Enabling Atomistic Modeling and Simulation of Complex Curved Cellular Membranes with xMAS Builder.

As more powerful high performance computing resources are becoming available, there is a new opportunity to bring the unique capabilities of molecular dynamics (MD) simulations to cell-scale systems. Membranes are ubiquitous within cells and are responsible for a diverse set of essential biological functions, but building atomistic models of cell-scale membranes for MD simulations is immensely challenging because of their vast sizes, complex geometries, and complex compositions. To meet this challenge, we have developed xMAS Builder (Experimentally-Derived Membranes of Arbitrary Shape Builder), which is designed to take experimental lipidomics and structural (e.g., electron microscopy and tomography) data as input and use them to build MD-ready models of cellular membrane systems. To test xMAS Builder's capabilities, we have used it to build two models (one ~12.0 million atoms and the other ~11.6 million atoms) of a test system with a representative complex lipid composition and geometry. The two models, which differed only in their lipid packing densities, both maintained their membrane integrity during an extended MD simulation (250 ns and 386 ns), but their highly divergent relaxation dynamics indicate that the proper packing density of curved membranes is determined by leaflet volume rather than surface area. These results suggest that xMAS Builder's algorithms produce high quality models and that simulation of these models will provide profound biophysical insights into the behavior of cellular membranes.