68 Deconvoluting the recognition of DNA shape from DNA sequence

Abstract

In previous work (Slattery et al., 2011; Joshi et al., 2007), we described the importance of the 3-D structure of the DNA double helix – DNA shape – in the recognition of DNA binding sites by the Hox family of transcription factors. In particular, we found that local minima in the width of the DNA minor groove create electronegative pockets that are binding sites for amino acids with positively-charged side chains such as arginine. Moreover, our data argued that DNA shape is a consequence of DNA sequence, and thus led to the idea that the recognition of specific DNA sequences by DNA binding proteins is mediated by both base-readout, typically in the major groove, and shape-readout (Rohs et al., 2010). The relationship between DNA shape and DNA sequence leads to a logical loop that is difficult to tease apart: if shape is a consequence of sequence, then is the recognition of a binding site by a transcription factor mediated by the sequence of base pairs or by the resulting shape of the DNA molecule? One prediction of the shape recognition model is that if the shape-detecting amino acids are mutated, the shape of the preferred binding sites would become less important. We tested this prediction by mutating the basic residues of Hox proteins known to insert into narrow regions of the DNA minor groove and carrying out in vitro SELEX-seq experiments. Consistent with the importance of DNA shape recognition by this family of homeodomain proteins, we found that DNA molecules selected to bind these mutants had a much smaller propensity to have local minor groove width minima compared with DNA molecules selected to bind the wild type proteins. Further, we were able to transfer the shape recognition properties of one Hox protein to another Hox protein by introducing residues that bind to narrow minor grooves. Together, these findings argue that the recognition of DNA shape is a key aspect of binding site selection by this family of DNA binding proteins.