Abstract
Biopolymers are macromolecules formed by the repetition of numerous subunits called monomers and derived from biomass. They constitute a class of materials used in various fields: pharmaceutical and cosmetic industry, tissue engineering, agro-alimentary and medicine. Their availability, price, non-toxicity and biodegradability make these materials a growing topic of interest to the community. The assembly of biopolymers in gels induced by an external factor (metal cations, pH and organic crosslinking agents) is one of their most interesting properties. The optimization of the physico-chemical, micro- and macroscopic properties of the gels necessitates understanding of factors that influence the gel equilibrium structures. Molecular modeling tools applied to study the inter- and intra-molecular interactions in biopolymer assemblies at short and long distances, such as covalent, electrostatic and the Van der Waals interactions or hydrogen bonds, can provide valuable information The effect of temperature and solvent also requires simulations on extended time and space scales. In this sense, a computational approach plays the role of "digital microscope".In this thesis, a multi-scale approach has been developed for the study of a group of polysaccharides: alginates. The formation of alginate/multivalent cation complexes (Alg/Mn+) is a determining step in the entire gelation process. The Density Functional Theory augmented with an empirical London dispersion term (DFT-D) applied to study various models of alginate complexes with divalent (Mn2+, Co2+, Cu2+ and Zn2+) and trivalent (Al3+, Cr3+, Ga3+, Fe3+, La3+ and Sc3+) metal cations led to the following conclusions. The tendency of binding energy of metal cations is independent on : the length and number of alginate chains, the presence of water in the first coordination sphere and the speciation of the metal. In all the hydrated structures, a monodentate mode of binding is established between the cation and the carboxylate groups. Thus, the interaction between cations and alginate depends only on local parameters which can be explained by a significant covalent contribution from the charge donation of carboxylic oxygen ligand atoms to Mn+. The importance of the binding strength for the complexation was confirmed by the DFT-D study of ion exchange using mixed La3+/Cu2+ alginate complexes. The Born-Oppenheimer Molecular Dynamics (BOMD) / Molecular Dynamics (MD) simulations of dimannuronates in water and lanthanum, and aluminum complexes revealed that water prevents the formation of intra-chain hydrogen bonds by solvating the functional groups. In the complexes of lanthanum dimannuronate and aluminum dimannuronate, the radial distribution functions of the carboxylic oxygen atoms confirm the establishment of a monodentate bonding mode. In all simulated complexes, no water molecule separates the metal ion from the ligand, thus forming inner sphere complex.As an alternative to MD simulations, a Monte Carlo parallel tempering method has been implemented in the deMon2k code. The QM/MM and QM/MC methods will also allow the examination of longer chains effect on the transition metal cations complexation by considering the temperature and the explicit aqueous solvent. These computational tools can be used straightforwardly for the study of other polysaccharides and biopolymers.