Biochemical Journal

Research article

A composite biochemical system for bacterial nitrate and nitrite assimilation as exemplified by Paracoccus denitrificans

Andrew J. Gates , Victor M. Luque-Almagro , Alan D. Goddard , Stuart J. Ferguson , M. Dolores Roldán , David J. Richardson


The denitrifying bacterium Paracoccus denitrificans can grow aerobically or anaerobically using nitrate or nitrite as the sole nitrogen source. The biochemical pathway responsible is expressed from a gene cluster comprising a nitrate/nitrite transporter (NasA), nitrite transporter (NasH), nitrite reductase (NasB), ferredoxin (NasG) and nitrate reductase (NasC). NasB and NasG are essential for growth with nitrate or nitrite as the nitrogen source. NADH serves as the electron donor for nitrate and nitrite reduction, but only NasB has a NADH-oxidizing domain. Nitrate and nitrite reductase activities show the same Km for NADH and can be separated by anion-exchange chromatography, but only fractions containing NasB retain the ability to oxidize NADH. This implies that NasG mediates electron flux from the NADH-oxidizing site in NasB to the sites of nitrate and nitrite reduction in NasC and NasB respectively. Delivery of extracellular nitrate to NasBGC is mediated by NasA, but both NasA and NasH contribute to nitrite uptake. The roles of NasA and NasC can be substituted during anaerobic growth by the biochemically distinct membrane-bound respiratory nitrate reductase (Nar), demonstrating functional overlap. nasG is highly conserved in nitrate/nitrite assimilation gene clusters, which is consistent with a key role for the NasG ferredoxin, as part of a phylogenetically widespread composite nitrate and nitrite reductase system.

  • nitrate reductase
  • nitrate transport
  • nitrite reductase
  • nitrogen assimilation
  • Paracoccus denitrificans


The importance of inorganic nitrate as a key nutritional component of the global nitrogen cycle, particularly for marine and freshwater autotrophic phytoplankton, is long recognized. This highly soluble anion can make a significant environmental impact, supporting accelerated biomass formation or ‘blooms’ in nitrate- and phosphate-polluted water courses that may have an important role as CO2 sinks. Accordingly, the biochemistry of nitrate assimilation has been well studied in cyanobacteria where assimilatory nitrate reduction is functionally linked to photosynthetic processes, and both nitrate and nitrite reductases use photosynthetically reduced ferredoxin as electron donor. In contrast, the utilization of nitrate by heterotrophic bacteria has received less attention. The ability of heterotrophic bacteria and archaea to metabolize nitrate or nitrite as the sole nitrogen source for growth is phylogenetically widespread. However, only a few physiological, genetic and biochemical studies have been performed, most notably early studies on Enterobacter aerogenes [1] and further studies on the γ-proteobacterium Klebsiella oxytoca [2], the photoheterotroph Rhodobacter capsulatus [3], the Gram-positive bacterium Bacillus subtilis [4] and the diazotroph Azotobacter vinelandii [5]. In contrast with cyanobacterial enzymes, the assimilatory reductases from most heterotrophic bacteria are thought to be dependent on the cytoplasmic reduced pyridine nucleotide pool, which enables them to be coupled to organic carbon catabolism [6]. In fact, previous studies suggest that heterotrophic bacterial species that can utilize nitrate for the biosynthesis of essential cellular components during growth may also be significant consumers of inorganic nitrogen globally, particularly in environments where there are high concentrations of dissolved organic carbon relative to dissolved organic nitrogen [79]. This is due to the high bioenergetic demand for reducing equivalents required for the assimilatory reduction of nitrate to ammonia, which requires eight electrons: Embedded Image(1) Embedded Image(2)

Consequently, assimilatory nitrate reduction is a good route for disposal of the excess reductant present in a reduced organic carbon pool [10]. However, the biochemical mechanism by which heterotrophic bacterial assimilatory nitrate reductases access this pool of cellular reductant is not well understood, particularly because analysis of the primary structure of these proteins suggests that they do not have a NAD(P)H-binding domain [2,11].

In addition to being a substrate for nitrogen assimilation in heterotrophs, nitrate can also be a substrate for anaerobic respiration, for example in denitrifying bacteria that can reduce nitrate, via nitrite, nitric oxide and nitrous oxide, to dinitrogen gas, and enterobacteria that can reduce nitrate to ammonium [1,12]. One of the paradigm heterotrophic denitrifiers, Paracoccus denitrificans, synthesizes two heterotrimeric respiratory ubiquinol/nitrate oxidoreductases, Nar and Nap (Figure 1). The membrane-bound enzyme (NarGHI) reduces nitrate as the first step of growth-linked anaerobic denitrification, whereas the periplasmic system (NapABC) serves to dissipate excess reducing equivalents formed during aerobic growth [12,13]. These enzymes have been studied at the biochemical level and derive electrons from the membrane-confined ubiquinol pool [14,15], which can be coupled to NADH generated from oxidative metabolism via the NADH-ubiquinone oxidoreductase. In P. denitrificans NarGHI, the active site for nitrate reduction is exposed to the cytoplasm. Therefore it is dependent on a nitrate transport protein NarK, a fusion protein of the two transmembrane domains NarK1 and NarK2, to deliver nitrate into the cell (Figure 1). These two functional components have putative roles in nitrogen oxyanion trafficking: NarK1 is a proposed proton-linked nitrate importer, and NarK2 is a putative nitrate/nitrite antiporter [16,17].

