Role of Thr66 in Porcine NADH-cytochromeb 5 Reductase in Catalysis and Control of the Rate-limiting Step in Electron Transfer*

Site-directed mutagenesis of Thr66 in porcine liver NADH-cytochromeb 5 reductase demonstrated that this residue modulates the semiquinone form of FAD and the rate-limiting step in the catalytic sequence of electron transfer. The absorption spectrum of the T66V mutant showed a typical neutral blue semiquinone intermediate during turnover in the electron transfer from NADH to ferricyanide but showed an anionic red semiquinone form during anaerobic photoreduction. The apparent k cat values of this mutant were ∼10% of that of the wild type enzyme (WT). These data suggest that the T66V mutation stabilizes the neutral blue semiquinone and that the conversion of the neutral blue to the anionic red semiquinone form is the rate-limiting step. In the WT, the value of the rate constant of FAD reduction (k red) was consistent with thek cat values, and the oxidized enzyme-NADH complex was observed during the turnover with ferricyanide. This indicates that the reduction of FAD by NADH in the WT-NADH complex is the rate-limiting step. In the T66A mutant, thek red value was larger than thek cat values, but thek red value in the presence of NAD+was consistent with the k cat values. The spectral shape of this mutant observed during turnover was similar to that during the reduction with NADH in the presence of NAD+. These data suggest that the oxidized T66A-NADH-NAD+ ternary complex is a major intermediate in the turnover and that the release of NAD+ from this complex is the rate-limiting step. These results substantiate the important role of Thr66 in the one-electron transfer reaction catalyzed by this enzyme. On the basis of these data, we present a new kinetic scheme to explain the mechanism of electron transfer from NADH to one-electron acceptors including cytochromeb 5.

fatty acid synthesis (4,5), cholesterol synthesis (6), and xenobiotic oxidation (7) as a member of the electron transport chain on the endoplasmic reticulum. In erythrocytes, this enzyme participates in the reduction of methemoglobin (8).
The outline of the catalytic cycle of the solubilized catalytic domain of NADH-cytochrome b 5 reductase (b5R) is understood as follows (3) (Scheme I). At first, two electrons are transferred from NADH to FAD by hydride (H Ϫ ) transfer. Then the twoelectron reduced enzyme-NAD ϩ complex (E-FADH Ϫ -NAD ϩ ) transfers two electrons to two one-electron acceptors one by one via the anionic red semiquinone form (E-FAD⅐ Ϫ -NAD ϩ ), and the reduced enzyme returns to the oxidized state. Strittmatter (9 -11) suggested that the reduction of FAD by NADH is the rate-limiting step in electron transfer catalyzed by b5R. Iyanagi et al. (3,12) found that the anionic red semiquinone of FAD in Pb5R is stabilized by binding of NAD ϩ . Kobayashi et al. (13) analyzed the conversion of the neutral blue to the red semiquinone in the presence of NAD ϩ using a pulse radiolysis technique. Meyer et al. (14) also demonstrated that NAD ϩ stabilizes the red semiquinone of the human b5R and modulates the electron transfer to b5. These studies suggest the importance of the anionic red semiquinone form of the b5R-NAD ϩ complex in the electron transfer.
The preliminary tertiary structure of human erythrocyte b5R (15,16), and the detailed tertiary structures of porcine and rat liver b5Rs at 2.1 Å resolution have been determined by x-ray crystallography (17)(18)(19)(20)(21). These structural studies revealed that NADH-cytochrome b 5 reductase belongs to the structurally related so-called "ferredoxin reductase family" (22,23) together with other flavoenzymes such as ferredoxin-NADP ϩ oxidoreductase (FNR) (22), phthalate dioxygenase reductase (24), flavodoxin reductase (25), NADPH-cytochrome P-450 reductase (26), and the cytochrome b reductase domain of nitrate reductase (27). Enzymes of this family contain a flavin-binding domain and a pyridine nucleotide-binding domain. The former domain has a highly conserved flavin-binding amino acid sequence motif, RXY(T/S). In the Pb5R, Arg 63 , Tyr 65 , and Thr 66 comprise this sequence motif (19). Using sitedirected mutagenesis, we demonstrated that the positive charge of Arg 63 is critical for the affinities of Pb5R for both NADH and NAD ϩ , and the specific arrangement between the side chain of Tyr 65 and FAD contributes to protein stability and electron transfer (28). Marohnic and Barber (29) also reported the effects of mutations of the corresponding Arg 91 in rat b5R.
