Putidaredoxin reductase, a new function for an old protein.

Properties of recombinant wild type (WT) and six-histidine tag-fused (His(6)) putidaredoxin reductase (Pdr), a FAD-containing component of the soluble cytochrome P450cam monooxygenase system from Pseudomonas putida, have been studied. Both WT and His(6) Pdr were found to undergo a monomer-dimer association-dissociation and were partially present as an NAD(+)-bound form. Although molecular, spectral, and electron transferring properties of recombinant His(6) Pdr to artificial and native electron acceptors were similar to those of the WT protein, the presence of eight additional C-terminal amino acid residues, Pro-Arg-His-His-His-His-His-His, had a crucial effect on the enzyme interaction with oxidized pyridine nucleotide. Under anaerobic conditions, NAD(+) induced in His(6) Pdr spectral changes indicative of flavin reduction and formation of the charge transfer complex between the reduced FAD and NAD(+). The reaction proceeded considerably faster in the presence of free histidine and thiol-reducing agents, such as dithiothreitol and reduced glutathione. In the presence of any of these three reagents, NAD(+) was capable of inducing reduction of the flavin in WT Pdr. Free thiol groups were identified as an internal source of electrons in the enzyme. The results showed that WT and His(6) Pdr were able to function as NAD(H)-dependent dithiol/disulfide oxidoreductases catalyzing both forward and reverse reactions, NAD(+)-dependent oxidation of thiols, and NADH-dependent reduction of disulfides. This function of the flavoprotein can be dissociated from electron transfer to putidaredoxin. Similarity of Pdr to the enzymes of the glutathione reductase family is discussed.

A FAD-containing putidaredoxin reductase (Pdr) 1 catalyzes the transfer of electrons from NADH to an iron-sulfur protein, putidaredoxin (Pdx), in the cytochrome P450cam monooxygenase system from Pseudomonas putida (1). Two-electron reduction of FAD with NADH is followed by the transfer of two single electrons from the flavin to Pdx. From Pdx, electrons are finally transferred to the terminal oxygenase cytochrome P450cam (CYP101) that catalyzes hydroxylation of camphor consuming two electrons and molecular oxygen per reaction cycle.
Pdr has a molecular mass of 45.6 kDa and contains one tightly bound FAD per single polypeptide. Redox measurements, where sodium dithionite was used as a reductant, demonstrated that the midpoint potential of Pdr is Ϫ285 mV (2). The reduction of Pdr by NADH is a stoichiometric process that results in the formation of a long-wavelength charge transfer complex without the appearance of a semiquinone intermediate (3). The reduction of Pdr with NADPH is approximately 3 orders of magnitude slower than with NADH (4). No charge transfer band could be detected when the flavoprotein was reduced with either NADPH or sodium dithionite. Using NADPH as a reductant, the midpoint potential of Pdr was calculated to be Ϫ315 mV (4). Titration of Pdr with NADPH in the presence of excess NAD ϩ resulted in a stoichiometric reduction. This was explained as a result of the very tight association of NAD ϩ with the reduced FAD and formation of the charge transfer complex that was able to shift the equilibrium of the reduction reaction toward product.
In P450cam monooxygenase, the reactions of NADH oxidation and Pdx reduction catalyzed by Pdr are tightly coupled. Investigation of the role of NAD ϩ /NADH in the catalytic reduction of Pdx has demonstrated that the pyridine nucleotide complex of Pdr is important in the enhancement of the reduction of Pdx (5). NAD ϩ was shown to be essential for the kinetically favored oxidation of Pdr from the two-electron reduced form to the semiquinone by oxidized Pdx. When Pdr was photoreduced in the presence of NAD ϩ , it transferred the first electron to Pdx at a rate that could account for catalytic turnover. In contrast, the rate of Pdx reduction by Pdr that was photoreduced in the absence of NAD ϩ was 3 orders of magnitude lower.
Due to instability, a semiquinone form of Pdr could not be detected and characterized by conventional spectrophotometric techniques. We have recently utilized laser flash photolysis techniques to produce and investigate the electron transferring properties of one-electron reduced species of the flavoprotein (6). Upon flash-induced reduction by 5-deazariboflavin semiquinone, Pdr was found to form a blue neutral FAD semiquinone (FADH ⅐ ). FADH ⅐ was unstable and partially disproportionated into fully oxidized and fully reduced flavin. The rate of FAD semiquinone disproportionation was enhanced by NAD ϩ . It was established that FADH ⅐ was capable of transferring an electron to Pdx with the rate constant that can account for catalytic turnover (6).
