UC Berkeley

Compartmentalization is a theme used throughout all kingdoms of biology to create functionally distinct units within a complex cellular environment. In addition to membrane-bound organelles, compartments can exist in the cell that are not membrane bound, yet are still physically distinct from surrounding space. The cell contains many of these membraneless organelles (nucleoli, stress granules, PML bodies, etc.), which are thought to be formed by liquid-liquid phase separation. We find that in early Drosophila embryos, a distinct chromatin domain (heterochromatin) forms by nucleating multiple Heterochromatin Protein 1a (HP1a) foci that grow individually, then fuse together into the final domain. This formation process is reminiscent of liquid-like fusion of nucleoli, which led us to comprehensively study whether heterochromatin could also be a phase-separated system within the nucleus. We utilized Fluorescence Correlation Spectroscopy (FCS) methods to investigate protein diffusion dynamics and determine if heterochromatin displays characteristics associated with phase separation. We find that, similar to other membraneless organelles, the heterochromatin domain is indeed capable of liquid-like fusion, is selectively permeable, and the hetero-euchromatic interface is a barrier to protein diffusion. Additionally, Drosophila HP1a protein is capable of liquid demixing in vitro and mediates domain formation in vivo. This work is the first to demonstrate that the heterochromatin domain is subject to phase separation principles, which suggests that phase interaction, rather than steric hindrance due to chromatin compaction, defines accessibility of heterochromatic areas. It has been suggested that phase separation could be a general organizing property for many membraneless organelles; therefore this model has broad implications for understanding the mechanism of nuclear and genome organization.