From Protein Structure Predictions to Dynamics

点击翻译按钮获取中文摘要

Professor Matthias Heyden at Arizona State University is supported by an award from the Chemical Theory, Models and Computational Methods (CTMC) Program in the Chemistry Division for the development of new methods to characterize and predict collective dynamics in proteins. The goal is to include dynamics and its effect on protein function in structure-based protein design processes and eliminate current limitations in the design of efficient enzyme catalysts. This promises to enable the design of tailored enzymes that catalyze novel reactions and replace energy and cost intensive processes in the synthesis of valuable compounds. The key development will be a series of new algorithms that capture anharmonic low-frequency motions of proteins in all-atom computer simulations. The computational nature of this project is further used to developed research opportunities for online undergraduate students that can be carried out remotely.Protein structure has long been regarded as the key determinant of protein function, for example, the ability of enzymes to catalyze a specific chemical reaction. However, an increasing body of evidence shows that structure alone is insufficient to explain the catalytic efficiency of many enzymes, whose activity instead relies on the co-existence of multiple distinct conformations and the transitions between them. While the accuracy of sequence-based protein structure predictions has seen dramatic recent improvements and strategies exist to tackle the inverse problem of designing amino acid sequences that fold into pre-selected structural motifs, our current ability to predict the functional dynamics of protein is limited even once the structure is known. All-atom molecular dynamics simulations provide a model with sufficient microscopic detail, but the computational costs associated with the sampling of conformational fluctuations limits their application to a small number of systems. On the other hand, coarse-grained models and inelastic network models are computationally efficient but lack microscopic detail. Here, we propose a new set of methods to extract collective protein degrees of freedom susceptible to large-scale motion from fluctuations in all-atom molecular dynamics simulations on timescales that are accessible to high-throughput simulations.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.