Examining the Accuracy of Monosaccharide Structures Calculated with Density Functional Theory by Experimental Comparisons

Abstract

Glycans are prominent in biological processes. Glycan structural characterizations can have significant implications for disease progression, e.g. cancer metastasis.1 Glycans are challenging to analyze with traditional chemical techniques due to a lack in diversity of functional groups; thus, computational modeling of glycans can help bridge this gap in knowledge regarding glycan structures. However, it is necessary that the computational models adequately represent carbohydrate structures and align with experimental results. Experimentally, mass spectrometry (MS) is one technique that is used for glycan characterization. In MS analyses, glycans are often analyzed as ion adducts. A group of computational methods and basis sets have been used to compare the energies of a subset of monosaccharides2, but this work has not been expanded to the ion adducts detected with MS, nor has it included experimental comparison. Herein, we optimized the geometry of both anomers of fucose and glucose in neutral, protonated, ammoniated, and sodiated forms using Density Functional Theory with the B3LYP or B3PW91 methods and 15 basis sets. Three basis sets (cc‐pVQZ, GEN, and GENECP) were unsuccessful in running calculations to completion. For the remaining computations, we examined values from the output files for variations between methods and basis sets. If computational methods yield accurate structures, the calculated values for molecular parameters, e.g. dipole moment, should be similar. Yet, if there are incorrect approximations in the basis set, the calculated electron density around the molecule will be incorrect and alter the values of the parameters using that specific functional. Greater variation occurred in the average dipole moments calculated for all methods and basis sets of fucose (standard deviation (SD) = 1.5) compared to glucose (SD = 0.68). The addition of ion adducts increased error in dipole moment calculations compared to the neutral molecules. The variation in these results suggests that some methods and basis sets are not reliable when modelling carbohydrates. Currently, molecular dynamics runs using GROMACS are being done to obtain over 50 different coordinate sets of each system for input into MOBCAL.3 MOBCAL will be used to obtain theoretical collisional cross sections (CCS) which will be compared with experimentally measured CCS by ion mobility MS. These results will illustrate which computational methods yield accurate carbohydrate structures, allowing future application of this work to address the biological utility of glycans.Support or Funding InformationBaylor University Office of the Vice Provost for Research and the Undergraduate Research and Scholarly Achievement Small Grant Program.This abstract is from the Experimental Biology 2019 Meeting. There is no full text article associated with this abstract published in The FASEB Journal.