With an oxystat, changes in the pattern of expression of FNR-dependent genes from Escherichia coli were studied as a function of the O-2 tension (pO(2)) in the medium, Expression of all four tested genes was decreased by increasing O-2. However, the pO(2) values that gave rise to half-maximal repression (pO(0.5)) were dependent on the particular promoter and varied between 1 and 5 millibars (1 bar = 10(5) Pa). The pO(0.5), value for the ArcA-regulated succinate dehydrogenase genes was in the same range (pO(0.5) = 4.6 millibars), At these pO(2) values, the cytoplasm can be calculated to be well supplied with O-2 by diffusion. Therefore, intracellular O-2 could provide the signal to FNR, suggesting that there is no need for a signal transfer chain, Genetic inactivation of the enzymes and coenzymes of aerobic respiration had no or limited effects on the pO(0.5) of FNR-regulated genes. Thus, neither the components of aerobic respiration nor their redox state are the primary sites for O-2 sensing, supporting the significance of intracellular O-2. Non-redox-active, structural O-2 analogs like CO, CN-, and N-3(-) could not mimic the effect of O-2 on FNR-regulated genes under anaerobic conditions and did not decrease the inhibitory effect of O-2 under aerobic conditions.