INTRODUCTION

Xanthobacter strain Py2 is one of several bacteria capable of aerobic growth using aliphatic alkenes (e.g. ethylene, propylene, and butylene) as sources of carbon and energy (1, 2). In these alkene-metabolizing bacteria, alkenes are oxidized to epoxides in reactions catalyzed by alkene monooxygenases as illustrated in Equation 1. In Xanthobacter strain Py2, epoxides are further metabolized by a CO₂-dependent carboxylation reaction, catalyzed by an epoxide carboxylase, that forms -keto acids as products (3, 4).

A variety of bacterial monooxygenases catalyze the incorporation of an oxygen atom, derived from molecular oxygen, into aliphatic hydrocarbon substrates. These include methane monooxygenase, which hydroxylates methane to methanol (5), propane monooxygenase, which hydroxylates propane to isopropyl alcohol (6), and alkane hydroxylase, which catalyzes the terminal hydroxylation of longer chain length (C₆ to C₁₂) n-alkanes (7). Interestingly, these enzymes have broad substrate ranges and can fortuitously catalyze the epoxidation of aliphatic alkenes in an identical manner to the reaction shown for alkene monooxygenase in Equation 1 (8-12). This common feature of reactivity toward alkenes suggests that the general class of hydrocarbon monooxygenases may share some similar biochemical and biophysical properties. Notably, however, alkene monooxygenases have a more restricted substrate specificity and are incapable of catalyzing the hydroxylation of saturated hydrocarbon substrates (1, 13, 14).

Recent in vitro studies of alkene monooxygenases from three bacteria, Nocardia corallina strain B-276 (14), Mycobacterium strain E3 (15, 16), and Mycobacterium sp. strain M156 (17) have revealed that these enzymes function as multiprotein complexes. This multicomponent nature is a feature shared by the aforementioned bacterial hydrocarbon monooxygenases that hydroxylate substrate molecules. The alkene monooxygenase complex from Nocardia corallina B-276 was recently purified and found to consist of three components as follows: a monomeric (40-kDa) NADH reductase containing FAD and a [2Fe-2S] cluster, a dimeric (53- and 35-kDa subunits) oxygenase containing two nonheme iron atoms, and a monomeric (14-kDa) coupling protein that contained no detectable cofactors (14). The alkene monooxygenase systems from Mycobacterium strain E3 and Mycobacterium sp. strain M156 were each resolved into two fractions using anion exchange chromatography that could be recombined with restoration of activity (16, 17). Of these systems, only the NADH reductase component from Mycobacterium strain E3 has been purified. This reductase was a monomeric protein (56 kDa) containing FAD and a [2Fe-2S] cluster (15).

In the present study, the purification and biochemical characterization of the alkene monooxygenase from Xanthobacter strain Py2 is reported. This particular alkene monooxygenase is of interest for several reasons. In whole cell suspensions, it has a broad substrate range for aliphatic alkenes, catalyzing the epoxidation of terminal and internal alkenes varying in chain length from ethylene to hexene (18). In addition, alkene monooxygenase will initiate the oxidation of a number of chlorinated alkenes, including trichloroethylene, cis- and trans-1,2-dichloroethylene, vinyl chloride, 1-chloropropylene, 1,3-dichloropropylene, and 2,3-dichloropropylene (13). These reactive compounds represent a class of environmental pollutants that impose potential health hazards to humans and for this reason there is a fundamental need to understand biological mechanisms for their metabolism. Alkene monooxygenase of Xanthobacter strain Py2 is also an inducible enzyme that is repressed during growth with conventional carbon sources (e.g. glucose) and induced upon exposure of cells to a range of aliphatic and chlorinated alkenes and epoxides (19). Finally, Xanthobacter Py2 is a Gram-negative bacterium that is genetically tractable (20), a feature that should facilitate the characterization of the genes encoding alkene monooxygenase and related proteins.

MATERIALS AND METHODS

Xanthobacter strain Py2 was grown in 15-liter semicontinuous cultures in a Microferm fermentor (New Brunswick Scientific) with propylene as the carbon source as described previously (4). Cells were harvested at an A₆₀₀ between 2.5 and 4.0 by tangential-flow filtration with a Pellicon system (Millipore Corp.) and stored at 80 °C. Frozen cell paste (100-200 g) was thawed and resuspended in 2 volumes of buffer (100 mM MOPS¹ (pH 7.2), 1 mM dithiothreitol, 5% (v/v) glycerol) containing 10 mg of deoxyribonuclease I. The cell suspension was passed three times through a French pressure cell at 110,000 kilopascal, and the lysate was clarified by centrifugation at 139,000 × g for 1 h at 4 °C. Immediately after centrifugation the glycerol concentration of cell-free extracts was increased to 15% (v/v).

