SEARCH

SEARCH BY CITATION

Keywords:

  • electron microscopy;
  • H2O-producing;
  • hexameric ring structure;
  • NADH oxidase;
  • thermophilic archaeon

Abstract

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

An NADH oxidase (NOX) was cloned from the genome of Thermococcus profundus (NOXtp) by genome walking, and the encoded protein was purified to homogeneity after expression in Escherichia coli. Subsequent analyses showed that it is an FAD-containing protein with a subunit molecular mass of 49 kDa that exists as a hexamer with a native molecular mass of 300 kDa. A ring-shaped hexameric form was revealed by electron microscopic and image processing analyses. NOXtp catalyzed the oxidization of NADH and NADPH and predominantly converted O2 to H2O, but not to H2O2, as in the case of most other NOX enzymess. To our knowledge, this is the first example of a NOX that can produce H2O predominantly in a thermophilic organism. As an enzyme with two cysteine residues, NOXtp contains a cysteinyl redox center at Cys45 in addition to FAD. Mutant analysis suggests that Cys45 in NOXtp plays a key role in the four-electron reduction of O2 to H2O, but not in the two-electron reduction of O2 to H2O2.

Abbreviations
CoADR

coenzyme A disulfide reductase

GR

glutathione reductase

Nbs2

5,5′-dithiobis-(2-nitrobenzoic acid)

NOX

NADH oxidase (EC 1.6.99.3)

NOXtp

Thermococcus profundus NADH oxidase

Thermococcus profundus is a thermophilic anaerobic archaeon belonging to the Thermococcaceae family that also includes Thermococcus kodakaraensis KOD1, a model thermophilic organism whose whole genome sequence has been reported [1]. As an anaerobe living in deep-vent environments, it seems likely that Thermococcus encounters high levels of oxygen stress in the water surrounding the vent [2]. In anaerobes, flavin-dependent NAD(P)H oxidases play an important role protecting organisms from oxidative stress [3].

NADH oxidase (NOX) is a member of the flavoprotein disulfide reductase family that catalyzes the pyridine-nucleotide-dependent reduction of various substrates, including O2, H2O2 and thioredoxin [4]. There are two types of NOX: those that catalyze the two-electron reduction of O2 to H2O2 and those that catalyze the four-electron reduction of O2 to H2O [5]. The physiological role of NOX is diverse, depending on its substrates and products in different organisms. In anaerobic mesophiles, NOX enzymes, such as those of Clostridium aminovalericum [6], Enterococcus (Streptococcus) and Lactococcus [3], are considered to be important enzymes in protecting against oxidative stress and in regenerating oxidized pyridine nucleotides through their capacity to reduce O2 to H2O without the formation of harmful reactive oxygen species. Some NOX proteins have also been purified and studied in (hyper)thermophilic organisms. NOX from Archaeoglobus fulgidus may be involved in electron transfer in sulfate respiration [7]. An H2O2-forming NOX functions as an alkyl hydroperoxide reductase in Amphibacillus xylanus [8]. Some NOX enzymes, such as those of Pyrococcus furiosus [9] and Thermotoga maritima [10], have been proposed to protect anaerobes from oxidative stress. In (hyper)thermophiles, the roles of some NOX enzymes remain to be elucidated [11].

NADH oxidase varies with the organism; however, these proteins generally share similar secondary structural folding [4,12]. An NOX from Thermus thermophilus is a homodimer as determined by X-ray crystallography [13]. Gel filtration chromatography indicated that NADH:flavin oxidoreductase from Eubacterium is composed of three identical subunits [14]; NOX in Clostridium thermohydrosulfuricum is probably made up of six subunits, as demonstrated by gel filtration [15]. In contrast, a heterogeneous NOX from Eubacterium ramulus is proposed to have an α8β4 assembly, as revealed by gel filtration and PAGE [16].

Two NOX enzymes from the Thermococcaceae family have been described. One is a novel enzyme in P. furiosus that produces both H2O2 (77%) and H2O (23%) [9]. The other NOX, in Pyrococcus horikoshii OT3, may function as a CoA disulfide reductase (CoADR) [17]; however, the function and structure of NOX in Thermococcus, a genus of the Thermococcaceae family, has not been clarified. In this study, we have cloned, overexpressed and purified a NOX that is composed of two cysteine moieties from T. profundus. We report its biochemical characterization and structure, and also used mutants to analyze its catalytic mechanism.

