Characterization of active recombinant his-tagged oxygenase component of Comamonas testosteroni B-356 biphenyl dioxygenase.

Biphenyl (BPH) dioxygenase oxidizes BPH to 2,3-dihydro-2,3-dihydroxybiphenyl in Comamonas testosteroni B-356. The enzyme comprises a two-subunit iron-sulfur protein (ISPBPH), a ferredoxin FERBPH, and a ferredoxin reductase REDBPH. REDBPH and FERBPH transfer electrons from NADH to an Fe-S active center of ISPBPH which activates molecular oxygen for insertion into the substrate. In this work B-356 ISPBPH complex and its alpha and beta subunits were purified from recombinant Escherichia coli strains using the His-bind QIAGEN system. His-tagged B-356 ISPBPH construction carrying a single His tail on the N-terminal portion of the alpha subunit was active. Its major features were compared to the untagged enzyme. In both cases, the native form is an alpha3beta3 heteromer, with each alphabeta unit containing a [2Fe-2S] Rieske center (epsilon455 = 8,300 M-1 cm-1) and a mononuclear Fe2+. Although purified His-tagged alpha subunit showed the characteristic absorption spectra of Rieske-type protein, reassociation of this enzyme component and His-tagged beta subunit to reconstitute active ISPBPH was weak. However, when His-tagged alpha and beta subunits were reassembled in vitro in crude cell extracts from E. coli recombinants, active ISPBPH could be purified on Ni-nitrilotriacetic acid resin.

Biphenyl dioxygenase (BPH dox) 1 catalyzes the first step of the bacterial BPH degradation pathway. The enzyme introduces molecular oxygen into the ortho-meta positions on one of the aryl rings to generate 2,3-dihydro-2,3-dihydroxybiphenyl. In a previous study (1), we have reported the purification and characterization of Comamonas testosteroni strain B-356-BPH dox system. The enzyme comprises three components which are: the terminal oxygenase, an iron-sulfur protein (ISP BPH ) made up of an ␣-subunit (M r ϭ 51,000) and a ␤-subunit (M r ϭ 22,000), encoded by bphA and bphE, respectively; a ferredoxin (FER BPH , M r ϭ 12,000) encoded by bphF; and a ferredoxin reductase (RED BPH , M r ϭ 43,000) encoded by bphG. FER BPH and RED BPH were found to be involved in electron transfer from NADH to ISP BPH (1). The Rieske center of the oxygenase component is then believed to receive the electron and pass it to a mononuclear Fe 2ϩ which activates molecular oxygen for in-sertion into the substrate (2,3).
The ISP BPH component has been purified from BPH-induced bacteria of strain B-356 (1) and from Pseudomonas sp. LB400 (4). Since active purified FER BPH and RED BPH were difficult to obtain from cell extracts of parental strains (1,4), these enzyme components were purified from Escherichia coli recombinant clones using the His-bind QIAGEN system (1). Both His-tagged FER BPH and His-tagged RED BPH from strain B-356 were able to transfer electrons from NADH to B-356-ISP BPH . However, purification of the individual active ISP BPH ␣ and ␤ subunits has not yet been reported.
Understanding the various factors that contribute to the strain selectivity pattern toward substrate should help the modeling of new mutants with increased ability to degrade BPH analogs such as polychlorinated biphenyls. The BPH dox reactivity pattern is a major determinant affecting the performance of bacterial polychlorinated biphenyl degraders. The BPH dox-congener selectivity pattern is partly affected by the position of attack on the aromatic ring. For example, the capacity of Pseudomonas sp. LB400 to preferentially degrade the ortho-substituted polychlorinated biphenyl congeners was attributed to its ability to oxygenate BPH at ring positions 3 and 4 in addition to 2 and 3. Haddock et al. (5) have recently shown that partially purified LB400-BPH dox was able to attack 2,2Ј,5,5Ј-tetrachlorobiphenyl (for which there is no available ortho-meta sites for oxygenation) in a 3,4-position suggesting that the same enzyme catalyzes both type of attacks. Using site-directed mutagenesis, Erickson and Mondello (6) have provided evidence that minor structural differences of the ISP BPH ␣ subunit component are responsible for major changes in the reactivity pattern of strain LB400-BPH dox.
