Seminar by Cecilia Elisabet Lövkvist

Exploring models of Polycomb-based epigenetic memory genome-wide

Cecilia Lövkvist (John Innes Centre, Norwich, United Kingdom)

During multicellular development, a single genome generates many different types of cells. Thereby, different gene expression patterns are introduced and maintained through epigenetic marks. Polycomb-group proteins are necessary for maintaining repressive epigenetic marks to keep the cell identity. 

The use of mathematical modelling to understand epigenetic memory and chromatin mediated gene regulation is increasing. However, most applications focus on maintenance of epigenetic states and the inheritance of the state through DNA replication and cell division. We have proposed a theoretical model for Polycomb target genes where transcription directly antagonizes Polycomb silencing. We now use the model combined with experiments to understand how Polycomb target genes can switch between epigenetic states. 

In both experiments in differentiated mammalian cells and simulations, we delete a Polycomb-group protein, Polycomb Repressive Complex 2 (PRC2), and thereafter restore its activity.  Earlier mathematical models would predict that PRC2 target genes in the deletion would switch to an active epigenetic state and remain in that state when the PRC2 activity is restored. However, our experiments show that some genes switch back to the silenced transcriptional state they have in the wild-type upon PRC2 reintroduction. We thereby have two types of genes, those that switch and remain (irreversible) or those that switch back (reversible).

We observe both irreversible and reversible genes in the simulations if we focus on two features of our theoretical model, the gene activation strength and the PRC2 activity.  Simulations further suggest that the transcription and the levels of repressive epigenetic marks in the PRC2 wild-type and PRC2 deletion observed experimentally, could already indicate if the gene will switch back or not upon PRC2 re-introduction.  This helps us to distinguish genes that are more likely to switch and understand the mechanisms behind epigenetic switching. It also shows that our theoretical model, with transcription antagonizing PRC2, can be extended and validated with experimental data.