The Thr 66 residue in Pb5R is positioned near both the N5 atom of the isoalloxazine ring of FAD and the potential binding site of the nicotinamide ring of NADH (20,28) (Fig. 1). This position corresponds to threonine or serine residues in the other members of the ferredoxin reductase family (22, 24 -27). Ser 96 in spinach leaf FNR is critical to the reductive halfreaction of FAD (30), and Ser 90 in the C-terminal Tyr 208 mutant of pea leaf FNR forms a hydrogen bond with the amide moiety on the nicotinamide ring of the pyridine nucleotide in both the enzyme-NADP ϩ and enzyme-NADPH complexes (31). However, b5R and cytochrome b reductase domain of nitrate reductase do not have an aromatic ring corresponding to that of the C-terminal Tyr 208 in pea leaf FNR, which contacts with the re-side of the isoalloxazine ring of FAD and moves away accompanied with the binding of nicotinamide (31). In addition, the main physiological role of leaf FNR is the reduction of the oxidized pyridine nucleotide, and the direction of the electron transfer between FAD and pyridine nucleotide is different from that of b5R. Therefore, it is considered that Thr 66 in Pb5R contributes to the reduction of FAD and/or the stabilization of the reduced FAD. Shirabe et al. (32) reported that mutations of the corresponding Thr 94 in human b5R affect oxidation of FAD, but the effects of the mutations on the properties of reduced FAD in the catalytic cycle have not been clarified.
To analyze the role of Thr 66 in catalysis in Pb5R, we replaced Thr 66 in Pb5R with serine (T66S), alanine (T66A), and valine (T66V) and analyzed the redox properties of FAD in the catalytic cycle using a stopped flow spectrophotometer. We present here that the conversion of the neutral blue to the red semiquinone intermediate and the release of NAD ϩ from the enzyme in the catalytic cycle were modulated by the mutations of Thr 66 in Pb5R. In addition, we present a new model of the reaction sequence of Pb5R containing the blue neutral semiquinone and the oxidized enzyme-NAD ϩ -NADH ternary complex.

EXPERIMENTAL PROCEDURES
Materials-Enzymes for recombinant DNA technology were from Takara and Toyobo. NADH and NAD ϩ were from Oriental Yeast. Wild type recombinant Pb5R (WT) was prepared as previously described (33).
Mutagenesis, Expression, and Purification of Mutant Pb5Rs-Alteration of the gene encoding Pb5R was carried out by site-directed mutagenesis using PCR by the methods described by Higuchi (34). Briefly, for the preparation of the mutant genes encoding the mutant proteins, two primary PCR products that overlap in sequence were first obtained from a DNA template, pU8Pb5R, which contains the gene encoding the WT (33). One product was generated with the forward primer 5Ј-TAG-GAGGTCATATGTCCACCCCGGCC-3Ј containing a NdeI site (underlined) and the mutagenic common reverse primer, 5Ј-GGGCCGAA-TGACCAG-3Ј, and the other was obtained with the forward mutagenic primer and the reverse primer 5Ј-CCGCCAAGCTTCTAGAAGGCGAA-GCAGC-3Ј containing a HindIII site (underlined). As the mutagenic forward primers, 5Ј-CTGGTCATTCGGCCCTACNNNCCCGTCTC-3Ј, which have a complementary nucleotide sequence to the 5Ј-end of the mutagenic forward primers, were used. In these primers, NNN are the bases corresponding to the 66th amino acid residue, and GCT, TCG, and GTG were used for the mutations to alanine, serine, and valine, respectively. The resultant two PCR products were mixed and reamplified with the forward and reverse primers. The resultant secondary PCR product was inserted into the plasmid pCW ori ϩ (35), a derivative of pHSe5 (36, 37), using the NdeI and HindIII sites to construct the expression plasmid for generating mutant proteins. The entire nucleotide sequences of the mutant genes were confirmed using an ABI PRISM 310 Genetic Analyzer. All of the mutant proteins were ex-pressed in the soluble fraction of Escherichia coli BL21 cells and purified using the same method as the WT (33). The purity of the mutant proteins was confirmed by SDS-PAGE using 12% polyacrylamide gels. The flavin bound to the mutant proteins was analyzed by TLC on a Kieselgel 60 F245 plate (Merck) (38). Purified mutant proteins were stored in 100 mM potassium phosphate (pH 7.0) containing 0.1 mM EDTA at Ϫ20°C until use.
