Evidence for two different electron transfer pathways in the same enzyme, nitrate reductase A from Escherichia coli Roger Giordani* and Jean Buc Laboratoire de Chimie Bacte ´ rienne, Institut Fe ´ de ´ ratif ‘Biologie Structurale et Microbiologie’, Centre National de la Recherche Scientifique, Marseille, France In order to clarify the role of cytochrome in nitrate reductase we have performed spectrophotometric and stopped-flow kinetic studies of reduction and oxidation of the cytochrome hemes with analogues of physiological quinones, using menadione as an analogue of menaquinone and duro- quinone as an analogue of ubiquinone, and comparing the results with those obtained with dithionite. The spectropho- tometric studies indicate that reduction of the cytochrome hemes varies according to the analogue of quinone used, and in no cases is it complete. Stopped-flow kinetics of heme oxidation by potassium nitrate indicates that there are two distinct reactions, depending on whether the hemes were previously reduced by menadiol or by duroquinol. These re- sults, and those of spectrophotometric studies of a mutant lacking the highest-potential [Fe-S] cluster, allow us to pro- pose a two-pathway electron transfer model for nitrate reductase A from Escherichia coli. Keywords: cytochrome b; electron transfer; Escherichia coli; nitrate reductase A; quinone. Nitrate can be used as an electron acceptor for anaerobic growth of Escherichia coli [1,2]. This oxidoreduction is catalysed by nitrate reductase A (EC 1.7.99.4). This enzyme is a membrane-bound complex of three subunits, designated a, b and c, coded by three genes, narG, narH and narJ, which form a single operon [3–5]. The a and b subunits are located on the cytoplasmic side of the membrane and form an ab complex that binds to the membrane by interacting with the c subunit, a very hydrophobic protein embedded inthemembrane[4]. Each subunit carries a different set of redox centers [5,6]. The 139 kDa a subunit contains the active site for the reduction of nitrate to nitrite and a molybdenum cofactor (molybdopterin guanine dinucleotide); a recent crystallo- graphic study has revealed the presence of a new [4Fe-4S] cluster [7] in the a subunit. The 58 kDa b subunit contains four [Fe-S] clusters which belong to two classes: a high- potential and a low-potential class. The high-potential class contains [4Fe-4S] and [3Fe-4S] clusters with redox potential of +180 mV and +130 mV, respectively. Redox potentials of )55 and )420 mV correspond to two [4Fe-4S] clusters in the low-potential class. The 20 kDa c subunit is a hydro- phobic cytochrome of type b carrying two hemes [8–10]. The electron transfer in membranous enzyme requires a quinone as an electron donor, and either ubiquinone (a benzoquinone) or menaquinone (a naphthoquinone) can fulfil the role [11]. A steady-state kinetic study of nitrate reductase was carried out by Morpeth et al.[12].Unfortu- nately, however, this study did not take account of the stoichiometry of the reaction and in consequence gave incorrect rate equations and led to hazardous conclusions. A previous kinetic study [13] suggested that duroquinol (a ubiquinone analogue) and menadiol (a menaquinone analogue) deliver their electrons at two different sites on the nitrate reductase. The loss of the highest-potential [4Fe-4S] cluster in a mutant form of nitrate reductase results in an enzyme devoid of menadione activity, but still retaining duroquinone activity. The existence of a specific site of reaction for each quinol, together with the differences in the effects on the two quinols produced by loss of an [Fe-S] cluster, suggested the possibility of two separate pathways for transfer of electrons from duroquinol and menadiol in nitrate reductase A [13]. EPR [9] and potentiometric [14] studies show the existence of two b type hemes in the c subunit, cyto- chrome b, of nitrate reductase A. The data from these studies support the assignment of the axial ligands to the low-potential heme (b L )(E m , 7 ¼ 20 mV) and to the high- potential (E m , 7 ¼ 120 mV) heme (b H ), respectively, located near the periplasmic and the cytoplasmic side of the membrane. Correct insertion of the two hemes into the c subunit requires anchoring to the ab complex [9,10]. Moreover, a kinetic study by stopped-flow of nitrate reductase with menadione only [15], exhibits four kinetic phases in the reduction of the hemes by menaquinol. According to the quinone type used, the two hemes of the cytochrome are probably involved in the transfer of electrons from the