Evaluating molecular simulations of protein dynamics using novel experimental data

Proteins are evolved molecular machines that carry out the essential chemical reactions necessary for life. Like machines designed by humans, proteins execute their functions through an orderly set of motions and fluctuations - their “reaction coordinate”. However, proteins are also marginally stable, with the expectation that functional dynamics are embedded in a small subspace of a high dimensional pattern of overall motions and configurational changes. One approach to studying intramolecular fluctuations is computational simulation of atomic trajectories using molecular dynamics. Recent advances in experimental protein dynamics now open up the ability to test, validate, and possibly improve the process of molecular dynamics (MD) simulations through direct comparisons between computational predictions and data. We present a comparative analysis between MD and experiment using data from X-ray diffraction studies reporting electric field-stimulated excited state motions. The analysis develops methods for simulating a protein crystal while carrying out rigorous evaluations of structural conformation ensembles, atomic strain, force field differences, and more. This work represents an opportunity for MD and initiates a path towards understanding the physics of protein function.