Although the substrate selectivity of strain B-356 distinguishes it from strain LB400, we have recently shown that strain B-356 dox is also able to oxygenate BPH at both 2,3-and 3,4-positions (1).
Other studies have suggested that additional determinants are affecting the reactivity pattern of various strains toward BPH analogs. For example, Hirose et al. (7) have found that the BPH dox components of Pseudomonas pseudoalcaligenes KF707 are to some extent, interchangeable with those of the Pseudomonas putida F1-toluene dioxygenase system encoded by the tod operon. Several of the recombinant enzymes tested were found to manifest new degradative abilities that neither of the original enzymes possessed. Their results suggest that the oxygenase ␤ subunit is specific to the system to which it belongs. However, to date, the role of the ISP BPH ␤ subunit in enzyme activity remains unknown.
At this time, it is clear that further investigation is required at the molecular level to identify the role of the BPH dox components regarding catalytic activity and substrate selectivity. However, because most of these enzyme components are unstable and cannot sustain the large number of manipulations needed for their purification, new approaches are needed to obtain purified active enzyme. In this work we have identified some of the major features of purified strain B-356 Histagged ISP BPH and of individual His-tagged ISP BPH subunits. We have also found that the His-tagged ISP BPH subunits can be reassembled in vitro to produce an active oxygenase component.
Plasmid DNA from E. coli, restriction endonuclease digestions, ligations, agarose gel electrophoresis, and transformation of E. coli cells were done according to protocols described by Sambrook et al. (9). Polymerase chain reactions (PCR) were performed using Pwo DNA polymerase following the method given by Boehringer Mannheim.
Purification Protocols-The procedures to obtain purified preparations of strain B-356-BPH dox ISP components from BPH-induced B-356 cells and to obtain purified His-tagged FER BPH and RED BPH from recombinant E. coli cells have already been described (1). The Histagged B-356-ISP BPH complex or individual His-tagged ISP BPH subunits were expressed in appropriate recombinant strains of E. coli M15[pREP4] and SG13009[pREP4] using a protocol similar to the one described for His-tagged RED BPH (1). The oligonucleotides used for PCR were chosen on the basis of the known DNA sequences of the genes to be amplified (11) and they were as follows: oligonucleotide I (BamHI), 5Ј-CGGGATCCGATGAGTTCGACTATGA-3Ј; oligonucleotide II (KpnI), 5Ј-GGGGTACCCCTCAGGGTTTGAGCGT-3Ј; oligonucleotide III (BamHI), 5Ј-CGGGATCCGAGATTATCCACTCCCT-3Ј; oligonucleotide IV (KpnI), 5Ј-GGGGTACCCCTCAAAAGAACACGCT-3Ј. To obtain purified His-tagged B-356-ISP BPH , the coding region of B-356-bphAE was PCR-amplified from a cloned DNA fragment using the oligonucleotides I and IV. To obtain the individual His-tagged ␣ and ␤ subunits of B-356-ISP BPH , the coding region of bphA or bphE were amplified using the oligonucleotides I and II, and III and IV, respectively.
The PCR products were digested with BamHI and KpnI. A 1.9kilobase DNA fragment (for bphAE), a 1.3-kilobase DNA fragment (for bphA), and a 0.6-kilobase DNA fragment (for bphE) containing the entire coding sequences were isolated and cloned into the compatible sites of pQE31. Constructions were such that the His tail added 13 amino acids (MRGSHHHHHHTDP) to the protein at its N-terminal portion. When both the ␣ and ␤ subunits were produced together in the same clone, the His tail was attached to the ␣ subunit only.
Protein Characterization-SDS-PAGE gels were developed according to Laemmli (12). Proteins were stained with Coomassie Brilliant Blue (9). M r values of native ISP BPH , His-tagged ISP BPH , and individual ␣ and ␤ subunits were determined by HPLC using a Perkin-Elmer Series 3 chromatograph and a Waters Protein Pak 300 SW column (7.8 ϫ 300 mm). The column was eluted at 1 ml/min with 100 mM potassium phosphate buffer, pH 7.0. The UV detector was a Perkin-Elmer LC65T and it was set at 280 nm. The column was calibrated with catalase (M r ϭ 232,000), bovine serum albumin (M r ϭ 67,000), albumin (M r ϭ 46,000), and RNase A (M r ϭ 13,700).