Results

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

Cloning and sequencing of the nox gene from T. profundus

In order to clone the T. profundus nox gene (NOXtp), we utilized a PCR-based DNA-walking method using the ClonTech genome-walker cloning kit, as described in Experimental procedures; the resulting DNA sequence comprised an ORF of 1329 bp, predicting a protein composed of 442 amino acids with a molecular mass of 48 611 Da. Figure 1 shows the nucleotide sequence of NOXtp and its flanking regions, together with the translated amino acid sequence. The 5′-flanking region of NOXtp contained a putative Archaea promoter with a TATA box and ribosome-binding site. The 3′-flanking region did not match with other Archaea genes, as judged by homolog searches in the NCBI database. Unlike NOX, which has only one conserved cysteine residue (Cys45) in its N-terminus [4], the amino acid composition of NOXtp revealed the presence of two cysteine residues, Cys45 and Cys122. Additionally, two conserved cofactor-binding domains were also identified in NOXtp. One was a FAD-binding domain containing the AMP-binding and FMN-binding motifs observed in enzymes belonging to the glutathione reductase (GR) family [18]. The other domain was a glycine-rich NAD-binding motif located between the AMP-binding and FMN-binding motifs (two FAD-binding domains) (Fig. 1). We propose that NOXtp belongs to the GR family, because of the high sequence identity of the cofactor-binding domains described above.

image

Figure 1.  Nucleotide sequence of the noxtp gene and predicted amino acid sequence of the gene product from Thermococcus profundus. The putative TATA-box and ribosome-binding site are underlined and in bold letters, respectively. The residues involved in FAD binding are shadowed in gray. The NAD-binding site is boxed. The cysteine residues are in bold italic.

Download figure to PowerPoint

Multiple sequence alignment (Fig. S1) revealed that Cys45 is located at a similar position to that of the cysteine residue in the conserved active site of NOX from P. horikoshii (also called CoADR) [17] and NOX and NADH peroxidase from Enterococcus faecalis [19,20]. Sequence analysis by clustal w showed that NOXtp shared a significant level of identity with NOX (CoADR) from P. horikoshii (80%) [17], NADH peroxidase from E. faecalis (28%) [20], and NOX from P. furiosus (36%) [9], Lactococcus lactis (30%) [21], Lactococcus sanfranciscensis (26%) [11] and E. faecalis (27%) [19] (Fig. S1). These proteins are generally composed of two identical subunits related by two-fold symmetry. Each subunit can be divided into a C-terminal dimerization domain and an N-terminal pyridine nucleotide disulfide oxidoreductase domain, which is actually a small NADH-binding domain with a large FAD-binding domain [4,12]. NOXtp has similar primary structure architecture to these proteins as determined by NCBI protein blast analysis.

Purification of native and recombinant NOXtp

In order to understand the oxygen detoxification mechanism of anaerobic microbes, we purified NOXtp from T. profundus by several chromatographic methods. The purified protein revealed a subunit with a molecular mass of approximately 50 kDa (Fig. S2). The N-terminal amino acid sequence of purified NOXtp from T. profundus was determined to be MERKRVVIIGGGAAG, which is highly similar to that of NOX in T. kodakarensis KOD1, P. furiosus, Pyrococcus abyssi and P. horikoshii OT3, belonging to the pyridine nucleotide disulfide oxidoreductase family. The purification of recombinant NOXtp from Escherichia coli was performed by ion exchange chromatography as described in Experimental procedures. SDS/PAGE analysis of recombinant NOXtp revealed a molecular mass of approximately 50 kDa, which is close to that of the purified protein from T. profundus (Fig. S2). However, gel filtration analysis under nondenaturing conditions showed that the purified NOXtp had a molecular mass of approximately 300 kDa (Fig. 2). These results indicated that NOXtp is a hexamer of 50 kDa subunits, in contrast to NOX proteins from thermophilic archaeans, which have been reported to be dimers or tetramers [13,17].

image

Figure 2.  Gel filtration chromatography profile of NOXtp purified from Es. coli. The purified protein was subjected to Superdex-200 gel filtration chromatography. Absorbance was measured at 280 nm. The x-axis shows the elution time. The standard proteins are ferritin (440 kDa), catalase (232 kDa), albumin (67 kDa) and ovalbumin (43 kDa).