The clarified cell-free extract was applied to a DEAE-Sepharose Fast Flow column (5.0 × 25 cm) equilibrated in 50 mM MOPS (pH 7.2), 1 mM dithiothreitol, and 15% (v/v) glycerol (buffer A) at a linear flow rate of 24 cm/h. After loading, the column was washed with 1300 ml of buffer A. The reductase component did not bind under these conditions and was collected in the flow-through and wash fractions. The column was further developed with a 3-liter linear gradient of 0 to 500 mM NaCl in buffer A, resulting in the elution of the remaining three alkene monooxygenase components. The oxygenase and small protein were localized in the fractions eluting between 200 and 250 mM NaCl, and the ferredoxin was localized in the fractions eluting between 370 and 410 mM NaCl.

Ammonium sulfate was added to the fractions containing the oxygenase and small protein to a final concentration of 1.5 M. After incubating at 4 °C for 30 min the sample was centrifuged at 139,000 × g for 40 min at 4 °C. The supernatant was applied to a Pharmacia HiLoad 26/15 phenyl-Sepharose column equilibrated in buffer A containing 1.5 M (NH₄)₂SO₄ at a linear flow rate of 60 cm/h. The column was then washed with 200 ml of buffer A containing 1.5 M (NH₄)₂SO₄. Under these conditions the small protein did not bind to the phenyl-Sepharose column and was recovered in the flow-through and wash fractions. The column was equilibrated with 200 ml of buffer A containing 300 mM (NH₄)₂SO₄ and then developed further with a 300-ml reverse linear gradient of buffer A containing 300 to 0 mM (NH₄)₂SO₄. After the completion of the gradient the column was washed with buffer A alone which eluted the oxygenase component. Fractions containing the oxygenase component were pooled and concentrated by ultrafiltration (YM30) to a volume of approximately 4 ml. The sample was then applied to a Pharmacia HiPrep 26/100 Sephacryl S-300 column equilibrated in buffer A containing 200 mM NaCl at a linear flow rate of 5.7 cm/h. Fractions containing the oxygenase component were pooled, concentrated by ultrafiltration, and frozen in liquid nitrogen.

The flow-through and wash fractions that contained the small protein from phenyl-Sepharose were dialyzed against two 4-liter changes of 50 mM MOPS (pH 7.2) containing 1 mM dithiothreitol and 10% (v/v) glycerol (buffer B). To concentrate the small protein, the dialyzed sample was diluted 2-fold with buffer B and applied to a Pharmacia HiLoad 26/10 Q-Sepharose column equilibrated in buffer B at a linear flow rate of 45 cm/h. After washing the column with 100 ml of buffer B, the flow direction was reversed, and the small protein was eluted with buffer B containing 700 mM NaCl. Active fractions containing the small protein were pooled and further concentrated using ultrafiltration (YM3) to a volume of approximately 4 ml. The protein was then applied to a Pharmacia HiPrep 26/70 Sephacryl S-100 column equilibrated in buffer B plus 200 mM NaCl at a linear flow rate of 8 cm/h. Active fractions were pooled, concentrated by ultrafiltration, and frozen in liquid nitrogen.

Fractions containing the ferredoxin from the DEAE-Sepharose column were diluted with 3 volumes of buffer B and applied to a Pharmacia HiLoad 26/10 Q-Sepharose column equilibrated in buffer B. The column was washed with 500 ml of buffer B containing 200 mM NaCl. The ferredoxin was eluted with a 500-ml linear gradient of 200-450 mM NaCl in buffer B at a linear flow rate of 17 cm/h. Fractions containing ferredoxin activity were pooled and concentrated by ultrafiltration (YM10) to a volume of approximately 4 ml. The protein was then applied to a Pharmacia HiPrep 26/60 Sephacryl S-100 column, equilibrated in 50 mM MOPS (pH 7.2) containing 5% (v/v) glycerol and 200 mM NaCl, at a linear flow rate of 8 cm/h. Fractions containing ferredoxin activity were collected, concentrated by ultrafiltration, and frozen in liquid nitrogen.

The flow-through and wash fractions from DEAE-Sepharose that contained the reductase were brought to 20% (v/v) glycerol and 5 mM dithiothreitol. The sample was then applied to a column of Reactive Green 19 (2.5 × 10 cm), equilibrated in 50 mM MOPS (pH 7.2) containing 5 mM dithiothreitol and 20% (v/v) glycerol (buffer C) at a linear flow rate of 25 cm/h. The column was washed with 200 ml of buffer C. Reductase was eluted with a 400-ml linear gradient of 0 to 500 mM NaCl in buffer C. Fractions containing reductase activity were pooled and dialyzed against 4 liters of 50 mM MES buffer (pH 6.2) containing 3 mM dithiothreitol and 20% (v/v) glycerol (buffer D). The reductase was then applied to a column of SP-Sepharose Fast Flow (2.5 × 10 cm) equilibrated in buffer D, at a linear flow rate of 20 cm/h. After washing the column with 200 ml of buffer D, the column was developed with a 210-ml linear gradient of 0-150 mM NaCl in buffer D. Fractions containing reductase activity were pooled and concentrated by ultrafiltration (YM30) to a volume of 4 ml. The sample was then applied to a Pharmacia HiPrep 26/60 Sephacryl S-100 column equilibrated in 50 mM MOPS (pH 7.2) containing 3 mM dithiothreitol, 20% (v/v) glycerol, and 200 mM NaCl at a linear flow rate of 5.7 cm/h. Active fractions were pooled, concentrated by ultrafiltration, and frozen in liquid nitrogen.

