Eur. J. Biochem. 204, 1075-1082 (1992) 0FEBS 1992

Effect of KCl on the interactions between NADPH :cytochrome P-450 reductase and either cytochrome c, cytochrome b5 or cytochrome P-450 in octyl glucoside micelles Yukio NISIMOTO' and Dale E. EDMONDSON'

' Department of Biochemistry, Aichi Medical University, Nagakute, Japan

Department of Biochemistry, Emory University School of Medicine, Atlanta, USA

(Received October 21, 1991) - EJB 91 1413

Significant dissociation of FMN from NADPH :cytochrome P-450 reductase resulted in loss of the activity for reduction of cytochrome bS as well as cytochrome c and cytochrome P-450. However, the ability to reduce these electron acceptors was greatly restored upon incubation of FMN-depleted enzyme with added FMN. The reductions of cytochrome c and detergent-solubilized cytochrome bs by NADPH: cytochrome P-450 reductase were greatly increased in the presence of high concentrations of KCI, although the stirnulatory effect of the salt on cytochrome P-450 reduction was less significant. No apparent effect of superoxide dismutase could be seen on the rate or extent of cytochrome reduction in solutions containing high-salt concentrations. Complex formation of the flavoprotein with cytochrome c, which is known to be involved in the mechanism of non-physiological electron transfer, caused a perturbation in the absorption spectrum in the Soret-band region of cytochrome c, and its magnitude was enhanced by addition of KC1. Similarly, an appreciable increase in ellipticity in the Soret band of cytochrome c was observed upon binding with the flavoprotein. However, only small changes were found in absorption and circular dichroism spectra for the complex of NADPH :cytochrome P-450 reductdse with either cytochrome b5 or cytochrome P-450. It is suggested that the high-salt concentration allows closer contact between the heme and flavin prosthetic groups through hydrophobic-hydrophobic interactions rather than electrostatic-charge pairing between the flavoprotein and the cytochrome which causes a faster rate of electron transfer. Neither alterations in the chemical shift nor in the line width of the bound FMN and FAD phosphate resonances were observed upon complex formation of NADPH :cytochrome P-450 reductase with the cytochrome.

cross-linking studies [S, 91 have shown that the ionic-strength dependence of the kinetic rate constants for the electrontransfer reaction between P-450 reductase and cytochrome h5 (or P-450) is dominated by charged amino acid side chains which are localized predominantly in the vicinity of the FMN prosthetic group. Our earlier studies with this flavoprotein [lo, 111 also indicated that l-ethy1-3-[3-(dimethyhmino)propyl]carbodiimide (EDC) promotes specific formation of amide bonds between complementary charge pairs during the interaction of P-450 reductase and cytochrome h5 (or cytochrome c) in L-a-dimyristoylphosphatidylcholine (DMPC) liposomes. In each of the cross-linking reactions, the increased ionic strength of the medium leads to a substantial loss of the electrostatically stabilized complex between the flavoprotein and the cytochrome in detergent micelles as well as phospholipid vesicles. However, it has been reported that Aichi Medical University, Nagakute, Aichi, Japan 480-1 1 the reduction rates of cytochrome b5 and cytochrome c by PAbbreviations. P-450 reductase, liver microsomal NADPH : 450 reductase were markedly stimulated at high-salt concencytochrome P-450 reductase; P-450, microsomal cytochrome trations and that, under these conditions, the flavoprotein is P-450LM2, induced in rabbit liver by phenobarbital; EDC, l-ethylcapable of directly reducing these cytochromes [12, 131. The 3-[3-(dimethylamino)propyl]carbodiimide; 2'-AMP, adenosine 2'formation of a covalently cross-linked complex between Pphosphate; DMPC, L-a-dimyristoylphosphatidylcholine. 450 and the P-450 reductase shows a very low yield under the Enzymes. NADPH:cytochrome P-450 reductase (EC 1.6.2.4);

The detergent-solubilized hepatic NADPH :cytochrome P-450 reductase (P-450 reductase), a flavoprotein component of a liver microsomal mixed-function oxidase, contains 1 mol each of FAD and FMN/polypeptide chain of molecular mass 77 kDa 11 - 51. It is now understood that reducing equivalents are donated from NADPH to the FAD-binding site 16, 71, sequentially transferred to the FMN-binding site, then the FMN site serves as the electron donor to microsomal cytochrome P-450LM2 (P-450) and cytochrome c. During this electron transfer process, the FAD and FMN coenzymes functionally hold both semiquinone and hydroquinone redox levels. Electron transfer between flavoprotein and cytochrome is fundamental to many biochemical pathways, such as mitochondria] and microsomal electron-transfer chains. Previous Correspondence to Y. Nisimoto, Department of Biochemistry,

superoxide dismutase (EC 1.15.1.1).