Quite unexpectedly, our laser flash photolysis experiments revealed that under anaerobic conditions and in the presence of NAD ϩ , the flavin in the recombinant Pdr molecule, expressed with the C-terminal six-histidine tag (His 6 Pdr), became reduced without any external source of electrons. This indicates that, in addition to FAD, there must be another redox center(s) in Pdr. The fact that Pdr belongs to a class of bacterial oxygenase-coupled NADH-dependent ferredoxin reductases that have a structural fold similar to that of the enzymes of the glutathione reductase (GR) family (7) suggests that the second redox center in Pdr is likely to involve sulfhydryl groups. Although Pdr has six cysteine residues, it does not have putative disulfide redox centers encoded by CXXC or CXXXXC motifs, characteristic features of pyridine nucleotide-disulfide oxidoreductases. Despite the low amino acid sequence similarity between oxygenase-coupled NADH-dependent ferredoxin reductases and GR-family enzymes, these proteins are thought to have evolved from a common ancestor but, in the course of evolution, have lost (or GR family enzymes have acquired) the cysteine residues essential for the catalysis (7). In the present work, we demonstrate that Pdr is a bifunctional enzyme that, in addition to catalyzing electron transfer to Pdx, can function as NAD(H):dithiol/disulfide oxidoreductase. Lacking traditional disulfide redox centers, Pdr seems to be unique in this class of enzymes.
Cloning and Purification of Six-histidine Tag-fused Pdr-The sixhistidine tag-fused Pdr (His 6 Pdr) was cloned by PCR using the pET-Pdr plasmid as a template. The 5Ј-nucleotide primer was synthesized with an overhanging NdeI restriction site. The 3Ј-oligonucleotide encoded Lys-Ala-Glu-Leu-Ser-Ser-Ala-Pro-Arg-His-His-His-His-His-His for part of the C terminus followed by the ␣-thrombin cleavage site Ala-Pro-Arg, six histidines, and an EcoRI restriction site (GCGAAT-TCTCAGTGATGGTGATGGTGATGACGCGGGGCACTACTCAGTTC-AGCTTT). The NdeI/EcoRI fragment of the synthesized 1.3-kilobase fragment of DNA encoding His 6 Pdr was ligated into the unique NdeI/EcoRI sites of the expression plasmid pET, and the sequence was confirmed.
Transformation, cell growth, and cell lysis were performed as described previously for WT Pdr (6). The lysate was loaded on Ni 2ϩnitrilotriacetic acid resin equilibrated in 50 mM sodium phosphate, pH 7.6, containing 0.1 mM phenylmethylsulfonyl fluoride (buffer A) and washed overnight with 50 mM sodium phosphate buffer, pH 7.6, containing 100 mM NaCl, 10 mM ␤-mercaptoethanol, 15 mM imidazole, and 10% glycerol (buffer B). The protein was eluted with buffer B containing 80 mM imidazole. Fractions with the ratio of A 280 /A 454 less than 8.5 were combined, concentrated, and loaded on a Sephacryl S-100-HR column. His 6 Pdr was purified to homogeneity on phenyl-Sepharose (Amersham Biosciences) using a linear 30 to 0% ammonium sulfate gradient in buffer A. The protein with a ratio of A 280 /A 454 Ͻ 7.0 was used in the experiments.
Mass Spectroscopy Analysis-Molecular mass was measured by means of a matrix-assisted laser desorption ionization (MALDI-TOF) system, Voyager-DE TM PRO from Per Septive Biosystems. Sinapinic acid was used as matrix, and bovine serum albumin was used as a standard. For the analysis, Pdr (0.4 mg/ml) was mixed with sinapinic acid (10 mg/ml in H 2 O/CH 3 CN (1:1) plus 3% trifluoroacetic acid), using a 1:5 or 1:6 ratio.
Spectral and Anaerobic Experiments-All UV-visible spectroscopy was performed using a thermostatted Cary 3 spectrophotometer. Anaerobic experiments were carried out in a quartz titration cuvette. Solutions were repeatedly evacuated and flashed with purified nitrogen. Anaerobic samples included an oxygen scrubbing system of 100 -200 M protocatechuic acid and 0.02 units/ml protocatechuate dioxygenase (8). Concentrations of WT Pdr were calculated using an extinction coefficient of 10.4 mM Ϫ1 cm Ϫ1 at 455 nm (2). The extinction coefficient for oxidized His 6 Pdr was determined on the basis of FAD released from the protein after SDS denaturation (9). The concentration of free FAD was calculated using ⑀ 450 nm ϭ 11.3 mM Ϫ1 cm Ϫ1 .
