Preparation, characterization, and chemical properties of the flavin coenzyme analogues 5-deazariboflavin, 5-deazariboflavin 5'-phosphate, and 5-deazariboflavin 5'-diphosphate, 5'leads to5'-adenosine ester.

In order to facilitate interpretation of the deazaisoalloxazine system as a valid mechanistic probe of flavoenzyme catalysis, we have examined some of the fundamental chemical properties of this system. The enzymatic synthesis, on a micromole scale, of the flavin coenzyme analogues 5-deazariboflavin 5'-phosphate (deazaFMN) and 5-deazariboflavin 5'-diphosphate, 5' leads to 5'adenosine ester (deazaFAD) has been achieved. This latter synthesis is accomplished with a partially purified FAD synthetase complex (from Brevibacterium ammoniagenes), containing both phosphorylating and adenylylating activities, allowing direct conversion of the riboflavin analogue to the flavin adenine dinucleotide level. The structure of the reduced deazaflavin resulting from enzymatic and chemical reduction is established as the 1,5-dihydrodeazaflavin by proton magnetic resonance. Similarly, the C-5 position of the deazaflavins is demonstrated to be the locus for hydrogen transfer in deazaflavin redox reactions. Preparation of 1,5-dihydrodeazaflavins by sodium borohydride reduction stabilized them to autoxidation (t 1/2 approximately 40 h, 22 degrees C) although dihydrodeazaflavins are rapidly oxidized by other electron acceptors, including riboflavin, phenazine methosulfate, methylene blue, and dichlorophenolindophenol. Mixtures of oxidized and reduced deazaflavins undergo a rapid two-electron disproportionation (k = 22 M-1 S-1 0 degrees C), and oxidized deazaflavins form transient covalent adducts with nitroalkane anions at pH less than 5. Generalized methods for the synthesis of isotopically labeled flavin and deazaflavin coenzymes and their purification by adsorptive chromatography are given.     

Photoreduction of flavoproteins and other biological compounds catalyzed by deazaflavins

Deazaflavins have been found to act as potent catalysts in the photoreduction of flavoproteins in the presence of EDTA and other "photosubstrates". In distinction to the catalysis brought about by normal flavins which involves dark reaction of the photoreduced flavin catalyst, the mechanism of the catalysis by deazaflavins has been shown to involve unstable, strongly reducing radicals which are generated by photolysis of a preformed covalent dimer. By this new method it is possible to reduce not only flavoproteins but a variety of other redox proteins, including heme proteins and iron-sulfur proteins. By virtue of its great catalytic efficiency, it is possible to employ concentrations of deazaflavin sufficiently low as not to interfere with the spectral evaluation of the reduced proteins obtained.     

Enzyme-catalyzed redox reactions with the flavin analogues 5-deazariboflavin, 5-deazariboflavin 5'-phosphte, and 5-deazariboflavin 5'-diphosphate, 5' leads to 5'-adenosine ester.

The ability of 5-deazaisoalloxazines to substitute for the isoalloxazine (flavin) coenzyme has been examined with several flavoenzymes. Without exception, the deazaflavin is recognized at the active site and undergoes a redox change in the presence of the specific enzyme substrate. Thus, deazariboflavin is reduced catalytically by NADH in the presence of the Beneckea harveyi NAD(P)H:(flavin) oxidoreductase, the reaction proceeding to an equilibrium with an equilibrium constant near unity. This implies an E0 of -0.310 V for the deazariboflavindihydrodeazariboflavin couple, much lower than that for isoalloxazines. With this enzyme, both riboflavin and deazariboflavin show the same stereospecificity with respect to the pyridine nucleotide, and despite a large difference in Vmax for the two, both have the same rate-determining step (hydrogen transfer). Direct transfer of the hydrogen is seen between the nicotinamide and deazariboflavin in both reaction directions. DeazaFMN reconstituted yeast NADPH: (acceptor) oxidoreductase (Old Yellow Enzyme), and deazaFAD reconstituted D-amino acid:O2 oxidoreductase and Aspergillus niger D-glucose O2 oxidoreductase are all reduced by substrate at approximately 10(-5) the rate of holoenzyme; none are reoxidized by oxygen or any of the tested artificial electron acceptors, though deazaFADH-bound to D-amino acid:O2 oxidoreductase is rapidly oxidized by the imino acid product. Direct hydrogen transfer from substrate to deazaflavin has been demonstrated for both deazaFAD-reconstituted oxidases. These data implicate deazaflavins as a unique probe of flavin catalysis, in that any mechanism for the flavin catalysis must account for the deazaflavin reactivity as well.     

