2-Nitrobenzoate 2-Nitroreductase (NbaA) Switches Its Substrate Specificity from 2-Nitrobenzoic Acid to 2,4-Dinitrobenzoic Acid under Oxidizing Conditions

Abstract

2-Nitrobenzoate 2-nitroreductase (NbaA) of Pseudomonas fluorescens strain KU-7 is a unique enzyme, transforming 2-nitrobenzoic acid (2-NBA) and 2,4-dinitrobenzoic acid (2,4-DNBA) to the 2-hydroxylamine compounds. Sequence comparison reveals that NbaA contains a conserved cysteine residue at position 141 and two variable regions at amino acids 65 to 74 and 193 to 216. The truncated mutant Δ65-74 exhibited markedly reduced activity toward 2,4-DNBA, but its 2-NBA reduction activity was unaffected; however, both activities were abolished in the Δ193-216 mutant, suggesting that these regions are necessary for the catalysis and specificity of NbaA. NbaA showed different lag times for the reduction of 2-NBA and 2,4-DNBA with NADPH, and the reduction of 2,4-DNBA, but not 2-NBA, failed in the presence of 1 mM dithiothreitol or under anaerobic conditions, indicating oxidative modification of the enzyme for 2,4-DNBA. The enzyme was irreversibly inhibited by 5,5′-dithio-bis-(2-nitrobenzoic acid) and ZnCl₂, which bind to reactive thiol/thiolate groups, and was eventually inactivated during the formation of higher-order oligomers at high pH, high temperature, or in the presence of H₂O₂. SDS-PAGE and mass spectrometry revealed the formation of intermolecular disulfide bonds by involvement of the two cysteines at positions 141 and 194. Site-directed mutagenesis indicated that the cysteines at positions 39, 103, 141, and 194 played a role in changing the enzyme activity and specificity toward 2-NBA and 2,4-DNBA. This study suggests that oxidative modifications of NbaA are responsible for the differential specificity for the two substrates and further enzyme inactivation through the formation of disulfide bonds under oxidizing conditions.

INTRODUCTION

Nitroreduction, catalyzed by an NAD(P)H-dependent nitroreductase, is the essential first step in the catabolism of a variety of structurally diverse nitroaromatic compounds, such as nitrobenzoates, nitrotoluenes, and 3-nitrophenols. Biodegradation of 2-nitrobenzoic acid (2-NBA) by Pseudomonas fluorescens strain KU-7 is the first prokaryotic example of formation of a 3-hydroxyanthranilate intermediate (1, 2). Iwaki and colleagues (3) identified the nbaA and nbaB genes, which encode 2-nitroreductase (NbaA) and a mutase (NbaB) that mediate NADPH-dependent reduction of 2-NBA to 2-hydroxylaminobenzoic acid and a Bamberger-type rearrangement of 2-hydroxylaminobenzoic acid to 3-hydroxyanthranilate. In Arthrobacter protophormiae strain RKJ100, there is an alternative route of 2-NBA metabolism to form 3-hydroxyanthranilate and anthranilate (4, 5).

NbaA (GenBank accession number BAF56676.1) is a homodimeric NADH:flavin mononucleotide (FMN) oxidoreductase-like fold protein (3). It is similar to a putative flavin-containing protein (78% sequence identity; ABE46991.1) located on the Polaromonas sp. strain JS666 plasmid 1 (GI:91790731), and it includes a flavin reductase-like domain (Pfam accession number PF01613 in the Pfam database [http://www.sanger.ac.uk/Software/Pfam/]) (6). Structurally, it is related to the NADH:FMN oxidoreductase-like structural family (SCOP accession number b.45.1.2 or 50482; http://scop.berkeley.edu/) (7). Iwaki and colleagues (3) showed that Asn40, Asp76, and Glu113 in the conserved region of NbaA are necessary for binding to a divalent metal ion implicated in FMN binding, and that an insertion loop of 10 amino acids at positions 65 to 74 mediates NADPH binding. It exhibits a narrow specificity toward 2-NBA, which is different from other nitro/flavin reductases (8–10). However, the catalytic properties and substrate specificity of NbaA have not been characterized.

The aim of this study was to determine the catalytic properties and substrate specificity of NbaA. Recently, it was found that NbaA acts as a redox-sensitive protein which is able to form higher-order disulfide-bonded proteins under oxidizing conditions (11). We observed that the enzyme was able to reduce 2-NBA and 2,4-dinitrobenzoic acid (2,4-DNBA) at different lag times, and that those activities were gradually decreased during the formation of intermolecular disulfide bonds under aerobic conditions. The formation of intermolecular disulfide bonds was analyzed by SDS-PAGE and tandem mass spectrometry. In order to confirm the reduction of 2-NBA and 2,4-DNBA, the products derivatized with acetic anhydride were analyzed by thin-layer chromatography and high-performance liquid chromatography with tandem mass spectrometry. We performed deletion mutations in the variable regions (Δ65-74 and Δ193-216) to examine the role of those regions in the catalysis and specificity of NbaA. Site-directed mutagenesis was carried out to examine the possible roles of cysteines at positions 39, 103, 141, and 194 in oxidative modification of the enzyme.

MATERIALS AND METHODS

Materials and chemicals.

The following reagents were purchased from Sigma (St. Louis, MO): 5,5′-dithio-bis-(2-nitrobenzoic acid) (DTNB), dithiothreitol (DTT), 2,4- and 3,5-dinitrobenzoic acids (DNBA), EDTA disodium salt, N-ethylmaleimide (NEM), hydrogen peroxide (H₂O₂), isopropyl ß-d-thiogalactopyranoside (IPTG), ß-NAD 2′-phosphate, reduced form (NADPH), 2-, 3- and 4-nitrobenzoic acids (NBA), and riboflavin 5′-monophosphate (FMN) sodium salt dehydrate. Sequencing-grade porcine trypsin was obtained from Promega (Madison, WI). A fast protein liquid chromatography system, columns, and resins were supplied by GE Healthcare Life Sciences (Uppsala, Sweden). Protein mini-gel kits and assay reagents were supplied by Bio-Rad (Hercules, CA) and Pierce (Thermo Fisher Scientific Inc., Rockford, IL). All other solvents and reagents used were of analytical grade and were obtained from USB Corporation (Cleveland, OH), J. T. Baker (Phillipsburg, NJ), and Merck (Darmstadt, Germany).

Strains and culture conditions.

The complete DNA sequence of NbaA was PCR amplified and cloned into pSD80 as previously described (3). The truncated (Δ65-74 and Δ193-216) and site-specific mutants (C39A, C103A, C141A, C194A, and C141A C194A) were constructed using the PCR primers given in Table 1. The resulting plasmids were transformed into Escherichia coli strain BL21(DE3). Transformed E. coli strains were grown in LB broth at 37°C and 180 rpm in the presence of 100 mg liter⁻¹ ampicillin. When cells reached an optical density at 600 nm (OD₆₀₀) of ∼0.5, the recombinant proteins were induced with 0.2 mM IPTG for 1 h. An empty pSD80 plasmid-transformed strain was included under the same culture conditions as a control.

Table 1.