Preparation of the Solubilized Domain of Porcine Liver Cytochrome b 5 -The recombinant solubilized domain of porcine liver cytochrome b 5 (Pb5) was prepared as follows. The cDNA encoding the full-length porcine liver cytochrome b 5 was amplified from the previously described first strand cDNA, which was prepared from a total RNA preparation from porcine liver (33). The forward primer was 5Ј-GTTAAGAAATGGCCGAGGAGTCC-3Ј, which has an initiator methionine codon followed by the nucleotide sequence encoding the N-terminal tetrapeptide of natural bovine liver b5 (39). The reverse primer was 5Ј-CTTCGGTTACCTTCTTTTCTGACG-3Ј. This nucleotide sequence was complementary to the nucleotide sequence located 12-35 bases downstream after the stop codon in the cDNA of bovine liver b5 (39). The amplified DNA fragment was blunt-ended and inserted into the HincII site of plasmid pUC118, and plasmid pU8Pb5 was selected. Plasmid pU8Pb5 contained the nucleotide sequence encoding 133 amino acid residues from the N-terminal Ala 1 to the C-terminal Asn 133 of porcine liver b5. The deduced amino acid sequence was identical to that previously reported except for the difference at position 3 (40 -42). The deduced amino acid residue at position 3 was not glutamine but glutamic acid. The polypeptide containing 87 amino acid residues from Ala 7 to Lys 93 was prepared as recombinant Pb5. The cDNA encoding recombinant Pb5 was amplified from pU8Pb5 with the forward primer 5Ј-AGGAGGTCATATGGCCGTGAAGTATTACACC-3Ј, which has an NdeI site (underlined) containing an additional initiator methionine codon, followed by the nucleotide sequence encoding the N-terminal hexapeptide of recombinant Pb5, and the reverse primer 5Ј-CCGCCA-AGCTTCTACTTGGCAATCTTGATC-3Ј, which corresponds to the Cterminal peptide, a stop codon, and a HindIII site (underlined). The resultant fragment was inserted into pCW ori ϩ using NdeI and HindIII sites to construct pCPb5. E. coli TG1 cells containing pCPb5 were cultivated in Luria-Bertani medium containing 50 g/ml ampicillin at 37°C. When the absorbance at 600 nm was ϳ0.3, isopropyl-␤-D-thiogalactopyranoside was added to a final concentration of 0.2 mM, and the cultivation was continued for 14 h. The cells were lysed by sonication in 10 mM Tris-HCl (pH 7.4) containing 1 mM EDTA and 2 mM phenylmethylsulfonyl fluoride. The lysate was subjected to centrifugation at 18,000 ϫ g for 20 min, and the supernatant was separated on a Sephadex G-100 (Pharmacia) column equilibrated with 10 mM potassium phosphate (pH 7.0) (buffer A). Red colored fractions containing recombinant Pb5 were applied to an anion exchange resin (DE52) column equilibrated with buffer A, and the proteins were eluted with a linear gradient of potassium chloride from 0 to 0.4 M in buffer A. Fractions containing recombinant Pb5 were concentrated and desalted with a Sephadex G-25 (fine) (Pharmacia) column equilibrated with 100 mM potassium phosphate (pH 7.0). The purified recombinant Pb5 showed a single band of ϳ10 kDa on a 15% polyacrylamide gel after SDS-PAGE. The yield of the purified protein from 1 liter of culture fluid was 5.9 mg. The N-terminal amino acid sequence analyzed with a Shimadzu PSQ-1 protein sequencer was Ala-Val-Lys-Tyr-Tyr, and most of the additional N-terminal methionine residues were cleaved. The oxidized and reduced absorption spectra at 300 -700 nm and the ability to accept electrons from the Pb5R were almost identical to those of the natural Pb5 (43).
Protein Concentrations-The molar extinction coefficients of the mutant proteins at 460 nm (⑀ 460 ) were determined by a method similar to that described by Aliverti and Zanetti (44), as previously described (33). The molar concentration of the WT was determined from the absorbance at 460 nm using the molar extinction coefficient, 1.02 ϫ 10 4 M Ϫ1 cm Ϫ1 (33). The protein concentration of recombinant Pb5 was determined using the molar extinction coefficient of natural Pb5, 1.13 ϫ 10 5 M Ϫ1 cm Ϫ1 at 413 nm (45).
Spectral Analyses-Absorption spectra were measured on a Hitachi U-2010 spectrophotometer. CD spectra were measured on a Jasco J-700 spectropolarimeter. Fluorescent emission spectra were measured in 10 mM potassium phosphate (pH 7.0) at 25°C on a Hitachi F-3010 Fluorescence spectrophotometer. The excitation wavelength was 460 nm, and emission spectra at 470 -650 nm were observed.
Enzymatic Activity-Steady-state enzymatic activities were measured as previously described (33 Pb5 values, 300 M NADH was used as an electron donor. The reduction rate of ferricyanide was measured at 420 nm using a molar extinction coefficient of 1.02 ϫ 10 3 M Ϫ1 cm Ϫ1 . The reduction rate of recombinant Pb5 was measured at 556 nm using a difference in molar extinction coefficient of natural Pb5 between the oxidized and reduced states, 1.9 ϫ 10 4 M Ϫ1 cm Ϫ1 (43). The concentration of NADH was determined using a molar extinction coefficient, 6.3 ϫ 10 3 M Ϫ1 cm Ϫ1 at 340 nm.
