Exploration of the diaphorase activity of neutrophil NADPH oxidaseCritical assessment of the interaction of iodonitrotetrazolium with the oxidase redox components Alexandra Poinas1, Jacq
Trang 1Exploration of the diaphorase activity of neutrophil NADPH oxidase
Critical assessment of the interaction of iodonitrotetrazolium
with the oxidase redox components
Alexandra Poinas1, Jacques Gaillard2, Pierre Vignais1and Jacques Doussiere1
1 Laboratoire de Biochimie et Biophysique des Syste`mes Inte´gre´s, UMR 5092 CEA-CNRS, De´partement de Biologie Mole´culaire
et Structurale Grenoble, France;2De´partement de Recherche Fondamentale sur la Matie´re Condense´e SCIB-SCPM,
CEA-Grenoble, France
In the O2 generating flavocytochrome b, the
membrane-bound component of the neutrophil NADPH oxidase,
electrons are transported from NADPH to O2 in the
fol-lowing sequence: NADPHfi FAD fi heme b fi O2
Although p-iodonitrotetrazolium (INT) has frequently been
used as a probe of the diaphorase activity of the neutrophil
flavocytochrome b, the propensity of its radical to interact
reversibly with O2led us to question its specificity This study
was undertaken to reexamine the interaction of INT with the
redox components of the neutrophil flavocytochrome b
Two series of inhibitors were used, namely the flavin analog
5-deaza FAD and the heme inhibitors bipyridyl and
ben-zylimidazole The following results indicate that INT reacts
preferentially with the hemes rather than with the FAD
redox center of flavocytochrome b and is not therefore a
specific probe of the diaphorase activity of
flavocyto-chrome b First, in anaerobiosis, reduced heme b in activa-ted membranes was reoxidized by INT as efficiently as by O2 even in the presence of concentrations of 5-deaza FAD which fully inhibited the NADPH oxidase activity Second, the titration curve of dithionite-reduced heme b in neutro-phil membranes obtained by oxidation with increasing amounts of INT was strictly superimposable on that of dithionite-reduced hemin Third, INT competitively inhib-ited the O2uptake by the activated NADPH oxidase in a cell-free system Finally, the heme inhibitor bipyridyl com-petitively inhibited the reduction of INT in anaerobiosis, and the oxygen uptake in aerobiosis
Keywords: diaphorase; INT reductase; NADPH oxidase; neutrophils; flavocytochrome b
Upon activation, the neutrophil NADPH oxidase complex
generates the superoxide anion O2 from which are derived
microbicidal oxygen species, such as hydrogen peroxide and
hypochloride The active NADPH oxidase complex consists
of a membrane-bound flavocytochrome b made of two
subunits, gp91phox and p22phox (phox for phagocyte
oxidase), and water-soluble proteins of cytosolic origin
(p67phox, p47phox, p40phox and Rac 1/2) [1] A defect in
any of the genes encoding gp91phox, p22phox, p47phox or
p67phox results in chronic granulomatous disease (CGD)
[2] Physiological activation of NADPH oxidase can be
mimicked by using a cell-free system with flavocytochrome b,
p47phox, p67phox, Rac, GTP and arachidonic acid as basic
components The large subunit of flavocytochrome b,
gp91phox, contains all of the redox components necessary
for electron transfer from NADPH to O2, namely FAD and
two hemes [3–6] Like the yeast FRE1 reductase, the b
cytochrome of the mitochondrial bc1 complex and cyto-chrome b6 of the b6f complex in chloroplasts, gp91phox contains multiple hydrophobic domains, consistent with transmembrane a helices, and two pairs of histidine residues
in these hydrophobic domains, separated by 13 intervening amino acids (quoted from [7]) Based on these considera-tions, it has been postulated that the two hemes located in the N-terminal domain of gp91phox are coordinated by two pairs of histidine residues within two distinct a helices [7] One of them (heme 1) is close to the cytosolic face of the membrane, the other (heme 2) is on the opposite side of the membrane The C-terminal region of gp91phox, which consists of predominantly hydrophilic amino acid residues,
is extramembranous and exposed to the cytosol It contains binding sites for NADPH and FAD The FAD binding site
is thought to be in the close neighborhood of heme 1 The topographical assignment of the redox centers of gp91phox
in this model indicates that electrons are transported from NADPH to O2across the membrane via a chain of redox components in the following sequence: NADPHfi FADfi heme 1 fi heme 2 fi O2 Consistent with the presence of two distinct domains in gp91phox are reports showing that gp91phox may act as a diaphorase in the presence of appropriate electron acceptors such as dichlo-rophenol indophenol [8] or p-iodonitrotetrazolium violet (INT) [9–11] Not only the oxidase activity, but also the diaphorase activity required activation for full elicitation [8,9] From these studies emerged the idea that the electron flux along the redox components of flavocytochrome b is
Correspondence to J Doussiere DBMS/BBSI, CEA-Grenoble,
17 rue des Martyrs, 38054 Grenoble cedex 9, France.
Fax: +33 4 76 88 51 85, Tel.: +33 4 76 88 34 76,
E-mail: jdoussiere@cea.fr
Abbreviations: INT, P-iodonitrotetrazolium; NBT, nitroblue
tetrazolium; 5-deaza FAD, 5 deazaflavin adenine dinucleotide; SOD,
superoxide dismutase; CGD, chronic granulomatous disease; phox,
phagocyte oxidase.