PCR primers used for construction of wild-type and mutant NbaA

PCR primer	Nucleotide sequence	Description	Reference or source
pSD80F	5′-GAGCTGTTGACAATTAAT-3′	pSD80 vector primers for DNA sequencing	45
pSD80R	5′-AGGACGGGTCACACGCGC-3′
NbaA-F	5′-CGGAATTCATGACGCACATTGCAATGTCA-3′	PCR amplification of a full DNA sequence containing the nbaA gene	3
NbaA-R	5′-AAAACTGCAGTCAGGGAGTAATCGGAAAGA-3′
NbaA-64R-blunt	5′-ATAATGATCCACGGCGAT-3′	Construction of Δ65-74 mutant by blunt ligation of PCR products	This study
NbaA-75F-blunt	5′-AAAGACACGCTAAAA AACATC-3′
NbaA-R192	5′-GTCGGTGCTGCAGCATTAGTTGGGTCCTCC-3′	Construction of Δ193-216 mutant	This study
NbaA-C39A-R	5′-AAGCGCTATAAGGCGCGGCATTCGCCAACCCCTCACTGTTCAAACTT-3′	Site-directed mutagenesis of cysteine at position 39 into alanine (C39A)	This study
NbaA-139F	5′-TGCCGCGCCTTATAGCGCTT-3′
NbaA-399R	5′- AAACCATTCGCTCAGCAATA-3′	Site-directed mutagenesis of cysteine at position 103 into alanine (C103A)	This study
NbaA-C103A-F	5′-TATTGCTGAGCGAATGGTTTTGGCGGGTAGTGATTTCCCCTCTCATAT-3′
NbaA-C141A-R	5′-AATCGATGATTTTGTAGAGTTTCGCTTCCCACGCAATGGGTGCGTCGG-3′	Site-directed mutagenesis of cysteine at position 141 into alanine (C141A)	This study
NbaA-445F	5′-ACTCTACAAAATCATCGATT-3′
NbaA-557R	5′-AGTTGGGTCCTCCCAAGCGCC-3′	Site-directed mutagenesis of cysteine at position 194 into alanine (C194A or C141A/C194A)	This study
NbaA-C194A-F	5′-GGCGCTTGGGAGGACCCAACTATGCGCGAACCACCGACCGGGTTCGCC-3′

Protein purification.

Cells were harvested at 4°C at 3,000 × g for 15 min using a Sorval centrifuge and rotor. Washed cells were suspended in 3 volumes of buffer A containing 10 mM DTT, 1 mM EDTA, and 50 mM Tris-HCl (pH 7.4) and disrupted by 3 passages through a prechilled French pressure cell (maximum capacity, 3.5 ml). After centrifugation at 12,000 × g for 30 min, the supernatant was applied to a DEAE-Sepharose column (1.6 by 20 cm) equilibrated with buffer A at a flow rate of 2 ml min⁻¹, and bound proteins were eluted by a 40-min linear gradient to 0.6 M NaCl plus buffer A with collection of 2-ml fractions. Aliquots of each fraction were mixed into 50 mM sodium phosphate buffer (pH 7.4) containing 1 mM 2-NBA, 1 mM NADPH, 10 μM FMN, and 0.1 mM MnCl₂ to determine the rate of NADPH oxidation at 340 nm (ε₃₄₀ = 6.21 mM⁻¹ cm⁻¹) using a Shimazu UV-1800 spectrophotometer (Shimazu Co., Kyoto, Japan) at room temperature. Fractions showing more than half-maximum activity were combined and treated with 1 M ammonium sulfate prior to loading on a Phenyl-Sepharose column (1.6 by 10 cm) equilibrated with 1 M ammonium sulfate in buffer A at 1 ml min⁻¹. Bound proteins were eluted in a 40-min linear gradient to buffer A with collection of 1-ml fractions. Fractions containing more than half-maximum activity were combined as described above and concentrated to approximately 500 μl using Centriplus YM-30 centrifugal filter devices (Millipore Co., Bedford, MA) before application to a Superdex 200 column (1.6 by 60 cm) equilibrated with buffer A at a flow rate of 0.25 ml min⁻¹. Protein concentration was determined using a Pierce Coomassie Plus protein assay kit and bovine serum albumin as a standard.

Enzyme reconstitution.

Purified NbaA (final concentration, 0.1 μM monomer) was mixed with various concentrations of divalent metal chloride salts (CaCl₂, MgCl₂, and MnCl₂) in 50 mM sodium phosphate buffer (pH 7.4) containing 1 mM NADPH, 10 μM FMN, 1 mM 2-NBA, or 2,4-DNBA, and NADPH oxidation was monitored for 30 min at room temperature. The lag time for enzyme activity to 2-NBA or 2,4-DNBA was estimated by a logistic function, and the metal-binding affinity (K_0.5M) was estimated by the Hill equation as mentioned below. The enzyme turnover number (k_cat, s⁻¹) was calculated by dividing the maximum velocity, v_max, by the 0.05 μM concentration of NbaA homodimer as the basic catalyst.

Enzyme inhibition.

To examine the presence of reactive thiol/thiolate in NbaA, 0.1 μM NbaA monomer was treated with various concentrations of a sulfhydryl-blocking agent (DTNB) or a sulfhydryl complexing agent (ZnCl₂) in 50 mM sodium phosphate buffer (pH 7.4) containing 10 μM FMN and 0.1 mM MnCl₂ for 10 min before enzyme activity was measured by addition of 1 mM NADPH and 1 mM 2-NBA.

Determination of substrate specificity and optimal conditions.

Substrate specificities of wild-type and mutant NbaA proteins were analyzed with various concentrations of 2-/3-/4-NBA and 2,4-/3,5-DNBA in 50 mM Tris-HCl (pH 7.4) containing 1 mM NADPH, 10 μM FMN, and 0.1 mM MnCl₂. The lag time for equilibration of enzyme activity with a substrate was estimated by a logistic function fit on the continuous NADPH oxidation (ΔA₃₄₀) curve, which was monitored every 2 s for 30 min. A midpoint rate of NADPH oxidation (v) by substrate concentration [S] was used for curve fitting with the Hill equation: v = (v_max × [S]^h)/(K_0.5S^h + [S]^h), in which v_max, K_0.5S, and h are the apparent maximum velocity, half-saturation constant, and Hill coefficient for substrate binding, respectively. Temperature and pH optima of NbaA reacting with 2-NBA were determined by varying the temperature in 50 mM sodium phosphate buffer at pH 7.4 or by varying the buffer pH at 25°C. In order to examine substrate-induced effects on the enzyme activity and differential specificity for 2-NBA and 2,4-DNBA, reconstituted NbaA at a final concentration of 0.1 μM monomer in 50 mM sodium phosphate buffer (pH 7.4) containing 0.1 mM MnCl₂ and 10 μM FMN was preincubated at ambient conditions for 30 min with 0.5 mM either 2-NBA or 2,4-DNBA or with 1 mM NADPH. The enzyme reaction was then started with the remaining substrate.

Analysis of metabolites.

