The NADH oxidase from Pyrococcus furiosus

Implications for the protection of anaerobic hyperthermophiles against oxidative stress


E. J. Crane, Department of Chemistry, Henson School of Science and Technology, Salisbury University, Salisbury, MD, 21801, USA. Fax: + 1 410 677 5043, Tel.: + 1 410 548 2945, E-mail:


A wealth of H2O-producing NADH oxidase (NOX) homologues have been discovered in the genomes of the hyperthermophilic Archaea, including two homologues in the genome of Pyrococcus furiosus which have been designated as NOX1 and NOX2. In order to investigate the function of NOX1, the structural gene encoding NOX1 was cloned from the genome of P. furiosus and expressed in Escherichia coli, and the resulting recombinant enzyme (rNOX1) was purified to homogeneity. The enzyme is a thermostable flavoprotein that can be reconstituted only with FAD. rNOX1 catalyzes the oxidation of NADH, producing both H2O2 and H2O as reduction products of O2 (O2 + 1–2NADH + 1–2H+ → 1–2NAD+ + H2O2 or 2H2O). To our knowledge, this is the first NADH oxidase found to produce both H2O2 and H2O. The enzyme exhibits a low Km for NADH (< 4 µm), and shows little or no reaction with NADPH. Transcriptional analyses demonstrated that NOX1 is constitutively expressed regardless of the carbon source and a single promoter was identified 25 bp upstream of the nox1 gene by primer extension. Although P. furiosus is a strict anaerobe, it may tolerate oxygen to some extent and we anticipate NOX1 to be involved in the response to oxygen at high temperatures.


glutathione reductase


NADH oxidase


NADH peroxidase


coenzyme A disulfide reductase






2,2′azino- bis(3-ethylbenzthazoline-6-sulfonic acid)

The H2O-producing NADH oxidases are members of the flavoprotein disulfide reductase superfamily of enzymes in the subclass usually represented by glutathione reductase (GRX). Within this GRX family, the NADH oxidases are members of the recently discovered subgroup of disulfide reductases that contain a single redox active cysteine at their active site [1]. The three current members of this subgroup are NADH oxidase (NOX; O2 + 2NADH + 2H+ → 2H2O + 2NAD+), NADH peroxidase (NPX; H2O2 + NADH + H+ → 2H2O + NAD+), and coenzyme A disulfide reductase (CoADR; CoA-S-S-CoA + NADPH + H+ → 2CoASH + NADP+). All of these enzymes contain an absolutely conserved single redox active cysteine, corresponding to Cys42 of the Enterococcus faecalis NPX (Fig. 1), rather than the cysteine disulfide usually found at the active site of the disulfide reductases. All of the confirmed members of this subgroup have been found in heme-deficient facultatively anaerobic bacteria with the exception of the NOX from the parasitic eucaryote Giardia duodenalis[2]. NPX and NOX are believed to play a role in oxidative stress defense and in the regeneration of oxidized pyridine nucleotides for these organisms.

Figure 1.

Multiple sequence alignment of P. furiosus NOX1 to known NADH oxidases/peroxidases. The alignment was performed with clustal w. The GenBank accession numbers for the other NADH oxidases are as follows: E. faecalis NOX (P37061), E. facaelis NPX (P37062), Streptococcus mutans NOX (JC4541), Brachyspira (Serpulina) hyodysenteriae NOX (AF060815), and Staphylococcus aureus CoADR (AF041467).

During the genome sequencing of thermophilic archaea and bacteria, homologues of NADH oxidase were found in a wide range of species including Pyrococcus horikoshii[3], Archaeaglobus fulgidus[4], Methanococcus jannaschii[5] and Thermotoga maritima[6]. This result was somewhat surprising in light of the fact that most of these organisms are strict anaerobes, a class of organisms that have not been found to possess NOXs. It should be noted that the H2O-producing NOXs are distinct from the H2O2-producing NAD(P)H oxidases which have been characterized from several thermophilic sources [7–9]. Many of these enzymes appear to be members of the thioredoxin reductase subclass of the disulfide reductases [10].

