Reconstitution and Characterization of Aminopyrrolnitrin Oxygenase, a Rieske N-Oxygenase That Catalyzes Unusual Arylamine Oxidation*

Rieske oxygenases catalyze a wide variety of important oxidation reactions. Here we report the characterization of a novel Rieske N-oxygenase, aminopyrrolnitrin oxygenase (PrnD) that catalyzes the unusual oxidation of an arylamine to an arylnitro group. PrnD from Pseudomonas fluorescens Pf5 was functionally expressed in Escherichia coli, and the activity of the purified PrnD was reconstituted, which required in vitro assembly of the Rieske iron-sulfur cluster into the protein and the presence of NADPH, FMN, and an E. coli flavin reductase SsuE. Biochemical and bioinformatics studies indicated that the reconstituted PrnD contains a Rieske iron-sulfur cluster and a mononuclear iron center that are formed by residues Cys69, Cys88, His71, His91, Asp323, His186, and His191, respectively. The enzyme showed a limited range of substrate specificity and catalyzed the conversion of aminopyrrolnitrin into pyrrolnitrin with Km = 191 μm and kcat = 6.8 min–1. Isotope labeling experiments with 18O2 and H218O suggested that the oxygen atoms in the pyrrolnitrin product are derived exclusively from molecular oxygen. In addition, it was found that the oxygenation of the arylamine substrates catalyzed by PrnD occurs at the enzyme active site and does not involve free radical chain reactions. By analogy to known examples of arylamine oxidation, a catalytic mechanism for the bioconversion of amino pyrrolnitrin into pyrrolnitrin was proposed. Our results should facilitate further mechanistic and crystallographic studies of this arylamine oxygenase and may provide a new enzymatic route for the synthesis of aromatic nitro compounds from their corresponding aromatic amines.

leaky cell membranes and concomitant breakdown of biosynthetic activity followed by cessation of cell growth (7).
The cloning and characterization of a 5.8-kb DNA region that encodes the pyrrolnitrin biosynthetic pathway was reported (8). This DNA region confers the ability to produce pyrrolnitrin when expressed heterologously in Escherichia coli and contains four genes, prnABCD, each of which is required for pyrrolnitrin production (9). A hypothetical biochemical pathway for the synthesis of pyrrolnitrin has been proposed by van Pée et al. (10). In this pathway shown in Fig. 1, the first step is the chlorination of tryptophan by PrnA at the 7 position to form 7-chlorotryptophan, followed by rearrangement of the indole ring to a phenylpyrrole ring and decarboxylation by PrnB to form monodechloroaminopyrrolnitrin. This intermediate is chlorinated a second time by PrnC to form aminopyrrolnitrin, which, in the last step of the pathway, undergoes oxidation of the amino group in aminopyrrolnitrin to a nitro group by PrnD to form pyrrolnitrin.
The biosynthesis of pyrrolnitrin is one of the best examples of enzyme-catalyzed arylamine oxidation. Although arylamine oxygenases seem to be widely distributed within the microbial world and used in a variety of metabolic reactions (11)(12)(13)(14)(15)(16), PrnD represents one of only two known examples of arylamine oxygenases or N-oxygenases involved in arylnitro group formation, the other being AurF involved in aureothin biosynthesis (16,17). The nitro group plays an important role in their activity. Pyrrolnitrin analogues containing an amino group instead of a nitro group have no inhibiting effect on the growth of Neurospora crassa (18). Sequence analysis suggested that PrnD is a Rieske oxygenase consisting of a consensus Rieske [2Fe-2S] cluster-binding motif and a mononuclear non-heme Fe(II)-binding motif DXXHXXXXH (8). Rieske oxygenases are widespread in nature and catalyze a diverse set of oxidation reactions including cis-dihydroxylation, monohydroxylation, desaturation, sulfoxidation, and O-and N-dealkylation (19,20). However, because of the difficulties in protein expression and purification, arylamine oxygenases involved in arylnitro group formation including PrnD have never been characterized and assigned a defined function.
Here we report for the first time the functional expression, purification, reconstitution, and characterization of a novel Rieske N-oxygenase, PrnD, that catalyzes unusual arylamine oxidation. To explore the mechanism of its reaction further, we describe in vitro oxygenation of the precursor arylamine to an arylnitro product by purified and reconstituted PrnD, which has one Rieske-type [2Fe-2S] center and one mononuclear iron site/monomer. We also suggest the origin of oxygen in the nitro group and the putative mechanism for arylamine oxidation by PrnD.
Construction of the pMAL-c2x-prnD Expression Plasmid-The prnD gene was amplified by PCR from P. fluorescens Pf-5 genomic DNA using two oligonucleotide primers 5Ј-GGGGATCCATGAACGACATTCA-ATTGGATCAAG-3Ј (BamHI restriction site is underlined) and 5Ј-G-GAAGCTTTCACCGCTCACTTGCGACGCG-3Ј (HindIII restriction site is underlined). The PCR amplification was carried out using Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA) to minimize potential point mutations introduced by PCRs under standard conditions (21). The PCR program was 2 min at 96°C followed by 30 cycles of 1 min at 94°C, 1 min at 48°C, 2 min at 72°C, and a final elongation of 7 min at 72°C. The PCR products were cleaved by BamHI and HindIII and purified using QIAEX II gel purification kit (Qiagen). The purified product was cloned into the BamHI-and HindIII-digested expression vector pMal-c2x. The resulting pMal-c2x-prnD is under the control of the tac promoter and expresses PrnD as a fusion protein to the C terminus of the maltose-binding protein (MBP). 3 The cloned PrnD gene was confirmed to be free of point mutations by DNA sequencing at the Biotechnology Center of the University of Illinois using the Big Dye TM Terminator sequencing method and an ABI PRISM 3700 sequencer (Applied Biosystems, Foster City, CA).