two specific binding sites and the two specific pathways suggested [13] for transport of electrons from the two quinol types to [Fe-S] clusters, molybdenum cofactor and nitrate. To clarify the role of the hemes we have made spectral and kinetic studies of the reduction of these hemes with various analogues of physiological Correspondence to J. Buc, Laboratoire de Chimie Bacte ´ rienne, 31 chemin Joseph-Aiguier, B.P. 71, 13042 Marseille Cedex 20, France. Fax: + 33 491 7189 14, E-mail: buc@ibsm.cnrs-mrs.fr Abbreviations: HOQNO, 2-n-heptyl-4-hydroxyquinoline-N-oxide. Enzyme: nitrate reductase A (EC 1.7.99.4). *Present address:Faculte ´ de Pharmacie, Universite ´ de la Me ´ diterrane ´ e, 13385 Marseille Cedex 05, France. (Received 17 March 2004, revised 8 April 2004, accepted 14 April 2004) Eur. J. Biochem. 271, 2400–2407 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04159.x quinones and compared the results with those obtained with dithionite. Oxidation of reduced hemes was followed after addition of potassium nitrate. In view of the rapidity of the reduction and oxidation reactions, the kinetic studies needed to be made with a stopped-flow apparatus. Experimental procedures Reagents and chemicals Menadione (2-methyl-1,4-naphthoquinone), and duroqui- none (tetramethyl-p-benzoquinone) were purchased from Aldrich. Juglone (5-hydroxy-1,4-naphthoquinone), plumb- agin (5-hydroxy 2-methyl-1,4-naphthoquinone), coen- zyme Q 0 (2,3-dimethyl 5-methyl-p-benzoquinone) and benzyl viologen were from Sigma. All other chemicals were of the highest grade of purity commercially available, and were supplied either by Prolabo or Merck. Bacterial strains and plasmids The strains used in this study were LCB2048 (DNRA,DNRZ) as strain devoid of nitrate reductase [16], pVA700, pJF119EH(narGHIJ)Ap r , as overexpressed wild- type plasmid [17] and pVA700-C16, pJF119EH(nar- GH[C16A]IJ)Ap r , as plasmid lacking the high-potential [4Fe-4S] cluster [17]. Growth conditions Growth conditions were those described previously [17]. All strains were grown anaerobically at 37 °ConTYmedium supplemented with glucose (2 gÆL )1 ). Expression of nitrate reductase was induced by isopropyl thio-b- D -galactoside (0.2 l M ) (pVA700). The antibiotics ampicillin (50 mgÆL )1 ) and chloramphenicol (10 mgÆL )1 )wereused. Preparation of subcellular fractions Cells were harvested during the exponential phase of growth and suspended in 100 m M potassium phosphate buffer pH 6.6 and disrupted at 69 MPa by passage through a French press. The resulting suspension was centrifuged (for 15 min at 18 000 g) to sediment unbroken cells. The supernatant was further centrifuged for 90 min at a maxi- mum of 120 000 g, and the new supernatant was discarded while the pellet was retained. All of these procedures were performed at 4 °C. The pellets, containing nitrate reductase as a complete membrane-bound complex, were resuspended in a small volume of buffer and stored at )80 °C until use. Quantification of nitrate reductase The concentration of nitrate reductase was estimated by reference to the percentage of enzyme in the total protein. This percentage was determined from the percentage of nitrate reductase antigen present in nitrate reductase membranous preparations measured by rocket immuno- electrophoresis [18] as described previously [19]. Proteins were estimated by the technique of Lowry et al. [20] using bovine serum albumin as standard. The amount of over- expressed enzymes was about 10-times that in the parent strain MC4100. With the plasmids used here, which express the genes stoicheiometrically, about 90% of the enzyme is membrane-bound (further details in [17]). Enzyme assays Nitrate reductase activity with benzyl viologen as substrate was measured spectrophotometrically [21] by nitrate- dependent oxidation of reduced benzyl viologen, with the precautions described previously [22]. Nitrate reductase activities with quinols as substrates were measured by a spectrophotometric method as des- cribed previously [22]. Spectra These were obtained with a Hitachi U-2000 spectro- photometer, thermostated at 37 °C and connected to a personal computer. All experiments were performed in 50 m M potassium phosphate buffer pH 6.6. A 1.6 mL quartz cuvette closed by a Teflon cap with a central hole was used to allow the different reagents to be added with a microsyringe. Thorough mixing in the cuvette was achieved by displacement of glass beads. As the mem- branes were from strains with nitrate reductase over- expressed, amounts of membrane low enough to avoid