Multivalent Signaling Clusters have Unique Sizes Determined by Membrane Localization and Excluded Volume: the Nephrin/Nck/N-WASP System

Abstract

The formation of molecular clusters, aggregates, liquid droplet phases or granules in cells are all consequences of non-stoichiometric multivalent interactions. The importance of these variably sized structures (which we have termed “pleomorphic molecular ensembles”) in many cell biological mechanisms has become increasingly appreciated. But the biophysical principles underlying their assembly and dynamics are not well understood. Here, we investigate some of these principles with the help of coarse grained modeling and simulations using SpringSaLaD software. We used the Nephrin/Nck/N-WASP system because of its importance in maintaining the actin cytoskeleton in kidney podocytes and because it has been well characterized with in vitro studies, including the demonstration of liquid droplet phases. Nephrin is a membrane bound receptor with three phosphotyrosines on its cytoplasmic domain that may each bind to a single SH2 domain on Nck. Nck additionally has three SH3 domains each of which can bind to any of five proline rich motifs within N-WASP. The model explicitly accounts for excluded volume of each of these domains; binding rates between domains are derived from experiment. We simulated systems with 35 or 280 molecules in volumes of 106 nm3 and 8X106 nm3, respectively. Remarkably, the average cluster size at steady state did not depend on the total number of interacting molecules. It did depend on structural details, especially the length of the linkers between binding domains. In particular, comparing the SpringSaLaD results to those from a non-spatial simulation showed that the excluded volume prevented the development of very large clusters. The cluster size distribution histogram contained several peaks, suggesting preferred stoichiometries could influence downstream signaling. We also showed that confinement of nephrin to the membrane significantly increased the average cluster size. (Supported by NIH grant P41 GM103313).