Towards understanding multiscale interactions in biomolecular condensate formation using multimodal experimental pipelines and simulations with supercomputers

Since their discovery, biomolecular condensates (BMCs) have been intensely studied, and have recently received even more attention as a result of observations indicating that their incorporation into functional cellular structures called membrane-less organelles can be driven solely by spontaneous, passive liquid-liquid phase separation events determined by multivalent macromolecular interactions; these transient functional droplets in the cell sequester specific molecules and allow diffusion of others, in some cases creating a localized environment that can accelerate the rate of in vivo biochemical reactions by two orders of magnitude. Many recent studies have been motivated to discover the rules responsible for the formation of BMC morphology, which include the influence of nucleotide or amino acid sequences of intrinsically disordered regions in protein and RNA molecules which form them, and indicate that simple, phenomenological models describing their assembly and aging, with no inclusion of atomistic details, are not sufficient to explain their formation and behavior. Therefore, understanding the details responsible for these phenomena requires a multi-pronged approach and multi-scale understanding of assembly, specifically, from atomic to micron scales, and the associated dynamics ranging from picoseconds to seconds. Here we describe our work combining multiscale simulations using large, supercomputing-enabled workflows, with multiple experiments including microscopy, dynamic light scattering, neutron scattering and x-ray scattering, in order to help uncover new details about the processes driving BMC assembly.