turbidity could be used. Reduction of quinone analogues Quinone analogue solution (1.7 mL of 20 m M , in ethanol) were added to 70 mg of zinc powder and 60 lLof5 M HCl, after mixing and sedimentation of the zinc the reduced analogue solution could used for two or three hours. This technique, used in organic chemistry [23], is very efficient and is easier to use than that with KBH 4 ,whichwas previously employed in nitrate reductase assays [22]. Stopped-flow kinetics Stopped-flow kinetic measurements were made with a Hi-Tech Scientific FF61 apparatus (Salisbury, UK) con- nected to a personal computer. The stopped-flow appar- atus was equipped with a 1 cm light-path quartz cuvette. To minimize oxygen leaks, the drive system was sub- merged in a thermostatically controlled circulating water bath at 37 °C. The stopped-flow system was thoroughly flushed with anaerobic buffer (50 m M potassium phos- phate, pH 6.6) immediately before the experiments were started. Solutions were equilibrated to assay temperature before experiments. To measure the absorbance at 560 nm, measurement of baselines were established at each wavelength with anaerobic buffer, and the values obtained were typically within 5% of those obtained from the absorbance spectra of the same solutions recorded on the Hitachi U-2000 spectrophotometer. Analysis of data A personal computer was used to fit experimental data to appropriate equations by nonlinear least squares, using a Newton–Gauss algorithm [24]. Ó FEBS 2004 Two electron transfer pathways in nitrate reductase (Eur. J. Biochem. 271) 2401 Results Spectrophotometric studies with quinone analogues Spectra were measured between 500 and 600 nm in order to follow the redox state of the cytochrome by following the characteristic peak at 560 nm. If one takes as reference the reduction by dithionite, a nonphysiological reducing agent that reduces the enzyme completely and nonspecifically (heme, [Fe-S] clusters and molybdedum cofactor) one can observe in the overexpressed wild-type strain, a difference between reduced and oxidized spectra, according to whether the subsequent reductant is menadione or duroquinone (Fig. 1). In both cases the maximum amplitude at 560 nm is less than that obtained with dithionite, which seems to indicate that the reduction of the hemes of the cytochrome varies according to the analogue of quinone used, and in neither case is it complete. Use of a strain devoid of nitrate reductase (LCB 2048) makes very clear the lack of any signal at 560 nm (Fig. 1, trace 4). Reduction by analogues is performed by addition of small quantities of reduced analogue until there is no further change in signal. To reduce quinone analogues, the reducing agent used was KBH 4 followed by zinc. Zinc in the presence of HCl has been preferred, as it gives a more reliable and stable reduction in one step with the advantage of not diluting the reducing solutions, thus allowing the addition of smaller quantities for the same reducing effect. In any event, the results obtained are equivalent whatever the mode of reduction of analogues of quinone used. The reduction of membranes with dithionite is performed by addition of dithionite powder. These experiments have been carried out using a range of concentrations in membrane at the limit of turbidity of the solution (0.193–2.09 l M ). In all cases one obtains similar results (Fig. 1) that show the variation of the amplitude of the signal at 560 nm between reduced and oxidized forms by the nitrate according to the concentration in nitrate reductase with the three electron donors used (dithionite, menadiol and duroquinol). In the concentration range used, the amplitude of absorbance reduced minus oxidized is linearly correlated with the concentration of nitrate reductase (Fig. 2) whatever the reducing agent used, the correlation coefficients of least- square regressions being 0.98, 0.98 and 0.96 for dithionite, menadiol and duroquinol, respectively. One can determine the slopes of the three straight lines obtained with the three electron donors. These slopes correspond to the differences in molecular extinction coefficients of reduced and oxidized forms of the hemes, and therefore to apparent molecular extinction coefficients corresponding to maximal reduction of the various electron donors: 0.069 l M )1 Æcm )1 for dithionite, 0.059 l M )1 Æcm )1 for menadiol and 0.031 l M )1 Æcm )1 for duroquinol. These values have induced us to verify with other analogues whether there are differences in amplitudes of the peak at 560 nm between reduced and