1076 conditions previously used [lo], and the effect of KC1 on the rate of P-450 reduction appears to be less significant. CD spectral perturbations have previously been shown to occur upon complex formation between the redox flavoprotein, ferredoxin : NADPH oxidoreductase, and either ferredoxin or rubredoxin [14, 151 and between flavodoxin and cytochrome c' [16]. In this paper, we report the occurrence of significant changes in the absorption and CD spectral properties of the Soret band of cytochrome c upon its binding with P-450 reductase. The spectral perturbations become larger when the ionic strength of the medium is increased by the addition of KCl. "P-NMR spectroscopy is also a powerful approach to examine the environments of phosphorus groups in proteins and has been effectively used to study the flavin coenzyme and protein-bound phosphorus residues in flavodoxin [17- 191, glucose oxidase [20] and bovine milk xanthine oxidase [21]. In addition, Otvos et al. [22] and Bonants et al. [23] have recently investigated the 31P-NMR spectral properties of the bound flavin phosphates of the P450 reductase. Upon either cytochrome c, cytochrome hs or P-450 binding with the reductase, the changes in chemical shift and line broadening due to the FMN and FAD phosphate resonances are measured. Present studies demonstrate that FMN serves as an electron donor flavin to cytochrome b5 as well as to cytochrome c' and P-450. Thus, it is suggested that differences in the structural and geometric features, occurring upon complex formation between the flavoprotein and each cytochrome, are evidently reflected in the rate of electron transfer and spectral alterations in the Soret region of the hemeprotein.

enzyme contained flavin, 15.5 nmol/mg protein and the ratio FMN/FAD was 0.96. The reductase prepared by this method was electrophoretically 95% or more pure, contained a few percent of proteolytically cleaved species and had a molecular mass of about 68 kDa.

MATERIALS AND METHODS

Isolation of detergent-solubilized cytochrome b5 and P-450 from microsomes of rabbit liver Cytochrome h5 was purified from rabbit liver microsomes as described by Ito and Sat0 [24] to a specific heme content of 45 nmol/mg protein. The cytochrome b5 concentration was calculated from the reduced minus oxidized absorption difference at 423 nm and 409 nm, using a difference absorption coefficient of 185 mM-' . cm-' [25]. The isolated protein electrophoresed as only one band with a molecular mass of 16 kDa and deoxycholate in the sample was removed by dialysis prior to reaction with the P-450 reductase in the presence of 35 mM n-octyl glucoside. A homogeneous preparation of cytochrome P-450 (P-450LM2) was prepared by the method of Imai and Sat0 [26] from liver microsomes of phenobarbitaltreated rabbits. The P-450 used in this study was electrophoretically homogeneous and contained 16.8 nmol heme/mg protein. The P-450 concentration was determined from the difference spectrum, reduced CO complex minus nm = 91 mM-' . cm-' [25]. reduced,

Materials NADPH, riboflavin, FMN, FAD, cytochrome c (horse heart), Triton N-101, glycerol, adenosine 2'-phosphate (2'AMP), KBr, deoxycholate and n-octyl glucoside and dithiothreitol were purchased from Sigma, DEAE-Sepharose CL-6B and 2',5'-ADP-Sepharose 4B from Pharmacia LKB and hydroxyapatite (Bio-gel HT hydroxyapatite) from BIORAD. FAD and FMN concentrations were determined spectrophotometrically using 11.3 x lo3 M - ' .cm-' (pH 7.0) at 450 nm and 12.2 x lo3 M-' . cm-' (pH 7.0) at 445 nm, respectively. The reductase concentration was determined using a molar absorption coefficient of 21.4 mM-' . cm-' at 450 nm [5]. Triton N-101 was finally removed from the purified preparation of reductase by hydroxyapatite chromatography equilibrated with 0.025 M potassium phosphate, pH 7.0, containing 10% (by vol.) glycerol. All other reagents were of the best grade commercially available. Purification of detergent-solubilized P-450reductase from rabbit liver

Microsomes were prepared from fresh rabbit liver by the method Of 'yanagi and [l] with Some modifications [4]. P-450 reductase was purified from deoxycholate and Triton-N-101-solubilized microsomes by ion-exchange chromatography of DEAE-Sepharose CL-6B followed by affinity chromatography o n 2',5'-ADP-Sepharose and hydroxyapatite chromatography. The NADPH :&tochrome creduc;tase activity of the preparation was 64 pmol cytochrome c reduced . min-' . mg-' protein when assayed at 25°C in 0.1 mM cytochrome c and 0.05 M potassium phosphate, pH 7.2. The

Preparation of FMN-depleted reductase FMN-depletion from the P-450 reductase was carried out by a slight modification of the method of Vermilion and Coon [6].The following procedure was performed at 3 'C in the dark. The native holoenzyme was diluted to a final concentration of approximately 9.6 pM (0.75 mg protein/ml) with 2M KBr in 0.1 M Tris/acetate, pH 8.2, containing 20% (by vol.) glycerol and 0.1 mM dithiothreitol in a final volume of 200 ml. Care was taken to exclude light from all enzyme solutions containing KBr. This KBr-treated P-450 reductase was dialyzed for 7 d against 2 12 M KBr-containing Tris/acetate, pH 8.2, then the enzyme was dialyzed overnight against 2 10.025 M phosphate, pH 7.2, containing 0.1 mM EDTA and 20% (by vol.) glycerol. The enzyme solution was finally applied to a column (1.2 cm x 8 cm) of hydroxyapatite previously equilibrated with 0.025 M phosphate, pH 7.2, containing 20% (by vol.) glycerol. The column was washed with 100 mlO.025 M phosphate, pH 7.2, and the reductase was eluted with 0.1 M phosphate, pH 7.2. The eluted fractions were pooled and concentrated to about 2 ml. On the basis of the activities for NADPH :cytochrome c reductase and NADPH :ferricyanide reductase, the preparation of FMN-depleted reductase contained 98% or more of the FAD present in the native enzyme but only 3% of the original amount of FMN [ 5 ] .

CD measurements Circular dichroism spectra were measured on an AVIV 60DS instrument interfaced to an AT and T model 6300 computer, Quartz used included a split cell with two cornpartmerits each of 0.45 cm path length. A]] CD spectral experiments were carried out at 10°C in 0.025 M phosphate, pH 7.0, containing 0.1 mM EDTA, 10% (by vol.) glycerol and 35 mM n-octyl glucoside.