The iron content of protein was evaluated by the ferrozine colorimetric method modified by Batie et al. (13). Acid-labile sulfide was determined according to the method of Fogo and Popowsky (14) as modified by King and Morris (15). Protein concentrations were estimated by the methods of Lowry (16) and Bradford (17) using bovine serum albumin as standard.
BPH Dox Assays-Enzyme assays were performed in a 200-l volume as described previously (1), except that the reaction was stopped by adding 400 l of methanol. After centrifugation for 30 s in a microcentrifuge, 50 l of the supernatant were injected into a Hewlett-Packard ODS Hypersil II (5 m) reverse phase column (4 mm ϫ 25 cm) that had been equilibrated with water:methanol:acetonitrile (50:25:25). The column was eluted for 5 min at 1 ml/min with the same solvent system, followed by a 3-min linear gradient to methanol:acetonitrile (60:40) and then eluted for 5 min with the final solvent mixture. The activity was evaluated from measurement of substrate disappearance and metabolite production. They were detected using a Perkin-Elmer LC95 UV/ visible detector set at 306 nm for 2,3-dihydro-2,3-dihydroxybiphenyl or 254 nm for biphenyl. Confirmation of the identity of the metabolites was obtained by gas chromatographic-mass spectrometric analysis using protocols described previously (1). Kinetic parameters, K m and V max , were estimated by varying BPH concentrations between 125 and 5000 M in the assay system described above with all other parameters being identical.

Comparative Features of His-tagged B-356-ISP BPH with B-356-ISP BPH -The
His-tagged B-356-ISP BPH construction was such that a single (His) 6 tail was added to the N-terminal portion of the ISP BPH ␣ subunit. Both subunits were expressed in isopropyl-1-thio-␤-D-galactopyranoside-induced E. coli recombinant cells as demonstrated by SDS-PAGE. Moreover, the presence of a single His tail was sufficient to obtain a purified preparation comprising both the His-tagged ␣ and ␤ subunits through a one-step purification performed on the Ni-nitrilotriacetic acid resin. This suggests that strong bonds are involved in the association between the subunits in the native protein and that the His tail of the ␣ subunit did not significantly affect this association. Moreover, as shown below, purified Histagged ISP BPH was active in the BPH dox assay with added His-tagged FER BPH and His-tagged RED BPH .
The yield of purified His-tagged ISP BPH was in the order of 1.5 mg/g cell paste (wet weight), which is slightly lower then the 3 mg/g cell paste obtained for ISP BPH from B-356. The best preparations showed two major bands on SDS-PAGE (Fig. 1). Their M r values were estimated to be 53,600 and 25,200, which correspond to the M r values of the ISP BPH ␣ (plus the His tail addition) and ␤ subunits, respectively. The M r of the native ISP BPH and His-tagged ISP BPH were estimated by HPLC gel filtration and they were found to be of 234,000 and 186,000, respectively. These values indicate that the native conformation of B-356 ISP BPH is ␣ 3 ␤ 3 and corroborate previously published data obtained for strain LB400-ISP BPH (4). The lower native M r of His-tagged ISP BPH remains unexplained. However, the difference in migration in the gel could be attributed to an electrostatic interaction between the His tail and the diol residues of the stationary phase.
The purity of different preparations of His-tagged ISP BPH were evaluated by scanning the SDS-PAGE gels, and these values were used as a correction factor to evaluate the actual protein concentration of His-tagged ISP BPH . Using these data, the average ⑀ 455 value was calculated to be 8,300 M Ϫ1 cm Ϫ1 (based on the determination on seven preparations). Purified His-tagged ISP BPH contained 2.6 Ϯ 0.5 iron and 1.9 Ϯ 0.1 sulfur per mol of ␣/␤ heterodimer. Therefore, the number of iron atoms per mol of protein for His-tagged ISP BPH preparations was close to the theoretical value of 3 iron that should be expected for a [2Fe-2S] Rieske-type center carrying a mononuclear Fe 2ϩ . As further evidence that the enzyme had retained most of its mononuclear Fe 2ϩ , we found that BPH dox specific activity of His-tagged ISP BPH preparations increased by only 30 -40% when an excess of iron was added to the reaction mixture.
