Conventional NMR structure determinations suffer from poor resolution of side chains, leading to the perception that X-ray crystal structures are more accurate. Current NMR chemical shift-based dihedral angle restraints and distance restraints from nuclear overhauser enhancements (NOEs) are insufficient for defining many side chain positions in a protein. We have developed robust methodology for determining precise side chain dihedral angles for protein residues with rigid side chains, versus rotamer populations for flexible side chains. We distinguish between rigid and mobile residues using 1 H-NMR transverse relaxation rates. We then fit quantitative J coupling data (3-bond J couplings for N-HB, CO-HB, and HA-HB) to the appropriate model. We demonstrate our methodology on a 131-residue chimeric protein construct of human cardiac troponin complex. We were able to obtain excellent side chain data for all structured residues in our protein construct. Most side chains are mobile, being able to access at least two major side chain rotameric states. 36/131 residues were deemed to have a single rotameric state that we restrained to +/− 10 degrees, and 14 residues were deemed capable of rotameric jumps but had a dominant rotamer that we restrained to +/− 20 degrees. In the absence of dihedral angle restraints, there were 9 residues that were constrained to a single completely wrong rotameric state, as is the common fear in NMR structure determination. The combined use of 1 H relaxation and quantitative J coupling analysis leads to NMR structure determinations of unprecedented quality in terms of side chain structure determination. Use of this new methodology allows solution NMR to discern side chain structure and dynamics under physiologic conditions better than X-ray crystallography.
Dumont et al. (Sun,) studied this question.