(Received 24 September 2001, revised 30 November 2001, accepted
3 January 2002)
Trang 2regulated both at the level of the NADPH-FAD portion of
the electron transfer chain and at the level of heme b [12,13]
This idea was supported by the finding that neutrophils of a
CGD (91X+) patient with a point mutation (Arg54Ser) in
the gp91phox subunit of flavocytochrome b, although
unable to reduce O2into O2 despite the presence of heme b
in gp91phox, retained the capacity to reduce INT [14] It was
also shown that electron transfer from NADPH to INT and
from NADPH to O2 could be activated independently of
each other, depending on the presence of p67phox and
p47phox [10,12] From these data it appeared that
measu-rement of INT reductase could be taken as an index of the
diaphorase activity of flavocytochrome b However, a recent
paper [15] called attention to the possibility of the
nonen-zymatic univalent reduction of tetrazolium salts, particularly
INT by O2 with concomitant production of the tetrazolium
radical In addition, the INT radical in an aerated medium
can reduce O2to O2 [16,17] These observations suggested
that under certain circumstances INT did not probe the
diaphorase activity of the NADPH oxidase The present
paper describes experiments in which INT and O2 were
compared for their ability to accept electrons from activated
flavocytochrome b, using neutrophil membranes pretreated
with 5-deaza FAD, an FAD analog inefficient in electron
transfer in flavocytochrome b, and with benzylimidazole
and bipyridyl as heme inhibitors The results show that INT
is able to directly oxidize reduced heme b
E X P E R I M E N T A L P R O C E D U R E S
Materials
NADPH, ATP, GTPcS were from Boehringer; horse heart
cytochrome c type III, arachidonic acid,
dimethanesulfox-ide, diisopropyl fluorophosphate, benzylimidazole and
hemin were from Sigma INT was from Amresco 5-Deaza
FAD was a gift from V Massey, Medical School Ann
Harbor Michigan (USA)
Biological preparations
Neutrophil membranes and cytosol were prepared from
bovine neutrophils in saline phosphate buffer (NaCl/Pi)
composed of 2.7 mM KCl, 136.7 mM NaCl, 1.5 mM
KH2PO4 and 8.1 mM Na2HPO4, pH 7.4 supplemented
with 1 mMdiisopropyl fluorophosphate and 1 mMEDTA
[13] Protein concentration was assayed with the BCA
reagent using BSA as standard Purified flavocytochrome b
in detergent was obtained as reported previously [18]
Preparation of INT radical
INT (60 mg) was solubilized in 1 mL of a mixture of
dimethyl sulfoxide/H2O (2 : 1, v/v) The oxidized INT was
reduced by a few grains of sodium dithionite The pale
yellow solution became rapidly orange, which is typical of
the INT radical This was immediately followed by four
sequential extractions of INT by 2 mL chloroform After
each extraction, the mixture was centrifuged at 1000 g for
2 min The chloroform solutions containing INT were
collected and pooled After evaporation under a flow of
nitrogen, the dry residue (57 mg) was taken up in 2 mL
dimethyl sulfoxide and kept at)20 °C under argon
Assay of oxidase and INT diaphorase activities NADPH oxidase activity was assayed in a cell-free system [13], either by measurement of the rate of production of O2
or the rate of O2 uptake INT diaphorase activity was assayed by the rate of reduction of INT into formazan in the presence of superoxide dismutase (SOD) or under anaero-biosis In all cases, the assay of oxidase activity was preceded
by an activation step at room temperature Briefly, mem-branes obtained from resting neutrophils were mixed with
2 mMMgSO4and an optimal amount of arachidonic acid After 5 min, cytosol from resting cells (an amount corres-ponding to 10· that of membrane protein), 20 lMGTPcS,
500 lMATP and 2 mMMgSO4were added, and incubation was continued for another 5 min In the case of O2 measurement, 10–20 lg aliquots of membrane protein were used in a final volume of 20–50 lL of NaCl/Pi Following activation, the suspension was transferred to a photometric cuvette containing 200 lM NADPH and either 100 lM cytochrome c or 100 lM INT in 2 mL NaCl/Pi Cyto-chrome c reduction was recorded at 550 nm (e¼ 21.1 mM )1Æcm)1) [19], and INT reduction at 500 nm (e¼ 11 mM )1Æ1 cm)1) [20] After 2–3 min, 50 lg of SOD was added to quench O2 In all preparations, cytochrome c reduction was inhibited to > 95% by the addition of SOD, indicating that O2 was the main product of reduction of O2
In contrast to the reduction of cytochrome c which is monoelectronic, reduction of INT to formazan requires two electrons For normalization of the data, the activities were calculated as lmol e–transferredÆmin)1Æmg membrane pro-tein)1 When reduction of INT was conducted under anaerobiosis, the cuvette containing the medium was sealed with gas-tight rubber stoppers, into which two needles were inserted One of the needles was used for flushing nitrogen, the other for gas evacuation
When the oxidase activity was assayed by the rate of O2 uptake, the suspension of activated particles was transferred
to an oxygraphic cuvette containing 1.5 mL NaCl/Pi supplemented with 250 lMNADPH and, when indicated, INT or heme inhibitors at different concentrations The quantity of neutrophil membranes used in the oxygraphic assays was 10· that used in the photometric assay For measurement of the Kmof activated oxidase for O2, the O2 concentration of the medium was decreased to 30–40% of the initial value by controlled N2bubbling prior to NADPH addition Below 40–50 lM, the oxygraphic traces curved inward The rates of oxygen uptake were deduced from the slopes of the tangents to the oxygraphic traces, and the contact points of the tangents with the curves were used to determine the O2 concentrations at which O2 uptake proceeds [18] All experiments were repeated two or three times, and the reported results are representative of at least two experiments
Optical spectra Absorption spectra of clear solutions were recorded at room temperature with an Uvikon 930 spectrophotometer In the case of turbid suspensions, a double beam PerkinElmer 557 spectrophotometer was used Reduction was achieved with
a few grains of sodium dithionite The amount of heme b in neutrophil membranes was determined from difference spectra (dithionite-reduced vs oxidized) The molar
Trang 3extinc-tion coefficients (De) were 106 mM )1Æcm)1at 425 nm (Soret
Peak) and 21.6 mM )1Æcm)1at 558 nm [21] The amount of
heme b in neutrophil membranes varies from 0.35 to
0.65 nmolÆmg protein)1 depending on preparations In
experiments where reduced heme b was reoxidized by
sequential additions of INT (cf Fig 8), the difference
spectra relative to the Soret band were recorded and the
extent of reoxidation was assessed by the decrease of the
peak of absorbancy by reference to a base-line (see Fig 8)
EPR spectra
EPR spectra were recorded with a X-band Bruker EMX
spectrometer equipped with an Oxford Instruments
ESR-900 continuous flow helium cryostat
R E S U L T S
Effect of NADPH oxidase activation and oxygen
on the rates of INT reductase and O2–production
The experiment illustrated in Fig 1 shows the effects of
arachidonic acid and O2on the rates of electron transfer
from NADPH to cytochrome c and INT Arachidonic acid,
an efficient amphiphile commonly used to elicit the
