Old yellow enzyme (NADPH oxidoreductase) in the free and complexed state was thoroughly investigated by the following techniques : absorption, circular dichroism, fluorescence/phosphorescence and electron paramagnetic resonance spectroscopy and fluorescence/phosphorescence decay measurements, applied over a wide range of temperature (7 -293 K). The data obtained were interpreted by comparison with results from similar measurements on free FMN, existing spectral data on isoalloxazine model systems and theoretical data.The results clearly demonstrate the inadequacy of a simple phenolate-FMN donor-acceptor charge-transfer complex to explain the phenomena occurring upon the addition of phenols to old yellow enzyme. Instead it was found that the phenolate anion interferes strongly with an existing tight complex between FMN and the apoprotein, probably an H-bonded structure in which FMN is tautomerized and interacts with an L-chiral center. This is concluded from a separate electronic transition with an origin at 496 nm, thus far not recognized as such, and the circular dichroism observed.The emission is dominated by that of free FMN, although protein-bound FMN seems also to become luminescent in glassy solution at 143 K. A second fluorescence/phosphorescence emission appears upon excitation of both native and complexed old yellow enzyme in the ultraviolet. This emission is quenched by the addition of phenol to the enzyme, shows a large (3000-cm-') blue shift on going to a low-temperature glass and is tentatively assigned to excimers of nucleic acids. Long-wavelength excitation with a synchronously pumped, mode-locked Rhodamine 6-G dye laser revealed a third, extremely weak emission in both native old yellow enzyme and its complexes. It decays with a lifetime of about 3 ns at 143 K. Electron paramagnetic resonance spectra revealed the presence of a low amount of an unpaired spin in old yellow enzyme. Owing to an unusual relaxational behaviour it could only be observed below 15 K and, again, the signal was measured in both the native enzyme and its complexes. Possible assignment and consequences of this observation are discussed.In frozen aqueous solutions of the enzyme-phenolate complex, a phase transition was discovered at which the colour of the complex reverted to that of the native enzyme. Subsequent melting restored the original colour.The observed phenomena and existing literature data lead to the conclusion that the only model from which no apparent inconsistencies emerge is that of a very complicated network of hydrogen-bonded structures in the protein. These involve several, partly unknown, chromophores. Phenols interfere with this network, leading to the formation of the long-wavelength absorption band in old yellow enzyme.The spectral changes induced by the binding of a lowmolecular-weight compound to old yellow enzyme (NADPH oxidoreductase) were initially regarded to be sufficiently mysterious to refer to this unknown compound as 'regreening factor' [I]. After the identification of this 'regreening factor'...