Abstract

This thesis is devoted to the study of the conformation of monosacchrides in six-membered ring form. The main goal is to develop and apply new computational tools to investigate conformational properties and to improve the description of carbohydrates in the framework of molecular dynamics simulations. In the field of monosaccharides, modeling the system within the molecular dynamics framework presents troublesome aspects. The most important issue is that some force fields (e.g., the chosen gromos 45a4 parameter set) fail in reproducing the conformational preferences of the sugar constituents, with the appearance of unphysical conformations. This lack stems from the fact that the conformational behavior, dominated by few structures, generates a severe bottleneck: the non-ergodicity of the system by any practical means. This aspect explains the interest in free energy calculations, and methods exist, such as umbrella sampling or metadynamics, that allow to accelerate the sampling of different conformations by adding bias forces. In general, accelerated sampling methods are based on the choice of collective variables (CVs), which is of particular importance for the proper reconstruction of free energy landscapes. In the field of conformational analysis, suitable CVs have to be considered to describe non-planar, puckered conformations of cyclic structures. One of the main goals of this work is the enhancement of the gromos 45a4 force field for carbohydrates, with respect to the ability to describe ring conformation (that is, puckering) of six-membered rings. To this end, the development of efficient computational tools for the investigation of the general puckering problem are presented. In particular, we indicate how to exploit the capabilities of the metadynamics algorithm applied to the investigation of puckered ring conformers, exploring also different parametrizations of puckered structures to assess their respective advantages as collective variables for metadynamics.