Biosynthesis of nucleotides, flavins, and deazaflavins in Methanobacterium thermoautotrophicum

The biosynthesis of deazaflavins, flavins, ribonucleotides, and selected amino acids was studied in Methanobacterium thermoautotrophicum by incorporation of 13C-labeled acetate and pyruvate. 13C enrichments were monitored by 13C and 1H NMR spectroscopy. The biosynthesis of ribonucleotides follows the standard pathway. The xylene ring of riboflavin is formed from two pentose moieties in agreement with studies in yeasts and eubacteria. The pyrimidine ring and the ribityl side chain of the deazaflavin chromophore of coenzyme F420 are derived from the purine nucleotide pool. The phenolic ring and C-5 of the deazaflavin system are supplied by the shikimate pathway. A hypothetical mechanism for the assembly of the deazaflavin chromophore from 5-amino-6-ribitylamino-2,4-(1H,3H)-pyrimidinedione and 4-hydroxyphenyl-pyruvate is proposed.     

Comparison of fungal laccases and redox mediators in oxidation of a nonphenolic lignin model compound.

Several fungal laccases have been compared for the oxidation of a nonphenolic lignin dimer, 1-(3, 4-dimethoxyphenyl)-2-(2-methoxyphenoxy)propan-1,3-diol (I), and a phenolic lignin model compound, phenol red, in the presence of the redox mediators 1-hydroxybenzotriazole (1-HBT) or violuric acid. The oxidation rates of dimer I by the laccases were in the following order: Trametes villosa laccase (TvL) > Pycnoporus cinnabarinus laccase (PcL) > Botrytis cinerea laccase (BcL) > Myceliophthora thermophila laccase (MtL) in the presence of either 1-HBT or violuric acid. The order is the same if the laccases are used at the same molar concentration or added to the same activity (with ABTS [2, 2'-azinobis (3-ethylbenzothiazoline-6-sulfonic acid)] as a substrate). During the oxidation of dimer I, both 1-HBT and violuric acid were to some extent consumed. Their consumption rates also follow the above order of laccases, i.e., TvL > PcL > BcL > MtL. Violuric acid allowed TvL and PcL to oxidize dimer I much faster than 1-HBT, while BcL and violuric acid oxidized dimer I more slowly than BcL and 1-HBT. The oxidation rate of dimer I is dependent upon both kcat and the stability of the laccase. Both 1-HBT and violuric acid inactivated the laccases, violuric acid to a greater extent than 1-HBT. The presence of dimer I or phenol red in the reaction mixture slowed down this inactivation. The inactivation is mainly due to the reaction of the redox mediator free radical with the laccases. We did not find any relationship between the carbohydrate content of the laccases and their inactivation. When the redox potential of the laccases is in the range of 750 to 800 mV, i.e., above that of the redox mediator, it does not affect kcat and the oxidation rate of dimer I.     

Characterization of covalent Ene adduct intermediates in "hydride equivalent" transfers in a dihydropyridine model for NADH reduction reactions

A study of the reactions of an NADH model, 1,4-di(trimethylsilyl)-1,4-dihydropyridine, 7, with a series of α,β-unsaturated cyano and carbonyl compounds has produced the first direct evidence for an obligatory covalent adduct between a dihydropyridine and substrate in a reduction reaction. The reactions were monitored by NMR spectroscopy. In all reactions studied, the covalent adduct was the first new species detected and its decomposition to form products could be observed. Concentrations of adducts were sufficiently high at steady-state that their structures could be determined directly from NMR spectra of the reaction mixtures; adduct structures are those expected from an Ene reaction between 7 and the substrate. This first reaction step results in transfer of the C(4) hydrogen nucleus of 7 to the substrate and formation of a covalent bond between C(2) of the dihydropyridine ring and the substrate α-atom. Discovery of these Ene-adduct intermediates completes the spectrum of mechanisms observed in NADH model reactions to span those with free radical intermediates, no detectable intermediates and now covalent intermediates. The geometry of the transition state for formation of the Ene adduct is compared with those of theoretical transition state models and crystal structures of enzyme-substrate/inhibitor complexes to suggest a relative orientation for the dihydropyridine ring and the substrate in an initial cyclic transition state that is flexible enough to accommodate all observed mechanistic outcomes.     