Incubations were performed in sealed serum vials (9-ml) containing buffer (50 mM MOPS (pH 7.2)), NADH (5 mM), propyne (3 ml), and cell-free extract (27 mg of protein) in a total volume of 1 ml. Inactivation of the alkene monooxygenase was optimized by incubating the vials with shaking for 45 min at 30 °C. The vials were then repeatedly evacuated and flushed with air on a vacuum manifold. Residual propyne and small molecules were removed by passing the extract over a Sephadex G-25 (1.5 × 5 cm) column equilibrated in MOPS buffer (50 mM (pH 7.2)).

Alkene monooxygenase activity was monitored by following the time course of propylene depletion in sealed assay vials (9-ml) containing a source of enzyme (cell-free extract, column fractions, or purified components), NADH (5 mM), propylene (2200 nmol), and buffer (50 mM MOPS (pH 7.2)). All assays were performed in a shaking water bath at 30 °C in a total volume of 1 ml. Assays were initiated by the addition of NADH after a 2-min preincubation with all other components. Alkene monooxygenase activity was monitored by periodically removing samples from the head space of the reaction vials and quantifying the amount of propylene remaining using gas chromatography as described previously (21).

NADH:ferredoxin oxidoreductase activity was measured spectrophotometrically at 25 °C in a Shimadzu model UV160U spectrophotometer by monitoring the time-dependent decrease in absorbance at 460 nm resulting from the reduction of the iron-sulfur cluster of the purified ferredoxin (₄₆₀ = 7,230 M¹ cm¹). Assays were performed in 2-ml (1-cm path length) anaerobic quartz cuvettes that had been modified by fusing a serum bottle-style quartz top (7 × 13 mm at mouth), allowing the cuvettes to be sealed with a red rubber serum vial stopper. Stoppered cuvettes containing buffer (50 mM MOPS (pH 7.2)), NADH (5 mM), and oxidized ferredoxin (0.59 mg) in a total volume of 1 ml were made anoxic by repeated cycles of evacuation and flushing with argon on a vacuum manifold. At the desired times during the course of assays, a small amount of reductase, oxygenase, small protein, or cell-free extract was added to cuvettes with a microsyringe.

All UV/visible spectra were obtained at 25 °C using a Shimadzu model UV160U spectrophotometer. EPR spectra were obtained using a Bruker model ESP 300 spectrometer equipped with an Air Products LTR3 liquid helium cryostat. A standard solution containing CuSO₄ (1 mM) and EDTA (10 mM) was used for spin quantitation of the metal centers in protein samples. Circular dichroism spectra were recorded at 25 °C using an AVIV model 62DS spectropolarimeter with a 2-nm spectral bandwidth and a scan step every 1 nm. Background spectra were recorded for each cuvette with buffer alone and were subtracted from sample data. Dithionite-reduced samples of purified ferredoxin were prepared by the addition of sodium dithionite from a stock solution (100 mM, prepared anaerobically in 50 mM MOPS (pH 7.2)) to samples made oxygen-free by repeated cycles of evacuation and flushing with argon on a vacuum manifold. The final concentration of dithionite in samples was 1 mM.

Electrochemical redox titrations were performed in a stirred cell in an argon-filled glove box as described by Ryle et al. (22). The cell contained purified ferredoxin (0.33 mg/ml) in 100 mM MOPS buffer (pH 7.0) that contained 250 mM NaCl and the following redox mediators, each present at a concentration of 0.025 mM: 2-anthroquinonsulfonic acid, indigo disulfonate, methylene blue, toluidine blue, phenazine methosulfate, and 2,6-dichlorophenolindophenol. Samples poised at defined potentials were removed and frozen in calibrated EPR tubes in liquid nitrogen. EPR spectra were recorded at 12 K and 1.01 milliwatt microwave power. The fraction of reduced ferredoxin was determined by measuring the peak-to-peak height for the g = 1.92 region of the EPR spectrum. The reduction potential is reported with respect to the normal hydrogen electrode.

Native molecular weights were estimated by gel filtration using a Pharmacia Superose 12 HR 10/30 column equilibrated in 50 mM MOPS (pH 7.2) containing 200 mM NaCl and 0.05% CHAPS. The column was calibrated using -amylase, alcohol dehydrogenase, bovine serum albumin, ovalbumin, carbonic anhydrase, and cytochrome c. Polypeptide molecular weights were also determined using a matrix-assisted laser desorption ionization-time of flight mass spectrometer (Micromars Inc.) operating in the linear mode at an acceleration voltage of 20 kV. Bovine serum albumin and myoglobin were used as the internal standards. Spectra were obtained for protein samples that had been dialyzed in deionized and distilled H₂O. Apparent molecular weights of purified polypeptides using denaturing electrophoresis were determined by comparison with the R_f values for molecular weight standard proteins. The standards were myosin (200 kDa), -galactosidase (116.2 kDa), phosphorylase b (97.4 kDa), bovine serum albumin (66.2 kDa), ovalbumin (45 kDa), carbonic anhydrase (31 kDa), soybean trypsin inhibitor (21.5 kDa), lysozyme (14.4 kDa), and aprotinin (6.5 kDa). Quantitative amino acid analysis was performed by the Protein/Nucleic Acid Shared Facility at the Medical College of Wisconsin, Milwaukee, WI. Multi-elemental metal analysis was performed on an inductively coupled plasma atomic emission spectrophotometer at the Utah State University Soil and Plant Analysis Laboratory. Total iron content was determined colorimetrically by complexation with Ferene S (3-2-pyridyl-5,6-bis(2-(5-furylsulfonic acid))-1,2,4-triazine) as described by Haigler and Gibson (23).