Download figure to PowerPoint

Structure of NOXtp

Gel filtration analysis under nondenaturing conditions revealed that purified NOXtp has a molecular mass of ∼ 300 kDa, corresponding to a hexamer with 50 kDa subunits (Fig. 2). This structure is different from that of other homologous NOX proteins, which consist of dimers or tetramers as revealed by X-ray crystallographic studies [12,13,22]. In order to clarify the oligomeric structure of NOXtp, we performed electron microscopy using purified NOXtp. The electron micrographs of the negatively stained NOXtp oligomers showed a uniform distribution of the ring-shaped structure in the top-view orientation (Fig. 3A). In total, 939 well-stained particles were translationally aligned, and were subjected to multivariate statistical analysis [23]. The eigenimages obtained from the translationally, but not rotationally, aligned images revealed a six-fold rotational symmetry (Fig. 3Ba). Using the 10 most significant eigenvectors, nine classes were discriminated on the basis of similarity of features after rotational alignment without symmetrization. Most class averages showed a star-shaped structure with six-fold symmetry with heavy stain accumulation in its center (Fig. 3Bb). In particular, the two class averages shown in panels 3 and 6 of Fig. 3Bb exhibited an obvious deviation from the star-like structure. This could result from incomplete stain embedding of the particle or from an unintentional inclination during preparation or microscopy. In order to analyze further the rotational symmetry of the top-on-view images, the same dataset was separated into many classes (10–30) using different eigenimages (10–20). The dataset was also aligned with an arbitrarily chosen reference and separated according to the similarity of features in the eigenimages. The resulting class averages revealed no other statistically significant symmetry (data not shown). We found no evidence for the existence of a NOXtp protein with intrinsically lower symmetry, at least at the resolution employed.

image

Figure 3.  Electron micrograph and structural analysis of NOXtp. (A) Purified NOXtp was absorbed onto the grids as described in Experimental procedures. The electron micrograph of the protein was then obtained by negative staining with 2% uranyl acetate. (B) Multivariate statistical analysis of NOXtp. (a) The average (AV) of 939 translationally, but not rotationally, aligned particles with end-on orientation and the 10 most significant eigenimages (numbers 1–10) are shown. In (b), the nonsymmetrized class averages (numbers 1–9) were derived from rotationally aligned images using the 10 most significant eigenvectors. The numerals shown in the top right corner of the class averages are the number of particles seen in each class. (C) The average of the side-on view of NOXtp (939 particles). (D) A schematic model for the assembly of NOXtp complexes. The diameters of the cavity, middle ring and outer ring are 4, 15 and 19 nm, respectively.

Download figure to PowerPoint

The average of the 939 top-on views revealed a star-shaped structure (Fig. 3C), which contained a middle region of high density with heavy stain accumulation in its center. The average view also revealed that the density of the complex was not homogeneous; the density increased towards the middle, such as seen in the valosine-containing protein-like ATPase from Thermoplasma acidophilum complex, which is composed of two stacked ring structures of different diameters [24]. The upper (or middle) ring of the NOXtp complex has a region that is denser than that of the outer ring, indicating the presence of a cavity in the complex with a width of approximately 4 nm. The diameters of the outer ring and the middle ring (Fig. 3D) were approximately 19 and 15 nm, respectively. These projected images as well as the gel filtration analysis indicated that NOXtp predominantly exhibits a hexameric star-shaped structure, in contrast to the structure recently reported by Kuzu et al. [22], which suggested a tetrameric structure for NOX from Lactobacilluus brevis.

NOXtp is an FAD-dependent NADH and NADPH oxidase

On the basis of the amino acid sequence, NOXtp contains two FAD-binding domains. The isoalloxazine ring system in FAD has been suggested to induce light absorbance in the UV and visible spectral range, giving rise to the yellow appearance of flavin and flavoproteins [25]. We performed light absorbance analysis to confirm NOXtp binding to FAD. Purified NOXtp from Es. coli has absorption maxima at 378 and 456 nm, with a shoulder at 480 nm, which are characteristic spectral features of proteins with bound flavin cofactors (Fig. 4A). The absorbance behavior also allowed the determination of the number of flavin molecules bound per mole of NOXtp subunit [17,25,26]. A stoichiometry of 0.7–0.9 mol FAD per mol NOXtp subunit was determined from the absorbance at 460 nm.

image

Figure 4.  Activity assays of NADH and NADPH oxidase. (A) Visible spectra of NOXtp (solid line), apo-NOXtp (dashed line) and the C45A mutant (dotted line). The absorbance was measured in 50 mm sodium phosphate buffer (pH 7.2) at 25 °C. (B) FAD effect on NAD(P)H oxidase activity. An activity assay was performed as described in Experimental procedures. The solid line shows the NADH oxidase activity of NOXtp purified from Es. coli (□), reconstituted NOXtp (○), and apo-NOXtp (△). The dashed line shows the NADPH oxidase activity of NOXtp from Es. coli (□), reconstituted NOXtp (○), and apo-NOXtp (△). (C) Optimal temperature of NAD(P)H oxidase activity. The assay was performed at the indicated temperatures in 50 mm potassium phosphate buffer (pH 7.2). NADH and NADPH oxidase activity are shown by a solid line and a dashed line, respectively. The squares show the measured temperature points. (D) Optimal pH of NAD(P)H oxidase activity. Different buffers were used in this assay. Sodium phosphate was used at pH 6.0, 6.6, 7.2 and 7.7; Hepes buffer and Tris buffer were used at pH 8.0 and 8.5; sodium borate buffer was used at pH 9.0. These buffers were used at a concentration of 50 mm. NADH and NADPH oxidase activity are shown by a solid line and a dashed line, respectively. The squares show the measured pH points.