Stopped Flow Measurements-Rapid reaction was analyzed with a Photal RA-401 stopped flow spectrophotometer (Otsuka Electronics) equipped with a Lauda RMS thermostatically regulated circulating water bath in 10 mM potassium phosphate (pH 7.0) at 25°C. The spectral changes of enzymes during and after the electron transfer from NADH to ferricyanide were analyzed as follows. Equal volumes of the enzyme solution containing potassium ferricyanide and the NADH solution were rapidly mixed, and the rapid scan spectra at 460 -800 nm and the time courses of the absorbance changes at 460, 530, and 620 nm were measured. The initial concentrations of NADH, oxidized enzyme, and potassium ferricyanide in the reaction mixtures were 1 mM, 10 M, and 150 M, respectively. Under these experimental conditions, a turnover phase and a subsequent reduction phase after the consumption of ferricyanide were observed. The rate constant for the absorbance change after the turnover (k) was determined by single exponential curve fitting of the data as previously described (33).
Spectral changes of the enzymes during the reduction with NADH were analyzed with rapid scan spectra and time courses of the absorbance changes at 460 nm in the presence and absence of 1 mM NAD ϩ . The enzyme solution, which contains or does not contain NAD ϩ , was rapidly mixed with the NADH solution. The rapid scan spectra in the region at 420 -560 and at 560 -730 nm were measured separately and were joined at 560 nm. In the measurement of the rapid scan spectra, the gate time was 4 ms. The rate constants of reduction (k red ) in the presence (k red ϩNAD ) and absence of 1 mM NAD ϩ (k red ϪNAD ) were determined by single exponential curve fitting of the data at 460 nm.
Photoreduction-The flavin cofactor in the enzymes was photoreduced in an anaerobic cuvette containing 40 M protein, 1 M 5-deazariboflavin, 5 mM EDTA, 200 M NAD ϩ , ϳ0.5 M indigodisulfonate, and 10 mM potassium phosphate (pH 7.0). The solutions were made anaerobic by successive flushing with oxygen-free argon gas with gentle agitation for more than 50 min. The absorption spectra were observed before and after illumination at 25°C with a 300 W halogen lamp at room temperature.
Measurement of Dissociation Constants-The dissociation constant of the oxidized enzyme for NAD ϩ (K d NADϩ ) was determined by measuring the perturbation of the flavin spectrum as previously described (28). Approximately 40 M of oxidized mutant proteins were titrated with NAD ϩ in 10 mM potassium phosphate (pH 7.0) at 25°C. After the successive addition of NAD ϩ into both the sample and reference cells, the absorption spectra were measured. The K d NADϩ values were determined by direct curve fitting of the difference in the absorbance at the wavelength, where the absorbance change caused by the addition of NAD ϩ was largest in the 400 -550-nm region, with the theoretical equation for a 1:1 binding mechanism taking into account the dilution.

Purification of Mutant Proteins-All of the purified mutant
Pb5Rs showed a single band on an SDS-PAGE gel that was located at the same position as the WT. The yields of the purified T66S, T66A, and T66V mutants from 1 liter of culture fluid were 15.7, 20.3, and 14.5 mg, respectively. In all mutant proteins, the only bound flavin detected on the TLC plate was FAD. The molar extinction coefficients of the T66S, T66A, and T66V mutants at 460 nm were similar to that of the WT and were 1.05 ϫ 10 4 , 1.01 ϫ 10 4 , and 1.06 ϫ 10 4 M Ϫ1 cm Ϫ1 , respectively.
Spectral Properties of Oxidized Proteins-The absorption spectrum of the T66S mutant was almost identical to that of the WT (Fig. 2A, panel a). The absorption spectra of the T66A and T66V mutants were also similar to that of the WT, but slight spectral changes were observed ( Fig. 2A, panels b and c). The absorption spectrum of the T66V mutant was blue-shifted by ϳ3 nm in comparison with that of the WT. The intensities of the peaks at 390 nm were slightly decreased in the spectra of the T66A and T66V mutants. The CD spectra of the T66A, T66S, and T66V mutants were also similar to that of the WT (Fig. 2B). All of the mutant proteins showed a significant decrease in the intensity of the fluorescence emission of FAD. The intensities of the fluorescence emission spectra of free FAD, the WT, and the T66S, T66A, and T66V mutants at 524 nm were 38.9, 1.3, 1.5, 1.4, and 1.3, respectively (spectra are not shown). These data show that the mutations did not change the overall structure of the oxidized form of the enzyme.
Steady-state Kinetic Parameters and Dissociation Constant for NAD ϩ -The apparent steady-state kinetic parameters and the K d NADϩ values are shown in Table I. The K m NADH values of the T66S, T66A, and T66V mutants were similar to that of the WT, indicating that the mutations in these proteins did not affect the apparent affinity for NADH. The k cat NADH values of the T66S and T66A mutants were almost identical to that of the WT, but that of the T66V mutant was ϳ10% of that of the WT. The K m Fe values of the three mutants were similar to that of the WT. The changes in the k cat Fe values of the three mutants were similar to those in the k cat NADH values. Only the mutation of Thr 66 to valine significantly affected the electron transfer from NADH to ferricyanide.
