This project proposes to develop new network-free spatial modeling software at the mesoscale - occupying the niche between detailed molecular dynamics and cellular reaction-diffusion systems. Specifically, we plan to address spatial scales in the range of 50nm ? 2m and temporal scales within the range of 100s to 10s. Examples of systems that would benefit from modeling tools at this ?mesoscale? are receptor signaling platforms and clusters (e.g. the immune synapse or the post-synaptic density), cell adhesion complexes, lipid rafts, chromatin organization, cytoskeletal dynamics, and nucleoprotein phases. Our approach builds on the foundation of the SpringSaLaD software, which uses a Langevin dynamics formalism to model multi-molecular interactions with explicit excluded volumes. It permits spatial simulations of combinatorially complex processes such as clustering and polymerization. The approach is an amalgam of kinetic and molecular modeling, in that it derives probabilities of reactions from both coarse structural features of the molecules and macroscopic biochemical parameters such as on and off rate constants, diffusion coefficients and allosteric state transition rates. Currently SpringSaLaD represents molecules as spherical ?sites? connected by linear linkers, modeled as stiff springs. Molecules can diffuse and react either in a rectangular volume or be anchored to a planar patch of membrane.
Our Specific Aims propose to dramatically expand the scope of SpringSaLaD by allowing more realistic representation of structural details and expanding the range of biophysical mechanisms that can be modelled. To better account for the influence of membrane curvature on clustering and the possibility of assemblies that span across thin processes such as filopodia or endocytic invaginations, we will implement methods for Brownian dynamics along curved membranes. We will develop methods to derive the arrangement of spherical sites and linkers from more realistic 3D molecular data, including atomic coordinates. We will develop new optional schemes to better account for the rigidity of molecular structures at this coarse-grained level and, separately, the flexibility of linker domains; the latter will help us represent the influence of intrinsically disordered domains. We will develop statistical and analytical methods to analyze simulation results and build lumped models so as to bridge from this mesoscale to the full cell scale. Finally, we propose to support mechanochemistry by accounting for local force experienced by a site and appropriately altering probabilities for unbinding (i.e. off rates), binding (on-rates) and the tension at membrane surfaces. Ultimately, the SpringSaLaD functionality will be incorporated within the Virtual Cell software system.
We will develop a new biologist friendly technology for modeling and simulation at the scale between molecular and cellular and thus bridge a gap in currently available software. Such computational modeling can provide insights into complex cellular processes that underlie signaling systems and become aberrant in disorders such as kidney failure, autism, cancer and diabetes.
