Inhibition of cytochrome P-450 activity in rat liver microsomes by the naturally occurring flavonoid, quercetin

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 240, No. 1, July, pp. 345-357, 1985

Inhibition

of Cytochrome the Naturally ROSEMARY

Department

P-450 Activity in Rat Liver Microsomes Occurring Flavonoid, Quercetin’

L. SOUSA

AND

Biological

Sciences,

of Applied

Cambridge,

Received

October

MICHAEL

A. MARLETTA’

Massachusetts

Massachusetts

24, 1984, and in revised

by

Institute

of

Technology,

02139

form

February

25, 1985

The kinetic characteristics and mechanism of flavonoid inhibition of cytochrome P-450-mediated reactions were examined in rat liver microsomes, using the naturally occurring flavonoid, quercetin (3,3’,4’,5,7-pentahydroxyflavone). Quercetin inhibited the 0-deethylation of ethoxyresorufin in P-naphthoflavone-induced microsomes by 15-80% at concentrations of lo-250 nM. The pattern of inhibition was dependent on quercetin concentration. Quercetin also inhibited p-nitroanisole demethylation and benzo(a)pyrene hydroxylation, but did not change the proportions of the individual benzo(a)pyrene metabolites in comparison to controls. Specific steps in the P-450 reaction pathway were tested for sensitivity to quercetin inhibition. The Km values of the P-450 substrates tested were increased in the presence of quercetin; competition for and/or alteration of the substrate binding site contributes to the mechanism of inhibition. In experiments under anaerobic, carbon monoxide-saturated conditions, quercetin did not inhibit cytochrome P-450 reduction by NADPH-cytochrome P-450 reductase. The cumene hydroperoxide-supported 0-deethylation of ethoxyresorufin was inhibited by quercetin (15-60% inhibition at concentrations of 50-300 nM), suggesting that quercetin may interfere with the formation or breakdown of the oxygenated heme complex. Stoichiometry experiments established that quercetin is a potent uncoupler of P-450 reactions, elevating the rates of HzOz formation almost twofold. St:ructure/activity studies indicated that certain other naturally occurring flavonoids .were at least as potent inhibitors of ethoxyresorufin deethylation as quercetin. These findings are of interest in light of the significant dietary exposure of the human population to the tlavonoids. o 19% Academic PESS, I~C.

Flavonoids are one of the most prevalent class of compounds in nature and are widely distributed in edible plants (2). Various aspects of their biological activity have received attention and considerable interest has focused on the ability of

flavonoids to modulate cytochrome P-450 activity. Early studies reported that the synthetic flavonoid, ANF3 (Fig. l), exhib-

3 Abbreviations used: ANF, 7,8-benzoflavone, anaphthoflavone; BNF, 5,6-benzoflavone, @-naphthoflavone; AHH, aryl hydrocarbon hydroxylase; 3-MC, 3-methylcholanthrene; PB- (BNF-) microsomes, liver microsomes from phenobarbital(or BNF-) treated rats; B(a)P, benzo(a)pyrene; DCPIP, dichlorophenolindophenol; CHP, cumene hydroperoxide; DMSO, dimethyl sulfoxide; galangin, 3,5,7-trihydroxyflavone; kaempferol, 3,4’,5,7-tetrahydroxyflavone; morin, 2’,3,4’,5,7-pentahydroxyflavone, dihydroquercetin, 3,3’,4’,5,7-pentahydroxyflavanone.

‘This work was supported by NIH Grant 5-POlES00597. Portions of this work were presented at the 75th Annual Meeting of the American Society of Biological Chemists, St. Louis, MO., 1984 (1) and at the 188th National Meeting of the American Chemical Society, Philadelphia, Pa., 1984. z To whom correspondence should be addressed at Room 56-229, Massachusetts Institute of Technology, Cambridge, Mass. 02139. 345

0003-9861/85 Copyright All rights

$3.00

0 1985 by Academic Press, Inc. of reproduction in any form reserved.

SOUSA

AND

OH

QUERCETIN

BNF

ANF

FIG.

I. Flavonoid

structures.

ited preferential inhibition or stimulation of in vitro microsomal AHH activity, depending on the pretreatment of the animals and the tissue of microsomal isolation (3). In comparison, another synthetic flavonoid, BNF, when administered in vivo, induces the synthesis of specific P-450 isozymes with high AHH activity (4). It has been found that, in general, flavonoids possessing hydroxyl groups often inhibit P-450 activity in vitro, while flavonoids lacking hydroxyl groups can act to stimulate P-450 activity directly (5). Conney and co-workers have focused on the mechanisms by which ANF stimulates P-450 activity in vitro (6-8), and have also studied the stimulation of zoxazolamine metabolism by flavone, a naturally occurring flavonoid, in vitro, and in vivo (9, 10). In contrast, little is known regarding the mechanism of inhibition of P-450 activity by hydroxylated flavonoids. Quercetin (Fig. 1) is one of the most abundant of the naturally occurring hydroxylated flavonoids, and dietary exposure is not trivial (11). Previous work has shown that quercetin inhibits B(a)P hydroxylase activity in liver microsomes from control and 3-MC-treated rats (3), and in human

