A homologous set of plasmids expressing tet, lacY, and melB, genes encoding integral cytoplasmic membrane proteins, and tWlC and ampC, genes encoding proteins for export through the cytoplasmic membrane, was constructed for studying the effects of transcription and translation of such genes on the hypernegative supercoiling of plasmids in Escherichia coli cells deficient in DNA topoisomerase I. The results support the view that intracellular bacterial DNA is anchored to the cytoplasmic membrane at many points through cotranscriptional synthesis of membrane proteins or proteins designated for export across the cytoplasmic membrane; in the latter case, the presence of the signal peptide appears to be unnecessary for cotranscriptional membrane association.The phenomenon of hypernegative supercoiling of plasmids in Escherichia coli topA mutants lacking DNA topoisomerase I was first reported by Pruss (37). The widely used cloning vector pBR322 isolated from topA null mutants was found to exhibit an extremely heterogenous distribution in its linking number, with a large fraction of the topoisomers more than twice as negatively supercoiled as the same plasmid isolated from isogenic topA+ strains. This topAdependent hypernegative supercoiling is plasmid specific: topoisomers of pUC19, a shortened derivative of pBR322, exhibit only minor differences in their linking numbers when isolated from isogenic topA mutant and topA+ strains. Dissection of the pair of plasmids pBR322 and pUC19 led Pruss and Drlica (38) to conclude that transcription of tet, the gene encoding the tetracycline resistance marker in pBR322, is necessary for the hypernegative supercoiling of the plasmid in the absence of DNA topoisomerase I; various deletions within the tet region also show that a functional product of the gene is not necessary for the phenomenon.The findings of Pruss and Drlica (38) provided a key experimental link between transcription and DNA supercoiling. Theoretical considerations on a plausible relation between transcription and DNA mechanics, however, can be traced two decades back. The idea that transcription might require a swivel in the DNA template to facilitate the rotation of the DNA relative to the RNA polymerase was first discussed in the 1960s (29; see also reference 13). With the discovery of E. coli DNA topoisomerase I in 1971, then known as the o protein (52), the possibility that this enzyme is involved in transcription was raised in this connection (53). More recently, it was postulated that a highly negatively supercoiled loop might form in the DNA template when the RNA polymerase is in contact with a templatebound regulatory protein (54).In 1987, Liu and Wang (25) proposed a twin-supercoiled-* Corresponding author.domain model of transcriptional supercoiling to account for all known experimental findings, including both the hypernegative supercoiling phenomenon described above and the observation of Lockshon and Morris (26) that inhibition of DNA gyrase in E. coli leads to the formation of highly positively superc...