Structure of flavin adducts with acetylenic substrates. Chemistry of monoamine oxidase and lactate oxidase inhibition.

The photoreaction of flavoquinones (lumiflavin, riboflavin, FMN etc. and their 3-alkylated derivatives) with propargylamine-type acetylenic substrates, R4 -Calpha identical to Cbeta -CgammaHR3 -NR2R1, yields a mixture of two adducts,which result from covalent Calpha fixation of the Cgamma-deprotonated substrate to either position C(4a) or N(5) in the flavin nucleus. The N(5) adduct is a dihydroflavin-5-trimethine-cyanine with very intense (xi greater than 20000 M-1 cm-1) absorption maxima in the region 380-450 nm depending on the R1,R2. This absorption allows recognition of minute amounts of this species of flavocyanine even in complex mixtures. Flavocyanines can be reconverted to starting flavin by base. It spectral properties are identical with those obtained for the pargyline-flavin inhibitor complex from bovine kidney or pig liver monoamine oxidase.     

The peroxidase-dependent activation of butylated hydroxyanisole and butylated hydroxytoluene (BHT) to reactive intermediates. Formation of BHT-quinone methide via a chemical-chemical interaction

The food antioxidants butylated hydroxyanisole (BHA) and butylated hydroxytoluene (BHT) are shown to be metabolized to covalent binding intermediates and various other metabolites by prostaglandin H synthase and horseradish peroxidase. BHA was extensively metabolized by horseradish peroxidase (80% conversion of parent BHA into metabolites) resulting in the formation of three dimeric products. Only two of these dimers were observed in prostaglandin H synthase-catalyzed reactions. In contrast to BHA, BHT proved to be a relatively poor substrate for prostaglandin synthase and horseradish peroxidase, resulting in the formation of a small amount of polar and aqueous metabolites (23% conversion of parent BHT into metabolites). With arachidonic acid as the substrate, prostaglandin H synthase catalyzed the covalent binding of [14C]BHA and [14C]BHT to microsomal protein which was significantly inhibited by indomethacin and glutathione. The covalent binding of BHA and its metabolism to dimeric products were also inhibited by BHT. In contrast, the addition of BHA enhanced the covalent binding of BHT by 400%. Moreover, in the presence of BHA, the formation of the polar and aqueous metabolites of BHT was increased and two additional metabolites, BHT-quinone methide and stilbenequinone, were detected. The increased peroxidase-dependent oxidation of BHT in the presence of BHA is proposed to occur via the direct chemical interaction of BHA phenoxyl radical with BHT or BHT phenoxyl radical. These results suggest a potential role for phenoxyl radicals in the activation of xenobiotic chemicals to toxic metabolites.     

Nucleophilic addition reactions of free and enzyme-bound deazaflavin.