production of O2 in a cell-free system of NADPH oxidase
activation, was added at increasing concentrations to
neutrophil membranes which were further supplemented
with neutrophil cytosol, GTPcS and ATP After completion
of activation, the rates of cytochrome c reduction and INT reduction were measured either in an aerated medium or in
a N2saturated medium The data were expressed in terms
of lmol e–transferred min)1Æmg membrane protein)1, cor-recting for the fact that cytochrome c is reduced by one electron and INT by two electrons In the aerated medium (Fig 1A), the cytochrome c reductase activity, referred as oxidase activity, and the INT reductase activity both peaked
at a concentration of 1.2–1.3 lmol arachidonic acidÆmg membrane protein)1, the rates of electron transfer being 1.00 lmol and 0.78 lmol e–min)1Æmg membrane pro-tein)1, respectively Both activities differed in their sensitiv-ity to SOD The INT reductase was inhibited 50% by SOD, whereas reduction of cytochrome c was inhibited by
> 95%, indicating that reduction of cytochrome c was essentially due to the superoxide anion O2 generated by the NADPH oxidase
In the oxygen-free medium (Fig 1B) INT was reduced, but not cytochrome c Thus, in contrast to INT, cyto-chrome c does not capture electrons from a redox center of flavocytochrome b The rate of reduction of INT was even higher in anaerobiosis than in aerobiosis (0.94 lmol vs 0.78 lmol e–transferred min)1Æmg membrane protein)1), and nearly the same as the rate of O2 production in aerobiosis This result means that the electron transfer step from NADPH to the redox center from which electrons are captured by INT controls the rate of the overall electron transfer from NADPH to O2
The SOD-sensitive reduction of INT by the activated NADPH oxidase in an aerated medium previously des-cribed and asdes-cribed to the reduction of INT by O2 generated by the oxidase activity of flavocytochrome b [9,11] deserves some comments An alternative explanation
is that INT radicals generated by direct capture of electrons from reduced flavocytochrome b interact with O2 to gen-erate oxidized INT (INTox) and O2 [15,17] according to reaction 1: INT•+ O2« INTox+ O2 Eliminating O2 with SOD displaces the reaction to the right, with formation
of INTox In this mechanism, the superoxide O2 is no longer considered as the product of reduction of O2at the heme level of flavocytochrome b, but rather as the product
of reduction of O2 by INT, and the SOD-dependent inhibition of INT reduction appears to be an indirect effect
On the other hand, the INT radical may generate by dismutation the fully reduced INTredaccording to reaction 2: INT•+ INT•+ H+fi INTred+ INTox
At low concentrations of INT, reaction 1 predominates, whereas at high concentrations of INT reaction 2 (which is second order with respect to the INT concentration) is favored Thus, the balance of INT depends not only on the presence of O2, but also on the concentration of INT This may explain why the SOD-dependent sensitivity of INT reduction fluctuates depending on experimental conditions For example, in a recent report, the extent of inhibition of INT reduction by SOD was limited to 10% [9] compared with 50% in the present paper (Fig 1)
Characterization of the INT radical The reduction of tetrazolium salts into formazan, which involves the overall transfer of two electrons per molecule, proceeds by stepwise addition of individual electrons [22]
Fig 1 Reduction of O 2 and INT by neutrophil membranes activated in a
cell-free system Effect of increasing concentrations of arachidonic
acid Neutrophil membranes (20 lg protein) were incubated at room
temperature with increasing concentrations of arachidonic acid, up to
3 lmolÆmg protein)1, and 5 m M MgSO 4 After 5 min, cytosol (200 lg
protein) was added, together with 0.5 m M ATP and 10 l M GTPcS.
The final volume was adjusted to 50 lL with NaCl/P i and incubation
was continued for a further 5 min The whole sample was transferred to
a photometric cuvette in 2 mL NaCl/P i supplemented with 200 l M
NADPH and either 100 l M cytochrome c (d) or 100 l M INT (j),
depending on the measurement of the oxidase activity (O 2
produc-tion) or the INT reductase activity The assays were carried out in
aerobiosis (A) and in anaerobiosis (B) as described in Experimental
procedures Production of O 2 as a reducing agent was quenched by
addition of 50 lg of superoxide dismutase to the medium of the
photometric cuvette containing either cytochrome c (s) or INT (h)
(A).
Trang 4This process involves the formation of a tetrazolinyl free
radical which was found by EPR spectroscopy to be
relatively stable at 25°C in hydrophobic media, even in the
presence of O2 The tetrazolinyl radical can also accumulate
by oxidation of formazan or by disproportionation of a
mixture of formazan and the tetrazolium salt The optical
spectra illustrated in Fig 2A were recorded before and after
addition of a small amount of sodium dithionite to a
solution of INT in a mixture of dimethyl formamide and
water 1 : 1, v/v) The recorded spectrum of oxidized INT
above 400 nm was flat (Fig 2A, trace a) Following
addition of sodium dithionite, an orange color rapidly
developed, with a maximal optical absorbance at 449 nm
(Fig 2A, trace b) In a few minutes, the color changed to red
(Fig 2A, trace c), corresponding to a new spectrum with
two maxima at 500 nm and 550 nm, which was
character-istic of the monoformazan, i.e the fully reduced product of
INT [15,17] In the following aeration of the medium, the
absorbance of the spectral bands decreased, but the shape of
the spectrum remained the same (Fig 2A, trace d), which
means that the fully reduced INT became reoxidized It was
concluded that the transient absorbance at 449 nm
corres-ponded to a partially reduced state of INT, most likely to
the INT radical
The rate of transition from the oxidized state to the fully
reduced state upon addition of sodium dithionite depended
on the medium When the solvent was water, the
red-colored formazan accumulated in a few seconds In contrast, in a mixture of dimethyl formamide and water
of 2 : 1 (v/v), the INT radical was stable for more than
10 min (Fig 2B), the stability of the INT radical increasing with the increase in dimethyl formamide concentration Detergents such as Triton X-100 or SDS at a final concentration of 1% (w/v) also stabilize the INT radical (data not shown) This observation corroborates data showing that the nitroblue tetrazolium (NBT) radical obtained by reduction of oxidized NBT by silver amalgam was stabilized in dimethoxyethane [23] The stabilizing effect
of SDS was not encountered with arachidonic acid which
we routinely used as an activator of the NADPH oxidase The first derivative EPR spectrum at 293 K of INT reduced by sodium dithionite in dimethyl formamide shows a radical structure centered at g¼ 2.00 (Fig 2C, upper trace) The spectrum was characterized by a 10-line pattern, nine of which resemble those of the EPR spectrum
of the 2,3,5 triphenyltetrazolium chloride radical [23] The 10-line pattern of the EPR spectrum of the INT radical could be simulated by assuming a structure containing four equivalent nitrogen atoms and a supernumerary atom with
