Previous isolation and analysis of E. coli RNA polymerase (EC 2.7.7.6) binding sites on X DNA had demonstrated the existence of a sigma-dependent process of recognition of A-T-rich DNA sequences. We have now extended this finding to T5 and T7 DNA and have provided evidence for the double-strandedness of the isolated binding sites. The possible equation of these sites to the genetically defined promoters is discussed.We have reported (1) the isolation of Escherichia coli RNA lpolymerase binding sites on X DNA by taking advantage of their resistance to nuclease digestion. Their nucleotide composition was enriched in A-T (57%) when a low polymerase-to-DNA ratio was used, although at high ratio it was not very different from that of total X DNA (50%). This observation reflected the existence of at least two types of sites (2), the ones with the highest affinity being selected for at low polymeraseto-DNA ratios. Indeed, when using tRNA as a competitor for RNA lolymerase during the digestion step, we observed (3) that, when the binding was done with holoenzyme at 370, a fraction of protected DNA fragments was resistant to displacement by tRNA, as assayed by the filter retention technique (4). This resistant fraction was enriched in A-T. In contrast, no resistance was observed with core enzyme or with holoenzyme incubated at 00. By further purification on acrylamide gels, we were able to resolve two populations of protected DNA fragments (5). The first type of fragments, referred to as peak I, contained 45-52 nucleotide residues, was considerably enriched in A-T (up to 67%), and, most importantly, was observed only when sigma was present during the binding step. The second type of fragments (peak II), containing 7-10 nucleotide residues, showed no dependence on sigma nor was it enriched in A-T. On the basis of its sigma dependence the first population of protected fragments was operationally defined as promoters. The sigma-dependent selection of A-T-rich DNA sequences by RNA polymerase along with an additional temperature requirement in order to form a stable complex, as demonstrated with T7 (6) and X DNA (3), fits well with the idea of a local melting of DNA. (1) The enrichment in A-T observed in the "promoters" (peak-I fragments) decreased with increasing polymerase-to-DNA ratios. This observation had been interpreted in the following way: at high ratios, peak I is contaminated with nonspecific binding sites of average base composition. A]-though of lower affinity than specific ones, these sites still bind tightly enough not to be displaced by DNase at high concentration of polymerase. Therefore, nonspecific sites of average composition should be detectable as peak-I molecules even in the absence of sigma.(2) Although the nucleotide composition of peak-I fragments was consistent with their being double stranded, this point had to be unequivocally established. Because of the high number of T5 promoters (10), making them more easily available in quantities possible to work with, the experiments designed to answ...
