Computational simulations and methods for biomedical applications

Abstract

Author(s): Felberg, Lisa E. | Advisor(s): Head-Gordon, Teresa | Abstract: The interplay of molecular simulations and experiment have been instrumental in the last decade in establishing quantitative understanding of the physics underlying molecular processes relevant to applications involving interfaces, polymers and biopolymers and the water solvent used to solvate them. Water in confinement is present in a range of biologi- cal environments on the cellular level, important for stimuli responsive polymers and their use as drug delivery vehicles, and is ubiquitous in many technological applications such as water desalination. In the first half of this dissertation, I have used a range of experimen- tal and simulation techniques to study water in nanoscale confinement for water between graphene sheets and water trapped in stimuli-responsive star diblock polymers. Using MD simulations to study water confined between two graphene walls, it was observed that dif- ferent phases of water could be created as a function of the two-dimensional density and graphene wall flexibility such as square and hexagonal ice. Additionally, at incommensurate 2D densities, the flexible walls were found to bend, creating a coexistence in the system between n- and (n+1) water interlayers. Using small angle X-ray scattering and MD simula- tions, I determined that by varying the hydrophilic block chemistry of the star polymer arms using poly-ethylene glycol (PEG), poly-2,methyl-oxazoline (POXA) and a highly branched polycarbonate-based polymer with a pendant hydrophilic group (PC1), only the PEG sys- tem displayed thermosensitivity over the temperature range observed due to reduction in water entropy, while an increased sidechain length and charge density leads to decreased sol- vent interactions. In addition to temperature sensitivity, pH sensitivity of acidic, basic and neutral polymers was studied in non-degradable nanogel star polymers. Using small angle X-ray scattering, it was found that the acidic (PMAA) and basic (PDMAEMA) polymers exhibited sharp transitions between expanded and collapsed states, with apparent pKas that were qualitatively different from the reported monomer pKas. By modulating the fraction of basic or acidic groups in the hydrophilic region, we were also able to change the apparent pKa of the star polymer.Despite the large success of traditional atomistic simulation methods, their applicability to systems of large size and long time scales are prohibitive. Therefore there exists a need to develop novel methods to overcome this obstacle. In the second half of my thesis I present an efficient method to evaluate the electrostatic interactions in large molecular systems based on the linearized Poisson-Boltzmann equation (LPBE) , which allows for the simulation of much larger systems at a fraction of the cost. I have developed a robust software imple- mentation of the fully analytical LPBE model, PB-AM, which solves for the complete mutual polarization potential of a system comprised of an arbitrary number of molecules with ar- bitrary charge distributions in a screened environment with each molecule represented as a single, spherical, low dielectric cavity. I also developed a software implementation of the semi-analytical LPBE solution, PB-SAM that extends the analytical model that repre- sents a molecule as a collection over overlapping rigid spheres that better describes a detailed molecular boundary. The software is available as stand-alone code, with automated installa- tion using CMake and is also available as a part of the distributed and open source software in the highly popular Adaptive Poisson-Boltzmann Solver (APBS) package. Both software implementations have new features of a simple application programming interface, can pro- duce electrostatic potential visualizations in two and three dimensions, and run Brownian dynamics schemes with a variety of applications.