Determination of the redox potentials and electron transfer properties of the FAD- and FMN-binding domains of the human oxidoreductase NR1 Robert D. Finn 1 , Jaswir Basran 2 , Olivier Roitel 2 , C. Roland Wolf 1 , Andrew W. Munro 2 , Mark J. I. Paine 1 and Nigel S. Scrutton 2 1 Biomedical Research Centre, University of Dundee, Ninewells Hospital and Medical School, Dundee, UK; 2 Department of Biochemistry, University of Leicester, UK Human novel reductase 1 (NR1) is an NADPH dependent diflavin oxidoreductase related to cytochrome P450 reduc- tase (CPR). The FAD/NADPH- and FMN-binding domains of NR1 have been expressed and purified and their redox properties studied by stopped-flow and steady-state kinetic methods, and by potentiometry. The midpoint reduction potentials of the oxidized/semiquinone ()315 ± 5 mV) and semiquinone/dihydroquinone ()365 ± 15 mV) couples of the FAD/NADPH domain are similar to those for the FAD/NADPH domain of human CPR, but the rate of hydride transfer from NADPH to the FAD/NADPH domain of NR1 is % 200-fold slower. Hydride transfer is rate-limiting in steady-state reactions of the FAD/NADPH domain with artificial redox acceptors. Stopped-flow studies indicate that hydride transfer from the FAD/NADPH domain of NR1 to NADP + is faster than hydride transfer in the physiological direction (NADPH to FAD), consistent with the measured reduction potentials of the FAD couples [midpoint potential for FAD redox couples is )340 mV, cf )320 mV for NAD(P)H]. The midpoint reduction potentials for the flavin couples in the FMN domain are )146 ± 5 mV (oxidized/semiquinone) and )305 ± 5 mV (semiquinone/dihydroquinone). The FMN oxidized/semi- quinone couple indicates stabilization of the FMN semiqui- none, consistent with (a) a need to transfer electrons from the FAD/NADPH domain to the FMN domain, and (b) the thermodynamic properties of the FMN domain in CPR and nitric oxide synthase. Despite overall structural resemblance of NR1 and CPR, our studies reveal thermodynamic simi- larities but major kinetic differences in the electron transfer reactions catalysed by the flavin-binding domains. Keywords: novel reductase 1; cytochrome P450 reductase; flavoprotein; potentiometry; kinetics. Human novel reductase 1 (NR1) is a new member of the growing family of diflavin reductases that contain both FAD and FMN prosthetic groups [1]. In mammalian systems, cytochrome P450 reductase (CPR) was the first diflavin reductase isolated [2,3], followed by the isoforms of nitric oxide synthase (NOS [4,5]); and methionine synthase reductase (MSR [6]). Bacterial members of the family include flavocytochrome P450 BM3 (CYP102 [7]); and sulfite reductase [8]. CPR is the most extensively character- ized member of the mammalian diflavin reductases. In eukaryotic cells, type II cytochromes P450 are located in the endoplasmic reticulum, where they receive electrons from CPR. Like all members of the diflavin reductase family, CPR accepts electrons from NADPH, an obligatory 2-electron donor. These electrons are then transferred in a finely coupled stepwise manner to various physiological redox acceptor proteins, in the case of CPR to the P450 enzymes bound to the endoplasmic reticulum [9]. CPR is a 78-kDa membrane-bound flavoprotein and is likely to have evolved by the fusion of two ancestral genes encoding proteins related to ferredoxin-NADP + reductase (FNR) and flavodoxin (Fld) [10], bringing the two flavins (FAD and FMN) in close proximity for electron transfer. The enzyme also transfers electrons to cytochrome b 5 [11], haem oxygenase [12], and the fatty acid elongation system [13]. CPR can also reduce a number of artificial redox acceptors [14,15] and drugs [16–20], and may also have a role in the generation of reactive oxygen species in the cell. The recent cloning and expression of a cDNA encoding protein NR1 in insect cells has established functional similarities with CPR [1]. As with CPR, human NR1 catalyses the NADPH-dependent reduction of cytochrome c and various other electron accepting compounds. How- ever, overall the enzymatic activities are substantially less than those seen for CPR. NR1 also supports the NADPH- dependent reduction of the quinone antineoplastic agent Correspondence to N. S. Scrutton, Department of Biochemistry, University of Leicester, University Road, Leicester LE1 7RH. Fax: + 44 116 252 3369, Tel.: + 44 116 223 1337, E-mail: nss4@le.ac.uk or M. J. I. Paine, Biomedical Research Centre, University of Dundee, Ninewells Hospital and Medical School, Dundee, DD1 9SY. Fax: + 44 1382 669993, Tel.: + 44 1382 496420, E-mail: m.j.paine@dundee.ac.uk Abbreviations: CPR, cytochrome P450 reductase; DCPIP, 2,6-dichlo- rophenolindophenol; MSR, methionine synthase reductase; NOS, nitric oxide synthase; NR1, novel reductase 1; FDR, ferredoxin NADP + reductase; FLD, flavodoxin. (Received 20 November 2002, revised 15 January 2003, accepted 21 January 2003) Eur. J. Biochem. 270, 1164–1175 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03474.x doxorubicin, and menadione [1], a functional property also shared by the NOS family of enzymes [21,22]. The biological role of NR1 is unknown, but it is expressed at high levels in a wide range of cancer cell lines suggesting a role in the metabolic activation of bioreductive drugs. The lack of a membrane anchor in NR1 suggests that reduction of the P450 enzymes attached to the endoplasmic reticulum is an unlikely physiological role. In this paper we have characterized in detail the redox and electron transfer properties of the component flavin- binding domains of protein NR1. These studies have enabled us to make detailed comparison with equivalent studies performed on the flavin-binding domains of human CPR [23,24]. Despite the structural similarity of NR1 and CPR inferred from alignment of protein sequences, we demonstrate major functional differences between the two enzymes. Studies with the isolated FAD-domain reveal that hydride transfer from NADPH to FAD is substantially impaired in NR1, accounting for the poor catalytic rates in steady-state studies with various redox acceptors. The reduction potentials of the FAD and FMN redox couples in NR1 are similar to those in CPR and NOS, suggesting that impaired