oxidized forms. Fig. 1. Difference spectra of membranous nitrate reductase. Mem- branous enzyme was reduced by (1) dithionite, (2) menadiol and (3) duroquinol. Reductions were obtained by addition of small amounts of reduced analogues or dithionite until no further change in the spectrum. Oxidations were performed by addition of nitrate. Trace (4) was obtained with a strain lacking nitrate reductase (LCB 2048) and dithionite as reductant. The membranous nitrate reductase concen- tration was 1.2 l M in all cases. Fig. 2. Variation of absorbance difference between reduced and oxidized membranous nitrate reductase vs. enzyme concentration. Membranous nitrate reductase was reduced by (1) dithionite, (2) menadiol and (3) duroquinol. Lines were obtained by least-squares fitting, the slopes of these lines being consistent with the apparent molecular extinction coefficients. The absorbances were measuered at 560 nm. 2402 R. Giordani and J. Buc (Eur. J. Biochem. 271) Ó FEBS 2004 We have used analogues of type ubiquinone (lapachol and Coenzyme Q 0 and decylubiquinone), and analogues of type menaquinone (plumbagine and juglone). These studies have been undertaken with various enzyme concentrations to permit us to obtain the apparent molecular extinction coefficients (Table 1). Except for decylubiquinone, which induces flocculation in the cuvette, we obtained a difference of amplitude of the peak at 560 nm between the form reduced by analogues and the form oxidized by nitrate, and in all cases this difference was less than that obtained with dithionite. The differences of molecular extinction coefficients for reduced and oxidized forms can be grouped into two classes according to their values, the analogues of ubiquinone and those of menaquinone, these last having larger differences of coefficients than the first. This is not surprising, because menaquinols are the preferred electron donors for nitrate reductase in anaerobic conditions [25]. These analogues (lapachol, plumbagine, etc.), unlike menadiol and duroquinol, give spectra with significant baselines in the absence of membrane, necessitating corrections to the spectra. In addition, the specific absorption of analogues at 560 nm complicates the kinetic study of reduction of the cytochrome according to the choice of menadiol and duroquinol for spectral and kinetic studies. Influence of HOQNO on spectra The presence of the menaquinone analogue 2-n-heptyl-4- hydroxyquinoline-N-oxide (HOQNO) in the medium inhib- its reduction of the cytochrome by menadiol, but does not inhibit reduction by duroquinol (Fig. 3). It is probable that the nucleus of the HOQNO molecule, similar to that of duroquinone, i.e. a nucleus of type benzoquinone, specific- ally inhibits the site for menadiol, as predicted by previous studies showing cross-inhibition of menadiol and duro- quinol on the binding of these electrons donors to the cytochrome b of nitrate reductase [13]. These results with HOQNO explain those of Rothery et al. [23], who found a site of interaction between the cytochrome and the quinones by studying inhibition of the nitrate reductase activity by HOQNO. It has therefore clarified the binding site of the menaquinols, but for technical reasons it cannot reveal the site for the ubiquinols, because HOQNO does not inhibit binding of the ubiquinols. Kinetic studies by stopped-flow We have studied oxidation by nitrate of the cytochrome of nitrate reductase, preliminarily reduced by an analogue of quinone, menadiol or duroquinol. We have made experiments at various concentrations of electron donors and nitrate, in each case for several ranges of time (0.5–5 s). The traces obtained are shown in Fig. 4. They are monophasic and fit a decreasing exponential equation. The amplitudes obtained agree with those obtained during spectrophotometric studies (Fig. 1). The time constants corresponding to apparent rate constants are obtained by fitting Eqn (1): DAbs ¼ a þ b e Àct ð1Þ where, DAbs is the variation of absorbance, a and b are amplitude parameters and c is the time constant. The time constants are different according to whether the enzyme has been reduced by menadiol or by duroquinol; the results are listed in Table 2. Whatever the experimental conditions, the apparent rate constant is 1 s )1 with menadiol, whereas it is 2 s )1 with duroquinol. This doubling of the apparent rate constants indicates that there are two distinct reactions, and so oxidation of the cytochrome by nitrate follows separate pathways according to the nature of the electron donor. The kinetics