Download figure to PowerPoint

As NOXtp contains FAD as a prosthetic group, apo-NOXtp was prepared by hydrophobic interaction chromatography under acidic conditions (pH 3.5) with saturated NaBr buffer [26,27], in order to confirm the function of FAD. The absorption spectrum of apo-NOXtp did not show any significant absorbance in the visible region, revealing that FAD was indeed absent (Fig. 4A). To determine whether FAD was required for the enzymatic activity of NOXtp, holoprotein and apoprotein activities were assayed. The NADPH oxidase activity of NOXtp was also measured, as described previously for NOX (CoADR) from P. horikoshii [17] and NOX from L. sanfranciscensis [12], which show high similarity to NOXtp and accept both NADH and NADPH as cofactors. The activity of the reconstituted enzyme, which was accomplished by incubating equimolar concentrations of apomonomers and FAD at room temperature for 5 min [26], was also measured. These assays revealed that NADH oxidase activity was slightly higher than that of NADPH oxidase, and FAD significantly restored the oxidase activity of apo-NOXtp (Fig. 4B). These results clearly indicated that NOXtp is an FAD-dependent NADH and NADPH oxidase, in contrast to NOX enzymes from other thermophilic archaeons, which only exhibit activity towards NADH [9–11].

To further determine the function of NOXtp, the steady-state kinetic parameters of NOXtp with either NADH or NADPH as the reducing substrate were measured at pH 7.2. NOXtp could catalyze NADH and NADPH oxidization with kcat values of 6.2 ± 0.5/s and 2.5 ± 0.3/s, respectively. The steady-state kinetic parameters of NOXtp were similar to those of NOX (CoADR) from P. horikoshii (Table 1). On the basis of the Km, both enzymes preferred NADPH as the substrate for oxidase activity, indicating that NOX (CoADR) from P. horikoshii and NOXtp belong to similar enzyme families. The optimal temperature for the NADH and NADPH oxidase activity of NOXtp was near 70 °C (Fig. 4C), which is lower than the optimal growth temperature (80 °C) of this organism. The optimal pH was between 7.5 and 8.0 for both NADH and NADPH oxidase activity (Fig. 4D).

Table 1.   Steady-state kinetic parameters of NOXtp and NOX (CoADR) from Pyrococcus horikoshii (50 mm potassium phosphate buffer, pH 7.2, 75 °C). Data shown are means of triplicate determinations ± SD.
ParameterNOXtp-NADH oxidaseNOXtp-NADPH oxidaseCoADR-NADH oxidaseCoADR-NADPH oxidase
  1. a From reference [17].

Kmm)53.1 ± 2.812.1 ± 1.173a13a
kcat (s−1)6.2 ± 0.52.5 ± 0.38.2a2.0a

NOXtp preferentially produces H2O

The product of O2 reduction is an important factor in evaluating the physiological function of NOXs [10]. For instance, NOX from P. furiosus, which may protect anaerobic thermophiles against oxidative stress, can produce both H2O2 and H2O [9]. In order to determine the product of the NAD(P)H oxidase activity of NOXtp, reactions containing 100 μm NAD(P)H were performed [all NAD(P)H consumed] according to the published method [9], and H2O2 was quantified using a peroxi-DETECT kit from Sigma (St Louis, MO, USA). When NADPH oxidation was performed at 80 °C, approximately 7% of the NADPH supplied was used to produce H2O2, and 2% of the NADH was recovered as H2O2 under the same conditions (Fig. 5D). These results demonstrated that NOXtp produces predominantly H2O using NADH and NADPH as electron donors.

image

Figure 5.  Comparisons of activity, products and structures between NOXtp and the mutants. (A) Electron micrographs of NOXtp (a), NOXtpC45A (b), NOXtpC122A (c) and NOXtpC45A/C122A (d). The bar represents 100 nm. (B) Native PAGE of wild-type NOXtp and its mutants. Lanes 1–4 correspond to (a), (b), (c) and (d) in (A), respectively; lane 5 is the molecular weight marker. The lower part shows the corresponding proteins determined by SDS/PAGE. (C) Specific activity of wild-type NOXtp (a), NOXtpC45A (b), NOXtpC122A (c) and NOXtpC45A/C122A (d) with NADH (bar 1) and NADPH (bar 2) as substrates. (D) The amount of H2O2 produced by NOXtp (a), NOXtpC45A (b), NOXtpC122A (c) and NOXtpC45A/C122A (d) when 100 μm NADH (bar 1) or 100 μm NADPH (bar 2) was oxidized.