Expression and Purification of the Fusion Protein-Overnight cultures of BL21 (DE3) cells transformed with the pMal-c2x-prnD vectors were diluted 1:200 into LB medium supplemented with ampicillin (100 g/ml) and grown at 30°C until absorbance at 600 nm (A 600 ) reached ϳ0.6. Then protein expression was induced by the addition of isopropyl-␤-D-1-thiogalactopyranoside to 0.1 mM final concentration, and the incubation was continued for 3 h. The cells were harvested by centrifugation at 4°C for 20 min at 6,000 ϫ g, rinsed with phosphate-buffered saline, and frozen and stored at Ϫ20°C. The yield was ϳ3 g of bacterial wet weight/liter of culture. A bacterial lysate was prepared by thawing and resuspending cells from a 1-liter culture in 40 ml of buffer containing 20 mM Tris-HCl, pH 7.5, 1 mM dithiothreitol, 100 mM NaCl, 1 mM phenylmethanesulfonyl fluoride, and 0.1 mM EDTA (starting buffer); this suspension was treated with 10 mg of lysozyme for 30 min. The suspension was sonicated on ice with a Branson sonicator for 1 min five times at 1-min intervals. After centrifugation for 20 min at 10,000 ϫ g, the supernatant was adjusted to 0.5 M NaCl and loaded onto an amylose resin column (3 mg of fusion protein/ml of amylose). The column was washed with 10 volumes of 0.5 M NaCl in starting buffer and with 20 volumes 20 mM Tris-HCl, pH 7.8, 100 mM NaCl, 1 mM dithiothreitol, before eluting with 10 mM maltose in the latter buffer. For cleavage with factor Xa, the dialyzed MBP-prnD fusion protein was incubated with factor Xa (1 g/200 g of fusion protein) for 16 h at 4°C. Maltose was removed by hydroxyapatite resin, and MBP was removed after cleavage by a second adsorption to the amylose resin. Fractions from the amylose column were then concentrated with Centricon-10 ultrafiltration units (Amicon), adjusted to 100 mM NaCl, 20 mM Tris, l mM dithiothreitol, pH 7.8.
Construction of the pTKXb-prnD Plasmid and Its Expression and Purification-The PCR product obtained by using the oligonucleotide primers 5Ј-GGCATATGAACGACATTCAATTGGATCAAG-3Ј (NdeI restriction site is underlined) and 5Ј-GGCTCGAGTCACCGC-TCACTTGCGACGCG-3Ј (XhoI restriction site is underlined) was cleaved by NdeI and XhoI and purified using QIAEX II gel purification kit (Qiagen). The purified product was cloned into the NdeI-and XhoIdigested expression vector p6xHTKXb119 (22) to give pTKXb-prnD, which was transformed into E. coli BL21 (DE3). In this expression plasmid, the prnD gene was placed under the transcriptional control of the promoter P BLMA from Bacillus licheniformis, which is constitutively induced in E. coli (23).
For the purification of His 6 -tagged TKXb-PrnD protein, 3 g of E. coli cells cultured for 24 h were resuspended in 10 ml of binding buffer (5 mM imidazole, 0.5 M NaCl, 20 mM Tris-HCl, pH 7.9) supplemented with 25 g/ml DNase I and disrupted by the same method as described above. After clarification by centrifugation for 30 min at 20,000 ϫ g and 4°C, the volume of the crude extract was adjusted to 10 ml with binding buffer. The crude extract was then loaded at a flow rate of 25 ml/h onto a 2.5-ml HisBind Resin column (Novagen, Madison, WI), which had been activated with NiSO 4 and equilibrated with binding buffer as described by the manufacturer. The sample was washed with 10 column volumes of binding buffer, followed by washing with 6 column volumes of wash buffer (60 mM imidazole, 0.5 M NaCl, 20 mM Tris-HCl, pH 7.9).
An additional wash step with a buffer containing 100 mM imidazole was performed for 6 column volumes prior to PrnD elution with 200 mM imidazole, 0.5 M NaCl, 20 mM Tris-HCl, 20% glycerol, pH 7.9. The protein solution was stored at Ϫ20°C until further use.
Site-directed Mutagenesis of pTKXb-PrnD-Site-directed mutagenesis was carried out by using a QuikChange site-directed mutagenesis kit from Stratagene. The pTKXb-prnD plasmid was used as the DNA template. Two conserved Cys residues (Cys 69 and Cys 88 ) in the iron-sulfur cluster and two Asp residues (Asp 323 and Asp 333 ) were mutated to Ala individually. The plasmids containing the correct mutant genes were then used to transform E. coli BL21(DE3), and colonies selected by kanamycin resistance were used for protein expression. The pTKXb-PrnD mutants were expressed according to the same procedure for the wild type enzyme as described above.
In Vitro Reconstitution of the Iron-Sulfur Clusters into the PrnD Proteins-The reconstitution procedures (24 -27) were carried out in an anaerobic chamber (Jacomex, Dagneux, France). All of the solutions were incubated anaerobically for 2 h before the beginning of each experiment. The purified PrnD was diluted to 10 -20 M with N 2 -sparged 20 mM Tris-HCl buffer (pH 7.8) containing 10 mM dithiothreitol, 200 mM NaCl, 0.5 mM EDTA, 1 mM phenylmethanesulfonyl fluoride, and 10% (v/v) glycerol. ␤-Mercaptoethanol was added to the protein solution at 1.0% (v/v), and the solution was gently mixed and left for 90 min. Fe(NH 4 ) 2 (SO 4 ) 2 and Na 2 S were added to the solution at a final concentration of ϳ1.0 mM. The reconstitution proceeded for 2-3 h. The solution was diluted 8-fold by the addition of N 2 -sparged 20 mM Tris-HCl buffer (pH 7.8), further diluted 2-fold by the addition of 20 mM Tris-HCl buffer (pH 7.8) equilibrated in air, and then dialyzed against the same buffer. Excess Fe 2ϩ and S 2Ϫ were removed by a desalting column (Bio-Rad; 10-DG). The reconstituted PrnD protein was concentrated on Centricon-30 ultrafiltration units (Amicon, Beverly, MA) to a final volume of 0.2 ml and stored at 4°C or for longer periods at Ϫ80°C. The iron-sulfur cluster incorporation was analyzed by EPR spectroscopy.
