Measuring and Simulating Protein Electron Paramagnetic Resonance Spectroscopy

Abstract

Electron paramagnetic resonance (EPR) spectroscopy allows one to investigate local molecular structure and dynamics over a wide range of time scales. For proteins, site-directed spin labeling (SDSL) and EPR can reveal key aspects of the structural and dynamical hierarchies that describe their complex behavior. However, due to EPR's high sensitivity to many properties of the spin label and the surrounding environment, interpretation of its complex spectra can be a formidable task. As a result, modeling and simulation are essential to bridging the gap between EPR spectra and their interpretation. In this thesis, two investigations that use modeling and simulation for EPR are summarized: one in the context of experimentation and one in the context of computation. The experimental work in Ch. 3 uses SDSL and several EPR measurement techniques to investigate the interactions between two membrane proteins that are crucial for heart muscle function. Computational simulation and simple diffusion model-based global fitting of the EPR spectra are employed to discern a model for the effects of phosphorylation on these interactions. The computational work contained in Ch. 4 goes beyond simple diffusion models, performs all-atom molecular dynamics (MD) simulation of spin-labeled proteins, and creates a set of MATLAB programs to that can use MD simulation data to simulate EPR spectra. Additionally, since MD simulations are very time-intensive, two prominent approximate methods are compared, which attempt to coarse-grain the simulated spin label dynamics and reduce the required length of MD trajectories.