Enhancer Journal Club: Polycomb potentiates meis2 activation in midbrain by mediating interaction of the promoter...
Much of the tissue specific gene-expression required for development appears to be driven by cis-regulatory elements like enhancers, but I think the field has a long way to go in understanding how these elements work mechanistically. One nice paper on this subject has recently been published by Haruhiko Koseki's lab out of from RIKEN. The abstract is as follows:
Polycomb-group (PcG) proteins mediate repression of developmental regulators in a reversible manner, contributing to their spatiotemporally regulated expression. However, it is poorly understood how PcG-repressed genes are activated by developmental cues. Here, we used the mouse Meis2 gene as a model to identify a role of a tissue-specific enhancer in removing PcG from the promoter. Meis2 repression in early development depends on binding of RING1B, an essential E3 component of PcG, to its promoter, coupled with its association with another RING1B-binding site (RBS) at the 3' end of the Meis2 gene. During early midbrain development, a midbrain-specific enhancer (MBE) transiently associates with the promoter-RBS, forming a promoter-MBE-RBS tripartite interaction in a RING1-dependent manner. Subsequently, RING1B-bound RBS dissociates from the tripartite complex, leaving promoter-MBE engagement to activate Meis2 expression. This study therefore demonstrates that PcG and/or related factors play a role in Meis2 activation by regulating the topological transition of cis-regulatory elements.
So in short, the authors contend that the Meis2 gene is bound by the polycomb-group protein RING1B, which represses it's expression. It also makes a DNA loop to contact another nearby RING1B site. When the gene needs to be activated during development, a tissue specific enhancer forms a new DNA loop to contact the Meis2 promoter and the contact between the promoter and the downstream RING1B site are lost. I think the most interesting feature of this paper is the isolation of the intermediate complex with both the RING1B site and the enhancer interacting with the promoter at the same time. Without this time point captured, one might assume that the repressive RING1B interaction is lost before the new activating interaction with the enhancer is gained.
Another very interesting feature is that the new interaction appears to depend on RING1B in some way. The implications of this are a little unclear as far as I can tell, as the first interaction (RING1B site - promoter) is also dependent on the presence of RING1B. When the authors knock out RING1B before the intermediate "triplet" interaction is formed, they no longer detect the repressive interaction, the intermediate or the activating interactions. To me this leaves unanswered whether the formation of the new enhancer-promoter interaction is really dependent on the physical presence of RING1B at some of these interacting sites, or whether it is simply dependent on the prior interaction between the promoter and the repressive site. Another very interesting question is at what stage the gene becomes active - is it when the enhancer begins to interact with the promoter in the intermediate complex or only when the repressive site is evicted? This could perhaps be answered by DNA/RNA FISH and might give some insight into the molecular mechanisms that might be involved here.
Overall, it's a very interesting paper which leaves a lot of new questions to be picked apart. I think these are exactly the kind of questions that could be answered with new Zinc-Finger/TAL technology in the next few years. For example, what would happen if you used a Zinc Finger to recruit the enhancer to the repressive complex outside of it's normal developmental complex? Is the interaction itself sufficient for evicting the RING1B site or is there another necessary signal? Just one example of how I think this technology could really help us start to unravel what is really going on in these sorts of systems.