The apparent K m Pb5 and k cat Pb5 values were evaluated using recombinant Pb5. The K m Pb5 value of the T66S mutant was similar to that of the WT. The K m Pb5 values of the T66A and T66V mutants were 6.3-and 5.6-fold of that of the WT, respectively. The k cat Pb5 value of the T66S mutant was similar to that of the WT, but the k cat Pb5 values of the T66A and T66V mutant were 34 and 4.2% of that of the WT, respectively. The mutations of Thr 66 to alanine and valine affected not only the rate of the electron transfer but also the apparent affinity for Pb5. The K d NADϩ value of the T66V mutant was 5.1-fold of that of the WT, whereas those of the T66S and T66A mutants were similar to that of the WT. The substitution of Thr 66 by valine caused a decrease in the affinity for NAD ϩ .
Rapid Scan Spectra during and after Turnover-The spectra of the WT and mutant enzymes during and after the turnover were directly analyzed by rapid scan analysis (Fig. 3). In this experiment, a turnover phase and a subsequent reduction phase by NADH after the consumption of ferricyanide were observed.
In the case of the WT and the T66S and T66A mutants, the rapid scan spectra in the turnover phase showed significant absorption at 460 -540 nm in addition to the broad absorption around 620 nm (Fig. 3A, panels a-c). Absorption at these wavelengths decreased simultaneously (Fig. 3B, panels a-c). The rate constant (k) values were consistent with the k cat values (the k cat NADH and k cat Fe values) (Table II). These data indicate that in the WT and the T66S and T66A mutants, the spectrum observed in the turnover phase is due to the intermediate, which contains oxidized FAD, and that the conversion of the intermediate is the rate-limiting step.
In the turnover phase, the T66V mutant showed peaks at 530 and 620 nm but did not show significant absorption at ϳ460 nm (Fig. 3A, panel d). This spectral shape was obviously different from those of the WT and the other mutants and was characteristic of a neutral blue semiquinone form. Spectra similar to that of a neutral blue semiquinone, which has peaks at ϳ500 and 600 nm, were observed in the flavodoxins and their mutants (46 -48). Accompanied with the consumption of ferricyanide, both peaks disappeared simultaneously (Fig. 3B,  panel d). The k value at 530 nm was almost identical to that at 620 nm (Table II). These data indicate that in the T66V mutant, the spectrum observed in the turnover phase is due to the neutral blue semiquinone and that the conversion of the neutral blue semiquinone is the rate-limiting step.
Photoreduction-The enzymes were anaerobically photoreduced in the presence of NAD ϩ to analyze the statically stable semiquinone form (Fig. 4). The enzymes showed biphasic absorbance changes at 375 and ϳ460 nm by successive photoreduction and resulted in the spectra of fully reduced forms. In the case of the WT, absorbance at 460 nm was decreased, and peaks at 375 nm and near 530 nm were increased after photoreduction for 5 min (Fig. 4a). This indicates the formation of the anionic red semiquinone (3). The peaks near 375 nm, which were characteristic of the anionic red semiquinone, were observed in the spectra of the T66S, T66A, and T66V mutants after photoreduction for 3, 5, and 4 min, respectively (Fig. 4,

b-d).
In these three mutants, the stable one-electron reduced form was an anionic red semiquinone. Reduction of Enzymes without Turnover-The WT and mutant enzymes were reduced with NADH in the presence and absence of NAD ϩ to identify the intermediates observed in the rapid scan spectra of the WT and the T66S and T66A mutants during the turnover with ferricyanide (Figs. 5 and 6). In the presence of 1 mM NAD ϩ , more than ϳ70% of the T66V mutant and 90 -95% of the other enzymes are in the oxidized enzyme-NAD ϩ complex as judged from the K d NADϩ values in Table I, and these complexes were reduced with NADH. In these experiments, only the last 10 -20% of the reduction phases were observed just after the dead time of the instrument (Figs. 5A and 6A).
In both the presence and the absence of NAD ϩ , spectra of the WT observed during the reduction with NADH showed a large absorption peak at 460 nm with a shoulder at 490 nm and a broad absorption around 620 nm, all of which decreased simul-taneously (Figs. 5B, panel a, and 6B, panel a). These spectral features were identical to that observed in the turnover phase (Fig. 3A, panel a). In the WT, the values of the rate constant of reduction (k red ) in the presence of NAD ϩ (k red ϩNAD ) and in the absence of NAD ϩ (k red ϩNAD ) were in good agreement with the halves of k cat values and were consistent with the k cat values (Table II). These data suggest that the spectrum of the WT observed in the turnover phase is due to the oxidized WT-NADH complex. The absorbance changes at 460 nm, which were observed after the flow was stopped, both in the presence and absence of NAD ϩ , were ϳ0.1 and 0.06, respectively (Figs.  5A, panel a, and 6A, panel a). These differences in the absorbance changes indicate that the reduction of the oxidized WT-NADH complex was delayed in the presence of 1 mM NAD ϩ .
