Bacteria, Entropy and the Need for Speed
Julian C. Shillcock, Blue Brain Project, Ecole Polytechnique Fédérale de Lausanne
Abstract: Certain pathogens infect a cell by harnessing entropic forces that arise from membrane fluctuations to transport toxic proteins into a cell where they interfere with protein synthesis and kill the cell. Within the cytosol so-called intrinsically-disordered proteins (IDP) spontaneously phase separate into fluid droplets called biomolecular condensates. In healthy cells, these droplets form and dissolve under cellular control to carry out a variety of functions, but in chronic diseases such as Ahlzeimer’s and ALS they undergo a slow transition from a healthy, fluid state to a disease state in which their constituent proteins form rigid fibrils. IDPs lack a unique, folded state, and are conformationally dynamic. They behave as flexible polymers with multiple, weak, distributed binding sites by which they aggregate into fluid networks. The common feature to these phenomena is that correlations occur on length scales much larger than the individual molecules, which makes coarse-grained simulations ideal for their study. Here we show how dissipative particle dynamics (DPD) simulations reveal the entropic forces between membrane-bound nanoparticles, and how flexible, self-associating polymers reveal the dynamic organization of biomolecular condensates. Coarse-grained simulations like DPD are required for these problems because the length and time scales of interest are far beyond those accessible to atomistic molecular dynamics. Even for coarse-grained simulations, the biologically interesting problems are on the edge of what is currently feasible, and consume hundreds of core-weeks for one data point. Although high performance parallel computing has been available for many years, most parallel simulation techniques require all processors to update their states locked in step. This is inefficient unless every processor has the same amount of computation to do at all times. Additionally, to be efficient, each processor must spend most of its time on calculations not requiring messaging otherwise it idles while waiting for (temporally expensive) messages. POETS (Partially Ordered Event Triggered Systems) eliminates these barriers by reducing each message to a small size whose cost of transmission is essentially nil and making messaging asynchronous. By a fortunate convergence of distinct research fields, the computational requirements of DPD lie within the performance envelope of POETS. This promises to open up new areas of research in biological physics that were previously inaccessible, and reveal how molecular interactions give rise to cellular complexity.
Biography: Julian Shillcock received his PhD in 1995 at Simon Fraser University in Canada for work on Monte Carlo simulations of liquid crystal phase transitions and the elastic properties of fluid and polymerized membranes. He was a Group Leader at the Max Planck Institute of Colloids and Interfaces, Germany for five years applying coarse-grained simulation techniques – principally Dissipative Particle Dynamics (DPD) and Brownian Dynamics - to equilibrium and dynamic properties of fluid lipid membranes and actin filaments. A major target of this research was to reveal the molecular rearrangements that occur during vesicle fusion. During this time, he developed a parallel DPD code that has been used in academic and industrial research groups. He was an Associate Professor at MEMPHYS in the Department of Physics and Chemistry, University of Southern Denmark for four years. He joined the Blue Brain Project in 2011, and uses mesoscale simulation techniques and theoretical analysis to study the dynamics of cellular processes on mesoscopic length and time scales. Current projects include developing theoretical and simulational frameworks to explain the formation of biomolecular condensates formed of intrinsically disordered proteins. The neuronal post-synaptic density is an example of such a condensed phase, and they have been implicated in neurodegenerative diseases including Alzheimer’s and Parkinson’s disease.