A Detailed Comparison Between The NMR Ensemble, Two X-ray Models And Computational Predictions Of Motions For A Designed Sugar Binding Protein

Abstract

Coarse-grained elastic network models with single point representation of amino acids are becoming increasingly popular for describing conformational flexibility and equilibrium dynamics of proteins. In particular, the Gaussian network model (GNM) predictions have been fairly successful in interpreting the residue-level root-mean-square variations in residue positions inferred from NMR ensembles of structural models for a given protein and the fluctuations in residue positions indicated by crystallographic B-factors. Here, we carried out a detailed analysis for a designed sugar binding protein whose structure was solved in two crystal forms by X-ray crystallography and by NMR. Comparison with experimental data and results from molecular dynamics simulations confirm that the GNM predicts well the equilibrium dynamics of this protein and correlates better with the NMR derived data than crystallographic B-factors. The results further stipulate the importance of examining multiple structures determined by different methods as well as performing both analytical and numerical studies, toward gaining an accurate understanding of the type and range of conformational motions accessible to a given protein under native state conditions.