Figure 1 Genetic and proposed functional organisation of the assimilatory nitrate reductase system from P. denitrificans

(A) The nas gene cluster includes seven open reading frames, nasTSABGHC. The gene products have the following putative roles: two regulatory components (nasS and nasT); two nitrogen oxyanion transporters (nasA and nasH); nitrate and nitrite reductases (nasC and nasB respectively); and a ferredoxin (nasG). (B) Three distinct nitrate reductase systems are present in P. denitrificans. The periplasmic (NapABC) and membrane-bound (NarGHI) enzymes are ubiquinol-dependent respiratory nitrate reductases. NarK, a nitrate importer, moves nitrate into the cytoplasm and also exports nitrite, the product of nitrate reduction, to the periplasm to support respiratory denitrification. For assimilatory nitrate reduction, electrons probably derived from NADH within the FAD-containing nitrite reductase (NasB) flow to the nitrate reductase (NasC), possibly via the small ferredoxin (NasG). In the Nas system there are two transporters, NasA and NasH, which are predicted to be involved in nitrate and nitrite transport respectively. Note that there is no common nomenclature for assimilatory nitrate and nitrite reductase genes in prokaryotes. In the case of nitrite reductase genes, the ‘nir’ prefix is quite widely used, but since P. denitrificans has a separate respiratory nitrite reductase nir gene cluster, this term would be inappropriate. Likewise the ‘nar’ prefix is sometimes used for the assimilatory nitrate reductase genes, but this would also be inappropriate for P. denitrificans, which also has a nar gene cluster encoding the respiratory nitrate reductase system. Hence we have adopted the ‘nas’ prefix for the genes encoding the assimilatory nitrate and nitrite reductase system of P. denitrificans. The term ‘nH+’ indicates that the number of protons (n) moved across the membrane is not known.

Paracoccus species can also assimilate nitrate via a third cytoplasmic reductase that has not yet been characterized, but is known to be distinct from the two respiratory systems [18]. In general, a bacterial nitrate assimilation system (Nas) involves a cytoplasmic molybdenum-dependent nitrate reductase that reduces nitrate to nitrite (eqn 1), which is further reduced to ammonium by a sirohaem-dependent nitrite reductase (eqn 2) [2]. Similarly to the respiratory Nar system, the cytoplasmic assimilatory system is also dependent on nitrate transport into the cell. However, a major biochemical conundrum is that it is not clear how the assimilatory nitrate reductase is coupled to NADH oxidation. This is because primary sequence analysis of bacterial assimilatory nitrate reductases suggests that, in contrast with the nitrite reductases, they do not possess an NADH-binding domain [2,11]. The absence of such a site is unimportant for photoautotrophic metabolism in cyanobacteria, where the electron donor is ferredoxin that is photoreduced by photosystem I [19], but it is an important issue in organoheterotrophic metabolism, where the nitrate reductase needs to be coupled to NADH released from oxidative metabolism of organic substrates. In the present study, we identify a key role for a putative Rieske-type [2Fe-2S] ferredoxin in P. denitrificans that is widely conserved in other bacterial Nas systems and show that this protein is essential for coupling of NADH oxidation to both nitrate and nitrite reduction. A three-component ferredoxin–nitrate–nitrite reductase system is proposed, where the ferredoxin mediates electron transfer from a single NADH oxidizing site within the nitrite reductase to the sites of nitrate and nitrite reduction present in the nitrite reductase and nitrate reductase components, respectively. Bioinformatic analysis of nas gene clusters suggests that this is a widespread mechanism amongst heterotrophic bacterial species and so provides the biochemical link between the nitrate reductase and the cytoplasmic NADH pool. In addition, we demonstrate a degree of biochemical overlap between the assimilatory Nas system and the respiratory Nar system at the level of nitrate transport and reduction.


Bacterial strains, media and growth conditions

All bacterial strains and plasmids used in the present study are listed in Table 1. P. denitrificans was routinely cultured under aerobic conditions at 30°C in either LB (Luria–Bertani) broth or a defined mineral salts medium with 50 mM succinate as the carbon source [20,21]. Ammonium chloride (10 mM), potassium nitrate (20 mM), potassium nitrite (10 mM) or sodium L-glutamate (5 mM) were used as nitrogen sources. For aerobic batch culture, 50 ml volumes were rotated at 250 rev./min in 250 ml flasks. During anaerobic growth, nitrate (30 mM) or nitrite also acted as the respective terminal electron acceptors for cellular respiration. The Escherichia coli strains were cultured aerobically in LB broth at 37°C. Cell growth was followed by measuring the attenuation (D) of cultures at 600 nm. Antibiotics were used at the following final concentrations: 100 μg/ml ampicillin, 25 μg/ml Km (kanamycin), 100 μg/ml rifampicin, 25 μg/ml spectinomycin, 60 μg/ml streptomycin, 10 μg/ml tetracycline, 20 μg/ml gentamycin or 50 μg/ml chloramphenicol. Bacterial growth curves shown in the Figures are presented as arithmetic plots to minimize distortion of small changes in biomass during the early stages of the growth curve. For determination of growth rates, curves were analysed as semi-log plots from which the apparent maximum growth rate, μmax (app), was determined from the gradient of the exponential growth phase.

View this table:
Table 1 Bacterial strains and plasmids used in the present study

Apr, ampicillin resistance; Cmr, chloramphenicol resistance; Gmr, gentamycin resistance; Kmr, Km resistance; Mob+, mobilization element; Rifr, rifampicin resistance; Specr, spectinomycin resistance; Smr, streptomycin resistance; Tcr, tetracycline resistance; Tra+, transfer gene.