31P-NMRspectra The 31P-NMR spectra of P-450 reductase were recorded either at 202.4 MHz on a General Electric GN-500 or at

1077 161.7 MHz on a Jeol GSX-400 spectrometer using 10-mm NMR tubes. Spectra were measured at 15 "C with broad-band proton dccoupling, 50000-230000 data points, a 45" flip angle, and a 1.72 pulse repetition rate. Free-induction decays were subjected to a 20-Hz exponential line broadening for all the spectra and the chemical shifts were determined relative to an external standard, 10% trimethyl phosphate. Samples were dissolved in 0.025 M Tris/acetate, pH 7.7, containing 10% (by vol.) glycerol, 20% D 2 0 and 35 mM n-octyl glucoside and were treated with Chelex to remove any contaminating paramagnetic metal ions. Fluorescence measurements

Fluorophotometric assays of total flavin was performed on the neutralized supernatant following precipitation of the protein and liberation of flavin using 20% trichloroacetic acid. Fluorescence measurements were carried out using a Hitachi fluorescence spectrophotometer, model 650-60, with an excitation wavelength of 450 nm and an emission wavelength of 525 nm. The fluorescence intensity was measured at both pH 7.7 and pH 2.6, thus allowing calculation of both FAD and FMN contents [27]. The emission spectrum of tryptophan residues in the reductase was recorded in a sample compartment thermostated at 15 "C. The excitation wavelength and spectral bandwidth were fixed at 295 nm and 5 nm, respectively.

100-

cS

Li

2 z

ao60-

P

40-

20 01

I

1.0

I

0.8

0.6

0.4

I

02

J

0

FMN/FAD (mol ratio)

Fig. 1. Effect of FMN removal and replacement on HADPH-linked cytochrome c, cytochrome b5 and P-450 reductase activities. KBrtrcated P-450 reductase preparations were pooled and finally dialyzed against 0.05 M Tris/acetate, pH 7.7, containing 0.1YOTriton N-101. 10% (by vol.) glycerol and 0.1 mM EDTA. The molar ratios of FMN/ FAD in the preparations were 0.95,0.36,0.28, 0.22 and 0.03, respectively. The rates of cytochrome c (squares), cytochrome b5 (circles) and P-450 (triangles) reduction in the presence (closed symbols) and absence (opened symbols) of added 2 nmol FMN were measured in 0.05 M phosphate, pH 7.2 at 36°C. All assay mixtures containing 0.02 nmol P-450 reductase, 95 nmol cytochrome c (or 80 nmol cytochrome b5 or 100 nmol P-450), 35 pmol n-octyl glucoside, with or without added FMN were incubated for 20 min at 20 'C. and the reaction was started by the addition of 100 nmol NADPH. The changes in cytochrome c, cytochrome b5and P-450 reductase activities Cross-linking of reductase and cytochrome c (or cytochrome b5) were indicated taking nativc enzyme activities as 100% (64 pmol cytochrome c reduced . min-' . (mg protein)-', 54.4pmol in n-octyl glucoside micelles cytochrome b5 reduced . min-' . (mg protein)-' and 1.85 pmol The cross-linking reaction between the two redox proteins NADPH oxidized . min-' . (mg protein)-', respectively). NADPH was carried out in the presence or absence of KC1 according oxidase activity by P-450, incubated with P-450 reductase, n-octyl to the methods described by Nisimoto and Lambeth [lo]. glucoside and 1 mM benzphetamine, was determined by an absorption change at 340 nm.

Before starting the reaction with 10 mM EDC, either cytochrome c (6.6 nmol) or cytochrome b, (5.1 nmol), P-450 reductase (2 nmol), n-octyl glucoside (7.35 pmol) and NADPH (0.21 pmol) were incubated in 0.025 M Tris/acetate, pH 7.6, containing 10% (by vol.) glycerol at 25°C for 5 min. Reaction mixture (20 pl) was added to 20 pl dissociation buffer (1 YOSDS, 1YO2-mercaptoethanol, 0.4 M sodium phosphate, pH 7.6 and 0.01 YObromophenol blue), mixed by vortexing and heated for 10 min at 100°Cfor subsequent analysis by SDSjPAGE as described by Rudolph and Krueger [28]. A constant amount of protein (60 pg or 20 pg) was loaded onto each gel lane. Measurements of protein and activities

Protein was assayed by the method of Lowry et al. [29] using bovine serum albumin as a standard. Enzymatic activities of P-450 reductase were measured at 25°C in 0.05 M potassium phosphate, pH 7.2. The rate of cytochrome c reduction by NADPH was followed using absorbance changes at 550nm, with the absorption coefficient of 18.5mM-' . cm- ' [30]. Cytochrome 6, reductase activity in the presence of 35 mM n-octyl glucoside was determined from the absorption changes, using an absorption coefficient of 100 mM-' . cm-' at 424 nm 1311. After incubation of reductase with P-450 in the presence of 35 mM n-octyl glucoside and 1 mM benzphetamine at 25°C for 10 min, the reductive reaction was started by addition of 0.1 mM NADPH. The P450 reductase activity was expressed as micromol NADPH oxidized . min-' . (mg reductax-'.