Download figure to PowerPoint

Cys45 but not Cys122 functions as the nonflavin redox center

NADH oxidase in members of the Thermococcaceaee family, such as T. kodakaraensis KOD1, P. horrikoshii, P. abyssi and P. furiosus, have only one conserved cysteine residue, Cys45; however, the sequence of NOXtp (Fig. 1) revealed that it contains two cysteine residues, Cys45 and Cys122. As cysteines are important residues for NOX enzyme activity, we replaced Cys45 and Cys122 with alanines to analyze the function of these two residues. After purification using the same method as that used for the wild-type enzyme, the number of cysteines in the three mutant enzymes (NOXtpC45A, NOXtpC122A and NOXtpC45A/C122A) was examined using Ellman’s method (Table 2). The single mutants, NOXtpC45A and NOXtpC122A, contained about one cysteine, and the double mutant contained no cysteines. These data confirmed the identity of the mutants and also indicated that the nonmutated cysteine remained in its native state. The visible absorption spectra showed that the three mutants contained tightly bound FAD (Fig. 4A, NOXtpC45A only shown – the other two mutants produced similar absorbance spectra). Electron microscopy and native PAGE showed no significant difference between wild-type NOXtp and its mutants (Fig. 5A,B). All of the data indicated that the disulfide bond was not responsible for hexameric oligomerization and that substitution of Cys45 and Cys122 with alanine did not result in major changes in NOXtp quaternary structure.

Table 2.   Determination of the sulfhydryl contents of wild-type and mutant NOXtp using Ellman’s reagent. Data shown are means of triplicate determinations ± SD.
ProteinNo. cysteines per protein
NOXtp1.84 ± 0.31
NOXtpC45A0.85 ± 0.34
NOXtpC122A0.84 ± 0.18
NOXtpC45A/C122A0.14 ± 0.07

In order to determine the catalytic mechanism of NOXtp, NAD(P)H oxidase assays were performed with the three mutants under the same conditions as used for the wild-type. The results showed that the C122A mutant had similar NADH and NADPH oxidase activity to that of the wild-type protein; however, the C45A mutant and the double C45A/C122A mutant had < 10% of the NAD(P)H oxidase activity of the wild-type protein (Fig. 5C). These results are similar to those obtained with a NOX from E. faecalis, where a serine substitution of its active site residue Cys42 (C42S) resulted in approximately 3% of the activity of the wild-type under the same conditions [4,28]. Considering these results, Cys45 may provide the essential second redox center in addition to the flavin. We further examined the products of NOXtp and its mutants. NAD(P)H oxidation was allowed to go to completion, and the amount of H2O2 formed in the reaction was quantified using the peroxi-DETECT kit. The NOXtpC122A mutant produced a similar amount of H2O as the wild-type under the same conditions and with the same substrates (Fig. 5D). Oxidation of NADH and NADPH by NOXtpC45A and NOXtpC45A/C122A led to the formation of about one equivalent of H2O2 (Fig. 5D), demonstrating that H2O2 production by these two mutants is stoichiometric with NADH and NADPH oxidation. The activity and product assays using the wild-type and mutants clearly demonstrated that Cys45 participates in the direct four-electron reduction of O2 to H2O, and the Cys45 mutation alters the reaction to produce H2O2 instead of H2O.

Discussion

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

In this study, we have demonstrated that NOXtp has a hexameric ring-shaped structure. Gel filtration under nondenaturing conditions revealed that NOXtp is composed of six subunits. Moreover, upon electron microscopic analysis, NOXtp was found to predominantly exhibit a hexameric structure that contained a middle region of high density with heavy stain accumulation in its center. However, the crystal structure of NOX from L. sanfranciscensis revealed a dimeric form with an N-terminal oxidoreductase domain and a C-terminal dimerization domain [12]. NPX from Streptococcus faecalis, catalyzing the conversion of H2O2 to H2O, was reported to be a homotetrameric structure [29]. These two mesophilic proteins show different types of subunit oligomerization and low sequence identity (Fig. S1), but each of their subunits shows high structural similarity and their folding patterns are similar to that of GR [12,29]. In contrast, NOX from Thermoanaerobium brockii was found to have a hexameric quaternary structure by gel filtration [15]. Electron microscopic analysis has revealed that NOXtp has a hexameric ring-shaped structure composed of two stacked rings of different diameters (19 and 15 nm respectively) that encompass a central opening; this is the first hexameric NOX determined by electron microscopy. Significantly, this structural feature of NOXtp is highly similar to that of valosine-containing protein-like ATPase from Th. acidophilum, an archaeal member of the AAA family (ATPases associated with a variety of cellular activities) [24]. In addition, the structure of the cysteine mutants, NOXtpC45A, NOXtpC122A and NOXtpC45A/C122A, was the same as that of the wild-type. Thus, it appears that a disulfide bond does not participate in the oligomerization and quaternary structure of NOXtp.