a spin of 1/2, namely a proton (Fig 2C, bottom trace), with isotropic hyperfine splitting constants of 0.5 mT between nitrogen atoms and 0.75 mT for the supernumer-ary proton
The specific chemical properties of INT and more particularly the stability of its radical in hydrophobic media may explain differences in the efficiency of electron capture by INT depending on experimental conditions of the assay of NADPH oxidase, for example the nature of the detergent used in the cell-free assay or the membrane concentration
Compared effects of the two heme inhibitors, benzylimidazole and bipyridyl on the optical and EPR spectra of hemin and flavocytochromeb
In a preliminary experiment, we found that benzylimidazole and bipyridyl inhibited not only the production of O2 assayed by the SOD-sensitive reduction of cytochrome c, but also the reduction of INT with the same efficiency Half inhibition was obtained with 2–3 mMbipyridyl and 5–7 mM benzylimidazole Inhibition was largely reversed by dilution, indicating that it was not due to denaturation of flavocyt-ochrome b Complementary experiments using optical and EPR spectra were carried out to assess the specificity of the effects of bipyridyl and benzylimidazole on the heme(s) of flavocytochrome b
Hemin was chosen as a model to test by spectral modifications the ability of benzylimidazole and bipyridyl
to react with heme iron Because hemin solutions in detergent are not turbid, absolute spectra of oxidized and reduced hemin in the absence and presence of inhibitors were recorded directly (Fig 3A, trace a, control, and trace
b, presence of benzylimidazole) The difference spectra of dithionite-reduced hemin plus benzylimidazole and dithi-onite-reduced hemin plus bipyridyl minus reduced hemin exemplify typical changes in the spectra consisting of the appearance of well defined peaks at 428 nm, 530 nm and
560 nm in the case of benzylimidazole (Fig 3A, trace c) and
at 437 nm in that of bipyridyl (Fig 3A, trace d) When the same heme inhibitors were added to neutrophil membranes
Fig 2 Evidence for accumulation of a stable INT radical during
reduction of INT by sodium dithionite The INT radical was prepared as
described in Experimental procedures (A) Optical spectra showing the
progressive reduction of INT by sodium dithionite, in a mixture of
dimethyl formamide and H 2 O (1 : 1, v/v), from a fully oxidized state
(a) to a fully reduced state (c) (after 5 min) via a semireduced state
corresponding to the INT radical (b) (after 2 min) The spectrum taken
15 min after aeration (d) has a shape similar to that of the fully reduced
INT, but its size was significantly decreased (B) Spectra of the purified
INT radical in a mixture of dimethyl formamide and H 2 O (2 : 1, v/v),
before (solid line) and after addition of sodium dithionite (dotted line).
(C) Upper trace: first derivative EPR spectrum of the INT radical
recorded at room temperature Microwave power 2 mW; modulation
frequency 100 kHz, modulation amplitude 0.5 mT, microwave
fre-quency 9.660 GHz Bottom trace: simulation of the EPR spectrum
shown in the upper trace was obtained by assuming four equivalent
nitrogen atoms coupled at 0.5 mT and one proton coupled at 0.75 mT.
Trang 5in the absence of arachidonic acid, the heme spectrum of
flavocytochrome b was not modified (Fig 3B, trace a)
Neither was it in the presence of arachidonic acid alone in
the absence of inhibitors (Fig 3B, trace b) However, when
the heme inhibitors were added to the neutrophil
mem-branes in the presence of arachidonic acid used at a
concentration that elicited maximal oxidase activity,
signi-ficant spectral modifications were recorded These
modifi-cations consisted in a decrease of the Soret peak
accompanied by a slight blue shift and in a decrease of the
a peak (Fig 3B, trace c, benzylimidazole, and trace d,
bipyridyl) Moreover, addition of benzylimidazole and
bipyridyl resulted in opposite modifications of the sizes of
the a and c peaks of heme b Benzylimidazole decreased the
a/c peak ratio from 4.5 to 3.4 whereas bipyridyl increased it
from 4.5 to 6.1, suggesting different types of constraint
applied to the hemes by the two inhibitors
Binding of benzylimidazole to the heme iron of hemin
and to the heme iron of purified flavocytochrome b was
assessed by EPR spectroscopy (Fig 4) Addition of
benzy-limidazole to hemin resulted in the decrease of the high spin
signal at g¼ 6.0 (Fig 4, trace d), characteristic of the
pentacoordinated form of the iron atom of the heme, and in
the concomitant emergence of a low spin signal with
components at g1¼ 2.97 and g2¼ 2.25 (trace e) Purified
flavocytochrome b in detergent displayed a mixture of
penta- and hexacoordinated forms of heme b (Fig 4, trace
a) The high spin signal at g¼ 6.0 similar to that of hemin accounted for the pentacoordinated form of the heme iron The hexacoordinated form was represented by two low spin
g1signals at g¼ 3.28 and g ¼ 2.85 The g2components are probably associated at g¼ 2.20 The signal at g ¼ 4.3 was due to adventitious ferric species [24] Thus, even in the absence of arachidonic acid, a fraction of purified flavocyt-ochrome b is pentacoordinated and capable of reacting with
O2 or with heme ligands The high spin fraction was significantly increased by addition of 100 lMarachidonic acid, whereas the low spin signals were totally erased (data not shown) in accordance with previous results [18] The high spin signal of purified flavocytochrome b (see control trace a) was decreased by addition of 25 mM benzylimidaz-ole (trace b), and nearly abolished at 50 mM benzylimidaz-ole (trace c), a concentration which also fully inhibited the NADPH oxidase activity Concomitantly with the disap-pearance of the high spin signal at g¼ 6.0, a low spin signal with components g1 and g2 at g¼ 2.97 and g ¼ 2.25 emerged at positions similar to those observed in the case of the hemin/benzylimidazole complex (trace e), probably due
to the binding of benzylimidazole as an axial ligand to the heme iron in hemin or in flavocytochrome b The two low spin signals g1at g¼ 3.28 and g ¼ 2.85 initially present in purified flavocytochrome b were not altered upon addition
of 50 mMbenzylimidazole This behavior is reminiscent of the absence of effect of benzylimidazole on the optical spectrum of resting neutrophil membranes in the absence of arachidonic acid Thus, the fraction of purified flavocyto-chrome b characterized by low spin signals contains a hexacoordinated heme iron unable to react with benzylim-idazole; this fraction, calculated by integration, represents roughly half of the total amount of flavocytochrome b
Fig 4 Effect of benzylimidazole on the EPR spectra of isolated cytochrome b and hemin Traces a–c are EPR spectra of purified flavo-cytochrome b (45 l M ) in solution in 20 m M P i , 20% glycerol, 0.5 M
NaCl and 0.1% Triton X-100, pH 7.4 Trace a corresponds to control flavocytochrome b, trace b to flavocytochrome b treated with 25 m M
benzylimidazole and trace c to flavocytochrome b treated with 50 m M
benzylimidazole Traces d and e correspond to hemin (1 m M ) in DMF untreated and treated with 50 m M benzylimidazole, respectively (A) and (B) show the high spin and low spin regions of the EPR spectra, respectively.