of reduction of the cytochrome is very complex. In time ranges from 0.5 to 10 s, one observes traces corresponding to the sum of three exponentials at least. These traces are too complex to be analysed with precision. The residual plots, difference between experimental data and calculated values obtained after fitting, shown in Fig. 4, Table 1. Apparent molecular extinction coefficients between reduced and oxidized membranous nitrate reductase for different reductants. e values were obtained by least-squares fitting, like those presented in Fig. 2. Reductant E m (mV) e (l M )1 Æcm )1 ) Artificial reductant Dithionite 0.069 ± 0.002 Menaquinone Menadione )1 0.059 ± 0.004 analogues Plumbagine )74 0.043 ± 0.001 Juglone +33 0.037 ± 0.005 Lapachol )179 0.046 ± 0.005 Ubiqinone Duroquinone +35 0.031 ± 0.04 analogues Coenzyme Q 0 +10 0.011 ± 0.03 Fig. 3. Variation of reducing of membranous nitrate reductase by quinol analogues vs. HOQNO concentration. The enzyme was reduced with duroquinol (j)ormenadiol(d). The nitrate reductase concentration was 0.6 l M . Lines were obtained by least-squares fitting. Ó FEBS 2004 Two electron transfer pathways in nitrate reductase (Eur. J. Biochem. 271) 2403 were distorted at the origin. These deviations were due to the time needed for a homogenous system to be reached after mixing of the membrane suspension and nitrate solution during the beginning of measurement for each experiment. Apart from these artifacts the distributions of the residuals show a good agreement between experimental data and the plot obtained by fitting. The traces obtained are comparable to these obtained by Zhao et al. [15]. These authors found four phases in the reduction of the cytochrome by menadiol, phases arbitrarily attributed to particular steps in the scheme that they proposed for the mechanism of reduction. Unfortunately, however, this scheme does not agree with the results from steady-state kinetics that we have obtained in a previous study [13]. There we showed that the catalytic mechanism implied an intermediate complex incompatible with binding of the electron donor and allowing it to give its two electrons successively before being released. Spectrophotometric studies with a mutant lacking the highest-potential iron sulphur cluster Use of a mutant (pVA700-C16) whose nitrate reductase has lost the highest-potential [4Fe-4S] cluster has allowed us to confirm the existence of these two pathways of electron transfer in the enzyme. Figure 5 shows reduction of the cytochrome of this mutant by dithionite, menadiol or duroquinol, followed by reoxidation by nitrate. The cyto- chrome is unambiguously reduced, regardless of the electron donor. On the other hand, although the enzyme reduced by duroquinol is fully oxidized by nitrate, that reduced by menadiol is not oxidized. This result agrees with the conclusions from the steady-state kinetics [13] which showed that in the case of this mutant no menadione oxidase activity was observed even though duroquinone oxidase activity was present. The result obtained with dithionite confirms this conclusion; indeed, even through the reduction is total, the oxidation is only partial, with the fraction of oxidation corresponding to that observed for duroquinol. These results indicate therefore that the [Fe–S] clusters do not have the same role in electron transfer in nitrate reductase. The [4Fe-4S] cluster of high-potential would therefore be the indispensable relay for the transfer of electrons when menadiol is the electron donor, and the [3Fe-4S] cluster would be implied in the transfer of electrons when duroquinol is used. Fig. 4. Oxidization of membranous nitrate reductase previously reduced by quinone analogues. Absorbance changes observed, at 560 nm, after mixing enzyme (3.14 l M ) reduced by menadiol (A) or duroquinol (B) with an equal volume of nitrate (40 m M )at37°C. The traces are fitted to Eqn (1) DAbs ¼ a + b e )ct which corresponds to a decreasing exponential. The apparent kinetic constants were 1 s )1 for enzyme reduced by menadiol and 2 s )1 for enzyme reduced by duroquinol. The distributions of residuals after fitting data from traces (A) and (B) to Eqn (1) are shown, respectively, in plots (C) and (D). Table 2. Comparison of kinetic contants of oxidation by nitrate of nitrate reductase hemes reduced by menadiol or duroqinol. Values of kinetic constants were averages of several experiments in different time ranges (0.5, 1, 2 and 5 s). The membranous nitrate reductase concen- tration was in all cases 1.57 l M . [Quinol] (m M ) [KNO 3 ](m M )k(s )1 ) Menadiol 0.768 100 1.065 ± 0.034 0.576 0.964 ± 0.042 0.384 1.043 ± 0.075 0.768 20 1.184 ± 0.068 0.576 1.029 ± 0.046 0.384 0.932 ± 0.100 Mean 1.027 ± 0.062 Duroquinol 0.768 100 1.830 ± 0.017 0.576 2.069 ± 0.018 0.384 2.120 ± 0.055 0.768 20 1.841 ± 0.092 0.576 2.026 ± 0.067 0.384 2.006 ± 0.019 Mean 1.988 ± 0.094 2404 R. Giordani and J. Buc (Eur. J. Biochem. 