Toward direct determination of conformations of protein building units from multidimensional NMR experiments part II: a theoretical case study of formyl-L-valine amide.
Chemical shielding anisotropy tensors have been determined for all twenty-seven characteristic conformers of For-L-Val-NH2 using the GIAO-RHF formalism with the 6-31 + G* and TZ2P basis sets. The individual chemical shifts and their conformational averages have been compared to their experimental counterparts taken from the BioMagnetic Resonance Bank (BMRB). At the highest level of theory applied, for all nuclei but the amide proton, deviations between statistically averaged theoretical and experimental chemical shifts are as low as 1-3%. Correlated chemical shift plots of selected nuclei, as function of the respective phi, psi, chi1, and chi2 torsional angles, have been generated. On two-dimensional chemical shift-chemical shift plots, for example, 1H(NH)-15N(NH) and 15N(NH)-13Calpha, regions corresponding to major conformational clusters have been identified, providing a basis for the quantitative identification of conformers from NMR shift data. Experimental NMR resonances of nuclei of valine residues have been deduced from 18 selected proteins, resulting in 93 1Halpha-13Calpha chemical shift pairs. These experimental results have been compared to relevant ab initio values revealing remarkable correlation between the two sets of data. Correlations of 1Halpha and 13Calpha values with backbone conformational parameters (phi and psi) have also been found for all pairs (e.g. 1Halpha/phi and 13Calpha/phi) but 1Halpha/psi. Overall, the appealing idea of establishing backbone folding of proteins by employing chemical shift information alone, obtained from selected multiple-pulse NMR experiments (e.g. 2D-HSQC, 2D-HMQC, and 3D-HNCA), has received further support.