Fundamentally different strategies for transcriptional regulation are revealed by analysis of binding motifs
Correspondence: (Login to view email address)
- Harvard Medical School, Systems Biology, Boston, MA, USA
- MIT, Harvard-MIT Division of Health Sciences and Technology, Cambridge, MA, USA
PDF (2.9 MB)
- Document Type:
- Manuscript
- Date:
- Received 27 December 2008 20:53 UTC; Posted 28 December 2008
- Subjects:
- Genetics & Genomics, Molecular Cell Biology, Bioinformatics, Evolutionary Biology
- Abstract:
To regulate a particular gene, a transcription factor (TF) needs to bind a specific genome location. How is this genome address specified amid the presence of ~106-109 decoy sites? Our analysis of 319 known TF binding motifs clearly demonstrates that prokaryotes and eukaryotes use strikingly different strategies to target TFs to specific genome locations; eukaryotic TFs exhibit widespread nonfunctional binding and require clustering of sites in regulatory regions for specificity.
Discussion
- Votes:
-
2 votes
- Comments:
-
2 comments
Thanks for the feedback, VRF! To answer your questions:
1. We don’t assume that only one gene is regulated by a particular combination of factors. Our calculation assumes that a functional cluster should have enough binding sites in a given stretch that the cluster is unlikely to arise by chance, but makes no assumption about the number of functional clusters in a genome.
2. Yes—we did use the average information content of a site for this calculation, which is ~12 bits, which, if there were no degeneracy in binding, would correspond to a 6 bp binding site.
Your point about the partial occupancy of arrays of sites raises a good question about the relationship between the “addressing” and functionality of clusters of binding sites, which remains to be explored.
Thank you, ZW
- (Login to share with a colleague)
Additional information
- License:
- This document is licensed to the public under the Creative Commons Attribution 3.0 License
- How to cite this document:
-
Wunderlich, Zeba and Mirny, Leonid. Fundamentally different strategies for transcriptional regulation are revealed by analysis of binding motifs. Available from Nature Precedings <http://hdl.handle.net/10101/npre.2008.2688.2> (2008)
- Version info:
-
Other versions of this document in Nature Precedings
Version number Document title Date v1 Posted 22 December 2008 Other versions of this document elsewhere on the web
VR Falkenberg on 30 December 2008 15:12 UTC
Hello, I am very impressed by your approach, and I have to admit the math is a little difficult for me to wrap my head around, but I have some comments from the perspective of a transcription expert…
1. Does the final calculation for multicellular eukaryotes with different binding sites (requiring 16-19) assume only one gene has any given combination of factors regulating it (only one location in genome)?
2. For the calculation for multicellular eukaryotes with the same binding site (requiring 7-9) was the reported “average” consensus sequence length (6 nt) used for this calculation?
I guess my experience with promoter regulation does not indicate this many sites are required for the expression of a given gene, in many cases, there are 16-19 available TFBS, but all of them are not always required to be bound for transcription. The way I thought of it, combinatorial regulation is very essential in that the available array of TFBS allows for transcription in many different cellular backgrounds (ie: in the blood six Sp1 sites are required, in the lungs it takes three Sp1 sites and one e-box—all of which are found in the same regulatory region, but not all of which are required for expression)
In addition, and I guess in support of above, when people do promoter studies, the removal of sites frequently leads to a reduction in transcription rather than an all or nothing phenomenon in functional promoters. I agree with the observation that spurrious TFBS do not typically lead to spurrious transcription (as you say it takes more than one in multicellular eukaryotes) but I am not sure your calculations provide an indication of the required sites because it is possible that evolutionary changes are largely responsible for preventing inappropriate transcription. (ie: when a non-functional site is bound by TFs and leads to aberrant transcription, then an adjacent repressor site is selected for or removal of the site is selected for and that inhibits the aberrant transcription)
I think using the 6nt average for multicellular eukaryotes contributes to the over-estimation. Frequently common TFs such as Sp1 have these short consensus sequences, but I have yet to find a promoter that has no binding sites for longer consensus TFs
HTH—Good luck