In the case of the T66S mutant, the spectrum observed during the reduction with NADH in the absence of NAD ϩ showed an obvious shoulder at 490 nm and a broad peak of absorption at ϳ620 nm (Fig. 6B, panel b). This spectral feature   Fig. 3B at 460, 530, and 620 nm, reduction rate constants (k red ) evaluated from the absorbance changes at 460 nm in the presence of the NAD ϩ shown in Fig. 5A (k red ϩNAD ), and in the absence of the NAD ϩ shown in Fig. 6A  was identical to that observed in the turnover phase (Fig. 3A,  panel b). In the T66S mutant, the k red ϩNAD and k red ϪNAD values were in good agreement with the halves of k cat values (Table  II). These data indicate that the spectrum observed in the turnover phase is due to the oxidized T66S-NADH complex.
The shoulder at 490 nm was not observed in the spectrum during the reduction with NADH in the presence of NAD ϩ (Fig.  5B, panel b). The absence of the shoulder was probably due to the formation of the oxidized T66S-NADH-NAD ϩ ternary complex. Such a ternary complex is considered in the T66A mutant  Table II Insets, magnified views of the spectra at 580 -720 nm. The directions of the spectral changes are indicated with arrows. The spectra and the absorbance changes were measured in 10 mM potassium phosphate (pH 7.0) at 25°C. also (see below). The absorbance changes at 460 nm, which were observed after the flow was stopped, both in the presence and absence of NAD ϩ , were ϳ0.12 and 0.08, respectively (Figs .  5A, panel b, and 6A, panel b).
The spectra of the T66A mutant observed during reduction with NADH in the presence of NAD ϩ showed absorption peaks at 460 nm without an obvious shoulder at 490 nm (Fig. 5B, panel c) as seen in the spectrum in the turnover phase (Fig. 3A,  panel c). In the T66A mutant, the k red ϩNAD value was roughly half of the k cat values and was consistent with the k cat values (Table  II). However, in the absence of NAD ϩ , an obvious shoulder at 490 nm was observed in the spectra of the T66A mutant (Fig.  6B, panel c). This shoulder was observed neither in the spectrum in the turnover phase (Fig. 3A, panel c) nor in the spectrum of the oxidized T66A-NAD ϩ complex (Fig. 7A, panel c). In the T66A mutant, the k red ϪNAD value was 1.7-fold of the k red ϩNAD value and larger than the halves of k cat values (Table II). The absorbance changes at 460 nm, which were observed after flow was stopped, both in the presence and absence of NAD ϩ , were ϳ0.13 and 0.045, respectively (Figs. 5A, panel c, and 6A, panel c). These differences in the absorbance changes are consistent with the result that the k red ϪNAD value was larger than the k red ϩNAD value. These results suggest that the spectrum of the T66A mutant in the turnover phase (Fig. 3A, panel c) is neither the spectrum of the T66A-NADH complex nor that of the T66A-NAD ϩ complex. Probably, the spectrum of the T66A mutant observed in the turnover phase is due to the T66A-NADH-NAD ϩ ternary complex, and the release of NAD ϩ from the ternary complex is the rate-limiting step. It is considered that the ternary complex has such a structure that both NADH and NAD ϩ bind simultaneously to the oxidized T66A mutant mainly with the 5Ј-ADP-ribose moiety of NADH and with the oxidized nicotinamide moiety of NAD ϩ , respectively. In the case of the T66V mutant, the spectra observed during reduction with NADH in the presence and absence of NAD ϩ showed a large absorption peak at 460 nm, with a shoulder at 490 nm and a broad region of absorption around 620 nm, all of which decreased simultaneously (Figs. 5B, panel d, and 6B, panel d). These spectral features are similar to those of the WT and are different from the spectrum observed in the turnover phase (Fig. 3A, panel d). The k red ϩNAD and k red ϪNAD values of the T66V mutant were larger than the k cat values of this mutant and roughly similar to the k red ϩNAD and k red ϪNAD values of the WT and T66S mutant (Table II). These results indicate that the substitution of Thr 66 by valine had very little effect on the reduction of the enzyme.