hydride transfer is attributed to less favour- able alignment of the nicotinamide coenzyme with FAD rather than a thermodynamic effect. Possible reasons for the poor hydride transfer rates in NR1 are discussed. Experimental procedures Chemicals and reagents All chemicals were purchased from Sigma (Poole, Dorset, UK) and all enzymes from Life Technologies Inc. (Paisley, UK), except where stated. The synthesis of A-side deuter- ated NADPD was as described in our previous work with NOS [25]. Expression constructs for the NR1 flavin-binding domains The cDNA encoding NR1 was cloned previously from MCF7 cells by RT-PCR [1]. The construction of an expression clone (pFAD-PET) suitable for production of the NR1 FAD/NADPH domain has been described [1]. The NR1 FMN domain construct was generated by PCR amplification using Pfu polymerase (Stratagene) and using the oligonucleotides 5¢-TGGAATCCATATGCCGAG CCCGCAGCTTCTG-3¢ and 5¢-GGAATTCCGGATCC TTAGGGCAGGGGGACTCC-3¢ as forward and reverse primers, respectively. Following amplification, the PCR product was cloned into pCR Blunt (Invitrogen) and sequenced to verify clone integrity. The NR1 FMN domain coding sequence was subcloned into the unique NdeI/XhoI sites of pHRT (patent pending). The resulting plasmid termed pHRT-NR1 FMN was used for expression of the FMN domain. Recombinant protein expression and purification For expression of the various domains in Escherichia coli, strain BLR (DE3/pLysS) containing the appropriate plasmid strains were grown overnight at 37 °CinLB broth containing ampicillin (50 lgÆmL )1 )andchloram- phenicol (34 lgÆmL )1 )toaD 600 of 0.4–0.8. Isopropyl thio-b- D -galactoside was then added (0.5 m M and 1 m M for the FAD-domain and FMN-domain constructs, respectively) to initiate protein expression, and cultures were incubated for a further 12 h at 30 °C. Cells were harvested by centrifugation (5000 g, 20 mins) and resus- pended in binding buffer (20 m M sodium phosphate buffer, pH 8.0, 500 m M NaCl, 5 m M imidazole and 10% glycerol). The NR1-FAD domain was purified over nickel agarose and 2¢,5¢-ADP Sepharose as described previously [1]. For purification of the FMN domain, cell suspensions were lysed by incubating at 30 °C for 15 min in the presence of 100 lgÆmL )1 lysozyme, followed by 30 min at 4 °Cinthe presence of 0.1% Triton X-100. The lysates were sonicated (MSE probe, several short bursts at high power) and centrifuged (40 000 g,30min,4°C). The supernatants were filtered through a 0.45-lm filter before being loaded on a Hi-trap nickel column. The column was washed sequentially with binding buffer and binding buffer containing 20 m M imidazole. The bound protein was eluted with binding buffer containing 350 m M imidazole. The eluted NR1 FMN domain was exchanged into Ôthrombin cleavage bufferÕ (20 m M Tris/HCl pH 8.4, 150 m M NaCl, 2.5 m M CaCl 2 ), and rebound to the nickel resin. Cleavage was performed at 4 °C overnight in the presence of thrombin at a concentration of 0.5 UÆmg )1 of protein. Cleaved NR1-FMN domain was separated from the nickel bound fusion tag by centrifugation (2000 g, 5 min). NR1-FMN domain was exchanged into 20 m M Tris/HCl buffer, pH 8.0, and loaded onto a Hi-Trap Mono Q column equilibrated with 20 m M Tris/HCl buffer, pH 8.0. The column was washed sequentially with 20 m M Tris/HCl buffer, pH 8.0, containing 50 m M NaCl and then 20 m M Tris/HCl buffer pH 8.0, containing 100 m M NaCl. NR1-FMN domain was eluted from the column with 20 m M Tris/HCl buffer, pH 8.0 containing 200 m M NaCl. Glycerol was added to 20% before the purified protein was stored at )70 °C. During purification, protein con- centrations were determined by Bradford analysis using Bio-Rad reagents and bovine serum albumin as a protein standard. Steady-state enzyme assays Reduction of prototypical cytochrome P450 reductase substrates dichlorophenol indophenol [DCPIP, 0–100 l M , e 600 ¼ 22 000 M )1 Æcm )1 ] and ferricyanide (0–250 l M , e 420 ¼ 1020 M )1 Æcm )1 ) was carried out using NR1 or CPR FAD/NADPH domain (700 pmol and 60 pmol enzyme, respectively) in 50 m M potassium phosphate buf- fer, pH 7.0, at 25 °C. The final assay volume was 1 mL. Apparent K m values for NADPH for the various FAD/ NADPH domains were determined by measuring the rate of potassium ferricyanide reduction at 25 °Cin50m M potassium phosphate buffer, pH 7.0, essentially as described previously for CPR [26]. Ferricyanide concentration was saturating at 250 l M . The NADPH concentration range was 0.5 l M to 100 l M . In attempts to reconstitute cytochrome c reductase activity, 100 pmol of either NR1 or CPR FAD/NADPH Ó FEBS 2003 Redox properties of human NR1 (Eur. J. Biochem. 270) 1165 domain was mixed with various amounts of the NR1 FMN domain, ranging from 0 to 2 nmol, in 50 m M potassium phosphate buffer, pH 7.0. The rate of reduction of cytochrome c in the presence of 200 l M NADPH and 100 l M horse heart cytochrome c at 25 °Cwasthen determined at 550 nm (e 550 ¼ 22 640 M )1 Æcm )1 )ona Varian Cary UV50 BioI scanning spectrophotometer. Stopped-flow kinetic studies Single turnover stopped-flow kinetic studies were per- formed using an Applied Photophysics SX.18 MV stopped-flow spectrophotometer. Measurements were car- ried out at 25 °Cin50m M potassium phosphate buffer, pH 7.0. Protein concentration (NR1 FAD domain) was 13 l M (reaction cell concentration) for photodiode array experiments and 4 l M (reaction cell concentration) for measurements in single wavelength mode. The sample- handling unit of the stopped-flow instrument was con- tained within a Belle Technology glove-box to maintain anaerobic conditions. All buffers were made oxygen-free by evacuation and extensive bubbling with argon before use. Prior to stopped-flow studies, protein samples were treated with potassium hexacyanoferrate to effect com- plete oxidation of the domains, and excess cyanoferrate was removed by rapid gel filtration (Sephadex G25). Treatment with hexacyanoferrate did not affect the kinetic behaviour of the domains. Stopped-flow, multiple-wavelength absorption studies were carried out using a photodiode array detector and X - SCAN software (Applied Photophysics Ltd). Spectral deconvolution was performed by global analysis and numerical integration