Simulating Chemical Processes From Brownian Diffusion to Binding Thermodynamics

Molecular recognition is a fundamental part of chemical processes, especially those relevant to biology. It refers to the process by which two molecules diffuse and eventually bind with one another to form a complex. This can be broken down into to broad aspects: kinetics and thermodynamics. Kinetics refers to the motion of molecules and the rates of their reactions with each other. Thermodynamics refers to the transfers of energy that drive the reaction when molecules bind together. The work in this dissertation uses computational methods to study both aspects of molecular recognition in a range of systems, and it includes the application of existing methods and development of new tools for simulation and analysis.A strong focus is given to protein-ligand systems, in which the ligand is an inhibitory drug designed to shut down function of its target protein. The concept of developing drugs (inhibitors) that bind with and disrupt the activity of their targets is the basis for much of modern medicine, and has had incredible success. The development of more effective inhibitors is a constant challenge. The physical and chemical principles that predict an inhibitor’s effectiveness have a complex interplay, and an understanding of these principles is challenging and highly sought after. Two projects described here use molecular dynamics (MD) simulations to elucidate these principles through better understanding inhibitor binding thermodynamics. These techniques are applied to a host of inhibitors for the carcinogenic CDK8 protein and to inhibitors of a protease (PLpro) of the recent SARS-CoV2 virus. Novel inhibitors of PLpro are also developed and validated. The other work described herein studies molecular binding kinetics in both natural and engineered systems. Brownian dynamics simulation software is described which has been developed by the author and other group members. Its goal is to provide a robust tool with which researchers can study molecular recognition and association in a range of systems under varied conditions. This program, called GeomBD3, has been applied to study association kinetics and mechanisms in enzyme bioconjugates, protein-ligand systems, and nucleic acid biosensors. These studies are included in this dissertation, and we thus demonstrate that Brownian dynamics simulations can aid in rational bio/chemical engineering design efforts and supplement experimental analysis