Analytical methods

For preparation of subcellular fractions, P. denitrificans strains were cultured in mineral salts medium with nitrate and glutamate in a BioFlo IV fermenter (New Brunswick Scientific) that was maintained at 30°C, pH 7.2, with dissolved O2 > 95%. Upon reaching a D600 value of ~0.6, a 10 litre culture volume was harvested and cells were fractionated to prepare the cytoplasmic fractions (see the Supplementary online data at for details). Assimilatory nitrate or nitrite reductase activity was assayed spectrophotometrically at pH 7.5 in the presence of the following electron donors: 100 μM NADH, 100 μM NAD(P)H or dithionite-reduced MV (Methyl Viologen) at 1 mM. Activity assays were performed in quartz cuvettes of 1 cm path length. The reactions were followed by measuring the decrease in absorbance observed over time at 340 nm for NADH (or NADPH)-dependent oxidation rates or at 600 nm when monitoring re-oxidation of the reduced Viologen cation radical [22]. Experiments were initiated by addition of NaNO3 and NaNO2, as required. NADH-dependent assays performed on cytoplasmic cell fractions showed a constant background oxidation that was proportional to the amount of extract used. This rate was unaffected when experiments were performed under strict anaerobic conditions and addition of NAD+ up to ~1 mM did not affect the rate of NADH-dependent nitrate or nitrite reduction. MV-dependent rates were obtained under strict anaerobic conditions, using stock solutions of sodium nitrite and dithionite prepared and stored anaerobically at 4°C prior to use [15]. The following molar extinction coefficients, ϵ340=6.22×103 M−1·cm−1 (NADH) and ϵ600=13.7×103 M−1·cm−1 (MV) were used to calculate specific activities. Extracellular nitrate concentration was determined using a Dionex ICS-900 HPLC system fitted with a DS5 conductivity detector and AS40 automated sampler. Samples were diluted into analytical reagent grade water (Fisher Scientific; total nitrogen <0.1 p.p.m.) and passed through a 0.2 μm syringe filter prior to injection on to a 2 mm×250 mm IonPac® AS22 analytical carbonate eluent anion-exchange column. Nitrite concentration present in the extracellular medium was determined colourmetrically as described by Nicholas and Nason [23]. For separation of nitrate and nitrite reductase activities by anion-exchange chromatography, a DEAE-Sepharose™ (GE Healthcare) column matrix was equilibrated in 5 mM L-ascorbate, 5 mM EDTA and 50 mM Tris/HCl, pH 7.5. The column was loaded with cytoplasmic extract [obtained from P. denitrificans WT (wild-type) cells grown with nitrate as the sole nitrogen source], washed with two column vol. and then developed with a linear gradient of 0–0.5 M NaCl, over one column vol., at 1.5 ml·min −1 flow rate.

Samples for 2D-PAGE (two-dimensional PAGE) were prepared from P. denitrificans WT cells that were washed twice with 50 mM Tris/HCl, pH 8.0, and resuspended in 10 mM Tris/HCl (pH 9.0) buffer solution containing DNaseI, RNaseA and protease inhibitors prior to lysis by sonication. Unbroken cells and cellular debris were removed by centrifugation (14000 g) for 20 min at 4°C. Supernatants were then subjected to ultracentrifugation (40000 g) for 1 h at 4°C to obtain soluble protein extracts. Protein concentration was quantified using the 2-D Quant kit (GE Healthcare) according to the manufacturer's instructions and precipitated in 10% (w/v) trichloroacetic acid solution by incubation on ice for 2 h, followed by centrifugation (14000 g) for 20 min. Protein pellets were washed twice in the following solutions: 50 mM Tris/HCl, pH 8.0; 50 mM Tris (pH unadjusted); and 80% (v/v) ice-cold acetone. The final resuspension was in 800 μl of solubilization buffer that contained 7 M urea, 2 M thiourea, 4% (w/v) CHAPS, 1% (w/v) DTT (dithiothreitol), 1.5% (v/v) IPG buffer (pH range 4–7; GE Healthcare) and a trace of Bromophenol Blue. This mixture was vortex-mixed for 2 h and centrifuged (14000 g) for 30 min, after which the supernatant was recovered for use. Immobiline DryStrips (11 cm in length, pH range 4–7 from GE Healthcare) were rehydrated with 350 μg of protein for 12 h in ceramic strip holders (GE Healthcare). IEF (isoelectric focusing) of samples was then performed in an IPGphor ceramic manifold, covered with Plusone DryStrip cover fluid. Sample strips were focused for 20000 V/h in an IPGphor isoelectric focusing system (GE Healthcare). Following IEF, strips were equilibrated as previously described [24] and applied to 12.5% polyacrylamide gels formed with 30% acrylamide/bis solution, 37.5:1 or 29:1 ratio (Bio-Rad). The second dimension separation was performed by using the Hoefer SE600 system (GE Healthcare) and protein spots were visualized using Coomassie Blue staining (2 g·l−1 Coomassie Brilliant Blue G250 and 0.5 g·l−1 Coomassie Brilliant Blue R250 in 5% methanol, 42.5% ethanol and 10% glacial acetic acid). Triplicate 2D-PAGE separations were generated for each sample condition and gels were imaged using the GS-800 calibrated densitometer (Bio-Rad).

Protein identification was performed in the UCO-SCAI proteomics facility (University of Córdoba, Córdoba, Spain), a member of the ProteoRed network. Protein spots of interest were excised automatically in a ProPic station (Genomic Solutions) and subjected to automated digestion with trypsin according to standard protocols in a ProGest station (Genomic Solutions). Peptide fragments were analysed in a 4700 Proteomics Analyzer MALDI–TOF/TOF (matrix-assisted laser-desorption ionization–tandem time-of-flight) mass spectrometer (Applied Biosystems) with an accelerating voltage of 20 kV, in the m/z range 800–4000 (reflectron mode and delayed extraction were enabled, and the elapse time was 120 ns). Proteins were identified by peptide mass fingerprinting and confirmed by MS/MS (tandem MS) analysis of the three most abundant peptide ions. The MASCOT search engine (Matrix Science) was used for protein identification over the non-redundant NCBI (National Center for Biotechnology Information) database of proteins. The confidence in the peptide mass fingerprinting matches (P<0.05) was based on the MOWSE (molecular weight search) method score (higher than 65) and C.I. (confidence interval) >99.8%, and confirmed by the accurate overlapping of the matched peptides with major peaks present in the mass spectrum.

