Flavins have been recognized as being capable of oneand two-electron transfer processes. They are thought to contribute to oxidative stress through their ability to produce superoxide. Flavoproteins play an important role in soil detoxification processes.
Original Description:
Original Title
The chemical and biological versatility of Riboflavin.pdf
Flavins have been recognized as being capable of oneand two-electron transfer processes. They are thought to contribute to oxidative stress through their ability to produce superoxide. Flavoproteins play an important role in soil detoxification processes.
Flavins have been recognized as being capable of oneand two-electron transfer processes. They are thought to contribute to oxidative stress through their ability to produce superoxide. Flavoproteins play an important role in soil detoxification processes.
Versatility of Riboflavin V. Massey' Department of Biological Chemistry, University of Michigan, Ann Arbor, MI 48 109-0606, U.S.A. Abstract Since their discovery and chemical characteriz- ation in the 193Os, flavins have been recognized as being capable of both one- and two-electron transfer processes, and as playing a pivotal role in coupling the two-electron oxidation of most or- ganic substrates to the one-electron transfers of the respiratory chain. I n addition, they are now known as versatile compounds that can function as electrophiles and nucleophiles, with covalent intermediates of flavin and substrate frequently being involved in catalysis. Flavins are thought to contribute to oxidative stress through their ability to produce superoxide, but at the same time flavins are frequently involved in the reduction of hydro- peroxides, products of oxygen-derived radical reactions. Flavoproteins play an important role in soil detoxification processes via the hydroxylation of many aromatic compounds, and a simple flavo- protein in liver microsomes catalyses many re- actions similar to those carried out by cytochrome P450 enzymes. Flavins are involved in the pro- duction of light in bioluminescent bacteria, and are intimately connected with light-initiated re- actions such as plant phototrapiam and nucleic acid repair processes. Recent reports also link them to programmed cell death. The chemical versatility of flavoproteins is clearly controlled by specific interactions with the proteins with vdhich they are bound. One of the main thrusts of current research is to try to define the nature of these interactions, and to understand in chemical terms the various steps involved in catalysis by flavo- protein enzymes. Introductieamd history Long before riboflavin was recognized as a vitamin and was characterized chemically, some of its more spectacular manifestations drew attention. Key words: electron transfer, FAD, flavins, flavoproteins, FMN. Abbreviation used: ETF, electron-transfer flavoprotein. 'e-mail massey@umich.edu 283 Delivered at King's College, London, on I9 November 1999, and at the University of Leeds, on 22 November I999 VINCENT MASSEY One of these is described in graphic terms [l] in the account of the historic voyage of the HMS Challenger, the first systematic exploration of the world's oceans, begin- in 1874 : 'There was no moon, and although the night was perfectly clear and the stars shone brightly, the lustre of the heavens was fairly eclipsed by that of the sea. The unbroken part of the surface appeared pitch black, but wherever there was the least ripple the whole line hoke into a brilliant crest of clear white light. Near the ship the black interspaces predominated, but as the distance increased 0 2000 Biochemical Society n P) Biochemical Society Transactions (2000) Volume 28, part 4 the glittering ridges looked closer, until toward the horizon, as far as the eye could reach, they seemed to run together and melt into one continuous sea of light. What was described was, of course, the bioluminescence of marine bacteria originating from the enzyme luciferase, acting on reduced flavin and oxygen. I will deal later with the chemistry of reduced flavins and oxygen, but let us begin with a brief history of flavins and their chemical and physical properties. Approximately 110 years ago an English chemist by the name of A. Wynter Blyth reported in the Transactions of the Chemical Society his work on the chemical composition of cows milk [2]. One of the components that he isolated was a bright yellow pigment, which he called lacto- chrome, which was later shown to be the com- pound that wenow know as riboflavin. After Wynter Blyths report, almost half a century passed before any significant new de- velopment occurred with his yellow pigment. Then, in the late 1920s and early 193Os, a tremen- dous flurry of research took place. Yellow pig- ments with bright greenish fluorescence were isolated from a wide variety of sources. Interest in them became intense when the yellow pigment was recognized to be a constituent of the vitamin B complex, and isolation of the vitamin was helped enormously by the realization that the potency of the vitamin was correlated with the green fluor- escence. Some of the foremost chemists of the time, Richard Kuhn in Heidelberg and Paul Karrer in Zurich, engaged in what became a bitter race to determine the structure and prove it by chemical synthesis. They both succeeded almost concurrently [3,4], and the name riboflavin was given to replace the variety of previous names, such as lactoflavin and ovoflavin, which were merely descriptive of the source from which it had been isolated. The name, of course, derives from the ribityl side chain and the yellow colour of the conjugated ring system (Figure 1). Quite concurrently with this chemical work, Otto Warburg in Berlin had been carrying out his pioneering work on the mechanism of biological respiration, the process by which small substrates derived from food are converted by living cells into usable energy, employing molecular oxygen as the final oxidant in a chain of reactions. In the process of these studies, Warburg isolated a yellow protein from yeast which was shown to catalyse the oxidation of another vitamin cofactor, now known as NADPH, using oxygen as the second substrate [S]. Soon after its discovery, Hugo Theorell, a Swedish biochemist, found that if the enzyme was precipitated with ammonium sulphate at a pH around 2, the protein precipitate was white and the supernatant was yellow. Neither the