Combined NMR, grid search/MM3 and metropolis Monte Carlo/GEGOP studies of two L-fucose containing disaccharides: α-L-Fuc-(1,4)-β-D-GlcNAc-OMe and α-L-Fuc-(1,6)-β-D-GlcNAc-OMe

Complete proton NMR data provided a firm experimental basis to infer the conformational properties of the Lewis type disaccharide α-L-Fuc-(1,4)-β-D-GlcNAc-OMe 1 and the N-glycoprotein type disaccharide α-L-Fuc-(1,6)-β-D-GlcNAc-OMe 2 in aqueous solution. Relaxed potential energy maps from systematic grid searches (GS) using the MM3 force field and Metropolis Monte Carlo (MMC) simulations employing the GEGOP force field were used to calculate corresponding ensemble average NMR data such as ID transient NOE curves and vicinal coupling constants J(H,H). For the disaccharide with a flexible (1,6) linkage (2), the relaxed potential energy surface based on the MM3 force field was calculated with three variable dihedral angles, φ, ψ and ω. R-factors derived from a comparison of experimental and theoretical NOE data allowed an evaluation of the quality of the conformational models derived from the calculations. Seven inter glycosidic 1D transient NOE curves were measured for each of both disaccharides, 1 and 2. Overall R-factors of approximately 17% for the 1 -4 linked disaccharide and 20% for the 1-6 linked disaccharide indicate a very satisfying agreement between theoretical calculations, both GS/MM3 and MMC/GEGOP, and experimental NOE data, indicating that the gross conformational picture developed is realistic. On the other hand, a close inspection of individual NOE curves and vicinal coupling constants ³J(H5,H6-pro-R) and ³J(H5,H6-pro-S) with corresponding theoretical values revealed shortcomings of the computational methods applied. Firstly, the conformational equilibrium around the C5-C6 bond of the GIcNAc unit in disaccharide 1 is not correctly described by the GS/MM3 method. This can lead to a false interpretation of NOE data involving protons attached to C6 of GlcNAc. Secondly, relaxation of ring geometry (GS/MM3) was found to have a measurable improvement of intra glycosidic NOEs in both disaccharides, 1 and 2. In summary, the conformational models derived for disaccharides 1 and 2 represent a starting point for further analysis of potential conformational changes that may occur upon binding of these compounds to specific receptor proteins such as lectins or antibodies.