P. furiosus is a strict anaerobe that grows heterotrophically by the fermentation of carbohydrates and peptides. While they are strict anaerobes, it is likely that Pyrococcus might encounter high levels of oxygen from the cool, high-pressure waters surrounding the vent. Recent studies have suggested that these organisms may have evolved at least a crude antioxidant defense system [11], which may involve the NOX1 enzyme discussed in this report. We have cloned the nox1 gene from P. furiosus, and overexpressed the enzyme. Reported below is the characterization of the rNOX1 from P. furiosus, the first enzyme of this class to be characterized in the hyperthermophiles.

Materials and methods

Growth of microorganisms

P. furiosus (DSM 3638) was grown in a sea salts medium as described previously [12]. Maltose (10 mm), pyruvate (40 mm), or tryptone (5 g·L−1) was included as primary carbon and energy source. Escherichia coli JM109(λDE3) and XL-1 were grown at 37 °C in TYP medium (16 g·L−1 yeast extract, 16 g·L−1 tryptone, 5 g·L−1 NaCl, and 2.5 g·L−1 K2HPO4). When required, the following antibiotics were included in the medium; kanamycin (50 µg·mL−1), ampicillin (50 µg·mL−1), and tetracycline (15 µg·mL−1).

Cloning and expression of the gene encoding the NADH oxidase

The structural gene encoding NOX1 (PF_1430634) was amplified from P. furiosus genomic DNA with the oligonucleotides BG436 (5′-CGCGCCATGGCAAGAATAGTTGTTATAGGTTCGGG-3′), and BG381 (5′-CGCGGGATCCTTATCATAACTTTCTCATAGC-3′). The N-terminus of NOX1 was identified based on the presence and proper spacing of the ribosome binding site, annotation from the genome sequence, and its good agreement with those of NOX homologs from mesophilic sources that had been characterized in detail. PCR amplification was carried out using Pfu polymerase (Stratagene), and the resulting 1.3-kb PCR product was cloned in the NcoI and BamHI sites of pET-24d. The resulting plasmid, pLUW772 was transformed into E. coli BL21(λDE3). For expression of the recombinant enzyme the strain JM109(λDE3) was used.

Overexpression, purification of rNOX1 and quaternary structure determination

rNOX1 was overexpressed in E. coli JM109(λDE3). Five liters of TYP media (containing kanamycin 50 µg·mL−1) in a fermentor was started with a 2% inoculation of overnight culture and grown at 37 °C. When the culture reached an D600 of 1.0 (≈ 3 h) isopropyl β-d-thiogalactopyranoside was added to a final concentration of 1 mm. The culture was incubated for an additional 5 h, and cells were collected by centrifuging at 15 000 g for 15 min. The cells were washed with 50 mm Tris buffer pH 7.50, and collected by centrifugation (15 000 g, 15 min). The resulting cells were frozen at −70 °C until purification. After thawing, the cells were resuspended in 50 mm Tris pH 7.50 (90 mL total volume) and sonicated by 3 × 1-min sonication treatments of 7-mL aliquots of the cell solution. Cell debris were removed by centrifugation at 15 000 g for 15 min. FAD was added to a final concentration of 1 mm, and the supernatant was then incubated at 90 °C for 20 min and placed on ice. The heat-denatured protein was removed by centrifugation at 15 000 g for 30 min and (NH4)2SO4 was added to the supernatant to bring it to a concentration of 1.25 m (NH4)2SO4. The following chromatography steps were performed using an AKTA low pressure chromatography system (FPLC) from Pharmacia Biotech. The supernatant was applied to a 10-mL phenyl-Sepharose HiLoad column (Pharmacia Biotech) equilibrated with 1.25 m (NH4)2SO4, 50 mm Tris pH 7.50, 1.25 m (NH4)2SO4. The column was washed with the equilibration buffer until excess FAD and protein had eluted. After washing with 25 mL 1.0 m (NH4)2SO4, 50 mm Tris pH 7.50, rNOX1 was eluted by a 120-mL linear gradient from 1.0 to 0 m (NH4)2SO4, with NOX1 eluting at ≈ 0.5 m (NH4)2SO4. Fractions containing rNOX1 were pooled and concentrated and exchanged into 50 mm Tris pH 7.50 by ultrafiltration. The concentrate (total volume 20 mL) was loaded and run on a 5-mL Hi-Trap Blue column (Pharmacia Biotech) that had been equilibrated with 50 mm Tris pH 7.50. The column was washed with 50 mm Tris pH 7.50, and NOX1 was eluted with a 36-mL linear gradient from 0 to 1 m KCl, 50 mm Tris pH 7.50, with rNOX1 eluting at ≈ 0.6 m KCl. Fractions containing rNOX1 were pooled and concentrated. The concentrated rNOX1 (total volume 2 mL) was loaded on a 120-mL Sephacryl S-200 HR (Pharmacia Biotech) size-exclusion column that was run at 0.5 mL·min−1. Fractions containing rNOX1 were pooled. The procedure yields 20.0 mg of enzyme (16% yield) with most of the loss occurring on the phenyl-Sepharose column. The purified protein gives a single peak during size-exclusion HPLC and single band on SDS/PAGE. Fig. 2 shows the UV/visible spectrum of rNOX1, which shows the characteristic spectrum of a flavoprotein. The A450/A280 ratio for purified and reconstituted rNOX1 is 5.83. The native molecular weight of rNOX1 of 97.1 kDa (Table 1) was determined by HPLC size-exclusion chromatography on a Suprasil column. Molecular masses were determined by comparison to standards of 670 kDa (thyroglobulin), 158 kDa (gamma globulin), 44 kDa (ovalbumin) and 17 kDa (myoglobin).