271) Ó FEBS 2004 Discussion Steady-state kinetic studies [13] have shown the existence of two specific binding sites for menadiol and duroquinol. This situation recalls that described for fumarate reductase from Escherichia coli, where the separation of oxidative and reductive activities suggests there are two quinol binding sites and that electron transfer occurs in two one-electron steps at these sites [26]. The presence of these two sites has been corroborated by crystallographic studies, with two binding sites termed Q P and Q D indicating their positions proximal or distal to the site of fumarate reduction; the use of the quinol-binding site inhibitor HOQNO shows that the inhibitor blocks binding of MQH 2 at the Q P site [27]. Study of the reduction and oxidation of the cytochrome has brought us to the conclusion that the two hemes do not have the same role. In particular, kinetic results with the stopped-flow apparatus for the oxidation of the cytochrome show that there are two distinct reactions, two apparent kinetic constants differing by a factor of two. Previous studies [13], as well as reduction and oxidation spectra by nitrate of the cytochrome of a mutant strain having lost the highest-potential [4Fe-4S] cluster, show that b clusters do not play the same role in the transfer of electrons between the cytochrome and the molybdenum cofactor. This [4Fe-4S] cluster is the indispensable relay when the donor is menadiol. The [3Fe-4S] cluster may be the relay when the donor is duroquinol. We can therefore conclude that the transfer of electrons up to the molyb- denum cofactor in nitrate reductase follows two distinct pathways according to the nature of the electron donor, as illustrated in Fig. 6. As the reduction of nitrate to nitrite requires two electrons, there must necessarily be two successive bindings of quinone, with transfer of one electron to the hemes, then to the [Fe-S] cluster, to be finally accumulated at the level of the molybdenum cofactor to be able to undertake the catalytic reaction. The location of the hemes has been studied by Rothery et al. [28,29]. These authors showed that the low-potential heme b L , located on the periplasmic side of the membrane, is associated with a single quinol-binding site. However, the technique used to determine binding sites of quinols, inhibition by HOQNO (a structural analogue of menaqui- none), did not allow them to see the binding site of ubiquinol because HOQNO is not an inhibitor for the ubiqinones. Indeed, cocrystallization of quinol fumarate reductase with HOQNO shows that HOQNO can inhibit the menaquinol binding site but not the ubiquinol binding site [27]. Consequently, the low-potential heme located on the periplasmic side of the membrane appears to be associated with a binding site for menaquinols, but it is highly probable that the high-potential heme b H , located on the cytoplasmic side, was associated with the binding site for ubiquinols. This conclusion agrees with structural data that indicate a cavity containing two distinct regions directly adjacent to the two hemes of c [7]. The four [Fe-S] clusters of b are positioned in two pairs due to the fact of their coordination to the polypeptide chain, each high-potential cluster being associated with a low-potential cluster [10]. The crystal structure of b shows two structural domains. Each domain contains a high- potential [Fe-S] cluster and low-potential [Fe-S] cluster, the two being sandwiched between two helices on one side, and Fig. 5. Spectra of membranous nitrate reduc- tase cytochrome of mutant devoid of the high- est-potential [4Fe-4S] cluster. Reduced spectra are shown as solid lines and oxidized spectra as dotted lines. The nitrate concentration was in all cases 0.95 l M . Fig. 6. Proposed mechanism for electron transfer in nitrate reductase. Menadiol and ubiquinol are symbolized by MQ and UQ, respectively. The molybdenum cofactor and the two hemes are denoted CoMo, b L (low-potential heme) and b H (high-potential heme). This scheme is consistent with the structural data of Bertero et al.[7]. Ó FEBS 2004 Two electron transfer pathways in nitrate reductase (Eur. J. Biochem. 