Fig 3 Effects of benzylimidazole and bipyridyl on optical spectra of
hemin and membrane bound flavocytochrome b (A) Traces a1 and a2:
absolute spectra of oxidized and dithionite-reduced hemin (10 l M ),
respectively Traces b1 and b2: absolute spectra of oxidized and
dithionite-reduced hemin in the presence of 5 m M imidazole Trace c:
difference spectrum of reduced hemin plus 5 m M benzylimidazole
against reduced hemin (10 l M ) Trace d: Difference spectrum of
reduced hemin plus 10 m M bipyridyl against reduced hemin (20 l M ).
(B) Traces a–d: difference spectra (dithionite-reduced minus oxidized)
at room temperature of neutrophil membranes in NaCl/P i
(1 mg proteinÆmL)1 equivalent to 0.65 nmol heme b) Trace a,
control; trace b, membranes supplemented by arachidonic acid
(1.3 lmolÆmg protein)1); trace c, reduced membranes plus arachidonic
acid treated for 5 min with 40 m M bipyridyl; trace d, reduced
membranes plus arachidonic acid treated for 5 min with 25 m M
benzylimidazole.
Trang 6Dependence of NADPH oxidase inhibition by 5-deaza FAD
on the activation state of flavocytochromeb
The effect of 5-deaza FAD, a flavin analog and an obligate
two-electron donor [25–27] was tested on the NADPH
oxidase activity and the INT reductase activity of
flavocyto-chrome b in the cell-free system (Fig 5) Its inhibitory effect
depended on the step at which it was added to the medium
When 5-deaza FAD was added to the activated cell-free
system, the NADPH oxidase and the INT reductase were
hardly inhibited On the other hand, when 5-deaza FAD
was preincubated with neutrophil membranes together with
arachidonic acid 5 min prior to addition of the other
components of the activation system, namely cytosol,
GTPcS and ATP, both the elicited NADPH oxidase and
INT reductase activities were inhibited efficiently
(Ki¼ 25 lM), and the extent of inhibition was the same
for the two activities The inhibition caused by 5-deaza
FAD was prevented by addition of a 10· excess of FAD
(data not shown) Together these results suggest that
arachidonic acid induces by itself some structural
modifica-tions in flavocytochrome b which result in the release of the bound FAD and its replacement by 5-deaza FAD When the oxidase is fully activated, these modifications do not occur any more It is noteworthy that, in contrast to the flavin analogs, the heme inhibitors benzylimidazole and bipyridyl were equally effective when added either to the activated cell-free system or to the membranes before NADPH oxidase activation
Effect of 5-deaza FAD and heme inhibitors
on the reduced state of flavocytochromeb
in activated neutrophil membranes
In the experiment of Fig 6 conducted in anaerobiosis, 5-deaza FAD was preincubated with neutrophil membranes and arachidonic acid 5 min before the addition of cytosol, GTPcS and ATP, a condition required for the optimal inhibitory effect of the flavin analog on the oxidase activity
In the control assay (Fig 6, trace a), addition of NADPH resulted in an abrupt rise of heme b reduction, followed by a plateau which corresponded to 40–50% of the full reduction obtained with sodium dithionite After the reduction plateau had been reached, 10 nmol O2(dissolved in buffer) were added, which resulted in an abrupt, but limited, reoxidation of heme b, rapidly counteracted by the electron flux issued from NADPH After a new redox equilibrium had been attained, two other redox cycles were initiated by addition of 5 and 10 nmol of INT Reoxidation of reduced heme b with 10 nmol INT was twice that with 5 nmol INT, indicating proportionality over this range of INT concen-tration Moreover, 5 nmol INT (a mediator which is reduced by a pair of electrons) are able to oxidize the same amount of reduced heme b as 10 nmol O2, which speaks in favor of the idea that reduced heme b is the source of electrons for both INT and O2 If the major source of electrons for INT were reduced FAD, assuming a back flow
of electrons from reduced heme b to FAD, the above stoichiometry would be different The experiment was completed by addition of sodium dithionite in limiting amounts just sufficient to reduce 95% of the heme components of flavocytochrome b Addition of 10 nmol O2 (dissolved in buffer) followed by 10 nmol INT resulted in oxidation cycles of heme b similar to those obtained in the presence of NADPH alone as reducing agent
Trace b (Fig 6) refers to the effect of 5-deaza FAD used
at a concentration that inhibited 80% of the oxidase activity The rate and the extent of the NADPH-dependent reduction of heme b were both largely decreased This explains why the redox cycles initiated by addition of
O2and INT were smaller in size, compared to the control (trace a); yet the dithionite-reduced heme b was reoxidized
by O2and INT to virtually the same rate and extent as in the control, despite the block imposed by 5-deaza FAD at the level of the flavin redox center of flavocytochrome b With a higher concentration of 5-deaza FAD, which inhibited
> 95% of the NADPH-dependent reduction of heme b, INT and O2were still able to reoxidize the dithionite reduced heme b (trace c, Fig 6) The two later traces (d and e, Fig 6) refer to the action of the heme inhibitors, benzylimidazole and bipyridyl Both inhibitors strongly interfered with the rate and extent of reoxidation of the heme b reduced in the presence of NADPH Following addition of sodium dithi-onite, the cycles of reoxidation of heme b initiated by O and
Fig 5 Dose-dependent inhibition of the NADPH oxidase and INT
reductase activities of neutrophil membranes by the flavin analog 5-deaza
FAD Oxidase activation was performed as described in Experimental
procedures and in the legend of Fig 1 5-deaza FAD was added with
arachidonic acid and MgSO 4 to the membrane suspension (20 lg
protein) After 5 min at room temperature, oxidase activation was
elicited by addition of cytosol and GTPcS After an additional 5 min
incubation, the oxidase activity was measured as O 2 production (d)
and the INT reductase activity by the rate of reduction of INT (j).
A parallel experiment was carried out in which 5-deaza FAD was
incubated for 5 min with the membrane suspension after the cell-free
activation of the NADPH oxidase; O 2 production (s), INT reductase
activity (h).
