Tuning a Nitrate Reductase for Function - The Journal of Biological ...

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2004 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 279, No. 31, Issue of July 30, pp. 32212–32218, 2004 Printed in U.S.A.

Tuning a Nitrate Reductase for Function THE FIRST SPECTROPOTENTIOMETRIC CHARACTERIZATION OF A BACTERIAL ASSIMILATORY NITRATE REDUCTASE REVEALS NOVEL REDOX PROPERTIES* Received for publication, March 9, 2004, and in revised form, May 24, 2004 Published, JBC Papers in Press, May 28, 2004, DOI 10.1074/jbc.M402669200

Brian J. N. Jepson‡, Lee J. Anderson§, Luis M. Rubio¶, Clare J. Taylor‡, Clive S. Butler储, Enrique Flores¶, Antonia Herrero¶, Julea N. Butt‡§, and David J. Richardson‡** From the Centre for Metalloprotein Spectroscopy and Biology, Schools of ‡Biological Sciences and §Chemical Sciences, University of East Anglia, Norwich NR4 7TJ, United Kingdom, the ¶Instituto de Bioquı´mica Vegetal y Fotosı´ntesis, Consejo Superior de Investigaciones Cientı´ficas-Universidad de Sevilla, E-41092 Seville, Spain, and the 储School of Cell and Molecular Biosciences, University of Newcastle, Newcastle upon Tyne NE2 4HH, United Kingdom

Bacterial cytoplasmic assimilatory nitrate reductases are the least well characterized of all of the subgroups of nitrate reductases. In the present study the ferredoxindependent nitrate reductase NarB of the cyanobacterium Synechococcus sp. PCC 7942 was analyzed by spectropotentiometry and protein film voltammetry. Metal and acid-labile sulfide analysis revealed nearest integer values of 4:4:1 (iron/sulfur/molybdenum)/molecule of NarB. Analysis of dithionite-reduced enzyme by low temperature EPR revealed at 10 K the presence of a signal that is characteristic of a [4Fe-4S]1ⴙ cluster. EPRmonitored potentiometric titration of NarB revealed that this cluster titrated as an n ⴝ 1 Nernstian component with a midpoint redox potential (Em) of ⴚ190 mV. EPR spectra collected at 60 K revealed a Mo(V) signal termed “very high g” with gav ⴝ 2.0047 in air-oxidized enzyme that accounted for only 10 –20% of the total molybdenum. This signal disappeared upon reduction with dithionite, and a new “high g” species (gav ⴝ 1.9897) was observed. In potentiometric titrations the high g Mo(V) signal developed over the potential range of ⴚ100 to ⴚ350 mV (Em Mo6ⴙ/5ⴙ ⴝ ⴚ150 mV), and when fully developed, it accounted for 1 mol of Mo(V)/mol of enzyme. Protein film voltammetry of NarB revealed that activity is turned on at potentials below ⴚ200 mV, where the cofactors are predominantly [4Fe-4S]1ⴙ and Mo5ⴙ. The data suggests that during the catalytic cycle nitrate will bind to the Mo5ⴙ state of NarB in which the enzyme is minimally two-electron-reduced. Comparison of the spectral properties of NarB with those of the membranebound and periplasmic respiratory nitrate reductases reveals that it is closely related to the periplasmic enzyme, but the potential of the molybdenum center of NarB is tuned to operate at lower potentials, consistent with the coupling of NarB to low potential ferredoxins in the cell cytoplasm.

Nitrate is a widely used and readily available source of inorganic nitrogen for plants and microorganisms (1). Fixed inorganic nitrogen is mainly supplied to natural environments either from human agricultural or industrial activities or from * This work was supported by Grants P13842 and B17233 from the United Kingdom Biotechnology and Biological Sciences Research Council. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ** To whom correspondence should be addressed. Tel.: 44-1603593250; Fax: 44-1603-592250; E-mail: [email protected].

biological nitrogen fixation. Most of it is converted to nitrate by nitrifying bacteria, and the nitrate then serves as a nitrogen source for assimilation or as a respiratory electron acceptor. Bacterial nitrate reductases are molybdoenzymes that can catalyze the two-electron reduction of nitrate to nitrite and can be classified into three groups according to their localization and function (2). Respiratory membrane-bound nitrate reductases are generally integral membrane protein complexes with the active site located on the cytoplasmic face of the cytoplasmic membrane and are constituted by subunits (e.g. NarI and NarH) that mediate electron transfer from the quinol pool to the catalytic subunit, NarG, which contains a bismolybdopterin guanine dinucleotide (bis-Mo-MGD)1 cofactor and a [4Fe-4S] cluster (3, 4). These membrane-bound nitrate reductases couple quinol oxidation by nitrate to the generation of a transmembrane proton electrochemical gradient, and their synthesis is normally unaffected by ammonium but repressed by oxygen (2). Periplasmic nitrate reductases are also linked to quinol oxidation in respiratory electron transport chains, but they have a range of different functions including the disposal of reducing equivalents during aerobic growth and nitrate respiration in nitrate-limited environments (5, 6). In periplasmic nitrate reductases, electrons from quinol are generally passed through one or two cytochrome c-containing proteins (NapC and NapB) to the catalytic subunit, NapA, that contains a bis-Mo-MGD cofactor and a [4Fe-4S] cluster (7–10). As with the membrane-bound nitrate reductases, synthesis of the periplasmic nitrate reductase is insensitive to ammonium, but the enzyme can be expressed either anaerobically or aerobically depending on the organism (2). The periplasmic and membrane-bound nitrate reductases are catalytically quite distinct (2). This distinctness is reflected in their molecular structures that show that cysteine provides a thiol ligand to the molybdenum atom in the periplasmic enzymes, whereas aspartate provides one or two oxygen ligands to the molybdenum atom in the membrane-bound enzymes (3, 4, 9, 10). As a result the two enzyme groups exhibit quite distinct Mo(V) EPR spectra (8). In contrast to the wealth of spectroscopic and structural data on the two groups of respiratory nitrate reductases, no detailed spectropotentiometric analysis has been carried out on any bacterial assimilatory nitrate reductase, despite the importance of such enzymes in the eutrophication and subsequent spoilage of many fresh and marine water systems (1). Genes or 1 The abbreviations used are: bis-Mo-MGD, bismolybdopterin guanine dinucleotide; NarB, Synechococcus sp. PCC 7942 assimilatory nitrate reductase; PGE, pyrolytic graphite edge; PFV, protein film voltammetry; mT, millitesla.