SDS-PAGE (14% total gel; 2.7% cross-linker running gel) was performed in a Mini-Protean II apparatus (Bio-Rad) following the Laemmli procedure (24). Electrophoresed proteins were visualized by Coomassie Blue staining. The flavin cofactor was extracted from purified reductase by the addition of 0.5% (v/v) trichloroacetic acid followed by boiling (5 min). The flavin was identified by high pressure liquid chromatography as described previously (25). A molar absorptivity of ₄₅₆ = 11.3 mM¹ was used for quantitation of FAD content. Protein concentrations were determined using either a modified biuret assay with bovine serum albumin as the standard (26) or extinction coefficients that were calculated based on the results obtained from quantitative amino acid analysis of protein samples.

CHAPS detergent, NADH, and Ferene S were purchased from Sigma. Propylene (99% minimum) was purchased from Matheson gas products. Propyne (>84%) was obtained from Columbia Organic Chemicals Co.

RESULTS

Consistent with the results of a previous study (1), cell-free extracts of Xanthobacter strain Py2 prepared from propylene-grown cells oxidized propylene to propylene oxide in an O₂- and NADH-dependent fashion. Alkene monooxygenase activity was quantitatively recovered in the soluble fraction after sedimentation of membranes by high speed centrifugation, indicating that the enzyme is a soluble protein. The addition of membranes to the soluble fraction did not increase or decrease alkene monooxygenase activity.

The soluble fractions of cell-free extracts prepared from different batches of cells had highly variable specific activities, ranging from 5 to 55 nmol of propylene oxidized/min/mg. In contrast, the specific activities of alkene monooxygenase in the whole cell suspensions used to prepare cell-free extracts was consistently in the range of 90 to 110 nmol/min/mg. Cell-free extracts from different batches of cells were prepared identically, so it is presently unclear why the in vitro activities are so variable. The addition of cysteine and ferrous iron, which stimulate activity of some other bacterial oxygenase systems (27, 28), had no effect on activity when added to the lysis buffer or to the buffer used to assay alkene monooxygenase. Likewise, lysis under anaerobic conditions, the addition of other divalent metal ions (Zn²⁺, Cu²⁺, and Mg²⁺), and the addition of protease inhibitors (phenylmethylsulfonyl fluoride, leupeptin, or aprotinin) had no effect on in vitro activity. The addition of glycerol had a stabilizing effect on alkene monooxygenase activity and was included for the purification of each component. Most preparations of cell-free extracts had activities of 15-25 nmol/min/mg; for consistency, the purification procedures described below made use of cell-free extracts with specific activities in this range.

Alkene monooxygenase was resolved into four components obligately required for reconstitution of activity. These four components were individually purified on the basis of their ability to complement the other three components in reconstituting alkene monooxygenase activity. Activity assays were performed under conditions where three components were present at saturating concentrations so that specific rates could be reported relative to the concentration of the fourth component, which was limiting in the assay. The biochemical properties of each component are described separately below.

Active fractions eluting off the initial DEAE-Sepharose column at salt concentrations between 370 and 410 mM NaCl were brown-red in color, suggesting that this component may be a ferredoxin and play a role in electron transfer. A summary of the three-step protocol used for the purification of the ferredoxin is presented in Table I. As shown in Fig. 1, the purified protein ran on SDS-PAGE as two closely migrating bands with apparent molecular masses of 19.5 and 18.7 kDa. The Coomassie Blue staining intensities of the two bands are not equal; the higher molecular weight band is present in an approximately 15-fold excess based on staining intensity. The protein was chromatographed over several additional columns (i.e. S-200 and Superose 12 gel filtration, Reactive Green affinity) with no further increase in specific activity nor change in the relative intensities of the two bands. When subjected to electrophoresis on a nondenaturing gel, the ferredoxin migrated as a single band. This band could be cut from the nondenaturing gel and then electrophoresed under denaturing conditions, revealing the same two bands (data not shown). The inclusion of protease inhibitors (phenylmethylsulfonyl fluoride, leupeptin, EDTA, and EGTA) in the lysis and chromatography buffers used for the purification of the ferredoxin did not change the doublet pattern seen for the purified protein on denaturing gels (data not shown).

