermal titration calorimetry, molecular dynamics (MD) simulations, and nuclear magnetic resonance spectroscopy, we elucidate the binding modes and conformational changes of p53 fragments upon interaction with two of its negative regulators: MDM2 and MDMX. Our results reveal distinct differences in the binding thermodynamics and structural adaptations of these complexes, providing insights into the molecular mechanisms underlying the regulation of p53 by its negative regulators and paving the way for the development of new therapeutics. Finally, in Chapter 4, we explore propagating MD trajectories using machine-learned velocity updates to address the computational limitations of ab initio MD (AIMD). AIMD provides accurate simulations but its application to IDPRs is hindered by high computational costs and challenges in modeling large systems and biologically relevant timescales. To address these challenges, we explore using neural networks to predict velocity updates based on a particle’s historical velocities. Using GROMACS, we test this method on ten classical isolated harmonic oscillators and demonstrate its high accuracy in predicting velocity updates while conserving essential physical properties. This in vacuo proof-of-concept provides a foundation on which the method may be extended to coupled harmonic oscillators and the condensed phase. This dissertation advances our understanding of IDPs and IDPRs through the development of innovative computational tools and application of detailed experimental analyses combined with MD simulations. The findings provide valuable insights into the functional roles of protein intrinsic disorder and offer new methodologies for the identification and characterization of IDPRs.

This document is currently not available here.

Share

COinS