In this study we present the first systematic computational three-dimensional scan of carbohydrate hydrophobic patches for the ability to interact through CH/π dispersion interactions. The carbohydrates β-d-glucopyranose, β-d-mannopyranose and α-l-fucopyranose were studied in a complex with a benzene molecule, which served as a model of the CH/π 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⋅⋅⋅π 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⋅⋅⋅π distance for the CH/π dispersion interaction is approximately (2.310±0.030) Å, and the angle defined as carbon-hydrogen-benzene geometrical centre is (180±30)°. 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.

In this article we present the first systematic study of the additive properties (i.e. degree of additivity) of the carbohydrate-aromatic moiety CH-π dispersion interaction. The additive properties were studied on the β-D-glucopyranose, β-D-mannopyranose and α-L-fucopyranose complexes with the naphthalene molecule by comparing the monodentate (single CH-π) and bidentate (two CH-π) complexes. All model complexes were optimized using the DFT-D approach, at the BP/def2-TZVPP level of theory. The interaction energies were refined using single point calculations at highly correlated ab initio methods at the CCSD(T)/CBS level, calculated as E + (E(CCSD(T))-E(MP2))(Small Basis). Bidentate complexes show very strong interactions in the range from -10.79 up to -7.15 and -8.20 up to -6.14 kcal mol(-1) for the DFT-D and CCSD(T)/CBS level, respectively. These values were compared with the sum of interaction energies of the appropriate monodentate carbohydrate-naphthalene complexes. The comparison reveals that the bidentate complex interaction energy is higher (interaction is weaker) than the sum of monodentate complex interaction energies. Bidentate complex interaction energy corresponds to 2/3 of the sum of the appropriate monodentate complex interaction energies (averaging over all modeled carbohydrate complexes). The observed interaction energies were also compared with the sum of interaction energies of the corresponding previously published carbohydrate-benzene complexes. Also in this case the interaction energy of the bidentate complex was higher (i.e. weaker interaction) than the sum of interaction energies of the corresponding benzene complexes. However, the obtained difference is lower than before, while the bidentate complex interaction energy corresponds to 4/5 of the sum of interaction energy of the benzene complexes, averaged over all structures. The mentioned comparison might aid protein engineering efforts where amino acid residues phenylalanine or tyrosine are to be replaced by a tryptophan and can help to predict the changes in the interactions. The observed results also show that DFT-D correctly describes the CH-π interaction energy and their additive properties in comparison to CCSD(T)/CBS calculated interaction energies. Thus, the DFT-D approach might be used for calculation of larger complexes of biological interest, where dispersion interaction plays an important role.

A method is described for determining 2,2'-i m i no bisetha no1 (diet ha nola m i ne) and tetra hyd ro-I ,4-isoxazi n (morpholine) in a system containing copper(1l). By using a more powerful complex-forming agent, copper(l1) is removed from the sample, and the compounds to be determined are extracted with acid and separated on a strongly acidic sulphonated cation-exchange column with phosphate buffer as the mobile phase. The relative error of the method is less than 5%, the detection limit being 20 p.p.m. for each amine.