Figure 2.

Spectrum of 16.0 µm rNOX1 in 50 mm Tris pH 7.50.

Table 1. Biophysical and kinetic properties of the P. furiosus rNOX1.
Predicted subunit molecular mass48.1 kDa
Determined subunit molecular mass50.0 kDa
Native molecular mass97.3 kDa
k cat (absence of substrate-level FAD)4.8 s−1
k cat (saturating substrate-level FAD)11.1 s−1
Km app FAD44 µm
Km app NADH< 4 µm
Km app O2> 110 µm
H2O2 produced/NADH consumed0.61

Enzyme assays

Thermophilic NADH oxidase assays were conducted in a Beckman-Coulter Diode-Array spectrophotometer fitted with a cuvette holder; the temperature was controlled by a circulating water bath. Temperatures given for assays are the actual cuvette temperatures, and all pH values were determined at room temperature. In the general NADH oxidase assay, 800 µL 50 mm sodium phosphate buffer pH 7.50, which had been preheated in a heat block, was added to a preheated 1.5 mL quartz cuvette. Substrate(s) and enzyme were then added (total assay volume 830 µL) and the oxidation of NADH was monitored at 340 nm. In the standard assay and enzyme-specific activity determinations, the temperature was 75 °C and the concentrations of NADH and FAD were 100 µm. The nonenzymatic rate of NADH/FAD thermal decomposition was subtracted from each assay. For experiments involving the determination of O2/NADH consumption ratios a 1.5-mL cuvette was filled completely with buffer and sealed with a coverslip that was lightly coated with vacuum grease, preventing gas exchange with the atmosphere. The assays were run until O2 was depleted as determined by the endpoint of the oxidase reaction. For experiments at saturating O2 concentrations, the buffer was bubbled with O2 gas for 15 min prior to the assay in a heat block set to the assay temperature.

H2O2 determinations

Hydrogen peroxide determinations were performed basically as outlined by Makinen and Tenovuo [13] with the following modifications: a stock solution containing 22 mg·mL−1 ABTS dye and 100 U·mL−1 horseradish peroxidase was prepared, and a standard curve from 0 to 88 µm H2O2 in 1.00-mL aliquots in the standard assay buffer was prepared. Standard NADH oxidase reactions in 800 µL volumes were run to completion (NADH limiting), and 200 µL of buffer was added. The reactions were cooled to room temperature, and assayed for H2O2 production in parallel with the standard curve by the addition of 150 µL of the stock assay mixture, incubation for 30 min, and determination of absorbance at 725 nm. Control experiments showed that the loss of hydrogen peroxide during the high temperature assay and subsequent cooling were < 5%.

