Purification and Characterization of 2-Hydroxybiphenyl 3-Monooxygenase, a Novel NADH-dependent, FAD-containing Aromatic Hydroxylase from Pseudomonas azelaica HBP1*

2-Hydroxybiphenyl 3-monooxygenase (HbpA), the first enzyme of 2-hydroxybiphenyl degradation inPseudomonas azelaica HBP1, was purified 26-fold with a yield of 8% from strain HBP1 grown on 2-hydroxybiphenyl. The enzyme was also purified from a recombinant of Escherichia coliJM109, which efficiently expressed the hbpA gene. Computer densitometry of scanned slab gels revealed a purity of over 99% for both enzyme preparations. Gel filtration, subunit cross-linking, and SDS-polyacrylamide gel electrophoresis showed that the enzyme was a homotetramer with a molecular mass of 256 kDa. Each subunit had a molecular mass of 60 kDa containing one molecule of noncovalently bound FAD. The monooxygenase had a pI of 6.3. It catalyzed the NADH-dependent ortho-hydroxylation of 2-hydroxybiphenyl to 2,3-dihydroxybiphenyl. Molecular oxygen was the source of the additional oxygen of the product. The enzyme hydroxylated various phenols with a hydrophobic side chain adjacent to the hydroxy group. All substrates effected partial uncoupling of NADH oxidation from hydroxylation with the concomitant formation of hydrogen peroxide. 2,3-Dihydroxybiphenyl, the product of the reaction with 2-hydroxybiphenyl, was a non-substrate effector that strongly facilitated NADH oxidation and hydrogen peroxide formation without being hydroxylated and also was an inhibitor. The apparentK m values (30 °C, pH 7.5) were 2.8 μm for 2-hydroxybiphenyl, 26.8 μm for NADH, and 29.2 μm for oxygen. The enzyme was inactivated byp-hydroxymercuribenzoate, a cysteine-blocking reagent. In the presence of 2-hydroxybiphenyl, the enzyme was partly protected against the inactivation, which was reversed by the addition of an excess of dithiothreitol. The NH2-terminal amino acid sequence of the enzyme contained the consensus sequence GXGXXG, indicative of the βαβ-fold of the flavin binding site and shared homologies with that of phenol 2-hydroxylase from Pseudomonas strain EST1001 as well as with that of 2,4-dichlorophenol 6-hydroxylase from Ralstonia eutropha.

Recently, the microbial degradation of 2-hydroxy-and 2,2Јdihydroxybiphenyl has become important to researchers involved in the desulfurization of coal and petroleum, since it was reported that 2-hydroxy-and 2,2Ј-dihydroxybiphenyl are the end products of the bacterial desulfurization of dibenzothiophene, a major sulfur-containing component of fossil fuels (14 -16). Furthermore, the two compounds inhibit the dibenzothiophene-degrading activity in cell-free extracts of Rhodococcus erythropolis D-1 (17).
Previously, we partly purified the 2-hydroxybiphenyl 3-monooxygenase for producing metabolites necessary for elucidating the degradation pathway of 2,2Ј-dihydroxybiphenyl in P. azelaica HBP1 (3). We now report on the purification and characterization of the enzyme. 2-Hydroxybiphenyl 3-monooxygenase is a novel flavin-containing, NADH-dependent aromatic hydroxylase with a broad substrate spectrum. Its relationship to other members of the group of phenolic hydroxylases is discussed.

Bacterial Strains and Culture Conditions
P. azelaica HBP1 was cultured on 2-hydroxybiphenyl as described previously (2). Cultures were grown in Erlenmeyer flasks on a rotary shaker (150 -170 rpm) at 30°C. The purity of the cultures was regularly tested by streaking them out on mineral salts medium plates that were prepared by adding 2-hydroxybiphenyl as a concentrated methanolic solution to the hot mineral medium containing 15 g/liter agar. Recombinants of Escherichia coli JM109 were cultivated in LB medium supplemented with ampicillin (100 g/ml) at 37°C (18).