The detailed methodologies for cell fractionation and the construction and complementation of the nas mutants are described in the Supplementary online data. Primers used in the present study are given in Supplementary Table S1 (at


Nitrate and nitrite as nitrogen sources for assimilation by P. denitrificans

In the absence of ammonium, P. denitrificans was able to grow aerobically using nitrate (Figure 2A) or nitrite (Figure 3A) as the sole nitrogen source with values for μmax (app) of 0.27 and 0.23 (±0.01) h−1 respectively at pH 7.2 (see Supplementary Table S2 at Approximately 10 mM of nitrate or nitrite was consumed during batch culture with 50 mM succinate present as carbon substrate (Figures 2B and 3B). A maximum D value of 2.15 (±0.05) was reached after ~16 h, corresponding to a growth yield of ~100 mg of dry weight cells/mM N, which was comparable with yields obtained from cultures assimilating nitrogen from ammonium under carbon-sufficient growth conditions. With nitrate as the sole nitrogen source, a transient accumulation of ~1 mM nitrite in the extracellular medium was observed during growth (Figure 2C). To probe the biochemical basis of this growth physiology, spectrophotometric solution assays were performed on subcellular fractions from P. denitrificans using reduced pyridine nucleotides as electron donor and either nitrate or nitrite as electron acceptor. Cytoplasmic fractions prepared from cells cultured with nitrate present as the sole nitrogen source showed clear nitrate- and nitrite-dependent NADH-oxidation rates (Figure 4). These activities were not detected in either periplasmic or membrane fractions prepared from the same cell cultures, or in any cell fraction when NADPH was used in place of NADH as electron donor. In addition, no activity above a stable background non-specific oxidation rate, which was proportional to the amount of cell extract used, was detected in cytoplasmic fractions from P. denitrificans cells when ammonium was the sole nitrogen source (Figure 4). This pattern of activity is consistent with the regulation of other bacterial nas systems that are subject to ammonium repression and nitrate induction [2,5,19,31].

Figure 2 Aerobic growth of P. denitrificans strains with nitrate

(A) Aerobic growth of P. denitrificans WT (□), nasA (○), nasH (△) and nasA nasH (▽) strains with nitrate present as the sole nitrogen source. Extracellular nitrate (B) and nitrite (C) concentrations are shown for the growth of strains described in (A). The results shown are the average of triplicate determinations.

Figure 3 Aerobic growth of P. denitrificans strains with nitrite

Aerobic growth of P. denitrificans WT (□), nasA (○), nasH (△) and nasA nasH (▽) strains with nitrite present as the sole nitrogen source. Cells were cultured at pH 7.2 (A) and pH 9.2 (C), during which extracellular nitrite concentration was determined (shown in B and D respectively). The results shown are the average of triplicate determinations.

Figure 4 Nitrate and nitrite reductase activity of cytoplasmic fractions of P. denitrificans

Fractions prepared from cells grown with either nitrate (continuous lines) or ammonium (broken lines) as the sole nitrogen source were assayed for nitrite (A) and nitrite (B) reductase activity. Spectrophotometric assays were performed at pH 7.5 in the presence of NADH (100 μM). The reaction was initiated by addition of either nitrate or nitrite and followed by measuring the decrease in absorbance observed over time at 340 nm.

Analysis of the P. denitrificans genome ( reveals the presence of three gene clusters that probably code for different nitrate reductase systems. The genes for the respiratory systems Nar and Nap are located on chromosome 2 (accession number NC_008687) and plasmid 1(accession number NC_008688) respectively. These clusters have been characterized previously in the closely related organism Paracoccus pantotrophus [12,16,32,33]. In P. denitrificans, a third putative nitrate reductase gene is present within a cluster on chromosome 2 that is predicted to encode both the regulatory and structural elements for a cytoplasmic nitrate and nitrite reductase system (Pden_4455–4449). This gene cluster comprises seven open reading frames, nasTSABGHC (from 5′ to 3′) and spans some 10 kilobases from base pairs 1657840 to 1667370. A non-coding region of ~200 bases divides this cluster into two distinct functional units (Figure 1A). The larger coding region, i.e. nasABGHC, which is the focus of the present study, codes for putative redox proteins and substrate transport proteins and lies downstream of two genes encoding a putative nitrate and nitrite-responsive two-component regulatory system, nasT and nasS [5]. The role of nasABGHC in the assimilation of nitrogen from nitrate was confirmed by insertion of a Km resistance marker into the nasA gene (nasAΔ::Km*). For this, the configuration of the resistance cassette was selected such that transcriptional terminators were present to cause a polar effect and prevent expression of all genes collectively transcribed downstream of nasA. Significantly, the nasAΔ::Km* mutant lost the capacity of the WT organism for aerobic growth with both nitrate and nitrite and thus established the importance of the nasABGHC region in the assimilation of both nitrogen sources.

Biochemical properties of cytoplasmic NADH-dependent assimilatory nitrate and nitrite reduction

Spectrophotometric assays were performed to define the kinetic properties of the NADH-dependent nitrate reductase and nitrite reductase activities present in cytoplasmic fractions of P. denitrificans. Activity was measured as the concentration of NADH was varied at fixed saturating concentrations of nitrate or nitrite (Figure 5A, open symbols), and conversely as the concentration of nitrate or nitrite was varied at fixed saturating NADH (Figure 5A, solid symbols). In all cases, enzyme activity varied in accordance with the Michaelis–Menten description and Hanes analysis was performed to define kinetic constants for nitrate (Figure 5B) and nitrite (Figure 5C) reduction. Values for Vmax of 111 or 302 (±12) nmol·min−1·mg of protein−1 were determined for NADH oxidation with nitrate or nitrite present as electron acceptor respectively. The Vmax expressed as a function of NADH consumed for the nitrite reductase reaction was ~3-foldhigher than that for the nitrate reductase reaction. However, when the electron stoichiometries of each reaction are taken into account, i.e. 2e/NO3 and 6e/NO2, the rates of nitrate and nitrite reduction are evenly matched, which would minimise the accumulation of toxic nitrite in the cytoplasm. The values of Km for the reduction of nitrate and nitrite were 17±4 and 5±2 μM respectively. In contrast, Km values determined for NADH oxidation during the nitrate reductase or nitrite reductase reactions were in good agreement, at 51±6 and 58±8 μM respectively.