Construction of pQE-ssuE Expression Plasmid-For the production of SsuE reductase, the ssuE gene was placed under the control of the T5 RNA polymerase promoter using vector pQE-80L. The ssuE gene was amplified by PCR from E. coli XL1 Blue genomic DNA using the two oligonucleotide primers 5Ј-GGCATATGCGTGTCATCACCCTGG-C-3Ј (NdeI restriction site is underlined) and 5Ј-GGAAGCTTTTA-CGCATGGGCATTACCTC-3Ј (HindIII restriction site is underlined). The 672-base pair PCR product was digested with NdeI and HindIII, and the resulting fragment was ligated in pGEMT-easy to form plasmid pGEMTssuE. For the production of SsuE as an N-terminal His 6 -tagged fusion protein, the NdeI-HindIII SsuE-encoding fragment from plasmid pGEMTssuE was ligated into NdeI-HindIII-digested pQE-80L to form plasmid pQE-ssuE.
Production and Purification of SsuE Protein-E. coli BL21(DE3) containing the appropriate overexpression plasmid pQE-ssuE was grown at 30°C and 220 rpm in a 2-liter Erlenmeyer flask containing 400 ml of growth medium. To minimize the formation of insoluble protein aggregates, which were observed when protein production was carried out at 30°C, cultures grown to an A 600 of 0.5 were cooled to 16°C, induced by the addition of isopropyl-␤-D-1-thiogalactopyranoside to a final concentration of 100 mM and incubated for a further 5 h at 16°C with constant shaking (220 rpm). The cells were collected by centrifugation for 20 min at 6000 ϫ g and 4°C, washed in an excess of 20 mM Tris-HCl buffer, pH 7.9, and stored at Ϫ20°C as frozen pellets until further use. About 4 g of fresh weight of cells was collected from a 1000-ml culture. SsuE protein was purified as described under "Construction of the pTKXb-prnD Plasmid and Its Expression and Purification." Reconstitution of PrnD Activity-PrnD activity was routinely assayed with HPLC. The assay mixture (final volume, 0.5 ml) contained 100 M NADPH, 3 M FMN, 500 M substrate, SsuE and PrnD to a SsuE/PrnD molar ratio of 4.0 in 20 mM Tris-HCl, pH 7.8, and was stirred at 30°C. The reactions were started by the addition of PrnD to the reaction mixture and analyzed by HPLC. One unit of activity was defined as the amount of enzyme forming 1 mol of product/min at 30°C under standard assay conditions, calculated from the rate of substrate depletion.
Activity Assay of PrnD with Superoxide Dismutase, Xanthine Oxidase, and Hydroxyl Radical Scavengers-Superoxide dismutase was added to some of the incubation mixtures. After a 5-min preincubation at 30°C with shaking, the reaction was started by the addition of aminopyrrolnitrin (1 mM). The reaction was run for 30 min and was terminated by pouring the incubation mixture into 10 ml of ice-cold dichloromethane that had previously been bubbled with N 2 for 20 min and analyzed by HPLC. Incubation mixtures for the xanthine/xanthine oxidase system in 20 mM Tris-HCl buffer (pH 7.8) contained 75 M of xanthine, 5 units of xanthine oxidase, and 1 mM of aminopyrrolnitrin in a final volume of 300 l. Catalase, mannitol, or Me 2 SO were also included in some incubation mixtures. After a 5-min preincubation at 30°C, the reaction was started by the addition of substrate, allowed to proceed for 30 min, and stopped and analyzed as described above. The production of pyrrolnitrin was linear under these conditions for at least 30 min.
Protein Electrophoresis and Immunoblotting-SDS-PAGE and immunoblot analyses were performed largely as described by Sambrook et al. (21). SDS-PAGE was performed on a Mini-PROTEAN II system (Bio-Rad) with 12% polyacrylamide gels (30% acrylamide, 0.8% bis-acrylamide stock) under denaturing conditions. Protein concentrations were measured using the method of BCA (Sigma) according to the manufacturer's instructions with bovine serum albumin as a standard. Samples for SDS-PAGE were incubated, typically 5 min at 90 -100°C, with an equal volume of denaturing sample buffer. Broad range molecular weight standards were purchased from Bio-Rad. Proteins from gels were transferred electrophoretically onto a polyvinylidene difluoride membrane (Amersham Biosciences). Nonfat dry milk was used as the blocking agent. The His⅐Tag monoclonal primary antibody (Novagen, Madison, WI) and the secondary mouse IgG alkaline phosphatase conjugate (Sigma) were used at 1:10,000 and 1:20,000 dilutions, respectively. Cross-reacting protein bands were visualized with 5-bromo-4chloro-3-indolyl phosphate (0.17 mg/ml) and nitro blue tetrazolium (0.33 mg/ml) purchased from Sigma.
Enzyme Kinetics and pH Profiles of the Kinetic Parameters-Kinetic parameters determined in atmospheric oxygen were obtained by fitting the data to the Michaelis-Menten equation. To determine the kinetic constants for oxygen, the oxygen was removed from the enzyme reaction mixture by alternately filling and evacuating the vessel containing the reaction mixture with pure nitrogen. Different oxygen concentrations were bubbled through the substrate solutions for at least 10 min to obtain final oxygen concentrations of 0.03, 0.06, 0.12, and 0.6 mM. The concentrations of the arylamines varied from 20 M to 5 mM.
For assays at different pH values, the reactions were performed in the following buffers (50 mM) and pH values: sodium citrate (pH 5.0 -5.5), potassium phosphate (pH 6.0 -6.8), Tris/HCl (pH 7.0 -9.0), and glycine/NaOH (pH 9.5-10.0). To construct the pH profiles, the kinetic parameters k cat and k cat /K m for aminopyrrolnitrin were determined between pH 5.0 and 9.5, and the pH dependence of Y (k cat /K m ) was fitted to a bell-shaped curve described by Equation 1, which describes a bell-shaped curve with a slope of ϩ1 at low pH and a slope of Ϫ1 at high pH, where H is the proton concentration, K 1 and K 2 are the dissociation constants for the groups that ionize at low and high pH, respectively, and Y H is the pH-independent plateau value of Y at intermediate pH.