RNA isolation, Northern analysis, and primer extension

Isolation of total RNA from P. furiosus and Northern analyses were carried out as described previously [12]. The probe for the Northern analysis was the PCR product obtained using BG436 and BG381. The 5′ end of the nox transcript was mapped using the oligonucleotide 5′-GGGAGAATATTGCATTGTTTCTTCT-3′, which binds 97-bp downstream of the ATG start codon, and was labeled with a fluorescent label, IRD800, by the manufacturer (MWG). Total RNA was isolated from P. furiosus grown with 10 mm maltose, 40 mm pyruvate, or 5 g·L−1 tryptone as the primary carbon and energy source. The primer extension reaction was carried out using the Reverse Transcription System (Promega) according to the manufacturer's instructions with the following modifications: 5.0 pmol oligonucleotide was hybridized with 15 µg total RNA in the buffer supplied in a 10-µL volume by heating at 70 °C for 10 min before allowing to slowly cool to room temperature. The reaction was started by the addition of dNTPs (1 mm final), MgCl2 (5 mm final), RNAsin (20 U final) and AMV-RT (22.5 U final) in a final volume of 20 µL. The reaction was incubated at 45 °C for 30 min at which point the reaction volume was adjusted to 50 µL with 10 mm Tris pH 8.5, and 1 µL of RNaseA (5 mg·mL−1) was added, and incubated at 37 °C for 10 min. The sample was then precipitated with ethanol and the pellet was dissolved in 3 µL loading buffer; 1 µL was subsequently electrophoresed on a sequencing gel in parallel with a sequencing reaction obtained with the same oligonucleotide.


General characterization of the P. furiosus NOX1

The structural gene encoding NOX1 (PF_1430634) consists of 1314 bp and encodes a protein of 438 residues with a predicted molecular mass of 48 kDa. blast and tfasta analysis of NOX1 revealed a significant level of identity to putative NADH oxidases from hyperthermophiles and bacterial NADH oxidases from mesophilic sources. Of particular interest was the identity to the well characterized NOXs from mesophilic organisms with the highest levels of identity found with the NADH oxidases from E. faecalis (28%), Streptococcus mutans (26%), and Brachyspira (Serpulina) hyodysenteriae (24%). As shown in Fig. 1, the P. furiosus NOX1 contains a cysteine which corresponds to the single redox active cysteine of the E. faecalis NOX and NPX, as well as considerable identity in areas that have been shown to be important for NADH and FAD binding in these enzymes. P. furiosus, P. horikoshii and P. abyssi each contain two NOX homologues in their genome. These NOXs (NOX1 and NOX2) are ≈ 30% identical in each of the species, and the individual NOXs show 70–80% identity between species, suggesting that NOX1 and NOX2 are different, although related, enzymes. This is supported by our characterization of the recombinant NOX2 from P. horikoshii, which is considerably different from the P. furiosus NOX1 (D. R. Harris & E. J. Crane, unpublished data). The recombinant NADH oxidase (rNOX1) was purified 12-fold with a yield of 16% and a specific activity of 12 U·mg−1. rNOX1 had a subunit molecular weight of 50 kDa as determined by SDS/PAGE. HPLC size-exclusion chromatography on a Suprasil column gave a single peak and a calculated molecular mass of 97.1 kDa, consistent with the enzyme existing as a dimer. The majority of the members of the disulfide reductase family are dimeric, with the exception of NPX from E. faecalis, which is tetrameric. During initial procedures in which the enzyme was not reconstituted until after purification, it was determined that the recombinant enzyme is ≈ 20% in the holo form prior to reconstitution (by a comparison of the FAD concentration, approximated assuming an extinction coefficient of 11 300 absorbance units·m−1·cm−1 at 450 nm, to the protein concentration as determined by the Bradford method). The enzyme was fully reconstituted with FAD but FMN was unable to reconstitute the rNOX1. The spectrum of reconstituted rNOX1 is that of a typical FAD-containing disulfide reductase, with λmax values in the 300–800 nm range at 373 and 446 nm (Fig. 2). The spectrum also shows the broad long wavelength charge-transfer at > 500 nm observed with NOX and NPX from E. faecalis[14,15]. There is a small (3.8% of absorbance at 373 nm) but significant reversible change in the 300–800 nm enzyme spectrum between 25 and 75 °C, as would be expected for a thermophilic enzyme (data not shown).