32212

This paper is available on line at http://www.jbc.org

Bacterial Assimilatory Nitrate Reductase enzymes for assimilatory nitrate reduction have been described from various bacterial sources (11), including several strains of cyanobacteria (12), Klebsiella oxytoca (13), Bacillus subtilis (14), and Azotobacter vinelandii (15). In A. vinelandii and cyanobacteria the enzymes are monomeric cytoplasmic or thylakoid-associated proteins that contain molybdenum, iron, and acid-labile sulfide and use flavodoxin, in the case of A. vinelandii (15), or ferredoxin and flavodoxin, in the case of the cyanobacteria (16, 17), as the physiological electron donor. The synthesis of assimilatory nitrate reductases in bacteria is insensitive to aerobiosis but inhibited by ammonium, contrasting with both the respiratory nitrate reductase groups (11, 12). Cyanobacteria are a highly significant group of phototrophic bacteria in ecological terms as they make a dominant contribution to carbon and nitrogen cycling in many environments on Earth (1). They carry out oxygenic photosynthesis using a photosynthetic apparatus that is similar to that of the chloroplasts of eukaryotic algae and higher plants, of which the cyanobacterial system is considered to represent the phylogenetic ancestor (12). Cyanobacteria preferentially use inorganic nitrogen for growth, being able as a group to assimilate nitrate, nitrite, ammonium, and dinitrogen. The assimilation of nitrate by cyanobacteria takes place in the cytoplasm. Thus, the process first requires the entrance of nitrate into the cell, for which a high affinity active transport system exists. Transport is then followed by the two-electron reduction of nitrate to nitrite catalyzed by the assimilatory nitrate reductase and the further six-electron reduction of nitrite to ammonium by a sirohemecontaining nitrite reductase. The ammonium is then incorporated into carbon skeletons, mainly utilizing the glutamine synthetase (GS)/glutamine:2-oxoglutarate amidotransferase (GOGAT) pathway (12, 18 –20). In recent years, a wealth of knowledge on the genetics of nitrate reduction in cyanobacteria has been obtained, mainly through studies in the unicellular strain Synechococcus sp. PCC 7942 (21–23). Recently the heterologous expression, purification, and kinetic characterization of the ferredoxin-dependent nitrate reductase, NarB, of Synechococcus sp. PCC 7942 was reported (24), which used metal and chemical analysis to confirm the presence of an iron-sulfur center and a bis-MoMGD cofactor and identified cysteine residues important in binding the cluster. However, the molecular nature of the cluster (i.e. [3Fe-4S] or [4Fe-4S]) could not be unambiguously determined. The present paper provides a spectropotentiometric and film voltammetric analysis of NarB, the first such study for any bacterial assimilatory nitrate reductase. This work resolves the nature of the cofactors and reveals some intriguing redox properties of the enzyme. These results are considered in the light of recent studies on quinol-dependent membranebound and periplasmic respiratory nitrate reductase systems and the distinct physiological role of NarB in ferredoxin-dependent nitrate assimilation. MATERIALS AND METHODS

Organisms, Growth Conditions, and Plasmids—The Escherichia coli strain DH5␣ (Stratagene) was used for expression of recombinant protein. E. coli was grown in Luria-Bertani medium. The plasmid pCSLM85, encoding NarB with a hexahistidine tag fused to its C terminus, has been described previously (24). Expression of NarB—E. coli strain DH5␣ was used for the expression of pCSLM85. Cultures were grown overnight to saturation in LuriaBertani (LB) medium containing 100 ␮g/ml ampicillin. An aliquot of the overnight culture was diluted 1:25 in LB medium and grown at 37 °C. At A600 of 0.6, isopropyl-␤-D-thiogalactopyranoside was added to a final concentration of 1 mM, and the cells were grown at 37 °C for 3 h. The induced cells were harvested, washed, and collected by centrifugation. Enzyme Assays and Protein Determinations—Nitrate reductase activity was measured spectrophotometrically by substrate-dependent oxidation of reduced methyl viologen (⑀600 nm ⫽ 13,700 M⫺1 cm⫺1). As-

