The H3K9me3 heterochromatin integrity and function are sustained by H3K9me3 methyltransferases-HP1 dependencies [TT-Seq]

H3K9me3-heterochromatin, compacted by HP1 proteins, restricts transcription factor binding and functions as a barrier to cell fate changes in development and reprogramming1,2. To investigate principles underlying H3K9me3-heterochromatin maintenance in mammalian cells, we applied degron engineering3,4 on mouse ESCs to conditionally deplete all three H3K9me3-methyltransferases within one hour and found an unexpectedly dynamic nature of heterochromatin maintenance. Following H3K9me3 methyltransferases degradation, HP1ß dissociates from heterochromatin within three hours, and transcription factors quickly access the de-compacted heterochromatin, activating transcription of their targets, which then promotes H3K9me3 decay. Both passive dilution and active removal contribute to the H3K9me3 decay. Mathematical modeling reveals distinct H3K9me3 heterochromatin states with diverse chromatin features which influence initial H3K9me3 levels and the stability of H3K9me3 domains, thereby expanding our understanding of distinct heterochromatin domains. Rampant transposable element de-repression interferes with transcription of poised early developmental and signaling genes outside H3K9me3-domains, highlighting the intricate roles of dynamic H3K9me3-heterochromatin maintenance in safeguarding developmental programs. Our findings deconvolve the hierarchical relationships between H3K9me3 marks, methyltransferases, HP1 and transcription factors to reveal principles of H3K9me3-heterochromatin maintenance and how its remodeling enables transcription and cell fate change. Overall design: A time-course experiment in which histone lysine 9 methyltransferases are ablated in mESCs and the expression of genes, repeats, and other transcriptional units is examined genome-wide by nascent RNA TT-seq. The time-course runs 48hrs with collection points at 0hrs, 3hrs, 12hrs, 24hrs, and 48hrs. Three biological repeats were sequenced in two separate sequencing runs (of the same libraries) for additional genome coverage at each time point. The 0hrs time point is taken to be the baseline against which decay is measured. A S. cerevisiae total RNA spike-in is employed to properly adjust sample signal between different time points.