When it comes to transcription factor binding, DNA methylation has a reputation as a trouble maker. When most people think of DNA methylation, they think of it blocking proteins from being able to bind. But does it deserve the bad rap? Though there are some examples of DNA methylation promoting binding of certain transcription factors (TF), there are more examples of those that are blocked. While we’ve seen systematic analyses of the effects of DNA methylation across many transcription factors in plants, animals have been late to the party. As such, we really don’t know the effect of DNA methylation on most of our precious TFs.
To address this knowledge gap, the lab of Jussi Taipale from the Karolinska Institutet in Sweden undertook the largest ever assessment of DNA methylation on 542 human TFs. The talented team used a method called HT-SELEX (High-Throughput Systematic Evolution of Ligands by EXponential enrichment). HT-SELEX uses an array of DNA binding proteins of interest and floods it with random DNA sequences. The proteins then bind their preferred motifs. The authors adapted this approach by methylating the DNA in a parallel experiment to get each protein’s preference for methylated DNA also. And in an elegant move, they combined the methylated and unmethylated libraries, performed the selection, and then bisulfite treated the output giving a quantified score for the preference of each protein at each CpG. They assessed a library of 542 (nearly half) of all human TFs; here’s what they found:
- About 60% of TFs were affected by methylated CpG sites
- Major classes of TFs (including bHLH, bZIP, and ETS) were blocked by methyl-CpG
- One third of TFs were positively affected by methyl-CpG, called methyl-plus, these were enriched for roles in embryonic/organismal development
- Most methyl-plus TFs were in the homeodomain family, which includes the HOX proteins
- Using x-ray crystallography, the homeodomain proteins bind methyl-CpG via direct hydrophobic interactions with the cytosine methyl-group
The size of the “methyl-plus” TF class identified in this paper is quite surprising. The methyl-plus proteins, such as the homeodomain family, may serve important roles during cellular reprogramming, when DNA methylation acts as a barrier. Given that more than half of TFs are affected, the methylation state of the DNA should be considered whenever TFs are discussed. This work is also a major step in mapping the binding motifs of all human TFs.
For all the details on this work, check out the full paper in Science, May 2017.