Table I. Purification of the four alkene monooxygenase components Step Volume Total protein Total activitya Specific activity Yield Fold purification ml mg milliunits milliunits/mg % x-fold Oxygenase   Extract 417 6,721 111,414 19 100 1   DEAE-Sepharose 205 1,765 57,398 33 52 1.7   Phenyl-Sepharose 140 205 32,896 161 30 8.5   S-300 60 111 28,413 240 26 12.6 Reductase   Extract 407 17,559 2,050,711 168 100 1   DEAE-Sepharose 1,000 2,243 1,043,100 465 51 3   Reactive Green 19 140 262 873,642 3,336 43 20   S-Sepharose 52 33 451,534 13,589 22 81   S-100 7 3.6 213,158 30,030 10 179 Small protein   Extract 770 19,846 408,639 20 100 1   DEAE-Sepharose 240 7,109 317,981 44 78 2.2   Phenyl-Sepharose 5 38 113,643 3,034 28 152   S-100 8.5 28 85,863 4,257 21 213 Ferredoxin   Extract 300 12,943 38,430 6.4 100 1   DEAE-Sepharose 195 220 29,833 135 78 21   Q-Sepharose 47 65 24,771 380 65 59   S-100 4.2 6.8 15,303 2,249 40 351 a A unit of activity is defined as 1 µmol of propylene degraded per minute at 30 °C.

Mass spectrometry was used to provide a more accurate estimation of the molecular weight of the ferredoxin. Two peaks were obtained for the purified protein with intensities similar to the banding pattern observed on SDS-PAGE. The major peak had a molecular weight of 13,335 and the minor peak had a molecular weight of 12,852. Using gel filtration chromatography, a molecular mass for the ferredoxin of 26 kDa was estimated. From these data it appears that the ferredoxin is a dimer of 13-kDa subunits. The protein is most likely a homodimer based on the relative intensities of the peaks observed under denaturing conditions. Possibly, a small population of monomeric units has undergone modification (e.g. cleavage) leading to the banding patterns observed.

Quantitative amino acid analysis was performed on the ferredoxin and used as the basis for reporting analytical data for the purified protein. Using a subunit molecular mass of 13,300 daltons, the amino acid composition of the ferredoxin was: 14 Asp + Asn, 2 Thr, 4 Ser, 20 Glu + Gln, 8 Pro, 14 Gly, 7 Als, 16 Val, 1 Met, 5 Ile, 12 Leu, 2 Tyr, 4 Phe, 7 His, 4 Lys, 2 Arg, and 3 Cys (determined after performic acid oxidation). Based on the results of amino acid analysis, the biuret protein assay was found to overestimate the protein concentration by a factor of 1.1. The iron content was determined to be 2.02 mol iron per mol of ferredoxin monomer by colorimetric analysis using Ferene S, and 2.13 mol iron per mol of ferredoxin monomer by plasma emission spectroscopy. Plasma emission spectroscopy also revealed the presence of variable levels of zinc (0.2 to 1.3 mol zinc per mol of ferredoxin monomer) in individual preparations. Dialysis versus , -dipyridyl resulted in the complete removal of zinc from ferredoxin samples with no effect on specific activity or iron concentration, indicating that the zinc is adventitiously bound.

The UV/visible spectra of the ferredoxin under various conditions are presented in Fig. 2. The air-oxidized protein (Fig. 2, solid line) had absorption maxima at 277, 323, and 454 nm and an absorption shoulder at 575 nm, and the dithionite-reduced protein (Fig. 2, dashed line) had absorption maxima at 432 and 515 nm. The difference spectrum for oxidized minus dithionite-reduced protein had absorption maxima at 460 and 370 nm (Fig. 2, inset). These various spectral features are characteristic of a class of iron-sulfur proteins (ferredoxins) containing a Rieske-type [2Fe-2S] center (29-31). Extinction coefficients of 7,230 M¹ ·cm¹ at 460 nm and 26,863 M¹·cm¹ at 280 nm were calculated for the ferredoxin monomer.

The addition of NADH to air-oxidized ferredoxin resulted in no changes in the UV/visible spectrum. However, the addition of a small amount of cell-free extract along with NADH resulted in the complete reduction of ferredoxin (Fig. 2, dashed-dotted line), indicating that NADH-dependent ferredoxin reduction is mediated by a reductase activity. The nature of this reductase activity is investigated in more detail below.

Air-oxidized ferredoxin was EPR silent, whereas dithionite-reduced ferredoxin had an EPR spectrum at 10 K displaying rhombic symmetry and with g values of 2.016, 1.918, and 1.776 (Fig. 3). These spectral features are characteristic of proteins that contain Rieske-type iron-sulfur centers (29, 31, 32). Spin quantitation of the EPR signal from reduced sample gave 1.06 ± 0.03 spins per mol of ferredoxin monomer.

The circular dichroism spectra of air-oxidized and dithionite-reduced ferredoxin are presented in Fig. 4. The features of these spectra are very similar to those reported for several proteins that contain Rieske-type [2Fe-2S] clusters (29, 30, 33). Together, the metal analysis and the UV/visible, EPR, and CD spectra of alkene monooxygenase ferredoxin provide strong evidence for the presence of one [2Fe-2S] Rieske-type center for each 13,300 molecular weight monomeric unit of this protein.