RESULTS Effects of FMN on NADPH-dependent cytochrome c, cytochrome b5 and Pa50 reductase activities As shown in Fig. 1, a remarkable dissociation of FMN from the reductase was observed in the presence of 2M KBr at pH 7.0. However, loss of FAD was minimal under the conditions used, and the ability of the enzyme to interact with pyridine nucleotide was not significantly altered by depletion of FMN. The five preparations of FMN-depleted P-450 reductase used in the present experiments contained more than 95% of the FAD present in the native P-450 reductase, but only 3 - 36% ofthe original amount of FMN. Consistent with the earlier studies [5, 61 using cytochrome c and P-450, the activities for reducing these cytochromes by the FMN-depleted reductase were lost in parallel with decreasing concentrations of FMN in the protein, but the ability to reduce the two electron acceptors was significantly restored after incubation with added FMN. Thus, these results suggested the direct involvement of reduced FMN with either cytochrome c or P-450. Similarly, the FMN-depleted preparation lost the activity for the reduction of cytochrome b5. The relative activity corresponded closely to the amount of FMN remaining in the preparation. When the FMN-depleted reductase was assayed after incubation with FMN, NADPH :cytochrome b5 reductase activity was considerably elevated, although the restoration of the activity was not as high as that of NADPH :cytochrome c reductase. About half of cytochrome-

1078 >

7

30 >

t

- 20

v)

a

z y

B

- f K2 0

-10

3

- 5 Y

I"-!

LT

OLE a

or 0

0.25

0.50

0.75

lo

KCI CONCENTRATION (M)

Fig. 2. Effect of KC1 on NADPH-dependent cytochrome c, cytochrome bS and P-450reductase activities. Each concentration of KCI (0.050.73 M) in 1 ml 0.05 M Tris/acetate, pH 7.6, containing 35 mM noctyl glucoside, was incubated with reductase (0.02 nmol) and either cytochrome c, cytochrome h5 or P-450 at 25°C for 20 min. The activities for cytochrome c reductase (0),cytochrome b, reductase ( 0 )and NADPH oxidase ( A ) were assayed by adding 100 nmol NADPH to the incubation mixture. In parallel with the activity assay, the fluorescenceintensities of the enzyme-bound flavins and tryptophan residues were measured after incubation of the reductase (2.95 nmol protein) at the indicated concentrations of KCI at 25°C at 525 nm, when excited at for 20 min. Flavin fluorescence (0) 450 nm, was measured at pH 7.6. For emission measurements of tryptophan residues of the reductase (m),the excitation and emission wavelengths were 295 nm and 336 nm, respectively, as fixed by a monochromator with a bandwidth of 2 nm.

h5-binding or P-450-binding sites on the reductase are probably considered to be destroyed by the procedure, suggesting that a more stringent structure of the reductase is required to interact with the two hemeproteins than that of cytochrome c. In agreement with earlier reports [5], FMN-depleted reductase could donate electrons to ferricyanide whose reductive rate corresponded closely to that of the holoenzyme. These results basically support the idea that not only the reaction with cytochrome c or P-450, but also the interaction with cytochrome h5 proceeds via a n FMN-dependent pathway and that the FAD-binding domain is essential to accept electrons from NADPH. Activity changes of reductase in the presence of high-salt concentrations

As shown in Fig. 2, NADPH-dependent reductions of cytochrome c and cytochrome b5,catalyzed by the reductase, were markedly increased a t high concentrations of KC1 as earlier observed upon addition of monovalent [12, 13, 321 or divalent cations [33] to the reaction mixture. Phillips and Langdon I321 explained the salt effect in terms of blocking charge pairing between NADP+ and the P-450 reductase, thereby hastening dissociation of the NADP', a rate-limiting step. The effect of KCI on reductase at neutral p H may also be considered as a localized perturbation of the enzyme, focused primarily at or near the active site(s). Thus, this perturbation appears to result in a significant increase in the reactivity of the flavin-binding domain toward cytochrome c and cytochrome h5. Also, in the presence of high concentrations of salt, the overall conformation of the protein backbone was reported to be largely unchanged during 20 min incubation at 25 "C, since fluorescence intensities of flavin and tryptophan

-I

1

I -0.02

1

350

400

450

1

I 500

550

1

600

WAVELENGTH (nm)

Fig. 3. Changes of difference absorption spectrum (mixture minus sum) dependent on the molar ratio of reductase to cytochrome c. Spcctra of P-450 reductase/cytochrome c mixture minus the sum of individual spectra were obtained in 0.025 M Tris/acetate, pH 7.6, containing 10% glycerol (by vol.) and 35 mM n-octyl glucosidc. Spectral changes were measured as a function of the concentration of added P-450 reductase to a constant concentration ofcytochrome c (1 3 pM). Curve 1. base line without reductase. Curves 2-6, difference spectra of cytochrome c reacted with the P-450 reductase whose concentrations were 2.5, 5, 7.5, 10 and 15 pM. Cell path length was 0.40 cm.