MARLETTA

liver microsomes (5). In addition, aromatase (estrogen synthetase) activity in human placental microsomes is inhibited by quercetin (EL!), as is zoxazolamine metabolism in vitro (9). However, the mechanism for this inhibition is not well understood. The results we report here consist of two parts. The first is a kinetic characterization of the inhibition by quercetin in liver microsomes from rats that had been pretreated with BNF. We found quercetin to be a potent inhibitor of cytochrome P450 activity in these microsomes, showing a complex inhibition pattern that is dependent on quercetin concentration. Since even induced microsomes contain several P-450 isozymes (13), we determined whether specific isoenzyme inhibition could account, at least in part, for the complex kinetics by testing the effects of quercetin on the metabolite profile of B(a)P. B(a)P, a polycyclic aromatic hydrocarbon, is an efficient substrate for BNFinduced microsomes, and P-450 mediated metabolism produces a number of oxidized products whose occurrence can be related to microsomal isoenzyme composition. The second part of our investigation was directed toward understanding the point or points in the P-450 cycle (Fig. 5) at which quercetin acts, along with structure/activity aspects of the inhibition. The potency of quercetin as a P-450 inhibitor, and its prevalence in nature, present the possibility of in vivo modulation of P-450 activity and may lead to the development of specific and potent P-450 inhibitors. MATERIALS

AND

METHODS

Preparation and solubilization of microsomes. Male Long Evans rats, 50-75 g (Charles River Breeding), were used as a source of microsomes. Liver microsomes from control rats, or rats that had been pretreated with BNF (80 mg/kg i.p. for 3 days) or PB (1 g/liter drinking water for 5 days) were prepared according to the methods described in Ryan et al. (14). Protein concentrations, measured by the Lowry procedure (15), were 48, 26, and 24 mg/ml, respectively, for the PB-, BNF-, and control microsomes. Specific contents, determined by the method of Omura and Sato (16), were 3.02, 1.35, and 0.41 nmol P-450/mg protein for the respective microsomes. Solubilized microsomes were prepared by the method of West et al. (1’7).

P-450

INHIBITION

NADPH-dependent 0-deethylation of ethmyreso&n Ethoxyresorufin was a generous gift from Professor C. Walsh and Dr. Alan Klotz of MIT. The 0-deethylation reaction was monitored at room temperature in a tungsten-lighted room by a direct fluorometric assay previously described (18) with minor modifications. Briefly, the following reaction mixture was added directly to a fluorometer cuvette: 1.97 ml of 0.1 M potassium phosphate buffer, pH 7.8, 4 ~1 of ethoxyresorufin (prepared in DMSO to give final concentrations of 0.05-0.5 PM), and microsomes (16 fig from BNF-treated rats, 48 pg from PB- and control rats). A baseline was recorded on a Farrand Model Mark I Spectrofluorometer with X,, = 572 nm and h., = 586 nm. A 20-~1 aliquot of 25 mM NADPH (Sigma) (final concentration, 0.25 mM) was added to the mixture to initiate the reaction, and the increase in fluorescence was recorded. The reaction was linear for at least 110 min for the higher ethoxyresorufin concentrations, although only the first minute of the reaction was used to calculate the kinetic data; all substrate concentrations gave linear rates during this first minute. When inhibitor was used, 4 ~1 in DMSO (final concentrations: lo-250 nM for quercetin, 12.5-100 nM for other flavonoids) was added to the fluorometer cuvette along with the reaction mixture; incubations without flavonoid contained 4 ~1 DMSO. The final DMSO concentration in the incubation (0.4%) did not affect the reaction. The fluorometer was calibrated with a solution of 25 nM resorufin and although quenching curves were determined initially, quenching was negligible, so that experimental points did not require adjustment. Cumene hydrqmmmide-dependent Ckkethylation of ethmyresow$n Cumene hydroperoxide (CHP) was obtained from K&K Biochemicals. The incubations were similar to those described for the NADPHdependent reaction, except that 0.25 mM CHP was added to the fluorometer cuvette to initiate the reaction. Ethoxyresorufin concentrations ranged from 0.5 to 7.5 PM, quercetin concentrations were 50-300 nM, and BNF’-microsomes (0.34 nmol cytochrome P-450) were used. Control experiments were performed to ensure that quercetin would not simply inhibit the binding of CHP to P-450: the formation of a spectral complex at 440 nm between CHP and P-450 (19, 20) was monitored by repeated scanning of this spectral region in a Perkin-Elmer Model 553 spectrophotometer, for a period of 10 min. in the presence and absence of quercetin. 0-Demethylation of pnitroanisole. This reaction was monitored in BNFor PB-microsomes in a manner similar to that of Zannoni (21). The following components were added directly to a spectrophotometer cuvette for a final volume of 1.0 ml: 70 mM sodium phosphate buffer, pH 7.8, 0.34 mg PB-microsomal protein (1.0 nmol P-450) or 0.3 mg BNFmicrosomes (0.4 nmol P-450), and p-nitroanisole

