DESIGN PRINCIPLES AND IMPLICATIONS OF NON-CONSENSUS PROTEIN-DNA BINDING

In my talk I will describe the new effect of non-consensus protein-DNA binding in the absence of specific base-pair recognition. This entropy-dominated effect stems from multiple, relatively weak interactions between transcription factors (TFs) and non-specific DNA binding sites enriched with repetitive DNA sequence elements. I will describe a high-throughput assay that enables us to accurately measure the free energy of such non-consensus protein-DNA binding. Strikingly, the measured magnitude of the non-consensus protein-DNA binding effect is of the same order as specific, consensus protein-DNA binding. This suggests that non-consensus protein-DNA binding might play a dominant role in establishing genome-wide TF-DNA recognition specificity in living cells. I will describe two specific examples that we have analyzed. First, we measured TF-DNA binding preferences for six human TFs with thousands of short genomic sequences containing specific binding sites for these TFs embedded in their natural genomic context. Second, I will show that experimentally measured RNA polymerase pausing rate in the E. coli bacteria is dominated by non-consensus protein-DNA binding. Taken together, our findings strongly suggest that some key established concepts for design principles of protein-DNA binding recognition should be reevaluated.