residues of the reductase were not much increased and nearly complete retention of the catalytic activity could be observed. However, prolonged incubation for more than 2 h at 25°C in the presence of 0.2 M KC1 caused remarkable increases in the fluorescence intensities of both the bound flavin and tryptophan residues. These alterations appear to be less significant without salt in the incubation mixture. Although it is known that reductase rapidly generates the superoxide anion at highsalt concentrations, the reductions of both cytochrome c and cytochrome b5 are not dependent upon superoxide anion production, but occur via direct interaction between P-450 reductase and the hemeprotein [13]. The activity for P-450 reduction, in contrast, becomes elevated a little upon addition of KC1, but superoxide dismutase shows no inhibitory effect on its activity in medium containing 0.5 M KCl. These results suggest that the reduction of P-450, catalyzed by P-450 reductase, also occurs via a direct electron transfer between the two redox components. Reductase-cytochrorne interactions in octyl glucoside micelles Pronounced spectral changes at 550 nm, 520 nm and in the Soret band region of cytochrome c are seen upon mixing with P-450 reductase, suggesting a perturbation in the environment of the heme c prosthetic group on complex formation with the flavoprotein (Fig. 3). The P-450 reductase at these wavelengths has very low absorption coefficients, allowing a negligibly small effect on the spectral alteration of cytochrome c. The magnitude of the spectral perturbation was measured as a function of the concentration of added P450 reductase while the cytochrome c concentration was held constant at 13 pM. Obviously, the size of the difference absorption signal is proportional to the amount of added P-450 reductase and reaches its maximum at a ratio of about one mol/mol cytochrome c, which supports a 1 : 1 complex formation between the two redox proteins. This stoichiometry is

1079

300 I*

350

WAVELENGTH (mn)

> 1.2 . c

1 3

H8

9 -I Q

’;

o.8: O.’V

0.0 hwfi.-04.

.--;

400

450

500

550

6d0

WAVELENGTH (mn)

c

1..*,. ,

,

j

consistent with the previously obtained results that a 1:l electron transfer; therefore, it is possible that a correlation covalently cross-linked complex is predominantly formed be- exists between the extent of heme c perturbation upon P-450 tween cytochrome c and P-450 reductase in the presence of reductase binding and the kinetics of electron transfer. The results shown in Fig. 5A demonstrate that the EDCEDC [ll]. However, no appreciable spectral alteration due to complex formation was seen by mixing either cytochrome h5 mediated 1: 1 cross-linked complex formation (89 kDa band) or P-450 with the flavoprotein. Thus, it is apparent that there between cytochrome c and P-450 reductase is gradually deis little change in the environments of the hemes of cytochrome creased in accordance with a rising concentration of KCl in the reaction mixture. This may be due to a weakened electrostatic h5 and P-450 upon interaction with P-450 reductase. Similarly, evidence for an alteration in the conformation protein- protein interaction in the medium with high ionic of heme c as a result of binding of cytochrome c with P- strength. Likewise, the cross-linking reaction between 450 reductase is also obtained by measurement of the CD cytochrome b, and P-450 reductase was also inhibited by spectrum. Fig. 4A and B show a comparison between the sum addition of KCl (Fig. 5B) despite the fact that the electronof individual CD spectra of P-450 reductase and cytochrome transfer rate between the two redox components is greatly c and the spectrum of a 1 : l molar ratio mixture of each stimulated by the salt. These data suggest that the differences component. It is clear that an appreciable perturbation has in binding geometries and mechanisms of the two interacting occurred upon mixing the two redox proteins and the amount proteins can be considered due to the presence and absence of cytochrome c complexed with P-450 reductase reached of KCI in the medium. In the absence and presence of added maximum at an excess of added reductase (fivefold molar KC1, the distinctive cross-linking band, consisting of 1 mol excess of cytochrome c). In Fig. 4 C the difference CD spec- each of P-450 and P-450 reductase was not seen. trum is plotted, which is obtained by computer subtraction of the spectra shown in Fig. 4 A and B. It is apparent that the ‘P-NMR spectrum of P-450 reductase-cytochrome complex major spectral effect occurs around the Soret band region of heme c (approximately 418 nm) and thus suggests a conforThe 31P-NMR spectrum of the oxidized form of n-octylmational perturbation of the heme c prosthetic group on com- glucoside-solubilized reductase is shown in Fig. 6 A. The peak plex formation with P-450 reductase. Fig. 4 D indicates that furthest downfield at 0.16 ppm is assigned to bound FMN almost similar changes in the hemeprotein interaction would phosphate, since the confirmation of this assignment is be anticipated upon addition of either holoenzyme or FMN- provided by comparing the 31P-NMR spectrum of the depleted reductase. Thus, FMN depletion suggests little or no holoenzyme with that of FMN-depleted reductase [23]. The effect on the flavoprotein - cytochrome c interaction. Fur- small resonance observed at about 1.55 ppm may be attributhermore, the extent of the spectral perturbation in the Soret table to the FMN which exists in rather loose binding environband region of cytochromc c complexed with reductase is ments in the reductase, and it is noted that this FMN phosintensified by increased KCl concentrations (data not shown). phate signal at 1.55 ppm becomes prominent when native No marked spectral changes for cytochrome c at a high con- reductase (77 kDa) is gradually converted to the proteolyzed centration of KCl (0.25 M) are observed without addition of form (68 kDa) by adventitiously contaminated protease(s). P-450 reductase, suggesting that the increased alterations in The two most upfield resonances at - 11.2 ppm and the heme - protein interaction, caused by high ionic strength, 15.2 ppm are also assigned to the two phosphorus nuclei of appear to be a result of specific binding between the two the phosphate moiety of another prosthetic flavin, FAD. In protein components. High ionic strength of the medium pro- addition to these expected phosphorus resonances based on duces larger spectral changes and shows an enhanced rate of the flavin coenzymes present in the reductase, we observed

1080

I

5

5

~

"

'

I

'

'

~

'

1

~

"

'

l

0

-5 -10 CHEMICAL SHIFT (PPm)

O

CHEMICAL SHIFT(PPm)

-5

-10

'

.

