Advancing our understanding of lipid bilayer interactions: a molecular dynamics study

Abstract

In recent years, advances in computer architecture and lipid force field parameters have made Molecular Dynamics (MD) a powerful tool for gaining atomistic resolution of biological membranes on timescales that other tools simply cannot explore. With many key biological processes involving membranes occurring on the nanosecond timescale, MD allows us to probe the dynamics and energetics of these interactions in molecular detail. Specifically, we can observe the interactions taking place as a peptide or protein comes into contact with a lipid bilayer, and how this may shape or alter the bilayer either locally (changes in headgroup orientation, lipid fluidity) or in bulk (lipid demixing, membrane curvature). The resolution achieved through atomistic MD can be directly compared with other tools such as NMR and EPR to gain a full perspective of how these biological systems behave over different timescales. As my background is in computational physics, this thesis not only looks into broadening our understanding of various interactions with biological membranes, but also into the development of construction and analytical software to assist in my research and benefit others in the field. One aspect of biological membranes that could vastly benefit from MD simulations is that of antimicrobial peptides (AMPs). These peptides primarily target and destroy microbes by permeabilising the cell membrane through a variety of proposed mechanisms, where each mechanism relies on the AMP to adopt specific conformations upon contact with bacterial membranes. In this thesis, I present an investigation into the interactions between a synthetic AMP and an inhibitor peptide designed to regulate antimicrobial activity through the formation of a coiled coil structure, which restricts the AMP from adopting new conformations. Simulations captured the spontaneous formation of coiled coils between these peptides, and specific residues in their sequences were identified that promote unfolding. This knowledge may lead to better design of coiled coil forming peptides. Another aspect of biological membranes that can be explored with MD is the interactions between model bacterial membranes and amphipathic helices, such as the MinD membrane targeting sequence (MinD-MTS). This 11-residue helix is responsible for anchoring the MinD protein to the inner membrane of Bacillus subtilis and plays a crucial role in bacterial cell division. MinD is known to exhibit sensitivity to transmembrane potentials (TMVs), whereby its localisation and binding affinity to bacterial membranes are disrupted upon removal of the TMV. Simulations revealed rapid insertions of MinD-MTS peptides into the headgroup region of a model bacterial membrane. Analytical software was constructed to measure the membrane properties of the lipids surrounding inserted MinDMTS peptides, which revealed splayed lipid tails and suggests the MinD-MTS may be capable of inducing membrane curvature. Additional simulations were conducted to investigate the influence of a TMV on model bacterial membranes, where software was constructed to measure changes in membrane properties. An analysis of these simulations suggests that a TMV is capable of lowering the transition temperature of a model bacterial membrane by a few degrees, yielding increased fluidity in the lipids and increased perturbations on the membrane surface. Finally, another aspect of biological membranes that can be explored through MD is that of electroporation. This induction of transient water pores in cell membrane provides an exciting aspect for drug delivery applications into cells, whereby electric fields are applied to cells to increase the uptake of therapeutic drugs. Simulations of membranes with high voltage TMVs were conducted that sought to investigate the implications of electroporation across a variety of bilayer compositions at different temperatures. Software was constructed to measure changes in membrane and system properties, which revealed that pore formation occurred at the same threshold voltage for different bilayer compositions in the fluid phase (~1.9 V) and a higher voltage for DPPC bilayers in the gel phase (~2.4 V). The TMV was found to be highly dependent on the area per lipid (APL), implying that bilayers with bulkier lipids or those transitioning from gel to fluid will experience smaller TMVs and fewer pore formations. These simulations also revealed lipid flip-flopping through pores, where charged lipids tended to translocate in the direction of the electric field to produce an asymmetrically charged bilayer. Finally, simulations utilising charged peptides with membranes yielded electroporation effects, whereby the charged peptides generate an identical TMV to those produced by an ion imbalance of equal magnitude. This suggests that charged peptides, such as AMPs, may be capable of permeabilising cell membranes through electroporation mechanisms.