NADH oxidase catalyzes the transfer of electrons from reduced pyridine nucleotides to O2 [2,4]. Here we have demonstrated that NOXtp can efficiently reduce O2 to produce H2O using both NADH and NADPH as electron donors. In addition, the activity and product assays of the wild-type and mutants showed that Cys45 is the active site residue and that Cys122 does not function in the NADH and NADPH oxidase activity. These results indicate that Cys45 participates in the direct four-electron transfer reduction of O2 to H2O, and that the Cys45 mutant alters the reduction to produce H2O2 instead of H2O, similar to NOX in E. faecalis [28]. NOX in E. faecalis belongs to a group of enzymes that use a cysteine sulfenic acid as the nonflavin redox center. These enzymes are found in Enterococcus and Streptococcus, which are aerotolerant anaerobes, where they play an important role in O2 tolerance [4]. For example, H2O-forming NOX-deficient mutants of Streptococcus pyogenes are unable to grow under high-O2 conditions, revealing the importance of NOX-scavenging activity against harmful O2 [30]. We therefore propose that NOXtp may decompose O2 in the anaerobe T. profundus.

The predominant production of H2O by NOXtp is in contrast to the exclusive production of H2O2 by most NOXtp homologs in thermophiles, such as NOX in A. fulgidus, Desulfovibrio gigas, Thermot. maritima and Thermoanaerobium brockii [10,11,15,31]. Previously, the production of H2O2 was considered to be the distinctive property of NOX proteins from thermophiles [10,11], with the exception of NOX from P. furiosus, which produces both H2O2 (77%) and H2O (23%) [9]. To our knowledge, NOXtp is the first NOX to be purified from thermophilic microorganisms that can catalyze electron transfer from NADH and NADPH to O2 and predominantly produce H2O. NOXtp is therefore better for removing O2 than other reported O2-scavenging systems, which must employ intermediates to reduce H2O2 produced by NAD(P)H oxidases, such as in D. gigas, where rubredoxin and neelaredoxin act as intermediates [31]. As NOXtp and the mesophilic enzymes that decompose injurious O2 belong to the same group (discussed above), and NOXtp reduces O2 to H2O directly, we propose that NOXtp may play an important role in O2 removal or aerobic tolerance in thermophilic anaerobes.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

Purification of NOXtp from T. profundus

Thermococcus profundus cells (8 L) were grown at 80 °C as reported previously [32]. After harvesting, the cells were dissolved in 20 mm potassium phosphate buffer (pH 6.5), containing 5 mm MgCl2, 0.5 mm EDTA, 1 mm dithiothreitol and 10% glycerol (PMEDG buffer), and disrupted by sonication. The homogenates were centrifuged at 10 000 g for 30 min. The supernatant was loaded on a phosphocellulose column that had been equilibrated with PMEDG buffer. After being washed completely, the proteins were eluted by 100, 200, 300, 400, 500 and 1000 mm NaCl in a stepwise gradient, and the eluates in 200 mm NaCl were dialyzed with 50 mm Tris buffer (pH 8.0) containing 400 mm NaCl. The sample was then loaded on an amino-benzimide column equilibrated with the same buffer. Unabsorbed proteins on the resin were collected and dialyzed with PMEDG buffer, concentrated using a centricon (Millipore, Billerica, MA, USA), and stored at −80 °C. Eluates in all steps were checked by transmission electron microscopy. The protein concentration was determined by the Bradford method, and BSA was used as standard.

SDS/PAGE and N-terminal sequencing

The purified enzyme was subjected to SDS/PAGE and electroblotted onto poly(vinylidene difluoride) membranes. The visible band was excised and applied to a protein sequence analyzer (Korea Basic Science Institute, Daejeon, Korea).