methods using PROKIN software (Applied Photophysics Ltd). In single wavelength studies, flavin reduction by NADPH was observed at 454 nm or 600 nm. Transients at 454 nm were found to be mono- phasic and were analysed by fitting to a standard single exponential expression. For studies of electron transfer from 2 electron-reduced FAD/NADPH domain to NADP + , the FAD/NADPH domain was titrated to the 2-electron level with sodium dithionite. The reduced FAD/ NADPH domain was then mixed rapidly with NADP + . Reaction transients at 454 nm were monophasic and fitted using a single exponential expression. In studies of hydride transfer, the concentration of coenzyme was always at least 10-fold greater than enzyme concentration to ensure pseudo first order conditions. Reduction of the FAD/NADPH domain by NADPH was also analysed using fluorescence detection. Enzyme concentration was 4 l M and oxidation of NADPH was monitored by fluorescence emission at 450 nm (excitation 340 nm). Emission bands were selected using a bandpass filter (Coherent Optics; 450 nm #35–3367). Tryptophan emission was monitored at 340 nm (excitation 295 nm). A bandpass filter (Coherent Optics; # 35–3003) was used to select fluorescence emission. Electron transfer from NR1 FAD/NADPH domain to the FMN domain was monitored using a sequential mixing protocol in the stopped-flow instrument. In the first mix the FAD/NADPH domain (2 l M ) was reduced with NADPH (2 l M ). Following an appropriate delay time to allow reduction of the FAD the solution was mixed with varying concentrations of the FMN domain (5–20 l M ). Reactions were followed at 454 nm and a double exponential process best described the absorbance change. Potentiometry Redox titrations were performed in a Belle Technology glove-box under a nitrogen atmosphere. All solutions were degassed under vacuum with argon. Oxygen levels were maintained at less than 2 p.p.m. The protein was applied to a Bio-Rad Econo-Pac 10DG desalting column in the anaerobic box, pre-equilibrated with degassed 100 m M potassium phosphate (pH 7.0) buffer, to ensure removal of all traces of oxygen. The protein solutions (typically 50–100 l M in 5–8 mL buffer, both in the presence and absence of 10% v/v glycerol) were titrated electrochemically according to the method of Dutton [27] using sodium dithionite as reductant and potassium ferricyanide as oxidant. Dithionite and ferricyanide were delivered in 0.2 lL aliquots from concentrated stock solutions (typically 10–50 m M ). Mediators were added to facilitate electrical communication between enzyme and electrode, prior to titration. Typically, 2 l M phenazine methosulfate, 5 l M 2-hydroxy-1,4-naphthoquinone, 0.5 l M methyl viologen, and 1 l M benzyl viologen were included (to mediate in the range between +100 to )480 mV, as described previously [23,28]). At least 15 min was allowed to elapse between each addition to allow stabilization of the electrode. Spectra (250–750 nm) were recorded using a Cary UV-50 Bio UV-Visible scanning spectrophotometer. The electroche- mical potential of the solution was measured using a Hanna pH 211 meter coupled to a Pt/Calomel electrode (Thermo- Russell Ltd) at 25 °C. The electrode was calibrated using the Fe 3+ /Fe 2+ EDTA couple as a standard (+108 mV). A factor of +244 mV was used to correct relative to the standard hydrogen electrode. Data manipulation and analysis were performed using ORIGIN (Microcal). For the FMN and FAD/NADPH domain titrations, absorbance values at wavelengths of 454 nm (close to the absorption maximum for oxidized flavin) and 585 nm or 600 nm (near the absorption maximum for the blue semiquinone form of flavin) were plotted against potential. Data for the titration of the individual NR1 FAD/NADPH and FMN domains were fitted to Eqn (1), which represents a 2-electron redox process derived by extension to the Nernst equation and the Beer–Lambert law, as described previously [23,28] A ¼ a10 ðEÀE 0 1 Þ=59 þ b þ c10 ðE 0 2 ÀEÞ=59 1 þ 10 ðEÀE 0 1 Þ=59 þ 10 ðE 0 2 ÀEÞ=59 ð1Þ In Eqn (1), A is the total absorbance, a, b and c are component absorbance values contributed by the relevant flavin in the oxidized, semiquinone and reduced states, respectively. E is the observed potential, and E 1 ¢ and E 2 ¢ are the midpoint potentials for oxidized/semiquinone and semiquinone/reduced couples, respectively, for the relevant flavin. In using Eqn (1) to fit the absorbance-potential data for the single-flavin systems (i.e. the isolated FAD/NADPH and FMN domains), the variables were unconstrained, and regression analysis provided values in close agreement to those of the initial estimates. 1166 R. D. Finn et al. (Eur. J. Biochem. 270) Ó FEBS 2003 Results Expression and purification of the NR1 domains We have previously expressed and purified full-length NR1 from insect Sf9 cells [1], but the yields of recombinant enzyme are insufficient for the detailed biophysical analyses described in this paper. We have therefore attempted to develop E. coli-based expression systems to generate suffi- cient quantities of recombinant enzyme for biophysical studies. Repeated attempts to express soluble full-length NR1 in E. coli were unsuccessful, but using a similar approach to that taken with CPR [23,24,29], we have dissected NR1 using recombinant DNA methods (see above) and expressed the individual FMN and FAD/ NADPH domains separately, in soluble form (Fig. 1) that are suitable for kinetic and thermodynamic studies. The NR1-FMN domain was constructed from residues 1–174, and encodes a polypeptide with a calculated molecularmassof% 20 kDa, while the NR1-FAD/ NADPH domain encodes a % 48 kDa peptide spanning residues 194–597. The peptide sequence between residues 175–193 that forms the interdomain linker region was absent from the constructs as it led to insolubility of both domains. The presence of this linker region might explain the problems we experienced with the expression of soluble full-length NR1 in E. coli. The recombinant domains were His-tagged to facilitate affinity purification by nickel agarose chromatography, and purified to homogeneity (Fig. 2A). In the case of the NR1-FAD/NADPH domain, a second 2¢,5¢-ADP-Sepharose affinity step was also incor- porated in the purification