Kinetic characterization of the rNOX1 NADH oxidase activity

rNOX1 shows significant NADH oxidase activity in both the presence and the absence of additional substrate-level FAD at 75 °C at pH 7.50. The enzyme has a broad pH range, with little difference in activity over the range pH 5.50 (50 mm Mes buffer) to 8.50 (50 mm Mops buffer) (73 and 81% of the activity observed at pH 7.50, respectively). rNOX1 shows a low Km for NADH of < 4 µm, whereas the substrate-level FAD-dependent portion of the activity shows a Km for FAD of 44 µm. kcat for the oxidase reaction in the absence of substrate-level FAD is 4.8·s−1, while kcat for the reaction in the presence of substrate-level FAD is 11.1·s−1. From the sealed cuvette NADH/O2 ratio determinations discussed below, the Km for O2 must be > 110 µm, and is higher than the concentration of O2 present in the standard assay. The enzyme shows no observable activity with NADPH, and FMN is not able to substitute for FAD in the substrate-level FAD-dependent portion of the reaction.

Thermoactivity and thermostability

Fig. 3A shows the thermoactivity of the NADH oxidase reaction of rNOX1 at pH 7.50 in sodium phosphate buffer in both the presence and the absence of additional FAD. In comparison to other thermostable enzymes, rNOX1 shows relatively broad temperature specificity, with significant activity below 50 °C. The FAD (substrate-level)-dependent reaction appears to show stronger temperature dependence than the FAD (substrate level)-independent reaction. As expected, the enzyme is remarkably thermostable, with a 30% loss of activity after 120 h at 85 °C and a 75% loss of activity after 24 h at 95 °C (incubations were 2 µm enzyme in 50 mm Tris pH 7.50, in the absence of FAD).

Figure 3.

Activity of rNOX1 is dependent on temperature, solvent and FAD concentration. (A) Temperature-dependence of rNOX1 NADH oxidase activity, using the standard assay buffer (50 mm sodium phosphate, 100 µm NADH pH 7.50) in the presence (▪) and absence (●) of FAD. (B) Activity of rNOX1 in the presence of several organic solvents in the standard assay buffer. ▪, Ethanol at 55 °C; ○, dimethylsulfoxide at 75 °C; □, dimethylformamide at 75 °C (C) The dependence of the ratio of NADH consumed to H2O2 produced on FAD concentration. Standard assay conditions were used, except that the NADH concentration was 37 µm and the FAD concentration was varied as indicated.

Activity in the presence of organic solvents

The ability of redox enzymes to catalyze industrially important reactions makes them of interest from a biotechnological point of view [16,17], and the thermophilic nature of NOX1 makes it especially so. As these reactions often require organic solvents, rNOX1 was tested for activity in the presence of several organic solvents. rNOX1 is active in the presence of high concentrations of several organic solvents, as shown in Fig. 3B. The enzyme shows significant activity in dimethylsulfoxide and dimethylformamide at 75 °C, and ethanol at 55 °C. The enzyme is remarkably stable and active in ethanol, showing higher activities at moderate concentrations (25%) of ethanol. While rNOX1 shows lower specific activities in the solvents indicated (except for ethanol), the linear appearance of the assays indicates that this is due to an intrinsic lower activity in the presence of the solvents and is not due to denaturation or inactivation of the enzyme during the assays.