The reduction potential of the iron-sulfur center of this ferredoxin was determined by reductive titration using EPR spectroscopy. As shown in Fig. 5, the data conformed to the Nernst equation for a one-electron event with a calculated midpoint potential of 49 ± 10 mV. This midpoint potential falls in the range of 150 to +350 mV reported for other proteins containing Rieske-type iron-sulfur centers (34).

The specificity of the alkene monooxygenase ferredoxin in alkene epoxidation was investigated by asking whether other iron-sulfur proteins could substitute in this regard. Neither spinach ferredoxin or Azotobacter vinelandii ferredoxin I were capable of substituting for the alkene monooxygenase ferredoxin component.

Active fractions that did not bind to the initial DEAE-Sepharose column exhibited NADH:ferricyanide oxidoreductase activity, suggesting that this component may be the NADH:acceptor oxidoreductase component of the alkene monooxygenase complex. The reductase was purified 178-fold with the enrichment of a single polypeptide that migrated on SDS-PAGE with an apparent molecular weight of 40,000 (Fig. 1, lane 6). Mass spectrometry provided a more accurate molecular mass estimate for the reductase of 35.5 kDa. The native molecular mass was estimated to be 27 kDa using gel filtration chromatography, suggesting that the reductase is a monomer.

The reductase catalyzed the reduction of the artificial electron acceptors cytochrome c, potassium ferricyanide, and dichlorophenolindophenol. NADPH was also utilized as an electron donor by the reductase, but with rates approximately 3-fold lower than with NADH. As shown in Fig. 6, the UV/visible absorption spectrum of the reductase had absorbance maxima at 274, 391, 421, and 455 nm, features that are similar to a number of well-characterized reductases that contain flavin and [2Fe-2S] cofactors (23, 27, 35). The flavin prosthetic group was identified as FAD, which was present in a stoichiometric ratio of 1.3 mol of FAD per mol of reductase. Iron analysis revealed the presence of 1.3 mol of iron per mol of reductase, which is somewhat lower than the 2 mol of iron predicted for a 2Fe-2S cluster. The reductase was the least stable of the four components, and some loss of iron and activity may have occurred during purification.

As mentioned above and shown in Fig. 2, the addition of NADH and a small amount of cell-free extract to the oxidized ferredoxin resulted in the reduction of the [2Fe-2S] center to the same level as was observed upon addition of the chemical reductant dithionite. It seems logical to predict that the reductase that purifies as a component of the alkene monooxygenase complex catalyzes the transfer of these electrons. To test this hypothesis, a catalytic amount of purified reductase was added to the oxidized ferredoxin in the presence of NADH. The resultant decrease in absorbance at 460 nm was diagnostic of the time-dependent reduction of the [2Fe-2S] center of the ferredoxin. As shown in Fig. 7A, the iron-sulfur center of the ferredoxin was rapidly reduced in the presence of purified reductase. The rate of reduction was proportional to the amount of reductase added (compare traces 2 and 3 in Fig. 7A) providing specific rates of ferredoxin reduction of 41.9 and 45.7 µmol of ferredoxin reduced per min/mg of reductase, respectively. The addition of the other purified components of alkene monooxygenase (the oxygenase and small protein) in the presence of NADH did not lead to reduction of the ferredoxin, demonstrating that these components do not have reductase activity.

We have noted that cell-free extracts prepared from Xanthobacter strain Py2 grown with glucose or acetone as carbon sources have high levels of NADH:ferricyanide oxidoreductase activity, yet are unable to substitute for the reductase purified in this study in reconstituting alkene monooxygenase activity with the other components. This suggests that the reductase component of alkene monooxygenase is highly specific and an inducible component of the system. To further investigate the specificity of the reductase, we asked whether the ubiquitous reductase proteins present in cell-free extracts from glucose- or acetone-grown cells could substitute for the alkene monooxygenase reductase component in mediating the reduction of the ferredoxin. As shown in Fig. 7B, trace 1, the addition of cell-free extracts from acetone- or glucose-grown cells resulted in a very slow rate of NADH-dependent ferredoxin reduction. In contrast, the addition of a corresponding amount of cell-free extract from propylene-grown cells resulted in the rapid NADH-dependent reduction of the ferredoxin (Fig. 7B, trace 2). This result verifies the specific and inducible nature of the alkene monooxygenase-linked reductase purified in this study.

Further fractionation of the active fraction that eluted from the initial DEAE-Sepharose column between 200 and 250 mM NaCl revealed the presence of two separable components. These components were resolved by phenyl-Sepharose chromatography as described under "Materials and Methods". The component that was not retained by phenyl-Sepharose was purified 212-fold with the enrichment of a single polypeptide that had an apparent molecular mass of 14.7 kDa (Table I and Fig. 1, lane 4). Mass spectrometry provided a molecular mass estimate for the protein of 11.1 kDa. The native molecular mass was estimated to be 14.9 kDa by gel filtration chromatography, indicating that the protein is a monomer. The UV/visible absorption spectrum of the purified protein did not have any discernible features in the wavelength range from 300 to 800 nm (Fig. 6). No metals were present in purified preparations of the protein. Likewise, no detectable EPR signals were associated with the small protein.