scheme, taking advantage of its nucleotide binding capacity of the resin. Following purifi- cation, the His-tag was removed from the NR1-FMN domain by protease cleavage, but routinely not from the NR1-FAD/NADPH domain as the His-tag was ineffi- ciently cleaved from this domain. The presence of the His- tag on the FAD/NADPH domain had no apparent effect on catalytic activity. The purified domains were yellow, indicating the presence of bound cofactor, and they displayed UV-visible absorp- tion spectra characteristic of flavin containing enzymes (Fig. 2B). The fully oxidized FMN domain had absorbance maxima at 376 and 454 nm, and the FAD/NADPH domain absorbed maximally at 376 and 453 nm. The FMN domain was stable for several weeks upon storage at )20 °C at high concentration (>100 l M ). However, the FAD/NADPH domain appeared less stable and had a tendency to aggregate over time, particularly if subjected to Fig. 1. Schematic overview of the domains generated for this study in the context of full-length NR1. Fig. 2. Expression of NR1 domains in E. coli. (A) SDS/12% PAGE analysis of purified NR1-FMN (FMN) and NR1-FAD/NADPH (FAD) domains (2 lgÆlane )1 ). (B) Absorption spectra of purified NR1- FMN and NR1-FAD/NADPH domains (approx. 7 and 8 l M protein, respectively). Ó FEBS 2003 Redox properties of human NR1 (Eur. J. Biochem. 270) 1167 cycles of freezing and thawing. In addition, potentiometric studies (see below) revealed a strong tendency for the FAD/ NADPH domain to aggregate at low potentials. Steady-state enzyme activities The catalytic activity of the NR1-FAD/NADPH domain was examined and compared with the CPR FAD/NADPH domain. The CPR FAD/NADPH domain retains trans- hydrogenase activity, and is capable of reducing a range of electron acceptors [29]. To assess the functional activity of the NR1-FAD/NADPH domain, we measured the specific activities for ferricyanide and DCPIP (Table 1). The ferricyanide and DCPIP reductase activities of the NR1- FAD/NADPH domain were approximately 29-fold and 5-fold lower than CPR-FAD/NADPH, respectively, con- sistent with differences in activity levels found between the full-length enzymes [1]. We hypothesized that differences in catalytic activity between CPR and NR1 might be explained by differences in the apparent affinity for NADPH. Thus, the apparent K m values for NADPH were determined by further steady-state kinetic assays of potassium ferricyanide reduction. The ferricyanide was maintained at saturating concentration (250 l M ), and NADPH concentration was varied between 0.5 and 100 l M . In this system, the apparent K m for NADPH was lower for the FAD/NADPH domain of NR1 (1.08 ± 0.12 l M ) than for the FAD/NADPH domain of CPR (2.62 ± 0.40 l M ). However, both domains clearly bind NADPH tightly, and differences in catalytic activity are evidently not associated with NADPH binding. Mixing the FAD/NADPH and FMN domains of NR1 under the conditions described in Experimental procedures did not stimulate to any considerable extent (<10%) the cytochrome c reductase activity of the system over that observed for the isolated FAD/NADPH domain, which had a k cat of 2.2 ± 0.25 min )1 (0.037 ± 0.004 s )1 ). Intact CPR-like diflavin reductase enzymes typically have high levels of cytochrome c reductase activity, with electron transfer to the cytochrome mediated by the FMN cofactor. A more substantial increase in specific activity for cyto- chrome c reduction was observed through mixing the human CPR FMN domain with its FAD/NADPH domain (% 16-fold) (29). However, the domain mixture still recon- stituted only % 2% of the activity of the full length CPR. In thecaseofNR1,thek cat for the full length enzyme is low (1.3 s )1 ) and is clearly gated largely by the hydride transfer event [1]. While domain reconstitution in NR1 only has a marginal effect on cytochrome c reduction rate, the reconstituted activity is % 3% of that in full length NR1, similartotheratioachievedwithCPRdomains. To explore further the influence of the hydride transfer step on catalytic activity of the FAD/NADPH domains of NR1 and CPR, we compared ferricyanide reduction rates using both NADPH and A-side deuterated coenzyme (NADPD) under saturating substrate conditions (100 l M NADPH/D and 250 l M ferricyanide) under standard conditions described in Experimental procedures. In the steady-state, ferricyanide reduction was slower for both enzymes using NADPD as reductant. A deuterium isotope effect of 2.5 was observed on CPR FAD/NADPH domain- catalysed ferricyanide reduction, and of 3.5 for the NR1 FAD/NADPH-catalysed process. The value of the KIE on steady-state activity for NR1 FAD/NADPH domain is consistent with that observed in stopped-flow studies of hydride transfer to the flavin (see below) and with this event being rate-limiting in reductive catalysis using artificial electron acceptors. Potentiometric analysis of the component domains Anaerobic spectroelectrochemical titrations of the isolated FAD/NADPH and FMN domains of NR1 enabled the determination of the midpoint reduction potentials for the oxidized/semiquinone (E 1 ) and semiquinone/hydroquinone (E 2 ) flavin couples. Data for both domains were fitted to the 2-electron Nernst function described in Experimental pro- cedures [23,28]. The FMN domain did not aggregate to any extent over the course of % 5–8 h required to complete the titrations. However, for the FAD/NADPH domain of NR1, considerable aggregation and precipitation of the protein occurred within 1 h of initiating the titrations, and this precipitation was accelerated at more negative poten- tials (<350 mV). To correct for baseline shifts in the titrations for both NR1 FAD/NADPH and FMN domains, spectra were manipulated by subtracting absorp- tion at 800 nm in each sample (where there is negligible absorption contribution from flavins in any redox state) across the entire spectrum. In the case of the FAD/NADPH domain, attempts were made to account for turbidity caused by protein aggregation by multiplying individual spectra by a correction factor (1–1/k;wherek is the absorption wavelength). However, correction did not improve to any large extent spectra for which the A 800 value had increased above approximately 0.05 units. To enable accurate