Figure 5 Kinetic properties of NADH-dependent assimilatory nitrate and nitrite reduction

(A) NADH-dependent nitrate (squares) and nitrite (circles) reductase activities present in cytoplasmic fractions from P. denitrificans grown with nitrate as the sole nitrogen source. Filled and open symbols show: (i) NADH oxidation rates observed with changing substrate concentration against fixed NADH (200 μM); and (ii) changing NADH concentration against fixed substrate concentration (200 μM) respectively. (B and C) Hanes analysis of nitrate reductase (B) and nitrite reductase (C) activities presented in (A). Values for Km of 17±2 and 5±2 μM were derived for nitrate and nitrite reduction respectively. Similar Km values for NADH of 51±8 and 58±8 μM were determined with nitrate or nitrite as electron acceptor respectively. From (i) Vmax values for nitrate and nitrite reduction were 95±10 and 86±10 units respectively, and (ii) Vmax values of 111±12 and 302±12 units were determined for NADH oxidation with nitrate and nitrite respectively (1 unit≡1 nmol·min−1·mg of protein−1).

The elution of nitrate and nitrite reductase activities was monitored when the cytoplasmic fraction was subject to anion-exchange chromatography. Column fractions were assayed for nitrate and nitrite reductase activity using the non-physiological electron donor reduced MV, which has been shown to donate electrons to both nitrate and nitrite reductases (either directly to the active sites or via electron-transferring iron–sulfur centres [14,15,34]). The two activities did not co-elute, with a large peak of MV-dependent nitrate reductase activity eluting at ~0.2 M NaCl, and the major peak of MV-dependent nitrite reductase activity eluting at a higher ~0.3 M NaCl (Figure 6A). Significantly, the nitrite reductase peak retained NADH-dependent nitrite reductase activity, but the nitrate reductase peak did not (Figure 6B). A small protein population, eluting at ~0.28 M NaCl, retained both NADH-dependent and MV-dependent nitrate and nitrite reductase activities.

Figure 6 A representative activity-elution profile observed during anion-exchange chromatography

Panels (A) and (B) show MV- and NADH-dependent nitrate (filled symbols) or nitrite (open symbols) reductase activities present in column fractions respectively. The DEAE–Sepharose™ column matrix was equilibrated in 5 mM L-ascorbate, 5 mM EDTA and 50 mM Tris/HCl, pH 7.5. The column was loaded with a cytoplasmic extract from P. denitrificans grown with nitrate as the sole nitrogen source, washed with two column vol. and then developed with a linear gradient of 0–0.5 M NaCl, over one column vol. at 1.5 ml·min−1 flow rate.

Genetic basis for a composite NADH-linked NasBGC nitrate and nitrite reductase system

Analysis of the P. denitrificans NasC primary amino acid sequence suggests that it binds an N-terminal [4Fe-4S] cluster and a molybdenum-containing cofactor and so shares a similar general organization to the assimilatory nitrate reductase of cyanobacteria (NarB) and the catalytic unit of the structurally-defined respiratory periplasmic nitrate reductases (NapA) [3537]. It is also predicted to contain an additional C-terminal region of ~200 residues that may bind a [2Fe-2S] cluster, as also proposed for the nitrate reductase from K. oxytoca (NasA) [2,31,38,39] (see Supplementary Figure S1 at Significantly, none of these nitrate reductases has a canonical NADH-binding domain. A P. denitrificans strain mutated in nasC lost the capacity for aerobic growth with nitrate as the nitrogen source [μmax (app) <0.01 h−1, maximum D < 0.05], but retained the ability to grow using nitrite and displayed similar growth kinetics to WT [μmax (app)=0.23±0.02 h−1, maximum D of 1.21±0.05 was reached after ~16 h] (Figure 7A and Supplementary Table S2). This finding is consistent with NasC being the sole assimilatory nitrate reductase present during aerobic growth, reducing nitrate to nitrite, but playing no further role in the subsequent reduction of nitrite to ammonium. The nasC mutant could be grown in the presence of glutamate and nitrate [μmax (app)=0.25±0.02, maximum D=1.3±0.1]. Under these conditions, cytoplasmic fractions from WT cells display both MV- and NADH-dependent nitrate reductase activities, but both activities were absent in the nasC mutant (Table 2).

Figure 7 Growth curves for the nasC and nasA strains under aerobic and anaerobic conditions

(A) Aerobic growth of the nasC strain with either nitrate (□) or nitrite (○) present as the sole nitrogen source. (B and C) Anaerobic growth of WT (□), nasA (○) and nasC (Δ) strains with either nitrate (B) or nitrite (C) present as the sole nitrogen source. The results shown are the average of triplicate determinations.

View this table:
Table 2 NADH- and reduced MV-dependent nitrate and nitrite reductase activity of cytoplasmic fractions prepared from P. denitrificans nasB, nasG and nasC mutants grown in the presence of nitrate plus glutamate

Activity was measured in nmol·min−1·mg of protein−1. *Activity decayed with t1/2 of ~50 min. n.d., not detectable.

In order for the assimilation of nitrogen from nitrate to proceed, the cytoplasmic nitrite generated by NasC must be further reduced to ammonium. NasB shares significant sequence homology to flavin, Fe-S and sirohaem-containing nitrite reductases [2,11,40] (see Supplementary Figure S2 at The P. denitrificans NasB polypeptide is predicted to contain N-terminal FAD- and NADH-binding domains, whereas highly conserved central and C-terminal sequence regions contain the cysteine residues required for iron–sulfur cluster co-ordination and the nitrite/sulfite reductase ferredoxin half-domain associated with sirohaem binding (see Supplementary Figure S2). A P. denitrificans nasB mutant was unable to grow aerobically with either nitrite or nitrate as the sole nitrogen source. In order to confirm that these growth defects were not caused by a downstream effect of the gene disruption process, the nasB strain was complemented with a pEG276-nasB expression construct. The presence of this plasmid allowed the nasB mutant to grow to near WT levels with either nitrate or nitrite as the sole nitrogen source (Supplementary Table S2) [μmax (app)=0.20±0.02 h−1, maximum D=1.23±0.05]. However, in the absence of the pEG276-nasB expression construct, no growth of the nasB mutant was observed, despite prolonged incubation for several days. This observation excludes the possibility of growth recovery through spontaneous mutation that may up-regulate a cryptic nitrite reductase. Cytoplasmic fractions obtained from the nasB mutant, grown in the presence of glutamate and nitrate, were assayed for NADH-dependent nitrate reductase and nitrite reductase activity. In contrast with WT cytoplasmic fractions, nitrite reductase activity was not detected in this mutant, consistent with the loss of the assimilatory nitrate reductase, NasB (Table 2). Significantly, nitrate reductase activity was also absent with NADH present as electron donor. Analysis of the NasC primary structure suggests that it lacks the NADH- and FAD-binding domains present in NasB that would be required for self-contained coupling of NADH oxidation to nitrate reduction (see Supplementary Figures S1 and S2). Therefore the absence of NADH-dependent nitrate reduction in the nasB mutant suggests a model in which the NADH-dehydrogenase domain of NasB provides electrons for both nitrite reduction by the sirohaem domain of NasB and nitrate reduction by NasC (Figure 1B). Such a model is consistent with the similar Km values for NADH observed with nitrate or nitrite in cytoplasmic extracts, and also the loss of NADH-nitrate reductaseactivity when nitrate reductase is separated from nitrite reductase by anion-exchange chromatography. Further evidence in support of this model was forthcoming from measuring nitrate reductase activity in the NasB mutant using the artificial electron donor MV, which can donate electrons directly to the nitrate reductase. Significantly, MV-dependent nitrate reductase activity was detected in the nasB mutant, which confirmed the presence of a functional NasC that is unable to couple to NADH oxidation in the absence of NasB.

