Enhancer Journal Club: Rev-Erbs repress macrophage gene expression by inhibiting enhancer-directed transcription

Update 08/08/2013: I've added a plain english summary of this paper at ScienceGist. Find it here: http://www.sciencegist.com/p/179

Two super interesting papers about enhancers came out in Nature last week from the Glass and Rosenfeld labs at UCSD. I'm not sure what to make of them, so I thought I would start a blog, and ask the internet what it thinks!

So the first of the two papers is by Lam et. al. and comes from the Glass lab. They were looking at two transcriptional repressors (Rev-Erb-α and Rev-Erb-β) and found some surprising effects on enhancers.

Here's the abstract:

Rev-Erb-α and Rev-Erb-β are nuclear receptors that regulate the expression of genes involved in the control of circadian rhythm, metabolism and inflammatory responses. Rev-Erbs function as transcriptional repressors by recruiting nuclear receptor co-repressor (NCoR)-HDAC3 complexes to Rev-Erb response elements in enhancers and promoters of target genes, but the molecular basis for cell-specific programs of repression is not known. Here we present evidence that in mouse macrophages Rev-Erbs regulate target gene expression by inhibiting the functions of distal enhancers that are selected by macrophage-lineage-determining factors, thereby establishing a macrophage-specific program of repression. Remarkably, the repressive functions of Rev-Erbs are associated with their ability to inhibit the transcription of enhancer-derived RNAs (eRNAs). Furthermore, targeted degradation of eRNAs at two enhancers subject to negative regulation by Rev-Erbs resulted in reduced expression of nearby messenger RNAs, suggesting a direct role of these eRNAs in enhancer function. By precisely defining eRNA start sites using a modified form of global run-on sequencing that quantifies nascent 5' ends, we show that transfer of full enhancer activity to a target promoter requires both the sequences mediating transcription-factor binding and the specific sequences encoding the eRNA transcript. These studies provide evidence for a direct role of eRNAs in contributing to enhancer functions and suggest that Rev-Erbs act to suppress gene expression at a distance by repressing eRNA transcription.

Essentially they first map the genomic locations of these two repressors and find that they colocalise primarily with monomethylated histone H3 lysine 4 (H3K4me1 - a known marker of enhancers). When they knock out the proteins in blood cells, they see around 140 genes increasing expression and 70 decreasing (and since these are known repressors, it makes sense that removing them primarily leads to upregulation of genes.) What is perhaps more surprising is that only three of the 141 genes which increase their expression level show Rev-Erbs binding at the promotor, indicating that their binding to enhancers may be regulating other genes. They test a few of these regions in luciferase assays (a standard assay for enhancer activity), to verify that they are bona fide enhancers, and they use GRO-seq (a method for mapping nascent transcription) to show that the regions outside of known genes which are marked by both Rev-Erbs binding and H3K4me1 are transcribed in both directions.

So far so good, there are a few examples of activators/repressors binding to enhancers instead of promotors, and the idea of enhancers being transcribed to produce eRNAs (enhancer RNAs) also been floating around for a few years. What's really interesting is that when they knock out Rev-Erbs they see increased transcription of eRNAs from sites that were previously bound by the repressors. So now they have seen that transcription increases at both the enhancers where the protein was bound and the nearby genes, the obvious question is whether and how transcription at the two sites are linked. One can imagine a few scenarios:

1. No causation: The increased transcription of both genes and eRNAs are both caused by some other effect
2. The act of transcription of the gene induces transcription of the enhancer: If the enhancer moves to contact the gene when the Rev-Erbs is knocked out, the increased proximity to active RNA polymerase could induce the increased transcription of the enhancer
3. The act of transcription of the enhancer induces transcription of the gene: Equally, the removal of the repressor could increase the transcription of the enhancer, so when it is brought close to the promoter of the nearby gene, transcription of that gene is also activated
4. The product of transcription of the enhancer induces transcription of the gene: The RNA produced from the enhancer could somehow be activating the gene.

Of these scenarios, number three is how a lot of people have conceptualised transcription at enhancers for a while. The more interesting option is number four, where it is actually the RNA product of the enhancer which has an active role, not just the fact that the enhancer is transcribed.

To establish that enhancer activity really causes the change in activity at the nearby gene, they use two independent approaches (siRNAs and antisense oligonucleotides) to reduce the eRNAs, and observe that when the eRNA product is not present the activation of the nearby gene is decreased. This nicely establishes that there is a causal relationship at work, and rules out scenario number one. It also makes it unlikely that transcription at the gene inducing transcription of the enhancer is having a major effect (since if this was the case, activity of the gene would not be affected by removing the eRNA product). It would have been nice to rule this out more formally by using siRNAs against the gene and demonstrating that the level of transcription of the enhancer is not affected, but I don't think it's a crucial experiment. What it does not rule out is that transcription of the enhancer may be the important factor, since siRNAs in particular are known to repress the promoters of their targets as well as simply reducing the level of mature RNA product.

Finally, the most interesting experiment of the paper. The authors return to dissecting one particular putative enhancer region in a luciferase assay, where they add different fragments of the Mmp9 enhancer region into a plasmid expressing luciferase at a low level from a minimal promoter. The enhancer region is made up of a central core containing binding sites for a number of different transcription factors, a region which codes for the antisense eRNA and a region which codes for the sense eRNA. The entire region does activate the expression of the luciferase, as expected. Normally, people think of enhancers as functioning by recruiting transcription factors, so one would expect that the core region on it's own (which contains the transcription factor binding sites) would be sufficient to activate the luciferase. However, the authors find that the core region on it's own does not have enhancer activity in their assay. In fact the activity is restored only by the addition of the core region plus the sense eRNA (but not the antisense eRNA). One could argue here that the region which codes for the eRNA actually contains some transcription factor binding sites that we simply aren't able to detect with conventional methods, so finally the authors also show that if you insert the core region plus the reversed sequence of the sense eRNA, you still get no activation. This is important because any transcription factors would bind equally well to the normal or the reversed sequence, but the eRNA which is produced would be completely different. They also use RT-PCR to demonstrate that in the context of the luciferase assay the enhancer region is producing the eRNAs they expect it to be producing.

Taken together, this is a really interesting demonstration that in the case of the two specific enhancers they investigate, the enhancer activity really depends on the actual product of transcription (the eRNA) and not just generally on transcription of the enhancer region in a non sequence specific manner. Of course, as ever, there are some points that this paper does not address. Is this eRNA dependent mechanism really the one which is operating in vivo? The siRNA and antisense oligo experiment makes this very likely, but really the killer experiment only demonstrates the effect in the context of an artificial luciferase assay, so there is at least a possibility that it is the act of transcription and not the eRNA product which is important at the endogenous locus. If it is the real in vivo mechanism, is it a general one or is it more widespread? The effect is only explored in detail at 2 of around 1000 identified enhancer regions, so something completely different could be going on at the other 998.

Most importantly though, in my opinion the paper does not address whether these regions are really enhancers, and not simply new non-coding RNA genes. The activity of an enhancer should require close proximity to its target genes, but this is not demonstrated for any of the enhancers studied in the paper. If it really is the specific eRNA which is activating the luciferase promoter, could you get this effect by expressing the eRNA from a different plasmid, or do they really need to be in conctact? For me, this is the most interesting question, but fortunately this aspect is explored in more detail in the second of these two papers - which I'll be discussing next time!

Paper reference: http://www.nature.com/nature/journal/v498/n7455/full/nature12209.html

Update: Journal club on the second paper is now up!