Regulation of the operon encoding ribonucleotide reductase: role of the negative sites in nrd repression.

Expression of the nrd genes was previously shown to be controlled by both positive and negative regulation (C. K. Tuggle and J. A. Fuchs, EMBO J. 5:1077-1085, 1986). Two regions, one located 5' and one located 3' of the nrd promoter (nrdP), were identified as negative regulatory sites since deletion of these sequences increased nrd expression. These regions of DNA have sequence similarities, and a looping mechanism was proposed to explain the requirement for two distinct sites in nrd repression. To investigate the role of these sequences in regulating nrd, a gel electrophoresis assay was used to detect the proteins that bind to the nrd regulatory sites. A protein that bound to restriction fragments containing the negative regulatory sites but not to other DNA fragments was identified in cell extracts and was partially purified. DNase I footprinting experiments showed that the binding protein protects the 5' negative site previously identified in vivo. The 3' negative site also identified in vivo was not required in vitro for high-affinity protein binding to the 5' site, but lower-affinity binding to this site could be detected. Specific binding to the 5' site was found to be elevated approximately 10-fold in crude extracts from thymine-starved cells as compared with that in extracts from unstarved cells. This higher activity was also evident in purified preparations, suggesting that thymine starvation increases the expression of the negative regulatory protein. The finding that a purified protein preparation binds both negative regulatory sites indicates that this preparation contains the nrd repressor protein or proteins. Insertion of 37 base pairs (3.5 helix turns) of DNA at a HpaII site or 35 base pairs (3.3 turns) at a MnlI site between the 5' regulatory sites and nrdP abolished the increase in nrd expression resulting from thymine starvation in vivo, but negative regulation appeared to be less affected than when either negative site was deleted. Insertion of DNA in these constructs was shown not to affect repressor binding in vitro, indicating either that a simple model of DNA looping to bring equivalent operator sites into physical proximity does not explain repression at nrd or that the distance between sites is sufficient that helical turns are of little importance.