On the other hand, the ISP BPH preparations obtained from B-356 cells were somewhat more altered than the His tail preparations. For example, based on the ⑀ 455 value of 8,300 M Ϫ1 cm Ϫ1 , we calculated that on average, 85% of the protein in the preparations had retained an intact Rieske center. This estimation was confirmed by the observation that these ISP BPH preparations contained 1.7 Ϯ 0.2 iron and 1.7 Ϯ 0.3 sulfur atom/mol of ␣/␤ heterodimer. It thus appears that the ISP BPH preparations obtained from strain B-356 had lost all their mononuclear iron and that a portion of the Rieske center was destroyed.
ISP BPH remained active for months at Ϫ70°C. However, the His-tagged ISP BPH was not as stable. We observed that older preparations of His-tagged ISP BPH that were reduced in activity showed a different subunit association pattern where the preparation contained ␣ monomer and ␤ homodimer with minor amounts of ␣ 3 ␤ 3 . On the other hand, ␣ 3 ␤ 3 heterodimer was the only form observed in the case of B-356 ISP BPH . This suggested that ␣ 3 ␤ 3 is the only active form of ISP BPH .
We found that a 20-min preincubation of His-tagged ISP BPH with 5 mM dithiothreitol on ice can restore activity of older preparations. The same phenomenon was observed with FER BPH . However, it is not clear whether this reactivation occurred by restoring the Rieske center or by changing some other feature of the molecule. On the other hand, fresh preparations were at their optimal level of activity when the required supplemental mononuclear Fe 2ϩ was added to the assay. Addition of dithiothreitol did not significantly affect their activity.
Kinetic parameters of the recombinant versus parental protein were compared when ISP BPH in the reaction mixture containing His-tagged RED BPH and His-tagged Fer BPH was replaced by His-tagged ISP BPH . For both preparations, 40°C was the optimal temperature for BPH dox activity and the reaction was optimal at pH 5.5-6.0. When the proportion of the enzyme components were varied in the assay, the reaction was optimal for equimolar amounts of each one. Under the conditions described under "Experimental Procedures," the K m and V max for the BPH dox activity were, respectively, 94 M and 1.5 nmol min Ϫ1 g Ϫ1 for ISP BPH and 100 M and 1.9 nmol min Ϫ1 g Ϫ1 for His-tagged ISP BPH .
In previous work, gas chromatography-mass spectroscopic analysis of metabolites produced from B-356-BPH dox reaction suggested that both 2,3-and 3,4-dihydro-dihydroxybiphenyl were produced from BPH (1). Both metabolites were also produced when B-356-ISP BPH was replaced by the recombinant ISP BPH in the BPH dox assay (results not shown).
Purification and Characterization of Individual His-tagged ISP BPH Subunits-Individual His-tagged ␣ and His-tagged ␤ subunits were expressed in E. coli clones in appropriate constructs. The level of expression of the two subunits was very high as observed on SDS-PAGE gels of urea-solubilized cells (data not shown). However, the yield of purification of ␣ and ␤ subunits was low (0.4 and 0.2 mg/g cell paste (wet weight), respectively), presumably because of the association of a por-tion of the proteins with inclusion bodies or membrane fraction. A similar observation was reported by Suen and Gibson (18) for the expression of Pseudomonas sp. NCIB 9816 -4 naphthalene dioxygenase in E. coli.
