Interflavin electron transfer in human cytochrome P450 reductase is enhanced by coenzyme binding Relaxation kinetic studies with coenzyme analogues Aldo Gutierrez 1,2 , Andrew W. Munro 1 , Alex Grunau 1,2 , C. Roland Wolf 3 , Nigel S. Scrutton 1 and Gordon C. K. Roberts 1,2 1 Department of Biochemistry and 2 Biological NMR Centre, University of Leicester, UK; 3 Biomedical Research Centre, University of Dundee, Ninewells Hospital and Medical School, Dundee, UK The role of coenzyme binding in regulating interflavin electron transfer in human cytochrome P450 reductase (CPR) has been studied using temperature-jump spectros- copy. Previous studies [Gutierrez, A., Paine, M., Wolf, C.R., Scrutton, N.S., & Roberts, G.C.K. Biochemistry (2002) 41, 4626–4637] have shown that the observed rate, 1/s, of interflavin electron transfer (FAD sq ) FMN sq fi FAD ox ) FMN hq ) in CPR reduced at the two-electron level with NADPH is 55 ± 2 s )1 , whereas with dithionite- reduced enzyme the observed rate is 11 ± 0.5 s )1 ,sug- gesting that NADPH (or NADP + ) binding has an important role in controlling the rate of internal electron transfer. In relaxation experiments performed with CPR reduced at the two-electron level with NADH, the observed rate of internal electron transfer (1/s ¼ 18 ± 0.7 s )1 )is intermediate in value between those seen with dithionite- reduced and NADPH-reduced enzyme, indicating that the presence of the 2¢-phosphate is important for enhancing internal electron transfer. To investigate this further, tem- perature jump experiments were performed with dithionite- reduced enzyme in the presence of 2¢,5¢-ADP and 2¢-AMP. These two ligands increase the observed rate of interflavin electron transfer in two-electron reduced CPR from 1/s ¼ 11 s )1 to 35 ± 0.2 s )1 and32±0.6s )1 , respectively. Reduction of CPR at the two-electron level by NADPH, NADH or dithionite generates the same spectral species, consistent with an electron distribution that is equivalent regardless of reductant at the initiation of the temperature jump. Spectroelectrochemical experiments establish that the redox potentials of the flavins of CPR are unchanged on binding 2¢,5¢-ADP, supporting the view that enhanced rates of interdomain electron transfer have their origin in a con- formational change produced by binding NADPH or its fragments. Addition of 2¢,5¢-ADP either to the isolated FAD-domain or to full-length CPR (in their oxidized and reduced forms) leads to perturbation of the optical spectra of both the flavins, consistent with a conformational change that alters the environment of these redox cofactors. The binding of 2¢,5¢-ADP eliminates the unusual dependence of the observed flavin reduction rate on NADPH concentra- tion (i.e. enhanced at low coenzyme concentration) ob- served in stopped-flow studies. The data are discussed in the context of previous kinetic studies and of the crystallo- graphic structure of rat CPR. Keywords: coenzyme binding; cytochrome P450 reductase; electron transfer; flavoprotein; temperature-jump relaxation spectroscopy. Members of the cytochrome P450 mono-oxygenase super- family catalyse the hydroxylation of a wide range of physiological and xenobiotic compounds; in eukaryotes the type II cytochromes P450, located in the endoplasmic reticulum, play a key role in drug metabolism. NADPH- cytochrome P450 reductase (CPR; EC 1.6.2.4) catalyses electron transfer to these type II cytochromes P450 [1–5]. CPR is a 78-kDa enzyme containing one molecule each of FAD and FMN [6]. Sequence analyses [5] and X-ray crystallographic studies of rat CPR [7] have revealed that CPR consists of an N-terminal membrane anchor, respon- sible for its localization to the endoplasmic reticulum, and three folded domains: an FAD- and NADPH-binding domain related to ferredoxin-NADP + reductase, a flavo- doxin-like FMN-binding domain and a connecting ÔlinkerÕ domain. In addition to the cytochromes P450, CPR can donate electrons to cytochrome b 5 [8], haem oxygenase [9], the fatty acid hydroxylation system [10] and a number of artificial redox acceptors [11,12] and drugs [13–17]. CPR is related to a number of mammalian diflavin reductases including the isoforms of nitric oxide synthase [18], methionine synthase reductase [19] and protein NR1 Correspondence to N. S. Scrutton, Department of Biochemistry, University of Leicester, University Road, Leicester LE1 7RH, UK. Fax: + 44 116 252 3369, Tel.: + 44 116 223 1337, E-mail:nss4@le.ac.ukortoG.C.K.Roberts,BiologicalNMR Centre, University of Leicester, University Road, Leicester LE1 7RH, UK. Fax: + 44 116 223 1503, Tel.: + 44 116 252 2978, E-mail: gcr@le.ac.uk Abbreviations: CPR, NADPH-cytochrome P450 reductase; 2¢,5¢-ADP, adenosine 2¢,5¢-bisphosphate; 2¢-AMP, adenosine 2¢-monophosphate. Enzymes: NADPH-cytochrome P450 reductase (CPR; EC 1.6.2.4). Note: In this paper the term ÔintactÕ CPR refers to soluble CPR, containing the FMN-binding, FAD-binding and ÔlinkerÕ domains and lacking only the N-terminal membrane-anchoring peptide sequence. (Received 14 March 2003, revised 17 April 2003, accepted 24 April 2003) Eur. J. Biochem. 270, 2612–2621 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03633.x [20]. It also shares some mechanistic and structural similar- ity with prokaryotic diflavin reductases such as cytochrome P450 BM3 [21] and sulfite reductase [22]. The mechanism of electron transfer in mammalian CPRs has been studied in detail by stopped-flow kinetic and potentiometric methods [23–26]. The reduction potentials of the flavin couples have been determined, and key intermediates in the reaction sequence have been identified. Moreover, the roles of specific residues in flavin reduction by NADPH have been elucidated by mutagenesis methods [25,27,28]. The rate of interflavin electron transfer in mammalian CPRs has been determined by flash photolysis [29] and relaxation kinetic methods [28]. We recently demonstrated [28] that this rate is relatively slow (55 s )1 ) in human CPR, despite the very close proximity (3.85 A ˚ ) of the flavin cofactors observed in the crystal structure of the rat enzyme [7], and