Hannah Reeves Research Talk (2020 Larock Conference, Physical/Analytical Presentation #1)

Title: The impact of conformational sampling on first-principles calculations of vicinal J-couplings in carbohydrates

Abstract: Carbohydrates play key roles in cellular function including cellular signaling, metabolism, and structure. Studies of carbohydrate structure often employ NMR experiments to measure vicinal J-coupling constants, which are highly correlated with dihedral angles of interest such as glycosidic linkages. Empirical Karplus relations are used to map dihedral angles to J-couplings, but their disadvantages include the relative scarcity of X-ray structures needed to fit the Karplus parameters, and it is unclear whether existing crystal structures accurately represent carbohydrates in biological environments.

With computational molecular dynamics simulations, a model of a molecule can be simulated in 3-D space. To determine the accuracy of the model, we can simulate J-couplings using the Karplus relation and make comparisons with the experiment, but the Karplus relation itself introduces another source of error, making it hard to directly validate the simulation accuracy. An alternate approach would be to use density functional theory (DFT) to compute the J-couplings for simulated structures, but the accuracy of such an approach needs to be validated as well. Therefore, I conducted (DFT) calculations of J-couplings on a series of rigid carbohydrates with known dihedral angles.

The predicted J-couplings from DFT calculations were compared to the experimentally-derived J-couplings. To probe the importance of modeling solvent and conformational flexibility, I compared calculations of the molecule in gas phase vs. in implicit solvent, and a single energy minimized structure vs. averaging over multiple frames of a DFT MD simulation. Our results provide important guidelines for comparing MD simulations of carbohydrate structures with NMR experiments.