Molecular origins of the thermodynamic properties, phase separation behaviour and structure of biomolecules in aqueous solutions

Biological systems are immensely complex with many components, intricate structures and sensitive phase behaviour. To gain a better understanding and move towards the modelling of biological systems, molecular modelling and simulation techniques are used, in this thesis, to study the phase behaviour of two biocompatible polymers, elastin-like polypeptides (ELPs) and polyethylene glycol (PEG), and a class of simple lipid molecules, monoglycerides, in aqueous solution. The simulation length and time scales required to study polymer phase behaviour and lipid self-assembly, let alone actual biological phenomena, remain out of reach of atomistic simulations. Removing atomistic degrees of freedom, coarse-graining (CG), enables the exploration of larger spatio-temporal scales. The challenge, however, is to develop CG force fields that are accurate, transferable and predictive. In this thesis, statistical association fluid theory (SAFT) methodologies are demonstrated to be viable tools for the study of biocompatible polymers and simple lipids: being capable as an equation of state (EoS) of extrapolating experimental data to a wide range of conditions in a physically consistent manner for water-ELP mixtures, and as a CG strategy for the study of structural properties of aqueous PEG and monoglycerides in simulation. For water and ELP mixtures, the SAFT EoS accurately reproduces experimental phase boundaries at low pressures giving credence to the phase behaviour predictions of other molecular weight ELPs over wider ranges of system conditions. At high pressures, the global phase behaviour predicted by the SAFT EoS shows bimodal liquid critical points and multiple re-entrant liquid-liquid immiscibility regions; features indicative of a new type of binary fluid phase behaviour. For the simulation of PEG and monoglycerides, the SAFT EoS is used to fit CG force fields to thermodynamic data. The non-bonded force fields from SAFT are combined with bonded potentials obtained from atomistic simulations to create semi-flexible models to study structural and interfacial properties not obtainable from the SAFT theory. For mixtures of water and PEG, temperature-dependent parameters are used and fine-tuned in simulation to model the closed-loop liquid-liquid equilibrium. The parameters are found to be transferable to PEG molecules of different molecular weight and simulations provide predictions of the radius of gyration, Flory exponents, surface pressures, surface thickness and surface adsorption concentration profiles; all consistent with experiments. The same methodology is applied to develop CG force fields for the amphiphilic monoglycerides in water. The model is used to study the effect of the hydrocarbon tail length and degree of tail saturation on the self-assembly of monoglycerides into mesophases in simulation, which predict a breadth of structures including lamellar, hexagonal, bicontinuous and micellar phases. The work in this thesis represents a step towards the application of SAFT methodologies to biologically relevant systems.Open Acces