The pH profile for k cat was constructed in a point-to-point manner.
Analytical Methods-Optical spectra were recorded on a Varian Cary 100 Bio UV-visible spectrophotometer. 1 H NMR spectra (500 MHz) were measured with a DXR 500 AVANCE spectrometer from Bruker Instruments (Karlsruhe, Germany). EPR samples were prepared in an anaerobic chamber. The protein solution was transferred to an EPR tube, and after 5-10 min of incubation at room temperature, the samples were quickly frozen and stored in liquid nitrogen until EPR analyses. EPR measurements were performed on a Varian E-122 X-band spectrometer equipped with an Air Products Helitran cyrostat. Other parameters for the measurement include 2-milliwatt microwave power, 2-G modulation amplitude, and 9.08-GHz microwave frequency. The spectra were acquired at 15 K. Spin concentrations were determined by double integrating base line-corrected spectra. For calibration, a sample containing 1 mM CuSO 4 in a 20% glycerol solution was run at the same power and temperature used for the protein samples. Mononuclear iron content was determined as described previously (28). Potentiometric redox titrations monitored by EPR were preformed using sodium dithionite anaerobically in 20 mM buffer at pH 7. PrnD protein concentration was 56 M. The following redox mediators were used: methyl viologen, benzyl viologen, neutral red, anthraquinone-2-sulfonate, anthraquinone-1,5-disulfonate, 2-hydroxy-1,4-naphthoquinone, 2,5dihydroxy-p-benzoquinone, phenazine methosulfate, menadione, and duroquinone, each at a concentration 50 M.
Acid-labile sulfide was determined as described elsewhere (29). Enzyme reaction products were analyzed and purified by an Agilent 1100 Series HPLC System. The sample was eluted on a ZORBAX SB-C8 Column (4.6ϫ 150 mm, Agilent). HPLC parameters were as follows: 25°C; solvent A, 1% acetic acid in water; solvent B, methanol; gradient, 5% B for 2 min; then to 100% B in 18 min and finally maintain at 100% B for 2 min; flow rate 1.0 ml/min; detection was by UV spectroscopy at 254 nm.
Reactions in the Presence of 18 O 2 -Two vials, one containing aminopyrrolnitrin and one containing the holo-PrnD reaction mixture, were degassed by application of a vacuum and flushed with argon for three times. The anaerobic holo-PrnD solution was transferred to the substrate vial containing aminopyrrolnitrin. The argon was removed by application of a vacuum, and finally 18 O 2 was allowed to enter into the vial. After incubation for 1 h at 30°C, the reaction was analyzed by HPLC coupled to an electrospray ionization mass spectrometer (TSQ Quantum; ThermoFinnigan, San Jose, CA) in positive ion mode. A linear gradient of MeOH (0 -75%) in 0.1% aqueous acetic acid was used.

RESULTS
Cloning and Heterologous Expression of PrnD-One of the most critical limitations for biochemical and mechanistic studies of PrnD to date has been the lack of an efficient method to obtain large amounts of purified enzymes. Indeed, soluble expression of PrnD has never been reported. In the present study, the gene encoding PrnD was amplified from P. fluorescens Pf-5 genomic DNA using PCR and cloned into the expression vectors pTKXb119 and pMal-c2x to give the plasmids pTKXb-prnD and pMal-c2x-prnD, respectively. Among them, the constitutive expression of the plasmid pTKXb-prnD with the N-terminal His 6 tag in E. coli BL21(DE3) led to the soluble production of a protein of expected size ϳ41 kDa that cross-reacted with a His 6 tag-specific antibody but only at a low level requiring immunological detection ( Fig. 2A,  lane 4). The overall protein yield of pTKXb-PrnD (below 0.1 mg of purified PrnD obtained per liter of culture) was too low to be used for characterization of PrnD. Instead, a large amount of purified PrnD was obtained from the cleavage of MBP-PrnD fusion protein followed by affinity chromatography as described under "Experimental Procedures." Induction of the plasmid pMal-c2x-prnD with 0.1 mM isopropyl-␤-D-1-thiogalactopyranoside at 20°C led to the overproduction of soluble MBP-PrnD fusion protein of expected size ϳ85 kDa (Fig. 2B,  lane 4). Induction of empty plasmid pMAL-c2x produced ϳ40 kDa proteins of the size expected for MBP alone (Fig. 2B, lane 2). After purification, MBP was cleaved from PrnD by Factor Xa treatment and removed. This procedure resulted in apparently homogenous PrnD protein as judged by SDS-PAGE (Fig. 2B, lane 7). The molecular mass observed in SDS-PAGE corresponded to the calculated mass of the protein (41 kDa). A protein yield of ϳ7 mg of pure PrnD/liter of culture was obtained. By using gel filtration chromatography, the molecular mass of native PrnD was determined as 86 kDa, showing that the protein is homodimeric in solution.
Reconstitution of PrnD with Reductase and Rieske [2Fe-2S] Cluster-Rieske oxygenases are typically comprised of two protein components: a terminal oxygenase containing a Rieske iron-sulfur cluster and a non-heme iron active site and a reductase containing flavin (30). Thus, a reductase was deemed necessary for the function of PrnD. Because the flavin reductase was known to be nonspecific (31), the unknown reductase for PrnD from the Pseudomonas strain was substituted by flavin reductase SsuE from E. coli. Induction of the plasmid pQE-ssuE in E. coli BL21(DE3) led to soluble SsuE protein production of expected size ϳ25 kDa, which was purified by nickel-nitrilotriacetic acid affinity chromatography (Fig. 3). However, PrnD protein cleaved from MBP-PrnD did not show any activity in combination with SsuE reductase, NADPH, and FMN. In addition, we were unable to detect the characteristic g ϭ ϳ1.9 EPR signal corresponding to Rieske proteins directly from PrnD cleaved from MBP-PrnD overproduced in E. coli. We therefore pursued in vitro reconstitution of the Rieske [2Fe-2S] cluster with reduced iron and sulfide as described under "Experimental Procedures." This approach allowed assembly of the characteristic Rieske [2Fe-2S] cluster into the overproduced PrnD protein as demonstrated by the following spectroscopic analyses. In addition, the color of the reconstituted PrnD proteins was brown, indicating the presence of an ironsulfur cluster.