For the purpose of protein purification, large amounts of cells were produced in aerated 20-liter carboys equipped with a magnetic stirring bar. Cells were harvested in the late exponential growth phase (with an A 546 of about 1.5) by centrifugation (15 min at 6,000 ϫ g) at 4°C. The culture fluid was passed through a paper filter before centrifugation to remove remaining 2-hydroxybiphenyl crystals. Cells were washed twice with an excess amount of triethanolamine-HCl buffer (10 mM, pH 7.5). Approximately 20 g of cell paste (wet weight) was obtained from one 20-liter batch. The cell paste was stored at Ϫ20°C until further use. Chemicals 2,3-Dihydroxybiphenyl was obtained from Wako Chemicals GmbH (Neuss, Germany). 2,5-Dihydroxybiphenyl, 2-propylphenol, 2-ethylphenol, and 2-sec-butylphenol were purchased from Aldrich-Chemie (Steinheim, Germany). 2-sec-Butylcatechol was prepared by the enzymatic conversion of 2-sec-butylphenol with 2-hydroxybiphenyl 3-monooxygenase and by subsequent purification of the product by means of a preparative HPLC apparatus. 18

Protein Purification
The enzyme purification apparatus was located in a laboratory at room temperature, but the buffer reservoirs and the sample collector vials were kept on ice. The flow rate was 1 ml min Ϫ1 for all chromatography steps, and the volume of the collected fractions was always 1 ml if not stated otherwise.
Preparation of Crude Cell Extract-Ten grams of cell paste was suspended in 50 ml of triethanolamine-HCl buffer (10 mM, pH 7.5). Crude cell extract was prepared by passing the cells through a French pressure cell (two passages at 20,000 p.s.i.) followed by ultracentrifugation (30 min at 50,000 ϫ g) at 4°C.
Protamine Sulfate Treatment-Crude cell extract was freed from DNA and basic proteins by slowly adding a 2% (w/v) protamine sulfate solution until a final concentration of 0.05 mg of protamine sulfate/mg of protein was reached. After stirring for 30 min at 4°C, the precipitated biopolymers were removed by centrifugation (30 min, 50,000 ϫ g, 4°C).
Hydrophobic Interaction Chromatography-Active fractions from the protamine sulfate treatment step were pooled, supplemented with (NH 4 ) 2 SO 4 to a final concentration of 0.9 M, and loaded onto a Fractogel TSK Butyl 650 S (Merck, Darmstadt, Germany) column (1 ϫ 15 cm) equilibrated with 0.75 M (NH 4 ) 2 SO 4 in 100 mM Na 2 HPO 4 buffer (pH 7.0). The column was washed with 2-3 volumes of the equilibration buffer, and the enzyme was eluted with a linear gradient from 0.75 to 0 M (NH 4 ) 2 SO 4 . The fractions containing the enzyme were desalted on Sephadex G-25 M columns (Pharmacia Biotech, Uppsala, Sweden).
Second Anion Exchange Chromatography-The desalted enzyme solution was supplemented with FAD to a final concentration of 0.3 mM and loaded onto an anion exchange column (Fractogel EMD TMAE-650) equilibrated with 10 mM triethanolamine-HCl buffer (pH 8.2). Isocratic conditions (0.3 M NaCl) were used to elute the 2-hydroxybiphenyl 3-monooxygenase.
Gel Filtration-The fractions containing active enzyme from the preceding step were incubated with 0.3 mM FAD for 30 min on ice. The solution was passed with a flow rate of 1.5 ml min Ϫ1 through a Superdex 200 gel filtration column (1.6 ϫ 60 cm; Pharmacia) equilibrated with 50 mM Na 2 HPO 4 buffer (pH 7.5). Pooled fractions were concentrated by ultrafiltration with Centricon concentrators (Amicon Inc., Beverly, MA). The concentrated solution of the purified enzyme (3-5 mg of protein/ml) was stored at Ϫ20°C until further use.

Analytical Methods
The disappearance of substrates and the formation of metabolites were monitored by high performance liquid chromatography (HPLC). Protein was removed from the samples (3 ml) by the addition of 20 l of 8.5% H 3 PO 4 and subsequent centrifugation. The samples were analyzed by injecting 20 l onto a computer-controlled Gynkotek high performance liquid chromatograph consisting of a Gina 50 automated injection module, a M480 G gradient pump, an on-line degasser, and an UVD 340 S photodiode array detector (Gynkotek, Germering, Germany). Reverse-phase separation was achieved on a Waters Nova-Pak C-18 column (Waters-Millipore, Milford, MA) by applying a linear gradient of 60 -70% B (A, 10 mM H 3 PO 4 ; B, 90% methanol, 10% 10 mM H 3 PO 4 ) with a flow rate of 0.6 ml min Ϫ1 .