Metabolites derived from the reduction of 2-NBA and 2,4-DNBA by NbaA were analyzed by thin-layer chromatography (TLC) and high-performance liquid chromatography with tandem mass spectrometry (HPLC-MS/MS) after following a modified procedure of derivatization of the unstable hydroxylamine groups with acetic anhydride (12). Reconstituted NbaA at a final concentration of 0.1 μM monomer was added to 10 ml of 50 mM sodium phosphate buffer (pH 7.4) containing 0.5 mM either of 2-NBA or 2,4-DNBA, 0.1 mM MnCl₂, and 10 μM FMN. After mixing with or without 1 mM NADPH, the reactions were carried out under aerobic or anaerobic conditions with monitoring NADPH oxidation for 50 min at room temperature. In order to remove dissolved oxygen, all buffers and solutions used were treated with sodium sulfite prior to mixing the enzyme and reactants. Aerobic or anaerobic conditions of the reaction mixtures were demonstrated by adding a redox indicator, leucomethylene blue for the so-called blue bottle reaction (13). The enzyme reaction was stopped by cooling in an ice bath, mixed with 2.5 ml of methanol, and acidified by slowly adding 7.5 ml of HCl. When acetic anhydride (7.5 ml) was added dropwise, the temperature was maintained at 0°C. After 30 min, the derivatization was stopped by the addition of sodium acetate (8.75 mg) dissolved in 15 ml of water. After repeated extraction with ethyl acetate, the combined organic layers were washed with water and dried over MgSO₄. The solvent was evaporated, and equal amounts (10 μl) of metabolites dissolved in 1 ml methanol were chromatographed on silica gel for separation with 2:2:1 (vol/vol/vol) hexane–t-butyl methyl ether-ethyl acetate. For HPLC-MS/MS analysis, a Thermo Accela UHPLC system was connected with a reverse-phase Hypersil Gold column (100 by 2.1 mm; particle size, 1.9 μm) to an LTQ-Velos Mass instrument. The column was equilibrated with 95% buffer A (0.1% formic acid in H₂O) plus 5% buffer B (0.1% formic acid in acetonitrile) at a flow rate of 150 μl min⁻¹. After injection of 10 μl sample prepared by 10-fold dilution of the concentrated metabolites in methanol, metabolites were eluted with a linear gradient of 5 to 80% buffer B over 30 min. The spray voltage was set with 5-kV negative polarity, and the temperature of the heated capillary was set to 275°C. Survey full-scan MS spectra (m/z 50 to 500) were acquired with 1 microscan and a resolution of 10,000 for determination of precursor ions and charge states, and MS/MS spectra of the 3 most intense ions from the survey scan were acquired with the following options: isolation width, ±1 Da; normalized collision energy, 35%; dynamic exclusion duration, 20 s. The mass spectral data were analyzed with Thermo Xcalibur software v. 2.1.

Analysis of NbaA oligomerization.

Oligomerization of NbaA was analyzed by size-exclusion chromatography at 0, 2, 6, and 12 h after incubation of PD10-desalted NbaA in 50 mM Tris-HCl buffer (pH 7.4) at an ambient temperature. Protein elution from a Superdex 200 column (1.6 by 60 cm) equilibrated with 50 mM Tris-HCl buffer (pH 7.4) at a constant flow rate of 0.25 ml min⁻¹ was monitored at 280 nm with collection of 0.5-ml fractions. In each fraction, the composition of oligomeric species of native protein was determined by native PAGE. The intensity of Coomassie-stained protein was calculated by local average volume using Molecular Dynamics ImageQuant, version 5.2 (GE Healthcare Life Sciences, Piscataway, NJ).

Determination of physicochemical factors for disulfide bond formation.

To examine physicochemical factors affecting disulfide bond formation, wild-type NbaA protein was incubated at a 1 or 10 μM concentration in various buffers (50 mM morpholineethanesulfonic acid-NaOH, pH 5.5 to 6.5; 50 mM HEPES-HCl, pH 6.5 to 8.1; 50 mM Tris-HCl, pH 7.8 to 9.7) containing DTT (0 to 10 mM), H₂O₂ (0 to 10 mM), or NaCl (0 to 1 M). During incubation at 25 or 37°C, subsamples were collected and treated with 10 mM NEM for 1 h in darkness. The compositions of the disulfide-bonded proteins were analyzed by nonreducing SDS-PAGE.

Mass spectrometry.

Disulfide-bonded isomer bands of NEM-treated proteins were excised from a gel and digested with trypsin for mass spectrometry (14). Dried peptide extracts were dissolved in 0.4% acetic acid and analyzed with a nanoflow LC-LTQ linear ion trap mass spectrometer (Thermo Fisher Scientific Inc.) with an Agilent Series 1200 nanoflow liquid chromatography system (Agilent, Santa Clara, CA) and a capillary column (75-μm inner diameter, 360-μm outer diameter, 15-cm length), which was packed in house with Magic C18AQ particles (5 μm; 200-Å pore size; Michrom Bioresources, Inc., Auburn, CA). The chromatographic conditions comprised a 45-min linear gradient from 5 to 40% acetonitrile (ACN) in 0.1% formic acid (FA), followed by a 5-min column wash in 80% CAN–0.1% FA and a 10-min column re-equilibration with 5% CAN–0.1% FA at a flow rate of 0.35 μl min⁻¹. The full mass scan was performed between m/z 300 and 2,000 and was followed by 5 data-dependent MS/MS scans with the following options: isolation width, ±1.5 m/z; normalized collision energy, 25%; dynamic exclusion duration, 30 s. Mass data were analyzed by an in-house Excel macro program with the options of average mass (m/z); precursor ion mass tolerance, 0.8 Da; fragment mass tolerance, 1 Da; and 2H loss (−2.02 Da) from all combinations of 2 tryptic peptides, each containing cysteine.

Western blotting.

Protein concentrations of wild-type and mutant NbaA proteins in whole-cell extracts were determined using a rabbit polyclonal anti-NbaA antibody and purified NbaA standard. Standard ECL reagents and films (GE Healthcare Life Sciences, Piscataway, NJ) were used for Western blot detection with horseradish peroxidase (HRP)-conjugated anti-rabbit IgG antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA).

RESULTS

Comparative sequence analysis of NbaA with flavin-containing reductase-like proteins.

In order to infer the enzymatic properties of NbaA, the amino acid sequence was reexamined by analyzing its phylogenetic relationship to the currently available GenBank protein database. The NbaA sequence (216 amino acids) is most similar to a putative flavin-containing reductase-like (FlaRed) protein (78% sequence identity; protein accession number ABE46991.1) retrieved from Polaromonas sp. strain JS666 plasmid 1 (GI:91700611:235953-236603). A phylogenetic tree of NbaA and homologous FlaRed proteins indicates its narrow occurrence among soil- and plant-associated species of alpha- and betaproteobacteria (Fig. 1A). Several strains, including Polaromonas sp. strain JS666, contain two or more homologous FlaRed protein-encoding genes (match scores of >100 and E values of <10⁻³⁰), and some bacteria carry FlaRed-coding plasmids. The sequence of NbaA shows 32% identity to the structurally characterized MTH152 (Protein Data Bank [PDB] accession number 1EJE) of Methanobacterium thermoautotrophicum strain ΔH (15). Sequence alignment shows a conserved cysteine (C141) and two variable regions, V1 and V2, corresponding to amino acids 65 to 74 and 193 to 216 in NbaA (Fig. 1B). In the NbaA polypeptide, there are 4 cysteines at positions 39, 103, 141, and 194.