Production of both H2O2 and H2O as products

A thermophilic NADH oxidase reaction containing 37 µm NADH (and no FAD) was allowed to go to completion (all NADH consumed) in an open cuvette, allowed to cool to room temperature, and assayed for H2O2 via a coupled reaction with horseradish peroxidase and ABTS dye [13]. This reaction shows that 24 µm H2O2 is produced for the 37 µm NADH consumed. FAD and NAD+ do not interfere with the horseradish peroxidase/ABTS dye H2O2 assay; however, care must be taken to ensure that all NADH is consumed as excess NADH re-reduces the oxidized ABTS dye and decreases the observed H2O2. Control experiments in which background reactions were run with 37 µm NADH and 100 µm FAD with no enzyme at 75 °C, cooled to room temperature, and lactate dehydrogenase and pyruvate were added to consume all remaining NADH showed that < 3 µm H2O2 is produced via nonenzymatic routes under the assay conditions. The NADH/H2O2 ratio of 0.61 suggests strongly that both H2O2 and H2O are being produced from the reduction of O2 by rNOX1, as the H2O2-producing reaction should consume 1 NADH per H2O2 produced, whereas the H2O-producing reaction would consume 2 NADH while producing no H2O2. It can be calculated that NOX1 produces 23% water and 77% H2O2 as products under the assay conditions given. Because a likely mechanism for H2O2 production would involve enzymatic reduction of substrate-level FAD, followed by nonenzymatic oxidation of free FADH2, the NADH/H2O2 ratio was calculated across a range of FAD concentrations. As shown in Fig. 3C, at a constant NADH concentration of 37 µm the amount of H2O2 produced shows only a slight increase between 0 and 140 µm FAD.

In order to confirm that the reaction produces both H2O and H2O2 products, the ratio of NADH consumed to oxygen consumed was determined. These reactions were performed in sealed cuvettes with no headspace at 75 °C with oxygen concentrations of 110 µm (ambient O2 concentration at 75 °C) and 550 µm (buffer saturated with O2). The reactions were allowed to proceed until all of the oxygen in the cuvette had been totally consumed, as determined by the endpoint of the oxidase reaction. When corrected for the background rate of thermal decomposition of NADH under anaerobic conditions, a ratio of 1.38 mole NADH consumed to each mole O2 present was obtained at the endpoint, confirming the result obtained using the H2O2 assay above. From the NADH/O2 ratio of 1.38 and based on the stoichiometries of the H2O and H2O2-producing reactions, it can be calculated that the enzyme is producing 28% H2O and 72% H2O2, which agrees closely with the result obtained above. From the initial, linear portion of these reactions kcatapp values of 4.0·s−1 in the presence of 110 µm O2 and 12.6·s−1 in the presence of 550 µm O2 were obtained, indicating that under the standard assay conditions the enzyme is not saturated for O2 and that the Km for O2 is > 110 µm.

Peroxidase and disulfide reductase activity

The P. furiosus NOX1 shows a moderate degree of identity with both the NADPH-dependent CoADR of Staphylococcus aureus and the NPX of E. faecalis. CoADR is an unusual disulfide reductase, catalyzing its disulfide reductase activity via a single cysteine residue at the active site [18]. This cysteine cycles between a mixed-CoA disulfide oxidized form and a free thiol reduced form [18]. NPX utilizes a single redox-active cysteine to catalyze the reduction of H2O2 to H2O; in this case the cysteine cycles between sulfenic acid (R-SOH, oxidized) and cysteine thiolate (R-S, reduced) forms [1]. Because of these identities, rNOX1 was assayed for disulfide reductase and peroxidase-like activities with a wide range of substrates. rNOX1 showed no significant disulfide reductase activity with NADH or NADPH and 5,5’-dithiobis(2-nitrobenzoic acid), oxidized glutathione, oxidized coenzyme A, or cystine. rNOX1 also showed no peroxidase-like activity with NADH or NADPH and hydrogen peroxide or cumene hydroperoxide.

