Phase-Separated Heterochromatin Domains Impart Mechanical Stiffness to the Nucleus

Heterochromatin association with the nuclear periphery is assumed to reflect its transcriptional silencing. However, recent work in mouse rod cells showed that peripheral localization is not required for gene silencing. Moreover, studies from our and other labs revealed that chromatin regions associated with the inner nuclear membrane play an integral role in the mechanical response of the nucleus. Since heterochromatin regions are spatially organized through liquid-liquid phase separation (LLPS), we asked whether phase-separated chromatin domains may have distinct mechanical properties and contribute to the emergent mechanical properties of the nucleus.

We use genetically-tractable S. pombe model to examine how altering heterochromatin formation or composition influences nuclear mechanics. We leverage image reconstruction software to quantitatively measure nuclear envelope fluctuations, a novel force spectroscopy that employs optical tweezers to directly measure viscoelastic properties of isolated nuclei, and microscopy approaches to monitor the dynamics of heterochromatin foci in living cells.

We find that both disrupting histone H3K9 methylation and loss of the histone H3K9me2/3 binding protein Swi6, an HP1 orthologue known to undergo LLPS, compromises nuclear stiffness. Accordingly, loss of the H3K9 demethylase Epe1 that leads to heterochromatin spreading increases nuclear stiffness. Mechanical response of the nucleus is also altered in cells expressing an allele of Swi6 with a mutation in its N-terminal domain critical for phase separation. This mutated protein is defective in transcriptional silencing, but not in dimerization or histone binding, suggesting that this Swi6 allele may uncouple the phase-separating properties of Swi6 from its other biochemical activities. Indeed, we found that cells expressing this Swi6 allele demonstrate altered accumulation of Swi6 in heterochromatic foci and defects in the coalescence of individual heterochromatin domains.

Altogether, these results argue that, beyond transcriptional effects, altering histone modifications has mechanical consequences. More generally, our work highlights that phase separated domains can perform mechanical work.