Additive empirical force field for hexopyranose monosaccharides

Abstract

We present an all-atom additive empirical force field for the hexopyranose monosaccharide form of glucose and its diastereomers allose, altrose, galactose, gulose, idose, mannose, and talose. The model is developed to be consistent with the CHARMM all-atom biomolecular force fields, and the same parameters are used for all diastereomers, including both the α-and β-anomers of each monosaccharide. The force field is developed in a hierarchical manner and reproduces the gas-phase and condensed-phase properties of small-molecule model compounds corresponding to fragments of pyranose monosaccharides. The resultant parameters are transferred to the full pyranose monosaccharides and additional parameter development is done to achieve a complete hexopyranose monosaccharide force field. Parametrization target data include vibrational frequencies, crystal geometries, solute -water interaction energies, molecular volumes, heats of vaporization, and conformational energies, including those for over 1800 monosaccharide conformations at the MP2/ cc-pVTZ//MP2/6-31G(d) level of theory. Though not targeted during parametrization, free energies of aqueous solvation for the model compounds compare favorably with experimental values. Also well-reproduced are monosaccharide crystal unit cell dimensions and ring pucker, densities of concentrated aqueous glucose systems, and the thermodynamic and dynamic properties of the exocyclic torsion in dilute aqueous systems. The new parameter set expands the CHARMM additive force field to allow for simulation of heterogeneous systems that include hexopyranose monosaccharides in addition to proteins, nucleic acids, and lipids.

Phosphatidylcholine reverse micelles on the wrong track in molecular dynamics simulations of phospholipids in an organic solvent

“…The motivation rises from questioning the validity of the increasingly common usage of biomolecular force fields in describing systems outside their parametrization environment. CHARMM36 force field performs very well in describing all the studied components individually or in aqueous environment, 37,38 and is currently one of the highest regarded lipid force fields, with good accuracy in water environment. [1][2][3][4] However, the FIG.…”

“…We chose these particular models because the lipid model has been validated also at low hydration with the water model, 37 verified parameters are available for both saturated and unsaturated lipids, 37 and force field consistent parameters for cyclohexane exist. 38 Furthermore, the accuracy of the lipid model compares favourably to other commonly used lipid force fields in a recent review, 1 or is among the best performing in several others 2,4 with specifically the deuterium order parameters most accurately reproduced. 4 In total, the force field is a very good candidate for good performance in describing the PC/cyclohexane system.…”

“…Cyclohexane parameters were taken from the compatible CHARMM carbohydrate force field of Ref. 38. Water was modelled using CHARMM's modified version of TIP3P water model, 36 which has Lennard-Jones (LJ) interactions also for the hydrogen atoms.…”

“…Hexadecane here shares the parameters of the PC lipid hydrocarbon tails in CHARMM36 lipid description 37 while two cyclohexane models were compared. One is by Guvench et al 38 (based on alkane parameters by Vorobyov et al 47 ) and the other has the aliphatic non-bonded parameters replaced by those of Yin and MacKerell 48 corresponding to reverting the non-bonded aliphatic parameters back to those of CHARMM27. Additionally, the cyclohexane by Vorobyov et al 47 was tested but as the models differ by only angle and dihedral parameters, no improvement over Guvench et al 38 model was found.…”

“…One is by Guvench et al 38 (based on alkane parameters by Vorobyov et al 47 ) and the other has the aliphatic non-bonded parameters replaced by those of Yin and MacKerell 48 corresponding to reverting the non-bonded aliphatic parameters back to those of CHARMM27. Additionally, the cyclohexane by Vorobyov et al 47 was tested but as the models differ by only angle and dihedral parameters, no improvement over Guvench et al 38 model was found. Initial configurations for these systems were prepared by first randomly placing the hexadecane and cyclohexane molecules inside a cubic box (10 × 10 × 10 nm 3 ).…”