Sulfhydryl Group Assay and Modification-Assays for free thiol groups in WT and His 6 Pdr were carried out under native or denatured conditions in 50 mM Tris acetate buffer, pH 7.5, in the absence or presence of 2% SDS or 4 M guanidine HCl. Before the reaction with DTNB, Pdr was reduced with an excess of DTT in 50 mM potassium phosphate buffer, pH 7.5, at 30°C for 30 min, and any excess of reducing agent was removed from the proteins by chromatography on Sephadex G-25 in the same buffer. The number of thiols in Pdr was quantitated by the addition of a large excess of DTNB to the protein solutions and following the formation of 2-nitro-5-thiobenzoate anion (TNB) at 412 nm. Extinction coefficients of TNB under native and denaturing conditions are 13.6 and 14.2 mM Ϫ1 cm Ϫ1 , respectively (10,11). Thiol modification was performed by incubation of 25 M Pdr with 500 M N-ethylmaleimide, para-chloromercuribenzoic acid, or CdCl 2 for 30 min under anaerobic conditions in the presence or absence of 1 mM NADH. After the reaction, excesses of the thiol-modifying agent and NADH were removed by gel filtration on Sephadex G-25.
Analytical Procedures-A K m value of His 6 Pdr for NAD ϩ was determined from a Lineweaver-Burk plot, where initial rate constants for NAD ϩ -induced flavin reduction were calculated from the linear intervals of the kinetic traces of the absorbance change at 455 nm versus time recorded under anaerobic conditions at different concentrations of oxidized pyridine nucleotide. NADH:DTNB oxidoreductase assays were performed in the reaction mixtures containing 4.3 mM DTNB in 150 mM potassium phosphate buffer, pH 8.0, at 25°C, in the absence and presence of different concentrations of histidine. During anaerobic and aerobic measurements, final enzyme concentrations were 0.1 and 3 M, respectively. The reactions were started by the addition of NADH, and the change of absorbance was recorded at 412 nm (⑀ ϭ 13.6 mM Ϫ1 cm Ϫ1 (11)). The initial rate constants for the reaction of DTNB reduction catalyzed by Pdr were determined from the slopes of linear intervals of the kinetic traces. For the reactions carried out under aerobic conditions, fast phase of the kinetic traces was used for calculations. A low rate conversion of DTNB to TNB occurred in the absence of enzyme, and this was subtracted from the rate in the presence of enzyme.
Pdx reduction activity was measured at 455 nm using an extinction coefficient of 2.5 mM Ϫ1 cm Ϫ1 as previously described (12). Specific reductase activities were calculated employing molar extinction coefficients of 6.22 mM Ϫ1 cm Ϫ1 at 340 nm, 1.02 mM Ϫ1 cm Ϫ1 at 420 nm, and 21 mM Ϫ1 cm Ϫ1 at 600 nm for NADH, ferricyanide, and DCIP, respectively. Flavin fluorescence intensity was measured using a Hitachi F-4500 fluorescence spectrophotometer ( ex ϭ 455 nm; em ϭ 522 nm). Laser flash photolysis experiments were carried out as described previously (6).

Spectral and Molecular Properties of Recombinant WT and
His 6 Pdr-To facilitate purification and to obtain large quantities of Pdr for mechanistic and structural studies, we have cloned and expressed in Escherichia coli a six-histidine tagfused enzyme. Incorporation of a sequence of six histidine residues and a ␣-thrombin cleavage site to the amino terminus of Pdr resulted in no expression of an active flavoprotein, most likely due to interference of the peptide with the FAD binding site located in the N-terminal part of the protein. In contrast, the presence of the His 6 tag at the C terminus did not affect binding of the flavin cofactor, and, as a result, high levels of expression of the properly folded enzyme were achieved (up to 70 nmol/1 g of cells).
During purification procedures, no difference in chromatographic behavior of WT and His 6 Pdr was found. It was noticeable that both proteins eluted from the gel filtration column in two bands (Fig. 1). A slower moving band corresponded to a molecular mass of ϳ46 kDa and contained 95% of the enzyme. A band of greater mobility containing the rest of the enzyme corresponded to a molecular mass of 93 kDa. SDS-PAGE analysis showed that under nonreducing conditions, both bands consisted of a mixture of 46-and 93-kDa proteins, whereas only one band corresponding to the 46-kDa protein could be seen on the gel in the presence of DTT (Fig. 1, inset). This suggests that Pdr undergoes a monomer-dimer association-dissociation in solution. Further evidence for such a reversible association was obtained from mass spectrometry analysis that required considerably lower enzyme concentrations. Two peaks were present in the MALDI-TOF mass spectrum of Pdr corresponding to the monomeric and dimeric protein (Fig. 2). As seen in inset A, peak I had a shoulder corresponding to a protein that was ϳ670 Da heavier than Pdr. We hypothesized that this fraction may represent NAD ϩ -bound Pdr. Indeed, when His 6 Pdr was reduced with NADH and quickly reoxidized prior to the mass spectral analysis, the portion of the fraction with the mass of 47,304 Da was significantly increased (Fig. 2, inset B).