is limited by conformational change and regulated by coenzyme binding [28]. This rate is lower (10 s )1 ) for dithionite-reduced enzyme, indicating that nicotinamide coenzyme optimizes interflavin electron transfer. We also showed that mutation of W676, which is located over the re-face of the FAD in CPR [7,30], has a major effect on the kinetics of interflavin electron transfer in both the ÔforwardÕ and ÔreverseÕ direc- tions. We have now extended our relaxation kinetic experiments to include studies of internal electron transfer in the presence of NADPH, NADH and fragments of nicotinamide coenzymes. We show that binding of the 2¢-phosphate of NADPH, and to a lesser extent other regions of the reducing coenzyme, specifically enhance the rate of internal electron transfer. Our studies point to a complex mode of regulation of electron transfer reactions in human CPR, triggered by coenzyme binding and involving conformational change over a relatively large distance, and highlight further the key role of conformational gating in biological electron transfer reactions. Experimental procedures Materials NADPH, NADH, 2¢,5¢-ADP, 2¢-AMP, nicotinamide 1,N 6 -ethenoadenine dinucleotide phosphate and sodium dithionite were from Sigma. 2¢(3¢)-O-(trinitrophenyl)adeno- sine 5¢-monophosphate was from Molecular Probes. All other chemicals were of analytical grade. Protein purification Human fibroblast CPR (lacking the N-terminal membrane- anchoring region) and the functional FAD-binding domain were expressed in Escherichia coli BL21(DE3)pLysS from appropriate pET15b plasmid constructs, and purified as described previously [24]. The FAD-binding domain con- struct includes the so-called linker domain [7]. Temperature-jump and stopped-flow kinetic methods Temperature-jump experiments, using a TJ-64 temperature- jump instrument (Hi-Tech Scientific), were performed under anaerobic conditions according to the method recently described [28], and the heating time was estimated to be 4 ls using a standard phenolphthalein-glycine buffer test (T-jump users’ manual, Hi-Tech Scientific, Salisbury, UK). The initial temperature was 20 °C. CPR samples (140 l M ) were prepared in an anaerobic glove box (Belle Technology Ltd). Reduction of the enzyme to the two- electron level by different electron donors (dithionite, NADH or NADPH) was monitored optically as described previously [28]. Typically, 20 transients were collected and averaged for each reaction condition. Optical artefacts caused by the high voltage discharge through the cell were accounted for by control measurements with the oxidized form of the enzyme, as described previously [28]. Relaxation transients were fitted to a monophasic process using Hi-Tech software dedicated to the T-jump instrument. Stopped-flow experiments with fluorescence detection were performed using an Applied Photophysics SX.18 M V stopped-flow instrument. Tryptophan fluorescence studies were performed using an excitation wavelength of 295 nm. The emission band was selected using a WG320 cut-off emission filter (Coherent Optics) in combination with a UG-11 filter (Coherent Optics) to block stray visible light. The photomultiplier voltage was kept constant during the measurements. The oxidized form of the CPR enzyme (10 l M ) was used in these experiments, which were performed in 50 m M potassium phosphate buffer (pH 7.0) at 25 °C. Inhibition studies The inhibition constant for 2¢,5¢-ADP was determined by using steady-state kinetic assays with cytochrome c as electron acceptor and NADPH as electron donor. Reaction mixtures contained 7 n M CPR, 50 m M potassium phos- phate buffer (pH 7.0), 50 l M cytochrome c and variable concentrations of NADPH and of inhibitor 2¢,5¢-ADP; reactions were initiated by making microlitre additions of NADPH to the reaction mix. Assays were performed at 25 °C and the initial velocity of the reaction was measured by reduction of cytochrome c 3+ at 550 nm (De ¼ 21.1 m M )1 Æcm )1 ) using a Cary-300 UV/visible spec- trophotometer. Data were first analysed graphically using double-reciprocal plots (1/v i vs. 1/[NADPH]) to determine the type of inhibition. Initial velocity values were then fitted to the equation describing competitive inhibition v i ¼ V max S K M 1 þ I K i  þ S hi ð1Þ by nonlinear regression analysis using the GRAFIT software package [31]. Potentiometric titrations Spectroelectrochemical titrations were performed within a Belle technology glove box under a nitrogen atmosphere (oxygen maintained at < 5 p.p.m.) in 100 m M potassium phosphate buffer (pH 7.0) containing 10% (v/v) glycerol (titration buffer), at 25 ± 2 °C, essentially as described previously [26]. Anaerobic titration buffer was prepared by flushing freshly prepared buffer with oxygen-free argon. CPR protein samples (typically % 50–80 l M )admittedto the glove box were de-oxygenated by passing through a Bio- Rad EconoPack 10 DG desalting column pre-equilibrated Ó FEBS 2003 Electron transfer in human cytochrome P450 reductase (Eur. J. Biochem. 270) 2613 in the anaerobic titration buffer. The final CPR concentra- tion used for the redox titrations was 60 l M . Solutions of benzyl viologen, methyl viologen, 2-hydroxy-1,4-naphtho- quinone and phenazine methosulfate were added to a final concentration of 0.5 l M as redox mediators for the titrations. Absorption spectra (300–800 nm) were recorded on a Cary UV50 Bio UV-visible spectrophotometer external to the glove box, with absorption signals relayed to the instrument from an absorption probe (Varian Inc.) immersed in the protein sample, via a fibre optic cable. The electrochemical potential was monitored using a Hanna instruments pH/voltmeter coupled to a Russell Pt/calomel electrode immersed in the CPR solution. The electrode was calibrated using the Fe(II)/Fe(III)-EDTA couple (+108 mV) as a standard. The enzyme solution was titrated electrochemically using sodium dithionite as reductant and potassium ferricyanide as oxidant, as described by Dutton [32]. Duplicate titration data sets were collected for CPR in the presence and in the absence of 2¢,5¢-ADP (60 l M ). Approximately 100 different spectra were recorded during each redox titration, covering the range between around )400 mV and +200 mV (vs. normal hydrogen electrode). Reductive and oxidative titrations of the enzyme indicated that there were no hysteretic effects, and that the enzyme did not aggregate to any significant extent over the course of the 5–8 h required to complete the titrations. Data analysis was performed using Origin (Microcal), essentially as described previously [26]. Throughout the titrations, the enzyme remained soluble and corrections for turbidity were not required. The absorption vs. spectral data were fitted by using Eqn (2), which describes a two-electron redox process derived by extension to the Nernst equation and the Beer–Lambert Law, or Eqn (3), which represents the sum of two two-electron redox processes. Fitting procedures are described in detail in our previous studies [26,33]. 