Abstract:

We are interested in how asymmetric charge neutralization of DNA by proteins results in DNA bending. We have previously reported electrophoretic phasing experiments utilizing homodimer peptides derived from GCN4, a yeast basic zipper (bZIP) transcription factor. Here we report the results of experiments that examine the importance of peptide sequence context in DNA bending and test the hypothesis that peptide structural asymmetry causes electrophoretic anomalies in the absence of DNA bending. We prepared two new series of bZIP peptides that differed dramatically in overall size, structure, and peptide sequence near the DNA. The magnitude of apparent DNA bending is independent of the structure of the protein. This result reduces the concern that bZIP protein structure causes electrophoretic anomalies in the absence of DNA bending. In all cases, both the magnitude and direction of the apparent DNA bend angle are strongly dependent on the local peptide charge. We attempted to validate independently our results with a minicircle competition binding assay. Binding preferences of severalfold for properly phased circular versus linear DNA templates were predicted. However, no binding preferences were observed. We propose that the minicircle binding assay may be intrinsically insensitive to DNA bending or flexibility induced by the bZIP peptides studied, and we provide a unifying explanation for the discrepancies between the cyclization and electrophoretic experiments.