The purified proteins were detected on SDS-PAGE (Fig. 1). Gel filtration HPLC analysis showed that the native form of the purified His-tagged ␣ subunit was a monomer (M r ϭ 44,000), while the His-tagged ␤ subunits had a tendency to join together to produce homodimers plus large conglomerates. The UV visible absorbance spectra of purified ␣ subunit (but not ␤ subunit) was similar to the spectra of ISP BPH and His-tagged ISP BPH (Fig. 2). There was a slight shift of the peaks with maxima at 335 and 443 nm instead of 323 and 455 nm. After reduction with excess dithionite under aerobic conditions, the spectra resembled that of other reduced Rieske-type proteins. Assuming that the ⑀ 443 of this protein is equal to the ⑀ 455 of the ISP BPH complex, we have estimated that on average 60% of the His-tagged ␣ subunit of purified preparations were carrying an intact Rieske center. This is supported by the observation that the ratio of iron and acid-labile sulfur per mol of His-tagged ␣ subunit was 1.3 Ϯ 0.4 iron and 1.3 Ϯ 0.1 sulfur. Therefore, part of the Rieske center was either denatured during purification of the His-tagged ␣ subunit or was not formed in the E. coli cells. The altered enzyme subunit was not restored by adding dithiothreitol or iron. Contrary to the naphthalene dioxygenase-␤ subunit (19), ISP BPH -␤ subunit showed a broad peak between 300 and 500 (Fig. 2B). This peak was not modified under reduced conditions. When preparations of individually purified subunits were combined in vitro and tested in the BPH dox assay immediately or after 18 h of preincubation at 4°C, the activity recorded was only about 1% of the activity obtained with His-tagged ISP BPH preparation that had been assembled in vivo. However, when the crude cell extracts were mixed instead of the purified proteins, BPH dox activity was restored to a higher level ( Table I). The activity increased with time when the mixture was preincubated for 18 h at 4°C (results not shown) as it had been observed for the naphthalene dioxygenase (18). In order to find out which protein was deactivated during the purification, we mixed a crude E. coli lysate containing one His-tagged subunit with a purified preparation of the other His-tagged subunit. Using this approach, only trace amounts of activity were obtained when purified ␣ subunit was added to the lysate but a fair level of activity was observed when purified ␤ subunit was used (Table I). On the other hand, when the crude cell extracts containing the individual subunits were mixed and incubated for 18 h at 4°C and then purified together on Ni-nitrilotriacetic acid resin, the resulting His-tagged ISP BPH preparation was active (Table I). These results suggested that the ␣ subunit was denatured when it was purified alone. However, its prior association with the ␤ subunit protects the enzyme against deactivation. These data clearly showed that ␣ and ␤ subunits of ISP BPH can associate in vitro to restore the active BPH dox.

DISCUSSION
In this work we have reported some of the major features of strain B-356 His-tagged ISP BPH and compared them to B-356 ISP BPH . Most of the data presented here, including the ␣ 3 ␤ 3 arrangement of subunits are similar to the recently published properties of LB400 oxygenase component. It is not surprising because both enzymes are structurally closely related (11). At this time, however, we cannot explain the significant difference between the ⑀ 455 value of 8,300 M Ϫ1 cm Ϫ1 we have determined for B-356-ISP BPH and the one that was reported for strain LB400 ISP BPH .
Our data show that the main features that characterize B-356 ISP BPH carrying a His tail on the ␣ subunit are very similar to those found for the untagged parental protein obtained from strain B-356 cells. Moreover, because the steps required for purification are milder, the His-tagged enzyme component preparations were less altered, as reflected by their UV visible spectral data, the iron and acid labile sulfur content, and the specific activity of the enzyme.
Hydroxylating dioxygenases are multicomponent enzymes that catalyze the transfer of electrons from an electron donor, usually a reduced nicotinamide dinucleotide, to the hydroxylating center of the oxygenase component. Based on the number of components and on the number and type of [2Fe-2S] centers involved in electron transfer, hydroxylating oxygenases have been subdivided into three classes (2,3). BPH dox belongs to the class IIB dioxygenases which also includes benzene dioxygenase and toluene dioxygenase. These enzyme systems comprise three components, a reductase containing a flavin cofactor, a Rieske-type ferredoxin protein, and a terminal oxy-genase, which is a two-subunit iron-sulfur protein containing a Rieske-type [2Fe-2S] center. Absorbance spectra of purified B-356-BPH dox and LB400-BPH dox terminal oxygenases were typical of Rieske-type proteins (1,4). Moreover, sequence analysis has identified a conserved sequence C-R-H-(aa) 17 -C-S-Y-Hof a Rieske center on the ␣ subunit of both strains ISP BPH (11,20,21). However, confirmation of the presence of the Rieske center on the ␣ subunit of class IIB oxygenases has never been established. Suen and Gibson (19) have obtained a purified preparation of the class III naphthalene dioxygenase terminal oxygenase-␣ subunit. However, their purified protein was denatured and did not show a typical Rieske-type absorbance spectra. Therefore, use of His-tail protein provided, for the first time, spectral evidence that the Rieske center is located on the ␣ subunit of the terminal oxygenase.
