3

The Chemical and Biological

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