colourless apoprotein nor the yellow flavin could catalyse the oxidation of NADPH ; however, when they were mixed together at the appropriate pH, activity was reconstituted. The yellow pigment isolated from the enzyme was in its spectral and fluorescence properties indistinguishable from riboflavin, yet riboflavin would not reconstitute enzyme activity. Theorell quickly found that the difference was due to the enzyme flavin having a phosphate residue in ester linkage at the terminal hydroxy group of the ribityl side chain [6]. This form is known as flavin mononucleotide, FMN. Theorells work was an important historical development, as it showed the biochemical basis for the necessity of riboflavin as a vitamin, i.e. as a cofactor in enzyme catalysis. I n the intervening years it has of course been shown that most of the Figure I Structures of riboflavin, FMN and FAD Hd OH IroclrYouPnr . Rin9Syrt.m 0 2000 Biochemical Society 284 Chemical and Biological Versatility of Riboflavin water-soluble vitamins function in a similar fashion, as prosthetic groups of coenzymes for a wide variety of enzymes. A few years later, Hans Krebs recognized the existence of another flavo- protein, D-amino acid oxidase [7], and Warburg and Christian, in 1938, showed that this protein could be separated into apoprotein and flavin, in the same way as had been done with the yeast yellow enzyme [8]. Again, enzyme activity could be regained on mixing the apoprotein and the isolated flavin, but not with riboflavin or FMN. The structure of the new flavin was correctly deduced to be that of a condensation product of FMN and AMP, and finally proved in 1954 by Todd and his group in Cambridge by total synthesis [9]. This form is known as flavin adenine dinucleotide, FAD. When I started working in the flavin field some 45 years ago, there were only a handful of flavoprotein enzymes known. Now they number in the hundreds, and new ones are reported every month. Mostly they contain non-covalently bound FAD or FMN, and are specific for binding either of the two flavin forms, whatever Nature initially provided them with. This is readily under- standable from the more than 40 flavoprotein crystal structures now available, which reveal that the majority of the flavin-protein interactions are with the N-10 side chain, i.e. the ribityl side chain of FMN or FAD. Fi gure 2 Artificial flavins that have been used as flavin re- placements in flavoproteins etc. Thus it is a general phenomenon that, pro- vided one can remove the original flavin under conditions that avoid denaturing the protein, most flavoproteins will accept into their active sites chemically modified flavin ring structures, pro- vided that the N-10 side chain is the appropriate one. A large number of chemically modified flavins are available, with different substituents at various positions all around the isoalloxazine ring system, as illustrated in Figure 2. The property of reconstitution with modified flavins has proved to be of enormous value in the study of flavoproteins, since those with chemically reactive substituents, such as sulphur-containing ones, or photoreactive ones, such as 6- and 8- azidoflavins, can be used as probes of the flavin environment at that particular position, i.e. to assess whether the particular position is accessible to solvent or buried in the protein [lo]. I n this respect they serve as valuable reporters of protein or flavin dynamics, especially in conjunction with knowledge from more static crystal structures. Some of them, as I will illustrate later, are also good reporters of the overall charge of the active site. I n addition, there is a beautiful correlation of the redox potential of the flavin and the electron- withdrawing or electron-donating properties of the substituents, making possible determination of mechanism through linear free energy cor- relations with activity, substrate dissociation con- stants and other kinetic and thermodynamic properties (see [l l-141 for recent examples). Biological roles of flavins Because of their chemical versatility, flavins are involved in a host of biological phenomena. They play a central role in aerobic metabolism through their ability to catalyse two-electron dehydro- genations of numerous substrates and to par- ticipate in one-electron transfers to various metal centres through their free radical states. I n this capacity they frequently form parts of multi- redox-centre enzymes, such as the succinate and NADH dehydrogenases, xanthine oxidasel dehydrogenase, cytochrome P450 systems and the more recently recognized nitric oxide synthase. A riboflavin-binding protein is involved in the development of chicken and mammalian fetuses [15,16], and flavoproteins have recently been implicated as playing signal transduction roles in programmed cell death [ 171 and regulation of biological clocks, such as that involved in adjustment to jet lag [18]. They are intimately involved in soil detoxification of aromatic 285 0 2000 Biochemical Society Biochemical Society Transactions (2000) Volume 28, part 4 pollutants [19] and in several light-dependent processes, such as photosynthesis [20] and light- dependent repair of DNA damage - the photo- reduction of DNA dimers [21]. They have also been shown recently to be the blue-light receptors long known to be involved in plant phototropism - the directional response of growing plants towards the source of light [22]. Properties of flavins Before we consider the chemistry of some of these flavin-linked processes, I want to spend a little time describing some of their fundamental properties. I have already mentioned their ability to participate in one-electron transfer reactions, which automatically implies the existence of semi- quinone oxidation states. I n free solution, i.e. when not enzyme-bound, a mixture of oxidized and reduced Aavin very rapidly sets up an equilibrium in which a certain amount of flavin radical is formed (Scheme 1). With free flavins the equilibrium is very much to the left, so that at pH 7 only about 5 % radical is stabilized in an equimolar mixture of oxidized and reduced flavin. The semiquinone can exist in a Scheme I Flavin semiquinone equilibria FI, flavin: ox, oxidized; red, reduced. HH' H'- + H+ Fi gur e 3 Spectra of glucose oxidase in the oxidized, semi- quinone and fully reduced states Data are from [23]. Neutral Setniquinone ... "._ '.., neutral or anionic form, with a pK of - 8.5. On