If the NADH-binding domain of NasB also serves NasC, then this raises the question of how electron transfer between NasB and NasC might occur. NasG is a strong candidate for mediating such electron transfer since it is predicted to be a Rieske-type iron–sulfur protein in which all residues (2 cysteine and 2 histidine) essential for co-ordination of a [2Fe-2S] redox site are conserved (see Supplementary Figure S3 at A nasG mutant was unable to grow with either nitrate or nitrite as the sole nitrogen source, under aerobic conditions. Expression of nasG in trans from a pEG276-nasG expression construct complemented the growth deficiencies of the nasG strain observed with nitrate and nitrite [μmax (app)=0.22±0.07 h−1, maximum D=1.31±0.05), confirming specific disruption of nasG with no downstream effects (Supplementary Table S2). Unlike NasC and NasB, there is no direct enzymatic assay with which to establish the synthesis of NasG. However, 2D-PAGE of soluble extracts from P. denitrificans WT, grown aerobically in the presence of glutamate and nitrate, revealed a protein with the mass and charge characteristics (12.1 kDa and pI of 5.6) of NasG that was absent in the nasG mutant (Figure 8). Analysis by MS confirmed that this spot was the NasG polypeptide (MASCOT Protein Score=314; see Supplementary Table S3 at Cytoplasmic fractions prepared from cells of the nasG mutant grown with glutamate and nitrate were devoid of NADH-dependent nitrate reductase or nitrite reductase activities, but MV-dependent activities were detected (Table 2). These findings confirm that NasG is essential for both NADH-dependent nitrate and nitrite reduction, consistent with the protein mediating electron transfer from NADH to the active sites of both NasB and NasC (Figure 1B). It was also notable that the MV-dependent nitrite reductase activity was unstable, with a half-life of ~50 min following preparation of the cytoplasmic extract. This instability was reflected in the 2D-PAGE analysis of the WT and nasG strains. In the WT extract, a triad of protein spots of ~88 kDa and pI of ~5.6 were each established by MS to be NasB (MASCOT Protein Score=622; see Supplementary Table S3) (Figure 8A). As expected, this triad was absent in the nasB mutant. However, they could also not be identified in extracts from the NasG mutant, despite analysis of a number of samples at different protein loadings. In contrast, the NasG spot could be detected at approximately WT levels in 2D-PAGE analysis of the nasB strain (Figure 8). Thus NasG is stable in the absence of NasB, but may be required for NasB stability. It was notable that, in the nasB strain, a new protein spot was very prominent in cells grown with nitrate and glutamate. This spot was identified by MS to be a member of the Hsp20 family of chaperones that protect unfolded proteins from aggregation (MASCOT Protein Score=632; see Supplementary Table S3) [41]. This response may reflect the need to protect NasG as it is being assembled in the absence of its NasB partner or a response to protein damage by nitrosative stress arising from a lesion in cytoplasmic nitrite reduction.

Figure 8 2D-PAGE analysis of soluble extracts from P. denitrificans

Protein (350 μg) was applied to 11 cm strips. Isoelectric focusing was performed in the range 4–7. The second dimension was performed with 12.5% were polyacrylamide gels formed from 30% acrylamide/bis solution, 37.5:1 (A) or 29:1 ratio (B).

The contribution of NasA and NasH to nitrate and nitrite transport

Delivery of nitrate and nitrite from the external environment to the cytoplasmic NasBGC system requires that two nitrogen oxyanions are transported across the cytoplasmic membrane against a membrane potential that is negative on the inside of the membrane. NasA and NasH are good candidate transporters for facilitating assimilatory nitrogen oxyanion import. NasA is predicted to be a 12-transmembrane helix transporter of the MFS (major facilitator superfamily) (Figure 1B and Supplementary Figure S4 at P. denitrificans synthesizes another well-characterized MFS family member in the presence of nitrate under anoxic conditions, NarK, which serves to transport nitrate for the respiratory nitrate reductase system, NarGHI (Figure 1B) [16,17,33]. A non-polar nasA strain was constructed and was found to be strongly attenuated for aerobic growth with nitrate as the sole nitrogen source [μmax (app)=0.04±0.01 h−1, maximum D=0.38±0.05), demonstrating that NasA is the major nitrate transporter under these growth conditions (Figure 2A and Supplementary Table S2). Despite disruption of the nasA gene, this strain retained the ability to grow aerobically with nitrite as the sole nitrogen source, albeit with slower growth [μmax (app)=0.20±0.01 h−1, maximum D=1.99±0.05] and nitrite consumption kinetics than that observed for WT cells, implying that NasA may make a contribution to, but is not essential for, nitrite uptake into the cell (Figure 3 and Supplementary Table S2). Efforts to complement the nasA strain using the same complementation vector system used successfully for nasB and nasG was unsuccessful, possibly due to a failure to assemble the integral membrane protein. However, evidence that there were no polar effects of the nasA mutation on downstream gene expression is given by the detection of NADH-dependent nitrate (~10 nmol·min−1·mg of protein−1) and nitrite (~30 nmol·min−1·mg of protein−1) reductase activity at WT levels in cells grown on glutamate-nitrate medium, which would require transcription of nasBGC, which lies downstream of nasA.

