Different gene regulation strategies revealed by analysis of binding motifs

Abstract

Coordinated regulation of gene expression relies on transcription factors (TFs) binding to specific DNA sites. Our large-scale information-theoretical analysis of > 950 TF-binding motifs demonstrates that prokaryotes and eukaryotes use strikingly different strategies to target TFs to specific genome locations. Although bacterial TFs can recognize a specific DNA site in the genomic background, eukaryotic TFs exhibit widespread, nonfunctional binding and require clustering of sites to achieve specificity. We find support for this mechanism in a range of experimental studies and in our evolutionary analysis of DNA-binding domains. Our systematic characterization of binding motifs provides a quantitative assessment of the differences in transcription regulation in prokaryotes and eukaryotes.

Figure 1. Information theory as applied to DNA-binding motifs

(a) The concepts of minimal information required in theory and in DNA recognition and the consequences of information deficiency, which results in spurious hits. (b) The sequence logos for low- and high- information motifs, and the likelihood of a spurious hit to the motif in a ‘random’ genomic background.

Figure 2. Properties of binding motifs for bacteria, yeast, and multicellular eukaryotes

(a) The bar chart displays the minimum required information content for bacteria, yeast, and multicellular eukaryotes (red), and the mean information content of TF binding motifs (blue) for 98 bacterial [21], 124 yeast [22] and 123 multicellular [20] eukaryotic motifs. The error bars are ± 1 standard deviation for the information content, and for I_min, the error bars represent the variability in that quantity due to the range of genome sizes N. The blue dots in the chart indicate the average information content from several other transcription factor binding motif databases (Table S6). Below each series in the bar chart, we display an example of sequence logo for a binding motif with close to average information content. The chart demonstrates that bacterial transcription factor binding motifs informative enough to make spurious hits to the genomic background unlikely, in constant to yeast and multicellular eukaryotic motifs. (b) The distributions of information content of motifs from the three representative databases cited above. The ranges of required information (I_min) are marked in red. Most bacterial motifs have I > I_min, whereas almost all eukaryotic motifs do not. (c) Average properties of transcription factor binding motifs, the expected number and the spacing between the spurious sites per genome in bacteria, yeast and multicellular eukaryotes.

Figure 3. Membership of PFAM protein domain families, by kingdom

To explore the evolution of DNA-binding domains, we examined the membership of PFAM protein domain families. Each column in (a, b) represents a single PFAM family, and the size of the red or blue bar indicates the proportion of the family’s bacterial and eukaryotic members, respectively. (a), shows the membership of DNA-binding domains, demonstrating that by bacteria and eukaryotes share very few. As a control (b), we plot the composition of PFAM glycolysis and /or gluconeogensis enzyme families, which are shared between kingdoms. In (c), we show a Venn diagram, after removing the weakest 10% of hits to a PFAM family profile.