Physical determinants of multiphase organisation in multi-component protein/RNA condensates

The formation of biomolecular condensates via phase separation achieves spatiotemporal organisation within the cell by creating intracellular microenvironments that contribute to the regulation of chemical reactions. Internal subcompartmentalisation of biomolecules to form multiphase architectures has been observed in a number of biomolecular condensates, where immiscible liquid-like phases with different compositions coexist within the same compartment. Understanding the biophysical mechanisms that drive the formation of multiphase condensates is highly desirable since these condensates are responsible for mediating many fundamental biological processes. Here, we couple a genetic algorithm to our residue-resolution coarse-grained model, Mpipi, to evolve two-component protein mixtures towards increasing multiphasicity (i.e., demixing into two compositionally distinct liquid-like phases). Given a target protein, our method allows us to predict a partner protein that supports multiphasic compartmentalisation and investigate the underlying thermodynamics of the multiphasicity. We find that the balance of homo- and heterotypic interaction energies between the two components is important for controlling multiphase organisation: a large difference between homotypic interactions of the two components favours demixing, and the strength of heterotypic interactions should be small but comparable to that of the weaker homotypic interactions for multilayered condensates to form. Our approach for designing multilayered condensates achieves this balance of interaction energies by optimising the composition and patterning of the protein sequences. Lastly, since many multiphase condensates in cells contain RNA, we extend our approach to include RNA as a third component and investigate the role of protein-RNA interactions in stabilising condensates and multilayered architectures.