NasH is a putative member of the formate–nitrite transporter superfamily that includes NirC from E. coli (Supplementary Figure S5 at [4244]. Evidence suggests that NirC can move nitrite bi-directionally across the cytoplasmic membrane during anaerobic growth of E. coli [43]. NasH is thus a prime candidate for mediating nitrite uptake or export and so may be important for nitrite homoeostasis during nitrate assimilation. Perhaps surprisingly though, the P. denitrificans nasH mutant displayed similar growth kinetics and yields [μmax (app)=0.28±0.01 h−1, maximum D=1.79±0.09] to that observed for WT, when cultured aerobically with either nitrate or nitrite as the sole nitrogen source at pH 7.2 (Figures 2 and 3). The pattern of nitrite consumption from the extracellular medium during growth with nitrite as the sole nitrogen source was also similar in the WT and nasH strains (Figure 3B), as was the transient accumulation of ~1 mM nitrite observed during the mid- to late-exponential growth phase when nitrate was present as the sole nitrogen source (Figure 2C). A double nasA/nasH mutant also retained the ability to grow with nitrite as the sole nitrogen source, at pH 7.2 [μmax (app)=0.20±0.01 h−1, maximum D=1.72±0.05], displaying similar growth kinetics to the nasA strain (Figure 3A and Supplementary Table S2). Nitrite is a protonatable anion that exists in equilibrium with nitrous acid (NO2+H+↔HNO2), with a pKa value of 3.3 [17]. Thus, at pH 7 with an external nitrite concentration of 10 mM, the concentration of HNO2 is present in the low μM range. This acid could freely diffuse across the phospholipid bilayer, dissociate in the cytoplasm, and so deliver nitrite to the NasBGC complex without recourse to a specific nitrite uptake system. The extent to which this could occur will be decreased at higher pH values. It was therefore notable that when the nasH and nasA/nasH mutants were grown at pH 9.2 with nitrite as the sole nitrogen source, significant attenuation in growth and nitrite consumption was observed in both cases, thus demonstrating a role for NasH in nitrite import (Figures 3C and 3D and Supplementary Table S2).

Functional substitution of Nar components for Nas components in anaerobic nitrate assimilation

P. denitrificans could grow anaerobically, as well as aerobically, with nitrate or nitrite as the sole nitrogen source (Figure 7 and Supplementary Table S2). Under these growth conditions energy conservation is via nitrate and nitrite respiration through the anaerobically synthesized Nar and Nir systems respectively [12]. The Nar system includes the ubiquinol/nitrate oxidoreductase NarGHI, of which the NarG nitrate-reductase subunit is located at the cytoplasmic face of the respiratory membrane and a fusion protein of two NarK-type modules facilitates nitrate and nitrite movement across the membrane (Figure 1B). To investigate whether the NarG and NarK proteins could functionally substitute for NasC and NasA, the nas mutants were cultured under anaerobic conditions with nitrate present as the sole nitrogen source and electron acceptor, conditions that result in nar gene expression. Both the nasA and nasC mutants showed significant growth under these conditions (Figure 7B), despite being unable to grow aerobically with nitrate as the sole nitrogen source (Figures 2A and 7A). The nasC mutant was also able to grow anaerobically with nitrite present as both the sole nitrogen source and respiratory electron acceptor (Figure 7C). Unlike the nasC strain, however, neither the nasB or nasG mutants retained the ability to grow anaerobically with nitrate or nitrite, demonstrating that the NADH-dependent NasBG nitrite reductase was indispensible to both oxic and anoxic assimilation of nitrate and nitrite.


The present study has established that P. denitrificans NasC and NasB are both part of a cytoplasmic NADH-dependent assimilatory nitrate and nitrite reduction system. However, analysis of the primary structures of both proteins revealed that only NasB has a canonical FAD-dependent NADH-binding domain. A bioinformatic analysis of predicted gene products for nasC and nasB homologues from the diverse bacterial phyla suggests that this is a common feature (see Supplementary Figure S6 at There is no bacterial assimilatory nitrate reductase that we can identify that has a NADH-binding site. As such, this makes them quite distinct from plant and fungal assimilatory nitrate reductases in which NAD(P)H binding domains are ubiquitous [45]. In the present study, we have reported genetic and biochemical results that suggest the NADH-oxidizing FAD domain of NasB provides electrons for both nitrite reduction by the NasB ferredoxin:sirohaem active site and nitrate reduction by the NasC molybdenum cofactor active site. This is shown by: (i) loss of NADH-dependent nitrate reductase activity when the nitrate reductase is separated from the NADH-dependent nitrite reductase; (ii) the Km value for NADH determined with either nitrate or nitrite being the same in cytoplasmic fractions from WT cells; (iii) non-polar deletion of nasB resulting in loss of aerobic growth with nitrate as the sole nitrogen source; and (iv) non-polar deletion of nasB resulting in the loss of NADH-dependent nitrate reductase activity, but not MV-dependent nitrate reductase activity. We also show that the putative Rieske [2Fe-2S] ferredoxin NasG is required for these NADH-dependent electron transfer processes. In a nasG strain: (i) no growth is observed with either nitrate or nitrite as the sole nitrogen source; and (ii) NADH-dependent nitrate and nitrate reductase activities are absent in cytoplasmic extracts, but MV-dependent activities are detected. These findings imply that NasG can interact in the cytoplasm with both NasB and NasC, and lead us to propose a model in which it serves at the interface of a composite NADH-dependent nitrate and nitrite reductase system, NasBGC (Figure 1B).