The absorbance spectra of WT and His 6 Pdr were similar and had absorption maxima at 379 and 455 nm, a pronounced shoulder at 480 nm, and a small broad absorption peak in the long wavelength area (Fig. 3). The long wavelength absorption has not been reported for the native Pdr from P. putida (3). The extinction coefficient of His 6 Pdr at 455 nm was calculated to be 10.9 Ϯ 0.3 mM Ϫ1 cm Ϫ1 . Fluorescent yield upon flavin excitation in His 6 Pdr was low and similar to that of the WT protein, indicating that the flavin fluorescence in both proteins is quenched (data are not shown). The ferricyanide, DCIP, and Pdx reductase activities of His 6 Pdr were similar to those of the WT enzyme (Table I).
Redox Properties of WT and His 6 Pdr-The difference between WT and His 6 Pdr was first revealed during laser flash photolysis experiments. Compared with the WT enzyme, reduction of His 6 Pdr by 5-deazariboflavin (dRF) semiquinone proceeded with a rate constant 2 times smaller. Hyperbolic fits to the plots of k obs versus enzyme concentration for the reactions of Pdr with dRF semiquinone give limiting k obs values of 1.3 ϫ 10 4 and 6.2 ϫ 10 3 s Ϫ1 and dissociation constants (K d ) of 10 and 11 M for WT and His 6 Pdr, respectively (Fig. 4A). Like the WT enzyme (6), His 6 Pdr produced the blue neutral FAD semiquinone after one-electron reduction by dRF radical. However, FADH ⅐ of His 6 Pdr was less stable than that produced by the WT protein. Within the studied time interval, 50 -60% of the FAD semiquinone produced by His 6 Pdr, versus 20 -25% in the WT flavoprotein, disproportionated to form the fully oxidized and   fully reduced FAD (Fig. 4B). These results were the first indication of the ability of the C-terminal peptide to influence the redox properties of the flavin or/and to promote some dimer formation that leads to semiquinone disproportionation.
Further, during preparation of anaerobic solutions of Pdr in the presence of NAD ϩ and in the absence of external source of electrons, it was found that oxidized pyridine nucleotide was able to promote reduction of the flavin only in His 6 Pdr and not in WT. The bright yellow solutions of His 6 Pdr were gradually bleached because of FAD reduction. The NAD ϩ -induced reaction proceeded in two spectrally distinct phases (Fig. 5). Characteristic features of the first phase, defined by the appearance of an isosbestic point at 506 nm, were the gradual loss of 455-nm absorbance and generation of a long wavelength absorbance band (Fig. 5B). This was indicative of flavin reduction and charge transfer complex formation between the reduced flavin and NAD ϩ . In addition, a small absorption band appeared at 525 nm. At the end of the first phase, ϳ55% of the FAD was reduced. During the second phase, flavin reduction and charge transfer complex formation continued, and an isosbestic point at 527 nm was observed (Fig. 5C). In this phase, a small absorbance peak appeared at 340 nm, indicating that some NADH was produced. At the end of the reaction, which was complete within a few hours, ϳ85% of the flavin was reduced. The two distinct phases observed during the reaction indicate the possible involvement of three redox species in the autoxidation/reduction of Pdr. In addition to the fully oxidized and fully reduced flavin bound to NAD ϩ , a third intermediate could be formed when the FAD is half-oxidized and half-reduced. Such complex redox behavior of His 6 Pdr might be also due to conformational inequality of the species binding pyridine nucleotide, where, if a dimer forms, the reactivity of each subunit will be influenced by the ligand and redox state of the other subunit in the dimer. Spectral changes observed during anaerobic reaction were fully reversible. Opening of the anaerobic cuvette and exposure to air resulted in complete oxidation of His 6 Pdr, and the spectrum of the reoxidized flavoprotein was identical to that of the native enzyme (data not shown). The K m value for NAD ϩ was 100 M at pH 8.0. No NAD ϩinduced spectral changes were observed in the solutions of His 6 Pdr when the pH was lowered from 8.0 to 6.0. This means that proton release is mandatory for the completion of the reaction and that the basic residue that is present in the active site and can take proton up has pK a above 6.0.
Effect of Thiol-reducing Agents and Histidine on NAD ϩ -induced Flavin Reduction in Pdr-Although there are no CXXC or CXXXXC motifs characteristic for disulfide redox centers (13) in the amino acid sequence of Pdr, the results described above suggest that thiols provide the source of electrons for flavin reduction. Indeed, if in addition to NAD ϩ an excess of thiol-reducing agent such as DTT or GSH was present in the anaerobic cell, the flavin reduction in His 6 Pdr proceeded faster, was complete, and resulted in the formation of significant amounts of NADH (Fig. 6). It should be noted that DTT and GSH did not promote the FAD reduction itself but required the presence of NAD ϩ . The initial rate constants of the flavin reduction in His 6 Pdr measured from kinetic traces of the absorbance change at 455 nm versus time were equal to 0.3, 0.4, and 0.5 nmol of FAD reduced min Ϫ1 in the absence and presence of DTT or GSH, respectively (Fig. 6, inset A).