Trang 7INT were also significantly diminished, compared to the
control (trace a) In summary, using the flavin analog
5-deaza FAD and the heme inhibitors, bipyridyl and
benzylimidazole, to limit or to block the flux of electrons
along the redox centers of flavocytochrome b, it is possible
to demonstrate that INT is able to capture electrons from
the heme components of flavocytochrome b
Kinetics of inhibition of the oxidase activity by INT and bipyridyl
As INT appeared to capture electrons from the hemes of flavocytochrome b, we asked whether INT could compete with O2 Activated neutrophil membranes in a cell-free system in the presence of cytosol, GTPcS and arachidonic acid were placed in an oxygraphic cuvette containing the NaCl/Pimedium whose O2concentration had been previ-ously lowered to 80 lM by controlled bubbling of N2 Then, O2uptake was initiated by addition of a saturating concentration of NADPH (250 lM) Below 80 lMO2, the oxygraphic traces curved inwards In the portion of the traces corresponding to O2concentrations ranging between
80 lMand 20 lM, the rates of O2uptake were deduced from the slopes of tangents at different concentrations of O2 The assay was repeated with INT added to the medium at increasing concentrations In the absence of INT, the reciprocal plots corresponded to a straight line from which a
Kmvalue of 25 lMfor O2could be deduced (Fig 7A) At increasing concentrations of INT, the plots corresponded also to straight lines that intersected the 1/v axis at a common intercept, which was the same as that of the control curve, and the apparent Kmfor oxygen increased in proportion to the increase in INT concentration These features are typical of a competitive inhibition The Kifound for INT was 30 lM, a value which is of the same order as the Kmfound for O2[13]
Assuming that capture of electrons from reduced heme b
by INT was responsible for the competitive effect of INT on the oxidase activity of flavocytochrome b, it was inferred that a heme b inhibitor should competitively inhibit in anaerobiosis the reduction of INT by membranes of neutrophils activated in a cell-free system This was in fact the case with bipyridyl as shown in Fig 7B A similar competitive inhibition of O2uptake by bipyridyl was found when activated neutrophil membranes were incubated with NADPH in an aerated medium (Fig 7C) In both cases (Fig 7B,C) the calculated Kifor bipyridyl were the same, namely 2 mM± 0.2 mM Twice higher values for Kiwere obtained with benzylimidazole (data not shown)
INT-dependent reoxidation of dithionite-reduced heme b
in flavocytochromeb and dithionite-reduced hemin
In this experiment, we followed the stepwise oxidation of reduced flavocytochrome b and reduced hemin by sequen-tial additions of small amounts of INT Neutrophil membranes pretreated by arachidonic acid were placed in photometric cuvettes under a flow of nitrogen It is known that pretreatment of membranes by arachidonic acid modifies the spin state of heme b [18], but is not sufficient per se to elicit the oxidase activity of flabocytochrome b
A parallel spectrophotometric assay was carried out with hemin Hemin and flavocytochrome b were reduced to
95% by the addition of a limited amount of sodium dithionite To reoxidize hemin and the heme component of flavocytochrome b, INT was added by small increments to the anaerobic cuvettes Absorbance was recorded using a double wavelength spectrophotometer, between 380 nm and 480 nm for hemin and between 400 nm and 465 nm for flavocytochrome b, corresponding to the Soret peak of the two pigments Reoxidation of hemin or the heme of
Fig 6 Effect of 5-deaza FAD, bipyridyl and benzylimidazole on the O 2
and INT-dependent reoxidation of reduced heme b in activated
neutro-phil membranes Activated membranes (2 mg protein in 1.7 mL,
equivalent to 0.70 nmol heme b) in the cell-free system (arachidonic
acid, cytosol and GTPcS) were placed in a photometric cuvette The
medium (1.7 mL) was made anaerobic as described in Experimental
procedures The following compounds were injected into the medium
in minimal volumes in the following sequence: NADPH, 1 lmol; O 2 ,
10 nmol (dissolved in buffer); INT, 5 and 10 nmol; sodium dithionite,
1 lmol Trace a, control membranes; trace b, membranes
preincu-bated for 10 min with 40 l M 5-deaza FAD and arachidonic acid
fol-lowed by addition of cytosol and GTPcS; trace c, same conditions as in
trace b except that 5-deaza FAD was used at 80 l M ; trace d, same
conditions as in trace a with 10 m M benzylimidazole (BI) added to the
cuvette after reduction of heme b by NADPH; trace e, same conditions
as in trace a with 10 m M bipyridyl (Bipy) added to the cuvette after
reduction of heme b by NADPH After full reduction of heme b with
sodium dithionite, further cycles of oxidation were initiated by
addi-tion of O 2 and INT (respectively 10 and 5 nmol).