In spite of the fact that approximately 60% of the ISP BPH ␣ subunit in purified preparations had retained an intact Rieske center with [2Fe-2S], these preparations were only weakly active when mixed with purified or crude preparations of ␤ subunit. Structural modifications during the purification process to prevent reassociation with the ␤ subunit cannot be excluded. This is also supported by the observation that these preparations did not show ␣ 3 ␤ 3 heteromers on gel filtration (results not shown). The fact that the ␣ subunit in crude cell extracts can combine with exogenous ␤ subunit to generate an active complex might suggest the presence in cell extract of proteins or other constituents that interact with the ␣ subunit to protect their folding. However, further study is needed to understand the reasons for impaired activity.
The sequence of events that occurs at the active site of aryl dioxygenases to bind and activate the molecular oxygen and to attach it to the substrate is still unknown. The [2Fe-2S] Riesketype cluster found in the terminal oxygenase component is presumed to involve at least two Cys and two His residues inside this domain (2,13) to coordinate the iron atoms. In addition, other Tyr and His residues that are located farther away inside the protein are presumed to be involved in the coordination of a mononuclear Fe 2ϩ which is required for oxygen binding to the enzyme (3,13,22).
Batie et al. (13) have clearly shown that Pseudomonas cepacia phthalate dioxygenase contains two iron atoms coordinated to Cys and His residues. They also demonstrated that the presence of an additional mononuclear Fe 2ϩ is required for enzyme activity. Suen and Gibson (19) have also reported the presence of 6 iron and 4 acid labile sulfide atoms per mol of purified Pseudomonas sp. NCIB 9816 -4 naphthalene dioxygenase oxygenase component, where this enzyme was found to be an ␣ 2 ␤ 2 enzyme. Haddock and Gibson (4) have inferred the presence of a third mononuclear iron in LB400-BPH dox terminal oxygenase from the observation that Fe 2ϩ stimulated the enzyme activity. Because of the good quality of the B-356 His- a BPH dox activity was evaluated by measuring the meta-cleavage metabolite at 434 nm when excess purified B2,3D and B1,2O were added to the reaction mixture as described under "Experimental Procedures." The activity is reported per mg of protein in the preparation containing the ␣ subunit.
b Crude lysate of E. coli cells producing either His-tagged ␣ or His-tagged ␤ were combined, incubated 18 h at 4°C, then purified on Ni-nitrilotriacetic acid resin to obtain a purified preparation of a reconstituted His-tagged ISP BPH . tagged ISP BPH preparations, our data provides evidence that this enzyme contains 3 iron atoms and 2 acid labile sulfur atoms per mol of ␣␤ heterodimer.
It as been suggested that the ␤ subunit of the aryl dioxygenase oxygenase component is involved in substrate recognition. Hirose et al. (7) have used various hybrid clones of P. pseudoalcaligenes strain KF707-BPH dox genes and P. putida strain F1-toluene dox genes, to show the importance of bphA2 (bphE) in determining substrate specificity. They made the observation that bphA2 could not be replaced by todC2 to obtain functional BPH dox, but that bphA1 could be replaced by todC1. Whatever the role of the ␤ component, it is certainly not carrying any Rieske center. This is confirmed by sequence analysis of the gene (11) and from the UV visible spectra of purified His-tagged ISP BPH ␤ subunit. However, its association to the ␣ subunit is essential to protect the activity of the latter.
Our data clearly show that the assembly of the ISP BPH subunits can occur in vitro. We have reported for the first time the purification of an in vitro assembled ISP terminal oxygenase. Because the recombinant His-tagged components of the B-356-BPH dox have retained all properties of the parental protein, including the activity, the QIAGEN purification system appears, therefore, quite promising to further explore the various features of bacterial dioxygenases. Particularly, it opens the possibility of comparing reconstituted hybrid oxygenases containing subunits belonging to terminal oxygenases of different origin. Therefore, this system introduces a very useful new tool to further investigate the major features that distinguish the various aryl dioxygenase among them. Ongoing work in our laboratory is intended to exploit this.