The flavin cofactor was extracted from the protein by treating a 400-l sample of 2-hydroxybiphenyl 3-monooxygenase (2.6 mg ml Ϫ1 ) in 20 mM phosphate buffer (pH 7.2) with 100 l of 5% trichloroacetic acid for 5 min at room temperature followed by centrifugation at 14,000 rpm. HPLC analysis (isocratic conditions with an eluent consisting of 40% methanol and 60% 10 mM H 3 PO 4 (v/v)) of the supernatant allowed the identification of the cofactor by comparison of the retention times and the UV-VIS spectra to authentic FAD (retention time, 10.3 min) and FMN (retention time, 14.4 min).
The protein contents of the cell extracts and the purified protein fractions were measured with the Bio-Rad protein assay kit (Bio-Rad Laboratories, Mü nchen, Germany). Bovine serum albumin in the concentration range from 2 to 20 g ml Ϫ1 was used as a standard.

Labeling with 18 O 2
Incorporation of 18 O 2 into the products of the enzyme reaction was measured to confirm the monooxygenation reaction. The experiments were carried out in two 13.8-ml serum flasks, which contained 1 ml of the enzyme incubation mixture consisting of 30 milliunits of 2-hydroxybiphenyl 3-monooxygenase, 0.3 mM NADH, and 10 mM triethanolamine-HCl buffer (pH 7.5). Both flasks were sealed with butyl rubber stoppers. A 50:50 mixture of 18 O 2 to 16 O 2 in one of the flasks was obtained by injecting 2.6 ml of 18 O 2 into the head space (12.8 ml). The other flask contained 16 O 2 from air. The enzyme reaction was started by injecting 15 l of a methanolic solution of 2-hydroxybiphenyl (0.1 mM) through the rubber seals. After 15 min, the incubation mixtures were acidified with 5 l of H 3 PO 4 (8.5%) to a final pH of 2 and subsequently centrifuged to remove the precipitated protein. The metabolites were extracted from the supernatants with an equal volume of ethyl acetate and dried with anhydrous sodium sulfate. Samples were derivatized with N,O-bis(trimethylsilyl)trifluoroacetamide and subjected to gas chromatography-mass spectrometry analysis as described before (3).
Isoelectric focusing gels contained 7.5% acrylamide and a broad pH range ampholyte (Resolyte 3.5-10, BDH Laboratory Supplies, Poole, United Kingdom). Isoelectric focusing was performed in the same apparatus that was used for SDS-PAGE (20). The pH gradient was formed by filling the upper buffer chamber with catholyte solution (20 mM sodium hydroxide) and the lower buffer chamber with anolyte solution (10 mM phosphoric acid). Five micrograms of protein was loaded onto the isoelectric focusing gel, and the pI was estimated from the position of the protein band relative to the position of the bands of the marker proteins (broad pI calibration kit, 3.5-10; Pharmacia). The molecular mass of the pure enzyme was determined under native conditions by gel filtration on a Superdex 200 gel filtration column (1.6 ϫ 60 cm, Pharmacia) equilibrated with 20 mM phosphate buffer (pH 7.5). The column was calibrated with blue dextran 2000 and the following reference proteins (Pharmacia): thyroglobulin (669 kDa), ferritin (440 kDa), catalase (232 kDa), aldolase (158 kDa), bovine serum albumin (67 kDa), ovalbumin (43 kDa), chymotrypsinogen A (25 kDa), and ribonuclease A (13.7 kDa).

Cross-linking
Covalent cross-linking of enzyme subunits was carried out by the method of Griffith (21) with glutaraldehyde as the reactive agent. Ten microliters of glutaraldehyde (0.5%, w/v) was added to 250 l (0.5 mg ml Ϫ1 ) enzyme solution and incubated for 12 h at room temperature. The treated protein sample was denatured with SDS and analyzed by SDS-PAGE.

NH 2 -terminal Amino Acid Sequence
The NH 2 -terminal amino acid sequence of the enzyme was determined by automated Edman degradation.

Kinetic Measurements
The specific activity of purified 2-hydroxybiphenyl 3-monooxygenase and of partly purified enzyme fractions was routinely measured in an spectrophotometric assay with NADH and was defined as the amount of NADH (mol) that the enzyme oxidized in the presence of 2-hydroxybiphenyl/min/mg of protein. The reaction was started by the addition of the substrate. The substrate-specific oxidation rates were corrected for nonspecific NADH oxidation in the absence of substrate (endogenous rate). The kinetic parameters V max and K m were calculated with a computer program (IGOR Pro; WaveMetrics Inc., Lake Oswego, OR). The calculation was based on weighted nonlinear regression analysis of the Michaelis-Menten model. We confirmed by Monte Carlo simulations that the least square estimators of the paramaters were close to being normally distributed and that the model was close to linear (22,23).