binding to a specific protein, this equilibrium can change dramatically. Some enzymes show essen- tially zero stabilization of semiquinone, while others give almost 100 % stabilization. With such cases, the protein may stabilize the neutral radical species over the whole range of pH values at which the enzyme is stable, i.e. the pK is shifted up significantly from 8.5. I n other cases, it is the semiquinone anion that is stabilized, i.e. the pK is decreased significantly. Fortunately, with some enzymes, of which glucose oxidase was the first example [23], the enzyme shows a pK in the observable range, permitting the identification of the spectral properties of both forms (Figure 3). With such large spectral differences between the various flavin oxidation states, it is clearly possible to monitor the events occurring in catalysis using the flavin itself as a reporter. All flavoprotein reactions involve two separate half- reactions, which can be monitored separately by rapid reaction techniques (Scheme 2). The majority of flavoprotein-reducing sub- strates are dehydrogenated in a two-electron re- duction step. The resulting reduced flavin is then re-oxidized by its oxidizing substrate, either in a two-electron step, as shown in Scheme 2, or in single one-electron steps, in which the flavin semiquinone would be observed as an intermedi- ate. I n some enzymes, molecular oxygen is the physiological substrate. Because of the general reactivity of reduced flavins with 0,, the reductive half-reaction is studied under anaerobic con- ditions by mixing enzyme from one syringe of a stopped-flow instrument with the substrate, AH,, from the other syringe. I n this way, reaction intermediates can be detected and rate constants of individual steps determined. The same thing can be done for the oxidative half-reaction. I n this case the enzyme is pre-reduced by a stoichiometric amount of substrate, or some chemical reductant Scheme 2 Reductive and oxidative half-reactions of flavoproteins FI, flavin; ox, oxidized; red, reduced. See the text for details. AH2Y F'ox VBH2 Chemical and Biological Versatility of Riboflavin such as dithionite, and then mixed in the stopped- flow instrument with the oxidizing substrate, oxygen or whatever it may be (B in Scheme 2). Scheme 3 shows the chemical reactions that occur in the reaction of reduced flavin with oxygen [24-281. The initial reaction is a one-electron reduction of 0, by the reduced flavin to yield a caged radical pair of neutral flavin radical and superoxide. This radical pair now has several alternative routes. I t can collapse into a flavin C4a peroxide, a nucleophile, which on protonation becomes the electrophilic hydroperoxide. The peroxide species may eliminate hydrogen peroxide to yield oxidized flavin, or there may be a second one-electron transfer from the radical pair to give the same products. The flavin peroxide species are involved in hydroxylation reactions, as we will see later. The third alternative route is the dissociation of the radical pair into its components, flavin radical and superoxide. The superoxide produced can react with peroxide to form hydroxyl radicals [29], and with nitric oxide to form peroxynitrite [30]. Production of hydroxyl radicals is commonly believed to be one of the major sources of oxidative stress and tissue damage, largely by reaction with lipids to produce lipid hydroperoxides. Recent work also suggests that peroxynitrite is an active agent in apotosis [31]. Now I would like to return to some examples of the information gained by replacing the native flavin with artificial flavins. A particularly good example is that provided by 8-chloroflavins and their reaction products. The 8-chloro substituent is readily eliminated by sulphur nucleophiles (RS-), yielding 8-SR- flavin and chloride ion [32]. Huge spectral changes accompany these reactions, which serve to make 8-C1-flavins very valuable probes of the protein environment around the flavin. Most 8-SR-flavins have absorption maxima in the 470 nm region, with molar absorption coefficients of approx. 25000 M-'.cm-'. Thus, with enzymes where the dimethylbenzene ring of the flavin is exposed to solvent, those enzymes in which the native flavin is replaced by 8-C1-flavin react rapidly with thio- phenol to give 8-thiophenyl-flavin enzymes, whereas those enzymes in which the flavin is buried in the protein fail to react, even over long periods [33]. Chloride is also displaced from 8-C1- flavins by Na,S, yielding 8-mercaptoflavins. Scheme 3 Reactions of reduced flavin with oxygen R R R ~y---+$OZ)y$O NH - Baeyer-Villiger reactions VI I V H 0 0 '0- 'OH 287 0 2000 Biochemical Society Biochemical Society Transactions (2000) Volume 28, part 4 These flavins exist in several easily distinguishable forms, the neutral 8-SH-flavin, with a spectrum similar to those of 8-SR-flavins (A,,, 450 nm; E 20000 M-' * cm-'), and the 8s- thiolate anion form (Amax 520-530 nm; E 30000 M-lacm-'), with a pKa for the transition of 3.8 [32]. On binding to many flavoproteins (in place of the native flavin), the spectrum of the 8-mercaptoflavin is shifted dra- matically (Amax 590-610 nm; E 30000 M-'*cm-') [33]. This form is due to the protein stabilizing a benzoquinoid resonance form, in which the nega- tive charge is localized in the N(l)C(2) =0 locus of the flavin, as shown in Scheme 4. Most flavoproteins of the oxidase class, which react rapidly with 0, to give H202 as product, stabilize 8-mercaptoflavin in this benzoquinoid resonance form, and subsequent crystal structures of such enzymes have indeed confirmed the exist- ence of positive charge in this region of the bound flavin. Such proteins also stabilize the flavin semiquinone anion, and many also facilitate the formation of flavin N(S)-sulphite adducts. A further correlative of protein stabilization of anionic flavin forms (including that of the reduced flavin) is to increase the redox potential of the bound flavin compared with that in free solution. A dramatic demonstration of the utility of such flavin replacement studies comes from work with xanthine oxidoreductase. This enzyme has been known for many years to exist in two interconvertible forms [34] : xanthine dehydro- genase, which oxidizes xanthine at the expense of reduction of NAD+, and xanthine oxidase, which cannot use NAD+as an electron acceptor, but instead employs molecular oxygen, forming both H202 and superoxide as products [35]. The de- hydrogenase form is converted into the oxidase form by oxidation of specific cysteine residues to cystine residues, and the oxidase form can be reconverted into the dehydrogenase by incubation with thiols such as dithiothreitol [36,37]. The redox potentials of the molybdopterin and iron- sulphur centres of the two enzyme forms are similar, but the flavin potential is lowered sub- stantially in the dehydrogenase form, due prin- cipally to stabilization by the protein of the neutral flavin semiquinone. On replacement of the native FAD by 8-mercapto-FAD, dramatic differences in the spectral properties of the flavin are found between the two forms. In the oxidase form the 8- mercapto-FAD is clearly stabilized as the benzo- quinoid resonance form, while in the dehydro- genase the neutral 8-SH-flavin form is found. By pH titration, the pK, of the mercaptoflavin was found to be - 5.5 in the oxidase form, only slightly perturbed from the free solution pKa of 3.8. However, in the dehydrogenase form, the pK, was increased to - 9.0 [37]. Thus it was clear that, in the oxidase form, the electrostatic potential of the protein around the flavin must be positive, whereas in the dehydrogenase it is probably largely negative. This prediction has been very nicely confirmed by the recent determination of the crystal structures of the two enzyme forms [38]. Catalytic versatility of flavoproteins Flavoprotein enzymes catalyse a large variety of different types of reactions. Many attempts have been made to achieve a rational classification of the different types of flavoproteins, depending on the type of chemical reaction catalysed, the nature of the reducing and oxidizing substrates, the physi- cochemical properties of the enzymes and, more recently, their structural motifs as determined by X-ray crystallography. None of these attempts has been entirely satisfactory. Nevertheless, it is clear that enzymes catalysing similar chemical reactions tend to have common characteristics which are particular to that group. The field is too vast to try to give a comprehensive review here, so I will concentrate on describing some enzyme groups with which I have had experience and where common properties seem to exist. See [39] for a more comprehensive review. Scheme 4 Thiolate and benzoquinoid resonance forms of 8-mercaptoflavins R h x - 530 nm hmax - 600 nm 0 2000 Biochemical Society 288 Chemical and Biological Versatility of Riboflavin Flavoproteins catalysing oxidation of a-hydroxyacids and a-amino acids These enzymes bring about dehydrogenation at the a-carbon atom of the substrate, yielding respectively the 2-0x0 (a-keto) acid or a-imino acid as primary product. The enzymes involved share many common characteristics, although there are also many differences, illustrating the difficulty of making a rational classification of flavoprotein enzymes. I n addition to the similarity of the reactions catalysed, both groups have also been shown to catalyse the anaerobic elimination of chloride from /?-chloro a-amino acids or /?- chloro a-hydroxyacids [40,41], and, in the case of D-aminO acid oxidase and /?-C1-a-aminobutyrate, without any sign of flavin reduction during the reaction [42]. Such reactions, and the charac- terization of a flavin N(5)-glycollyl adduct in the reaction of L-lactate mono-oxygenase with glycol- ate as substrate [43], led to the widespread ac- ceptance of the catalytic reaction proceeding via a carbanion mechanism, in which the primary step involves abstraction of the a-proton of the substrate by an enzyme base, as illustrated in Scheme 5. The anaerobic elimination of chloride could also be explained simply by this mechanism, as shown in Scheme 6. The carbanion mechanism received support from the crystal structures of two representative a-hydroxyacid-oxidizing enzymes, yeast L-lactate dehydrogenase (flavocytochrome b,) [44] and spinach glycolate oxidase [45], which show a striking similarity of amino acid residues sur- rounding the flavin, and a histidine residue which Scheme 6 Chloride elimination from /?-chloro a-amino acids or P-chloro hydroxyacids by a carbanion mechanism 11 +- Scheme 5 Possible carbanion mechanism for the dehydrogenation of a-hydroxyacids X 289 0 2000 Biochemical Society Biochemical Society Transactions (2000) Volume 28, part 4 could serve as the active-site base required for a carbanion mechanism. Remarkably, all the other known a-hydroxyacid-oxidizing enzymes have the same conserved amino acid residues as those of flavocytochrome b, and glycolate oxidase, and rapid-reaction kinetic experiments are consistent with them all proceeding by similar mechanisms. The determination of the crystal structures of the D-aminO acid oxidases of pig kidney [46,47] and the yeast, Rhodotorula gracilis [48], has sparked a re-evaluation of the carbanion mech- anism. Since these crystal structures show that there is no base present in the active site, with this enzyme at least, direct hydride transfer to the flavin is favoured and, by implication, the same mechanism also for the a-hydroxyacid oxidases [46,48]. While this may indeed be true, it is as well to point out that, so far, there has been no satisfactory explanation for the chloride elimin- ations in terms of a hydride transfer mechanism. I t should also be emphasized that the two groups of enzymes differ in many ways. The hydroxyacid enzymes all contain FMN as prosthetic group, while the amino acid oxidases have FAD. The FMN enzymes bind substrate on the flavin si-face, while the FAD enzymes interact with substrate on the opposite re-face [49]. The FMN enzymes have the flavin aromatic ring well protected from solvent, while this portion of the FAD is quite accessible to solvent in D-amino acid oxidase [SO]. Common characteristics are the stabilization of anionic flavin forms, such as the anionic semi- quinone, N( 5)-sulphite adducts and the benzo- quinoid form of 8-mercaptoflavin [SO]. As an illustration of the difficulties in classification, all flavoprotein oxidases, not only those catalysing the oxidation of a-hydroxyacids, share these same characteristics [lo]. But so also does yeast D- lactate dehydrogenase, flavocytochrome b,, which re-oxidizes the reduced flavin by single-electron transfer to a protein-bound haem rather than to 0, [511. Flavoprotein disulphide reductases This is an ever-growing family of enzymes which contain an active-site disulphide in addition to the