determination of the FAD potentials, titrations of the NR1 FAD/NADPH domain were per- formed on samples of identical concentration over small ranges (100–150 mV) of the potential range, moving on to a fresh sample when turbidity proved excessive. In this way, data across the entire range were collected and were of suitable quality for determination of both redox couples for the FAD/NADPH domain flavin. Table 1. Apparent turnover numbers for DCPIP and ferricyanide reduction by the FAD/NADPH binding domains of CPR and NR1. Apparent turnover numbers were determined at 25 °Cin50m M potassium phosphate buffer, pH 7.0, as described in Experimental procedures, by monitoring the reduction of substrate (DCPIP or ferricyanide) at appropriate wavelengths (600 nm and 420 nm, respectively). Results are the mean and standard deviation of triplicate assays. For the determination of apparent k cat values, experiments were performed at a saturating concentration of NADPH (200 l M ), over a wide range of concentrations of the substrate (0–100 l M for DCPIP and 0–700 l M for ferricyanide). Values were determined by curve fitting to the Michaelis–Menten equation and are the mean and standard deviation of three separate experiments. Domain Substrate DCPIP (k cat , s )1 ) Ferricyanide (k cat , s )1 ) CPR FAD/NADPH 4.57 ± 0.18 65.33 ± 2.20 NR1 FAD/NADPH 0.86 ± 0.02 2.27 ± 0.13 1168 R. D. Finn et al. (Eur. J. Biochem. 270) Ó FEBS 2003 Visible absorption spectra collected during the redox titration of the NR1 FMN domain are shown in Fig. 3. The oxidized domain has typical flavin absorption maxima at 454 nm and 376 nm, and forms a neutral blue semiquinone during dithionite titration, indicating that the potential for the ox/sq couple is more positive that that for the sq/red couple. The semiquinone has absorption maximum at 585 nm, with a shorter wavelength maximum at 352 nm. Data from the absorption maximum of the oxidized flavin (A 454 ) and near the semiquinone maximum (A 600 )were fitted to the 2-electron Nernst function (Fig. 4), and yielded essentially identical data for the ox/sq ()152±4mV; )146 ± 5 mV, respectively) and sq/hq couples ()304 ± 8mV;)305 ± 5 mV) (Fig. 4). Absorption spectra collected during the redox titrations of the NR1 FAD/NADPH domain are shown in Fig. 5. As with the FMN domain, a neutral blue semiquinone is stabilized. However, the maximal intensity of the semiqui- none is much less than that observed for the NR1 FMN domain, suggesting that the midpoint reduction potentials for the FAD ox/sq and sq/hq couples are much closer together than those for the FMN. Typically for this class of diflavin enzymes, the potentials for the FAD/NADPH domains are rather more negative than those for the FMN domain, reflecting the direction of electron transfer from NADPH through FAD, then FMN and on to the final electron acceptor(s) [23,30]. The NR1 FAD/NADPH domain follows this trend, with midpoint reduction poten- tials of )315 ± 5 mV (ox/sq) and )365 ± 15 mV (sq/hq) derived from fitting titration data at the semiquinone absorption maximum (585 nm) to the 2-electron Nernst function (Fig. 6). The tendency of the NR1 FAD/NADPH domain to aggregate and precipitate during the redox titrations (and to do so particularly rapidly at more negative potentials) explains the larger error for the midpoint potential for the sq/hq couple. This aspect of FAD/ NADPH domain behaviour is shared also by the homo- logous FAD/NADPH domains of human CPR [23] and flavocytochrome P450 BM3 [28]. Aggregation of the NR1 FAD/NADPH domain was much less extensive in the absence of cofactor reduction. Stopped-flow kinetic studies Reduction of the FAD/NADPH domain of NR1 was investigated by stopped-flow methods using a photodiode array detector. Aggregation of the domain was not observed over the short time periods used in stopped-flow experi- ments. The spectral changes accompanying flavin reduction (Fig. 7) revealed the absence of major spectral change in the long wavelength region (550 nm to 650 nm). This contrasts with similar studies with the FAD/NADPH domain of human CPR where rapid absorption increases in this region, attributable to the formation of an oxidized enzyme- NADPH charge-transfer species, that accumulates prior to flavin reduction. Global fitting of the spectral changes for the NR1 FAD/NADPH domain indicated the presence of only one detectable kinetic phase corresponding to FAD reduction; FAD reduction proceeds with an observed rate constant of 1.07 ± 0.02 s )1 . In single wavelength studies at 454 nm the absorption changes reporting on FAD reduc- tion were monophasic, consistent with a single step kinetic model, and the observed rate of FAD reduction was found to be independent of coenzyme concentration in the pseudo first order regime (Fig. 8A; Table 2). Studies at 600 nm indicated that a spectroscopically distinct NADPH-E ox Fig. 3. Spectral changes during redox titration of the FMN domain of human NR1. Anaerobic spectroelectrochemical titration was performed as described in Experimental procedures. The oxidized FMN domain is shown as a thick solid black line, and has the highest absorption at 454 nm, with the second major band at 376 nm. The other spectrum shown by a thick solid line is that at which the blue FMN semiquinone is maximally populated. The spectral maxima for this species are located at approximately 585 nm and 352 nm. Spectra collected during addition of the first electron (oxidized-to-semiquinone transition) are indicated by thin, solid black lines. Spectra collected during addition of the second electron to the flavin (semiquinone-to-hydroquinone transition) are indicated by dotted lines. Isosbestic points for the ox/sq [1] and sq/hq [2] couples are locatedat approximately 501 nm and 434 nm, respectively. Approximately 100 spectra were collected across the relevant range of potentials. For clarity, only selected spectra are shown. Ó FEBS 2003 Redox properties of human NR1 (Eur. J. Biochem. 