32213

says (3-ml final volume) were carried out by mixing at 25 °C in anaerobic cuvettes containing 0.5 mM methyl viologen, 25 mM Hepes, 100 mM NaCl, 10% glycerol, pH 8.0, and nitrate, chlorate, selenate, tellurate, or nitrite. Methyl viologen was reduced by the addition of 0.5 mM sodium dithionite, and turnover was initiated by the addition of purified NarB (40 nM). Data were fitted to the Michaelis-Menten description of enzyme kinetics to determine Km and kcat for the tested substrates. Protein determination was by the methods of Markwell et al. (25) or Bradford (26). Purification of NarB—The cell pellet containing NarB was resuspended in 5 volumes (v/w) buffer A (25 mM Hepes, pH 8, 100 mM NaCl, 10 mM imidazole, and 10% glycerol). The cells were then lysed by French press disruption, and the crude extract was centrifuged at 40,000 ⫻ g for 30 min at 4 °C. The supernatant was chromatographed on a 25-ml TALON column charged with Ni2⫹ (BD Biosciences) in buffer A with a superimposed 200-ml linear gradient from 10 to 250 mM imidazole. The fractions were analyzed on SDS-polyacrylamide gels, and the fractions containing NarB were identified through specific interaction with antibodies against histidine tags. The peak fractions of NarB were pooled, buffer-exchanged to remove the imidazole, and concentrated using a Centricon Plus-20 concentrator (molecular mass cutoff, 30,000 Da; Millipore Corp.). The protein was then loaded onto a Superdex 200 HR 10/30 column (Amersham Biosciences) and eluted in 25 mM Hepes, pH 8.0, 100 mM NaCl, and 10% glycerol. All purification steps, except for gel filtration, were carried out at 4 °C. The protein was stored either at ⫺80 °C or in liquid nitrogen because of the lability of NarB at 4 °C for prolonged periods of time. Potentiometric Titrations—Chemical redox titrations were performed with the sample stirred in a glass cell housed in an anaerobic chamber and thermostatted at 4 °C with platinum working and calomel reference electrodes. Potassium ferricyanide was used to fully oxidize the protein to 400 mV, and dithionite was used as the reductant in the potential range of 50 to ⫺550 mV. Titrations were performed in 50 mM Hepes, pH 8.0, 100 mM KCl. Electrode-solution mediation was facilitated by the following mediators at a 10 ␮M concentration: neutral red, 2-hydroxy-1,4-naphthoquinone, safranine T, and phenosafranine. After equilibration at each potential, a 300-␮l sample was removed and frozen in liquid nitrogen, and the EPR spectrum was recorded. All potentials are given relative to the standard hydrogen electrode. Spectroscopy—Electron paramagnetic resonance spectra were recorded on a Bruker EMX system X-band spectrometer equipped with an Oxford Instrument ESR-9 liquid helium flow cryostat and a dual mode EPR cavity. Spin quantification of the EPR signals due to the [4Fe-4S] cluster and Mo(V) was estimated by double integration and comparison with a 2 mM Cu-EDTA standard, all measured under non-saturating conditions. Conditions of measurement are as described in the appropriate figure legends. EPR simulation was carried out using WINEPR Simfonia Version 1.25 (Bruker). Where simulations are presented, experimental and simulated spectra are aligned with a magnetic field range corresponding to a microwave frequency of 9.6800 GHz. UVvisible absorption spectra of purified NarB were obtained on a Hitachi U4001 UV-visible spectrometer. Protein Film Voltammetry—Protein film voltammetry was performed with a three-electrode cell configuration housed in an anaerobic chamber as described previously (27). All potentials are reported with respect to the standard hydrogen electrode. Films of NarB were prepared by coating freshly polished pyrolytic graphite edge (PGE) electrodes or gold electrodes coated with mercaptopropionic acid, with a submicroliter quantity of ⬃40 ␮M enzyme containing 2 mM neomycin. Metal and Sulfide Analysis—Iron and molybdenum content in purified nitrate reductase preparations was determined by electrothermic atomic absorption spectrometry. The iron content was also determined in colorimetric assays using bipyridyl (28), and acid-labile sulfide was determined as described by Rasmussen et al. (29). RESULTS

The Properties of the Synechococcus sp. PCC 7942 NarB Preparations Used in the Spectropotentiometric Analyses—The Synechococcus sp. PCC 7942 NarB was purified as a recombinant His6-tagged protein from E. coli strain DH5␣. The protein migrated as a single 78-kDa protein on SDS-polyacrylamide gels (Fig. 1A, inset) and was the only visible band in Coomassie Blue-stained gels of the samples used for spectroscopy and protein film voltammetry. The metal and acid-labile sulfur content of the preparations (mean of three determinations) was 4:4.2:1.2 (iron/sulfur/molybdenum)/NarB. The enzyme had a

32214

Bacterial Assimilatory Nitrate Reductase

FIG. 1. Spectral properties of Synechococcus sp. PCC 7942 NarB. A, the UV-visible absorption spectrum of NarB (20 ␮M) in 50 mM Hepes (pH 8.0). Solid line, air-oxidized NarB; dashed line, dithionitereduced NarB. Inset, SDS-PAGE analysis of Synechococcus nitrate reductase preparations purified from cells of E. coli strain DH5␣ (pCSLM85). Lane M, molecular mass markers (from top to bottom: 250, 150, 100, 75, 50, 37, and 25 kDa); lane 1, cell extract from uninduced cells; lane 2, cell extract from isopropyl-␤-D-thiogalactopyranoside-induced cells; lane 3, sample after TALON nickel affinity column; lane 4, purified sample after Superdex 200 HR 10/30 column. B, X-band EPR spectrum of NarB (50 ␮M) in buffer 50 mM Hepes (pH 8.0). Spectrum I, “air-oxidized” NarB; spectrum II, NarB reduced with 10 mM dithionite (gray line spectrum is a simulation of the reduced [4Fe-4S]⫹ cluster signal with parameters of g1,2,3 ⫽ 2.058, 1.951, 1.908); spectrum III, NarB reoxidized with 10 mM nitrate. The conditions of measurement were as follows: temperature, 10 K; microwave power, 2 milliwatts; microwave frequency, 9.68 GHz; and modulation amplitude, 1 mT.

