In vivo control of promoter and terminator efficiencies at a distance

In pBR329, the genes providing resistance to ampicillin (beta-lactamase, bla) and chloramphenicol (chloramphenicol acetyl transferase, cat) are encoded on the same strand. The bla gene lies downstream of the cat gene, separated by an intergenic sequence of 414 bp. The transcription starts of the two genes are 1090 bp apart. We have probed, in vivo, the effect on transcription of the bla gene, of the introduction, in front of the cat gene, of a series of synthetic promoters covering a large (over 60-fold) range of efficiency. The rising efficiency of the cat promoter has several important consequences for transcription of the bla gene. First, a strong (up to sevenfold) stimulation of the bla promoter is observed, together with a shift of the main bla transcription start site, 10 bp upstream. Furthermore, the relative efficiencies of the bla transcription terminators are reduced. Finally, because of a lesser relative efficiency of the cat transcription terminators as well, we observe enhanced intrusion into the bla gene of transcripts initiated at the cat promoter, some of them extending to the bla transcription terminator and beyond. The operon-like expression of the cat-bla gene tandem is controlled by the efficiency of the cat terminator, which in turn depends on that of the cat promoter. This demonstrates a direct link between the efficiencies of promoter and terminator. Upon inhibition of bacterial gyrase activity, i.e. relaxation of negative supercoiling action, bla expression increases sharply in pBR329, but remains almost unchanged in a plasmid (pBRGC-1) in which cat is under the control of a 6.5-fold stronger promoter. Therefore, under normal gyrase activity, the stimulation of the bla promoter in pBRGC-1 (relative to pBR329) appears to be linked to topological relaxation of its template in situ, in keeping with earlier in vitro observations. We propose that the relaxed state of pBRGC-1 in situ could be due to the decrease in the plasmid linking number, introduced by the 10-12 RNA polymerases that simultaneously transcribe the cat gene in that plasmid, compared with only one or two in pBR329. We find that the negative superhelical densities of both plasmids are almost identical when extracted from the cell. Therefore gyrase would not correct for the relaxed state of plasmid pBRGC-1 observed in situ.