Spectral Characterization of a Rieske [2Fe-2S] Cluster in Reconstituted PrnD- Fig. 4 shows the optical spectra of reconstituted PrnD. Although the spectrum of the reconstituted and air-oxidized PrnD shows absorption maxima at 278, 339, and 440 nm and a distinct shoulder at 570 nm, the spectrum of the reconstituted and dithionite reduced PrnD shows absorption maxima at 405 and 521 nm, which are most similar to those of the [2Fe-2S] proteins with the Rieske-type [2Fe-2S] cluster (30,32) and distinctively different from those of the bacteriumtype ferredoxins or rubredoxins (33). The extinction coefficient of oxidized PrnD at 440 nm, ⑀([2Fe-2S]) of 6.5 mM Ϫ1 cm Ϫ1 (mean from three different preparations), is similar to that at 458 nm for Thermus thermophilus Rieske protein (6.0 mM Ϫ1 cm Ϫ1 ) (32). The optical spectrum was not affected by dialysis against an EDTA-containing buffer. PrnD is also readily reducible by sodium ascorbate in 5 min, suggesting that the redox center has a reduction potential well above 0 V, as is the case for the respiratory Rieske-type [2Fe-2S] proteins (32) and much higher than those of the [2Fe-2S] centers of the plant-type ferredoxins (30). Coordination of Rieske clusters by His imidazole nitrogen as well as Cys sulfur ligands is thought to account in part for the high potentials (Ϫ140 to ϩ320 mV) of Rieske [2Fe-2S] centers relative to Ϫ270 to Ϫ460 mV in plant, vertebrate, and bacterial [2Fe-2S] ferredoxins (34, 35). Fig. 5A shows the X-band EPR spectrum of reconstituted PrnD at 15 K. Although the air-oxidized form of PrnD as it is isolated is diamagnetic (EPR silent) and does not elicit any adventitious Fe 3ϩ signal at a g of 4.3, indicating the absence of any rubredoxin-like iron center, the protein reconstituted and fully reduced by excess dithionite elicits a rhombic S ϭ 1 ⁄ 2 resonance with principal g x,y,z values of 1.78, 1.89, and 2.03 (Fig. 5A), which could be readily detected at temperatures at least up to 70 K, as in the cases of some [2Fe-2S] clusters (32,36). In addition, its unusually low g av value (1.90) is comparable with that of the Riesketype [2Fe-2S] clusters (g av ϭ ϳ1.91) (34, 37). The EPR spectra of reduced PrnD were very similar to those of the Rieske-type [2Fe-2S] protein from T. thermophilus (30), phthalate oxygenase, and iron-sulfur-containing oxygenases isolated from Pseudomonads (38 -42), indicating that PrnD contains a Rieske-type [2Fe-2S] center. The redox potential was determined through titration of PrnD with different redox mediators. The titration was followed through EPR, and a redox midpoint potential of 95 mV was found (Fig. 5B).
The metal analysis of reconstituted PrnD by inductively coupled plasma atomic emission spectrometry showed the presence of tightly bound iron that was not removable by dialysis for 2 days against 20 mM Tris-HCl buffer (pH 7.8) containing 1 mM EDTA. The iron contents of PrnD as isolated and PrnD after dialysis against EDTA were determined to be 2.9 and 2.1 mol of iron/mol of monomer, respectively. Reconstituted PrnD also contained 1.9 mol of acid-labile sulfide/mol of monomer (means from three different preparations), presumably indicating an sulfur/iron ratio of 1 for the protein after dialysis against EDTA.   Thus, these data indicate that PrnD contains one [2Fe-2S] cluster and one mononuclear non-heme iron center/41-kDa monomer.
Requirement of the Mononuclear Iron Site-The EPR signals in the reduced wild type PrnD were quantified by double integration and compared with a copper sulfate standard. To quantitate the spin concentrations associated with the EPR signals, a sample of the reconstituted and reduced PrnD with a protein and iron concentrations of 46 and 133 M, respectively, was used. The spin concentration was 45 M corresponding to the [2Fe-2S] (S ϭ 1 ⁄ 2) signal. Whereas total iron concentration was decreased to 96 M after dialysis against EDTA, the spin concentration was not changed after dialysis. The spin concentration of the [2Fe-2S] center is approximately equal to that of the PrnD protein (Ϯ10%). These results reinforce the conclusion that the [2Fe-2S] is a bona fide component of the enzyme. The mononuclear iron occupancy (43 M) determined by subtracting the Rieske iron content (90 M, each S ϭ 1 ⁄ 2, spin ϭ 2.0 iron) from the total iron (133 M) was ϳ90% of the theoretical value.
Enzyme with about 3 atoms of iron/PrnD monomer had full PrnD activity when assayed in the absence of Fe 2ϩ . The removal of the mononuclear ion by dialysis against EDTA from PrnD led to the significant loss of activity (Ͼ90%), and the incubation with Fe 2ϩ caused nearly maximal reactivation of PrnD, indicating the necessity of the mononuclear iron for enzyme activity. Other samples with intermediate Fe/PrnD ratios (between 2 and 3) were prepared by incubating EDTAtreated PrnD with various levels of Fe 2ϩ . The specific activity of these samples was proportional to the iron content over the range from 2 to 3 total atoms of iron/PrnD monomer. Taken together, these data indicate the requirement of mononuclear iron site for PrnD activity.