Fig 1.

(A) Phylogenetic tree of Pseudomonas fluorescens KU-7 NbaA (GenBank accession number BAF56676.1; shown in the shaded box) and homologous flavin-containing reductase [FlaRed]-like proteins clustered by the unweighted-pair group method using average linkages (UPGMA). The scale bar indicates 0.1 amino acid changes per branch length. (B) Partial sequence alignment of NbaA (BAF56676.1:KU-7) and representative homologues highlights the conserved Cys141 site and the variable V1 (amino acids 65 to 74) and V2 (193 to 216) regions. Two or more copies of homologous genes are present in several bacterial species (GenBank accession numbers): Azorhizobium caulinodans ORS571 (BAF88487.1 and BAF88982.1), Agrobacterium radiobacter K84 (ACM26428.1 and ACM26650.1), Bradyrhizobium japonicum USDA 110 (BAC48596.1, BAC48663.1, and BAC47690.1), Bradyrhizobium sp. strain BTAi1 (ABQ35041.1 and ABQ34396.1), Bradyrhizobium sp. strain ORS278 (CAL76387.1 and CAL75771.1), Burkholderia phytofirmans PsJN (ACD19612.1 and ACD21190.1), Burkholderia xenovorans LB400 (ABE34598.1 and ABE33052.1), Polaromonas sp. strain JS666 (ABE46991.1 [plasmid 1], ABE43004.1, and ABE45686.1), Pseudomonas fluorescens Pf0-1 (ABA74886.1 and ABA76199.1), Rhizobium etli CFN 42 (ABC91672.1 and ABC90974.1), Rhizobium etli CIAT 652 (ACE92008.1 and ACE91236.1), Rhizobium leguminosarum bv. trifolii WSM1325 (ACS57175.1 and ACS56319.1), Rhizobium leguminosarum bv. trifolii WSM2304 (ACI55925.1 and ACI55127.1), and Rhizobium leguminosarum bv. viciae 3841 (CAK08853.1, CAK06476.1, and CAK08013.1). FlaRed-coding plasmids were Agrobacterium vitis S4 pAtS4e (ACM40198.1), Burkholderia phymatum STM815 pBPHY01 (ACC74970.1), Polaromonas sp. strain JS666 plasmid 1 (ABE46991.1), Sinorhizobium meliloti 1021 pSymA (AAK65563.1), and Sinorhizobium medicae WSM419 pSMED02 (ABR64490.1).

Substrate specificities of NbaA and its mutants deleted for the variable regions.

To examine potential roles of the variable regions (V1 and V2), wild-type NbaA and truncated mutant proteins Δ65-74 and Δ193-216 were overproduced in E. coli cells and purified up to nearly 95% homogeneity as seen by SDS-PAGE (Fig. 2A). In size-exclusion gel chromatography, the elution peaks of NbaA (24.4 kDa) and the truncated mutant proteins, Δ65-74 (23.3 kDa) and Δ193-216 (21.7 kDa), were detected in the molecular mass range of 40 to 50 kDa, which corresponds to their predicted homodimer mass values.

Fig 2.

(A) Overexpression and purification of wild-type NbaA and the truncated mutant proteins Δ65-74 and Δ193-216. Lane 1, whole-cell extracts; 2, DEAE-Sepharose; 3, Phenyl-Sepharose; 4, Superdex 200 gel filtration. M_r, Bio-Rad Precision Plus molecular weight protein standards, in thousands (Lot 161-0375). (B and C) pH (B) and temperature optima (C) of NbaA for 2-NBA. Rates of NADPH oxidation at 340 nm were measured with 50 mM sodium phosphate buffers containing 0.1 μM NbaA monomer, 1 mM NADPH, 1 mM 2-NBA, 0.1 mM MnCl₂, and 10 μM FMN. (D) NADPH oxidation rates of wild-type NbaA (circles) and truncated mutants Δ65-74 (triangles) and Δ193-216 (squares) for the reduction of 2-NBA (open symbols) and 2,4-DNBA (closed symbols) in 50 mM sodium phosphate buffer (pH 7.4) at 25°C. Curves were fitted by the Hill equation as described in Materials and Methods.

Figure 2B and C show the optimal pH and temperature for in vitro 2-NBA reduction activity. The relative activity (%) of NbaA was optimal at pH 7.4 and 50°C. However, its activity sharply declined above pH 7.4 and 50°C. In this study, we determined the enzyme properties and substrate specificities of wild-type and mutant NbaA in 50 mM sodium phosphate buffer, pH 7.4, at 25°C. Under this condition, NbaA was able to reduce 2-NBA and 2,4-DNBA but not 3-NBA, 4-NBA, or 3,5-DNBA. The Δ65-74 mutant had 2-NBA activity similar to that of wild-type NbaA; however, the mutant displayed approximately 20-fold lower 2,4-DNBA activity than wild-type NbaA, whereas the Δ193-216 mutant had approximately 9-fold lower 2-NBA activity and no 2,4-DNBA activity (Fig. 2D). These results indicate that the V1 region is necessary for 2,4-DNBA but not 2-NBA, while the V2 region affects the catalysis of both 2-NBA and 2,4-DNBA.

Effects of divalent cations on NbaA reduction of 2-NBA and 2,4-DNBA.

Purified NbaA without addition of a divalent cation showed a significant delay (∼6 min) for the 2-NBA activity and no 2,4-DNBA activity for 30 min (Fig. 3A and B). Treatments of purified NbaA with chloride salts of divalent cations, including CaCl₂, MgCl₂, and MnCl₂, markedly enhanced the enzyme activity for both substrates by decreasing the lag times. Ca²⁺ and Mn²⁺ ions produced similar activities for 2-NBA; however, the metal ion binding affinities and lag times of the enzyme were significantly different from each other (Table 2). The Mn²⁺ ion had a higher binding affinity (K_0.5M = 25 μM) than the Ca²⁺ ion (K_0.5M = 330 μM), while the Mn²⁺-bound enzyme displayed approximately 3-times-lower activity and a longer lag time for 2,4-DNBA reduction than the Ca²⁺-bound enzyme. In contrast, NbaA had similar affinities for binding Mg²⁺ and Ca²⁺ ions, but the Mg²⁺ ion conferred much lower enzyme activities and longer lag times for both substrates than Mn²⁺ and Ca²⁺ ions. Furthermore, the Mg²⁺-bound enzyme was rapidly decreased (inactivated) after equilibration. The binding affinities of each cation to purified NbaA reacting with 2-NBA and 2,4-DNBA did not significantly differ, indicating that the divalent cations did not affect substrate specificity of NbaA to 2-NBA and 2,4-DNBA.

Fig 3.

