Extracting molecular details of the mechanism of promoter function in transcription is a subject of broad interest. We have chosen to look at one of the simplest aspects of this process, namely, the strength of strand pairing in the promoter region. Specifically, the stability of the DNA double helix was determined in the vicinity of the promoter by computing the free energy for strand separation as a function of dinucleotide free energy values taken from the calorimetric measurements of Breslauer et al. The stability of 168 E.coli promoter regions was studied within a window of +/- 250 nucleotides on either side of the transcription start sites. We found that for this set of promoters the -10 region was significantly the least stable. There is no correlation between the free energies and the rates of RNA polymerase-promoter open complex formation measured for 25 promoters. We also compare the free energies of 121 promoter mutations across the -35 and -10 consensus regions with the free energies of the corresponding wild type sequences. These pairwise mutant-wild type comparisons provide a particularly good test since the examined sequences differ only in one nucleotide to that all other sequence dependent effects are fixed. About 80% of the mutations in the -10 region that show increased/reduced promoter activity are less/more stable than the wild types. The observed high free energy peak, as well as the mutation data, stronglY support the conjecture that the instability, or melting properties, of the -10 region play a significant role in promoter function.