"

~

'

'

-15

-20

-15

-20

~

'

~

Fig. 6. 202.4 MHz 31P-NMR spectra of P-450 reductase. The native holoenzyme was dissolved in 0.025 M Tris/acctate, pH 7.6. containing 10% (by vol.) glycerol, 20% D 2 0 and 35 mM n-octyl glucoside. The temperature was 15°C. A, 0.171 mM oxidized enzyme without cytochrome, 50000 transients; B 0.166 mM oxidized enzyme mixed with 0.175 mM cytochrome c, 80000 transients. An exponential linebroadcning parameter of 20 Hz was applied for each spectrum. The narrow high resonance at approximately - 1.54 ppm in the spectra is probably due to a small amount of contaminating inorganic phosphate. The vertical arrows show chemical shifts of FMN phosphate and FAD phosphate resonances of P-450 reductase.

Fig. 5. Effect of KCI concentration on cross-linking of P-450reductase and either cytochrome c or cytochrome b5 in n-octyl glucoside micelles. 2 nmol P-450 reductase, 6.6 nmol cytochrome c (or 5.1 nmol cytochrome b,) and 7.35 pmol n-octyl glucoside were incubated with 10 mM EDC in 0.025 M Tris/acetate, pH 7.6, containing 10% (by vol.) glycerol at 25°C for 20 min. An appropriate amount of reaction mixture was subjected to SDS/PAGE. (A) Lane 1 shows P-450 reductase and cytochrome c which were mixed and incubated without EDC. Lanes 2 -6 show the reaction at 0,0.05,0.1,0.2 and 0.4 M KC1, respectively. Lane M indicates the marker proteins whose apparent molecular masses were 116 kDa (P-galactosidase), 97 kDa (phosphorylase b), 68 kDa (bovine serum albumin), 45 kDa (ovalbumin) and 14.4 kDa (lysozyme). respectively. The horizontal arrow X, A and B on the left side denote the positions of a major cross-linked spccies. detergent-solubilized native form of P-450 reductase and cytochrome c, respectively. The staining bands on the bottom indicate the dye front, bromophenol blue. (B) Experimental conditions were the same as shown in (A), except that cytochrome b5 was used instead of cytochrome c. The horizontal arrow X, A and B on the right side indicate the positions of a 1 :1 cross-linked band, P-450 reductase and cytocbrome h5. respectively.

the relatively sharp signal at - 1.54 ppm which is presently unassigned and probably due to some inorganic phosphate tightly bound to the protein during the purification procedure. Since these phosphorus residues cannot be washed from the reductase, it is necessary to examine whether or not they are covalently bound with the protein, as detected in flavodoxin [17 - 191. In contrast to the recent data reported by Otvos et al. [22] and Bonants et al. [23], our NMR spectrum of the

reductase did not show the marked phosphorus resonances due to residual 2'-AMP and phospholipids bound to the protein, since the sample of native reductase was fully washed with the Tris/acetate, containing 35 mM n-octyl glucoside before reaction with the cytochromes. It was thus worth observing whether complex formation of either cytochrome c, cytochrome b5 or P-450 with P-450 reductase would influence the 31P resonances of the two flavins. Neither alterations in the chemical shift nor in line broadening of the FMN phosphate are seen upon binding of cytochrome c (Fig. 6 B), therefore the magnetic environments of the flavin phosphate are unperturbed by the paramagnetic heme group. These observations are consistent with the results of flavodoxin [16] which show no major changes in the FMN conformation of flavodoxin upon binding with cytochrome c. Cytochrome b5 and P-450 also gave little or no observable changes on the line widths of the FMN resonances (data not shown). Likewise, "P-NMR spectra indicate no effect of cytochrome binding on the chemical shift of the bound FAD pyrophosphate, suggesting that no remarkable alteration in the FAD environments occurs on complex formation. The previous observation that the FAD phosphate and FMN phosphate resonances are not significantly perturbed upon binding MnZf suggested that the phosphate groups of both the bound FMN and FAD coenzymes are deeply buried in the interior of the protein [22] and the protein structure is rigid when FMN is present.

The P-450 reductase is a complex protein consisting of multiple structural domains which bind P-450, FMN, FAD and NADPH. Presumably these domains must interact in a highly specificmanner to allow effective electron transfer from NADPH to either cytochrome c, cytochrome bS or P-450.