Continuous spectrophotometric assays for NADPH-dependent NbaA reactions for the reduction of 2-NBA (A) and 2,4-DNBA (B). Absorbance at 340 nm was recorded after 0.1 μM NbaA monomer was immediately mixed for ∼3 s into 50 mM sodium phosphate buffer (pH 7.4) containing 10 μM FMN, 1 mM NADPH, and 1 mM 2-NBA or 2,4-DNBA in the presence of a metal chloride: 1 mM CaCl₂, 1 mM MgCl₂, and 0.1 mM MnCl₂. Untreated (no addition) and MnCl₂-treated NbaA were used to show effects of divalent cations and a reducing agent (1 mM DTT) on NbaA activities. (C) First-order relationship of NADPH turnover rates and levels with lag times for equilibration of metal-bound NbaA with 2-NBA and 2,4-DNBA. (D) Logistic curves for determination of half-maximal inhibitory concentrations (IC₅₀) of a thiol-blocking agent, 5,5′-dithio-bis-(2-nitrobenzoic acid) (DTNB), and a thiolate-complexing agent, ZnCl₂. Inhibitors were treated with 0.1 μM NbaA monomer in 50 mM sodium phosphate buffer (pH 7.4) containing 10 μM FMN and 0.1 mM MnCl₂ for 10 min before the enzyme activity was measured with 1 mM NADPH and 1 mM 2-NBA.

Table 2.

Turnover rates and metal binding affinities of NbaA

Substrate	Parametera result for:
	2-NBA		2,4-DNBA
	kcat (s−1)	K0.5M (μM)	kcat (s−1)	K0.5M (μM)
No addition	12 ± 1	NA	ND	NA
CaCl2	138 ± 37	330 ± 45	5.3 ± 0.3	310 ± 55
MgCl2	25 ± 4	290 ± 43	1.1 ± 0.6	280 ± 58
MnCl2	145 ± 26	25 ± 6	1.7 ± 0.3	28 ± 6

^aParameters estimated by the Hill equation fitted to the plot of the NADPH oxidation rate by metal ion concentration with the addition of 0.1 μM NbaA monomer into 50 mM sodium phosphate buffer (pH 7.4) containing 1 mM NADPH, 10 μM FMN, and 1 mM 2-NBA or 2,4-DNBA at 25°C are shown as the means ± standard deviations (SD). NA, not available; ND, not detected after 30 min.

It is worth noting that the reduction of 2-NBA and 2,4-DNBA by NbaA does not occur simultaneously: the reduction of 2-NBA occurs earlier than that of 2,4-DNBA. Moreover, the reduction of 2,4-DNBA, but not 2-NBA, failed in the presence of 1 mM DTT. Figure 3C shows that NADPH turnover rates (k_cat) of NbaA for 2-NBA and 2,4-DNBA are directly proportional to the first-order loss of enzyme activity, which is correlated with the type of metal ions. When assayed with 2-NBA, the enzyme was inhibited by a thiol-blocking agent, DTNB, and a thiolate-complexing agent, ZnCl₂, with 50% inhibitory concentrations (IC₅₀s) of 19 and 330 μM (Fig. 3D). This inhibition was not eliminated after PD10 desalting, indicating the irreversible inhibition of reactive thiol/thiolate groups of cysteine residues. It was suspicious that any modifications of cysteine residues on NbaA might induce conformational changes of the enzyme, leading to alterations in catalysis and specificity.

Effects of oxygen on metabolism of 2-NBA and 2,4-DNBA by NbaA.

When NbaA was preincubated with 0.5 mM either of 2-NBA or 2,4-DNBA or with 1 mM NADPH under aerobic conditions, resulting in blue color after the addition of a redox indicator of leucomethylene blue, the enzyme activities for the two substrates, 2-NBA and 2,4-DNBA, were significantly decreased (Fig. 4A). After 30 min of preincubation of reconstituted NbaA with any substrate, the addition of the remaining substrates resulted in approximately 30% decreased NADPH oxidation for 2-NBA and about 75% decreased activity for 2,4-DNBA compared to results without preincubation. Under anaerobic conditions maintained by reducing dissolved oxygen with sodium sulfite, NbaA showed NADPH oxidation only for 2-NBA, not for 2,4-DNBA. These results indicated that NbaA was sensitive to oxygen, by which the enzyme activity for 2-NBA and 2,4-DNBA was modified and further inactivated over time during the incubation under aerobic conditions.

Fig 4.

(A) Continuous spectrophotometric assays for the reduction of 2-NBA and 2,4-DNBA before or after 30 min of preincubation of 0.1 μM NbaA monomer with 1 mM NADPH or 0.5 mM either of 2-NBA or 2,4-DNBA in 50 mM sodium phosphate buffer (pH 7.4) containing 10 μM FMN and 0.1 mM MnCl₂ under aerobic (oxic) or anaerobic (anoxic) conditions and treated with Na₂SO₃. After a 50-min reaction started by the addition of the remaining substrates, aerobic or anaerobic conditions were indicated by adding a redox indicator, leucomethylene blue for the blue bottle reaction (13), as shown in the middle of the graphs. (B) Silica gel thin-layer chromatography of 2-NBA and 2,4-DNBA metabolites shows the reduction of substrate levels and the resulting product bands under UV light. Metabolites resulting from NbaA reactions with 0.5 mM 2-NBA or 2,4-DNBA in the presence or absence of NADPH under aerobic or anaerobic conditions were derivatized with acetic anhydride, as described in Materials and Methods. (C) Base peak chromatograms of 2-NBA metabolites resulting from NbaA reactions with 0.5 mM 2-NBA in the presence (gray peaks) or absence (dark peaks) of 1 mM NADPH under aerobic conditions. The lower panels show tandem mass spectra of m/z 166.10 and m/z 178.22 ions assigned to the fragments of 2-NBA and N-acetylanthranilic acid (product I). (D) Base peak chromatograms of 2,4-DNBA metabolites resulting from NbaA reactions with 0.5 mM 2,4-DNBA in the presence (gray peaks) or absence (dark peaks) of 1 mM NADPH under aerobic conditions. The lower panels show tandem mass spectra of m/z 211.07 and m/z 239.11 ions assigned to the fragments of 2,4-DNBA and 2-acetoxyamino-4-nitrobenzoic acid (product II).

In order to examine the reduction of 2-NBA and 2,4-DNBA by analytical TLC and HPLC-MS/MS, it was necessary to derivatize unstable metabolites of 2-NBA and 2,4-DNBA with acetic anhydride. The derivatization of 2-NBA metabolites allowed us to detect a fluorescent product band on silica gel under UV light (Fig. 4B). This product was produced from the reduction of 2-NBA under anaerobic as well as aerobic conditions in the presence of NADPH but not without NADPH, suggesting that NbaA could catalyze the reduction of 2-NBA with the consumption of NADPH in anaerobic conditions. From TLC, 2,4-DNBA levels appeared to be decreased by the NADPH consumption of NbaA under aerobic conditions but not in anaerobic conditions. It was similar to results as measured by a spectrophotometer. However, the reduction of 2,4-DNBA with NADPH under aerobic conditions did not show any fluorescent or visible product band on silica gel under UV light. Thus, it was uncertain whether 2,4-DNBA actually had been reduced.