271) 2405 a b-sheet on the other [7]. This structure is consistent with the idea that the [Fe-S] clusters are coupled in two redox units, each containing a low-potential and a high-potential [Fe-S]. Thus each structural domain of b may be a transfer unit for the electron transfer pathway in this enzyme. In our scheme we have integrated these structural data supposing that the two pairs of [Fe-S] cluster both participate. That recalls that two parallel electron pathways towards the a [4Fe-4S] cluster and the molybdenum cofactor with involvement of high-potential [Fe-S] had been suggested, supposing that the low-potential clusters play no redox role [17]. Note that HOQNO and stigmatellin inhibit nitrate- dependent heme reoxidation [28], suggesting the presence of a second dissociable Q-site (the Q nr site) between heme b H and the [3Fe-4S] cluster [29]. Likewise our results are not at variance with kinetic data for reduction of the enzyme by menadiol [15], suggesting the existence of two menadiol binding sites in the enzyme, one with higher affinity than the other, as well as inhibition data indicating the possibility of more than one menaquinol binding site in nitrate reductase [15]. The fact that nitrate is able to oxidize heme b H but not heme b L in the presence of an excess of quinol and HOQNO [30] is consistent with our results showing inhibition by HOQNO to the extent of reduction by menadiol and not by duroquinol, and thus inhibition of duroquinone oxidase activity. The existence of these two pathways of electron transfer may appear surprising, but nitrate reductase is one of the rare enzymes of quinone-oxidase type that can accept both menaquinol and ubiquinol as electron donor according to conditions of growth. Acknowledgements WeareindebtedtoDrF.Blascoforhishelpintheconstructionof mutant strains, to Dr G. Giordano for preparation of enzymes and to Dr J. Pommier for performing the rocket assays. The authors express their gratitude to Dr Wolfgang Nitschke (BIP, CNRS, Marseille) to have allowed us to use a stopped-flow apparatus. We thank Dr A. Cornish-Bowden for critical reading and correcting of the manuscript. References 1. Pichinoty, F. (1969) Les nitrates re ´ ductases bacte ´ riennes. I. Substrats, e ´ tat particulier et inhibiteurs de l’enzyme. Arch. Mikrobiol. 68, 51–64. 2. Ruiz-Herrera, J. & DeMoss, J.A. (1969) Nitrate reductase com- plex of Escherichia coli K-12: participation of specific formate dehydrogenase and cytochrome b 1 components in nitrate reduc- tion. J. Bacteriol. 99, 720–729. 3. Sodergren, E.J. & DeMoss, J.A. (1988) narI region of the Escherichia coli nitrate reductase (nar) operon contains two genes. J. Bacteriol. 170, 1721–1729. 4. Sodergren, E.J., Hsu, P.Y. & DeMoss, J.A. (1988) Roles of the narJ and narI gene products in the expression of nitrate reductase in Escherichia coli. J. Biol. Chem. 263, 16156–16162. 5. Blasco, F., Iobbi, C., Giordano, G., Chippaux, M. & Bonnefoy, V. (1989) Nitrate reductase of Escherichia coli: completion of the nucleotide sequence of the nar operon and reassessment of the role of the a and b subunits in iron binding electron transfer. Mol. Genet. 218, 249–256. 6. Guigliarelli, B., Asso, M., More, C., Augier, V., Blasco, F., Pommier, J., Giordano, G. & Bertrand, P. (1992) EPR and redox characterization of iron-sulphur centers in nitrate reductase from Escherichia coli. Evidence for a high-potential and a low-potential class and their relevance in the electron-transfer mechanism. Eur. J. Biochem. 207, 61–68. 7.Bertero,M.G.,Rothery,R.A.,Palak,M.,Hou,C.,Lim,D., Blasco, F., Weiner, J. & Strynadka, N.C. (2003) Insights into the respiratory electron transfer pathway from the structure of nitrate reductase A. Nat. Struct. Biol. 10, 681–687. 8. Hackett, N.R. & Bragg, P.D. (1982) The association of two distinct b cytochromes with the respiratory nitrate reductase of Escherichia coli. J. Bacteriol. 154, 719–727. 9. Magalon, A., Lemesle-Meunier, D., Rothery, R.A., Frixon, C., Weiner,J.H.&Blasco,F.(1997)Hemeaxialligationbythehighly conserved His residues in helix II of cytochrome b (NarI) of Escherichia coli nitrate reductase A (NarGHI). J. Biol. Chem. 272, 25652–25658. 10. Blasco, F., Guigliarelli, B., Magalon, A., Asso, M., Giordano, G. & Rothery, R.A. (2001) The coordination and function of the redox centres of the membrane-bound nitrate reductases. Cell Mol. Life Sci. 58, 179–193. 11. Wallace, B.J. & Young, I.G. (1977) Role of quinones in electron transport to oxygen and nitrate in Escherichia coli. Studies with a ubiA – menA – double quinone mutant. Biochem. Biophys. Acta 461, 84–100. 