1081

Insight into the nature of protein - protein interactions of redox components have come from studies of flavodoxin and cytochrome c [34, 351. Flavodoxin and cytochrome c interact through the complementation of charged amino acid residues near the partially exposed surfaces of their respective flavin and heme c prosthetic groups [36, 371. Chemical cross-linking data [38]and computer modeling [34] have shown that electrostatic linkages were favored between the lysines of the hemecontaining crevice of cytochrome c and the aspartate and glutamate residue side chains of Clostridium MP flavodoxin. According to the stereoscopic view of this model, two groups of acidic residues positioned on either side of the flavin ring of flavodoxin were localized to the segment at amino acid residues 55-65 on one side, and to residues 13 and 120 on the other [34]. These acidic groups are supposed to interact with basically charged residues near the heme c group. In addition, their CD and 31P-NMR data [I61 support the view that charge pairing of Clostridium MP flavodoxin with cytochrome c results in a marked perturbation of the heme crevice, which allows the benzenoid portion of the FMN to be closer to the heme. Study of the mechanism of electron transfer from microsomal P-450 reductase to either P-450 or cytochrome b 5 , has revealed that complex formation of the redox components due to complementary electrostatic interactions is of functional and structural importance [39 -421. Chemical modification of acidic amino acid residues demonstrated the complex formation due to electrostatic interaction to transfer electrons from P-450 reductase to both physiological (P-450 and cytochromc bs) and non-physiological (cytochrome c) electron acceptors [43, 441. The P-450 reductasc and P-450 form a functional electron-transfer complex through the interaction of carboxyl residues of the reductase and amino residues of P-450 [43,45]. Four discrete groups of acidically charged residues are observed between FMN-binding and FAD-binding segments of P-450 reductase [46]: Glu-Glu-Asp (residues 116- 119); Glu-Asp-Asp (residues 142- 147); Asp-Asp-Asp (residues 207 - 209); Glu-Glu-Asp (residues 21 3 - 21 5). In fact, chemical cross-linking in the presence of EDC was achieved in a study of the P-450-reductase -cytochrome-c complex and amino acid sequence analysis revealed that the cross-link occurred between Lysl3 of cytochrome c and one of six sidechain carboxyl residues (207-21 5) of reductase [ll]. Therefore, these acidic residues clusters of P-450 reductase involve possible candidates for electrostatic charge pairing with a positively charged group of either P-450 or cytochrome c. In addition, the results of the present study and previous chemical cross-linking data [lo] demonstrated that cytochrome b5 binding site(s) is/are also likely to be in very close proximity to the FMN domain of P-450 reductase, since FMN is essentially required to reduce cytochrome bs (Fig. 1) and the purified P-450-reductase - cytochrome-b5 cross-linked derivative can hardly reduce exogenously added cytochrome c or P-450 [47]. Previous modification studies have shown that the exposed side-chain carboxyl groups of Glu47, Glu48 and Glu52 on cytochrome b5 commonly participated in a protein -protein interaction with NADH : cytochrome b5 reductase [48] or P450 [49]. Although it is difficult at this time to specify the nature of the structural differences between cytochrome b5 and P-450 (or cytochrome c) binding regions on the P-450 reductase, on the basis of the tertiary structure of cytochrome h,, we suggest that several basic residues near the FMNbinding domain of P-450 reductase show probable chargepair interactions with the side-chain carboxyl groups of the

1.0

1

h

-2.0 110

130 150 170 190 SEQUENCE POSITION (amino acid residue)

210

Fig. 7. Hydropathy profile of FMN-binding domain of P-450 reductase. The hydropathy plot was calculated according to Kyte and Doolittle [50] using a window of 12 residues by a computer program. Significant hydrophobic moieties were observed at two regions; amino acid residues 125 - 138 and residues 155 - 173. Proposed tyrosine residues (vertical arrows, Y140 and Y178) in close proximity to the bound FMN are shown according to Porter and Kasper [46, 511.

above acidic residues and the exposed heme propionate of cytochrome b,. Blocking the carboxyl group of cytochrome b, or P-450 reductase by chemical modification resulted in great s t h u lation of electron transfer to the heme protein from P-450 reductase [49]. Furthermore, the divalent cations (Cazf and Mg2+) markedly enhanced the activities for reduction of cytochrome b5 as well as of cytochrome c [33]. Although these results must be more adequately explained, we suggest that the interaction between P-450 reductase and either cytochrome c, cytochrome b, or P-450 may involve at least two binding domains: a specifically charged-pair region and hydrophobically bound sites on P-450 reductase and the heme protein. As described above, regions containing multiple charged residues around the FMN-binding domain are considered to be possible candidates for the electrostatic interaction with the cytochrome. However, the presence of salt, diminishing the magnitude of electrostatic binding between these two redox proteins may adversely lead to stronger hydrophobic -hydrophobic interactions. In fact, two peptide portions (residues 125- 138 and 155- 173) in the FMN-binding domain exhibit significant hydrophobicity which can be assumed to be hydrophobic sites interacting with hydrophobic regions of heme proteins (Fig. 7). Thus, intermolecular binding between these hydrophobic moieties near the bound FMN and multiple hydrophobic amino acid residues surrounding the heme is intensified by KCl, and the two prosthetic groups may be able to assume a conformation which causes a faster rate of electron transfer. It is likely that, upon binary-complex formation with the P-450 reductase, conformational perturbation of heme is most conspicuous in cytochrome c. Although we do not have clear evidence, it was assumed that the proteindetergent structure in the micelles was not seriously affected by the salt concentrations we used. However, some studies are

1082 required to clarify whether or not the interaction between noctyl glucoside micelles and the redox components is changed by r ising the C1 con ntratio in the medium.

$e are g r a t e h to Dr. %nya Yo%kawa (Basic Research Laboratory of Himeji Institute of Technology) for his helpful discussion on our data and we are also thankful to Mr. Makoto Naruse (Aichi Medical University) for his technical assistance of 31P-NMR spectral measurements.

