TEMPERATURE-DEPENDENCE OF MOLECULAR-CONFORMATION, DYNAMICS, AND CHEMICAL-SHIFT ANISOTROPY OF ALPHA,ALPHA-TREHALOSE IN D2O BY NMR RELAXATION

A molecular dynamics study and solution-phase H-1 and C-13 chemical sh ift anisotropy determination of a symmetric cryogenic disaccharide, al pha,alpha-trehalose, has been performed in a temperature range between 264 and 350 K. Negligible temperature dependence of proton-carbon cou plings of the asymmetrically [1-C-13]-labeled trehalose suggest that t he averaged conformation of the interglycosidic linkage is centered ar ound dihedral angles phi = psi = -41 degrees with +/-5 degrees uncerta inty. Homonuclear NOE-s in the labeled trehalose support the dominance of similar conformation. Close to the slow and fast motional regime, alpha,alpha-trehalose can be considered as a spherical top, and global correlation times can be determined easily at extreme temperatures. T his allows the construction of an Arrhenius plot for the whole range o f temperature. The approach, which we call ''trouble-free'', yields E( a) = 28.2 kJ/mol for the activation energy of molecular reorientation. The model-free analysis of C-13 T-1, T-2, and NOE data showed a local maximum of the generalized order parameters with S-2 = 0.9 around 273 K. Monte-Carlo error analysis corroborated that this effect could be real; however, effective correlation times have relatively high error limits. Thermodynamically, the S-2 data can be interpreted in terms of changes in the Gibbs free energy due to increased or diminished spati al restriction of rapid CH fluctuations. Liquid state H-1 and C-13 che mical shift anisotropies were determined from the interference of dipo le-dipole and chemical shift anisotropy relaxation, In solution, chemi cal shift anisotropies cannot be separated from an inherent geometrica l factor, so a combined CSA, factor was used. Cross-correlated spectra l densities could be well fitted for the C-1,H-1 vector over the entir e temperature range with the ''trouble-free'' global correlation times . The resulting numerical values for CSA, were smaller compared to the model-free evaluation, due to the omission of internal fluctuations. The measured shift anisotropies were found to be independent from the selection of isotropic or anisotropic dynamical models. Apparent CSA(g ) factors were nearly constant in the entire temperature range except C-3, H-2, and H-3. Comparison with deuterium labeled [2,4,6-H-2]-treha lose proved that temperature-induced changes of the ABX-type strong co upling pattern (caused by the change of differential chemical shift of vicinal H-2 and H-3 protons) interfere with asymmetric multiplet rela xation and potentially lead to misinterpretation of CSA/DD relaxation rates.