Characterization of the Reconstituted PrnD and Its Reaction Product-The physiological substrate aminopyrrolnitrin was used to check whether the Rieske iron-sulfur cluster reconstituted PrnD was indeed able to catalyze an arylamine oxidation in combination with SsuE reductase in vitro. Although no activity was observed when aminopyrrolnitrin alone was used as a substrate, PrnD showed significant activity when SsuE, FMN, and NADPH were used for reconstitution, producing a compound that migrated on HPLC at the same retention time (15.94 min) as authentic pyrrolnitrin. To verify the identity of this compound produced by PrnD, it was purified by HPLC as mentioned under "Experimental Procedures," and this compound was unambiguously identified as pyrrolnitrin by 1 H NMR and high resolution electron ionization mass spectrometry. ( 1 H NMR (Acetone-d 6 , 500 MHz) ␦ (ppm) 6.91 (m, 1H), 7.02 (m, 1H), 7.65 (m, 4H); HR EI-MS, calculated for pyrrolnitrin (M ϩ ) 255.9806, found 255.9808.) 1 H NMR spectrum of the isolated product was identical to spectra previously reported for pyrrolnitrin (43). Taken together, these results clearly indicate that the product of the prnD gene is an arylamine oxygenase, aminopyrrolnitrin oxygenase.
Enzyme Kinetics and pH Profiles of the Kinetic Parameters-An increase in PrnD activity with increasing SsuE concentration was observed with saturation after the SsuE/PrnD molar ratio reaches ϳ3.0.
Thus, at a SsuE/PrnD molar ratio of 4.0, initial velocity studies were performed with aminopyrrolnitrin as a variable substrate in the presence of fixed concentration of PrnD (ϳ1 M). Values of K m ϭ 191 M for aminopyrrolnitrin and a k cat ϭ 6.8 min Ϫ1 (specific activity ϭ 0.08 unit/mg protein) were determined for PrnD. This heterologously expressed and reconstituted PrnD showed a comparable activity to the native PrnD enzyme in P. fluorescens Pf5 (specific activity ϭ 0.10 unit/mg protein), indicating that mononuclear ion atoms are present in high occupancy.
To give further insight into the mechanism of PrnD, the variation of kinetic parameters with pH was studied. The pH profiles of the kinetic parameters using aminopyrrolnitrin as a substrate are shown in Fig. 6. The effect of pH on the ionization of free P. fluorescens PrnD was visualized by the plot of log k cat /K m (M Ϫ1 s Ϫ1 ) versus pH. The pH profile of log k cat /K m for aminopyrrolnitrin oxygenation fits Equation 1, which describes a bell-shaped curve with slopes equal to 1, is bell-shaped, and is dependent on two ionizing groups, indicating that a group with a pK of ϳ6.8 on the acidic side of the curve has to be deprotonated and a group with a pK of ϳ8.9 on the basic side of the curve has to be protonated for activity. The fit indicated a correlation coefficient R 2 value of 0.986. In the log k cat versus pH plot, log k cat increased steadily up to pH 8.0 and then decreased with a slope less than 0.3.
Substrate Specificity-In air-saturated buffer at 30°C, the concentration of oxygen in solution is 247 M (43), i.e. ϳ10-fold the K O2 values (27 M or below) determined with arylamines (TABLE ONE), ensuring that most enzymes are saturated with oxygen. Consequently, both the apparent k cat and k cat /K m values determined at atmospheric oxygen with arylamines approximate well the values that would be measured by FIGURE 6. pH dependence of steady state kinetic parameters, log k cat (•), and log k cat /K m (E) of PrnD with aminopyrrolnitrin as substrate. PrnD activity was measured at 30°C in the presence of NADPH, FMN, and SsuE, as described under "Experimental Procedures."

TABLE ONE
Substrate specificity of aminopyrrolnitrin oxygenase (PrnD) Enxyme activity was measured as described under "Experimental Procedures" at varying concentrations of organic substrates in air-saturated 20 mM Tris-CCl (pH 7.8) at 30°C. Other substrate analogues mentioned under "Experimental Procedures" did not show any significant conversion into the corresponding nitro products. K m is the Michaelis constant for the organic substrate. K m under O 2 is the Michaelis constant for oxygen. varying the concentration of both organic substrate and oxygen. As expected based on this analysis, the k cat and k cat /K m determined with the arylamines at atmospheric oxygen were similar to those determined by varying the concentration of oxygen (data not shown). PrnD revealed a high substrate specificity toward physiological substrate aminopyrrolnitrin, p-aminobenzyl amine, p-aminobenzyl alcohol, and p-aminophenyl alanine. 4-Aminobenzamide was slightly converted, and none of 15 other compounds tested was converted. The more extensive kinetic analyses are consistent with a narrow range of substrate specificity, as indicated by k cat /K m that reflects the relative selectivity of the enzyme for different substrates (TABLE ONE). Interestingly, the specificity of the enzyme for arylamines seems to be independent of the size of the substrate, because the k cat /K m value for aminopyrrolnitrin was similar to those for p-aminobenzyl alcohol and p-aminophenyl alanine and 2-fold higher than that for p-aminobenzyl amine. Furthermore, substrate size does not significantly affect the overall enzymatic rate of turnover, as shown by the k cat values seen within the similar structure of arylamines and the different structure of arylamines. Effect of Superoxide Dismutase, Catalase, and Xanthine Oxidase on PrnD Activity-The binding of dioxygen to iron at the active site can produce superoxide by attracting one electron from the active site iron and can produce a peroxide by attracting one electron from the active site iron and one from the Rieske center. Such species could attack substrate bound in the substrate pocket and form a covalent bond to the substrate. Thus, the oxidation of an arylamine to an arylnitro group can be catalyzed via a nonenzymatic free radical chain reaction that is initiated and propagated by superoxide. If this were the case, one would expect the rate of oxygenation to decrease significantly when the enzymatic activity of PrnD is measured in the presence of superoxide dismutase, because the superoxide would be converted to oxygen and hydrogen peroxide. Consequently, the effect of superoxide dismutase on the reaction catalyzed by PrnD with substrates was determined. With both substrates tested, aminopyrrolnitrin and p-aminobenzyl amine, the rate of oxygenation did not change when superoxide dismutase was present in the reaction mixture at a concentration range of 10 -100 g/ml. The catalase and xanthine/xanthine oxidase systems, which generates both superoxide and hydrogen peroxide, were used to further investigate the involvement of superoxide or hydrogen peroxide in the oxygenation. Inhibition of pyrrolnitrin formation by catalase and xanthine oxidase was variable but was always less than 10% up to 100 g/ml.

