|Kelterer, Anne-Marie - TECHNISCHE UNIV. GRAZ|
Submitted to: Carbohydrate International Symposium
Publication Type: Abstract Only
Publication Acceptance Date: June 1, 2003
Publication Date: June 1, 2003
Citation: French, A.D., Kelterer, A., Johnson, G.P. 2004. Perfluorinated cellobiose as a theoretical model for cellulose and related disaccharides [abstract]. Carbohydrate International Symposium. p. 39. Technical Abstract: Ramachandrannn (phi, psi) plots for perfluorinated cellobiose were calculated at the HF/6-31G* and HF/6-311+G* levels of quantum chemical theory. They provide surprisingly good (but slightly different) predictions of the linkage conformations that are observed by crystallography on a wide variety of compounds having beta-(1,4)-linkages. The smaller number of rotatable exo-cyclic groups of the perfluorinated molecule compared to cellobiose permits an exhaustive examination of all exo-cyclic orientations at expensive quantum chemical levels, at least within the fourth of phi, psi space that is populated by crystal structures. In some ways, this finding echoes our results with simpler THP models [1,2]. The perfluorinated structure is sterically a better analog than those analogs, although the bulk of a hydroxyl group is still somewhat underrepresented by a fluorine atom. Crystal structure conformations were taken from three venues, including small molecule crystal structures of compounds having the same tetrahydropyrans (THP) ring backbone of cellobiose. Observed linkage data also came from cellodextrins and lactose moieties in complexes with protein crystal structures. A third set of structures comes from fiber diffraction studies of cellulose and its derivatives. The latter two groups have lower quality data, and separate considerations apply to data from all three venues. Still, almost all structures are found in a relatively restricted area of phi, psi space, surrounded by the 1 kcal/mol contour line. None of these model compounds has hydroxyl groups that could result in hydrogen bonding. The successful prediction of the observed structures of hydroxyl-bearing compounds indicates that hydrogen bonding is much less of an influence on crystallline conformation than is widely thought. The lecture will address the details of this analysis, including the possibility that the high electronegativity of the fluorine atom is affecting the torsional energies as well as the various considerations that might provide separate groups of conformations for different sources of crystal structures.