FAD prosthetic group, and involve a pyridine nucleotide as one substrate and a disulphide or dithiol as the other substrate (or alternatively mercuric ion liganded with thiols in the case of mercuric reductase). They share many structural and mechanistic similarities, and a wealth of information is available (see [S2] for a compre- hensive review). I n some cases, such as lipoyl dehydrogenase, the reaction is freely reversible, and physiologically the enzyme functions as part of an 0x0 acid oxidase multienzyme complex to re- oxidize protein-bound dihydrolipoic acid at the expense of reduction of NAD' to NADH. I n the case of glutathione reductase the physiological reaction (and thermodynamic equilibrium) is the utilization of NADPH to reduce oxidized gluta- thione. Although differing in detail from one enzyme to another, the same basic mechanism appears to apply. The reductive half of the reaction is illustrated in Scheme 7. I n a typical reaction, the reduced pyridine nucleotide binds rapidly and positions itself over the re-face of the flavin, so that hydride transfer from the pro(S) position of NAD(P)H occurs efficiently. The reduced flavin reacts rapidly with the active-site disulphide to form the flavin C4a- cysteinyl adduct and a thiol from the other sulphur of the active-site disulphide. This species, like the reduced flavin, is rarely observed experimentally, except when trapped by alkylation of one of the thiols [53] or by removal of the second thiol by mutagenesis [54]. The next step is the elimination of thiol from the cysteinyl adduct, Scheme 7 Reductive half-reaction of a typical flavoprotein disulphide reductase Em 0 2000 Biochemical Society 290 re-forming Yo Chemical and Biological Versatility of Riboflavin oxidized flavin and resulting in the charge-transfer complex of active-site thiolate and oxidized flavin characteristic of this group of flavoproteins. The oxidative half-reaction of the catalytic cycle does not involve the flavin, but consists of a series of thiol-disulphide interchange reactions with the second substrate, a low-molecular-mass disul- phide such as lipoic acid or oxidized glutathione, or a disulphide-containing protein such as thio- redoxin. I n the case of mercuric reductase this involves an intramolecular reaction with a C- terminal dithiol liganded with mercuric ion [55]. Flavoprotein mono-oxygenases Here again we have a broad group of enzymes which share many properties and mechanistic features. The major common property is use of NADH or NADPH to reduce the enzyme flavin (always FAD) and the formation of a flavin C4a peroxide on reaction of the reduced enzyme with molecular oxygen ; this is the reactive oxygen species that is responsible for oxygenation of the substrate. The flavin peroxide in its protonated form is a potent electrophile and it is this form that is employed in the subgroup of aromatic hydroxyl- ases, where a second hydroxy group is introduced into the aromatic substrate already containing one hydroxy group [56]. On the other hand, the peroxide anion, formed initially in the reaction of the reduced flavin with 0, (see Scheme 3), is a good nucleophile, and is employed by the second subgroup of mono-oxygenases in oxygen-in- sertion reactions, such as those catalysed by bacterial luciferase [57] and cyclohexanone mono- oxygenase [58]. The two subgroups each have distinctive common properties. I n the aromatic hydroxylases there is an exquisite control mech- anism to ensure that NAD(P)H is used only when substrate is present to be hydroxylated. A complex of oxidized enzyme and aromatic substrate is required for rapid reduction of the flavin by NAD(P)H ; differences of as much as 100 000-fold in rate constants have been observed in the absence and presence of substrate. The nucleophilic mono- oxygenases do not display this control mechanism, but have developed an alternative one whereby the C4a peroxide formed on reaction with 0, is somehow stabilized by the protein, to such an extent that it can be isolated by low-temperature chromatography in the case of bacterial luciferase [59], and is reactive only in the presence of the substrate for oxygen transfer. This feature is distinct from the situation with the aromatic hydroxylases, where the hydroperoxide is very 29 I unstable in the absence of the substrate to be hydroxylated. The reaction mechanism of p-hydroxy- benzoate hydroxylase has been investigated in great detail, and is summarized in Scheme 8. The basic mechanism shown appears to be followed by all members of this class [60]. These enzymes are delightful ones to work with experimentally, since practically every species in the catalytic cycle can be distinguished on the basis of characteristic absorbance and fluorescence properties, permit- ting in the case of p-hydroxybenzoate hydroxylase the estimation of the values of each of the rate constants shown in Scheme 8. Thus the step k, results in a long-wavelength-absorbing charge- transfer complex with NADPH as donor and oxidized flavin as acceptor. The step k, involves reduction of the flavin and is accompanied by large changes in absorbance and fluorescence, and results in another distinctive long-wavelength charge-transfer complex, in which reduced flavin is the donor and NADP+ the acceptor. The decay of this characteristic absorbance permits the de- termination of k,, the release of NADP+ to yield the reduced enzyme substrate complex. I t is this species that reacts with 0, to give intermediate (I), a complex of substrate and the flavin C4a-hydro- peroxide, which has a distinctive spectrum with a wavelength maximum at 380 nm, and is formed with an observed pseudo-first-order rate constant directly proportional to the oxygen concentration, yielding the value of k,. I t is from intermediate (I ) that the oxygen transfer reaction to substrate occurs, yielding intermediate (11), a complex of enzyme 4a-hydroxyflavin and the non-aromatic dienone form of the product. With some sub- strates, such as 2,4-dihydroxybenzoate and p- aminobenzoate, this species is sufficiently long- lived to permit characterization of its absorbance properties and determination of the rate constant k,. With these substrates the re-aromatization to give the hydroxylated product is sufficiently slow to determine the value of k,. With p-hydroxy- benzoate as substrate the non-aromatic intermedi- ate is not observed, presumably because it re- aromatizes faster