The supercoiling levels of plasmid DNA were determined from Escherichia coli which was grown in ways that are known to alter global patterns of gene expression and metabolism. Changes in DNA supercoiling were shown to occur during several types of these nutrient upshifts and downshifts. The most dramatic change in supercoiling was seen in starved cells, in which two populations of differentially relaxed plasmids were shown to coexist. Thus, some changes in the external nutritional environment that cause the cells to reorganize their global metabolism also cause accompanying changes in DNA supercoiling. Results of experiments with dinitrophenol suggested that the observed relaxations were probably not due to reduced pools of ATP. When rifampin was used to release supercoils restrained by RNA polymerase, the cellular topoisomerases responded by removing these new, unrestrained supercoils. We interpret these results as implying that the cellular topological machinery maintains a constant superhelical energy in the DNA except during certain growth transitions, when changes in metabolism and gene expression are accompanied by changes in DNA supercoiling.The negative supercoiling of DNA is essential to the maintenance of normal cell function, especially in such cellular processes as transcription, replication, and recombination (for reviews, see reference 7 and 38). Several Escherichia coli proteins exist that have the ability to alter the level of supercoiling by breaking and rejoining DNA (for a review, see reference 41). Presently, the known enzymes are DNA gyrase (11), which can use ATP to increase negative supercoiling, and topoisomerases I (40) and III (4), which can relax negatively supercoiled DNA. Since supercoiling in vivo is maintained by the opposing actions of these enzymes (5,29,30), their relative activities could be regulated to produce various levels of supercoiling. If such a change in supercoiling were to occur in response to physiological stimuli, it could contribute to the global regulation of cellular processes. In this study we investigated the extent to which the level of DNA supercoiling varies during growth of E. coli, with emphasis on how these changes could be involved in the global regulation of transcription.While the importance of DNA supercoiling in the maintenance of cellular processes has been established, changes in DNA supercoiling which could account for global regulatory responses remain to be demonstrated. There are several examples in which plasmids isolated from bacteria grown under different conditions have been reported to vary in the level of supercoiling, including cells grown at different temperatures (12) and cells grown to the stationary phase (34). It has been shown in additional studies (27,28) that the level of supercoiling of plasmids has some dependence on the local DNA sequence. One special case of control has been reported in which the level of chromosomal supercoiling differed in Salmonella typhimurium growing aerobically and anaerobically because of changes in ...

Transcription of the phage Mu com/mom operon is trans-activated by another phage gene product, C, a site-specific DNA binding protein. To gain insight into the mechanism by which C activates transcription, we carried out footprinting analyses of Escherichia coli RNA polymerase (= RNAP) binding to various com-lacZ fusion plasmids. KMnO4-sensitive sites (diagnostic of the melted regions in open-complexes) and DNase I-sensitive sites were located by primer-extension analysis. The results are summarized as follows: (i) in vivo, in the absence of C, RNAP bound in the wild-type (wt) promoter region at a site designated P2; in vitro DNase I-footprinting showed that P2 extends from -74 to -24 with respect to transcription initiation. This overlaps a known strong C-binding site (at -35 to -54). RNAP bound at P2 appeared to be in an open-complex, as evidenced by the presence of KMnO4-hypersensitive sites. (ii) In contrast, when C was present in vivo, RNAP bound in the wt promoter region at a different site, designated P1, located downstream and partially overlapping P2. RNAP bound at P1 also appeared to be in an open-complex, as evidenced by the presence of KMnO4-hypersensitive sites. (iii) Two C-independent mutants, which initiate transcription at the same position as the wt, were also analyzed. In vivo, in the absence of C, RNAP bound mutant tin7 (contains a T to G substitution at -14) predominantly at P1; in vitro DNase I-footprinting showed that P1 extends from -56 to +21. With mutant tin6 (a 63 base-pair deletion removing P2, as well as part of P1 and the C-binding site from -35 to -54), RNAP bound to P1 independent of C. We conclude that P1 is the 'functional' RNAP binding site for mom-transcription initiation, and that C activates transcription by promoting binding at P1, while blocking binding at P2.