In the presence of thiol-reducing agents, NAD ϩ also induced flavin reduction in WT Pdr. In the presence of an excess of DTT, reduction of FAD was slow, such that ϳ85% of the flavin was reduced over a period of 2 weeks, and practically no NADH formed (Fig. 7, spectrum b). In contrast, reaction of WT Pdr with NAD ϩ in the presence of GSH was considerably faster, was completed within 36 h, and resulted in full flavin reduction and formation of considerable amounts of NADH (Fig. 7, spectrum c). Since the reduction of WT Pdr required long periods of time, particular care was taken to ensure that the cuvette was anaerobic throughout the experiment. The amount of protocatechuic acid in the experiments involving WT Pdr was increased from 100 to 200 M. The ability of Pdr to accept electrons from thiol-reducing reagents and use them for NAD ϩ reduction (e.g. functioning as a thiol:NAD ϩ oxidoreductase) proves that the redox-active groups in the protein are free thiols.
Significantly longer reaction times with thiol-reducing agents in the experiments described above was not the only difference between WT and His 6 Pdr. Upon the addition of either GSH or DTT to His 6 Pdr, reduction of the flavin, formation of the charge transfer complex, and production of NADH were continuous. With WT enzyme, an initial decrease in absorbance at 455 nm was followed by an increase, indicating that some of the flavin was getting reduced and then partially reoxidized (Fig. 7, inset). It took ϳ25 h before irreversible flavin reduction, charge transfer complex formation, and NADH production began. In His 6 Pdr, therefore, the C-terminal peptide containing a six-histidine tag appears to greatly facilitate thiol oxidation and flavin reduction. We next clarified if free histidine could affect redox properties of Pdr. It appeared that histidine significantly stimulated NAD ϩ -induced flavin reduction in His 6 Pdr (Fig. 6, inset B). The initial rate constant of the reaction approached 0.7 nmol of FAD reduced min Ϫ1 at 200 mM histidine. Free histidine was also found to promote NAD ϩ -dependent electron transfer to FAD in WT Pdr. Although this reaction was not nearly as fast than that observed in His 6 Pdr, it was complete in 72 h and did not require the presence of sulfhydryl reducing agents (data not shown).
NADH-dependent DTNB Reductase Activity of Pdr-In order to elucidate whether Pdr can catalyze the reverse reaction, NADH:disulfide oxidoreduction, we measured NADH-dependent DTNB reductase activity of the enzyme. Both WT and His 6 Pdr were capable of reducing DTNB. The reaction was highly affected by oxygen. Under aerobic conditions, there was a lag period of a few minutes in the absorbance change at 412 nm (Fig. 8A). The continuation of the lag phase was dependent on DTNB, NADH, and protein concentrations. Under strict anaerobic conditions and saturating concentrations of the reaction components, no lag phase in the DTNB reduction was observed (Fig. 8A, inset). Moreover, the values of k cat for the reaction of disulfide reduction carried out under anaerobic conditions were 1 order of magnitude higher than those measured in the presence of oxygen (Table II). WT and His 6 Pdr had similar turnover numbers and K m values for both substrates. It should be mentioned that the K m values for NADH measured under anaerobic conditions were 2 orders of magnitude lower than those determined in the presence of oxygen.
There was no effect of Neurospora NADase on either the continuation of the lag phase or kinetic parameters of the reaction. In contrast, salts, such as ammonium and lithium sulfates, had a pronounced stimulating effect on DTNB reduction by both WT and His 6 Pdr. Histidine had a dual effect on the reaction (Fig. 8B). It stimulated DTNB reduction only by His 6 Pdr under aerobic conditions. In the presence of 1 mM histidine, the continuation of the lag phase was 2 times shorter, and the rate constant of the reaction was maximal and 50% larger than corresponding parameters measured in the absence of the amino acid (Fig. 8B, inset). In contrast, the DTNB reductase activities of WT and His 6 Pdr measured under anaerobic conditions were inhibited by histidine. The reason for this dual effect of histidine on the kinetics of DTNB reduction is not clear at the moment. Only under anaerobic conditions and at pH below 6.5, both WT and His 6 Pdr were capable of catalyzing NADH-dependent reduction of oxidized glutathione. The reaction was slow (Ͻ0.2 mol of NADH oxidized min Ϫ1 mol Ϫ1 Pdr) with complex kinetics. Elucidation of the detailed mechanism of the disulfide reduction catalyzed by Pdr requires further studies.