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 ð2Þ 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 þ d10 ðEÀE 0 3 Þ=59 þ e þ f10 ðE 0 4 ÀEÞ=59 1 þ 10 ðEÀE 0 3 Þ=59 þ 10 ðE 0 4 ÀEÞ=59 ð3Þ In these equations, A is the total absorbance; a, b and c are component absorbance values contributed by one flavin in the oxidized, semiquinone and reduced states, respectively, and d, e and f are the corresponding absorbance compo- nents associated with the second flavin. E is the observed potential in mV; E 1 ¢ and E 2 ¢ are the midpoint potentials for oxidized/semiquinone and semiquinone/reduced couples, respectively, for the first flavin, and E 3 ¢ and E 4 ¢ are the corresponding midpoint potentials for the second flavin. The complexity of the system (4-electron titration with probable overlap of midpoint potentials) necessitated the use of a two-stage fitting process. Our previous studies using intact CPR and the FMN and FAD domain of human CPR identified isosbestic points for the oxidized/semi- quinone and semiquinone/reduced transitions for both the FAD and FMN flavins at % 500 nm and 430 nm, respect- ively. For intact CPR, data fitting at these wavelengths yielded, within the uncertainty of the measurements, the same midpoint potentials for the redox couple contributing to the absorbance change whether Eqn (2) or Eqn (3) was used. In using Eqn (2), pairs of variables were set as constant and equal prior to fitting. Thus, to determine the oxidized/semiquinone couples, b ¼ c and e ¼ f (the absorp- tion values for the semiquinone and reduced forms of the first and second flavins, respectively, at the isosbestic point), and to determine the semiquinone/reduced couples, a ¼ b and d ¼ e. Midpoint potential values obtained from the fits at isosbestic points were then used as starting points to enable accurate fitting of absorption vs. reduction potential data for intact CPR at wavelengths near-maximal for the oxidized flavins (455 nm) and for the neutral blue semi- quinone species observed to accumulate during the redox titration (585 nm). The oxidized flavins were assumed to have equal absorbance coefficients, specifically a ¼ d.This is known to be the case from previous studies with the individual domains [26]. Reduction potentials for the system were determined using A 455 and A 585 data and Eqn (2) through an iterative process in which midpoint potentials associated with one isosbestic point were held fixed (e.g. FMN and FAD oxidized/semiquinone) while the other pair were varied. The process was then repeated with the resultant potentials determined for the second pair fixed and the first pair allowed to vary in the second round of analysis. This process was repeated iteratively until there was no further change observed. Data derived in this way are similar to those originally estimated from the fits at the isosbestic points [26]. Results Chemical relaxation reactions with NADH-reduced CPR The optical spectrum of CPR reduced by stoichiometric NADH is identical to that obtained with NADPH (Fig. 1A), indicating that binding of the coenzyme 2¢-phosphate does not affect the redox potentials of the flavins. Furthermore, the optical spectra obtained with nicotinamide coenzymes are very similar to the spectrum of CPR following reduction with stoichiometric sodium dithio- nite [28], indicating that the potential of the flavin couples is the sole determinant of the redox equilibrium within the enzyme under these conditions, and hence of the nature of the equilibrium perturbation following rapid elevation of temperature in a temperature-jump experiment. A detailed discussion of the electron distribution prior and subsequent to capacitor discharge in the temperature-jump instrument for NADPH- and dithionite-reduced enzyme can be found in our previous publication [28]. Briefly, two different two- electron reduced species of CPR are thought to predominate prior to temperature elevation. One species contains the high-potential blue semiquinone form of the FMN (FMN ox/sq ¼ )66 mV) and the semiquinone form of FAD (FAD ox/sq ¼ )283 mV). The second species contains oxidized FAD and the hydroquinone form of FMN (FMN sq/hq ¼ )269 mV). These two species are present in approximately equal concentration (Scheme 1), consistent with our measured values for the relevant couples of the FAD and FMN [26]. We have shown that perturbation of this equilibrium by a temperature jump leads to further and 2614 A. Gutierrez et al. (Eur. J. Biochem. 