To address this question, a sensitive HPLC-MS/MS instrument was employed to identify metabolites resulting from the reduction of 2-NBA and 2,4-DNBA with the consumption of NADPH by NbaA under aerobic conditions. From base peak chromatograms, the peaks of 2-NBA substrate were detected at m/z 166.11 in the retention time (RT) range of 12.29 to 12.66 min (Fig. 4C). From the reaction with NADPH for 50 min, the 2-NBA level was decreased by about 81% compared to that without NADPH. A tandem mass spectrum of 2-NBA contained a major fragment ion peak at m/z 122.14 which was assigned to the fragment from collision-induced dissociation (CID) of carboxyl anion from the molecular ion. When 2-NBA was reduced, a single peak of product I was detected at m/z 178.22 (RT, 15.69 min). This product generated 3 major fragment ions at m/z 134.13 [M-COO⁻], 135.13 [M-CH₃CHO], and 136.01 [M-CH₃CO] by collision-induced fragmentation. Based on this, it was identified as N-acetylanthranilic acid (CAS no. 89-52-1), which is likely formed from the derivatization of 2-hydroxylaminobenzoic acid with acetic anhydride. The emission of a blue fluorescence on silica gel supports the identification of product I as N-acetylanthranilic acid. During the same reaction time, the 2,4-DNBA level detected at m/z 211.07 and 18.73 min was decreased with NADPH by a third of the control level without NADPH (Fig. 4D). The resulting product II was detected at m/z 239.11 and 25.36 min. A tandem mass spectrum of 2,4-DNBA had a major fragment ion peak at m/z 167.10, which was assigned to the loss of carboxyl anion from the molecular ion. From CID fragmentation, product II generated a major fragment ion at m/z 195.06, which was assigned to the loss of carboxyl anion from the molecular ion (m/z 239.11). The product II at m/z 239.11 was identified as 2-(acetoxyamino)-4-nitrobenzoic acid that was likely formed from the derivatization of 2-(hydroxylamino)-4-nitrobenzoic acid with acetic anhydride. The 2-hydroxylamine metabolites most likely can be produced from the 2-nitroreduction of 2-NBA and 2,4-DNBA by NbaA as a specific enzyme generally shows excellent chemoselectivity and regioselectivity in metabolism of substrate. If NbaA would reduce otherwise the 4-nitro group of 2,4-DNBA to produce 4-hydroxylamino-2-nitrobenzoic acid, the 4-hydroxylamine group could be derivatized to the 4-(N-acetoxyacetamido) form, showing the molecular ion at m/z 282.2, because the described derivatization procedure completely acetylates the 4-hydroxylamine group at both N and O centers (12). However, we detected no m/z 282.2 ion signal from the base peak chromatogram of 2,4-DNBA reduction. When derivatized with acetic anhydride, 2-hydroxylamino-4-nitrobenzoic acid appeared to preferentially undergo O-acetylation rather than N-acetylation, which formed N-acetylanthranilic acid (product I) from 2-hydroxylaminobenzoic acid by the same procedure. Although N-acetylation is generally favored, some hydroxylamine groups of nitroaromatic compounds are known to undergo O-acetylation (16). These chemical analyses show that NbaA specifically transforms 2-NBA and 2,4-DNBA to the 2-hydroxylamine compounds with the consumption of NADPH.

Protein oligomerization.

During incubation of NbaA at room temperature, 2-NBA- and 2,4-DNBA-reducing activities gradually decreased as NbaA homodimer oligomerized to form higher oligomers, as observed by size-exclusion chromatography and native PAGE analyses (Fig. 5A). In size-exclusion chromatography, both activities were overlapped in the elution fractions of the NbaA homodimer but not of higher oligomers. Decreasing levels of NbaA homodimer during incubation were positively correlated with decreased NADPH turnover rates of 2-NBA and 2,4-DNBA (Fig. 5B). When enzyme kinetics for 2-NBA and 2,4-DNBA were evaluated by the Hill equation, oligomerization of NbaA resulted in a decrease in the apparent maximum rates without significantly changing the half-saturation constants and Hill slopes for substrate binding (Fig. 5C), suggesting that the protein oligomerization led to a decrease in enzyme concentration.

Fig 5.

(A) Size-exclusion chromatography and native PAGE analyses of NbaA oligomers during incubation of NbaA at pH 7.4 and 25°C. Molecular weight (M_r; in thousands) markers for size-exclusion chromatography were thyroglobulin (670,000), bovine-γ-globulin (158,000), chicken ovalbumin (44,000), equine myoglobin (17,000), and vitamin B₁₂ (1,300). Native PAGE gels were analyzed with the column-loaded samples (denoted as load) and elution fractions (2 ml each) between 40 to 68 min. Molecular weight (M_r) markers for native PAGE were 175,000, 83,000, 62,000, and 48,000. Protein elution peaks and bands are denoted with arrowed lowercase letters: a, homodimer; b, homotetramer; c, homohexamer; and d, homo-octamer. (B) Composition of NbaA oligomers (stacked bars) and their NADPH oxidation rates (line-symbol curves) with 1 mM 2-NBA (circles) or 2,4-DNBA (triangles) measured in 50 mM sodium phosphate buffer (pH 7.4) containing 1 mM NADPH, 0.1 mM MnCl₂, and 10 μM FMN. (C) Measurements of NADPH oxidation rate by concentration of 2-NBA (open symbols on solid-line curves) or 2,4-DNBA (closed symbols on broken-line curves) after incubation for 0 (circles), 2 (triangles), 6 (squares), and 12 h (diamonds).

Intermolecular disulfide bond formation of NbaA.

NbaA formed higher-order disulfide-bonded proteins under ambient conditions. The disulfide bond formation was similarly observed at 1 and 10 μM NbaA monomer concentrations, indicating a zero-order reaction independent of protein concentration (Fig. 6A). The disulfide bond formation occurred more rapidly at 37 than 25°C. In addition, the intermolecular disulfide bond formation of NbaA was largely increased by alkaline pH and H₂O₂ treatments (Fig. 6B). Using various buffers with a pH range of 5.5 to 9.7, the pK_a value of NbaA participating in the formation of intermolecular disulfide bonds was estimated at approximately pH 7.4 (Fig. 6C). At this point, the formation of intermolecular disulfide bonds causing enzyme inactivation was drastically increased by approximately 7-fold from 3.4% h⁻¹ below pH 7.4 to 24% h⁻¹ above pH 7.4, as the pH was changed by one unit at a time. These data indicate that NbaA can form inactive higher proteins through the formation of intermolecular disulfide bonds under physiological conditions.

Fig 6.

Nonreducing SDS-PAGE of disulfide-bonded NbaA proteins. (A) Effects of protein concentration (1 or 10 μM NbaA monomer) and temperature (25 and 37°C) at pH 7.0 for 1 h. A control sample (Ctrl) treated with 10 mM DTT at 60°C showed the reduction of disulfide-bonded proteins to NbaA monomer. (B) Distribution patterns of NbaA monomer and higher disulfide-bonded proteins after 1 h of incubation with or without 10 mM H₂O₂ at pH 7 or 10 at 25°C. (C) Relative levels of NbaA monomer (A1) and higher disulfide-bonded proteins (A2, A3, and >A3) after 1 h of incubation at various pH conditions at 25°C. The pK_a value was determined from a 10% decrease from the maximum level of NbaA monomer by fitting to a logistic curve.