Chemical Modification
2-Hydroxybiphenyl 3-monooxygenase (4 -4.5 M, in 20 mM phosphate buffer, pH 7.2) was incubated with various concentrations of p-hydroxymercuribenzoate at 25°C. Aliquots of 10 l were withdrawn from the incubation mixtures, immediately diluted in 990 l of phosphate buffer, and assayed for enzyme activity. Reactivation of the mercurated enzyme was followed after the addition of 1 mM dithiothreitol.

Purification of 2-Hydroxybiphenyl 3-Monooxygenase-2-Hy-
droxybiphenyl 3-monooxygenase (HbpA) from cells of P. azelaica strain HBP1 grown on 2-hydroxybiphenyl as the only carbon and energy source was purified 26-fold with a yield of 8% (Table I). The enzyme was brightly yellow and accounted for about 4% of the total cell protein in the crude cell extract. Protein samples from the final step of the purification were subjected to SDS-PAGE. The enzyme migrated as a single band. Computer densitometry of scanned slab gels revealed a purity of over 99% (Fig. 1A). To detect minor contaminating proteins in the sample, separate gels were stained with silver. No additional protein bands could be detected by silver staining.
At a temperature of 4°C, the activity of the purified enzyme remained constant for at least 12 h, and after 24 h approximately 90% of the activity was still present. The pure enzyme was stored at a concentration of 3.8 mg/ml in 50 mM phosphate buffer (pH 7.5) at Ϫ20°C. Under these conditions, it retained full activity for at least 6 months. Molecular Characteristics of the Purified Enzyme-The relative molecular mass of the pure enzyme under native conditions was determined to be 256 kDa by gel filtration. SDS-PAGE of the purified monooxygenase gave a single band with an estimated molecular weight of 60 kDa, suggesting a tetrameric quatenary structure of the enzyme. The subunits of the enzyme were covalently cross-linked with glutaraldehyde. SDS treatment of such a preparation and subsequent separation of its components by SDS-PAGE led to the appearance of three bands with estimated molecular masses of 259, 138, and 64 kDa (not shown). It is assumed that the bands reflect the tetrameric, dimeric, and monomeric forms of the enzyme. A protein band corresponding to a trimeric form was not found.
The pI was determined to be at pH 6.3 Ϯ 0.7 (95% confidence interval) by isoelectric focusing as described under "Material and Methods." Identification of FAD as the Prosthetic Group-The enzyme activity in the pooled fractions of the hydrophobic interaction chromatography step was increased by 30% when 10 M FAD was present in the assay. The addition of FMN to the assay mixture had no effect on the enzyme activity. These results suggested that FAD might be the prosthetic group of the monooxygenase. To isolate the cofactor, samples of the purified enzyme were treated with 5% trichloroacetic acid. HPLC anal-  ysis confirmed that the cofactor of the monooxygenase was indeed FAD, since the extracted compound co-chromatographed with authentic FAD and also had an identical UV-VIS spectrum (diode array detection).
Spectral Properties of the Monooxygenase-An enzyme preparation from the final purification step (gel filtration) was taken for spectroscopic analysis. The enzyme sample was judged to be free of unbound FAD, since the specific activity did not decrease after gel filtration and the addition of FAD to the eluted enzyme did not further stimulate activity. The UV-vis spectrum of the purified 2-hydroxybiphenyl 3-monooxygenase showed maxima at 382 and 452 nm and a minimum at 410 nm. This is characteristic of a flavoprotein (Fig. 2). The ratio of the absorbance at 250 nm relative to that at 450 nm was 6.2. Above 320 nm, the spectrum closely resembled that of authentic FAD except that the ratio of the absorbance at 375 nm to the one at 450 nm was 0.92 for the enzyme as compared with 0.82 for free FAD. The absorption at 450 nm of an enzyme solution with a protein content of 2.3 mg ml Ϫ1 (8.9 M) was 0.37 (Fig. 2). On the basis of a molar absorption coefficient (⑀ 450 ) of 11,300 M Ϫ1 cm Ϫ1 for the FAD moiety (25), it was estimated that 1 mol of enzyme contained 3.7 mol of FAD.
Isotope Labeling Experiment-To confirm the monooxygenation reaction for the formation of 2,3-dihydroxybiphenyl from 2-hydroxybiphenyl by 2-hydroxybiphenyl 3-monooxygenase, the enzymatic reaction was carried out in the presence of 18 O 2 . The reaction products were isolated, derivatized, and analyzed by gas chromatography-mass spectrometry. Fig. 3 shows the mass spectra of the TMS derivatives of 2,3-dihydroxybiphenyl from the incubation with 16 O 2 (A) and with a 1:1 mixture of 16 O 2 to 18 O 2 (B). The fragmentation patterns of the two spectra clearly show that one atom of dioxygen was incorporated into 2-hydroxybiphenyl during the course of the reaction.