DeazaFMN-containing glycolate oxidase has been prepared and shown to catalyze the stereospecific transfer of the alpha-hydrogen from substrate to enzyme-bound deazaFMN. The reaction of sulfite, cyanide, and hydroxylamine with several deazaflavin-containing enzymes (glycolate oxidase, D-amino acid oxidase, glucose oxidase, N-methylglutamate synthetase) and free deazaFMN has been examined. All the deazaflavin systems tested form reversible 1:1 complexes with sulfite and cyanide. The pH dependence of the reaction of free deazaFMN with cyanide indicates that cyanide anion is the reacting nucleophile. Hydroxylamine complexes are formed with deazaFMN glycolate oxidase and deazaFAD glucose oxidase. The effectiveness of the various nucleophilic reagents in complex formation decreases in the following order: sulfite greater than cyanide greater than hydroxylamine. The relative stability observed for the sulfite and cyanide complexes formed with various deazaflavin systems (glycolate oxidase greater than D-amino acid oxidase greater than free deazaFMN) follows the same trend observed for the stability of the sulfite complexes formed with the corresponding flavin system. A correlation is also observed between the reduction potential (E'o) of the deazaflavin system (glycolate oxidase (- 170 mV) greater than D-amino acid oxidase (-240 mV) greater than free deazaFMN (-178 mV) and the stability of the deazaflavin-nucleophile complexes. The following evidence indicates that deazaflavin systems are generally more susceptible toward nucleophilic attack than corresponding flavin system: (a) with the exception of glucose oxidase, the dissociation constants for the deazaflavin-sulfite complexes are at least 1 order of magnitude less than the corresponding flavin sulfite complexes; (b) the least reactive nucleophile, hydroxylamine, does not form a complex with any of the flavin systems. In the case of cyanide, a complex is formed only with native glycolate oxidase, which is the flavin-containing system most susceptible to attack by the more reactive sulfite. Formation of the various (deaza)flavin-nucleophile complexes is characterized by a bleaching of the longer wavelength absorption band of the chromophore and increases in absorption below the isosbestic point of the reaction in the near-ultraviolet region of the spectrum. These results are consistent with the formation of covalent adducts via attack of the various nucleophiles at position 5 of (deaza)flavin. The reaction with cyanide provides the first example of a reversible addition of carbanion to enzyme-bound (deaza)flavin.     

Screening of potential substrates or inhibitors of cytochrome P450 17A1 (CYP17A1) by electrochemical methods

The electrochemical redox reactions of the recombinant form of human cytochrome P450 17A1 (CYP17A1) were investigated. The hemoprotein was immobilized on the electrode modified with a biocompatible nanocomposite material based on the membrane-like synthetic surfactant didodecyldimethylammonium bromide (DDAB) and gold nanoparticles. Analytical characteristics of DDAB/Au/CYP17A1 electrodes were investigated using cyclic voltammetry, square wave voltammetry, and differential pulse voltammetry. Analysis of electrochemical behavior of cytochrome P450 17A1 was performed in the presence of a natural substrate, pregnenolone (1), known inhibitor, ketoconazole (2), and in the presence of synthetic derivatives of pregnenolone: acetyl pregnenolone (3), cyclopregnenolone (4), and tetrabromo-pregnenolone (5). Ketoconazole, the azole inhibitor of cytochromes P450, blocked catalytic current in the presence of the substrate, pregnenolone (1). Compounds 3–5 did not demonstrate substrate properties towards the electrode/CYP17A1 system. Compound 3 did not influence catalytic activity with pregnenolone, whereas compounds 4 and 5 demonstrated some inhibitory activity. Thus, electrochemical redox reactions of CYP17A1 may serve as an adequate substitution of the reconstituted system, which requires additional redox partners for manifestation of catalytic activity of hemoproteins of the cytochrome P450 superfamily.     

Kinetic comparison of reduction and intramolecular electron transfer in milk xanthine oxidase and chicken liver xanthine dehydrogenase by laser flash photolysis

A comparative study using laser flash photolysis of the kinetics of reduction and intramolecular electron transfer among the redox centers of chicken liver xanthine dehydrogenase and of bovine milk xanthine oxidase is described. The photogenerated reductant, 5-deazariboflavin semiquinone, reacts with the dehydrogenase (presumably at the Mo center) in a second-order manner, with a rate constant (k = 6 x 10(7) M-1 s-1) similar to that observed with the oxidase [k = 3 x 10(7) M-1 s-1; Bhattacharyya et al. (1983) Biochemistry 22, 5270-5279]. In the case of the dehydrogenase, neutral FAD radical formation is found to occur by intramolecular electron transfer (kobs = 1600 s-1), presumably from the Mo center, whereas with the oxidase the flavin radical forms via a bimolecular process involving direct reduction by the deazaflavin semiquinone (k = 2 x 10(8) M-1 s-1). Biphasic rates of Fe/S center reduction are observed with both enzymes, which are due to intramolecular electron transfer (kobs approximately 100 s-1 and kobs = 8-11 s-1). Intramolecular oxidation of the FAD radical in each enzyme occurs with a rate constant comparable to that of the rapid phase of Fe/S center reduction. The methylviologen radical, generated by the reaction of the oxidized viologen with 5-deazariboflavin semiquinone, reacts with both the dehydrogenase and the oxidase in a second-order manner (k = 7 x 10(5) M-1 s-1 and 4 x 10(6) M-1 s-1, respectively). Alkylation of the FAD centers results in substantial alterations in the kinetics of the reaction of the viologen radical with the oxidase but not with the dehydrogenase. These results suggest that the viologen radical reacts directly with the FAD center in the oxidase but not in the dehydrogenase, as is the case with the deazaflavin radical. The data support the conclusion that the environments of the FAD centers differ in the two enzymes, which is in accord with other studies addressing this problem from a different perspective [Massey et al. (1989) J. Biol. Chem. 264, 10567-10573]. In contrast, the rate constants for intramolecular electron transfer among the Mo, FAD, and Fe/S centers in the two enzymes (where they can be determined) are quite similar.     