Spectral Changes Caused by Binding of NAD ϩ , 5Ј-ADPribose, and NADH-To analyze the contribution of the 5Ј-ADPribose moiety to the binding of pyridine nucleotide, NAD ϩ , 5Ј-ADP-ribose, and NADH were added to the oxidized enzymes as shown in Fig. 7. The spectral changes of the T66S and T66A mutants caused by the addition of NAD ϩ were large, whereas those of the WT and T66V mutant were small (Fig. 7A). These spectral changes are due to the formation of the oxidized enzyme-NAD ϩ complexes. The features of the spectral changes caused by the addition of NAD ϩ were similar, but the degrees of the spectral changes were different. The difference spectra of the WT and mutant enzymes showed a positive peak around 513 nm and negative peaks around 459 and 490 nm. These spectra of the oxidized enzyme-NAD ϩ complexes were changed  Table II Insets, magnified views of the spectra at 580 -720 nm. The directions of the spectral changes are indicated with arrows. The spectra and the absorbance changes were measured in 10 mM potassium phosphate (pH 7.0) at 25°C. by the addition of 5Ј-ADP-ribose, which lacks a nicotinamide moiety. The spectra of the WT and mutant enzymes obtained after the addition of 5Ј-ADP-ribose were similar, and their difference spectra showed negative peaks around 470 and 500 nm. These spectra were almost identical to those obtained after the addition of 5Ј-ADP-ribose to the oxidized enzymes (Fig. 7B). These spectral changes caused by the additions of NAD ϩ and 5Ј-ADP-ribose in this order indicate that NAD ϩ in the oxidized enzyme-NAD ϩ complexes was released by the addition of 5Ј-ADP-ribose, resulting in the formation of the oxidized enzyme-5Ј-ADP-ribose complexes. The spectra of the oxidized enzyme-5Ј-ADP-ribose complexes were hardly changed by the addition of NAD ϩ (Fig. 7B), indicating that the affinity of the oxidized enzymes for NAD ϩ is lower than that for 5Ј-ADP-ribose. The spectra obtained after the additions of both NAD ϩ and 5Ј-ADPribose were almost identical to those of the enzyme-5Ј-ADPribose complexes (Fig. 7). The resultant enzyme-5Ј-ADP-ribose complexes were reduced by the addition of NADH and resulted in the spectra of the reduced enzyme-NAD ϩ complexes, which showed a broad absorption of the charge transfer complex at 500 -800 nm. The 5Ј-ADP-ribose moiety is necessary for both the binding of pyridine nucleotide and the release of NAD ϩ from the oxidized enzyme-NAD ϩ complexes. DISCUSSION In this study, we replaced Thr 66 in Pb5R with serine, alanine, and valine and analyzed the effects of the mutations on the electron transfer catalyzed by Pb5R. The absorption, CD, and fluorescence spectra of the mutant proteins indicate that the amino acid substitutions did not change the overall structure of Pb5R. The T66A and T66V mutants maintained catalytic activity, indicating that the hydroxyl group of Thr 66 is not essential for electron transfer. However, mutations of Thr 66 in Pb5R affected catalysis.
The T66V mutant exists as the neutral blue semiquinone form during the turnover (Fig. 3), and the conversion of this SCHEME II form is the rate-limiting step. However, the anaerobic photoreduction spectra of the T66V mutant in the presence of NAD ϩ indicated the presence of the anionic red semiquinone form (Fig. 4). It is considered that the neutral blue semiquinone form converts to the red semiquinone form during turnover, and this step is the rate-limiting step. In the T66V mutant, only the hydroxyl group of the Thr 66 in the WT was replaced with a methyl group. The blue neutral semiquinone of FAD (FADH⅐) has a hydrogen atom on the N5 atom of the isoalloxazine ring, and release of a proton from the N5 position is required for the conversion of the blue neutral semiquinone to the anionic red semiquinone (49). The substitution of Thr 66 by valine is considered to be unfavorable for the effective proton release from the N5 position of the neutral blue semiquinone of FAD. Kobayashi et al. (13) observed the conversion of the blue semiquinone to the red semiquinone form of the WT using pulse radiolysis below pH 6.5, and only the stable red semiquonone appeared at 200 s after pulse above pH 7.0. These data suggest that at pH 7.0 the blue semiquinone form of the WT is unstable, and the conversion of the blue to red semiquinone is very fast.
The spectrum of the T66A mutant, which was observed during the turnover (Fig. 3A, panel c), was assigned to be the oxidized enzyme-NADH-NAD ϩ ternary complex. Those of the WT and the T66S mutant (Fig. 3A, panels a and b) were assigned as the oxidized enzyme-NADH complexes, but the spectra of the T66S mutant observed during the reduction with NADH in the presence of 1 mM NAD ϩ were assigned as the oxidized T66S-NADH-NAD ϩ ternary complex (Fig. 5B, panel  b). These data suggest that the oxidized enzyme-NADH complex is produced via the oxidized enzyme-NADH-NAD ϩ complex, and these complexes could be involved in the catalytic cycle of Pb5R.