Effect of SH Group Modification on Redox Properties of Pdr-In order to further investigate the role of thiols in Pdr catalysis, we have studied the effect of SH group modification on redox properties of the enzyme. Under native and reducing conditions, only 3 of 6 cysteine residues in WT and His 6 Pdr react with DTNB (Table III). Upon denaturation of the protein with 2% SDS or 4 M guanidine HCl, two more SH groups become accessible for DTNB. The sixth cysteine in Pdr seems to be nonreactive or inaccessible under these conditions. Cd 2ϩ , para-chloromercuribenzoic acid (PCMB), and N-ethylmaleimide (NEM) were used for thiol modification. Cadmium ion acts like arsenite by bridging the nascent thiols (14), whereas molecules of PCMB and NEM react with a single SH group. Reactions of Pdr with thiol-modifying agents were carried out in the presence and absence of NADH. When an excess of CdCl 2 was added to the solution of Pdr, the protein irreversibly precipitated, presumably due to the formation of multiple intermolecular bridges via the cadmium ion and aggregation of the protein. No protein precipitation occurred during Pdr modification with PCMB or NEM, and absorbance spectra of modified proteins were similar to those of unmodified enzymes (Fig. 9).
However, the thiol-modifying agents affected the interaction between Pdr and NADH. The recombinant flavoprotein could not be completely reduced even by a large excess of NADH. Approximately 13 and 27% of WT and His 6 Pdr, respectively, remains oxidized under anaerobic conditions in the presence of a 40-fold excess of reduced pyridine nucleotide. In contrast, the addition of NEM to the partially reduced enzyme resulted in completion of flavin reduction and in a slight increase of a charge transfer complex between the reduced FAD and oxidized pyridine nucleotide (Fig. 9, spectra b and c). This result can best be explained if Pdr and NAD ϩ form a tight complex. As mentioned previously, this complex was detected by mass spectroscopy analysis. In addition, complete reduction of Pdr by NADH in the presence of Neurospora NADase also demonstrates that part of the flavoprotein is tightly bound to oxidized pyridine nucleotide. By hydrolyzing oxidized pyridine nucleotide produced during reaction of the flavoprotein with NADH, NADase permits the reaction to proceed to completion to yield the fully reduced enzyme (14). When this enzyme was present in the reaction mixture, the addition of an excess of NADH to Pdr resulted in complete FAD reduction followed by disappearance of the charge transfer complex between the reduced flavin and oxidized pyridine nucleotide (Fig. 10). If cysteine residues are involved in or promote Pdr-NAD ϩ complex formation, then reaction of thiol groups with the sulfhydryl-modifying reagents would result in a release of NAD ϩ and its replacement with NADH. The fact that NAD ϩ does not induce any changes in the absorbance spectra of either PCMB-or NEM-modified His 6 Pdr supports this view.
Although SH group modification completely eliminated an inducing effect of NAD ϩ on the electron flow from thiols to FAD, it did not affect the ability of Pdr to transfer electrons to K 3 Fe(CN) 6 , DCIP, and Pdx, indicating that interaction sites of the flavoprotein with these electron acceptors and NADH were not perturbed by the modifying agents. Furthermore, there was no inhibitory effect of thiol modification on the DTNB reductase activity of Pdr. It is possible that either modification was reversed upon reoxidation of the enzyme or sulfhydryl groups involved in the thiol-disulfide exchange were not affected by NEM and PCMB.

Properties of Recombinant Pdr-Although
Pdr was discovered more than 3 decades ago (1), it has not been studied very   extensively, in contrast to other proteins in the P450cam monooxygenase system. It was difficult to obtain large amounts of the pure enzyme due to its low expression and undeveloped purification procedures (15,16). Introduction of chromatography on phenyl-Sepharose as the last purification step significantly improved the quality of purified Pdr (6). 2 In order to further facilitate the purification procedure we have cloned and expressed Pdr in E. coli as a six-histidine tag fused protein.
Engineering of His 6 Pdr and its comparison with the WT enzyme enabled us to reveal new properties and a new function of this old flavoprotein. First, we have found that recombinant Pdr is capable of forming dimers in solution. The dimer formation, not reported for native Pdr, was detected by gel filtration, electrophoresis, and mass spectrometry methods and occurred in the wide range of protein concentration (0.4 -40 mg/ml). The fact that Pdr dimers were sensitive to the sulfhydryl-reducing agent treatment suggests that cysteine residues may be involved in the dimerization process. Elucidation of whether or not this reversible association-dissociation is specific and important for Pdr catalysis requires further investigations.
Second, mass spectrometry analysis revealed that oxidized Pdr has a high affinity for NAD ϩ and, even after long purification procedures, is partially present in solution in the NAD ϩbound form. This may explain why Pdr does not bind to affinity resins, such as 5Ј-AMP-Sepharose, used for purification of NAD ϩdependent enzymes. Existence of an NAD ϩ -bound Pdr also explains the inability of a large excess of NADH to fully reduce the flavoprotein over a period of a few hours. It is possible that NADH would be able to slowly displace NAD ϩ from the active site of Pdr. However, complete and fast flavin reduction by reduced pyridine nucleotide occurred only if NADase or thiolmodifying agents were present in the reaction mixture.