Optimal pH and Temperature Conditions for Enzymatic Activity-The effect of temperature on the activity of the 2-hydroxybiphenyl 3-monooxygenase was investigated over the range of 10 -50°C. The temperature optimum of the enzymatic reaction was at 33°C. The monooxygenase retained more than 70% of its activity in the temperature range from 27 to 40°C. Above 50°C, enzymatic activity was completely absent. The monooxygenase maintained more than 80% of its activity in the pH range between pH 7.2 and 7.8 and showed a maximum activity at pH 7.5. Beyond pH 7.8, the activity of the 2-hydroxybiphenyl 3-monooxygenase declined abruptly with increasing pH.
Substrate Specificity and Uncoupling Effects of Substrates and Products-Incubation of 2-hydroxybiphenyl with the purified 2-hydroxybiphenyl 3-monooxygenase led to the formation of a reaction product that was isolated and identified as 2,3dihydroxybiphenyl by gas chromatography-mass spectrometry analysis. The substrate specificity of the 2-hydroxybiphenyl 3-monooxygenase was investigated with a variety of orthosubstituted phenols. Table II shows initial velocities of substrate conversions determined by independent measurements of NADH oxidation, oxygen consumption, and substrate disappearance at 30°C and pH 7.2. It is evident that for all substrates the stoichiometric coefficients for NADH oxidation (or oxygen consumption) and hydroxylation were not equal. This indicates that the substrates partially uncoupled oxygen activation from hydroxylation with the resultant reduction of both atoms of oxygen to form hydrogen peroxide. The formation of hydrogen peroxide was shown by the addition of catalase at the end of the oxygen uptake experiments. For 2-hydroxybiphenyl as the substrate, 15% of the consumed oxygen was recovered after the addition of catalase, which means that according to the stoichiometry of the catalase reaction 30% of the consumed oxygen was diverted to hydrogen peroxide. With 2-sec-butylphenol, which was more rapidly metabolized than 2-hydroxy-

2-Hydroxybiphenyl 3-Monooxygenase from P azelaica HBP1
biphenyl, 47% of the consumed oxygen was released as hydrogen peroxide. The uncoupling determined by the catalasedependent oxygen production exceeded the uncoupling determined by the measurement of initial reaction velocities. Subsequent experiments showed that the product of the hydroxylation of 2-hydroxybiphenyl, 2,3-dihydroxybiphenyl was a strong effector of 2-hydroxybiphenyl 3-monooxygenase and stimulated stoichiometric NADH and oxygen consumption without undergoing hydroxylation. This explained the elevated amounts of hydrogen peroxide observed after the complete conversion of 2-hydroxybiphenyl.
Measurement of NADH oxidation with low concentrations of 2-hydroxybiphenyl and 2-sec-butylphenol revealed the contribution of the reaction products to the consumption of NADH (Table III). After the complete consumption of 10 M 2-hydroxybiphenyl, the NADH oxidation did not cease but continued at a lower rate (1.93 mol min Ϫ1 mg of protein Ϫ1 ). The residual activity was attributed to the uncoupling effect of 2,3-dihydroxybiphenyl, which was formed during the reaction, because it was in good agreement with the oxidation rate obtained with 10 M 2,3-dihydroxybiphenyl as the substrate. The same observation was made with 10 M 2-sec-butylphenol and an eqimolar concentration of the hydroxylated product 2-sec-butylcatechol. Residual activities remaining after the conversion of low substrate concentrations were also observed with 2,2Ј-dihydroxybiphenyl, and 2-propylphenol but, interestingly enough, not with 2,5-dihydroxybiphenyl. It became obvious from these ex-periments that the apparent NADH oxidation rate, when determined by the standard assay method, was always the sum of two oxidation rates, one effected by the substrate and the other one by the product. For calculations concerning the kinetic constants, we therefore only used rates derived from the measurement of initial velocities. This guaranteed that product inhibition and product uncoupling did not interfere with the measurements.
Besides the aromatic compounds, oxygen and NADH were also substrates of the 2-hydroxybiphenyl 3-monooxygenase reaction. In the absence of an aromatic substrate the oxidation of NADH amounted to 0.14 Ϯ 0.04 mol min Ϫ1 mg of protein Ϫ1 in the standard assay. NADPH could replace NADH as the electron donor for the reaction, but its apparent K m value was much higher (Table IV).