kcat of 80 s⫺1 when assayed using reduced methylviologen as the electron donor and nitrate as the electron acceptor. The Km for nitrate was 0.05 mM yielding a kcat/Km of 1,600 s⫺1 mM⫺1. Chlorate served as an alternative substrate for which the kcat was lower than for nitrate (5 s⫺1) and the Km was much higher (2.5 mM), giving a kcat/Km of 2 s⫺1 mM⫺1, which is 800⫻ lower than that for nitrate. Selenate was also able to serve as an alternative substrate with a similar Km to chlorate (2.5 mM) but with a much lower kcat of ⬃1 s⫺1 (kcat/Km ⫽ 0.4 s⫺1 mM⫺1). NarB turnover was sensitive to inhibition by azide and cyanide. Azide inhibition could be modeled for a simple competitive inhibitor with a KI of 13 mM. Cyanide was a much more potent inhibitor, but the pattern of inhibition was more complex and suggestive of mixed inhibition that could be defined by an I50 of 0.01 mM. NarB was also stimulated by chloride anions exhibiting a 2- and 4-fold increase in nitrate reductase activity in the

presence of 100 and 500 mM NaCl, KCl, or NH4Cl. Spectropotentiometric Characterization of the NarB IronSulfur Cluster—The UV-visible spectrum of air-oxidized NarB exhibits broad absorption shoulders at around 320, 400, and 450 nm (Fig. 1A). The features at 400 – 450 nm decrease in intensity on reduction with dithionite and are indicative of the presence of an iron-sulfur cluster, consistent with previous studies on this enzyme (24). The EPR spectrum of air-oxidized NarB showed weak features in the g ⫽ 2.1 region that are characteristic of the presence of a [3Fe-4S]1⫹ cluster, but integration of this signal suggested that it maximally accounted for 0.05 spin/NarB (Fig. 1B, spectrum I). However, on reduction with dithionite an intense rhombic signal developed with simulated parameters of g1,2,3 ⫽ 2.058, 1.951, 1.908. Such signals are characteristic of a S ⫽ 1⁄2 [2Fe-2S]1⫹ or [4Fe-4S]1⫹ cluster. However, the temperature dependence (the signal did not persist above 40 K) and the metal/sulfide ratio of approximately 4:4:1 (iron/sulfur/molybdenum) suggest that the signal arises from a [4Fe-4S]1⫹ cluster. Spin quantification revealed that the signal accounted for ⬃1 spin/NarB showing that the iron-sulfur content of the NarB could be entirely accounted for by this signal. An additional signal was observed at g ⬃ 2, which also increased in intensity in the dithionite-reduced sample (Fig. 1B, spectrum II). This signal arises from Mo(V) and was studied in more detail at 60 K (see below). The addition of nitrate to reduced enzyme resulted in the disappearance of the [4Fe4S]1⫹ EPR signal, showing that the cluster is acting as a functional redox center facilitating the reduction of the added nitrate (Fig. 1B, spectrum III). NarB samples were electrochemically poised at a range of potentials between 400 and ⫺400 mV. The [4Fe-4S]1⫹ EPR signal developed over the potential range of 0 to ⫺400 mV, and the titration could be fitted with a simple n ⫽ 1 Nernstian curve from which a midpoint potential of ⫺190 mV for the [4Fe-4S]2⫹/1⫹ couple could be derived (Fig. 2). Spectropotentiometric Analysis of the NarB Molybdenum Center—The EPR spectrum collected at 60 K of air-oxidized NarB was dominated by a rhombic and highly anisotropic signal with simulation parameters as follows: g1,2,3 ⫽ 2.0228, 1.9983, 1.9930; gav ⫽ 2.0047; anisotropy (g1 ⫺ g3) ⫽ 0.030; and rhombicity (g1 ⫺ g2/g1 ⫺ g3) ⫽ 0.82 (Fig. 3A). The signal was split by a weakly interacting I ⫽ 1⁄2 nucleus (A1,2,3 ⫽ 0.68, 0.75, 0.68 mT) that remained in spectra collected from enzyme exchanged into D2O. This signal was assigned to a Mo(V) species that is similar to a signal termed very high g and observed in Paracoccus pantotrophus periplasmic nitrate reductase (8), and this signal accounted for 10 –20% of the total molybdenum in NarB, depending on the preparation (four independent preparations were examined). Reduction of the enzyme with a 20-fold excess of dithionite led to the disappearance of this signal, but a new Mo(V) signal appeared that also accounted for ⬃20% of total molybdenum in the sample and that could be described by the following simulation parameters: g1,2,3 ⫽ 1.9970, 1.9902, 1.9820; gav ⫽ 1.9897; anisotropy (g1 ⫺ g3) ⫽ 0.015; and rhombicity (g1 ⫺ g2/g1 ⫺ g3) ⫽ 0.45 (Fig. 3B). The signal was split by two weakly interacting I ⫽ 1⁄2 nuclei, one of which was only resolved in the g1 feature (A11,2,3 ⫽ 0.65, 0.60, 0.50 mT and A21 ⫽ 0.22 mT) and is similar to a signal termed high g and observed in P. pantotrophus periplasmic nitrate reductase (8). Neither Mo(V) signal was present at significant levels when the NarB preparations were equilibrated with a redox mediator mixture under anaerobic conditions at electropositive potentials (⬃400 mV). As the potential was lowered to below 0 mV the high g signal began to appear and accounted for ⬃1 spin/mol of NarB when fully developed in samples poised below ⫺280 mV (Fig. 4A). No decrease in intensity was observed in samples poised at ⫺550 mV. Addition of nitrate to such samples


