Biological cellular function relies on the coordination of many information processing steps in whichspecific biomolecular complexes are formed. But how can proteins and ligands find their targets inextremely crowded cellular membranes or cytosol? Here, I propose that intracellular signal processingdepends upon the spatiotemporal order of molecules arising from “dynamical sorting”, i.e.no stable structure may exist at any time and yet the molecules might be ordered at all times. Ipropose to investigate the existence and the physicochemical driving forces of such dynamicalsorting mechanisms in selected neuronal synaptic membrane proteins that must be tightly coordinatedduring neurotransmission and recovery. To this end, massive atomistic and coarse-grainedmolecular simulations will be combined with statistical mechanical theory, producing predictions tobe experimentally validated through our collaborators.
The main methodological challenge of this proposal is the infamous sampling problem: Even withimmense computational effort, unbiased molecular dynamics trajectory lengths are at most microsecondsfor the protein systems considered here. This is insufficient to calculate relevant states,probabilities, and rates for these systems. Based on recent mathematical and computationalgroundwork developed by us and others, I propose to develop enhanced sampling methods thatwill yield an effective speedup of at least 4 orders of magnitude over current simulations. This willallow protein-ligand and protein-protein interactions to be sampled efficiently using atomistic models,thus having far reaching impact on the field.