Network Analysis of Molecular Dynamics Sectors in the p53 Protein

Abstract

Target-specific drug design necessitates the manipulation of multifaceted networks of biomolecules. At the level of proteins, such networks are composed of amino acids, and can be affected by biological processes such as allostery, point mutations and ligand binding. Binding an effector to an allosteric site, for instance, may limit the active site's binding affinity, thus inhibiting the protein's function. The complexity of amino acid interactions makes difficult the process of predicting protein behavior and network dynamics in the face of such intricate biological processes. In this study, we show that a combination of molecular dynamics (MD) simulations, network analysis, and spectral decomposition of pairwise interaction matrices progresses the ability to pick out targets residues with characteristics beneficial to the design of therapeutics. We observed in silico that the p53 tumor suppressor protein contains a dense network of cohesive residues which may have the ability to transduce longrange allosteric signals. By engineering point mutations to residues within this cohesive network, we found that disturbing the network's connectivity impacted the overall function of p53. The network was outputted via an updated implementation MD sectors, which analyzes and spectrally decomposes correlated motions between residues within an MD trajectory. The application of MD sectors to p53 could aid in the of design drugs which interact directly with sector residues. MD sectors has the potential to successfully pick out allosteric drug targets in a wide variety of biological systems in an accurate and costeffective manner. VI Contents ACKNOWLEDGEMENTS .

The Isoforms of p53: A Computational Approach to Molecular Dynamics Simulations and Protein Folding

“…The functional form of a force field with four parameters: bonded interactions, valence angles, torsions, and non-bonded interactions. 22 The above equation is utilized in our MD simulations for us to calculate an atom's position at a specific time. We use the positions of all of these atoms in tandem to get a sense for the structure and conformation of the whole system.…”

“…Figure 2.2.1:The functional form of a force field with four parameters: bonded interactions, valence angles, torsions, and non-bonded interactions 22. …”

Interconnected Protein Networks: Insights Towards CRIB-Par6 Protein Allostery Through a Graph-Theoretic Analysis

“…In the Thayer lab, SCA-inspired methodologies such as MD-sectors (Fabry, 2020;Lakhani et al, 2020) and more recently MD-Energetic Network Decomposition (MD-END) analysis (Abramson, 2021;Han, Abramson, & Thayer, 2022) have been developed to explore protein residue networks relevant to allosteric dynamics through covariance-based analyses rooted in first-principal statistics. This is achieved through the utilization of molecular dynamics (MD) simulations of protein systems of interest as input which can estimate the timedependent physical and energetic interactions of proteins and other macromolecules at a scientifically viable level of physical precision.…”

“…Since both x and z deviated from their starting positions at the same time, the distance covariance between them would be positive whereas the distance covariance between y and x or y would be minimal since it did not tend to move with either over time. (Fabry, 2020).…”

“…Analysis of this motional covariance data between residues thus compose a network of residues displaying high conformational and dynamic coupling potentially significant for the transmission of conformationally-based signalling pathways within CRIB-Par6 and its binding dynamics. The MD-sector methodology employed in the present study is based on an updated protocol by Jonathan Fabry (Fabry, 2020) which enables much greater integration for running MD-Sector calculations between the high performance computational resources on the HPCC and local computers. Such updates include utilizing the full range of viable residues within the protein for calculation of MD-sectors, a VPiCA Ratio Analysis (VRA) test for determination of sector cohesiveness, and a streamlined approach towards performing the majority of calculations within singular Jupyter notebook accessed on the HPCC rather than a set of scripts and programs stored locally on a specific computer (Fabry, 2020).…”