Kinetic Studies-The 2-hydroxybiphenyl 3-monooxygenase seemed to follow Michaelis-Menten kinetics with all the substrates tested. The apparent K m and V max values for the aromatic substrates and the reduced pyridine nucleotides were determined from weighted nonlinear regression analysis (Table IV). Reaction rates were independent of the order in which substrate and cofactor were added.
Product and Substrate Inhibition-The relationship between 2-hydroxybiphenyl concentration and enzyme activity was analyzed in the concentration range of 0.001-5 mM. Maximal activity was reached at a concentration of 0.05 mM. In the presence of 1 mM 2-hydroxybiphenyl, the enzyme was inhibited by 15%. With increasing substrate concentrations, the enzyme activity decreased (3 and 5 mM gave 25 and 47% inhibition, respectively). At the solubility threshold of 2-hydroxybiphenyl in water (700 ppm, 4.1 mM) the enzyme still maintained 65% of its maximal activity. The inhibitor constant K i for 2-hydroxybiphenyl was determined to be 6.5 mM.
To examine whether the monooxygenase was inhibited by the product of the reaction with 2-hydroxybiphenyl, we measured the initial velocities of the 2-hydroxybiphenyl-dependent NADH oxidation in the presence of different 2,3-dihydroxybiphenyl concentrations. The rate for the NADH oxidation effected by the addition of the aromatic substrate was corrected for the rate obtained with 2,3-dihydroxybiphenyl in the absence of substrate. Direct plots (initial velocity V versus substrate concentration [S]) revealed that, in the presence of increasing 2,3-dihydroxybiphenyl concentrations, V max decreased more and more (Fig. 4A). Depiction of the kinetic data in a double reciprocal plot (1/V versus 1/[S]) for different fixed concentrations of 2,3-dihydroxybiphenyl showed a pattern typical of a mixed type inhibition. The bunch of reciprocal plots with increasing fixed concentrations of 2,3-dihydroxybiphenyl intersected to the left of the 1/V axis above the [P] axis. The slopes a The proportion of uncoupling of hydroxylation with the different substrates was calculated as the deviation of the apparent oxygen consumption from the assumed 1:1 stoichiometry with respect to substrate conversion. Effects of Thiol Reagent, Metal Ions, and Chloride-A rapid inactivation of the enzyme (4.5-4.8 M) was observed in the presence of p-hydroxymercuribenzoate (pHMB), an efficient blocker of cysteine groups (Fig. 5). The inactivation of the enzyme was a reversible reaction, since the activity was restored by the addition of 1 mM dithiothreitol (Fig. 5A). The maximum extent of inactivation was dependent on the concentration of the inhibitor. An inactivation of 100% was achieved within 1 min upon the addition of 100 M pHMB, whereas an inactivation of 25% was measured in the presence of 10 M pHMB (Fig. 5B). Furthermore, we investigated the influence of effectors on the course of the inactivation reaction. 2-Hydroxybiphenyl (100 M) partly protected the enzyme from inactivation, whereas the presence of 2,3-dihydroxybiphenyl (100 M) did not protect the enzyme from inactivation by pHMB.
The extents of inhibition by heavy metal and chloride ions are listed in Table V. Upon the addition of 10 M of the heavy metal salts CuSO 4 , AgNO 3 , or HgCl 2 , the activity of 2-hydroxybiphenyl 3-monooxygenase ceased immediately, whereas the enzyme underwent a partial inhibition in the presence of chloride ions depending on the concentration. By varying the amounts of 2-hydroxybiphenyl, an uncompetitive inhibition type was found for chloride ions with respect to the substrate 2-hydroxybiphenyl (data not shown).
NH 2 -terminal Amino Acid Sequence-The NH 2 -terminal amino acid sequence of the purified 2-hydroxybiphenyl 3-monooxygenase from the recombinant of E. coli JM109 was determined by automated Edman degradation and compared with other previously described sequences of bacterial phenol hydroxylases (Fig. 6). The NH 2 -terminal amino acid sequence of the enzyme purified from wild type strain HBP1 was also determined (14 amino acids) and was 100% identical to the one of the the enzyme purified from the recombinant. The analyzed sequence of the monooxygenase (HbpA) contained the consen-sus sequence GXGXXG, indicating the fingerprint of an ␤␣␤fold (26).