Third, electron transferring activities of recombinant Pdr to artificial acceptors, ferricyanide and DCIP, determined in the present study (Table I) were more than 1 order of magnitude higher than those published previously for the native enzyme (3). Since the amino acid sequences of native P. putida and WT recombinant proteins are identical, this discrepancy between specific activities might be partially due to differences in measurement conditions and higher purity of the recombinant enzymes.
Although molecular, spectral, and electron transferring properties of recombinant WT and His 6 Pdr are similar, the presence of eight additional C-terminal amino acid residues has a dramatic effect on the enzyme. In our laser flash photolysis experiments, the C-terminal peptide hindered interaction of His 6 Pdr with the 5-deazariboflavin radical and significantly destabilized the FAD semiquinone (Fig. 4). Finally, and most importantly, engineering of the six-histidine tag fusion protein showed that Pdr can function as a pyridine nucleotide:disulfide oxidoreductase.
Similarity of Pdr with the Enzymes of the GR Family-In order to understand what processes occur upon binding of oxidized pyridine nucleotide to His 6 Pdr, it would be helpful to compare Pdr and lipoamide dehydrogenase, a member of the glutathione reductase family. The spectral changes observed upon NAD ϩ binding to the two-electron reduced form (EH 2 ) of lipoamide dehydrogenase, in which both flavin and redox active disulfide groups are present in the active site (13,17), were similar to those induced by oxidized pyridine nucleotide in His 6 Pdr and were explained as follows. In uncomplexed EH 2 , the oxidation-reduction potential of the active center dithiol/disulfide is considerably more positive than that of FAD (18), and electrons are located principally in the dithiols, whereas the flavin is largely oxidized. The long wavelength absorbance in the absorbance spectrum of EH 2 arises from charge transfer between thiolate anion and oxidized flavin and is responsible for the red color of this intermediate (13,19). The binding of NAD ϩ to EH 2 is thought to influence the equilibrium distribution of electrons between the flavin and the disulfide redox couple by increasing the amount of reduced flavin with resultant charge transfer from reduced flavin to bound oxidized pyridine nucleotide (17). The oxidation of EH 2 by NAD ϩ is very rapid and is over within milliseconds. It is accompanied by proton release and becomes thermodynamically unfavorable at low pH.
Similarity between the spectral changes induced by NAD ϩ in His 6 Pdr and lipoamide dehydrogenase led us to consider that Pdr might have redox-active dithiols. This was later confirmed when we found that Pdr can function as a pyridine nucleotide: dithiol/disulfide oxidoreductase catalyzing both forward and reverse reactions, NAD-dependent oxidation of thiols, and NADH-dependent reduction of disulfides. The lack of putative disulfide redox centers and the complexity of the thiol-disulfide exchange reactions catalyzed by Pdr indicate that there are some significant differences between the mechanism of function of His 6 Pdr and lipoamide dehydrogenase. The absence of a shoulder at 530 nm in the absorbance spectrum of uncomplexed Pdr, characteristic for thiolate to flavin charge transfer complex (13,19), and a considerably slower reaction of NAD ϩ with Pdr than with the EH 2 form of lipoamide dehydrogenase indicate that there must be a thermodynamic or/and conformational barrier(s) for migration of electrons from the dithiols to the flavin in Pdr. The GR family of enzymes have conserved nascent thiols encoded by CXXC or CXXXXC motifs (13). In the amino acid sequence of Pdr, the two closest cysteine residues are 15 amino acid groups apart. Our computer modeling studies using an oxygenase-coupled NADH-dependent ferredoxin reductase component in biphenyl dioxygenase from P. putida as a molecular model (7) show that potential dithiol couples in Pdr can be placed not closer than 10 Å from the isoalloxazine ring of FAD. The remote location of redox active sulfhydryl groups from the flavin and/or from each other might be one of the obstacles for the fast electron migration to the flavin. Binding of NAD ϩ to the protein, therefore, might influence flavindithiol interaction by bringing thiol groups together close enough so they could form disulfide or/and by bringing the 2 L. Koo and P. R. Ortiz de Montellano, unpublished results. dithiols closer to the flavin isoalloxazine ring.