Anti-cooperativity in the two electron oxidation of the S118C disulfide dimer of azurin

We have constructed a disulfide dimer of S118C azurin, in which two copper centers are coupled through a relatively short covalent pathway, and studied its electron transfer properties. The dimer exhibits intriguing mechanistic properties. Due to the strain in the molecule, caused by the limited accessibility of Cys118, anti-cooperativity occurs in the two step oxidation of the dimer with a difference in redox potential between the two half reactions of 33 mV. Upon oxidation, the dimer favours the semi-reduced over the fully oxidized state, as the Cu(I) site in the semi-reduced dimer is able to stabilize the strained dimer complex. The internal electron transfer is surprisingly slow, which could be partially due to an increase in reorganization energy.     

Mechanism(s) for the metabolism of mitoxantrone: electron spin resonance and electrochemical studies

Mitoxantrone has been reported to lack certain properties that characterize quinone containing antitumor agents that undergo enzymatic reduction. These properties are the stimulation of NADPH oxidation, the stimulation of oxygen consumption by microsomes and reductases and, the absence of oxygen free radicals during these reactions. Having these properties implies the presence of a futile redox cycle that requires the generation and the oxidation of a semiquinone free radical. It would follow that if mitoxantrone does not redox cycle in the presence of reductases, then the semiquinone free radical is not produced or, if it is formed, it reacts quickly to form diamagnetic products. However, using liver microsomes, there are reports of the formation of the mitoxantrone free radial anion. In this paper we investigated the mitoxantrone free radical anion generated electrochemically and found that in the presence of oxygen it behaved like other semiquinones. That is, it is oxidized to the parent compound (presumably generating oxygen free radicals), indicating the ability to redox cycle. The reduction potential to generate such free radical in aqueous medium is very high (-0.79 V) when compared to diaziquone (-0.36 V) and Adriamycin (-0.6 V). This suggests that mitoxantrone may not be a substrate for reductases. Under reductive conditions with purified NADPH cytochrome P-450 reductase which very easily reduces diaziquone and Adriamycin, mitoxantrone was not reduced. However, under the same conditions, mitoxantrone was oxidized by the prototype oxidase horseradish peroxidase with the production of a mitoxantrone free radical. This oxidation was accompanied by a drastic change in color and the formation of a dark precipitate. Because microsomes contain a variety of enzymes, we suggest that the previously observed free radical in microsomes is probably due to the oxidation of mitoxantrone. In this theory, this product is probably a polymer which would not require oxygen to be formed. Thus, under oxidative conditions, the mitoxantrone free radical cation will also display impaired redox activity.     

Covalent adducts between tRNA (m5U54)-methyltransferase and RNA substrates.

