A theoretical case study of type I and type II beta-turns

NMR chemical shielding anisotropy tensors have been computed by employing a medium size basis set and the GIAO-DFT(B3LYP) formalism of electronic structure theory for all of the atoms of type I and type II beta-turn models. The models contain all possible combinations of the amino acid residues Gly, Ala, Val, and Ser, with all possible side-chain orientations where applicable in a dipeptide. The several hundred structures investigated contain either constrained or optimized phi, psi, and chi dihedral angles. A statistical analysis of the resulting large database was performed and multidimensional (2D and 3D) chemical-shift/chemical-shift plots were generated. The (1)H(alpha-13)C(alpha), (13)C(alpha-1)H(alpha-13)C(beta), and (13)C(alpha-1)H(alpha-13)C' 2D and 3D plots have the notable feature that the conformers clearly cluster in distinct regions. This allows straightforward identification of the backbone and side-chain conformations of the residues forming beta-turns. Chemical shift calculations on larger For-(L-Ala)(n)-NH(2) (n=4, 6, 8) models, containing a single type I or type II beta-turn, prove that the simple models employed are adequate. A limited number of chemical shift calculations performed at the highly correlated CCSD(T) level prove the adequacy of the computational method chosen. For all nuclei, statistically averaged theoretical and experimental shifts taken from the BioMagnetic Resonance Bank (BMRB) exhibit good correlation. These results confirm and extend our previous findings that chemical shift information from selected multiple-pulse NMR experiments could be employed directly to extract folding information for polypeptides and proteins.