Simulation of Raman optical activity of multi-component monosaccharide samples.

Determination of the saccharide structure in solution is a laborious process that can be significantly enhanced by optical spectroscopies. Raman optical activity (ROA) spectra are particularly sensitive to the chirality and conformation. However, the interpretation of them is largely dependent on computational tools providing a limited precision only. To understand the limitations and the link between spectral shapes and the structure, in the present study we measured and interpreted using a combination of molecular dynamics (MD) and density functional theory (DFT) Raman and ROA spectra of glucose and mannose solutions. Factors important for analyses of mixtures of conformers, anomers, and different monosaccharides are discussed as well. The accuracy of the simulations was found to be strongly dependent on the quality of the hydration model; the dielectric continuum solvent model provided lower accuracy than averaging of many solvent-solute clusters. This was due to different conformer weighting rather than direct involvement of water molecules in scattering recorded as ROA. However, the cluster-based simulations also failed to correctly reproduce the ratios of principal monosaccharide forms. The best results were obtained by a combined MD/DFT simulation, with the ratio of α- and β-anomers and the -CH2OH group rotamers determined experimentally by NMR. Then a decomposition of experimental spectra into calculated subspectra provided realistic results even for the glucose and mannose mixtures. Raman spectra decomposition provided a better overall accuracy (∼5%) than ROA (∼10%). The combination of vibrational spectroscopy with theoretical simulations represents a powerful tool for analysing the saccharide structure. Conversely, the ROA and Raman data can be used to verify the quality of MD force fields and other parameters of computational modeling.