FIG. 2. Variation of Synechococcus nitrate reductase iron-sulfur center EPR signal intensity with potential. A, selected spectra from different potentials in the titration (arrows indicate decreasing potential). Spectral parameters are g1 ⫽ 2.058, g2 ⫽ 1.951, and g3 ⫽ 1.908. B, plot of normalized signal intensities at g1 (●, peak), g2 (‚, peak-to-trough), and g3 (䡺, trough) versus potential. The data are fitted with an n ⫽ 1 Nernstian curve with Em ⫽ ⫺190 mV. Conditions of measurement were as follows: NarB (50 ␮M) in 50 mM Hepes buffer (pH 8.0); temperature, 10 K; microwave power, 2 milliwatts; microwave frequency, 9.68 GHz; and modulation amplitude, 1 mT.

resulted in a decrease in intensity of the Mo(V) signal. The titration was fitted with an n ⫽ 1 Nernstian curve, from which a midpoint potential of ⫺150 mV for the Mo6⫹/5⫹ couple was derived (Fig. 4B). The very high g signal was not seen in the anaerobic titration and was only apparent in the air-equilibrated samples. Protein Film Voltammetry of NarB—Films of NarB on PGE electrodes gave rise to catalytic currents for which the essential features were retained during potential cycles after the second scan, with scan rates up to 20 mV s⫺1 and electrode rotation rates from 1000 to 5500 rpm. No faradaic response was observed in the absence of enzyme at all nitrate concentrations investigated. Thus the voltammograms can be considered to reflect steady-state behavior of the enzyme. Analysis of the cyclic voltammogram under substrate-limited conditions (e.g. 0.01 mM NaNO3 in Fig. 5A) shows that there is no significant current (and hence electron transfer to nitrate) until the applied potential is lowered to below ⫺250 mV. Decreasing the potential further then results in a sharp increase in current amplitude until a local maximum is reached at around ⫺450 mV. The presence of maxima and minima in the cyclic voltammograms using both PGE electrodes and gold electrodes coated

32215

FIG. 3. X-band EPR signals from the Mo(V) cofactor of Synechococcus nitrate reductase. A, air-oxidized spectrum of NarB (50 ␮M) in 50 mM Hepes buffer (pH 8.0) and simulated spectrum with parameters of g1,2,3 ⫽ 2.0228, 1.9983, 1.9930 and A1,2,3 ⫽ 0.68, 0.75, 0.68 mT. B, sample from A treated with 1 mM dithionite and simulated spectrum with parameters of g1,2,3 ⫽ 1.9970, 1.9902, 1.9820; A11,2,3 ⫽ 0.65, 0.60, 0.50 mT; and A21 ⫽ 0.22 mT. Conditions of measurement were as follows: temperature, 60 K; microwave power, 2 milliwatts; microwave frequency, 9.68 GHz; and modulation amplitude, 0.5 mT. The gray line spectra are simulations of the Mo(V) signals.

with mercaptopropionic acid cannot be due to substrate depletion because the feature is apparent on both the forward and reverse sweeps. Moreover, the local minimum is not due to product inhibition because addition of 5 mM nitrite did not enhance this feature. The amplitude of the catalytic current measured at ⫺450 mV was dependent on nitrate concentration, and this could be fitted to a simple Michaelis-Menten description to yield a Km for nitrate of 0.08 mM. This was similar to the Km obtained in solution state assays (0.05 mM) using dithionite-reduced methylviologen as the electron donor, which provides electrons at a similar potential. This provided evidence that forming a film on the electrode surface had not structurally perturbed the active site of the enzyme. Increasing the substrate concentration to 1 mM nitrate resulted in the local maxima and minima being masked from the catalytic wave form; this was primarily due to the catalytic current at the negative potential extreme continuing to increase at nitrate concentrations where the catalytic current at more positive potentials had reached a nitrateindependent value. The Km for nitrate measured at ⫺650 mV increased to 0.350 mM, suggesting a redox potential-dependent structural change in the catalytic site of NarB that decreases its affinity for nitrate. DISCUSSION

This work has presented the first spectropotentiometric study of a bacterial assimilatory nitrate reductase and now

32216


FIG. 4. Variation of Synechococcus nitrate reductase Mo(V) EPR signal intensity with potential. A, selected spectra from different potentials in the titration (arrows indicate decreasing potential). The spectral parameters are g1,2,3 ⫽ 1.9970, 1.9902, 1.9820. B, plot of normalized signal intensity (measured from the trough of the g3 feature) versus potential. The data were fitted with an n ⫽ 1 Nernstian curve with Em ⫽ ⫺150 mV. Conditions of measurement were as follows: NarB (50 ␮M) in 50 mM Hepes buffer (pH 8.0); temperature, 60 K; microwave power, 2 milliwatts; microwave frequency, 9.68 GHz; and modulation amplitude, 0.5 mT.

provides the opportunity for the first comparison of the redox properties of members from the three main subgroups of bacterial nitrate reductases, the assimilatory, periplasmic, and membrane-bound enzymes (2). The study of the Synechococcus sp. PCC 7942 NarB has shown that the enzyme binds a single [4Fe-4S]2⫹/1⫹ cluster. The only previous EPR study on a bacterial assimilatory nitrate reductase, that of A. vinelandii, failed to detect any iron-sulfur center EPR signals despite the presence of spectroscopic features characteristic of an iron-sulfur center in the UV-visible spectrum (30). In this context we did note that this center was rather labile and could not be detected in EPR spectra of NarB samples that had been stored for a few days at 4 °C and may have lost iron, which may be the reason for previous failure in detecting such a Fe-S cluster (24). The Km for nitrate (0.05 mM) is much lower than reported previously (in the range of 0.7–2.1 mM), which may reflect the full cofactor loading of these preparations (21). This lower Km is also a more physiologically relevant level for nitrate reduction, as the steady-state concentration of nitrate reported for nitrate-grown Synechococcus (previously called Anacystis nidulans) cells is 20 –30 ␮M (31). The presence of a [4Fe-4S] cluster and a bis-Mo-MGD cofactor on the catalytic subunit is a characteristic that Synechococcus sp. PCC 7942 NarB shares with the catalytic subunits of the respiratory membrane-bound nitrate reductase and the periplasmic nitrate reductases. However, it is the periplas-