than it is formed. I n this case only the rate constant of the hydroxylation step, k,, can be determined. The final intermediate (111), a complex involving the C4a-hydroxyflavin and product, then undergoes dehydration to give the oxidized flavin enzyme, ready for the next cycle of catalysis. An important finding from structural studies of p-hydroxybenzoate hydroxylase is that the 0 2000 Biochemical Society Biochemical Society Transactions (2000) Volume 28, part 4 flavin exists in two different conformations : one in which the flavin is mostly buried within the protein, and positioned ideally for hydroxylation of the substrate, and the other in which the flavin ring system is swung out by 30-40 towards a solvent-exposed position where it could not hydroxylate the substrate [61,62]. This inherent mobility of the flavin has been confirmed in solution studies by replacement of the native flavin by the photoreactive 6-azido-FAD [63], and appears to be an important component of catalysis, the movement of the flavin to the out position being required for reduction by NADPH and movement back to the in position being required for the hydroxylation reaction [64]. The possibility of this being a general phenomenon is raised by the recent report that the flavin of the related phenol hydroxylase, the reaction mechanism of which is very similar to that of p-hydroxybenzoate hydroxylase [65], is also found crystallographically in two similar buried and exposed conformations. Reductases, dehydrogenases and electron transferases This is a somewhat unsatisfactory categorization, since many enzymes that function in one-electron transfers also dehydrogenate an organic substrate, and similarly ones that reduce olefinic bonds also involve dehydrogenation of the primary reducing substrate, which in both cases is frequently a reduced pyridine nucleotide. There are many different enzymes in this broad grouping; I will limit discussion to a few individual cases. Acyl-CoA dehydrogenases These are widely occurring enzymes involved in the oxidation of fatty acids, which oxidize acyl- CoA thioesters to the corresponding enoyl-CoA esters. The reduced flavoprotein so formed is re- oxidized by successive one-electron transfers to another flavoprotein, the electron-transfer Aavo- protein (ETF), which in mammals is in turn re- Scheme 8 Reaction mechanism of p-hydroxybenzoate hydroxylase E-FAD,, &pH R .. 0 NADPH H OH 2 0 2000 Biochemical Society k2 I R NADP+ H o 292 Chemical and Biological Versatility of Riboflavin -FAD NADPH NADP -FUN \f, oxidized by a third flavoprotein, ETF-ubiquinone reductase [66]. Crystal structures of both mam- malian and bacterial acyl-CoA dehydrogenases are available, and permit an authoritative evaluation of the reaction mechanism, which involves re- moval of the substrate a-proton by an active-site aspartate residue and a concerted hydride transfer from the /?-position to the flavin N-5 position (Scheme 9) [66]. Cytochrome P450 reductase This enzyme is responsible for reducing the haem moiety of the widespread cytochromes of the P450 family. I t is unusual in containing equimolar amounts of FAD and FMN, which were both shown to be stabilized as the neutral semiquinone on reductive titration, but with very different redox potentials [67]. Studies on the enzyme in which the FMN was removed, but leaving the FAD intact [68], and in which the FMN was -FADH2 Scheme 9 Oxidation of substrate by acyl-CoA dehydrogenases I R Edo replaced by a series of artificial FMNs of different redox potentials [69], showed that it is the FAD that is reduced by NADPH. Rapid internal elec- tron transfer occurs between the two flavins, with the equilibrium distribution of flavin redox states being determined by their relative redox potentials [70], in a fashion similar to that which had been found for xanthine oxidase [71]. Reduction of cytochrome P450 is by electron transfer from the fully reduced FMN, and during steady-state turn- over the enzyme cycles predominantly between the one- and three-electron reduced states, with the FAD acting as an electron buffer, as shown in Scheme 10. 'Old Yellow Enzyme' Finally, I want to return to Warburg's 'Old Yellow Enzyme ', whose physiological function in yeast still remains unknown. Despite lack of knowledge of its true function, a lot is known about this enzyme. One of its most characteristic properties is its ability to form beautiful charge- transfer complexes with aromatic and hetero- aromatic compounds containing an ionizable hydroxy group [72]. The enzyme from brewer's bottom yeast, the source used by Warburg, and the similar enzyme from baker's yeast, Succhuro- myces cerevisiue, were shown to be a mixture of isoenzymes [73] derived by subunit association from two separate genes [74]. The expression of the protein from a single gene from brewer's Scheme 10 Catalytic turnover with NADPH-cytochrome P450 reductase - e -FAD --FMNK NADPH Priming 293 Turnover 0 2000 Biochemical Society Biochemical Society Transactions (2000) Volume 28, part 4 bottom yeast (OYEl) permitted the isolation of high-quality crystals and structural determination of the protein [75]. The wave number maximum of the long-wavelength band formed with different p-substituted phenols is correlated well with the Hammett 6p substituent of the phenol [76], and with the one-electron redox potential of the enzyme-bound flavin when the native FMN was replaced with a series of artificial Aavins [77]. These correlations were historically very import- ant in establishing the existence of charge-transfer complexes in flavoproteins. The crystal structure of the enzyme with bound phenol clearly shows the parallel stacking of phenol and flavin required for the observed charge-transfer interaction. The ligand-binding site is largely hydrophobic, but His-191 and Asn-194 form hydrogen bonds with the phenolic oxygen and thus contribute to the lowering of the phenolic pK, and its binding in the phenolate anion form. A breakthrough in determination of the func- tion of Old Yellow Enzyme came from the find- ing that it catalysed efficiently an NADPH- cyclohexenone reductase activity, in which the olefinic linkage, not the carbonyl function, was reduced [74]. I t quickly became evident that the enzyme was capable of reduction of a large number of a,/?