REFERENCES 1. lyanagi, T. & Mason, H. S. (1973) Biochemistry 12, 2297-2308. 2. Dignam, J. D. & Strobel, H. W. (1975) Biochem. Biophys. Res. Commun. 63,845 - 852. 3. Dignam, J. D. & Strobel, H. W. (1977) Biochemistry 16, 11161123. 4. Yasukochi, Y. & Masters, B. S. S . (1976) J . Biol. Chem. 251, 5337- 5344. 5. Vermilion, J. L. & Coon, M. J. (1978a) J . Biol. Chem. 253, 26942704. 6. Vermilion, J. L.&Coon, M. J. (1978b)J. Biol. Chem.253,88128819. 7. Iyanagi, T., Anan, F. K., Imai, Y. &Mason, H. S. (1978) Biochemistry 17, 2224- 2230. 8. Tdmburini, P. P., Jansson, I., Favreau, L. V., Backes, W. L. & Schenkman, J. B. (1986) Biochem. Biophys. Res. Commun. 137, 437 -442. 9. Tamburini, P. P. & Schenkman, J. B. (1987) Proc. Natf Acad. Sci. USA 84,ll-15. 10. Nisimoto, Y. & Lambeth, J. D. (1985) Arch. Biochem. Biophys. 241, 386-396. 11. Nisimoto, Y . (1986) J . Biol. Chem. 261, 14232- 14239. 12. Bilimoria, M. H. & Kamin, H. (1973) Annu. N Y . Acad. Sci. 212, 428-448. 13. Prough, R. A. & Masters, B. S. S. (1974) Arch. Biochem. Biophys. 165,263 - 267. 14. Zanetti, G. & Curti, B. (1984) in Flavins and Flavoproteins (Bray, R. C., Engel, P. C. & Mayhew, S. G., eds) pp. 179-182, de Gruyter, Berlin. 15. Przysiecki, C. T., Bhattacharyya, A. K., Tollin, G . & Cusanovich, M. A. (1985) J . Biol. Chem. 260, 1452-1458. 16. Tollin, G., Brown, K., De Francesco, R. & Edmondson, D. E. (1987) Biochemistry 26, 5042 - 5048. 17. Moonen, C. T. W. & Miiller, F. (1982) Biochemistry 21, 408414. 18. Edmondson, D. E. & James, T. L. (1979) Proc. Nut1 Acad. Sci. USA 76,3786 - 3789. 19. Edmondson, D. E. & James, T. L. (1982) in Flavins and Flavoproteins (Massey, V. & Williams, C. H., eds) pp. 111 118, Elsevier North Holland, Amsterdam. 20. James, T. L., Edmondson, D. E. & Husain, M. (1981) Biochemistry 20,617-621. 21. Davis, M. D., Edmondson, D. E. & Miiller, F. (1984) Eur. J . Biochem. 145.237 - 243.

22. Otvos, J. D., Krum, D. P. & Masters, B. S. S . (1986) Biochemistry 25, 7220 - 7228. 23. Bonants, P. J. M., Miiller, F., Vervoort, J. & Edmondson, D. E. (1990) Eur. J . Biochem. 190, 531 -537. 24. Ito, A. & Sato, R. (1976) J . Biol. Chem. 243,4922-4923. 25. Omura, T. & Sato, R. (1964) J. Biol. Chem. 239,2370-2378. 26. Imai, Y. & Sato, R. (1977) Biochem. Biophys. Res. Commun. 75, 420 -425. 27. Faeder, E. J. & Siegel, L. M. (1973) Anal. Biochem. 53,322- 336. 28. Rudolph, S. A. & Krueger, A. (1976) J . Chromatogr. 116, 462464. 29. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193, 265-275. 30. Van Gelder, B. F. & Slater, E. C. (1962) Biochim. Biophys. Actu 58, 593 - 595. 31. Strittmatter, P., Rogers, M. J. & Spatz, L. (1972) J . Biol. Chem. 247,7188-7194. 32. Phillips A. H. & Langdon, R. G. (1962) J. Biol. Chem. 237,2652 2660. 33. Tamura, M., Yoshida, S., Tamura, T., Saitoh, T. & Takeshita, M. (1990) Arch. Biochem. Biophys. 280, 313-319. 34. Weber P. C. & Tollin, G. (1985) J . Biol. Chem. 260, 5568- 5573. 35. Cheddar, G., Meyer, T. E., Cusanovich, M. A,, Stout, C. D. & Tollin, G. (1986) Biochemistry 25,6502-6507. 36. Simondsen, R. P., Weber, P. C., Salemme, F. R. & Tollin, G. (1982) Biochemistry 21,6366-6375. 37. Matthew, J. B., Weber, P. C., Salemme, F. R. & Richards, F. M. (1983) Nature 301, 169- 171. 38. Dickerson, J. L., Kornuc, J. J. & Rees, D. C. (1985) J . Biol. Chem. 260, 5175-5178. 39. Poulos, T. L. & Kraut, J. (1980) J . Biol. Chem. 255, 1032210330. 40. Salemme, F. R. (1977) Annu. Rev. Biochem. 46, 299 - 329. 41. Hackett, C. S. & Strittmatter, P. (1984) J. Biol. Chem. 259,32753282. 42. Bernhardt, R., Pommerening, K . & Ruckpaul, K. (1 987) Biochem. Int. 14, 823 - 832. 43. Tamburini, P. P., MacFarquhar, S. & Schenkman, J. B. (1986) Biochem. Biophys. Res. Commun. 134, 519- 526. 44. Nadler, S. G. & Strobel, H. W. (1988) Arch. Biochem. Biophys. 261,418 - 429. 45. Makower, A., Bernhardt, R., Rabe, H., Janig, G. R. & Ruckpaul, K. (1984) Biochim. Biophys. Acta 43, 1333- 1341. 46. Porter, T. D. & Kasper, C. B. (1985) Proc. Nut1 Acad. Sci. USA 82,973 - 977. 47. Nisimoto, Y. & Murakami, 0. H. (1988) Biochemistry 27,58695876. 48. Dailey, H. A. & Strittmatter, P. (1979) J . Bio/. Chem. 254, 53885396. 49. Tamburini, P. P., White, R. E. & Schenkman, J . B. (1985) J . Bid. Chem. 260,4007-4015. 50. Kyte, J. & Doolittle, R. F. (1982) J . Mol. Biol. 157, 105-132. 51. Porter, T. D. & Kasper, C. B. (1986) Biochemistry 25,1682- 1687.