The interaction of tRNA (m5U54)-methyltransferase (RUMT) with in vitro synthesized unmodified tRNA and a 17-base oligoribonucleotide analog of the T-arm of tRNA in the absence of AdoMet has been investigated. Binary complexes are formed which are isolable on nitrocellulose filters and are composed of noncovalent and covalent complexes in nearly equal amounts. The covalent RUMT-RNA complexes are stable to SDS-PAGE and migrate slower than free enzyme or RNA. Kinetic and thermodynamic constants involved in formation and disruption of noncovalent and covalent binary complexes have been determined and interpreted in the context of steady-state kinetic parameters of the enzyme-catalyzed methylation and 5-H exchange of substrate. The results show that the isolable covalent complex is kinetically incompetent as an intermediate for methylation. Isotope trapping experiments show that when AdoMet is added to preformed binary complex, all bound tRNA is converted to methylated product; thus, the covalent complexes are chemically competent to form products. We have concluded that, after a reversible binary complex is formed, the catalytic thiol adds to the 6-carbon of the U54 of tRNA. The initial adduct leaves the reaction pathway to protonation at carbon 5; the latter can deprotonate and re-enter the pathway to form methylated product. It is speculated that covalent binary RUMT-RNA adducts may serve as depots of enzyme-tRNA complexes primed for methylation, or in unknown roles with RNAs other than tRNA.     

Control of the Generation and Reactions of Free Radicals in Biological Systems by Kinetic and Thermodynamic Factors

Quantifiable redox properties are useful predictors of substrate reactivity in enzyme-catalysed redox reactions of e.g. nitroreductases or peroxidases. Redox properties may also control the rates of electron-transfer reactions between radical products of reduction and oxidation, and endogenous oxidants and reductants respectively. However, in numerous instances prototropic properties of substrate or radical may have profound kinetic consequences, protonation of radicals frequently slowing down electron-transfer reactions. Further, reactions which are thermodynamically extremely unfavourable may still proceed if radical products are removed from the pre-equilibrium efficiently. Thus kinetic considerations often outweigh the purely thermodynamico viewpoint.     

Control of the Generation and Reactions of Free Radicals in Biological Systems by Kinetic and Thermodynamic Factors

Quantifiable redox properties are useful predictors of substrate reactivity in enzyme-catalysed redox reactions of e.g. nitroreductases or peroxidases. Redox properties may also control the rates of electron-transfer reactions between radical products of reduction and oxidation, and endogenous oxidants and reductants respectively. However, in numerous instances prototropic properties of substrate or radical may have profound kinetic consequences, protonation of radicals frequently slowing down electron-transfer reactions. Further, reactions which are thermodynamically extremely unfavourable may still proceed if radical products are removed from the pre-equilibrium efficiently. Thus kinetic considerations often outweigh the purely thermodynamico viewpoint.     

Electrochemical and infrared spectroscopic analysis of the interaction of the Cu(A) domain and cytochrome c(552) from Thermus thermophilus

The hydrophobically guided complex formation between the Cu(A) fragment from Thermus thermophilus ba(3) terminal oxidase and its electron transfer substrate, cytochrome c(552), was investigated electrochemically. In the presence of the purified Cu(A) fragment, a clear downshift of the c(552) redox potential from 171 to 111mV±10mV vs SHE' was found. Interestingly, this potential change fully matches complex formation with this electron acceptor site in other oxidases guided by electrostatic or covalent interactions. Redox induced FTIR difference spectra revealed conformational changes associated with complex formation and indicated the involvement of heme propionates. This article is part of a Special Issue entitled: 17th European Bioenergetics Conference (EBEC 2012).     

Energetic feasibility of hydrogen abstraction and recombination in coenzyme B(12)-dependent diol dehydratase reaction.

Coenzyme B(12) serves as a cofactor for enzymatic radical reactions. The essential steps in all the coenzyme B(12)-dependent rearrangements are two hydrogen abstraction steps: hydrogen abstraction of the adenosyl radical from substrates, and hydrogen back-abstraction (recombination) of a product-derived radical from 5'-deoxyadenosine. The energetic feasibility of these hydrogen abstraction steps in the diol dehyratase reaction was examined by theoretical calculations with a protein-free, simplified model at the B3LYP/6-311G* level of density functional theory. Activation energies for the hydrogen abstraction and recombination with 1,2-propanediol as substrate are 9.0 and 15.1 kcal/mol, respectively, and essentially not affected by coordination of the substrate and the radical intermediate to K+. Since these energies can be considered to be supplied by the substrate-binding energy, the computational results with this simplified model indicate that the hydrogen abstraction and recombination in the coenzyme B(12)-dependent diol dehydratase reaction are energetically feasible.     