270) Ó FEBS 2003 transient oxidation of the FAD semiquinone form as it transfers an electron to the FMN sq to produce more of the FAD/FMNH 2 species. This is observed as a net loss of absorbance at 600 nm (i.e. loss of blue semiquinone signature). Transient oxidation of the FAD semiquinone shifts the equilibrium towards the FAD/FMNH 2 species (Scheme 1). The relaxation process therefore involves electron transfer between the two flavin cofactors (i.e. FAD sq –FMN sq (q)FAD ox –FMN hq ). With NADPH- reduced CPR, electron transfer occurs with an observed rate, 1/s,of55±2s )1 [28]. Although the initial equilibrium distribution is identical, the observed rate of internal electron transfer in CPR reduced with NADH (1/s ¼ 18 ± 0.7 s )1) is a factor of three less than the corresponding value for NADPH- reduced CPR (1/s ¼ 55 ± 0.5 s )1 ;Fig.1B),andmore similar to that seen for dithionite-reduced CPR (11 ± 0.5 s )1 ). The midpoint redox potential values for the NADPH/NADP + and NADH/NAD + couples are the same ()320mV,pH7.0,30°C), so that the altered rate of interflavinelectrontransferisnotrelatedtoachangein driving force. The presence of a phosphate group esterified attheribose2¢ position in the redox-inactive adenosine moiety of NADPH is the only chemical difference between the two coenzymes. The observation of a threefold reduc- tion in rate of interflavin electron transfer in NADH- reduced CPR thus suggests a role for coenzyme binding, and in particular for interactions made by the 2¢-phosphate of NADPH, in modulating interdomain electron transfer in CPR. Temperature-jump experiments performed with frag- ments of NADPH were performed to test this hypothesis and are described in the following sections. Binding of adenosine 2¢,5¢-bisphosphate to CPR and effects on interflavin electron transfer In the crystal structure of rat CPR [7], NADP + appears to be bound to the enzyme predominantly through inter- actions involving its adenine end, whose binding site is contained within the FAD-binding domain of the protein (Fig. 2). The electron density for the NMN portion of the coenzyme is poorly defined, and the position of the nicotinamide ring is conjectured to be different in each of the two independent molecules present in the crystal [7]. To investigate any effect of binding 2¢,5¢-ADP on the flavin environment, difference absorption spectra on binding 2¢,5¢- ADP were measured with both the isolated FAD-binding domain and the intact soluble CPR enzyme (see Note). Binding of a stoichiometric amount of 2¢,5¢-ADP to the isolated FAD-binding domain, in either the oxidized form or the form reduced to the one-electron level with dithionite, leads to clear changes in the absorption spectrum of the FAD, consistent with a perturbation of the cofactor environment (Fig. 3). Similar experiments performed with intact CPR (oxidized and two-electron reduced) show that the optical changes induced on binding 2¢,5¢-ADP (Fig. 3C,D) are different from those observed in the isolated FAD domain in both oxidized and reduced states. This is especially evident for the reduced forms (Fig. 3B,D) in the region 500–700 nm. Thus, in intact CPR the binding of 2¢,5¢-ADP to the FAD domain appears to modify not only the environment of the FAD but also that of the FMN Fig. 1. Relaxation transients observed with CPR reduced at the two- electron level with nicotinamide coenzymes. (A and C) Absorption spectra obtained after reduction of human CPR (140 l M )with NADPH (140 l M ). Conditions: 100 m M potassium phosphate buffer, pH 7.0, 25 °C. Identical spectral changes to those shown in (A) were also observed on adding 140 l M NADH to human CPR. (B) Tem- perature-jump difference absorption transients obtained for CPR reduced at the two-electron level with NADPH (upper transient) and NADH (lower transient). Conditions: thermostat temperature, 20 °C; temperature-jump, +7 °C; voltage discharge, 12.5 kV; 140 l M CPR, 100 m M potassium phosphate buffer, pH 7.0. For the difference absorption transient measured at 450 nm for CPR reduced with NADPH; 1/s ¼ 55 ± 2 s )1 . For the difference absorption transient measured at 450 nm for CPR reduced with NADH; 1/s ¼ 18 ± 0.7 s )1 . Difference absorption transients were generated by subtracting the transient obtained for the oxidized enzyme from that obtained for reduced CPR as described previously [28]. Ó FEBS 2003 Electron transfer in human cytochrome P450 reductase (Eur. J. Biochem. 270) 2615 cofactor. This structural perturbation transmitted between the two domains may play a role in the kinetics of internal electron transfer (see below). Titration with 2¢,5¢-ADP at protein levels sufficient for good spectral signals gave linear plots, implying that the K d is quite low. Thus, it was necessary to measure the inhibition constant for 2¢,5¢-ADP in steady-state CPR-catalysed reactions as a guide to its strength of binding to CPR. Steady-state kinetic studies with cytochrome c as electron acceptor demonstrated that 2¢,5¢-ADP is a competitive inhibitor with respect to NADPH. The K m,app for NADPH increases as the 2¢,5¢-ADP concentration is increased, without effect on V max,app (Fig. 4). The value of the inhibition constant for 2¢,5¢-ADP (K i ¼ 5.4 ± 1.3 l M )is, within error, identical to the K m,app for NADPH (6.7 ± 1.3 l M ), suggesting that this moiety is the major determinant of coenzyme binding to CPR. The rate of interflavin electron transfer for CPR reduced at the two-electron level by dithionite (1/s ¼ 11 ± 0.5 s )1 ) is increased approximately threefold to 35 ± 0.2 s )1 in the presence of a stoichiometric amount of 2¢,5¢-ADP, and this is accompanied by an increase in the amplitude of the absorbance change. This is consistent with the notion that the 2¢,5¢-ADP moiety of NADPH makes a major contribu- tion to the enhanced rate of internal electron transfer in NADPH-reduced CPR as compared to the dithionite- reduced enzyme. The absence of a comparable rate enhancement in NADH-reduced CPR suggests that the 2¢-phosphate of NADPH is a major factor in increasing the rate of interdomain electron transfer. It should be noted, however, that the binding of 2¢,5¢-ADP to dithionite- reduced enzyme does not lead to as fast an electron transfer rate as that observed in NADPH-reduced samples (35 ± 0.2 s )1 vs. 55 ± 0.5 s )1) . This suggests that other interactions (perhaps with the pyrophosphate bond) can also contribute to the enhancement of the rate of interflavin electron transfer. Similar temperature-jump experiments performed with 2¢-AMP are also consistent with this conclusion. The observed rate of interflavin electron transfer obtained with dithionite-reduced CPR in the presence of either 140 l M or 1800 l M 2¢-AMP (K i2¢-AMP ¼ 180 ± 11 l M ) is identical, within experimental uncertainty, to that obtained with 2¢,5¢-ADP (32 ± 0.6 s )1 and 35 ± 0.2 s )1 , respectively). Spectroelectrochemical titration Any perturbation in the redox potential of the relevant flavin couples on binding 2¢,5¢-ADP might account for the observed increase in the rate of internal electron transfer, and perhaps for the changes in the flavin absorption spectra. We have therefore determined the redox potentials of CPR in the absence and presence of 2¢,5¢-ADP. Electrochemical titration of CPR with dithionite has revealed that each of the four flavin couples in CPR can be distinguished and their reduction potentials determined [26]. Spectral changes accompanying a typical titration are shown in Fig. 5A and plots of absorbance at 455 nm vs. potential in Fig. 5B; in these experiments CPR concentration was 60 l M and 2¢,5¢-ADP was present in equimolar amounts. The oxidized enzyme has visible absorption maxima at 455 nm and 381 nm, with the semiquinone species having a maximum at 585 nm. Redox titrations performed in the presence of 2¢,5¢-ADP exhibited similar spectral changes to those observed in its absence. Plots of absorption vs. potential data for CPR in the presence and absence of 2¢,5¢-ADP were essentially identical (Fig. 5B), indicating that there was negligible difference in redox behaviour of the flavins in the presence of this ligand. While the midpoint reduction potential values for three of the four couples, estimated as described in Experimental procedures, are slightly more positive than those reported previously, all values are within error of the previously reported data set [26]. The redox potentials were not altered to any significant extent by addition of 2¢,5¢-ADP. Thus, the midpoint potentials for both the 2¢,5¢-ADP-bound and Fig. 2. The crystal structure of NADP 9 -bound rat CPR in its oxidized form [7]. (A) The overall polypeptide fold. The different structural domains are indicated. The FAD-domain contains the binding site for NADPH and is related to the FNR family of flavoproteins. For clarity, the NADPH-binding subdomain is also indicated. (B) The coenzyme- binding site. The main residues known to interact with the adenosine 2¢-phosphate group are shown (R298, R597 and K602). W677 (W676 in human CPR), which shields the Re-face of the FAD isoalloxazine ring position, is also shown. The nicotinamide moiety of NADP(H) is highly disordered in the crystal. 2616 A. Gutierrez et al. (Eur. J. Biochem. 270) Ó FEBS 2003 ligand-free forms of CPR are )76 ± 5 mV (FMN ox/sq); )248 ± 10 mV (FMN sq/hq); )260 ± 12 mV (FAD ox/ sq) and )361 ± 7 mV (FAD sq/hq). Our findings contrast with published work on cytochrome P450 BM3 , where NADP(H) binding to the reductase domain is thought to induce significant changes in the reduction potentials of the FAD[34].WeconcludethatinthecaseofCPRthe observed effects of the binding of coenzyme and coenzyme fragments in increasing the rates of electron transfer result largely from structural perturbations that affect the rate- limiting step(s) for electron transfer ) probably inter– domain interactions ) without altering the driving force. Effects of binding adenosine 2¢,5¢-bisphosphate on hydride ion transfer In previous work with CPR and the isolated FAD- domain we provided kinetic evidence for the existence of a second kinetically distinct noncatalytic NADPH-bind- ing site in the enzyme, occupation of which led to a two- to fivefold decrease in the rate of hydride ion transfer [24,25]. Similar observations have subsequently been made with the structurally related enzymes neuronal nitric oxide synthase [35] and the adrenodoxin reductase homologue FprA from Mycobacterium tuberculosis [36]. A general model has emerged from these kinetic obser- vations in which occupation of this second site by NADPH hinders the release of NADP + .Because electron transfer is reversible, hydride transfer leads to an equilibrium distribution of enzyme species comprising oxidized enzyme, oxidized enzyme bound to NADPH, and reduced enzyme bound to NADP + .Releaseof NADP + from the latter species will lead to further flavin reduction as the equilibrium distribution is shifted toward reduced enzyme (see [24,35] for a more detailed discus- sion). Thus, if the binding of NADPH to the noncata- lytic site hinders NADP + release from the catalytic site, the observed rate of FAD reduction will be decreased at Fig. 3. Difference absorption spectra for the binding of 2¢,5¢-ADP to the isolated FAD-domain and CPR. (A) FAD-domain. Difference absorption spectra (spectrum in the presence of 100 l M 2¢,5¢-ADP minus spectrum in the absence of 2¢,5¢-ADP) of oxidized FAD-domain (100 l M ). (B) As for panel A, but for the FAD-domain (100 l M ) reduced at the one-electron level (blue semiquinone species) with sodium dithionite. (C) CPR. Difference absorption spectra (spectrum in the presence of 50 l M 2¢,5¢-ADP minus spectrum in the absence of 2¢,5¢-ADP)ofoxidizedCPR(50l M ). (D) As for (C) but for CPR (50 l M ) reduced at the two-electron level (blue disemiquinone species) with sodium dithionite. Fig. 4. The inhibition of cytochrome c reductase activity by 2¢,5¢-ADP. Reaction mixtures contained 7 n M CPR, 50 m M potassium phosphate buffer, pH 7.0, 50 l M cytochrome c and various concentrations of NADPH and of the inhibitor 2¢,5¢-ADP. The assay temperature was 25 °C and the reaction rate was measured by reduction of cyto- chrome c 3+ . Five different concentrations of 2¢,5¢-ADP were used: 5, 10, 20, 40 and 80 l M corresponding to curves a, b, c, d and e, respectively. Ó FEBS 2003 Electron transfer in human cytochrome P450 reductase (Eur. J. Biochem. 