Mapping the population of protein conformational energy sub-states from NMR dipolar couplings.

The precision with which X-ray crystallography and nuclear magnetic resonance (NMR) have provided structural models of biologically active and inactive conformations of countless proteins belies an easily overlooked dilemma. Proteins are inherently dynamic, exhibiting conformational freedom on timescales from picoseconds to seconds, implicating structural rearrangements that are essential for their biological function. Classical structural biology determines static models, that afford little insight into the underlying conformational equilibrium. The role that structural dynamics play in biological processes can only be understood by characterizing all thermally accessible protein conformations and their populations. NMR spectroscopy is uniquely sensitive to the presence of conformational dynamics in solution. Residual dipolar couplings (RDCs) measured in weakly aligned proteins, scalar couplings, and chemical shifts, probe motions occurring on timescales faster than 100 s of microseconds. These parameters therefore offer general tools to characterize protein motion on physiologically important timescales. A common approach to the dynamic interpretation of RDCs is to combine experimental restraint terms with a classical potential-energy force field to develop a conformational ensemble in agreement with experimental data. RDCs have also been exploited to characterize the conformational space sampled by the protein backbone either by fitting experimental data to determine angular excursions of internuclear bond vectors, or in comparison with different levels of accelerated molecular dynamics (AMD) to describe the most appropriate ensemble. Comparison of motions modeled using the Gaussian axial fluctuation (GAF) model, with ensembles derived from restraint-free AMD, demonstrated that such methods can provide a convergent description of protein motion.