One possible role of the small protein might be to facilitate electron transfer between the reductase and ferredoxin components. To investigate this possibility, a stoichiometric amount of small protein was added to the ferredoxin in assays where the time course of ferredoxin reduction was followed (i.e. the assay conditions of Fig. 7). The addition of the small protein had no effect on the rate of ferredoxin reduction, indicating that the small protein does not have a role in this initial electron transfer reaction.

The active fraction retained by phenyl-Sepharose was purified 12.6-fold with enrichment of three polypeptides with apparent molecular masses of 53, 43, and 6 kDa (Fig. 1, lane 3). The relatively low-fold purification required to obtain a homogeneous preparation of the oxygenase indicates that this protein represents a sizable percentage of soluble protein present in cells grown with propylene as the carbon source and that it is the most abundant of the alkene oxygenase components. These observations agree with our previous study showing that polypeptides with molecular masses of 53 and 43 kDa (as well as several others) were newly synthesized, and accumulated to high levels, in glucose-grown Xanthobacter cultures induced for alkene monooxygenase activity (19).

Mass spectrometry provided more accurate molecular mass estimations of 58.1, 38.0, and 9.65 kDa for the three oxygenase polypeptides. The staining intensities of the three polypeptides on SDS-PAGE gave relative molar ratios of 1.0 (58.1-kDa band), 0.93 (38.0-kDa band), and 1.2 (9.65-kDa band), suggesting a minimal core complex consisting of one each of the three subunits with a molecular mass of 106 kDa. The molecular mass of the oxygenase was estimated to be 195 kDa by gel filtration, suggesting that the native enzyme has an ()₂ quaternary structure.

Metal analysis revealed the presence of iron as the only metal associated with alkene monooxygenase. Based on the results of quantitative amino acid analysis, the biuret protein assay was found to overestimate protein concentration by the fairly sizable factor of 1.33. Using the protein concentration determined from amino acid analysis, the stoichiometry of iron was calculated to be 3.8 mol of iron/mol of ()₂ complex or 1.9 mol of iron/mol of protomer.

The UV/visible absorption spectrum of the oxygenase component is presented in Fig. 6. The lack of any significant absorption in the wavelength range from 300 to 800 nm indicates that the protein does not contain iron-sulfur or heme prosthetic groups. No EPR signals were detected for air-oxidized (as isolated) oxygenase or oxygenase reduced by the addition of various amounts of dithionite (0.2 eq to 1 mM dithionite). Likewise, no EPR signals were detected in a sample of the oxygenase reduced by the addition of NADH and catalytic amounts of the other three components (reductase, ferredoxin, and coupling protein).

The properties of the multimeric, iron-containing protein described above suggest that it contains the active site for oxygen activation and alkene epoxidation. To obtain more evidence supporting this proposal, the ability of this component to complement cell-free extracts in which alkene monooxygenase was inactivated by propyne was investigated. Propyne has previously been shown to be a specific, irreversible inactivator of alkene monooxygenase activity in cell suspensions of Xanthobacter strain Py2 (13). The inactivation is consistent with propyne acting as a mechanism-based inactivator in a manner analogous to that described for acetylene inactivation of methane (36) and ammonia monooxygenases (37). If this is the case, propyne should specifically inactivate the oxygenase component of alkene monooxygenase and not affect the activity of other required components. In the present analysis, no detectable alkene monooxygenase activity remained in cell-free extracts treated with propyne. Alkene monooxygenase activity could be fully restored by the addition of the oxygenase component to the extract (data not shown). To obtain 100% of the initial activity of the extract (i.e. the activity before propyne treatment), it was necessary to add an amount of purified oxygenase approximately 5-10-fold higher than the amount of oxygenase that would have been initially present in the extract. The need to add excess oxygenase for full reactivation may be due to inhibition exerted by the inactive oxygenase component still present in the extract, since the inactive component probably competes for binding to one or more of the other components. Significantly, the single addition of the other components (reductase, ferredoxin, or small protein) to the propyne-treated extract did not stimulate alkene monooxygenase activity to any detectable degree.

DISCUSSION

The results of the present work demonstrate that alkene monooxygenase from Xanthobacter strain Py2 is a multicomponent enzyme consisting of the following: 1) an NADH reductase, which provides the reductant for O₂ activation; 2) a Rieske-type ferredoxin, which is a mediating electron transfer protein; 3) an oxygenase, which contains the catalytic center for alkene epoxidation; and 4) a small protein of unknown function. The function of the small protein may be analogous to that of component B of the multicomponent soluble methane monooxygenases purified from Methylosinus trichosporium OB3b and Methylococcus capsulatus Bath (27, 38). Component B is a small protein (16 kDa), containing no cofactors, that dramatically affects the rate and specificity of turnover of methane monooxygenase (39). Component B is thus regarded to serve as an effector or regulatory protein. The small protein component of alkene monooxygenase, and small proteins recently identified as components of other bacterial oxygenase systems (14, 28), may have similar roles.