Substrate
Hydroxyl radicals produced either via a Haber-Weiss reaction or a Fenton-type reaction could also participate in the oxidation process. To test this possibility, incubations of aminopyrrolnitrin with the PrnD system were performed in the presence of mannitol and Me 2 SO, which have been found to be effective as hydroxyl radical scavengers (44). At the concentrations used in the experiments (0 -100 mM), neither mannitol nor Me 2 SO affected PrnD activity. Thus, it appears unlikely that the oxidation of aminopyrrolnitrin to pyrrolnitrin is mediated by hydroxyl radicals. Taken together, these results are consistent with the oxidation of the arylamine substrates catalyzed by PrnD occurring at the enzyme active site and not involving a free radical chain reaction.
Investigation of the Origin of Oxygen with 18 O 2 and H 2 18 O-To confirm whether indeed molecular oxygen was the substrate of the PrnD reaction and whether one or both oxygen atoms of O 2 were incorporated into the product, isotope labeling experiments with 18 O 2 were carried out. Incorporation of the label was analyzed by liquid chromatography coupled to electrospray ionization mass spectrometry analysis. Reconstituted PrnD was incubated with aminopyrrolnitrin under an atmosphere of 18 O 2 . A control incubation was carried out under the usual 16 O 2 atmosphere. The reaction product pyrrolnitrin was separated by HPLC and analyzed by electrospray ionization mass spectrometry. Fig. 7 shows the molecular ions obtained in a usual 16 O 2 atmosphere (Fig. 7A) and in 18 O 2 atmosphere (Fig. 7B). Unlabeled pyrrolnitrin showed the molecular ion at [MϩH] ϩ ϭ 257 under the 16 O 2 atmosphere. In contrast, most of the pyrrolnitrin produced under the 18 O 2 atmosphere showed the molecular ion at [MϩH] ϩ ϭ 261, indicating the incorporation of two 18 O atoms. Exchange of 18 O 2 with water was not observed, although it has been reported to take place with some other non-heme iron-containing oxygenases (45). PrnD was also incubated with aminopyrrolnitrin in labeled water H 2 18 O. A control incubation was carried out in the usual buffer. In contrast to 18 O 2 incorporation, pyrrolnitrin produced in H 2 18 O showed the molecular ion at [MϩH] ϩ ϭ 257. Hence, the oxygen atoms in the pyrrolnitrin product are derived exclusively from molecular oxygen O 2 , not from H 2 O.
Overall Reaction in PrnD-The consensus Rieske sequence motif CXHX 15-17 CXXH in PrnD identified by sequence alignment (see supplemental material) of over 80 genes sharing homology to PrnD suggests that the two atoms of iron are coordinated by Cys 69 and Cys 88 and His 71 and His 91 , respectively (Fig. 8A). Site-directed mutagenesis confirmed the indispensable role of Cys 69 and Cys 88 in catalysis because both C69A and C88A PrnD mutants lost their activity (below 0.1%), and the [2Fe-2S] EPR signal is completely absent from the spectra of two cysteine mutants. These results unambiguously confirm the requirement of two cysteine residues for iron sulfur cluster formation in PrnD. The other highly conserved motif in PrnD is DXXHXXXXH involved in the formation of a mononuclear non-heme Fe(II) catalytic site that is believed to be the site of dioxygen activation and substrate oxygenation in Rieske oxygenases (19,46). Sequence alignment between PrnD and the Rieske dioxygenase with a crystal structure, naphthalene dioxygenase ␣-subunit suggests that (i) His 186 , His 191 , and Asp 323 are involved in the formation of the recurring 2-His-1-carboxylate facial triad structural motif conserved in all known non-heme Fe(II) oxygenases as shown in Fig. 8A (19), and (ii) Asp 183 , the equivalent residue of D205 in naphthalene dioxygenase ␣-subunit (47), is involved in electron transfer from the Rieske cluster to the non-heme iron center during catalysis. The indispensable role of these two aspartate residues in catalysis was further confirmed by site-directed mutagenesis. Indeed, upon site-directed mutagenesis of Asp 323 and Asp 183 into alanine, their activity was significantly decreased for both D323A (ϳ3%) and D183A (below 0.1%) PrnD mutants. The spectra of the reduced two mutants (D323A and D183A) still exhibit the g x ϭ 1.78, g y ϭ 1.89, and g z ϭ 2.03 features, showing that the [2Fe-2S] center is present in two mutants. With D323A the amplitude of the [2Fe-2S] signal is essentially the same as in the wild type enzyme. With D183A the amplitude is slightly smaller, but g value resonances are still clearly discernible. The overall sequential enzyme reaction and electron transfer for PrnD was illustrated in Fig. 8B.

DISCUSSION
The multi-component oxygenases including Rieske oxygenase seem to be widely distributed within the bacterial world and used in a large variety of biosynthetic and metabolic reactions (19). However, whereas the flavin reductase component of these systems has been the subject of detailed mechanistic and structural studies, very little is known on the oxygenase component, where the oxidation takes place. Especially in the case of arylamine oxygenase including PrnD, there is no information regarding enzyme characteristics and structure. The involvement of PrnD in the biosynthesis of pyrrolnitrin was suggested in vivo by a gene inactivation experiment (48). However, because of the general difficulties in expressing and purifying PrnD enzyme, coupled with the lack of commercial substrates required for biochemical studies, so far PrnD has never been characterized and assigned to a defined function by in vitro oxygenation. The present results now allow the formulation of a detailed hypothesis for the formation of pyrrolnitrin.