FIG. 5. Protein film voltammetry of Synechococcus nitrate reductase. A, protein film voltammetry of NarB in 10, 250, and 1000 ␮M nitrate. Voltammograms were recorded at separate NarB films, and the responses subsequently were normalized to the current that each film displayed in 10 mM nitrate (at ⫺700 mV). The dashed line illustrates a typical voltammogram recorded at a freshly polished PGE electrode in 1000 ␮M nitrate in the absence of a NarB film. Voltammetry was performed in 50 mM Hepes, 100 ␮M EGTA, 2 mM neomycin, pH 8.0, at 30 °C with a scan rate of 5 mV s⫺1 and electrode rotation at 3000 rpm. B shows the dependence of current on nitrate concentration measured at ⫺450 and ⫺650 mV. SHE, standard hydrogen electrode.

mic nitrate reductase that NarB appears to be most closely related to in evolutionary terms (2), and this is confirmed by the catalytic and spectropotentiometric properties. In terms of the catalytic properties NarB appears highly specific for nitrate, the kcat/Km for nitrate being 800-fold higher than for chlorate. This is also a property of periplasmic nitrate reductases but not of membrane-bound nitrate reductase, where the kcat/Km for nitrate and chlorate are comparable (32, 33). Likewise both the Synechococcus sp. PCC 7942 NarB and the periplasmic nitrate reductases are relatively insensitive to competitive inhibition by azide (KINarB ⫽ 13 mM (this work); KINapA ⫽ 11 mM (8)). By contrast the membrane-bound nitrate reductases are highly sensitive to competitive inhibition by this anion (KI ⫽ 5 ⫻ 10⫺4 mM). In the membrane-bound nitrate reductase the iron-sulfur cluster is coordinated by three cysteines and one histidine, whereas in the periplasmic nitrate reductase it is coordinated by four cysteines. The latter is also predicted to be the case for NarB, in which the spacing of the conserved cysteines in the N-terminal region, which is predicted to bind the cluster, is

Bacterial Assimilatory Nitrate Reductase similar to that of NapA, and the conserved coordinating histidine present in membrane-bound enzymes is absent, as highlighted below for NarB, Rhodobacter sphaeroides NapA, and E. coli NarG.

32217

(Em) of the Mo6⫹/5⫹ couple is ⫺150 mV, and the Mo5⫹ state could not be further reduced at potentials down to ⫺550 mV. Precipitation of the enzyme in solution prevented reduction below this point; thus an upper limit of ⫺550 mV can be placed

SEQUENCE 1

Mutation of these four conserved cysteines in the N-terminal region of NarB showed that they are required for enzyme activity (24). The similarity in the NarB and periplasmic nitrate reductase iron-sulfur clusters is also reflected in rather similar redox properties, with the NarB cluster having a midpoint redox potential of ⫺190 mV that compares with ⫺160 mV for the equivalent cluster in the periplasmic nitrate reductase of Paracoccus denitrificans (7). The Mo(V) EPR signals identified in NarB also have strong similarity to signals detected in the P. pantotrophus periplasmic nitrate reductase and A. vinelandii assimilatory nitrate reductase and termed very high g and high g on the basis of the gav values (8, 16, 30, 34). The very high g signals are most likely to arise from inactive forms of the enzyme. In the periplasmic nitrate reductase they arise following exposure of cyanide-treated enzyme to oxidizing conditions, and in NarB they arise following exposure to air. In both cases the inactivation is reversible, with the NarB very high g signal not being apparent at any potential in samples maintained under strictly anaerobic conditions. The high g signal dominated in NarB samples maintained under anaerobic conditions and accounted for 1 spin/NarB when fully developed. The similarity of the high g signals in NarB and the periplasmic nitrate reductase probably reflects very similar Mo(V) coordination. The cysteine residue known to provide a thiol ligand to molybdenum in the structurally defined periplasmic nitrate reductases of R. sphaeroides (see below) and Desulfovibrio desulfuricans (9, 10) is conserved in NarB, and electron nuclear double resonance studies on the periplasmic nitrate reductase suggest that it is one of the non-exchangeable cysteine ␤-methylene protons that cause splitting of the Mo(V) EPR signals (35). By contrast, the aspartate residue known to provide one or two oxygen ligands to molybdenum in the structurally defined E. coli membrane-bound nitrate reductase (3, 4) is absent

SEQUENCE 2

in NarB as highlighted below. Thus, as discussed previously (24), NarB, like the periplasmic nitrate reductase (9), is likely to have a molybdenum coordination sphere that minimally comprises a dithiol provided by each molybdopterin guanine dinucleotide moiety and one thiol provided by the coordinating cysteine. One oxo/hydroxo/water group may also be present depending on the redox state of the enzyme. Such molybdenum coordination is consistent with preliminary extended x-ray absorption fine structure studies on NarB.2 The spectroscopic similarity in the Synechococcus sp. PCC 7942 NarB and periplasmic nitrate reductase Mo(V) signals does not appear to extend to the redox properties of the Mo6⫹/5⫹ and Mo5⫹/4⫹ couples. In NarB the midpoint redox potential

