(191dh) An Efficient Brownian Dynamics Approach for Modeling Multivalent Ligand-Receptor Assembly in the Cell Membrane

Multivalent ligand-receptor assembly is a key feature of many cell signal transduction systems. Nonetheless, the underlying mechanisms of ligand-receptor assembly in the plasma membrane of cells remain poorly understood. Current modeling and simulation techniques provide limited capabilities for the spatiotemporal modeling of multivalent species (molecules or complexes). A key barrier to modeling such systems is the multiscale nature of the problem. Multivalent molecules may interact via their subnanometer-scale binding features (domains or motifs). On the other hand, the time and spatial scales of assembly and diffusion of macromolecular species could be relatively large. In this work, we present a time-adaptive Brownian Dynamics (BD) simulation algorithm that enables efficient modeling of diffusion, reaction, and assembly of multivalent species. Combined with agent-based modeling, the BD approach provides a coarse-grained description of the structural details of multivalent molecules and their site-specific features. It enables motion of the multivalent molecules and their complexes due to lateral diffusion and rotation. The BD algorithm adaptively adjusts step sizes to capture the subnanometer scale site-specific interactions of the molecules and complexes, while enabling fast computation. We demonstrate these capabilities by developing models on multivalent antigen-receptor interactions in the plasma membrane of immune cells. Using the models, we provide analysis on the computational performance of the simulation approach both in the concentrated and dilute regimes. The models predict the distribution of receptor aggregate sizes in response to antigen molecules of variable valencies and structures. Consistent with experimental reports, the models also reveal and explain self-limiting antigen-receptor aggregation in the cell membrane.