Sites of predicted stress-induced DNA duplex destabilization occur preferentially at regulatory loci

Abstract

This paper describes a computational method to predict the sites on a DNA molecule where imposed superhelical stresses destabilize the duplex. Several DNA sequences are analyzed in this way, including the pBR322 and ColE1 plasmids, bacteriophage f1, and the polyoma and bovine papilloma virus genomes. Superhelical destabilization in these molecules is predicted to occur at small numbers of discrete sites, most of which are within regulatory regions. The most destabilized sites include the terminator and promoter regions of specific plasmid operons, the LexA binding sites of genes under SOS control, the intergenic control region of bacteriophage f1, and the polyadenylylation sites in eukaryotic viruses. These results demonstrate the existence of close correspondences between sites of predicted superhelical duplex destabilization and specific types of regulatory regions. The use of these correspondences to supplement string-matching techniques in the search for regulatory loci is discussed.