Prof. Eva Meirovitch

Prof. Eva Meirovitch


Protein function is determined by both structure and dynamics. NMR spin relaxation is the single most useful method for studying protein dynamics. Uniform isotope labeling with 15N and 13C renders almost every atom in the protein NMR active. We found that the commonly used method for translating experimental NMR spin relaxation data into an insightful dynamic picture, called model-free (MF), is oversimplified. This implies often inaccurate, and at times misleading, dynamic pictures.

To improve the analysis, we developed a new approach for treating NMR spin relaxation data in protein. The theory underlying this approach, the Slowly Relaxing Local Structure (SRLS), has been developed by Freed and co-workers at Cornell University in the course of the years. This is a many-body stochastic coupled rotator approach. The rotators represent particles reorienting stochastically in solution. In the context of biological macromolecules, in particular proteins, the particles of interest are moieties such as the 15N-1H bond, or the 13CH2D methyl group, which bear the NMR nucleus observed (15N in the former case and 2H in the latter case). These nuclei represent “the probe”.

Each “probe” is involved in the global motion of the protein and in site-specific motions. Thus, both hydrodynamic information and intra-molecular flexibility can be elucidated.  The mutual correlation (coupling) between these motions, the appropriate symmetry of the physical quantities involved, and realistic local geometry are not accounted for properly in the MF approach. SRLS accounts for all of these important factors.

SRLS has been shown to be significantly more accurate and physically insightful than MF. So far our efforts focused on the dynamic properties of the protein backbone using 15N-1H as probe, and the dynamic properties of protein side chains using the methyl group 13CH2D as probe. A new picture of protein dynamics has emerged. This is a consequence of the fact that using over-simplified methods can lead not only to quantitatively inaccurate, but also to qualitatively erroneous results.

Efforts to extend the capabilities of SRLS are underway. Work related with new applications is in progress.  We also contemplate combining SRLS with molecular dynamics simulations.  The development of effective user-friendly software for distribution in the protein dynamics community is pursued.

It should be noted that the field of research described lies at the interface of Magnetic Resonance, NMR spin relaxation, liquid dynamics, statistical mechanics, structural biology and protein NMR.