After 26-fold purification, the enzyme was more than 99% pure as judged by analysis of SDS gels. On the basis of determinations of the relative molecular mass carried out under denaturing and under native conditions as well as by crosslinking experiments with glutaraldehyde, we suggest that under physiological conditions, the enzyme was a 256-kDa tetramer consisting of four subunits, each with a relative molecular mass of 60 kDa. The majority of the flavoproteins so far described consist of subunits with a relative molecular mass in the range of 60 -70 kDa (31) and are monomers or dimers with the exception of melilotate hydroxylase (32) and 2,4-dichlorophenol 6-hydroxylase (33), which are tetramers. The visible absorption spectrum of 2-hydroxybiphenyl 3-monooxygenase was typical of a flavoprotein. The prosthetic group was noncovalently bound FAD. Some loss of FAD during the purification of the enzyme was observed at an ionic strength higher than 0.5 M. The molecular ratio of FAD to protein was 3.7. Therefore, each subunit of the monooxygenase contained one molecule of FAD. The purified 2-hydroxybiphenyl 3-monooxygenase catalyzed the NADH-dependent hydroxylation of 2-hydroxybiphenyl, forming 2,3-dihydroxybiphenyl as the product. Experiments with 18 O 2 proved the enzymatic incorporation of one oxygen atom of molecular oxygen into the substrate. As with many other flavoprotein aromatic hydroxylases, the additional hydroxy group was introduced in the ortho-position with respect to the existing hydroxy group. Although the enzyme regioselectively hydroxylated only the C-3 position of the substrate, it had a relaxed substrate specificity with respect to the hydrophobic side chain, since it was able to hydroxylate various 2-alkyl-and 2-arylphenols. As a general feature, substrates of the 2-hydroxybiphenyl 3-monooxygenase had a 2-R-phenol structure, where R is a hydrophobic carbon moiety. The substrates of the 2-hydroxybiphenyl 3-monooxygenase were different from those of other flavoprotein aromatic hydroxylases acting on substituted phenols. An exception was 2-methylphenol, which is a common substrate for several phenol 2-hydroxylases (34) as well as for 2,4-dichlorophenol 6-hydroxylase from Acinetobacter sp. (33). The values of the apparent Michaelis constants for the substrates (Table IV) were similar to the ones reported for the substrates of 2,4-dichlorophenol 6-hydroxylase (33,35) and phenol 2-hydroxylase (36). Many aromatic hydroxylases have a preference for NADPH as the electron donor for the reduction of the flavin molecule, whereas some can utilize NADH (31). 2-Hydroxybiphenyl 3-monooxygenase utilized NADH as well as NADPH, and the values of the apparent Michaelis constants for the two substrates, 2-hydroxybiphenyl and 2,2Ј-dihydroxybiphenyl, were similar with either reduced pyridine nucleotide. However, the K m values for NADPH were markedly higher than those for NADH. A relaxed cofactor specificity was demonstrated for other flavoprotein aromatic hydroxylases, which oxidize NADH and NADPH with similar efficiency (37). The binding of 2-hydroxybiphenyl to 2-hydroxybiphenyl 3-monooxygenase enhanced the oxidation of NADH by the enzyme. The factor of this stimulation in rate is generally on the order of 10 3 to 10 4 with flavoprotein hydroxylases (38). For 2-hydroxybiphenyl 3-monooxygenase with 2-hydroxybiphenyl as the substrate, we only observed a 20-fold stimulation. Therefore, 2-hydroxybiphenyl 3-monooxygenase was a rather strong NADH oxidase in the absence of the aromatic substrate. Uncoupling of oxygen reduction from the hydroxylation reaction, which is observed with substrate effectors as well as with non-substrate effectors, leads to the formation of hydrogen peroxide and is a common feature of flavin monooxygenases (38). With 2-hydroxybiphenyl 3-monooxygenase, partial uncoupling occurred in the presence of each substrate, but the extent of uncoupling varied with the substrate. The same phe-  3 0 NaCl (10 mM) 3 6 NaCl (50 mM) 7 5 NaCl (100 mM) 8 9 nomenon was reported for the substrates of phenol 2-hydroxylase from Trichosporon cutaneum (36). However, the amount of hydrogen peroxide produced in the presence of different effectors was not always consistent with the expected amount that we calculated from the disappearance of oxygen and substrate. We showed that in the case of the enzyme substrates 2-hydroxybiphenyl and 2-sec-butylphenol, the products of the 2-hydroxybiphenyl 3-monooxygenase reaction also contributed to the increase of hydrogen peroxide through uncoupling (Table  III). A striking feature of the 2-hydroxybiphenyl 3-monooxygenase was the fact that the product of the reaction with 2-hydroxybiphenyl, 2,3-dihydroxybiphenyl, was a remarkably efficient non-substrate effector. 2,3-Dihydroxybiphenyl significantly stimulated the consumption of NADH and oxygen without undergoing hydroxylation. Furthermore, 2,3-dihydroxybiphenyl inhibited the conversion of 2-hydroxybiphenyl by mixed inhibition. In the case of 4-hydroxybenzoate 3-hydroxylase, 3,4dihydroxybenzoate, the product of the hydroxylation of 4-hydroxybenzoate, acts also as a non-substrate effector (39). Product inhibition as well as the formation of toxic amounts of hydrogen peroxide by 2,3-dihydroxybiphenyl probably did not significantly impair growth of strain HBP1. The 2,3-dihydroxybiphenyl dioxygenase, the next enzyme of the 2-hydroxybiphenyl degradation pathway, has 1000-fold higher activity than the monooxygenase (3) and presumably keeps the intracellular pool of 2,3-dihydroxybiphenyl at a negligibly low level.
