Three-Dimensional Potential Energy Surface of Selected Carbohydrates' CH/π Dispersion Interactions Calculated by High-Level Quantum Mechanical Methods

In this study we present the first systematic computational three-dimensional scan of carbohydrate hydrophobic patches for the ability to interact through CH/pi dispersion interactions. The carbohydrates beta-D-glucopyranose, beta-D-mannopyranose and alpha-l-fucopyranose were studied in a complex with a benzene molecule, which served as a model of the CH/pi interaction in carbohydrate/protein complexes. The 3D relaxed scans were performed at the SCC-DFTB-D level with 3 757 grid points for both carbohydrate hydrophobic sides. The interaction energy of all grid points was recalculated at the DFT-D BP/def2-TZVPP level. The results obtained clearly show highly delimited and separated areas around each CH group, with an interaction energy up to -5.40 kcal mol(-1). The results also show that with increasing H center dot center dot center dot pi distance these delimited areas merge and form one larger region, which covers all hydrogen atoms on that specific carbohydrate side. Simultaneously, the interaction becomes weaker with an energy of -2.5 kcal mol(-1). All local energy minima were optimized at the DFT-D BP/def2-TZVPP level and the interaction energies of these complexes were refined by use of the high-level ab initio computation at the CCSD(T)/CBS level. Results obtained from the optimization suggest that the CH group hydrogen atoms are not equivalent and the interaction energy at the CCSD(T)/CBS level range from -3.54 to -5.40 kcal mol(-1). These results also reveal that the optimal H center dot center dot center dot pi distance for the CH/pi dispersion interaction is approximately (2.310 +/- 0.030) angstrom, and the angle defined as carbon-hydrogen-benzene geometrical centre is (180 +/- 30)degrees. These results reveal that whereas the dispersion interactions with the lowest interaction energies are quite strictly located in space, the slightly higher interaction energy regions adopt a much larger space.