SDS-PAGE and mass spectrometric analyses of disulfide-bonded isomers.

SDS-PAGE under nonreducing conditions showed that NbaA formed three isoforms of each disulfide-bonded species (Fig. 7A). To analyze the disulfide bond positions by mass spectrometry, the three isoform bands of dimer-sized NbaA, considered the initial products of intermolecular disulfide bond formation, were excised from the gel. The trypsin-digested isoforms resulted in different patterns of extracted ion chromatograms for the 3 disulfide-bonded peptides derived from random cross-linking of two tryptic peptides, ¹³²ITDAPIAWEC*K¹⁴² and ¹⁸⁸LGGPNYC*R¹⁹⁵ (Fig. 7B). The identities of these peptides were confirmed by tandem mass spectrometry (Fig. 7C). None of the other disulfide-bonded peptides theoretically derived from random cross-linking by the other cysteines at positions 39 and 103 were detected. SDS-PAGE and mass spectrometry revealed that cysteines at positions 141 and 194 mediate the formation of 3 isoforms by random disulfide bonding.

Fig 7.

(A) Patterns of disulfide-bonded NbaA protein isomers in nonreducing SDS-PAGE denaturing conditions. M_r indicates Bio-Rad Precision Plus protein standards in thousands (Lot 161-0375). (B) Extracted ion chromatograms (XIC) of precursor ions of 3 different disulfide-bonded dipeptides at a charge state (z) of +2 or +3. (C) Tandem mass spectra of 3 disulfide-bonded dipeptides. The assigned b and y ions generated by collision-induced fragmentation are labeled on the one-letter amino acid sequences above the spectra, and the major fragment ions are shown in the tandem mass spectra.

Site-directed mutagenesis.

To characterize the functional roles of cysteines in NbaA, each cysteine (C39, C103, C141, and C194) was replaced with alanine by site-directed mutagenesis. Wild-type NbaA and the site-specific mutants (C39A, C103A, C141A, C194A, and C141A C194A) yielded different protein expression levels (Fig. 8A, upper). Thus, it was necessary to determine their molar concentrations in whole-cell extracts using the known concentration of purified NbaA by Western blotting with anti-NbaA antibody (Fig. 8A, lower). The crude NbaA activity for 2-NBA reduction (k_cat = 84 ± 1 s⁻¹) decreased by approximately 1.6-fold compared to that of purified NbaA (138 ± 6 s⁻¹), whereas the crude enzyme activity for 2,4-DNBA (k_cat = 47 ± <1 s⁻¹) increased by approximately 9-fold compared to that of purified NbaA (5.3 ± <0.1 s⁻¹). There was an inverse relationship between enzyme activity and lag time such that the lag time for 2,4-DNBA was markedly decreased from 396 to 75 s, whereas the lag time for 2-NBA increased from 20 to 29 s (Fig. 8B). As a negative control, whole-cell extracts obtained from IPTG-induced cells containing empty plasmid resulted in no detectible NADPH oxidation. It appeared that whole-cell extract contained certain factors which enhance the enzyme activity for 2,4-DNBA at the expense of its 2-NBA activity. Therefore, we used the crude enzymes to accurately estimate the kinetic parameters, since a large amount of active enzyme was inactivated over the prolonged lag time.

Fig 8.

(A) Expression patterns of wild-type NbaA and the cysteine-to-alanine mutants. The lower panel shows molar concentrations of the crude enzymes in whole-cell extracts determined using the known concentrations of purified NbaA by Western blotting (WB). (B) Continuous spectrophotometric assays for the crude NbaA (0.1 μM monomer concentration) with addition of 1 mM 2-NBA or 2,4-DNBA to 50 mM sodium phosphate buffer (pH 7.4) containing 1 mM NADPH, 10 μM FMN, and 0.1 mM MnCl₂ under aerobic conditions. The same protein amount extracted from empty plasmid-transformed cells was used to show the absence of NADPH oxidation with 2-NBA or 2,4-DNBA under the same conditions. (C and D) 2-NBA (C) and 2,4-DNBA (D) kinetic curves of wild-type NbaA and the cysteine-to-alanine mutants. Each 0.1 μM monomer of the crude enzymes was used in the reactions.

C39A and C103A resulted in similar enzyme activities for 2-NBA and 2,4-DNBA (Fig. 8C and D); they exhibited strong positive cooperativity for 2-NBA (h, >2.8) and decreased the enzyme activity for 2,4-DNBA (Table 3). These mutations were similar to the truncation of the V1 region, which markedly reduced the enzyme activity for 2,4-DNBA but not 2-NBA (Fig. 2D), while C141A and C194 showed different patterns for 2,4-DNBA and 2-NBA. C141A decreased the enzyme activity for 2-NBA but not for 2,4-DNBA. This mutation did not significantly change the enzyme affinity to 2-NBA but enhanced the cooperativity for 2,4-DNBA. It appeared to enhance the specificity toward 2,4-DNBA at the expense of the 2-NBA activity. In addition, C194A could reduce the 2-NBA activity to a certain degree without a significant change in the 2,4-DNBA activity. However, the double C141A C194A mutation diminished both 2-NBA and 2,4-DNBA activities. These effects of C141A and C194A on the enzyme activities were similar to the differential specificity of NbaA which occurred during disulfide bond formation under aerobic conditions. This indicated that oxidative modifications on either C141 or C194 can alter the catalysis and specificity of the enzyme, and that further modifications of both residues could lead to inactivation of the enzyme, as did the disulfide bond formation. Based on the present data, we propose that NbaA exists as a transient intermediate between the reduced and oxidized forms of the four cysteines, and this is responsible for modulating the catalysis and specificity.

Table 3.

Kinetic parameters of wild-type NbaA and cysteine-to-alanine mutant NbaA for 2-NBA and 2,4-DNBA

NbaA type	Parametera result for:
	2-NBA			2,4-DNBA
	vmax (μM min−1)	K0.5S (μM)	h	vmax (μM min−1)	K0.5S (μM)	h
Wild-type	252 ± 46	602 ± 133	1.75 ± 0.22	141 ± 1	644 ± 3	6.75 ± 0.19
C39A	252 ± 9	231 ± 12	2.87 ± 0.42	7.3 ± 3.4	624 ± 225	3.34 ± 2.34
C103A	252 ± 15	137 ± 16	2.81 ± 0.82	23.9 ± 5.4	493 ± 203	1.25 ± 0.27
C141A	149 ± 29	417 ± 94	2.19 ± 0.57	145 ± 2	435 ± 17	15.3 ± 6.9
C194A	184 ± 11	149 ± 22	5.85 ± 2.56	160 ± 8	568 ± 18	5.89 ± 1.00
C141A C194A	61 ± 1	265 ± 4	6.54 ± 0.5147	50.1 ± 4.6	565 ± 38	4.26 ± 0.91

^aKinetic parameters estimated by the Hill equation fitted to the kinetic data shown in Fig. 8C and D are expressed as the estimated means ± SD. v_max, apparent maximum rate; K_0.5S, half-saturation constant for substrate binding; h, Hill coefficient.