Treatment of 2-hydroxybiphenyl 3-monooxygenase with pHMB, an effective modifier of cysteine residues, led to a rapid inactivation of the enzyme. The loss of enzyme activity could be reversed by the addition of dithiothreitol. With pHMB concentrations smaller than 100 M, the enzyme maintained residual activities. This indicates that only partial modification of the enzyme occurred. The substrate, 2-hydroxybiphenyl, partially protected 2-hydroxybiphenyl 3-monooxygenase from inactivation by the thiol reagent. The presence of 2,3-dihydroxybiphenyl had no effect on the inactivation reaction. Because 2-hydroxybiphenyl did not completely protect the enzyme from inactivation, it is not very likely that it masked a cysteine residue essential for catalysis. For 4-hydroxybenzoate 3-hydroxylase, chemical modification studies with different thiol reagents revealed that none of the five cysteine residues present in the enzyme is crucial for the enzymatic activity (40). Analysis of the three-dimensional structure of 4-hydroxybenzoate 3-hydroxylase confirmed the absence of cysteine residues in the active center (41). Selective cysteine-serine replacements in 4-hydroxybenzoate 3-hydroxylase also confirmed that the cysteine residues are not essential for catalysis and that mercuration of Cys-211 impairs binding of the aromatic substrate (42).
The inhibition of the monooxygenase by chloride ions was noncompetitive. Monovalent ions are noncompetitive inhibitors of melilotate hydroxylase (32), but on the other hand they are uncompetitive inhibitors of phenol 2-hydroxylase (43) and 3-hydroxyphenylacetate 6-hydroxylase (37). Heavy metal ions, which are known to build complexes with the flavoquinone form of free flavins (44), immediately abolished activity of 2-hydroxybiphenyl 3-monooxygenase.
The NH 2 -terminal amino acid sequence of 2-hydroxybiphenyl 3-monooxygenase contained the GXGXXG sequence, which is the core of a fingerprint of 11 amino acids at crucial positions within a stretch of 29 -32 amino acid residues (26). Proteins that contain a match to the fingerprint fold into an ADPbinding ␤␣␤-unit involved in FAD and NAD binding. The putative fingerprint regions of both 2-hydroxybiphenyl 3-monooxygenase and 4-hydroxybenzoate 3-hydroxylase did not exactly match the requested sequence. They deviated in the last two amino acid positions from the predicted fingerprint and did not have the obligatory acidic residue at position 32. However, the presence of this acidic residue seems not to be absolutely necessary, since 4-hydroxybenzoate 3-hydroxylase contains a ␤␣␤-fold in the FAD-binding domain at the NH 2 terminus of the enzyme despite the lack of this residue (41). The NH 2 -terminal sequence of 2-hydroxybiphenyl 3-monooxygenase had a high degree of sequence identity to that of 2,4dichlorophenol 6-hydroxylase from R. eutropha (28) and phenol 2-hydroxylase from Pseudomonas sp. strain EST1001 (27). Furthermore, 2-hydroxybiphenyl 3-monooxygenase shared many molecular and catalytic properties with the 2,4-dichlorophenol 6-hydroxylase from Alcaligenes sp. (33), which is also a tetramer composed of identical subunits with a relative molecular mass of 63 kDa.
We purified and described a novel flavin monooxygenase, which is able to hydroxylate a large number of 2-alkyl-and 2-arylphenols. It shares many similarities with other enzymes of the group of external flavoprotein aromatic hydroxylases but displays an exceptionally wide substrate spectrum for regioselective hydroxylation.