The proposed interplay of the three alkene monooxygenase components for which functions can be assigned is presented in Scheme 1. This scheme highlights the interesting observation that alkene monooxygenase uses a two-component electron transfer chain to provide reductant to the oxygenase. In this regard alkene monooxygenase differs from other bacterial oxygenases that catalyze the oxidation of short chain aliphatic hydrocarbons, most notably methane monooxygenase, which transfers electrons directly from the reductase to the oxygenase (27). The two-component electron transfer system of alkene monooxygenase is more typical of a broad class of dioxygenases that act on aromatic hydrocarbon substrates, including benzene dioxygenase from Pseudomonas putida (40), biphenyl 2,3-dioxygenase from Pseudomonas sp. strain LB400 (41), naphthalene dioxygenase from Pseudomonas sp. strain NCIB 9816 (42), toluene dioxygenase from P. putida (43), and halobenzoate 1,2-dioxygenase from Pseudomonas aeruginosa 142 (44). Each of these aromatic dioxygenases is a three-component enzyme system comprised of an NADH reductase, a Rieske-type ferredoxin and a terminal dioxygenase.

Scheme 1.

Recently, Fox and co-workers (31) showed that an aromatic monooxygenase, toluene-4-monooxygenase from Pseudomonas mendocina KR1, was comprised of four components as follows: a reductase, a Rieske-type ferredoxin, an effector protein, and the terminal hydroxylase. Toluene-4-monooxygenase, like alkene monooxygenase, thus combines elements of both the aromatic dioxygenases (two-component electron transfer scheme) and methane monooxygenase (regulatory or effector protein) systems (31). Alkene monooxygenase and toluene-4-monooxygenase are the only reported bacterial oxygenases to have this unique four-component structure.

Studies of three Gram-positive bacteria capable of growth with aliphatic alkenes as carbon sources have revealed that the alkene monooxygenases of these bacteria function as multicomponent enzymes as well (15, 16). The alkene monooxygenase from one of these bacteria, N. corallina strain B-276, was shown to be a three-component system consisting of an oxygenase, a reductase, and a small protein that lacked detectable cofactors (14). Notably, no electron-transferring ferredoxin was associated with alkene monooxygenase from N. corallina, highlighting a fundamental difference between the Xanthobacter and Nocardia systems. Another notable difference between the two systems is the subunit stoichiometry and quaternary structure of the oxygenase component. The oxygenase component of N. corallina alkene monooxygenase was comprised of two subunits with molecular masses of 53 and 35 kDa, apparently arranged in an quaternary structure (14). As discussed above, Xanthobacter alkene monooxygenase was comprised of three subunits arranged in an ₂₂₂ configuration.

The initial biochemical and spectroscopic characterization of the oxygenase component of Xanthobacter alkene monooxygenase suggests that this protein contains one diiron site per protomer or two diiron active sites per holoenzyme. Binuclear diiron centers have been implicated as serving as the oxygen-activating sites for a number of soluble bacterial hydrocarbon oxygenases, including methane monooxygenase (45), toluene-4-monooxygenase (31), and toluene-2-monooxygenase (28). Alkene monooxygenase from N. corallina contained two iron atoms per protein, suggesting that this enzyme contains a diiron center as well (14). Based on the data obtained to date, we are unable to state unequivocally that the irons of alkene monooxygenase form a binuclear center. Centers of this type do not have UV/visible absorbance and can only be observed by EPR in the mixed valent state (31). In initial experiments, we have been unable to obtain an EPR signal for the oxygenase component, suggesting that the divalent state is not readily formed by this putative diiron center. Similar results have been obtained for other diiron enzymes, for example toluene-4-monooxygenase and stearoyl-ACP desaturase, and more sophisticated spectroscopic techniques were necessary to define the diiron centers of these enzymes (31). Likewise, further experimentation is required to verify the nature of the iron site of Xanthobacter alkene monooxygenase.

In summary, a nonheme iron-containing alkene monooxygenase with a novel four-component structure has been purified and characterized from a Gram-negative, aliphatic alkene-utilizing bacterium. The properties of Xanthobacter alkene monooxygenase are surprisingly different from those reported recently for alkene monooxygenase from a Gram-positive bacterium (N. corallina). There are a number of potentially interesting questions relating to this class of enzymes that warrant further investigation now that the purified system is in hand. One of these is the remarkable stereospecificity exhibited by the enzymes; in cell suspensions, alkene monooxygenases from different alkene-utilizing bacteria catalyzed alkene epoxidation with a high degree of stereoselectivity for forming the corresponding (R)- or (S)- epoxides (46). It will be interesting to determine whether the purified enzyme exhibits this same stereospecificity and, if so, to investigate the molecular basis for the stereoselectivity. Alkene monooxygenase in whole cell suspensions of Xanthobacter Py2 is also capable of initiating the oxidation of chlorinated alkenes (13); it will also be interesting to investigate the mechanistic and kinetic details of xenobiotic transformations with the purified enzyme system.