Although PrnD has been suggested to contain [2Fe-2S] clusters, no experimental evidence has been available. In this study, we have succeeded in overexpressing the soluble forms of the PrnD protein in E. coli by co-expressing with MBP, which is known to help proteins of interest undergo proper folding in vivo. Incubation of the PrnD cleaved from MBP-PrnD fusion protein with reduced iron and sulfide under anaerobic conditions allowed highly efficient reconstitution (2.9 atoms of iron/ mol of PrnD subunit, more than 95% of the expected amounts) of the redox-active Rieske-type [2Fe-2S] center in vitro as demonstrated by the g x ϭ 1.78, g y ϭ 1.89, and g z ϭ 2.03 EPR spectrum (g av ϭ (g x ϩ g y ϩ g z )/3 Ϸ 1.90) detected at 15 K (Fig. 5). This spectrum is typical of the highly characteristic, rhombic EPR spectra displayed by Rieske [2Fe-2S] clusters (49,50). Plant-type [2Fe-2S] ferredoxins such as those found in cyanobacteria (25) and the ferredoxin centers of some bacterial dioxygenases (35) also display rhombic EPR spectra, but with g values at about g x ϭ 1.89, g y ϭ 1.96, and g z ϭ 2.05 (g av Ϸ 1.96). In addition, the redox midpoint potential of 95 mV (at pH 7) for the g av ϭ 1.91 [2Fe-2S] cluster from PrnD lies well in the range of potentials obtained so far for Rieske proteins (51,52).
During the reconstitution, proteins were allowed to refold in the presence of reduced iron and sulfide and finally exposed to aerobic buffer in the absence of reducing agents as suggested by Cheng et al. (26). Exposure to oxidizing conditions may stabilize the Rieske [2Fe-2S] cluster by allowing formation of the disulfide predicted by the previous studies (53)(54)(55). Although other proteins containing the Rieske [2Fe-2S] center such as the Nostoc Rieske protein (56) were reconstituted, this is the first report of reconstitution and characterization of a novel Rieske N-oxygenase PrnD catalyzing unusual arylamine oxygenation.
Substrate specificity of PrnD enzyme seems to be independent of the size of substrate, based on the effect of the substrate size on the k cat /K m and k cat values with several substrates. This independence of the size of substrate were also found in other oxygenases (57)(58)(59), suggesting that the position of the functional group may be more important than the size of the substrate. Thus, with PrnD, it can be proposed that substrate binding occurs at a hydrophobic site large enough to accommodate arylamine substrates with up to at least two aromatic rings, such as aminopyrrolnitrin.
The chloroperoxidase from P. pyrrocinia was thought to catalyze the oxidation of arylamine based on the finding that this enzyme catalyzed the in vitro oxidation of the amino group of aminopyrrolnitrin to the nitro group of pyrrolnitrin, strongly suggesting the involvement of this enzyme in pyrrolnitrin biosynthesis (60). However, definite proof that no haloperoxidase type of enzyme was involved in the biosynthesis of pyrrolnitrin was obtained by a gene disruption experiment. Even after disruption of the haloperoxidase gene in a P. fluorescens strain, this strain still produced pyrrolnitrin (48). Instead, this oxidation is more likely to be catalyzed by a Rieske oxygenase, as suggested by the homology of PrnD with these enzymes (8). Based on the results of the present studies, now it is clear that this oxidation is catalyzed by PrnD, a Rieske N-oxygenase as demonstrated by in vitro oxygenation of aminopyrrolnitrin into pyrrolnitrin.
In contrast to the halogenation reaction catalyzed by PrnA or PrnC, the catalytic mechanism of PrnD is still unknown. However, by analogy to the mechanisms for N-oxidation of arylamines catalyzed by cytochrome P-450 (15), chloroperoxidases (61), and chemical catalysts (62), a putative mechanism for bioconversion of aminopyrrolnitrin to pyrrolnitrin by PrnD may be proposed. The incorporation of the oxygen requires the activation by iron (63,64). Activated oxygen species or iron peroxo or hydroperoxo intermediates (28,65) at the non-heme catalytic site can attack nitrogen and following concerted oxygen transfer from the high valent iron-oxo species suggested in cytochrome P-450 (63,66), phthalate dioxygenase (63), putidamonoxin (63), naphthalene dioxygenase (28), and methane monooxygenase (64,67) to the nitrogen may produce aromatic nitro-metabolite pyrrolnitrin.
In summary, in the present study the gene coding for PrnD from P. fluorescens Pf5 was cloned and heterologously expressed in E. coli. The resulting enzyme that catalyzes unusual arylamine oxidation was purified, reconstituted, and found to be a homodimer containing 1 mol of Rieske [2Fe-2S] center and 1 mol of mononuclear iron site/mol of subunit. With arylamines such as aminopyrrolnitrin and p-aminobenzyl amine, substrate oxidation occurs at the enzyme active site. A steady state kinetic analysis showed that the preferred substrate for the enzyme is physiological aminopyrrolnitrin and that the enzyme has narrow substrate specificity. From a mechanistic standpoint, PrnD operates through a dioxygenase-like catalytic mechanism, in which molecular dioxygen is incorporated into the substrate based on the results of isotope labeling experiments with 18 O 2 . To our knowledge, this represents the first account in which arylnitro compounds are formed from arylamine substrate in vitro catalyzed by a Rieske N-oxygenase. Additionally, a single gene encoding an arylamine oxygenase has never been expressed in heterologous hosts and has never been characterized. The availability of large amounts of recombinant active PrnD will be instrumental for detailed mechanistic and structural studies aimed at a better understanding of the chemical mechanism of oxidation of arylamine catalyzed by PrnD. Our results should improve understanding of arylamine oxidation in biological processes. In addition, PrnD may be used to synthesize a variety of aromatic nitro compounds from the corresponding aromatic amines, thus adding a new powerful tool into the toolbox of industrial biocatalysis.