It is notable that many bacterial nas clusters that we have examined encode a NasG-like protein, in addition to an NADH-oxidizing sirohaem-dependent nitrite reductase, suggesting an important conserved function (Supplementary Figure S6). It is absent in the Synechococcus elongatus cluster (Supplementary Figure S6), but in this case the photoautotroph uses reduced ferredoxin generated from photosystem I to drive nitrate and nitrite reduction [19]. A nasG homologue is also absent from the nas cluster of Klebsiella species (Supplementary Figure S6) [39]. However, the NasB proteins from Klebsiella are larger and show an extended C-terminal regions of ~100 residues that share homology with NasG. In addition, in K. oxytoca, there is a gene in the nas operon that codes for a flavoprotein which may substitute for NasG or support it in enabling electron transfer from NADH to the nitrate reductase subunit. A homologue of this gene may also be playing such a role in B. subtilis [4], whereas in Klebsiella pneumoniae this gene appears to be fused to the gene for the nitrite reductase (see Supplementary Figure S6). Nevertheless, even taking these exceptions into account, the widespread distribution of NasG-type modules leads us to propose that the composite NasBGC system exemplified here by P. denitrificans is a very common feature of Nas systems in phylogenetically diverse bacteria. Although the interaction between NasC and NasBG does not resist anion-exchange chromatography, there is precedent for a stable interaction between a molybdenum cofactor-dependent enzyme and a Rieske-type ferredoxin. Intriguingly, the catalytic subunit of the arsenite oxidase from Alcaligenes faecalis forms a stable heterodimeric complex and co-crystallizes with its redox partner, a Reiske-type [2Fe-2S] protein that shares sequence similarity with NasG (~40%) [46,47]. Structural studies also revealed that the iron–sulfur centres present in each subunit are ideally situated, at <14 Å (1 Å=0.1 nm) (edge-to-edge) from one another across the heterodimeric interface, to facilitate rapid electron transfer that does not limit catalysis [47]. Thusthe NasC–NasG interaction may be a weaker manifestation of theprotein–protein interaction observed in the arsenite oxidase complex. The genetic evidence that NasG is a dedicated ferredoxin for the NasBC system and that the NasB protein is possibly unstable is perhaps surprising given that the P. denitrificans genome is predicted to encode a number of small cytoplasmic ferredoxins. However, precedent for this may be found in the E. coli cytoplasmic nitrite reductase system. E. coli is unable to grow aerobically with nitrite as the sole nitrogen source, but when grown under nitrate-rich anoxic conditions a respiratory nitrate reductase NarG and a sirohaem:ferredoxin-type nitrite reductase NirB operate in the cytoplasm to respire nitrate and detoxify the nitrite product [50]. E. coli can also grow anaerobically with nitrite as the sole nitrogen source under conditions where the nirB gene encoding this anaerobically inducible nitrite reductase is expressed [50,51]. E. coli nirB is a homologue of P. denitrificans nasB (65% similarity) and both genes are found upstream of their ferredoxin partners, E. coli nirD and P. denitrificans nasG respectively, which share 59% similarity (Supplementary Figure S6). Significantly, similar to the P. denitrificans nasG mutant, NADH-dependent nitrite reductase activity is also lost in an E. coli nirD mutant [52]. In both E. coli and P. denitrificans, the nirB/nasB and nirD/nasG loci lie immediately upstream of a gene encoding a transporter nirC/nasH (55% similarity) in their respective genomes (Supplementary Figure S6). In the present study, by growing P. denitrificans at high pH to minimize HNO2 formation, we have shown that nasH contributes to nitrite uptake. These observations show a clear genetic and biochemical link between two nitrite reductase systems with distinct primary functions in nitrite assimilation and detoxification.

The present study also looked at the rescue of nasA and nasC mutants under anaerobic growth conditions by the respiratory Nar system. This is consistent with functional overlap between two physiologically distinct systems whose cellular function commonly requires the import and cytoplasmic-based reduction of nitrate. In this respect, it is notable that an assimilatory nitrate reductase gene is absent from some assimilatory gene clusters in nitrate-assimilating bacteria where the respiratory nitrate reductase narG gene is present at a different genetic locus in the bacterium. A noteworthy example is Mycobacterium (Supplementary Figure S6), in which narG mutants cannot grow with nitrate as the sole nitrogen source [48,49]. The interchangeability between NasC and NarGHI in anaerobic assimilatiory nitrate reduction is also consistent with the dissociation of NasC from NasBG observed during anion-exchange chromatography, suggesting a modular arrangement in which NasBG can exist as a stable functional entity in the absence of NasC.


Andrew Gates, Victor Luque-Almagro, Alan Goddard, Stuart Ferguson, M. Delores Roldán and David Richardson designed the research. Andrew Gates, Victor Luque-Almagro and M. Dolores Roldán performed the research. Andrew Gates, Victor Luque-Almagro, Alan Goddard, Stuart Ferguson, M. Dolores Roldán and David Richardson analysed the data. Andrew Gates and David Richardson wrote the manuscript.


This work was supported by the Biotechnology and Biological Sciences Research Council [grant numbers BBE0219991 and BBD5230191]. D.J.R. is a Royal Society and Wolfson Foundation for Merit Award Fellow. M.D.R. thanks the Ministerio de Ciencia y Tecnología for supporting this work [grant numbers BIO2005-07741-C02-01 and BIO2008-04542-C02-01] and the Junta de Andalucía [grant number CVI1728]. V.M.L.-A. was the recipient of a postdoctoral fellowship from the Ministerio de Ciencia y Tecnología, Spain. We are also grateful to the U.S. Department of Energy for providing the funds to sequence the genome of Paracoccus denitrificans PD1222.


We thank Dr Niels-Ulrik Frigaard (Department of Biology, University of Copenhagen, Denmark) for providing the pSRA2 plasmid containing the streptomycin and spectinomycin resistance cassette and Dr Eva Pérez Reinado (Department of Biochemistry and Molecular Biology, University of Cordoba, Spain) for the mobilizable vector pSUP202*.

Abbreviations: 2D-PAGE, two-dimensional PAGE; IEF, isoelectric focusing; Km, kanamycin; LB broth, Luria–Bertani broth; MV, Methyl Viologen; MFS, major facilitator superfamily; μmax (app), apparent maximum growth rate; WT, wild-type


View Abstract