Author
 |
French, Alfred - Al |
|
Johnson, Glenn |
|
Submitted to: Carbohydrate International Symposium
Publication Type: Abstract Only Publication Acceptance Date: 6/23/2005 Publication Date: 8/31/2005 Citation: French, A.D., Johnson, G.P. Conformational study of cellobiose by qm theory. Carbohydrate International Symposium. 2005. Abstr. No. OP9. Interpretive Summary: Technical Abstract: To gain a better understanding of the effects of intramolecular forces on disaccharide shapes, three regions of the ',' space of cellobiose were analyzed with quantum mechanics. A central region, in which most crystal structures are found, was covered by a 9 x 9 grid of 20' increments of ' and '. Besides these 81 minimizations that were constrained in ' and ', we studied two central sub-regions and two regions at the edges of our maps of complete ',' space with unconstrained minimization, for a total of 85 target geometries. HF/6-31G* and single-point HF/6-311+G* calculations were used to find the lowest energies for each target. B3LYP/6-31+G* and single point B3LYP/6-311+G* energies for the four regional minima were also computed. As in the work of Stortz, it was important to test many different combinations of exo-cyclic group orientations (starting geometries). For each of the 85 target geometries, 155 different starting geometries were tried. At the HF/6-31G* level, 31 different starting geometries resulted in the lowest energy for one or more of the target geometries. Those starting geometries came from four different methods that were based on molecular mechanics energies. None of these 155 structures had all 10 exo-cyclic orientations within 30' of any other starting geometry. A routine from MacroModel furnished 19 of the 31 structures. Six were based on erroneous interpretation of the 26 structures reported by Strati et al. Four of the correct structures by Strati et al. contributed, and of the 58 structures generated with a facility in Chem-X, two corresponded to lowest energies. Although all four sets contributed to the adiabatic energy map, use of any single set gave errors of 6 to 10 kcal/mol in this fairly low-energy region of ',' space. At the three other levels of theory, other starting geometries gave the lowest energies. Each of the four levels of theory gave a different overall lowest energy structure that was similar to, but not the same as, the one found by Strati et al. These structures are in a region that is not well populated by crystal structures but they are stabilized by highly cooperative hydrogen bonds. The HF/6-31G* energy contours of the mapped central region were compatible with the observed crystal structures. In particular, two experimental conformations from protein-carbohydrate complexes correspond to lower energies than on empirical energy surfaces. Observed structures that lacked O3...O5' hydrogen bonds were about 1 kcal/mol above the map’s minimum, and relative energies for observed structures that have a pseudo two-fold screw axis were about 0.4 to 1.0 |