270) 2617 high NADPH concentrations, and this is in fact observed [24,25]. By contrast, in stopped-flow experiments per- formed with NADH a conventional saturable increase in the rate of flavin reduction with increasing NADH concentration is seen (Fig. 6B). This indicates that NADH does not bind to the second coenzyme-binding site and/or that NAD + dissociates too rapidly to limit the rate of hydride transfer. In light of the major importance of the 2¢-phosphate in coenzyme binding, the latter seems a likely explanation. In rapid-mixing stopped-flow experiments, preincubation of the isolated FAD-domain with stoichiometric 2¢,5¢-ADP eliminates this inhibitory effect observed at high NADPH concentration, presumably by preventing the simultaneous binding of two NADPH molecules (Fig. 6A and inset). In temperature-jump experiments with dithionite-reduced CPR in the presence of a 10-fold molar excess of 2¢,5¢- ADP, the observed rate of interflavin electron transfer (1/s ¼ 28 ± 0.5 s )1 ) is only slightly less than that observed with stoichiometric 2¢,5¢-ADP (1/s ¼ 35 ± 0.2 s )1 ). Com- bined, the stopped-flow and temperature-jump data indicate that increasing the concentration of 2¢,5¢-ADP above stoichiometric levels does not have any substantial addi- tional effect on the rate of internal electron transfer. These results suggest that the observed inhibitory effect of high NADPH concentrations on the reductive half-reaction is associated with the nicotinamide moiety of the coenzyme. The role of W676 in interdomain electron transfer Our previous studies have shown that microsecond tem- perature perturbation of CPR reduced at the two-electron level with NADPH yields two different relaxation processes [28]. The initial fast relaxation (1/s ¼ 2200 ± 300 s )1 )is not associated with electron transfer, but is attributed to local conformational changes in the vicinity of the cofactors and induced by NADPH binding [28]. It is not observed in the W676H mutant of CPR [28], suggesting that these changes may involve W676, a residue that stacks against the Re-face of the isoalloxazine ring of FAD in CPR (Fig. 1B) and which is involved in a series of conformational changes associated with the hydride transfer step [25]. This fast relaxation process is not observed in temperature-jump experiments with CPR reduced by NADH, nor in Fig. 5. Redox potentiometry studies. (A) Spectral changes during redox titration of human CPR. Human CPR (60 l M )was titrated electrochemically as described in Experimental procedures. Progressive reduc- tion of the enzyme leads to bleaching of the oxidized flavin spectrum and accumulation of neutral blue flavin semiquinones with absorption maximum at 585 nm. Positions of isosbestic points in the titration are indicated at 501 nm (oxidized/semiquinone couple for both flavins) and at 435 nm (semiquinone/ reduced couple). (B) Absorption vs. potential data for 2¢,5¢-ADP-bound and ligand-free forms of human CPR. Plots of absorption data at 455 nm (at the flavin maximum for the oxidized CPR) vs. applied potential are shown for both ligand-free (h)and2¢,5¢-ADP-bound (d) forms of human CPR. Enzyme concen- trations in all titrations were 60 l M . 2618 A. Gutierrez et al. (Eur. J. Biochem. 270) Ó FEBS 2003 experiments with dithionite-reduced CPR incubated with a stoichiometric amount of 2¢,5¢-ADP. That only NADPH (and not NADH, or 2¢,5¢-ADP) can elicit this rapid structural relaxation indicates additional complexity in the interaction of coenzymes with CPR. This suggests that the nicotinamide ring of NAD + interacts differently with the enzyme than that of NADPH or that NAD + is bound weakly to reduced CPR. It is possible that the conformational changes produced by occupation of the 2¢-phosphate binding site that increase the rate of electron transfer also affect the nicotinamide ring binding site and W676. Stopped-flow tryptophan fluores- cence experiments indicated a rapid but small increase (complete in about 2 ms after the mixing event; instrument dead-time 1 ms) in fluorescence emission on mixing oxidized CPR (10 l M ) with a stoichiometric amount of 2¢,5¢-ADP (Fig. 7A). However a similar transient is observed in control experiments involving rapid mixing of Fig. 6. Stopped-flow kinetic studies of flavin reduction in the presence of 2¢,5¢-ADP. (A) Fluorescence stopped-flow transients (excitation at 340 nm; emission at 450 nm) for NADPH oxidation by the FAD- domain. Conditions: 50 m M potassium phosphate buffer, pH 7.0, 25 °C, 10 l M FAD-domain, 200 l M NADPH. Transient a: mono- phasic fluorescence transient (k obs ,of3.2s )1 ) obtained for NADPH oxidation by the FAD-domain. Transient b: fluorescence transient obtained for NADPH oxidation by the FAD-domain preincubated with a stoichiometric amount of 2¢,5¢-ADP (10 l M ). Data are best described by a double-exponential expression, yielding values k obs1 and k obs2 of 26 s )1 and 2.5 s )1 , respectively. Inset: Dependence of the observed rates on NADPH concentration. s, n, Estimated values for k obs1 and k obs2 , respectively, for FAD-domain preincubated with stoichiometric 2¢,5¢-ADP. d, Estimated values for k obs in the absence of 2¢,5¢-ADP preincubation. Note the overlapping of n (k obs2 ) and d (k obs ). (B) NADH concentration dependence of the rate of flavin reduction in intact CPR. Absorption stopped-flow experiments. Reaction monitored at 450 nm. Conditions: 50 m M potassium phos- phatebuffer,pH7.0,25°C, 10 l M CPR. Fig. 7. Stopped-flow tryptophan fluorescence studies with oxidized CPR and 2¢,5¢-ADP. Conditions: 10 l M CPR, 10 l M 2¢,5¢-ADP, 50 m M potassium phosphate buffer pH 7.0, 25 °C. Excitation wavelength 295 nm. (A) Transient observed upon rapid-mixing of oxidized CPR and 2¢,5¢-ADP. (B) Control experiment. Fluorescence change observed when CPR was mixed against buffer alone. The transients observed in (A) and (B) are most likely to originate from fast pressure-pulse pro- pagation after the initial mixing event (apparatus dead time %1ms). (C) Transient a: same as in (B). Transient b: fluorescence kinetic transient observed during reduction of CPR with 10-fold excess NADPH (100 l M ). Mutagenesis studies demonstrated that these fluorescence fluctuations are related to changes in the environment of W676 (see Fig. 2), accompanying hydride transfer from the nicotind- amide nucleotide to the FAD cofactor [25]. Ó FEBS 2003 Electron transfer in human cytochrome P450 reductase (Eur. J. Biochem. 