Cloning of NOXtp from T. profundus

Polymerase chain reaction experiments with T. profundus genomic DNA as a template were performed using degenerate oligonucleotides (sense primer, 5′-GTA GTA ATA ATA GGA GGA GGA GCN GCN GGN ATG-3′; antisense primer, 5′-TAN ACT TTY TCN CAN SWN GTY TGC AT-3′; N = A, G, C and T; Y = C and T; W = A and T). The sense primer was designed from the known N-terminal sequence, and the antisense primer was from the conserved C-terminal sequence of NOX. The experiment using the two oligonucleotides afforded an amplificate of ∼ 1.3 kb, which was ligated into the pTOPO vector (Invitrogen, Carlsbad, CA, USA), transformed, and confirmed by sequencing. The resulting sequence was used for subsequent cloning. Full gene cloning of NOXtp was performed using the Universal Genomewalker kit (ClonTech, Mountain View, CA, USA). Briefly, the genomic DNA was digested with EcoRV, DrabI, PvuII and SspI separately, and ligated to the adaptor provided by the kit. PCR was performed with the adaptor primers (provided by the kit) and tail-specific primers (N-terminal genomewalker primer, 5′-TGG AGG TCT TTG CCG CGC TTT TTG AT-3′; C-terminal genomewalker primer, 5′-GGT GTG CAG GCT GTA AAT GCC GAG AT-3′), which corresponded to the known sequence detected by the degenerate primers. The PCR products were ligated into the pTOPO vector, transformed, and sequenced.

Expression and purification of NOXtp in Es. coli

The new primers (sense primer, 5′-CGCGCG CCATGG AGAGGAAACGCGTTGTTAT-3′; antisense primer, 5′-CGCG AAGCTT TAAAACTTTAGAACCCTG-3′) were designed on the basis of the sequence of the Genomewalker result (the underlined bases indicate the restriction enzyme site). The PCR product and pET28-(a) were digested by NcoI and HindIII and ligated. The ligation product was transformed into Es. coli BL21(DE3) by electroporation. Finally, the recombinant vector (pENOXtp) was confirmed by sequencing.

Recombinant Es. coli cells (2 L) were cultured in LB broth to a D600 nm of 1.0, and then induced with 1 mm isopropyl-thio-β-d-galactoside for 4 h. After harvesting, the cells were resuspended in PMEDG buffer and disrupted by sonication. After centrifugation (3000 g, 30 min), the supernatants were heated at 65 °C for 30 min, and then the denatured proteins were removed by centrifugation (3000 g, 30 min). The supernatants were loaded onto a phosphocellulose column that had been equilibrated with the same buffer. After being washed completely, the proteins were eluted with 200 mm NaCl. The purified protein was checked by SDS/PAGE and dialyzed with PMEDG buffer, concentrated, and stored at −80 °C.

Mutagenesis

The primers used for the single cysteine to alanine mutants were as follows: C45A, forward primer, 5′-ACG GAA TGG GTG AGC CAC GCT CCC GCC GGT ATC CCC TAC GTA GTT GAG GGT-3′; C45A, reverse primer, 5′-ACC CTC AAC TAC GTA GGG GAT ACC GGC GGG AGC GTG GCT CAC CCA TTC CGT-3′; C122A, forward primer, 5′-CCG CAG GTT CCG GCG ATA GAG GGC GCC CAC CTG GAA GGA GTA TTC ACA GCA-3′; and C122A reverse primer, 5′-TGC TGT GAA TAC TCC TTC CAG GTG GGC GCC CTC TAT CGC CGG AAC CTG CGG-3′. The PCR was performed using Pfu polymerase (Takara, Kyoto, Japan), and the cycling parameters were: 95 °C for 5 min (one cycle), 95 °C for 30 s, and 68 °C for 12 min (12 cycles). After amplification, the PCR mixture was digested with DpnI and then transformed into Es. coli BL21(DE3) by electroporation. The mutants were confirmed by DNA sequencing. The double cysteine mutants were produced by the same method, except that pENOXtpC45A was used as the template and C122A primers were used for the amplification. The mutant proteins were purified using the same method as used for wild-type purification.

Gel filtration chromatography

The sample (1 mg·mL−1) was loaded onto a Superdex-200 column (Amersham Biotech, Piscataway, NJ, USA). Standard proteins included ferritin (440 kDa), catalase (232 kDa), albumin (67 kDa) and ovalbumin (43 kDa).

Apo-NOX preparation

The purified NOXtp from Es. coli is a holoenzyme with FAD. The protein was dialyzed with 100 mm phosphate buffer (pH 7.2) containing 2.4 m (NH4)2SO4, 1 mm dithiothreitol and 0.5 mm EDTA, and then loaded on the hydrophobic interaction chromatography column equilibrated with the same buffer. FAD was eluted with equilibration buffer saturated with NaBr (pH 3.5). The column was balanced again with the equilibration buffer, and the apoprotein was eluted with 100 mm phosphate buffer [26,27]. Eluates were dialyzed with the PMEDG buffer described above, and stored at −80 °C.

