AbstractThe plasma membrane is a dynamic interface between a cell and its external environment. It is a complex structure composed of lipids, proteins, carbohydrates, and RNAs. Interactions between these constituent molecules give rise to biochemical processes such as cell-signalling, antigen presentation, and vesiculation. Part of this thesis seeks to con-tribute to our current understanding of the molecular origin of both physiological and pathophysiological phenomena in cellular membranes. This is done through use of coarse-grained and all-atom molecular dynamics (MD) simulations. The other part of this thesis aims to improve and add to the set of software tools currently used for analysing MD simulations of lipid membranes.
In the plasma membrane, specific lipid species are thought to aggregate into functional platforms known as ‘lipid-rafts’. The lipid-raft hypothesis posits that these small regions, nanometers in size, are comprised of highly ordered lipids in which cell-signalling proteins are embedded. If the aggregation of cell-signalling proteins within lipid-rafts is required for signal transmission, the breakdown of these structures would disrupt signalling pathways and in turn bring about cell death. In Chapter 3, I describe how the oxidation of cholesterol leads to the disruption of these highly ordered nanodomains in model membranes. I reveal three potential mechanisms by which nanodomain forma-tion is disrupted, and in doing so I provide a molecular level description of the means by which cholesterol oxidation may cause apoptosis in biological membranes.
Cholesterol and sphingomyelin are the two lipid species widely thought to be im-portant in the formation of lipid-rafts in mammalian plasma membranes. These two lipid species are known to have a high affinity for one another, and in complex lipid mixtures this leads to their co-localisation in highly ordered nanodomains. However, the precise origin of the preferential mixing of these two lipids is currently unknown. In Chapter 4, through an unsupervised clustering of cholesterol-sphingomyelin conformations I find there are four distinct modes of interaction between these lipid species. One of these modes is possible only with sphingomyelins — not other lipid species. This particular mode desolvates the hydrophobic core of cholesterol, thus reducing the free energy cost of exposing this core to the surrounding solvent. Therefore, I suggest this mode of interaction between cholesterol and sphingomyelin is the reason why cholesterol preferentially mixes with sphingomyelins over other lipid species.
In Chapter 5, I describe LiPyphilic - an open-source Python package I have created for analysing MD simulations of lipid membranes. LiPyphilic offers analyses that provide important structural and dynamical information about lipid membranes, but are not available in any other software. LiPyphilic is fast, fully-tested and easy to install. It is designed to be interoperable with the wider scientific Python stack, and was built following best practices in modern software development. The challenge now is to ensure the long-term sustainability of LiPyphilic by building a community of users and contributors to the project.
|Date of Award
|1 Jan 2022
|Chris Lorenz (Supervisor)