270) 1169 species did not accumulate prior to flavin reduction, as was seen for the isolated FAD/NADPH domain of CPR [24]. In studies with CPR [24], NOS [25] and the adrenodoxin reductase homologue FprA from Mycobacterium tubercu- losis (which is related structurally to the FAD/NADPH domains of the diflavin reductase family [32]; K. McLean, N. S. Scrutton & A. W. Munro, unpublished results) the observed rate of hydride transfer accelerates as the coen- zyme concentration is decreased to levels that are stoichi- ometric with the enzyme concentration. This unusual kinetic behaviour has been attributed to the presence of a second, regulatory coenzyme-binding site the occupation of which attenuates hydride transfer from the catalytic site at high NADPH concentrations. This behaviour is not observed with the FAD/NADPH domain of NR1 and highlights a major difference in the kinetic properties of NR1 compared with other diflavin reductase enzymes, despite the overall inferred structural similarity. The flavin reduction rate for NR1 FAD/NADPH domain is 0.27 ± 0.1 in studies performed with A-side deuterated coenzyme (NADPD; Fig. 8A, inset), yielding a kinetic isotope effect of 3.7. This is consistent with the absorption change at 454 nm reporting on the hydride transfer step and with hydride transfer being fully rate-limiting in steady-state turnover with ferricyanide as electron acceptor (see above). Reactions performed over an extended time base under aerobic conditions with absorption detection at 454 nm gave access to the flavin re-oxidation rate for stoichiometrically reduced FAD/ NADPH domain. In this case, re-oxidation occurred with an observed rate constant of 0.025 ± 0.0005 s )1 . Given the reduction potentials of the FAD ox/sq and FAD sq/hq couples of the isolated FAD/NADPH domain (midpoint potential for the 2-electron couple, E 12 ¼ )340±10mV)inrelationtothatofNADPH()320 mV) we undertook a study of the reverse hydride transfer reaction from dihydroquinone FAD/NADPH domain to NADP + . Enzyme was initially titrated to the 2-electron level with sodium dithionite under anaerobic conditions and mixed rapidly with NADP + . Absorption transients were monophasic at 454 nm (Fig. 8B), and the observed rate constants for FAD oxidation were independent of NADP + concentration (Table 2). The rate of hydride transfer is % 2.5-fold faster in the ÔreverseÕ direction and similar observations have been made with FAD/NADPH domain of human CPR, where the midpoint potential for the 2-electron flavin couple ()329 ± 7 mV) is also more negative than that for NADPH [23,24]. Fluorescence detection was also used in stopped-flow studies of enzyme reduction by NADPH. NADPH fluor- escence was used in our previous studies with human CPR and NOS to follow NADPH oxidation. However, reduction of the FAD/NADPH domain of NR1 by NADPH is not accompanied by a change in fluorescence emission at 450 nm following excitation at 340 nm for reasons that are as yet are unclear. Changes in tryptophan fluorescence emission do, however, accompany reduction of the FAD/ NADPH domain (Fig. 8C). Unlike with CPR FAD/ NADPH domain (which gives rise to a fluorescence decrease on flavin reduction), fluorescence transients dis- played an increase in fluorescence emission. The rapid increase in fluorescence observed with the CPR domain prior to flavin reduction, which reports on coenzyme binding, is not observed in the NR1 domain transients. Observed rate constants for the monophasic fluorescence increase with the NR1 FAD/NADPH domain are inde- pendent of coenzyme concentration and are similar in value to the rate constants determined from absorption measure- ments at 454 nm for flavin reduction (Table 2). The ability of the reduced FAD/NADPH domain to transfer electrons to the oxidized FMN domain was studied by sequential stopped-flow methods. In the first mix the FAD/NADPH domain was mixed with stoichiometric NADPH, and the reduced domain was then mixed with the oxidized FMN domain in a second mix. Reaction transients measured at 454 nm were biphasic and the observed rate constants calculated for both the fast and slow phases were independent of coenzyme concentration (Table 2). Technical difficulties owing to aggregation of the FAD/NADPH domain in dithionite titrations preven- ted detailed analysis of electron transfer between dithionite reduced FAD/NADPH domain and the oxidized FMN domain. The lack of a second order dependence of the observed rate for interdomain electron transfer as the Fig. 4. Absorbance vs. potential plots for the FMN domain of human NR1. (A)PlotofA 600 (near the blue semiquinone maximum) vs. reduction potential fitted to a 2-electron Nernst function, as described in Experimental procedures. (B) Plot of A 454 data (at the oxidized flavin maximum) from the same titration, also fitted to the 2-electron Nernst function. Midpoint reduction potentials for the ox/sq ()152±4mV and 146±5mV) and sq/hq ()304±8mV; )305 ± 5 mV) couples of the flavin determined from fits to both data sets are identical within error. 1170 R. D. Finn et al. (Eur. J. Biochem. 270) Ó FEBS 2003 concentration of the FMN domain is increased indicates that the reaction rate is controlled by some process other than collision of the two flavin-binding domains. Discussion The ability to dissect CPR into distinct functional domains has assisted in providing detailed structural and kinetic information about the redox and structural properties of CPR [24,33–35]. The results of this study show that NR1, which is related structurally to CPR, can be dissected into distinct functional domains. Owing to the difficulties in expressing the full-length protein, the ability to isolate individual flavin-binding domains has facilitated the deter- mination of both thermodynamic and kinetic properties of NR1, and as we have shown for CPR and P450 BM3 (23,24,28) the properties of the flavin binding domains of NR1 are likely to mimic the redox properties of the domains in full-length enzyme. In steady-state assays, the NR1-FAD/NADPH domain has significantly lower catalytic activity for prototypical reductase substrates compared to the CPR-FAD/NADPH domain, in agreement with previously reported findings for intact NR1 [1] (Table 1). Stopped-flow studies with this domain indicate that flavin reduction occurs relatively slowly (% 1s )1 ) as a monophasic process, and flavin reduction displays a KIE of 3.7 in reactions with A-side NADPD. The slow reduction of FADby NADPH is limiting in steady-state reactions as indicated