The primary photoreactions in the complex cytochrome-P-890-P-760 (bacteriopheophytin760) of Chromatium minutissimum at low redox potentials.

Experimental evidence for electron transfer, photosensitized by bacteriochlorophyll, from cytochrome c to a pigment complex P-760 (involving bacteriopheophytin-760 and also bacteriochlorophyll-800) in the reaction centers of Chromatium minutissimum has been described. This photoreaction occurs between 77 and 293 degrees K at a redox potential of the medium between -250 and -530 mV. Photoreduction of P-760 is accompanied by development of a wide absorption band at 650 nm and of an EPR signal with g=2.0025+/-0.0005 and linewidth of 12.5+/-0.5 G, which are characteristic of the pigment radical anion. It is suggested that the photoreduction of P-760 occurs under the interaction of reduced cytochrome c with the reaction center state P+-890-P--760 which is induced by light. The existence of short-lived state P+-890-P--760 is indicated by the recombination luminescence with activation energy of 0.12 eV and t 1/2 less than or equal to 6 ns. This luminescence is exicted and emitted by bacteriochlorophyll and disappears when P-760 is reduced. At low redox potentials, the flash-induced absorbance changes related to the formation of the carotenoid triplet state with t 1/2 = 6 mus at 20 degreesC are observed. This state is not formed when P-760 is reduced at 293 and 160 degrees K. It is assumed that this state is formed from the reaction center state P+-890---760, which appears to be a primary product of light reaction in the bacterial reaction centers and which is probably identical with the state PF described in recent works.     

Mechanism of action of methanol oxidase, reconstitution of methanol oxidase with 5-deazaflavin, and inactivation of methanol oxidase by cyclopropanol

Methanol oxidase isolated from Hansenula polymorpha contains two distinct flavin cofactors in approximately equal amounts. One has been identified as authentic FAD and the other as a modified form of FAD differing only in the ribityl portion of the ribityldiphosphoadenosine side chain. The significance of this finding is as yet unknown. Previous studies have shown that cyclopropanol irreversibly inactivates methanol oxidase [Mincey, T., Tayrien, G., Mildvan, A. S., & Abeles, R. H. (1980) Proc. Natl. Acad. Sci. U.S.A. 77, 7099-7101]. We have now established that inactivation is accompanied by covalent modification of the flavin cofactor. The stoichiometry of this reaction is 1 mol of cyclopropanol/mol of active flavin. The structure of the covalent adduct was determined by NMR, IR, and UV spectral studies to be an N5,C4a-cyclic 4a,5-dihydroflavin. Reduction of the covalent adduct with NaBH4 at pH 9.0 before removal from the enzyme converted it to the 1-(ribityldiphosphoadenosine)-substituted 4-(3-hydroxypropyl)-2,3-dioxoquinoxaline. Cyclopropyl ring cleavage accompanies inactivation, and covalent bond formation occurs between a methylene carbon of cyclopropanol and N5 of flavin. Methanol oxidase was also reconstituted with 5-deazaflavin adenine dinucleotide (dFAD). Reconstituted enzyme did not catalyze the oxidation of alcohols to the corresponding aldehydes, nor did reduced reconstituted enzyme catalyze the reverse reaction. Incubation of reconstituted enzyme with cyclopropanol resulted in an absorbance decrease at 399 nm, but no irreversible covalent modification of the deazaflavin cofactor. A reversible addition complex between cyclopropanol and dFAD is formed. The structure of that complex was not definitively established, but it is likely that it is formed through the addition of cyclopropoxide to C5 of dFAD. The failure of dFAD-reconstituted methanol oxidase to catalyze the oxidation of substrate, as well as the lack of reaction with cyclopropanol, supports a radical mechanism for alcohol oxidation and cyclopropanol inactivation. Methanol oxidase catalyzes the oxidation of cyclopropylcarbinol to the corresponding aldehyde. No ring-opened products were detected. The failure to form ring-opened products has been used as an argument against radical processes [MacInnes, I., Nonhebel, D. C., Orsculik, S. T., & Suckling, C. J. (1982) J. Chem. Soc., Chem. Commun., 121-122]. We present arguments against this interpretation.