Transcriptional analyses of the nox1 gene

nox1 appears to be expressed at high levels regardless of the carbon source. A single transcript of 1.5 kb was identified by Northern analysis (Fig. 5). The presence of maltose (10 mm), pyruvate (40 mm), or tryptone (5 g·L−1) had no effect on the expression levels. A similar expression pattern was also observed with the NADH oxidase from E. faecalis, which is also expressed constitutively at high levels regardless of the available carbon source (A. Claiborne, Wake Forest University, Winston-Salem, NC, USA, personal communication). A single major, apparent mRNA 5′ end was identified upstream of nox1 by primer extension (Fig. 3). Analysis of the upstream sequence identified the TATA-box found 25 bp upstream of the transcription initiation site and the TFB-responsive element or BRE element common to archaeal promoters.

Figure 5.

Transcriptional analyses of the nox1 gene. (A) Northern blot analysis of the P. furiosus nox transcript. P. furiosus was grown with the following compounds as primary carbon and energy source: 10 mm maltose (M), 40 mm pyruvate (P), or 5 g·L−1 tryptone (T). (B) Identification of the transcriptional start site for the nox by primer extension analysis. (C) Sequence of the nox promoter region. The transcriptional start site as determined by primer extension analysis is designated by an arrow. The TATA and BRE sequences are boxed. The putative ribosome binding site (SD) is shown and the final ATG of each sequence is the first codon of each gene.


rNOX1 represents the first H2O-producing NADH oxidase homologue of the Archaea to be characterized. While rNOX1 does in fact show a significant NADH oxidase activity, the kcat of 4.8·s−1 for the oxidase reaction in the absence of substrate-level FAD is much lower than kcat values usually seen for NADH oxidases. The E. faecalis enzyme, for example, has a kcat of 320·s−1 at 25 °C [19]. Given that the optimal growth temperature for Pyrococcus is 100 °C, the activity of the enzyme is probably much greater under physiological conditions; however, the turnover number at the highest temperature examined in these studies (88 °C) was only 5.5·s−1 (in the absence of substrate-level FAD). This may be due, however, to the sharp decrease in oxygen solubility as temperatures approach the boiling point of water. The actual oxygen concentrations that Pyrococcus itself might encounter in the high-pressure volcanic vent environment in which low temperature (2 °C) and high temperature water are continuously mixing is difficult to determine. The low Km for NADH does lead us to the conclusion that NADH is the probable reducing substrate for this enzyme.

The role of the substrate-level FAD in the FAD-dependent portion of the reaction is not clear. The behavior of the enzyme during size-exclusion chromatography indicates that FAD remains enzyme-bound at room temperature, while from the small, reversible change in the visible spectra seen between 25 and 75 °C, it can be calculated that at least 82% of the FAD remains in the enzyme-bound form at 75 °C. NOX1 may be acting as a FAD reductase, possibly through the enzymatic reduction of free FAD via an enzyme-bound FADH2. In such a mechanism, it seems likely that FMN would be able to substitute for FAD, which is not seen with rNOX1. Another possibility would be an exchange mechanism where enzyme bound FADH2 is released from the enzyme and replaced by a solution FAD, as is seen during FMN reduction in the bacterial luciferase reaction of Vibrio harveyi[20]. The detailed mechanism of both the FAD-dependent and independent reactions are currently being investigated and will be the subject of a future manuscript.

Perhaps the most puzzling aspect of the reactivity of rNOX1 is the observation that the enzyme produces both H2O and H2O2 in significant quantities; to our knowledge this is the first enzyme that has been shown to do this. A minimal mechanism is shown in Fig. 4 for such a two-path mechanism. This mechanism includes a peroxyflavin intermediate that has been observed during stopped-flow studies of the E. faecalis NOX [21]. The production of both products leads to contradictory conclusions regarding the oxidase activity of rNOX1. On the one hand, production of H2O indicates that rNOX1 functions to combat oxidative stress through the reduction of oxygen to water while at the same time regenerating oxidized pyridine nucleotide for fermentative carbohydrate and amino-acid consumption. On the other hand, the production of H2O2 by the NADH oxidase appears not to be physiologically useful. This has been the case with an array of enzymes with H2O2-producing NADH oxidase activities. These include the alkyl hydroperoxide reductase of Salmonella typhimurium[22], the NADH oxidase of Amphibacillus xylanus (which acts as part of an alkyl hydroperoxide reductase system) [23] and the NADH oxidase of Sulfolobus solfataricus (which appears to be a thioredoxin reductase) [10].