DISCUSSION

We purified and characterized 2-nitrobenzoate 2-nitroreductase (NbaA) from P. fluorescens strain KU-7; NbaA specifically transforms 2-NBA and 2,4-DNBA into the 2-hydroxylamine compounds. Phylogenetic analysis suggests that NbaA belongs to a new family of the flavin-containing reductase-like superfamily. Homologous genes are narrowly distributed in soil- and plant-associated agrobacteria, rhizobia, bradyrhizobia, and pseudomonads and often comprise two or more homologous genes in the genomic DNA and plasmids of some bacteria. The coexistence of multiple homologous genes is likely associated with horizontal gene transfer and might be related to niche-specific adaptations to different types of plants and soils (17). The specific properties of these gene products have not been characterized. In this study, we found that NbaA is able to reduce 2,4-DNBA as well as 2-NBA. 2,4-DNBA is formed by photolysis of the explosive 2,4,6-trinitrotoluene (18). Some bacteria were able to degrade 2,4-DNBA in a nitroaromatic pesticide-contaminated environment (19). However, no specific enzyme for 2,4-DNBA reduction has been identified. To our knowledge, this is the first report that NbaA is able to specifically reduce 2,4-DNBA as well as 2-NBA.

NbaA binds with a divalent cation and FMN cofactor. The FMN cofactor is required for catalysis; divalent cations, including Mn²⁺, Ca²⁺, and Mg²⁺, are not necessary for catalysis, rather they increase the reaction rate by reducing the lag time of the enzymatic activity. The divalent cation-induced effects on NbaA are similar to those of glutamine synthetases in E. coli and Mycobacterium smegmatis (20, 21). When the k_cat/K_0.5S values are measured for Mn²⁺-bound enzyme, the value for 2-NBA (110 μM⁻¹ min⁻¹) is much higher than that for 2,4-DNBA (0.44 μM⁻¹ min⁻¹). This is because the k_cat value for 2,4-DNBA decreases and K_0.5S increases after a prolonged lag time of 690 s, while the 2-NBA activity has a lag time of only 24 s. The kinetic behavior of NbaA for the 2 substrates could not be described by the substrate-induced fit model (22) and relevant conformational changes in E. coli argininosuccinate synthetase (23), taurine/alpha-ketoglutarate dioxygenase (24), membrane insertase YidC (25), and multidrug transporter EmrE (26). The large difference in lag times for equilibration of enzymatic activity toward 2-NBA and 2,4-DNBA and failure to reduce 2,4-DNBA in the presence of DTT or under anaerobic conditions suggest a redox-induced conformational change of the enzyme specific to 2,4-DNBA.

Cysteines are vulnerable to oxidation and can cross-link through disulfide bonds to other cysteines in close proximity, even under mild conditions (27). It is a rarely used amino acid that accounts for about 2% in eukaryotic proteins and about 1% in prokaryotic proteins (28). Locations of some cysteines in the catalytic sites are highly conserved (referred to as motifs) in iron-sulfur cluster-containing proteins (29), thioredoxin-like redox proteins (30), heme-binding proteins (31), and peroxiredoxin and flavoprotein reductase (32). Oxidation of such cysteines carries out biological functions in electron transfer, covalent sulfur bridge, and thiol-disulfide exchange reactions and triggers redox processes that control cell regulatory pathways and induction of a variety of proteins, including transcription factors, molecular chaperones, and protein tyrosine phosphatases (33). In addition, several molecular studies of noncatalytic cysteines have shed light on the functional role in protein stability, protein folding and refolding, regulation of reactive oxygen signal transduction, and substrate-binding affinity and activity of enzyme (34–38). The common mechanism underlying cysteine oxidation is exerted via conformational changes through the formation of disulfide or cyclic sulfenamide covalent bonds or sulfenic or sulfonic acids. For example, Bacillus subtilis ResA, a thioredoxin-like protein, is able to undergo a redox-induced conformational change between the reduced and oxidized states and uses alternative conformations to select the substrates (39). We indicate that alanine substitution of the four cysteine residues in NbaA at positions 39, 103, 141, and 194 alters the catalytic properties and substrate specificity of the enzyme for 2-NBA and 2,4-DNBA. This demonstrates that oxidative modifications on the cysteine residues of NbaA can induce the differential specificity for the two substrates, 2-NBA and 2,4-DNBA. The oxidation of either C39 or C103, located around the V1 region, may reduce the enzyme activity for 2,4-DNBA but not 2-NBA. On the contrary, the oxidation of either C141 or C194 in the V2 region may reduce the enzyme activity for 2-NBA but not 2,4-DNBA and may further modifications to form intermolecular disulfide bonds that inactivate the enzyme.

A prolonged lag time for 2,4-DNBA provides no catalytic advantage, since a large amount of the enzyme is inactivated through the formation of disulfide bonds. It is similar to treatment with a thiol-blocking agent (e.g., DTNB) or a thiolate-complexing agent (e.g., ZnCl₂). Size-exclusion chromatography and gel electrophoresis show that the NbaA homodimer is the basic catalyst, and that the enzyme is inactivated by the formation of higher oligomers via the formation of intermolecular disulfide bonds. Mass spectrometry proves the formation of three isoforms of disulfide-bonded proteins by random disulfide cross-linking between the two cysteines at positions 141 and 194. The random disulfide cross-linking of proteins with two cysteines is likely to increase the structural complexity of the oligomer population by the large asymmetry of isoform distribution. The rate of in vitro intermolecular disulfide bond formation of NbaA is markedly enhanced at ambient conditions above pH 7.4 or under oxidizing conditions. These factors are critical for regulating cytosolic protein by oxidative modifications under physiological conditions (40, 41). In cells, some redox proteins, such as the peroxiredoxin system and thioredoxin system, catalyze reversible oxidation and reduction reactions between cysteines and disulfide bonds in cytosolic protein (41–43). Actually, the intermolecular disulfide bond formation of NbaA was markedly increased by the exposure of E. coli cells to oxidative stress, while it was significantly reduced by a forced expression of thioredoxin A (11). Therefore, the catalytic activity and substrate specificity of oxygen-sensitive NbaA may be more effectively controlled with such redox systems in cells.

In conclusion, we presented here that NbaA is a specific enzyme for transformation of 2-NBA and 2,4-DNBA to the 2-hydroxylamine compounds. However, it does not react with 2-NBA and 2,4-DNBA simultaneously. The mutation studies on the two variable regions and cysteine residues support that the NbaA activities for the two substrates are modulated by oxidative modifications of the four noncatalytic cysteines, which may induce some conformational changes in the substrate-binding sites. The conformational change induced by reversible oxidative modification of noncatalytic cysteine is superficially effective to regulate substrate availability by allosteric regulation of both catalysis and specificity (37, 44). Further studies are needed to explore the redox properties of the cysteinyl residues and the mechanisms underlying the oxidative posttranslational modifications which lead to conformational changes in catalysis and specificity of the enzyme.