-unsaturated aldehydes and ketones, and additionally that it could catalyse dismutation reactions, such as that with cyclohexenone, where the product of anaerobic incubation of the enzyme with cyclohexenone was an equimolar mixture of cyclohexanone and phenol [78] (Scheme 11). The ability of the enzyme to be reduced by a saturated aldehyde or ketone is clearly dependent on the redox potential of the bound flavin, as shown in studies where the native flavin was replaced Scheme I I Cyclohexenone dismutase reaction catalysed by the Ol d Yellow Enzyme E, enzyme, FI, flavin; ox, oxidized; red, reduced. with the high-potential 8-cyano-FMN, which effectively converts the enzyme from an NADPH- enone reductase into an oxygen-dependent de- saturase [79]. The above reactions have been found to be highly stereospecific. The reduction of unsatur- ated carbonyl compounds, such as cinnam- aldehyde, was shown to involve a trans-addition across the double bond, with the P-hydrogen being derived from the pro-R-hydrogen of NADPH [via solvent-protected flavin N-( S)], and the a-hydro- gen being derived from solvent [78]. This stereo- specificity is conserved in the oxidation of satur- ated aldehydes and ketones by 8-CN-FMN-OYE The crystal structure suggested a possible candidate as a carrier for the hydrogen that is placed at the a-carbon atom of the saturated product, i.e. Tyr-196, which is ideally located to serve as a solvent-exchangeable acid for proton- ation of the a-carbon. This was confirmed by changing this residue to phenyalanine, whereby the reactivity of the mutant enzyme with cyclo- hexenone was decreased by a factor of 5 x lo5, and that with cinnamaldehyde by a factor of 1.1 x lo3 [80]. Modelling of cyclohexenone into the active site of OYE, with the carbonyl oxygen hydrogen- bonded to His-191 and Asn-194 in the same way as with phenolates, places the /?-carbon in an ideal position to receive a hydride from the flavin N-5, and the a-carbon in the correct position to receive a proton from Tyr-196 in a trans-addition reaction across the olefinic bond [80]. The role of Tyr-196 as proton donor to the a- carbon has also been demonstrated in a striking fashion in the NADPH-dependent reduction of 1791. Scheme I 2 Reduction of nitrocyclohexene by the Old Yellow Enzyme R 4v0- h R 0 2000 Biochemical Society 294 Chemical and Biological Versatility of Riboflavin nitrocyclohexene. Nitrocyclohexene is a very good substrate for the enzyme, with the nitro group functioning to activate the olefinic bond in an analogous fashion to that with the carbonyl group of aldehydes and ketones. In this case, however, a discrete intermediate, identified by its spectral properties as the nitronate form of nitrocyclo- hexane, is formed and released from the enzyme [81] (Scheme 12). The nitronate is a stable species and in free solution is only slowly protonated at the a-position to give nitrocyclohexane. The NADPH-dependent reduction of nitrocyclo- hexene proceeds just as fast with the Tyr-196 + Phe mutant enzyme as with wild-type enzyme, with the rapid accumulation of the nitronate product. The wild-type enzyme also catalyses efficiently the protonation of the nitronate, but the Tyr-196 + Phe enzyme is incapable of catalysing the conversion, thus confirming convincingly the role of Tyr-196 as an active-site acid in the overall reaction. While the physiological substrate (or sub- strates) of OYE has not yet been identified, it seems very likely that it is an a$-unsaturated compound. Since the cloning of the OYE1 gene from brewer's bottom yeast and of the 0 YE2 and 0 YE3 genes from S. cerevisiae [74,82], a number of related enzymes have been identified from plant and bacterial species. One of these, from Pseudomonas putida, functions as a morphinone reductase [83], and ones from plant sources as a 12-oxophytodienoic acid reductase [84,85]. Both of these reactions involve NAD(P)H-linked re- duction of a,&unsaturated carbonyl compounds. All members of the family contain FMN as the prosthetic group, and all catalyse the reduction of cyclohexenone. References I 2 3 4 5 6 7 8 9 10 I I Thompson, C. W. ( 1877) Voyage of the Challenger, vol. 2, p. 85, Macmillan & Co., London Blyth. A. W. (I 879) J . Chem. SOC. 35,530-539 Kuhn, R, Reinemund, K. and Weygand, F. (I 934) Ber. 67, 1460- I463 Karrer, P., Schopp, K. and Benz, F. (I 935) Helv. Chim. Acta I8,42&429 Warburg, 0. and Christian, W. ( 1933) Biochem. Z. 266, 37741 I Theorell, H. (I 935) Biochem. Z. 275, 344-346 Krebs, H. A. (I 935) Biochem. J . 29, 1620-1 625 Warburg. 0. and Christian, W. (I 938) Biochem. Z. 298, 150-155 Christie, S. M. H., Kenner, G. W. and Todd, A. R. ( I 954) J . Chem. SOC. 46-52 Ghisla, S. and Massey, V. ( 1986) Biochem. J. 239, I - I 2 Eckstein, J . W., Hastings, J . W. and Ghisla, S. ( 1993) Biochemistry 25, 8095-8 I02 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Miller, J . R. and Edmondson, D. E. (I 999) J. Biol. Chem. 274, Ortiz-Maldonado. M., Ballou, D. P. and Massey, V. ( 1999) Biochemistry 38,8 124-8 I37 Yorita, K., Misaki, M., Palfey, 6. A. and Massey, V. (2000) Proc. Natl. Acad. Sci. U.S.A. 97, 248G2485 Ramana Murty, C. V. and Adiga. P. R (I 982) Science 216, 191-193 White, H. B. and Menill, A. H. (I 988) Annu. Rev. Nutr. 8, 279-299 Susin. S. A., Lorenzo, H. K., Zamzami, N., Matzo, I., Snow, 6. E., Brothers G. M., Mangion, J ., J acotot, E., Constantini, P., Loeffler, M. et al. ( I 999) Nature (London) 397,44 I 445 Ishiwaka, T.. Todo, T.. Matsumoto. A,, Tanimuta, T.. Togashi, S. and Ueda, R ( 1999) in Flavins and Flavoproteins I999 (Ghisla, S., Kroneck P.. Macheroux. P. and Sund, H., eds), pp. 3 17-320, Rudolf Weber Agency for Scientific Publications, Berlin Dagley, S. (I 987) Annu. Rev. Microbiol. 4 I, 1-23 Zanetti, G. and Aliverti, A. (I 99 I ) in Chemistry and Biochemistry of Flavoenzymes, vol II (Muller, F., ed.), pp. 305-35 I , CRC Press, Boca Raton J oms, M. S. , Wang, 6. and J ordan, S. P. (I 987) Biochemistry Briggs, W. R., Christie, J . M., Knieb, E. and Salomon, M. ( I 999) in Flavins and Flavoproteins I999 (Ghisla, S., Kroneck P., Macheroux, P. and Sund, H., eds), pp. 299-308, Rudolf Weber Agency for Scientific Publications, Berlin Massey, V. and Palmer, G. ( I 966) Biochemistry 5, Massey. V., Palmer, G. and Ballou, D. P. ( 1973) in Oxidases and Related Systems (King, T. E., Mason, H. S. and Morrison, M., eds). pp. 2543. University Park Press, Baltimore Kemal, C., Chan, T. W. and Bruice. T. C. (I 977) J . Am. Chem. SOC. 99, 7272-7286 Bruice, T. C. (I 984) Isr. J . Chem. 24,566 I Palfey, B. A., Ballou. D. P. and Massey, V. ( 1995) in Reactive Oxygen Species in Biochemistry (Valentine, J. S., Foote, C. S., Liebman. J . and Greenberg, A,, eds), pp. 37-83, Blackie Academic, London Halliwell. B. ( 1995) in Reactive Oxygen Species in Biochemistry (Valentine, J . S.. Foote, C. S., Liebman, J. and Greenberg, A., eds). pp. 3 13-335, Blackie Academic, London Huie, R. E. and Padmaja. S. ( 1993) Free Radical Res. Commun. 