by the KIE of 3.5 observed for NR1 FAD/NADPH domain-catalysed ferricyanide reduction. The apparent turnover number with ferricyanide (2.27 s )1 ) is approximately twice the hydride transfer rate (% 1s )1 ) measured in stopped-flow studies, consistent with it being a one-electron acceptor. Comparable studies with CPR indi- cate more complex behaviour; in this case flavin reduction is biphasic (observed rate constants % 200 s )1 and % 3s )1 )and the kinetic mechanism for flavin reduction is shown in Scheme 1 (for further details and experimental data sup- porting the assignment of observed rate constants to kinetic phases see [24]). The fast phase (200 s )1 ) represents the rapid formation of an equilibrium between an oxidized enzyme- NADPH complex and reduced enzyme-NADP + complex (species CT2). The slow phase (% 3s )1 ) is attributed to the Fig. 6. Absorbance vs. potential plot for the FAD/NADPH domain of human NR1. Plot of A 585 (near the blue semiquinone maximum) vs. reduction potential was fitted to a 2-electron Nernst function, as des- cribed in Experimental procedures. The fit yields midpoint reduction potential values of )315 ± 5 mV for the oxidized/semiquinone cou- ple, and )365 ± 15 mV for the semiquinone/hydroquinone couple. Fig. 5. Spectral changes during redox titration of the FAD/NADPH domain of human NR1. Anaerobic spectroelectrochemical titration was performed as described in the Experimental procedures. For clarity, only selected spectra are shown. The highest intensity spectrum is that of oxidized FAD/NADPH domain, and is shown as a thick solid black line with absorption maxima at 376 and 453 nm. The isosbestic point for the oxidized-to-semiquinone transition [1] is located at approximately 501 nm. Solid lines indicate spectra recorded during addition of the first electron (ox/sq transition), whereas dotted lines shows spectra recorded during addition of the second electron (sq/hq transition). The tendency of the protein to aggregate at negative potentials, along with the low potential for the semiquinone/hydroquinone couple of the FAD ()365 ± 15 mV) prevented collection of useful spectral data at potentials below %)430 mV. Ó FEBS 2003 Redox properties of human NR1 (Eur. J. Biochem. 270) 1171 release of NADP + with concomitant displacement of the equilibrium distribution of enzyme species towards further reduction of the FAD (i.e. further transfer of hydride equivalents from NADPH to the FAD is gated by the release of NADP + ). A similar mechanism has also been suggested Fig. 7. Spectra. (A) Spectral changes accompanying the reduction of the FAD/NADPH domain of NR1 by NADPH. Conditions: 50 m M potassium phosphate buffer, pH 7.0; 25 °C. Protein concentration 13 l M ; NADPH concentration 130 l M . (B) Initial and end spectrum obtained from fitting to a single step kinetic model. Observed rate constant for FAD reduction 1.07 ± 0.01 s )1 . Fig. 8. Absorbance and fluorescence changes accompanying flavin reduction in the FAD/NADPH domain of NR1 by NADPH. Condi- tions: coenzyme concentration, 100 l M ; enzyme concentration 4 l M ; 50 m M potassium phosphate buffer, pH 7.0, 25 °C. (A) Monophasic absorption transient at 454 nm for reduction of the FAD/NADPH domain by NADPH and NADPD (inset); coenzyme concentration 100 l M . Observed rate constants calculated at different concentrations of NADPH are given in Table 2. (B) Monophasic absorption transient at 454 nm for oxidation of the FAD/NADPH domain by NADP + . Enzyme was initially reduced at the 2-electron level by titration with sodium dithionite in the presence of methyl viologen. Observed rate constants calculated at different concentrations of NADP + are given in Table 2. (C) Tryptophan fluorescence emission transient observed during the reduction of the FAD/NADPH domain with NADPH. Observed rate constants calculated at different concentrations of NADPH are given in Table 2. In all panels, the solid black line is the fit to the experimental data (shown in greyscale). 1172 R. D. Finn et al. (Eur. J. Biochem. 270) Ó FEBS 2003 for reactions of the reductase domain of NOS with NADPH [25]. The steady-state turnover value (65 s )1 )fortheFAD/ NADPH domain of CPR in reactions with NADPH and ferricyanide is much faster than the slow NADP + release step observed in stopped-flow studies. With CPR we suggest therefore that ferricyanide oxidizes the EH 2 NADP + form of the FAD/NADPH domain, and that subsequent release of NADP + from ENADP + occurs at a faster rate than from 2-electron reduced enzyme (i.e. EH 2 NADP + ) (Scheme 1). That enzyme oxidation occurs from the EH 2 NADP + species of the NADPH/FAD domain of CPR is also consistent with the KIE value of 2.5 observed in steady-state reactions with ferricyanide (i.e. hydride transfer and not NADP + release is rate-limiting). Although NR1 is structurally related to CPR and NOS, the rate of hydride transfer in NR1 (% 1s )1 )is substantially less than the rates in CPR (% 200 s )1 [24]); and NOS (% 200 s )1 for the first hydride transfer reaction in the FAD-FMN reductase domain [25]). In searching for a structural reason for the substantially reduced rates of hydride transfer in NR1 we note the absence of a cysteine residue that corresponds to Cys630 in CPR; the equivalent residue in NR1 is Ala549 [1]. In CPR, Cys630 forms part of a catalytic triad with Ser457 and Asp675, and mutagenesis studies with rat CPR have demonstrated a key role for this residue in hydride transfer from NADPH to FAD [36,37]. Our own studies with flavocytochrome P450 BM3 also indicate that mutation of the equivalent cysteine residue in this enzyme to alanine substantially decreases the rate of flavin reduction and has an adverse effect on the FAD reduction potential (O. Roitel, N. S. Scrutton and A. W. Munro, unpublished work). Potentiometric studies of the isolated FAD/NADPH and FMN domains of NR1 have allowed us to establish that both flavins stabilize neutral blue semiquinones, and to determine the midpoint reduction potentials for the four redox couples of NR1 (Fig. 9). These data indicate that the relative potentials of the flavins are ordered similarly to Table 2. Summary of observed rate constants from stopped-flow kinetic studies. All reactions were performed in 50 m