TOUCAN Tutorial - Return to bench

Is the Motif "C1" bound by a novel Transcription Factor ?

In the previous section, we learned how the program MotifSampler can be used to predict over-represented motifs in our set of 52 human promoter sequences, which show a common expression profile upon stimulation with Interleukin-1. We focused on one over-represented motif, termed "C1", and demonstrated that it occurs in 62 % of the sequences. We concluded with a series of questions that obviously come to our minds when thinking about how to proceed.

I would like to show you how we can approach the first 2 of these questions, namely if this sequence motif is actually recognized by a protein, and if the expression of this protein is also inducible by IL-1 treatment. For this purpose, we performed EMSA (Electrophoretic Mobility Shift Assay) analyses, using this C1 motif, and for comparison, the NF-kappaB consensus site as positive control (Mayer et al., 2004). These oligonucleotide probes were radioactively labelled and analyzed for their binding efficiency to nuclear extracts from IL-1 treated and control HUVEC. The results are displayed in the image below.

It can be clearly seen that a band was detected that bound to the double-stranded C1 oligonucleotide, showing a moderate but significant (1.8 fold) induction upon IL-1 stimulation. Please note that we also used oligonucleotides derived from the consensus site of the transcription factor FREAC7, also known as FOXL1 (Pierrou et al., 1994), which appeared in one of the predicted "modules" in this set of sequences (see section "ModuleSearcher"), but no binding was detected with the FREAC7 site. This, in fact, is not surprising as we realized later that our initial microarray data indicated a very low, if not undetectable, expression of this gene in endothelial cells.
To test for specificity, we generated a mutant C1 oligonucleotide by substituting those residues that were highly conserved in the "Logos" representation of the motif. Binding of the protein to the C1 element was inhibited by an excess of wild-type, but not of mutant oligonucleotide, demonstrating the specificity of binding. Lanes 10 and 11 show binding of NF-kappaB used as positive control. Fold induction was determined by densitometric quantification of the specific bands.

Obviously, two important questions remain to be addressed and will be the focus of future investigations:

What is the nature of this protein ?
Can it be demonstrated that predicted target promoters indeed are regulated by this factor ?


Conclusion

In this tutorial, we tried to demonstrate some of the powerful components which are integrated in the TOUCAN package for the analysis of regulatory patterns in clusters of co-expressing genes. We showed the procedures which are needed to predict known transcription factor binding sites, to evaluate their significance, to extract potential functional combinations of sites, and finally to predict "de novo" over-represented sequence patterns. The last point also strongly affirmed the concept of a direct inter-play between the experimental plan in the "wet-lab" and bioinformatics analyses.

Previous < > Next