Molecular determinants of condensate assembly in heterogeneous environments - Project Summary Biomolecules, including proteins, nucleic acids, and metabolites, must first colocalize in order for biochemical reactions to take place. As a result, spatially organizing biomolecules inside of living cells is a crucial require- ment for a wide variety of biological processes. Whereas traditional organelles utilize membranes to accomplish this task, more recently discovered biomolecular condensates instead achieve spatial organization via phase separation—a thermodynamically driven process that causes immiscible fluids to demix spontaneously. Over the past decade, an increasingly large number of biomolecular condensates have been discovered and implicated in essential biological functions including gene transcription, protein synthesis, and cell signaling. Biomolecular condensates are also believed to play roles in numerous pathological conditions, including neurodegenerative diseases and viral replication. Understanding the biophysical mechanisms that control the stability, molecular compositions, and assembly pathways of intracellular biomolecular condensates is therefore of critical impor- tance. However, fundamental questions regarding the regulation of these structures—particularly in the hetero- geneous intracellular environment—remain unanswered. My research program uses theoretical and computational methods to uncover the molecular determinants of biomolecular condensate compositional control and assembly mechanisms. Our approach to these fundamen- tal biophysical problems leverages our unique expertise in combining molecular simulation, optimization theory, and machine-learning strategies, and builds on our recent successes developing predictive theories of multi- component phase separation and self-assembly kinetics. Over the next five years, we will focus our efforts on answering two key biophysical questions underlying intracellular condensate assembly. First, to understand how protein/RNA condensates can perform distinct biological functions in heterogeneous environments, we are de- veloping computational models to map mixtures of protein/RNA sequences to distinct condensate compositions. Our innovative approach combines data-driven machine-learning models with optimization techniques, which will ultimately allow us to design protein and RNA sequences to rationally manipulate condensate-based spatial organization in vivo. Second, to understand the mechanisms by which certain nuclear condensates assemble at specific locations on chromatin, we are developing simulation models to predict the nucleation pathways of chromatin-associated condensates. Our simulation strategy will provide essential insights and quantitative pre- dictions of how histone modifications and transcription factors regulate the assembly of chromatin-associated condensates. Our approaches to these questions emphasize experimentally testable predictions, and we are working closely with experimental collaborators to validate the results of our models. In this way, our theoretical and computational studies will advance the state-of-the-art in biomolecular condensate research, and also pave with way for therapeutic applications based on manipulating this form of intracellular spatial organization.
