Submitted to: Carbohydrate Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/4/2000
Publication Date: N/A
Citation: N/A Interpretive Summary: Computerized molecular models of agricultural materials are important for the interpretation of other experiments that are intended to understand the basic nature of these materials. However, carbohydrate materials have been particularly difficult to model well and sucrose, common table sugar, is especially difficult. If sucrose is modeled successfully, then a new level of ability to understand molecular structure will have been reached. This paper shows that the observed structures of sucrose can be predicted based just on its chemical structure. Normally models are based on one of two different methods. Some are based on empirical forces taken from experimental results. Others are based on the first principles of physics, the so-called ab initio quantum mechanics method. Our accomplishment required the time-consuming (one year of computer time) quantum mechanics calculations to be combined with much faster empirical, molecular mechanics calculations. This work is primarily of interest to biologists trying to understand the role of sucrose in numerous biological functions, including biosynthesis of cotton cellulose, and to chemists who are trying to make new materials from sucrose. It is also of importance to those who are developing modeling software.
Technical Abstract: Despite the importance of sucrose, its crystallographically determined shapes had not been explained satisfactorily by computerized molecular modeling. Its inter-residue linkage, with axial and usually pseudo-axial glycosidic bonds, is somewhat flexible and previous efforts had not simultaneously calculated low energies for observed shapes and high energies for shapes that are not observed. In the present work, both quantum mechanics (QM) and molecular mechanics (MM) were used to produce a hybrid energy surface. Improved predictive ability resulted from using torsional energies that were calculated by HF/6-31G* QM for an analog of sucrose based on tetrahydropyran (THP) and tetrahydrofuran (TFH). Remaining contributions to the potential energy of sucrose were calculated with the MM program, MM3. The QM torsional energies based on that dimeric THP-O-THF analog predicted crystalline conformations better than those from individual 2-Me-O-THP and 2-Me-O-THF molecules, indicating substantial non-additivity. It is proposed that the QM values for the torsional energies from the dimeric analog were needed in part because of an over-lapping exo-anomeric effect. Prediction of the distribution of observable geometries was also enhanced by reducing the strength of hydrogen bonding, compared to values that are used for studies of molecules in vacuum. This is consistent with the absence of inter-residue, intramolecular hydrogen bonds in many crystalline sucrose moieties.