M potassium phosphate buffer, pH 7.0 at 25 °C. In studies with the isolated FAD/NADPH domain, protein concentration was 4 l M .Instudiesofinterdomainelectrontransfer, the FAD/NADPH domain was reduced with stoichiometric NADPH prior to a second mix with the FMN domain (see text for details). Errors are those from fitting to the average of at least five kinetic transients. FAD reduction (A 454 nm transient) Trp fluorescence emission FAD oxidation (A 454 nm transient) Interdomain electron transfer (A 454 nm transient) NADPH (l M ) k obs (s )1 ) NADPH (l M ) k obs (s )1 ) NADP + (l M ) k obs (s )1 ) FMN domain (l M ) k fast (s )1 ) k slow (s )1 ) 5 1.07 ± 0.01 4 1.97 ± 0.02 4 2.60 ± 0.02 5 1.20 ± 0.04 0.20 ± 0.02 15 1.07 ± 0.01 20 1.14 ± 0.01 10 2.74 ± 0.02 7.5 1.70 ± 0.04 0.12 ± 0.01 25 1.09 ± 0.01 40 1.12 ± 0.01 50 2.50 ± 0.01 10 1.29 ± 0.01 0.16 ± 0.01 50 0.99 ± 0.01 100 1.08 ± 0.01 100 2.31 ± 0.02 12.5 1.60 ± 0.02 0.10 ± 0.01 100 0.98 ± 0.01 200 1.02 ± 0.01 200 2.23 ± 0.02 15 1.59 ± 0.05 0.20 ± 0.01 200 1.00 ± 0.01 20 1.45 ± 0.02 0.12 ± 0.01 300 0.96 ± 0.01 Scheme 1. Kinetic mechanism for flavin reduction. A ox refers to the electron acceptor ferricyanide. Fig. 9. Flavin reduction potentials for members of the diflavin reductase enzyme family. The various midpoint reduction potentials for the oxidized/semiquinone (grey boxes) and semiquinone/hydroquinone couples (white boxes) of the FAD and FMN cofactors in the various diflavin reductases are shown diagrammatically. These are NR1 (this work), human cytochrome P450 reductase (CPR [23]), neuronal nitric oxide synthase (NOS [30]), and flavocytochrome P450 BM3 reductase (BM3 [28]). The midpoint reduction potential for the physiological reductant NAD(P)H ()320 mV) is shown as a dotted bar. Ó FEBS 2003 Redox properties of human NR1 (Eur. J. Biochem. 270) 1173 [...]... of the NR1 FAD domain (50 mV cf 85 mV for CPR FAD and 159 mV for NR1 FMN) explains the rather low intensity of the blue semiquinone signature at long wavelength that accumulates during redox titration The fact that the FAD potentials thermodynamically disfavour its reduction by NADPH is another likely factor in explaining the slow flavin reduction rate in NR1 As might be predicted on the basis of the. .. for human CPR However, the oxidized/semiquinone couple for the NR1 FMN is rather more negative than that determined for CPR ()146 mV cf )66 mV) In addition, the midpoint potential for the 2 -electron reduction of the NR1 FAD cofactor is also slightly more negative than that for CPR ()340 mV cf )329 mV) and that for NADPH ()320 mV) The smaller separation between the ox/sq and sq/hq midpoint potentials of. .. progenitors [9] The physiological role of FDR enzymes is to catalyse the reduction of NADP+ Thus, it appears likely that the role of the FLD domain in the diflavin reductases is to remove one (or both) electrons from the FAD hydroquinone, thus disfavouring the reverse reaction Our work with the NR1 flavin-binding domains has highlighted major kinetic differences in the kinetics of hydride transfer and intermediates... with CPR and NOS, despite the overall similar thermodynamic properties In future work, we intend to establish the physiological role of NR1, to obtain structural data for the FAD/NADPH and FMN domains, and to examine in greater detail the reasons underlying its slow, rate-limiting hydride transfer reaction Acknowledgements This work was funded by grants from the Medical Research Council, the Lister... Institute of Preventive Medicine and the Wellcome Trust The authors would like to thank the Biotechnology and Biological Sciences Research Council, the Medical Research Council and the European Union for financial support for these studies We are also grateful for helpful discussions with Professor Gordon Roberts and Dr Aldo Gutierrez (University of Leicester) N.S.S is a Lister Institute Research Professor... H-1,N-15 and C-13 NMR resonance assignment, secondary structure and global fold of the FMN-binding domain of human cytochrome P450 reductase J Biomol NMR 10, 63–75 34 Gutierrez, A., Doehr, O., Paine, M., Wolf, C.R., Scrutton, N.S & Roberts, G.C (2000) Trp-676 facilitates nicotinamide coenzyme exchange in the reductive half-reaction of human cytochrome P450 reductase: properties of the soluble W676H and. .. catalyze the activation of redox cycling and bioreductive anticancer agents Cancer Res 59, 1929–1934 Munro, A.W., Noble, M.A., Robledo, L., Daff, S.N & Chapman, S.K (2001) Determination of the redox properties of human NADPH-cytochrome P450 reductase Biochemistry 40, 1956–1963 Gutierrez, A., Lian, L.Y., Wolf, C.R., Scrutton, N.S & Roberts, G.C (2001) Stopped-flow kinetic studies of flavin reduction in human. .. reduction potentials, the dithionite-reduced NR1 FAD/NADPH domain catalyses NADP+ reduction approximately 2.5-fold faster than the NADPH reduces the enzyme FAD A similar phenomenon was observed for human CPR [24], and we consider that this behaviour reflects the evolutionary origins of the diflavin reductases, which have evolved from fusion of genes encoding ferredoxin NADP+-reductase (FDR) and flavodoxin... FMN and FAD in catalysis by liver microsomal NADPH- cytochrome P-450 reductase J Biol Chem 256, 266– 277 Dutton, P (1978) Redox potentiometry: determination of midpoint potentials of oxidation-reduction components of biological electron- transfer systems Methods Enzymol 54, 411–435 Daff, S.N., Chapman, S.K., Turner, K.L., Holt, R.A., Govindaraj, S., Poulos, T.L & Munro, A.W (1997) Redox control of the. .. kinetic studies of flavin reduction in human cytochrome P450 reductase and its component domains Biochemistry 40, 1964–1975 Knight, K & Scrutton, N.S (2002) Stopped-flow kinetic studies of electron transfer in the reductase domain of neuronal nitric oxide synthase: re-evaluation of the kinetic mechanism reveals new enzyme intermediates and variation with cytochrome P450 reductase Biochem J 367, 19–30 Vermilion, . Determination of the redox potentials and electron transfer properties of the FAD- and FMN-binding domains of the human oxidoreductase NR1 Robert. of both thermodynamic and kinetic properties of NR1, and as we have shown for CPR and P450 BM3 (23,24,28) the properties of the flavin binding domains of NR1