Protein side-chain dynamics as observed by solution- and solid-state NMR spectroscopy: a similarity revealed

In this paper, we seek to compare the internal dynamics of a small globular protein, SH3 domain from alpha-spectrin, in solution and in a crystalline state. The comparison involves side-chain methyl 13C R1 relaxation rates that are highly sensitive to local dynamics in the vicinity of the methyl site. To conduct the relaxation measurements, protein samples have been prepared using specially labeled alpha-ketoisovalerate precursors, resulting in selective incorporation of the 1H-13C spin pair in one or both methyl groups of the valine and leucine side chains. The sparse labeling pattern in an otherwise deuterated sample makes it possible to record high-resolution 13C, 1H solid-state spectra using magic angle spinning experiment with a MAS frequency of 22 kHz. Furthermore, this labeling scheme avoids proton-driven 13C-13C spin-diffusion effects, thus allowing for accurate measurements of 13C R1 relaxation in the individual methyl groups. While the relaxation response from a polycrystalline sample is generally expected to be multiexponential, we demonstrate both theoretically and experimentally that in this particular case the relaxation profiles are, in excellent approximation, monoexponential. In fact, solid-state relaxation data can be interpreted in a model-free fashion, similar to solution data. Direct comparison between the experimentally measured solid and solution rates reveals a strong correlation, r = 0.94. Furthermore, when solution rates are corrected for the effect of the overall molecular tumbling (quantified on the basis of the solution 15N relaxation data), the results are in one-to-one agreement with the solid-state rates. This finding indicates that methyl dynamics in the solution and solid samples are quantitatively similar. More broadly, it suggests that the entire dynamic network, including motions of side chains in the protein hydrophobic core and backbone motions, is similar. This result opens interesting possibilities for combined interpretation of solid- and solution-state relaxation data, potentially leading to a detailed characterization of internal protein dynamics on a wide range of time scales.