2 L. J. Anderson, C. S. Butler, J. Charnock, and D. J. Richardson, unpublished data.

on the Em of the Mo5⫹/4⫹ couple. By contrast, in the periplasmic enzyme from P. pantotrophus the high g Mo6⫹/5⫹ couple has a midpoint potential above 400 mV, with the Em of the Mo5⫹/4⫹ couple lying closer to that of the NarB Mo6⫹/5⫹ couple at around ⫺120 mV. The Em of the NarB Mo6⫹/5⫹ couple is also much lower that of the membrane-bound nitrate reductase, which lies in the range of 250 to ⬎450 mV depending on the pH (32, 36). This difference of ⬃400 – 600 mV in the Em of the Mo6⫹/5⫹ couple of the assimilatory and respiratory nitrate reductases may underlie differences in the operating potential of these enzymes as revealed by protein film voltammetry (32, 37, 38). In NarB the catalysis of nitrate reduction does not initiate until the potential falls below ⫺200 mV. By contrast, in the respiratory membrane-bound nitrate reductase PFV has revealed that catalysis is initiated below 150 mV and has an activity maximum at around 20 mV (32, 37). In PFV of the periplasmic nitrate reductases of P. pantotrophus3 and R. sphaeroides (38), catalysis is initiated at below ⬃100 mV and has an activity maximum of approximately ⫺100 mV (38). The difference in the operating potentials of the assimilatory and respiratory nitrate reductases can be rationalized by consideration of the physiological electron transport pathways to these enzymes. The respiratory enzymes operate on the oxidizing side of the menaquinol or ubiquinol pools, for which the Em values of ubiquinone/ubiquinol (UQ/UQH2) and (menaquinone/menaquinol (MQ/MQH2) are approximately 60 and ⫺80 mV, respectively (2). By contrast, NarB takes the electrons required for nitrate reduction from low potential ferredoxins, for which the [2Fe-2S]2⫹/1⫹ Em is approximately ⫺400 mV (39), and thus can provide a thermodynamic driving force comparable with that in PFV experiments where catalysis reaches a maximum at electrode potentials of approximately ⫺450 mV. Clearly the Synechococcus sp. PCC 7942 NarB enzyme would not be able to operate in a respiratory chain on the oxidizing side of the quinol pool and so is tuned for its operation in cytoplasmic reductive nitrate assimilation. The catalytic cycle of NarB cannot be forwarded with certainty at present; however, some issues can be considered in the light of the data presented. In the potential range at which NarB catalyzes nitrate reduction, the [4Fe-4S]2⫹/1⫹ clusters in the protein film will be almost entirely reduced, and the molybdenum ions will be predominantly in the Mo5⫹ state. Thus the NarB enzymes will hold minimally two electrons at this point. The low potential of the Mo5⫹/4⫹ couple (⬍⫺500 mV) makes it unlikely that this could be reached with physiological electron donors in the cell cytoplasm, although driving the enzyme into this redox state may account for the very low potential modulation of the catalytic wave observed in the cyclic voltammograms. However, it is most likely that under physiological conditions nitrate binds predominantly to the Mo5⫹ state. This binding would then raise the potential of the Mo5⫹/4⫹ couple allowing a second electron to pass from the reduced iron-sulfur cluster to the molybdenum ion, thus

3

A. Gates, D. J. Richardson, and J. N. Butt, unpublished data.

32218


providing the two electrons required for nitrate reduction. The midpoint potential of the nitrate-bound Mo5⫹/4⫹ couple cannot be derived from equilibrium spectropotentiometry, but from the midpoint of the catalytic wave collected under substrate-limited conditions in the PFV it can be predicted to be in the region of ⫺350 mV. The peak of activity at low potential revealed by PFV arises from a reversible attenuation of activity on further reduction of the enzyme. Similar behavior is observed in PFV of the respiratory nitrate reductases, although at much higher potentials. This behavior has been attributed to the reduction of Mo5⫹ to Mo4⫹, but the proximity of the [4Fe-4S] cluster reduction potential to the window where the redox switch is operating has made an unambiguous assignment difficult. In NarB the Em of the [4Fe-4S]2⫹/1⫹ couple lies well outside the window where the attenuation occurs. Thus the attenuation is assigned to the Mo5⫹/4⫹ couple in NarB, for which the upper limit is ⫺550 mV from spectropotentiometric titrations. In conclusion, the high level of sequence identity and the similarity between NarB and the periplasmic nitrate reductases suggest a close evolutionary relationship that places the assimilatory and periplasmic nitrate reductases in a distinct evolutionary class of bis-Mo-MGD enzymes to the membranebound nitrate reductase. This view is supported by the similarity of the Mo(V) EPR spectra of these two enzymes. At present it cannot be predicted with certainty whether the assimilatory enzyme evolved from the periplasmic respiratory enzyme through loss of the signal peptide for export or, viceversa, through the gain of the signal peptide for export. However, whatever the evolutionary order of events may be, it appears that the divergence of function was accompanied by an evolution of protein structure that has served to modulate the operating potential of the molybdenum center and to tune it to its physiological electron donation system. Despite different operating potentials and molybdenum coordination spheres in the various nitrate reductases, the mechanistic observations suggest the importance of Mo5⫹ binding nitrate and optimizing the catalytic cycle. REFERENCES 1. Falkowski, P. G. (1997) Nature 387, 272–275 2. Richardson, D. J., Berks, B. C., Russell, D. A., Spiro, S., and Taylor, C. J. (2001) Cell. Mol. Life Sci. 58, 165–178 3. Jormakka, M., Richardson, D., Byrne, B., and Iwata, S. (2004) Structure 12, 95–104 4. Bertero, M. G., Rothery, R. A., Palak, M., Hou, C., Lim, D., Blasco, F., Weiner, J. H., and Strynadka, N. C. (2003) Nat. Struct. Biol. 10, 681– 687