Based on these data, we present here a new model of the reaction sequence of b5R containing the neutral blue semiquinone form and the oxidized enzyme-NADH-NAD ϩ ternary complex as intermediates (Scheme II). This model contains the following processes: (i) formation of the oxidized enzyme-NADH complex (E-FAD-NADH), (ii) conversion of E-FAD-NADH to a form that has the ability to transfer H Ϫ (E-FAD-NADH*), (iii) H Ϫ transfer from NADH to FAD, (iv) the first one-electron transfer from the two-electron reduced enzyme complex (E-FADH Ϫ -NAD ϩ ), (v) rapid conversion of the neutral blue semiquinone form (E-FADH⅐ Ϫ -NAD ϩ ) to the anionic red semiquinone form (E-FAD⅐ Ϫ -NAD ϩ ), (vi) the second one-electron transfer from E-FAD⅐ Ϫ -NAD ϩ , (vii) formation of the oxidized enzyme-NADH-NAD ϩ ternary complex (E-FAD-NADH-NAD ϩ ) by binding of NADH, and (viii) release of NAD ϩ . Although there is no direct evidence for the conversion of E-FAD-NADH to E-FAD-NADH* (process (ii)), the existence of the E-FAD-NADH complex, which has no ability to transfer H Ϫ , is a reasonable assumption. This is because H Ϫ transfer itself is generally very fast, and a two-step mechanism for pyridine nucleotide binding has been proposed for the related family enzymes, nitrate reductase (50), phthalate dioxygenase reductase (51,52), and FNR (31). In the WT and the T66S mutant, the rate-limiting step is process (ii). In the T66A mutant, the rate-limiting step is process (viii), and the rate of process (i) is faster than that of process (viii). In the T66V mutant, the rate-limiting step is process (v). This model reasonably interprets the data presented here, but more investigations are required before this model is established.
Massey and Hemmerich (53) proposed that the neutral blue semiquinone form is an obligatory intermediate in flavoproteins, which are involved in one-electron transfers. Murataliev et al. (54,55) suggested that the so-called "air-stable" blue semiquinone form of the housefly NADPH-cytochrome P-450 reductase, which is also a one-electron transfer flavoenzyme, is inactive and different from the catalytically competent semiquinone form. The neutral blue semiquinone form of a one-electron transfer flavoenzyme may be less active in the one-electron oxidation than the anionic red semiquinone form. It is of interest that the rate-limiting step of the nonphysiological diaphorase activity of leaf FNR is the reductive half-reaction and that a stable neutral blue semiquinone form is produced by anaerobic photoreduction in the absence of NADP ϩ (30,31). It is reasonable that the less active neutral blue semiquinone form is required for leaf FNR, because during photosynthesis, the one-electron reduced form of leaf FNR must be protected from unfavorable oxidation to form the twoelectron reduced FAD (FADH Ϫ ), which is necessary for the two-electron reduction of NADP ϩ . In contrast, b5R may not require the less active neutral blue semiquinone intermediate, because the physiological direction of the electron transfer catalyzed by b5R is from NADH to one-electron acceptors and opposite from that of leaf FNR.
The hydroxyl group of the corresponding Ser 90 in the Cterminal mutant of pea leaf FNR forms a hydrogen bond with the amide moiety on the nicotinamide ring of the pyridine nucleotide in the NADP ϩ and NADPH complexes (31). Although the mutations of the Thr 66 in Pb5R hardly affect the K m NADH values (Table I), the possibility that Thr 66 in Pb5R interacts with the amide moiety on the nicotinamide ring of pyridine nucleotide cannot be excluded. It seems that the apparent K m NADH value of Pb5R is mainly dependent on the affinity for 5Ј-ADP or the 5Ј-ADP-ribose moiety of the pyridine nucleotide. We have previously reported that the R63A and R63Q mutants of Pb5R did not bind to 5Ј-ADP-agarose and suggested that Arg 63 in Pb5R assists the binding of NAD ϩ and NADH by reducing electrostatic repulsion between the negative charges on the phosphates of pyridine nucleotide and FAD (28). In addition, replacement of the Thr 153 and Thr 156 residues in Pb5R, which are positioned close to the potential binding site of the nicotinamide moiety of NADH (Fig. 1b), with serine, alanine, and valine residues also hardly changed the K m NADH values (data not shown).
In contrast to the K m NADH values, the apparent K m Pb5 values of the T66A and T66V mutants were significantly larger than that of the WT. It is considered that the structural imperfections in the b5R-NAD ϩ -Pb5 ternary complexes, which were caused by the mutations, increased the K m Pb5 values. Meyer et al. (14) suggested that NAD ϩ optimizes the human b5R-b5 complex and modulates the electron transfer.
In conclusion, direct evidence for the rate-limiting step of Pb5R was provided using stopped flow spectrophotometry. The rate-limiting steps in the catalytic cycles of the T66V mutant, the WT and the T66S mutant, and the T66A mutant were the conversion of the neutral blue to the red semiquinone, the reduction of FAD in the oxidized enzyme-NADH complexes, and the release of NAD ϩ from the oxidized T66A-NADH-NAD ϩ ternary complex, respectively. The conserved Thr 66 in Pb5R participates in the modulations of the semiquinone forms, the release of NAD ϩ from the enzyme, and the specific electron transfer to Pb5.