Trang 8flavocytochrome b was assessed by the decrease of the
absorbance (Fig 8, insert) Up to 80% heme reoxidation, a
linear relationship between the amount of added INT and
the absorbance decrease was observed (Fig 8) Above 80%
heme reoxidation, the curves departed from linearity The
two dose–response curves for hemin and flavocytochrome b
were virtually superimposable By extrapolation of the
linear portions of the curves to the abscissa, it could be
calculated that 0.45–0.48 mol INT was needed to reoxidize
1 mol hemin or 1 mol heme of flavocytochrome b, which is
consistent with the stoichiometry of 2 e–captured per INT
molecule in both cases Thus, like reduced hemin, the
reduced heme of flavocytochrome b is able to donate
electrons to INT A back transfer of electrons from reduced
heme b to FAD should not be significant because, as noted
above, the oxidase activity of neutrophil membranes treated
with arachidonic acid is dormant in the absence of cytosolic
factors [18] and, consequently, electron transfer between the
redox centers of flavocytochrome b is maintained at a very
low level In summary, the value of the stoichiometric ratio
of reduced INT to oxidized heme suggests that under our
experimental conditions electrons are essentially transferred
from the heme components of flavocytochrome b to INT
D I S C U S S I O N
The postulate that the flavocytochrome b component of the
neutrophil NADPH oxidase has the potential to display a
diaphorase activity stems from two different observations
First, artificial disruption of the electron transfer chain of
flavocytochrome b by detergents allows capture of electrons
upstream of heme b [8] Second, a mutated
flavocyto-chrome b from a CGD (91X+) patient, that was unable to
reduce O2into O2, was still able to carry electrons from
NADPH to INT This latter observation led to the belief
that INT is a suitable electron acceptor from the reduced
FAD component of flavocytochrome b and thereby a suitable probe of the diaphorase activity of flavocyto-chrome b [9] There were, however, in the meantime, reports that pointed to the rather complex behavior of INT [15,17]
We therefore decided to determine by classical techniques, using specific inhibitors of the electron flux in flavocyto-chrome b, which redox components interacted with INT The flavin analog, 5-deaza FAD was used to inhibit the flux
of electrons at the FAD level On the other hand, two efficient heme inhibitors, bipyridyl and benzylimidazole, were selected after validation by optical and EPR spectro-metric tests
The experiments described here led us to conclude that INT is able to capture electrons from the heme b components of the neutrophil flavocytochrome: (a) in activated membranes maintained in anaerobiosis, heme b reduced by NADPH was reoxidized by INT as efficiently
as by O2 (Fig 6) Reoxidation of heme b was insensitive
to 5-deaza FAD, which contrasted with the inhibitory effect of this compound on the reduction of heme b by NADPH (Fig 6) This result, which agrees with the well recognized function of 5-deaza FAD as a flavin inactive analog, rules out the possibility of a back reaction from heme to flavin and then from flavin to INT; (b) INT was able to reoxidize hemin reduced by sodium dithionite (Fig 7) As FAD was absent in this experiment, electrons were directly transferred from hemin to INT The comparative titrations by INT of dithionite-reduced hemin and dithionite-reduced heme b of flavocytochrome b ended with the same stoichiometry of roughly 0.5 mol INT reduced by 1 mol hemin or by 1 mol heme b in flavocytochrome b; (c) INT was able to compete with O2
in an aerobic medium, and also with the heme inhibitor, bipyridyl, in anaerobiosis (Fig 7) These results led us to conclude that the heme components of flavocytochrome b interact directly with INT
Fig 7 Kinetics of inhibition of the elicited NADPH oxidase activity in a cell-free system (A) and (C), aliquots of neutrophil membranes (150 lg protein in A) activated in the cell-free system were placed in the oxygraphic cuvette containing 1.5 mL NaCl/P i supplemented with different fixed concentrations of INT in A (d, zero; s, 30 l M ; j, 60 l M ; h, 90 l M ; _, 120 l M ) and bipyridyl in C (j, zero; d, 3 m M ; h, 7 m M ; s, 15 m M ) The initial concentration of O 2 (230 l M ) was decreased by about two-thirds by controlled bubbling of nitrogen prior to the assay of O 2 uptake initiated
by addition of 250 l M NADPH The rate of O 2 uptake was calculated from the slopes of the tangents to the oxygraphic traces (B) Activated membranes (40 lg protein per assay) were placed in a closed photometric cuvette containing 1 mL of a medium previously made anaerobic by N 2
bubbling The medium consisted of NaCl/P i supplemented with 250 l M NADPH and different concentrations of INT ranging from 20 to 200 l M The rate of INT reduction recorded at 500 nm was calculated using a molar extinction coefficient of 11 m M )1 Æcm)1 Bipyridyl was used at the following concentrations: j, zero; d, 3 m M ; h, 7 m M ; s, 15 m M
Trang 9Theoritically, INT might accept electrons from reduced
FAD in activated flavocytochrome b inasmuch as the FAD
binding site is believed to be located in the relatively
hydrophilic C-terminal half of the gp91phox subunit of the
flavocytochrome Experimental data show that this is not
the case, which raises the question of the capacity of INT to
probe specifically the diaphorase activity of neutrophil
NADPH oxidase It is however, not excluded that under
specific conditions, for example the presence of a detergent
or of a mutation such as Arg54Ser [14], the structural
arrangement of the FAD binding domain of
flavocyto-chrome b is modified, resulting in a loss of interaction of
FAD with heme 1 and in a facilitated access of INT to FAD Along this line, one may recall that a peculiar behavior of FAD in flavocytochrome b was recognized in the past [28], and explained in terms of a kinetic barrier between flavin and heme b A peculiar structural arrange-ment of the peptide chain in the FAD region of flavocyto-chrome b, possibly related to the oligomerization of the protein, might be responsible for its very unusual properties Finally, we would like to point out that some of the unusual properties of INT might be due to the stabilization of its radical in hydrophobic media, for example the lipid phase of membranes
A C K N O W L E D G E M E N T S
We thank V Massey for the gift of 5-deaza FAD, J Willison for careful reading of the manuscript and J Bournet-Cauci for excellent secretarial assistance This work was supported by funds from the Centre National
de la Recherche Scientifique, the Commissariat a` l’Energie Atomique, the Universite´ Joseph Fourier–Grenoble I, and the Association pour la Recherche sur le Cancer (9996).
R E F E R E N C E S
1 Babior, B.M (1999) NADPH oxidase: an update Blood 93, 1464– 1476.
2 Thrasher, A.J., Keep, N.H., Wientjes, F & Segal, A.W (1994) Chronic granulomatous disease Biochim Biophys Acta 1227, 1–24.
3 Segal, A.W., West, I., Wientjes, F., Nugent, J.H., Chavan, A.J., Haley, B., Garcia, R.C., Rosen, H & Scrace, G (1992) Cyto-chrome b-245 is a flavocytoCyto-chrome containing FAD and the NADPH-binding site of the microbicidal oxidase of phagocytes Biochem J 284, 781–788.
4 Doussie`re, J., Brandolin, G., Derrien, V & Vignais, P.V (1993) Critical assessment of the presence of an NADPH binding site on neutrophil cytochrome b558 by photoaffinity and immunochemical labeling Biochemistry 32, 8880–8887.
5 Doussie`re, J., Buzenet, G & Vignais, P.V (1995) Photoaffinity labeling and photoinactivation of the O2ÿ generating oxidase
of neutrophils by an azido derivative of FAD Biochemistry 34, 1760–1770.
6 YuL., Quinn, M.T., Cross, A.R & Dinauer, M.C (1998) Gp91 (phox) is the heme binding subunit of the superoxide-generating NADPH oxidase Proc Natl Acad Sci USA 95, 7993–7998.
7 Finegold, A.A., Shatwell, K.P., Segal, A.W., Klausner, R.D & Dancis, A (1996) Intramembrane bis-heme motif for transmem-brane electron transport conserved in a yeast iron reductase and the human NADPH oxidase J Biol Chem 271, 31021– 31024.
8 Doussie`re, J & Vignais, P.V (1992) Diphenylene iodonium as an inhibitor of the NADPH oxidase complex of bovine neutrophils Factors controlling the inhibitory potency of diphenylene iodo-nium in a cell-free system of oxidase activation Eur J Biochem.