18. 195-199 23515-23525 26, 68 10-68 I 6 3 I8 1-3 I89 30 Emster, L. ( 1993) in Active Oxygens, Lipid Peroxides and Antioxidants (Yagi, K., ed.), pp. 1-38, J apan Scientific Societies Press, Yokyo 3 I Cassina, A. M., Hodara, R., Souza, J . M., Thomson, L., Castro, L., Ischiropoulos, H., Freeman, B. A. and Radi, R. (2000) J . Biol. Chern. 275, in the press 32 Moore, E. G., Ghisla, S. and Massey, V. (I 979) J . Biol. Chem. 33 Massey, V., Ghisla, S. and Moore, E. G. (1979) J . Biol. Chem. 34 Batteli, M. G., Lorenzoni, E. and Stirpe, F. (I 973) Biochem. 1. 35 Hille, R. and Nishino, T. ( I 995) FASEB J . 9, 995- I003 36 Saito, T. and Nishino, T. ( 1989) J . Biol. Chern. 264, 37 Hunt, J . and Massey, V. (I 992) J . Biol. Chem. 267, 254, 8 173-8 I78 254,9640-9650 131, 191-198 10015-10022 21479-21485 295 0 2000 Biochemical Society Biochemical Society Transactions (2000) Volume 28, part 4 38 Enroth, C., Eger, 6. T., Pai, E. F., Okamoto, K., Iwasaka, T., Nishino, T., Hori, H. and Nishino, T. (I 999) in Flavins and Flavoproteins I999 (Ghisla, S., Kroneck P., Macheroux, P. and Sund, H., eds), pp. 783-786, Rudolf Weber Agency for Scientific Publications, Berlin 39 Palfey, 6. A. and Massey, V. ( 1996) in Comprehensive Biochemical Catalysis (Sinnott, M., ed.), pp. 83- 154, Academic Press, New York 40 Walsh, C. T., Schonbtunn, A. and Abeles, R H. (I 97 I) J . Biol. Chem. 246, 6855-6866 41 Walsh, C. T., Lockridge, O., Massey, V. and Abeles, R H. (I 973) J . Biol. Chem. 248, 7049-7054 42 Walsh, C. T., Krodel, E., Massey, V. and Abeles, R H. (I 973) J . Biol. Chem. 248, 19461955 43 Ghisla, S. and Massey, V. (I 980) J . Biol. Chem. 255, 44 Xia, Z. X. and Mathews, F. S. ( 1990) J . Mol. Biol. 2 12, 837463 45 Lindqvist, Y. and Bidnden, C. I. (I 989) J. Biol. Chem. 264, 3624-3628 46 Mattevi, A, Vanoni, M. A., Todone, F., Rini, M., Teplyakov, A., Coda, A, Bolognesi, M. and Curti, B. (I 996) Proc. Natl. Acad. Sci. U.S.A. 93,7496-750 I Setoyama, C. and Mium R (I 996) J. Biochem. (Tokyo) 120, 14-17 48 Pollegioni. L, Ghisla, S., Umhau, S., Molla G., Hams, C. M. and Pilone, M. S. (I 999) in Flavins and Flavoprotein I999 (Ghisla, S., Kroneck P., Macheroux, P. and Sund. H. eds), pp. 55 1-558, Rudolf Weber Agency for Scientific Publications, Berlin 49 Manstein, D., Massey, V., Ghida, S. and Pai, E. F. (1988) Biochemistry 27,230C-2305 50 Schopfer, L M., Massey, V. and Claibome, A. (198 I) J. Biol. Chem. 256,7329-7337 5 I Lederer, F. ( I99 I) in Chemistry and Biochemistry of Flavoenzymes (Muller, F., ed.), vol. II, pp. 153-242, CRC Press. Boca Raton 52 Williams, C. H. (I 992) in Chemistry and Biochemistry of Flavoenzymes (Muller, F., ed.), vol. Ill, pp. I2 1-2 I I, CRC Press, Boca Raton 53 Thorpe, C. and Williams, C. H. (I 98 I) Biochemistry 20, I 507- I 5 I 3 54 Miller, S. M., Massey, V., Ballou, D. P., Williams, C. H., Distefano, M. D., Moore, M. J . and Walsh, C. T. ( I 990) Biochemistry 29,283 1-284 I 55 Miller, S. M., Massey, V., Williams. C. H., Ballou, D. P. and Walsh, C. T. (I 99 I) Biochemistry 30,260C-26 I2 56 Entsch, B., Ballou, D. P. and Massey, V. (I 976) J . Biol. Chem. 25 I, 2550-2563 57 Macheroux, P.. Ghisla S. and Hastings, J. W. ( 1993) Biochemistry 32, I4 183- I4 I86 58 Ryenon. C. C., Ballou, D. P. and Walsh, C. T. (1982) Biochemistry 2 I, 2644-2655 59 Hastings, J. W., Balny, C., LePeuch, C. and Douzou, P. (I 973) Proc. Natl. Acad. Sci. USA 70, 3468-3472 60 Entsch, 6. and van Berkel. W. J . H. (I 995) FASEB J . 9, 476483 5688-5696 47 Mizutani, H., Miyaham I., Hirotsu, K., Nishina, Y., Shiga, K., 61 Schreuder, H. A,, Mattevi, A., Obmolova, G., Kalk K. H., Hol, W. G. J ., Van der Bolt, F. J . T. and van Berkel, W. J . H. (I 994) Biochemistry 33, I0 I 6 I - I0 I70 Ballou, D. P. and Ludwig, M. L. ( 1994) Science 266, I I C- I I4 Biochemistry 36, I 57 I 3- I 5723 Massey, V. ( 1999) Biochemistry 38, I 153- I I58 62 Gatti, D. L., Palfey, B. A., Lah, M. S., Entsch, B., Massey, V., 63 Palfey, 6. A., Ballou, D. P. and Massey, V. (I 997) 64 Palfey, 6. A., Moran, G. R., Entsch, B., Ballou, D. P. and 65 Enroth, C., Neujahr, H., Schneider, G. and Lindqvist, Y. 66 Thorpe, C. and Kim, J . -1. P. ( 1995) FASEB J . 9,7 18-725 67 lyanagi, T., Makino, N. and Mason, H. S. ( 1974) Biochemistry 68 Vermilion, J. L and Coon, M. J . (I 978) J . Biol. Chem. 253, 69 Vermilion, J. L, Ballou, D. P., Massey, V. and Coon, M. J . (I 98 I) J . Biol. Chem. 256, 266-277 70 Oprian, D. D. and Coon, M. J . (I 982) J . Biol. Chem. 257, 8935-8944 71 Olson, J . S., Ballou, D. P., Palmer, G. and Massey, V. (1974) J . Biol. Chem. 249, 4363-4382 72 Abramovitz, A S. and Massey, V. (I 976) J , Biol. Chem. 25 I, 73 Mium R, Yamano, T. and Miyake, Y. (I 986) J. Biochem. 74 Stott, K., Saito, K., Thiele, D. J. and Massey, V. (I 993) J. Biol. ( 1998) Sttucture 6, 605-6 I7 13, 1701-1710 88 12-88 I9 5327-5336 (Tokyo) 99,90 1-906 Chem. 268,6097-6 I06 75 FOX, K. M. and K~@J s, P. A. ( 1994) Structure 2, 1089- I I05 76 Massey, V., Schopfer, L. M. and Dunham, W. R (I 984) in Flavins and Flavoproteins (Btay, R C., Engel, P. C. and Mayhew, 5. G., eds), pp. I9 1-2 10, Walter de Gruyter, Berlin 77 Stewart R C. and Massey, V. ( 1985) J. Biol. Chem. 260, 78 Vaz, A. D. N., Chaktabotty, S. and Massey, V. (I 995) 79 Murthy, Y. V. S. N., Meah, Y. and Massey, V. (I 999) J. Am. 80 Kohli, R M. and Massey, V. (I 998) J. Biol. Chem. 273, 8 I Massey, V.. Meah, Y.. Xu, D. and Brown, B.J . ( 1999) in Flavins and Flavoproteins I999 (Ghisla, S.. Kroneck P.. Macheroux P. and Sund, H., eds), pp. 645-654, Rudolf Weber Agency for Scientific Publications, Berlin 82 Niino, Y. S., Chakrabotty, S., Brown, 6. J. and Massey, V. (I 995) J. Biol. Chem. 270, 1983- I99 I 83 French, C. E. and Bruce, N. C. (I 995) Biochem. J. 3 12, 67 1-678 84 Schaller, F. and Weiler, E. W. (I 997) J. Biol. Chem. 272, 28066-28072 85 Strassner, J., Furholz, A., Macheroux, P., Amrhein. N. and Schaller, A ( 1999) J. Biol. Chem. 274, 35067-35073 13639-1 3647 Biochemistry 34,42464256 Chem. SOC. I 2 I, 5344-5345 32763-32770 Received 2 May2000 0 2000 Biochemical Society 296