Another aspect that should be addressed is the mechanism of thiol-disulfide exchange reactions catalyzed by traditional disulfide oxidoreductases where the thiolate anion is the reactive nucleophile. The pK a values of protein sulfhydryl groups can be very different from the typical value of 8.5. In most cases, the low pK a has been attributed to the presence of a nearby positively charged center that stabilizes the thiolate anion electrostatically. There is at least one base, usually a histidine, close to one sulfur of disulfide in the active sites of disulfide oxidoreductases, such as lipoamide dehydrogenase, glutathione reductase, and thioredoxin reductase. In addition to unfavorably positioned dithiols, another reason for the considerably longer time of the reaction of NAD ϩ with Pdr than with lipoamide dehydrogenase (milliseconds versus hours) might be the protonation state of the redox active thiol(s) and, thus, the difference in flavin and dithiol/disulfide redox potentials. One possible reason that NAD ϩ induces FAD reduction only in His 6 Pdr and not in the WT protein is that the six-histidine tag acts as a base catalyst and, by shifting the dithiol/disulfide redox potentials, promotes disulfide formation and electron transfer to FAD. The finding that the addition of histidine to the reaction mixture stimulated FAD reduction in WT Pdr supports this view. Formation of the charge transfer complex between NAD ϩ and reduced flavin stabilizes the intermediate, a small portion of which is in equilibrium with a complex between reduced pyridine nucleotide and oxidized FAD. It is unknown whether the C-terminal peptide interacts with the redox-active thiols of the same or different Pdr molecule. In the GR family of enzymes, a catalytically important histidine located at the Cterminal end of one subunit of a dimer acts as an acid/base catalyst in the active site of another subunit (20, 21). The C-terminal peptide in His 6 Pdr may not be long enough to reach the active site of the same molecule, but it could approach the redox-active cysteines and promote thiolate formation in neighboring molecules. The time required for the formation of intermolecular complexes might be an additional factor that prolongs the reaction between NAD ϩ and the flavoprotein. The mechanism of NAD ϩ -induced FAD reduction in His 6 Pdr is summarized in Scheme 1.
Finally, an important property of the GR family disulfide oxidoreductases is that they can function only as dimers. In these enzymes, there is an intimate contact between the two monomers, and elements of each active site are drawn from both subunits. Besides, strong negative functional cooperativity or half-site reactivity is observed between the subunits in lipoamide dehydrogenase and mercuric ion reductase, where NAD ϩ /NADP ϩ binding in one subunit induces adduct formation in that subunit and promotes changes in the other subunit that preclude tight binding of a second molecule of oxidized pyridine nucleotide (22)(23)(24). This mechanism implies a fluid functional asymmetry of the dimer that occurs as a consequence of subunit communication (25). Since (i) Pdr is a structural homologue to the enzymes of the GR family and is capable of dimerizing in solution and (ii) the spectral changes occurring during anaerobic reaction of Pdr with NAD ϩ are clearly biphasic, it is tempting to suggest that Pdr dimers are also involved in the reaction with NAD ϩ and that binding of oxidized pyridine nucleotide to one of the subunits is preferential and affects reactivity of the other monomer. It is possible that in order to function as dithiol/disulfide oxidoreductase, Pdr also must form a dimer. Since under aerobic conditions there were only 5% of dimers in Pdr solutions and continuation of the lag phase observed in the reaction of dithiol/disulfide exchange catalyzed by the flavoprotein was protein concentration-dependent, we speculate that the dimer formation might be a limiting step of the reaction and, therefore, might be partially responsible for the lag phase.
Dual Function of Pdr-Cysteine 73 of Pdx has been identified as important in the interaction with Pdr, since its replacement with other amino acid residues resulted in more than 60% loss of activity toward the flavoprotein (12). One of the findings in this study was the lack of effect of the sulfhydryl group modification of Pdr on its Pdx reductase activity. This implies that the SH groups in Pdr are not involved in the complex formation with Pdx. Although Pdr has a marked disulfide/ dithiol oxidoreductase activity, the turnover number for this reaction and the affinity to the disulfide/dithiol substrates are considerably lower than corresponding parameters for the reaction of Pdx reduction. Taken together, our results indicate that the electron transferring and dithiol/disulfide oxidoreductase activities of Pdr can be distinguished. Pdr can be considered as a bifunctional enzyme that, under oxygen-and NADHsaturating conditions, acts as NADH-dependent ferredoxin reductase, supplying electrons for the P450cam monooxygenase, whereas in the oxygen-depleted cells it can function as dithiol/disulfide oxidoreductase, catalyzing various reactions depending on what pyridine nucleotide and dithiol/disulfide is available. Considering structural homology between Pdr and the GR family enzymes and some resemblance between mech-SCHEME 1. Mechanism of NAD ؉ -induced FAD reduction in His 6 Pdr. Models represent mechanisms for intramolecular (A) and intermolecular (B) interaction of the C-terminal peptide with the active site of His 6 Pdr. anisms of Pdr and lipoamide dehydrogenase, we hypothesize that Pdr might be a predecessor of the FAD-containing disulfide oxidoreductases. The mechanism of disulfide reduction, not effective and "rudimentary" in Pdr, could be later developed to perfection by the GR family enzymes that acquired a redoxactive disulfide close to FAD. It seems very unlikely that the ability of Pdr, a very sufficient ferredoxin reductase, to catalyze dithiol/disulfide oxidoreduction was developed later in the course of evolution.