270) 2619 CPR against buffer alone (Fig. 7B), ruling out any specific effect on the conformation of W676 induced by the binding of 2¢,5¢-ADP. Thus, although coenzyme binding is driven primarily by interaction of the 2¢,5¢-ADP moiety with CPR, the larger tryptophan fluorescence changes on mixing NADPH with CPR in stopped-flow experiments ([24,25]; Fig. 7C), attributed to environmental effects on W676, must derive from additional interactions of the coenzyme. We have previously rationalized the lack of a tryptophan fluorescence change observed in temperature-jump experi- ments with NADPH by suggesting that the equilibrium position of NADP + in the two-electron reduced enzyme is that observed in the reported crystal structure (i.e. the nicotinamide ring is not tightly associated with the enzyme and is distant from the isoalloxazine ring) [28]. Discussion Our previous temperature-jump studies with human CPR highlighted the importance of conformational change in limiting the rate of internal electron transfer and pointed to the role of the coenzyme in enhancing reaction rate [28]. We have now extended our temperature-jump studies to identify key interactions that are responsible for modulating the rate of internal electron transfer between the two domains of this enzyme. We have demonstrated that the role of the 2¢-phosphate of NADPH is to optimize electron transfer between the flavin cofactors. Occupation of the 2¢-phos- phate binding site by NADPH, 2¢,5¢-ADP or 2¢-AMP leads to altered rates of conformational change. These conform- ational changes are transmitted over a relatively large distance through the protein to optimize electron transfer betweentheflavins.Inthecaseof2¢,5¢-ADP binding, optical spectroscopy provides further evidence for a Ôlong-rangeÕ perturbation of the environments of the isoalloxazine rings of the flavins. This conformational change appears to involve significant domain movement, as it is sensitive to the effects of solution viscosity [28]. Reduction of the enzyme with NADH leads to a slower rate of internal electron transfer owing to the absence of the phosphate group in the 2¢-phosphate-binding site. However, electron transfer in the NADH reduced enzyme is faster than in dithionite-reduced CPR, indicating that other interactions made by the coenzyme play some role in optimizing electron transfer between the flavins. This is further suggested by the fact that the observed rate of internal electron transfer in the presence of 2¢,5¢-ADP and 2¢-AMP for dithionite-reduced enzyme is increased to % 60% the value seen for NADPH-reduced enzyme. In agreement with our kinetic work on human CPR, crystallographic studies of mutant forms of rat CPR also indicate a role for conformational change and domain re-orientation [30]. The multidomain structure of CPR is highly flexible in solution. The crystal structure strongly suggests that the so-called linker domain plays a key role in controlling the mutual orientation of the two flavin- containing domains. Our kinetic [24] and NMR studies (B. Hawkins, I. Barsukov, L Y. Lian, G. C. K. Roberts, unpublished data) indicate that the two isolated flavin- binding domains do not form a strong complex in solution, and electron transfer between the isolated domains is much slower than in the intact enzyme, with a second-order rate constant of 9.5 · 10 4 M )1 Æs )1 [24]. By tethering the domains with the linker region the electron transfer rate is optimized, but more importantly new opportunities are presented for regulating the rate of interflavin electron transfer by coenzyme binding to the NADPH/FAD domain. Multiple conformations of CPR in the absence of coenzyme have been suggested from earlier steady-state studies with the rat enzyme [37]. It was rationalized that binding of NADPH would then induce conformational change and that a unique conformation would be populated compatible with hydride transfer. Our use of equilibrium- perturbation methods with CPR has now provided evidence that coenzyme binding energy is utilized to optimize reaction steps other than the initial hydride transfer event. Given the conserved multidomain structure of other mem- bers of the diflavin reductase family, it seems reasonable to propose that conformational gating of electron transfer may also occur in these enzymes. The temperature-jump method may be useful for probing these aspects in related enzymes is currently in hand. Acknowledgements We thank C. Baxter and Dr Andrew Westlake for assistance in preparing Fig. 2. The work was funded by the Medical Research Council, the Wellcome Trust and the Lister Institute of Preventive Medicine. N.S.S. is a Lister Institute Research Professor. References 1. Strobel, H., Hodgson, A. & Shen, S. (1995) NADPH-Cytochrome P450 Reductase and its Structural and Functional Domains. In Cytochrome P450: Structure, Mechanism and Biochemistry (Ortiz de Montellano, P., ed), pp. 225–244. Plenum Press, New York, USA. 2. Vermillon, J.L. & Coon, M.J. (1978) Purified liver microsomal NADPH-cytochrome P450 reductase. J. Biol. Chem. 253, 2694–2704. 3. Lu, A., Junk, K. & Coon, M. (1969) Resolution of the cytochrome P-450-containing omega-hydroxylation system of liver micro- somes into three components. J. Biol. Chem. 244, 3714–3721. 4. Porter, T.D. (1991) An unusual yet strongly conserved flavo- protein reductase in bacteria and mammals. Trends Biochem. Sci. 16, 154–158. 5. Porter, T.D. & Kasper, C.B. (1986) NADPH-cytochrome P-450 oxidoreductase: flavin mononucleotide and flavin adenine dinu- cleotide domains evolved from different flavoproteins. Biochem- istry 25, 682–687. 6. Iyanagi, T. & Mason, H.S. (1973) Some properties of hepatic reduced nicotinamide adenine dinucleotide phosphate-cyto- chrome c reductase. Biochemistry 12, 2297–2307. 7. Wang, M., Roberts, D.L., Paschke, R., Shea, T.M., Masters, B.S.S. & Kim, J.J. (1997) Three-dimensional structure of NADPH-cytochrome P450 reductase: prototype for FMN- and FAD-containing enzymes. Proc. Natl Acad. Sci. USA 94, 8411–8416. 8. Enoch, H.G. & Strittmatter, P. (1979) Cytochrome b5 reduction by NADPH-cytochrome P-450 reductase. J. Biol. Chem. 254, 8976–8981. 9. Schacter, B.A., Nelson, E.B., Marver, H.S. & Masters, B.S.S. (1972) Immunochemical evidence for an association of heme oxygenase with the microsomal electron transport system. J. Biol. Chem. 247, 3601–3607. 