5. Ellington, M. J., Bhakoo, K. K., Sawers, G., Richardson, D. J., and Ferguson, S. J. (2002) J. Bacteriol. 184, 4767– 4774 6. Potter, L. C., Millington, P., Griffiths, L., Thomas, G. H., and Cole, J. A. (1999) Biochem. J. 344, 77– 84 7. Breton, J., Berks, B. C., Reilly, A., Thomson, A. J., Ferguson, S. J., and Richardson, D. J. (1994) FEBS Lett. 345, 76 – 80 8. Butler, C. S., Charnock, J. M., Bennett, B., Sears, H. J., Reilly, A. J., Ferguson, S. J., Garner, C. D., Lowe, D. J., Thomson, A. J., Berks, B. C., and Richardson, D. J. (1999) Biochemistry 38, 9000–9012 9. Dias, J. M., Than, M. E., Humm, A., Huber, R., Bourenkov, G. P., Bartunik, H. D., Bursakov, S., Calvete, J., Caldeira, J., Carneiro, C., Moura, J. J., Moura, I., and Romao, M. J. (1999) Structure Fold. Des. 7, 65–79 10. Arnoux, P., Sabaty, M., Alric, J., Frangioni, B., Guigliarelli, B., Adriano, J. M., and Pignol, D. (2003) Nat. Struct. Biol. 10, 928 –934 11. Lin, J. T., and Stewart, V. (1998) Adv. Microb. Physiol. 39, 1–30 12. Flores, E., and Herrero, A. (1994) in The Molecular Biology of Cyanobacteria (Bryant, D. A., ed) pp. 487–517, Kluwer Academic Publishers, Dordrecht, The Netherlands 13. Lin, J. T., Goldman, B. S., and Stewart, V. (1993) J. Bacteriol. 175, 2370 –2378 14. Ogawa, K., Akagawa, E., Yamane, K., Sun, Z. W., LaCelle, M., Zuber, P., and Nakano, M. M. (1995) J. Bacteriol. 177, 1409 –1413 15. Gangeswaran, R., and Eady, R. R. (1996) Biochem. J. 317, 103–108 16. Mikami, B., and Ida, S. (1984) Biochim. Biophys. Acta 791, 294 –304 17. Hirasawa, M., Rubio, L. M., Griffin, J. L., Flores, E., Herrero, A., Li, J., Kim, S. K., Hurley, J. K., Tollin, G., and Knaff, D. B. (2004) Biochim. Biophys. Acta 1608, 155–162 18. Luque, I., Flores, E., and Herrero, A. (1993) Plant Mol. Biol. 21, 1201–1205 19. Omata, T., Andriesse, X., and Hirano, A. (1993) Mol. Gen. Genet. 236, 193–202 20. Luque, I., Flores, E., and Herrero, A. (1994) Biochim. Biophys. Acta 1184, 296 –298 21. Rubio, L. M., Herrero, A., and Flores, E. (1996) Plant Mol. Biol. 30, 845– 850 22. Rubio, L. M., Flores, E., and Herrero, A. (1998) J. Bacteriol. 180, 1200 –1206 23. Rubio, L. M., Flores, E., and Herrero, A. (1999) FEBS Lett. 462, 358 –362 24. Rubio, L. M., Flores, E., and Herrero, A. (2002) Photosynth. Res. 72, 13–26 25. Markwell, M. A., Haas, S. M., Bieber, L. L., and Tolbert, N. E. (1978) Anal. Biochem. 87, 206 –210 26. Bradford, M. (1976) Anal. Biochem. 72, 248 –254 27. Butt, J. N., Anderson, L. J., Rubio, L. M., Richardson, D. J., Flores, E., and Herrero, A. (2002) Bioelectrochemistry 56, 17–18 28. Eady, R. R., Robson, R. L., Richardson, T. H., Miller, R. W., and Hawkins, M. (1987) Biochem. J. 244, 197–207 29. Rasmussen, T., Berks, B. C., Sanders-Loehr, J., Dooley, D. M., Zumft, W. G., and Thomson, A. J. (2000) Biochemistry 39, 12753–12756 30. Gangeswaran, R., Lowe, D. J., and Eady, R. R. (1993) Biochem. J. 289, 335–342 31. Lara, C., Romero, J. M., and Guerrero, M. G. (1987) J. Bacteriol. 169, 4376 – 4378 32. Anderson, L. J., Richardson, D. J., and Butt, J. N. (2001) Biochemistry 40, 11294 –11307 33. Craske, A. L., and Ferguson, S. J. (1986) Eur. J. Biochem. 158, 429 – 436 34. Bennett, B., Berks, B. C., Ferguson, S. J., Thomson, A. J., and Richardson, D. J. (1994) Eur. J. Biochem. 226, 789 –798 35. Butler, C. S., Fairhurst, S. A., Ferguson, S. J., Thomson, A. J., Berks, B. C., Richardson, D. J., and Lowe, D. J. (2002) Biochem. J. 363, 817– 823 36. Magalon, A., Asso, M., Guigliarelli, B., Rothery, R. A., Bertrand, P., Giordano, G., and Blasco, F. (1998) Biochemistry 37, 7363–7370 37. Elliott, S. J., Hoke, K. R., Heffron, K., Palak, M., Rothery, R. A., Weiner, J. H., and Armstrong, F. A. (2004) Biochemistry 43, 799 – 807 38. Frangioni, B., Arnoux, P., Sabaty, M., Pignol, D., Bertrand, P., Guigliarelli, B., and Leger, C. (2004) J. Am. Chem. Soc. 126, 1328 –1329 39. Bottin, H., and Lagoutte, B. (1992) Biochim. Biophys. Acta 1101, 48 –56