Enzyme assays

The NADH or NADPH oxidase activity of the recombinant protein was examined by time-dependent removal of NAD(P)H in aerobic conditions. The assays were performed in 50 mm sodium or potassium phosphate buffer (pH 7.2), 0.5 mm NAD(P)H and 100 mm NaCl at the indicated temperatures. The reaction was started by adding NOXtp in the amounts indicated. The rate of NAD(P)H consumption was measured by monitoring the decrease in A340 nm. One unit of activity was defined as the amount of enzyme catalyzing the oxidation of 1 μmol NADH per min at 75 °C in 50 mm potassium phosphate buffer (pH 7.2) and 0.5 mm NADH. To measure kinetic parameters, reaction rates were measured at a series of NAD(P)H concentrations, and the rates at various substrate concentrations were finally fitted by Lineweaver–Burk plots. The parameters (with standard deviation) were determined by three separate experiments.

Determination of the sulfhydryl content

The sulfhydryl contents were determined using Ellman’s reagent in anaerobic conditions according to a published method [17,33]. After the proteins and 5,5′-dithiobis-(2-nitrobenzoic acid) (Nbs2) were incubated for 15 min at 25 °C, A412 nm was monitored to estimate the number of cysteine residues present as protein/Nbs2 mixed disulfide. The sulfhydryl concentrations in these proteins were determined from a calibration curve created using known concentrations of standard l-cysteine solutions.

H2O2 detection

H2O2 was detected using the PeroxiDetect Kit (Sigma). Briefly, the assay was performed in 50 mm sodium phosphate buffer (pH 7.2), 100 μmol NAD(P)H, 1 mm EDTA, 100 mm NaCl and 0.2 nmol NOXtp. The reaction was allowed to go to completion. Reaction buffer (100 μL) was added to the kit. Peroxides convert Fe2+ to Fe3+ ions under acidic conditions. Fe3+ ions will then form a colored adduct with xylenol orange, which is observed at 560 nm. NAD(P)H will interfere with the H2O2 assay, so all of the NAD(P)H must be consumed completely.

Electron microscopy and image processing

Purified NOX was applied to glow-discharged carbon-coated copper grids. After the proteins had been allowed to absorb for 1–2 min, the grids were rinsed with droplets of de-ionized water, and stained with 2% (w/v) uranyl acetate. Electron micrographs were recorded with an FEI TECHNAI 12 microscope at a magnification of ×51 600 (nominal magnification of ×52 000) and an acceleration voltage of 120 kV.

Light-optical diffractograms were used to select the micrographs, to examine the defocus and to verify that no drift or astigmatism was present. Suitable areas were digitized as arrays of 1024 × 1024 pixels with leaf scan 45 at a pixel size of 20 μm, corresponding to 0.38 nm at the specimen level. For image processing, the semper [34] and em [35] software packages were used. From digitized micrographs, smaller frames of 64 × 64 pixels containing individual particles were extracted interactively. These images were aligned translationally and rotationally, using standard correlation methods [36,37]. An arbitrarily chosen reference was used for the first cycle of alignment and averaging, and the resulting average was used as a reference in the second refinement cycle. For analysis of the rotational symmetry of top-on-view images, the individual images were aligned translationally but not rotationally [38]. These aligned images were subjected to multivariate statistical analysis [39]. The resulting eigenimages represent all-important structural features of the original dataset. If the images had different rotational symmetries in the original dataset, the eigenimages would reveal the different symmetry axes. Moreover, these images can be distinguished and subsequently separated on the basis on eigenimages. The rotationally aligned images were classified on the basis of eigenvector–eigenvalue data analysis, and subsequent averaging was performed for each class separately. The average was finally symmetrized on the basis of angular correlation coefficients [40].

Acknowledgements

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

B. Jia, S. Lee, B. P. Pham, R. Yu and T. L. Le were supported by scholarships from the Brain Korea21 project in 2008, Korea. This work was supported by a grant from the MOST/KOSEF to the Environmental Biotechnology National Core Research Center (grant no. R15-2003-012-01003-0), and the Korea Research Foundation Grant funded by the Korean Government (MOEHRD) (grant no. KRF-2007-521-C00241), to G. W. Cheong.

References

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

Fig. S1. Multiple sequence alignment of Thermococcus profundus NADH oxidase (NOXtp) with the homolog NADH oxidases or NADH peroxidase (NPX).

Fig. S2. Purification of Thermococcus profundus NADH oxidase (NOXtp) from Escherichia coli and Thermococcus profundus.

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
FEBS_6665_sm_figs S1-S2.pdf519KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.