10. Ilan, Z., Ilan, R. & Cinti, D.L. (1981) Evidence for a new physiological role of hepatic NADPH: ferricytochrome (P-450) 2620 A. Gutierrez et al. (Eur. J. Biochem. 270) Ó FEBS 2003 oxidoreductase. Direct electron input to the microsomal fatty acid chain elongation system. J. Biol. Chem. 256, 10066–10072. 11. Yasukochi, Y. & Masters, B.S. (1976) Some properties of a detergent-solubilized NADPH-cytochrome c (cytochrome P-450) reductase purified by biospecific affinity chromatography. J. Biol. Chem. 251, 5337–5344. 12. Kurzban, G.P. & Strobel, H.W. (1986) Preparation and char- acterization of FAD-dependent NADPH-cytochrome P-450 reductase. J. Biol. Chem. 261, 7824–7830. 13. Keyes, S.R., Fracasso, P.M., Heimbrook, D.C., Rockwell, S., Sligar, S.G. & Sartorelli, A.C. (1984) Role of NADPH: cyto- chrome c reductase and DT-diaphorase in the biotransformation of mitomycin C1. Cancer Res. 44, 5638–5643. 14. Bligh, H.F., Bartoszek, A., Robson, C.N., Hickson, I.D., Kasper, C.B., Beggs, J.D. & Wolf, C.R. (1990) Activation of mitomycin C by NADPH: cytochrome P-450 reductase. Cancer Res. 50, 7789–7792. 15. Bartoszek, A. & Wolf, C.R. (1992) Enhancement of doxorubicin toxicity following activation by NADPH cytochrome P450 reductase. Biochem. Pharmacol. 43, 1449–1457. 16. Walton, M.I., Wolf, C.R. & Workman, P. (1992) The role of cytochrome P450 and cytochrome P450 reductase in the reductive bioactivation of the novel benzotriazine di-N-oxide hypoxic cytotoxin 3-amino-1,2,4-benzotriazine-1,4-dioxide (SR 4233, WIN 59075) by mouse liver. Biochem. Pharmacol. 44, 251–259. 17. Patterson, A.V., Barham, H.M., Chinje, E.C., Adams, G.E., Harris, A.L. & Stratford, I.J. (1995) Importance of P450 reductase activity in determining sensitivity of breast tumour cells to the bioreductive drug, tirapazamine (SR 4233). Br.J.Cancer72, 1144–1150. 18. Griffith, O.W. & Stuehr, D.J. (1995) Nitric oxide synthases: properties and catalytic mechanism. Annu. Rev. Physiol. 57, 707–736. 19. Leclerc, D., Wilson, A., Dumas, R., Gafuik, C., Song, D., Wat- kins, D., Heng, H.H., Rommens, J.M., Scherer, S.W., Rosenblatt, D.S. & Gravel, R.A. (1998) Cloning and mapping of a cDNA for methionine synthase reductase, a flavoprotein defective in patients with homocystinuria. Proc. Natl Acad. Sci. USA 95, 3059–3064. 20. Paine, M.J., Garner, A.P., Powell, D., Sibbald, J., Sales, M., Pratt, N., Smith, T., Tew, D.G. & Wolf, C.R. (2000) Cloning and characterization of a novel human dual flavin reductase. J. Biol. Chem. 275, 1471–1478. 21. Narhi, L.O. & Fulco, A.J. (1986) Characterization of a catalyti- cally self-sufficient 119,000-dalton cytochrome P-450 mono- oxygenase induced by barbiturates in Bacillus megaterium. J. Biol. Chem. 261, 7160–7169. 22. Gruez, A., Pignol, D., Zeghouf, M., Coves, J., Fontecave, M., Ferrer, J.L. & Fontecilla-Camps, J.C. (2000) Four crystal struc- tures of the 60 kDa flavoprotein monomer of the sulfite reductase indicate a disordered flavodoxin-like module. J. Mol. Biol. 299, 199–212. 23. Oprian, D.D. & Coon, M.J. (1982) Oxidation-reduction states of FMN and FAD in NADPH-cytochrome P-450 reductase during reduction by NADPH. J. Biol. Chem. 257, 8935–8944. 24. Gutierrez, A., Lian, L.Y., Wolf, C.R., Scrutton, N.S. & Roberts, G.C.K. (2001) Stopped-flow kinetic studies of flavin reduction in human cytochrome P450 reductase and its component domains. Biochemistry 40, 1964–1975. 25. Gutierrez, A., Doehr, O., Paine, M., Wolf, C.R., Scrutton, N.S. & Roberts, G.C.K. (2000) Trp-676 facilitates nicotinamide coenzyme exchange in the reductive half-reaction of human cytochrome P450 reductase: properties of the soluble W676H and W676A mutant reductases. Biochemistry 39, 15990–15999. 26. 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. 27. Dohr, O., Paine, M.J., Friedberg, T., Roberts, G.C.K. & Wolf, C.R. (2001) Engineering of a functional human NADH-depen- dent cytochrome P450 system. Proc. Natl Acad. Sci. USA 98, 81–86. 28. Gutierrez, A., Paine, M., Wolf, C.R., Scrutton, N.S. & Roberts, G.C.K. (2002) Relaxation kinetics of cytochrome P450 reductase: internal electron transfer is limited by conformational change and regulated by coenzyme binding. Biochemistry 41, 4626–4637. 29. Bhattacharyya, A.K., Lipka, J.J., Waskell, L. & Tollin, G. (1991) Laser flash photolysis studies of the reduction kinetics of NADPH: cytochrome P-450 reductase. Biochemistry 30, 759–765. 30. Hubbard,P.A.,Shen,A.L.,Paschke,R.,Kasper,C.B.&Kim,J.J. (2001) NADPH-cytochrome P450 oxidoreductase. Structural basis for hydride and electron transfer. J. Biol. Chem. 276, 29163–29170. 31. Leatherbarrow, R.J. (1992) Erithacus Software Ltd, Staines, UK. 32. Dutton, P.L. (1978) Redox potentiometry: determination of midpoint potentials of oxidation-reduction components of biological electron-transfer systems. Methods Enzymol. 54, 411–435. 33. Daff, S.N., Chapman, S.K., Turner, K.L., Holt, R.A., Govinda- raj, S., Poulos, T.L. & Munro, A.W. (1997) Redox control of the catalytic cycle of flavocytochrome P-450 BM3. Biochemistry 36, 13816–13823. 34. Murataliev, M.B. & Feyereisen, R. (2000) Functional interactions in cytochrome P450BM3. Evidence that NADP (H) binding controls redox potentials of the flavin cofactors. Biochemistry 39, 12699–12707. 35. 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. 36. McLean, K., Scrutton, N.S. & Munro, A.W. (2003) Kinetic and spectroscopic characterisation of the Mycobacterium tuberculosis adrenodoxin reductase homologue FprA. Biochem. J. 372, 317–327. 37. Sem, D.S. & Kasper, C.B. (1993) Enzyme–substrate binding interactions of NADPH-cytochrome P-450 oxidoreductase characterized with pH and alternate substrate/inhibitor studies. Biochemistry 32, 11539–11547. Ó FEBS 2003 Electron transfer in human cytochrome P450 reductase (Eur. J. Biochem. 270) 2621 . Interflavin electron transfer in human cytochrome P450 reductase is enhanced by coenzyme binding Relaxation kinetic studies with coenzyme analogues Aldo. UK The role of coenzyme binding in regulating interflavin electron transfer in human cytochrome P450 reductase (CPR) has been studied using temperature-jump