208, 61–71.
9 Cross, A.R., Yarchover, J.L & Curnutte, J.T (1994) The super-oxide-generating system of human neutrophils possesses a novel diaphorase activity Evidence for distinct regulation of electron flow within NADPH oxidase by p67-phox and p47-phox J Biol Chem 269, 21448–21454.
10 Cross, A.R & Curnutte, J.T (1995) The cytosolic activating fac-tors p47phox and p67phox have distinct roles in the regulation of electron flow in NADPH oxidase J Biol Chem 270, 6543–6548.
11 Li, J & Guillory, R.J (1997) Purified leukocyte cytochrome b558 incorporated into liposomes catalyzes a cytosolic factor dependent diaphorase activity Biochemistry 36, 5529–5537.
Fig 8 Effects of INT on reoxidation of dithionite-reduced hemin and
dithionite-reduced flavocytochrome b Neutrophil membranes in
NaCl/P i (2 mg proteinÆmL)1; 0.60 nmol heme bÆmg protein)1, total
volume 1.7 mL) were preincubated for 10 min at room temperature
with arachidonic acid, 1.5 lmolÆmg protein)1 (j) A solution of
freshly prepared hemin was used at the concentration of 10 l M in
NaCl/P i supplemented with 1% Triton X-100 (s) Neutrophil
membranes or hemin were placed in a photometric cuvette The
medium had previously been made anaerobic by bubbling N 2 , and
maintained anaerobic during the optical assays as described in
Experimental procedures After a few minutes of anaerobiosis, small
amounts of sodium dithionite solution were added to attain a level of
reduction of heme b and hemin of 90–95% Sequential additions of
small aliquots (0.2 lL) of a solution of INT previously made anaerobic
by N 2 bubbling were injected to the medium until heme b and hemin
were fully reoxidized After each addition the difference spectra
(reduced minus oxidized) were recorded As shown in the insert in the
case of flavocytochrome b, the loss of absorbance was measured
between the peak of absorbance of the fully reduced spectrum and a
base-line drawn between the isosbestic point at 416 nm and 450 nm.
The results were normalized in terms of percent of reduced heme and
plotted against the calculated ratio of added INT to oxidized heme
(mol/mol) A similar procedure was used in the case of hemin.
Trang 1012 Cross, A.R., Erickson, R.W & Curnutte, J.T (1999)
Simulta-neous presence of p47 (phox) and flavocytochrome b-245 are
required for the activation of NADPH oxidase by anionic
amphiphiles Evidence for an intermediate state of oxidase
activation J Biol Chem 274, 15519–15525.
13 Doussie`re, J., Bouzidi, F., Poinas, A., Gaillard, J & Vignais, P.V.
(1999) Kinetic study of the activation of the neutrophil NADPH
oxidase by arachidonic acid Antagonistic effects of arachidonic
acid and phenylarsine oxide Biochemistry 38, 16394–16406.
14 Cross, A.R., Heyworth, P.G., Rae, J & Curnutte, J.T (1995) A
variant X-linked chronic granulomatous disease patient (X91+)
with partially functional cytochrome b J Biol Chem 270, 8194–
8200 [published erratum in J Biol Chem (1995) 270, 17056].
15 Liochev, S.I., Batinic-Haberle, I & Fridovich, I (1995) The effect
of detergents on the reduction of tetrazolium salts Arch Biochem.
Biophys 324, 48–52.
16 Auclair, C., Torres, M & Hakim, J (1978) Superoxide anion
involvement in NBT reduction catalyzed by NADPH-cytochrome
P-450 reductase: a pitfall FEBS Lett 89, 26–28.
17 Auclair, C & Voisin, E (1986) Nitroblue tetrazolium reduction.
In CRC Handbook of Methods for Oxygen Radical Research
(Greenwald, R.A., ed.), pp 123–132 CRC Press, Boca Raton, FL,
USA.
18 Doussie`re, J., Gaillard, J & Vignais, P.V (1996) Electron transfer
across the O2ÿ generating flavocytochrome b of neutrophils.
Evidence for a transition from a low-spin state to a high-spin state
of the heme iron component Biochemistry 35, 13400–13410.
19 Greenwood, C & Palmer, G (1965) Evidence for the existence of
two functionally distinct forms cytochrome c monomer at alkaline
pH J Biol Chem 240, 3660–3663.
20 Pearse, A.G.E (1961) Histochemistry: Theoretical and Applied, 2nd edn Little, Brown and Co, Boston, MA, USA.
21 Cross, A.R., Higson, F.K., Jones, O.T.G., Harper, A.M & Segal, A.W (1982) The enzymatic reduction and kinetics of cytochrome b -245 of neutrophils Biochem J 204, 479–485.
22 Maender, O.W & Russel, G.A (1966) The formation of radical intermediates in formazan-tetrazolium salt systems J Org Chem.
31, 442–446.
23 Deguchi, Y & Takagi, Y (1967) ESR spectrum of 2,3,5-triphenyl tetrazolium Tetrahedron Lett 33, 3179–3180.
24 Gadsby, P.M & Thompson, A.J (1986) Low temperature EPR and near-infrared MCD studies of highly anisotropic low-spin ferrihaem species FEBS Lett 197, 253–257.
25 Light, D.R., Walsh, C., O’Callaghan, A.M., Goetzl, E.J & Tauber, A.I (1981) Characteristics of the cofactor requirement for the superoxide-generating NADPH oxidase of human polymorphonuclear leukocytes Biochemistry 20, 1468–1476.
26 Spencer, R., Fisher, J & Walsh, C (1976) Preparation, char-acterization, and chemical properties of the flavin coenzyme analogues 5-deazariboflavin, 5-deaza-riboflavin 5¢-phosphate, and 5-deazariboflavin 5¢-diphosphate, 5¢-5¢ adenosine ester Biochem-istry 15, 1043–1053.
27 Fisher, J., Spencer, R & Walsh, C (1976) Enzyme-catalyzed redox reactions with the flavin analogues 5-deazariboflavin, 5-deazariboflavin 5¢-phosphate and 5-deazariboflavin 5¢-diphos-phate, 5¢-5¢-adenosine ester Biochemistry 15, 1054–1064.
28 Koskhin, V., Lotan, O & Pick, E (1997) Electron transfer in the superoxide-generating NADPH oxidase complex reconstituted
in vitro Biochim Biophys Acta 1319, 139–146.