Activating a gene through pioneer transcripts

First up read yesterday's entry on Genomic Organization.

Now that you've done that, let's talk about a paper that appeared in Nature about a month ago. The article is entitled:

Stepwise chromatin remodelling by a cascade of transcription initiation of non-coding RNAs (link)

Superficially you would look at this title and exclaim Wow another function for non-coding RNAs! Well not exactly. It would seem that everyone is going ga-ga over these non-coding RNAs, but if you dig deeper, something else is going on. Note that I'm not saying that the paper is crap, in fact the results are VERY interesting, but you have to keep in mind that this paper is describing is how the act of transcribing non-coding RNA affects genomic organization.

But before we begin, let's dust off our lexicon. Here are some definitions that I did not bring up yesterday. Chromatin can be thought of as the configuration of DNA with its associated proteins, mainly nucleosomes. Chromatin remodelling refers to alterations in the packaging of this DNA so that the tightness and location of nucleosomes have been altered. As I described yesterday, these changes will affect how DNA binding proteins associate with the genome, which in turn modifies what regions are transcribed into RNA.You'll also remember that the theme of that post was that RNA Polymerase II (aka Pol II) and nucleosomes have a strange relationship. In fact Pol II can directly alter the modifications found on histones and can also influence how DNA is bound to its nucleosomes. So there is a constant conversation between chromatin structure and Pol II.

The Nature paper illustrates this principle nicely. It demonstrates how one gene, fbp1, is activated in response to glucose deprivation. Strangely, Pol II plays a big part in initiating gene activation by allowing the chromatin to be remodelled.

You see it would seem that when the gene is "inactive", Pol II transcribes very long RNAs that start well before the fbp1 gene. These "pioneer" transcripts cover a whole section just before and then continue past the gene and end at the normal termination site. These long mRNAs are even polyadenylated at their end. But they are weird. These RNAs are neither spliced nor translated into protein. The transcripts are found at a very low level and seem to be unstable (I'm inferring that the transcripts have a short half-life from some of their gels, but unfortunately the authors don't measure this parameter). When glucose levels are lowered, the long transcripts disappear and instead new shorter RNAs are made by Pol II molecules. These shorter RNAs begin at points closer to the gene's start site but again end at the gene's termination site. The short guys are however much more numerous as compared to the initial pioneer transcripts. Eventually very short transcripts are made. These transcripts start at the "consensus start site" of the fbp1 gene and are not only properly spliced but are translated into protein.

Now here is the cool part, if you genetically modify the yeast genome so that a transcriptional terminator is introduced in front of the fbp1 gene, you not only prematurely truncate these long pioneer transcripts but you prevent the production of all the shorter transcripts. Yes you prevent the stepwise activation of the gene.

So what is happening?

The authors demonstrate that during the generation of these pioneer transcripts, the tightness to which the DNA is bound to its nucleosomes is altered. This change is extremely apparent in regions of the genomic DNA present just before the gene. In some regions the connection is loosened, in other regions it is tightened. Furthermore, these alterations require certain DNA binding proteins(aka transcription factors or TFs) that are known to stimulate the production of fbp1 mRNA. This data suggests the following model:

1) Under high glucose levels Pol II sporadically transcribes specific long non-coding pioneer transcripts that are quickly degraded. During transcription the connection between the DNA and the nucleosome is loosened for very brief periods but this state is quickly reversed.
2) Glucose levels fall, and trough some signaling cascade, certain TFs are activated.
3) Now when Pol II synthesizes these pioneer transcript TFs quickly climb onto the loosened DNA and stabilize the new confirmation. Different parts of the DNA are loosened from the underlying nucleosomes, other regions are tightened.
4) The bound TFs along with the loosened DNA conformations promote Pol II to start synthesizing RNA at these new sites that lie closer to the gene. More RNA and shorter RNAs are made.
5) Due to the increase in Pol II, newer regions of the DNA are loosened up so that finally TFs can bind right at the gene's promoter.
6) TFs at the promoter activate the gene full blast and now fully processed fbp1 mRNA is made.

So here is a perfect example of how RNA Polymerase, transcription factors, DNA binding elements and nucleosomes all talk to eachother inorder to convert a signal (low glucose) into the activation of the fbp1 gene by a stepwise process.

Some questions:
- Are nucleosomes pushed around or are they modified on their histone tails?
- Why aren't the long transcripts spliced? Is there a link between nucleosomes (or even histone modifications) and splicing? (This has been hinted by other results.)
- Why are the long transcripts quickly degraded?

Ref:
Kouji Hirota, Tomoichiro Miyoshi, Kazuto Kugou, Charles S. Hoffman, Takehiko Shibata & Kunihiro Ohta
Stepwise chromatin remodelling by a cascade of transcription initiation of non-coding RNAs
Nature (08) 456:130-134