It has been shown that reduced flavocytochrome b2 not only catalyzes reduction of bromopyruvate [P. Urban, P.M. Alliel and F. Eur. J. Biochem. 134,275 -2811 but also transforms it into pyruvate in a reductive elimination process. The dehydrohalogenation reaction also takes place when oxidized enzyme acts on bromolactate, but the reaction is more difficult to observe under these conditions because of its low efficiency compared to the normal oxidative process. The maximal rates of pyruvate production from bromopyruvate and chloropyruvate differ by a factor of less than 10, whereas elimination from fluoropyruvate cannot be detected. These results support a mechanism in which the dehydrohalogenation reaction takes place from a carbanion intermediate of the normal reductive-oxidative pathway.The mechanism of flavoprotein-catalyzed dehydrogenation reactions has been under active investigation in recent years. For oxidations of the type H -4 -Z H + ,C = Z, the so-called carbanion mechanism is generally accepted, although not proven (for reviews, see [I, 21). The dehydrohalo genation of 3-chlorosubstrates by some dehydrogenases/oxidases constituted one of the first pieces of evidence in favour of the carbanion mechanism. The D-and L-amino-acid oxidases were shown to catalyze the non-oxidative transformation of 3-chloroalanine and 2-amino-3-chlorobutyrate to pyruvate and 2-ketobutyrate respectively [3, 41. Similarly, lactate oxidase from Mycohucterium smegrnatis catalyzed the non-oxidative transformation of 3-chlorolactate to pyruvate [S]. On the basis of available evidence, it was suggested that an a-carbanion was an intermediate common to the pathways of the normal oxidative reaction and of the non-oxidative halide ion elimination [3,5]. Other interpretations, however, could not be ignored underlined that the kinetic evidence did not rule out elimination from an Ered. P complex; in that case its occurrence did not allow deductions concerning the redox mechanism. Subsequently, in rapidreaction studies under anaerobic conditions, lactate oxidase appeared to be first reduced by 3-chlorolactate, then reoxidized before reaching the steady state [9]. Furthermore, 3-chlorolactate was not dehydrohalogenated by baker's yeast flavocytochrome b2 (a dehydrogenase/electron transferase) and kidney L-hydroxyacid oxidase, even though it was normally oxidized by the two enzymes [lo, 1 I]. In those two cases it was proposed that carbanion processing might be much faster than chloride ion elimination. It is, however, clear that one of the important pieces of evidence in favour of a carbanion mechanism was missing for those enzymes.In this paper we describe experiments which demonstrate that flavocytochrome b2 can actually catalyze the dehydro- halogenation of some 3-halogenosubstrates. The elimination is much easier to observe under the transhydrogenation conditions recently described by us [12], starting from reduced enzyme and a halogenopyruvate, than during the normal reaction between oxidized enzyme and the corresponding 3-halogen...