Figure 4.

Minimal two-path mechanism for the reactivity of rNOX1.

If NOX1 cycles through a cysteine-sulfenic acid, it is likely that Cys42 can be overoxidized to the sulfinic (R-SO2H) or sulfonic (R-SO3H) acid states, as seen with the E. faecalis NPX and NOX [24–26]. One possible explanation for the two products observed with rNOX1 would be the presence of two distinct enzyme populations: (a) a Cys-SOH containing enzyme, which would produce H2O as a product, and (b) a Cys-SO2H or -SO3H containing enzyme that would produce H2O2. The observed spectrum of rNOX1, however, appears to provide evidence against this explanation. In the case of NPX and NOX, a distinct spectral change upon over-oxidation is seen in the loss of the broad charge-transfer band at > 500 nm [24,25]. This band is also missing in the both the C42S and C42A mutants of NPX [27] and the C42S mutant of NOX [21]. The presence of a very similar appearing broad charge transfer band in rNOX1 at > 500 nm at least suggests that a large population of its putative cysteine redox center is intact.

While the ratio of H2O2 to H2O product changes very little with the concentration of substrate-level FAD, it is possible that this ratio changes with varying concentrations of NADH or O2. It is possible that at low or high concentrations of either of these substrates the amount of H2O2 produced may decrease to a level that would make the NADH oxidase reaction more obviously beneficial from the organism's point of view. One possible mechanism for the partial reduction of O2 to 2H2O would be through the initial 2e reduction of O2 to H2O2, followed by the reduction of this intermediate or first product to H2O by another 2e from NADH. The lack of any peroxidase activity for rNOX1 argues against this sort of mechanism. We cannot, however, rule out a mechanism in which H2O2 formed at the active site is further reduced to H2O, while the active site remains inaccessible to substrate-level H2O2. Further studies are currently underway examining the kinetics of rNOX1, as well as characterization of the rNOX2 of both P. horikoshii and P. furiosus.

Recently, Pyrococcus has been shown to have the makings of a formidable defense against oxidative stress. A superoxide reductase has been isolated from P. furiosus that reduces superoxide radicals to H2O2 via NAD(P)H and rubredoxin [11]. The H2O2 product of this reaction, while far less toxic than O2, would still require subsequent removal. It was previously suggested that an NADH peroxidase homologue such as NOX1 might play a role in the removal of H2O2. We have found, however, that rNOX1 does not clearly exhibit a significant peroxidase or alkyl hydroperoxide reductase activity. This does not rule out a role for NOX1 in a multienzyme peroxidatic complex, as is seen in the case of the alkyl hydroperoxide reductase activity of the AhpF/AhpC system of Salmonella typhimurium[22].

Jenney et al. did not detect a catalase activity in crude extracts of P. furiosus[11] and we have not been able to identify any catalase homologues in the P. furiosus genome. This result was not unexpected, as a catalase would simply regenerate oxygen. One or more peroxiredoxin (Prx)-type peroxidase(s) analogous to those characterized for a wide range of other organisms are, however, present in the Archaea [11,28,29]. Prx proteins, which catalyze the cysteine-dependent reduction of H2O2 and alkyl hydroperoxides, are present in both ‘1-Cys’ and ‘2-Cys’ versions, and homologues of both types have been found in the P. furiosis genome. Electron donors to Prxs in other organisms include thioredoxin (Trx), bacterial AhpF, or specialized small redox-mediating proteins [30,31]. No homologues of AhpF or Trx appear to be encoded in the P. furiosis genome, however, the possibility remains that an unidentified enzyme with an AhpF-type activity works with one or both Prx homologues in removing peroxides in this organism.


The authors thank L. B. Poole for helpful discussions and E. G. Senkbeil for assistance with size-exclusion HPLC.