Identification and characterization of cis-regulatory elements that target Polycomb in the mouse genome

Multicellular organisms consist of numerous cell types, each serving a specific function. Remarkably, almost all cells within an organism contain the same genetic information. Nevertheless, each cell type interprets this information differently, resulting in cell type specific gene expression patterns. These expression patterns define cellular function and are acquired upon lineage commitment of a pluripotent cell. Once acquired, these patterns can be stably maintained throughout subsequent cell divisions. For example, upon differentiation of a stem cell pluripotency-associated genes need to be silenced, while lineage-specific genes have to be activated. The maintenance and propagation of these expression patterns is thought to be mediated at least in part via the posttranslational modification of chromatin components. These covalent modifications are deposited by specialized enzymes that modify specific histone. However, while many of the enzymes responsible for establishing these marks have been identified, how they are targeted to specific loci remains unclear.
Polycomb-group (PcG) proteins represent key regulators of gene expression, especially during early development where they play key roles in the stable repression of developmental regulators. They form several complexes that mediate the modification of distinct histones. For example, the PRC2 complex mediates trimethylation of histone H3 at lysine 27 (H3K27me3), a chromatin mark essential for proper development of both flies and mammals. However, despite the importance of this modification, it remains elusive how H3K27me3 is targeted to specific loci. In Drosophila melanogaster, it has been demonstrated that transcription factors (TFs) play a major role in guiding PcG complexes to specific DNA elements, termed Polycomb responsive elements (PREs). Efforts to identify similar DNA elements in mammals have proven less successful, with only a handful of PREs known today. Furthermore, it is unclear whether the correlation between TF binding and PcG recruitment observed in D. melanogaster is indeed reflecting a direct physical interaction or rather an indirect crosstalk involving other factors.
In this study, we aimed to investigate the mechanisms that facilitate PRC2 recruitment and deposition of its associated H3K27me3 mark in mammals. We hypothesized that recruitment of PcG complexes to specific loci is encoded within the target DNA sequence either in the form of TF binding sites or other sequence queues. To test this, we employed a reductionist approach and inserted a set of endogenous PRC2 targets in mouse embryonic stem (ES) cells into a defined ectopic locus. We then examined whether these ectopically inserted DNA sequences could recapitulate the H3K27me3 levels observed at endogenous loci. Indeed, all of the tested elements were able to reconstitute endogenous PRC2 and H3K27me3 patterns. Further dissection of these elements revealed that DNA sequences rich in CpG dinucleotides and as short as 220 bp are sufficient to establish an H3K27me3 domain. Furthermore, we found that cell-type specific recruitment is determined by the transcriptional state of the target locus. In particular, transcriptional activity regulated by TF binding to a proximal cis-regulatory element can efficiently block the acquisition of H3K27me3. Finally, by systematically mutating the identified recruiter elements we demonstrate that DNA methylation directly prevents the recruitment of H3K27me3 to the underlying DNA sequence.
Taken together, we propose a model whereby PRC2 recruitment and H3K27me3 deposition defines a default chromatin signature at transcriptionally inactive and unmethylated genomic regions. Furthermore, we show that TFs are involved in the recruitment of PRC2 by controlling the transcriptional activity of the target locus. This study therefore provides novel insights into the relationship between different gene regulatory mechanisms and broadens our understanding of the crosstalk between TFs and epigenetic modifications.