12. Morpeth, F.F. & Boxer, D. (1985) Kinetic analysis of respiratory nitrate reductase from Escherichia coli K12. Biochemistry 24, 40–46. 13. Giordani, R., Buc, J., Cornish-Bowden, A. & Cardenas, M.L. (1997) Kinetics of membrane-bound nitrate reductase A from Escherichia coli with analogues of physiological electron donors. Different reaction sites for menadiol and duroquinol. Eur. J. Biochem. 250, 567–577. 14. Rothery, R.A., Blasco, F., Magalon, A., Asso, M. & Weiner, J.H. (1999) The hemes of Escherichia coli nitrate reductase A (Nar- GHI): potentiometric effects of inhibitor binding to NarI. Bio- chemistry 38, 12747–12757. 15. Zhao, Z., Rothery, R.A. & Weiner, J.H. (2003) Transient kinetic studies of heme reduction in Escherichia coli nitrate reductase A (NarGHI) by menaquinol. Biochemistry 42, 5403–5413. 16. Blasco, F., Pommier, J., Augier, V., Chippaux, M. & Giordano, G. (1992) Involvement of the narJ or narW gene product in the formation of active nitrate reductase in Escherichia coli. Mol. Microbiol. 6, 221–230. 17. Guigliarelli, B., Magalon, A., Asso, M., Bertrand, P., Frixon, C., Giordano, G. & Blasco, F. (1996) Complete coordination of the four Fe-S centres of the b subunit from Escherichia coli nitrate reductase. Physiological, biochemical, and EPR characterization of the site-directed mutants lacking the highest or lowest potential [4Fe-4S] clusters. Biochemistry 35, 4828–4836. 18. Graham, A., Jenkins, H.E., Smith, N.H., Mandrand, M.A., Haddock, B.A. & Boxer, D.H. (1980) The synthesis of formate dehydrogenase and nitrate reductase proteins in various fdh and chl mutants of Escherichia coli. FEMS Microbiol. Lett. 7, 145–151. 19. Buc,J.,Santini,C.L.,Blasco,F.,Giordani,R.,Cardenas,M.L., Chippaux,M.,Cornish-Bowden,A.&Giordano,G.(1995) Kinetic studies of a soluble ab complex of nitrate reductase A from Escherichia coli.Useofvariousab mutants with altered b sub- units. Eur. J. Biochem. 234, 766–772. 20. Lowry, O.H., Rosebrough, N.J., Farr, A.L. & Randall, R.J. (1951) Protein measurement with the folin phenol reagent. J. Biol. Chem. 193, 265–275. 21. Jones, R.W. & Garland, P.B. (1977) Sites and specificity of the reaction of bipyridylium compounds with anaerobic respiratory enzymes of Escherichia coli. Biochem. J. 164, 199–211. 2406 R. Giordani and J. Buc (Eur. J. Biochem. 271) Ó FEBS 2004 22. Buc, J. & Giordani, R. (1998) A spectrophotometric method for kinetic studies with quinone-dependent oxidoreductases. Appli- cation to detection in membranes of nitrate reductase activity with menadione and duroquinone as electron donors. Enzyme Microb. Technol. 22, 165–169. 23. Rothery, R.A., Chatterjee, I., Kiema, G., McDermott, M.T. & Weiner, J. (1998) Hydroxylated naphthoquinones as substrates for Escherichia coli anaerobic reductases. Biochem. J. 332, 35–41. 24. Cleland, W.W. (1967) The statistical analysis of enzyme kinetic data. Adv. Enzymol. 29, 1–34. 25. Polglase, W.J., Pun, W.T. & Withaar, J. (1966) Lipoquinones of Escherichia coli. Biochim. Biophys. Acta 118, 425–426. 26. Westenberg, D.J., Gunsalus, R.P., Ackrell, B.A.C. & Cecchini, G. (1990) Electron transfer from menaquinol to fumarate reductase anchor polypeptide mutants of Escherichia coli. J. Biol. Chem. 265, 19560–19567. 27. Iverson, T.M., Luna-Chavez, C., Croal, L.R., Cecchini, G. & Rees, D.C. (2002) Crystallographic studies of the Escherichia coli quinol-fumarate reductase with inhibitors bound to the quinol- binding site. J. Biol. Chem. 277, 16214–16130. 28. Magalon, A., Rothery, R.A., Lemesle-Meunier, D., Frixon, C., Weiner, J.H. & Blasco, F. (1998) Inhibitor binding within the NarI subunit (cytochrome b nr )ofEscherichia coli nitrate reductase A. J. Biol. Chem. 273, 10851–10856. 29. Rothery,R.A.,Blasco,F.,Magalon,A.&Weiner,J.H.(2001)The diheme cytochrome b subunit (NarI) of Escherichia coli nitrate reductase A (NarGHI): structure, function and interaction with quinols. J. Mol. Microbiol. Biotechnol. 3, 273–283. 30. Rothery, R.A., Blasco, F. & Weiner, J.H. (2001) Electron transfer from heme b L to the [3Fe-4S] cluster of Escherichia coli nitrate reductase A (NarGHI). Biochemistry 40, 5260–5268. Ó FEBS 2004 Two electron transfer pathways in nitrate reductase (Eur. J. Biochem. 271) 2407 . Evidence for two different electron transfer pathways in the same enzyme, nitrate reductase A from Escherichia coli Roger Giordani* and Jean Buc Laboratoire. than the other, as well as inhibition data indicating the possibility of more than one menaquinol binding site in nitrate reductase [15]. The fact that nitrate