BY

FLAVONOIDS

347

(0.05-2.0 mM final concentrations). The gnitroanisole was delivered from ethanol solutions such that the final ethanol concentration did not exceed 0.2%. The reaction was initiated by the addition of NADPH (0.3 mM final concentration) and the increase in absorbance at 403 nm corresponding to pnitrophenol formation was monitored at room temperature on a Perkin-Elmer Model 553 spectrophotometer using an empirically determined extinction coefficient of 14.7 mM-’ cm-‘. When present, quercetin was added from DMSO solutions to give final concentrations of 3.0-12.5 pM for incubations with PB-microsomes and 0.25-5.0 pM for BNF-microsomal assays; otherwise, samples contained only the DMSO vehicle. DMSO did not affect the reaction. Hydraxylutim of B(a)P-Flumvme.~ric assay. In a total volume of 1.0 ml were added 0.1 M potassium phosphate buffer, pH 7.4, containing 0.1 mM EDTA and 3 mM MgQ, 100 j.tM B(a)P (final concentration) delivered in acetone, 58 pg BNF-microsomal protein, and, when present, quercetin, delivered from acetone solutions, ranging in concentration from 0.01 to 500 PM. Samples without quercetin contained acetone only. Reactions were initiated with 0.25 mM NADPH and incubated at 37°C for 5 min, followed by extraction with 2 ml hexane:acetone (3:1), then back extraction into 2 ml of 0.5 N KOH. Fluorescent phenolic metabolites were measured in a Farrand Model Mark I spectrofluorometer calibrated to 0.1 pM 3-OH B(a)P, with X,, = 396, &,,, = 522. Background fluorescence due to quercetin was not observed. HP.92 analysis of B(a)P metabolites. Immediately before use [3H]B(a)P (NEN, sp act, 78.9 Ci/mmol) was purified by TLC on a silica G glass TLC plate (Analtech) with a benzene mobile phase. B(a)P was located by fluorescence and scraped from the plate for further purification on a silica column (Biosil, 200-325 mesh; Bio-Rad, Richmond, Calif.) (in a loml glass pipet) using a benzene mobile phase. Pure unlabeled B(a)P was added simultaneously to the column to give a final specific activity of 50 mCi/ mmol. The column fraction containing the [3H]B(a)P peak (as determined by liquid scintillation counting) was dried under a stream of nitrogen and redissolved in a small volume of acetone so that each incubation contained a B(a)P concentration of 10 pM at a specific activity of 50 &i/pmol. The incubations, in a total volume of 1.0 ml, contained 0.1 M potassium phosphate buffer, pH 7.4, 0.1 mM EDTA, 3 mM MgC12, 116 pg BNF-microsomal protein, and [‘Hp(a)P. The reaction was initiated with 15 ~1 of 75 mM NADPH, and was incubated in a 37°C water bath for 30 min. When inhibitor was used, it was added from acetone solutions for final concentrations of 10 nM to 10 pM. After the incubation, authentic unlabeled B(a)P metabolite standards (provided by the National Cancer Institute) were added to the reaction mixture followed by immediate extraction with three 2-ml volumes of ethyl acetate:acetone (3:l). By adding the

348

SOUSA

AND

standards at this point, HPLC recovery from the work-up procedure could be quantified and results could be adjusted accordingly. The combined extracts were dried with anhydrous MgS04 and filtered, and solvent was evaporated under nitrogen. The metabolites were then redissolved in 0.4 ml methanol, and the methanol concentration was adjusted to 60% with water. Metabolites and parent compound were separated on a Waters HPLC system with an Ultrasphere ODS (4.6 X 250 mm) column (Altex Inc., Berkeley, Calif.). Methanol concentration was increased linearly from 60 to 100% over 40 min. Flow rate was 1.0 ml/min with a column temperature of 25°C. Absorbance at 270 nm was monitored with a Micromeritics Model 786 variable-wavelength detector. Fractions were collected directly into 20-ml scintillation vials, 7 ml of Aquasol II (NEN, Boston, Mass.) was added to each vial, and radioactivity counted using a Beckman Model LS-3150~ liquid scintillation counter. Binding spectra of quwcetin. A Hitachi Model 557 dual-wavelength, double-beam spectrophotometer with split-compartment cuvettes was used to measure difference spectra of quercetin with BNF-, PB-, and control microsomes according to the method described by Jefcoate (22). Quercetin concentrations ranging from 6.25 to 200 fiM were used. Quercetin was dissolved in DMSO, and reference cuvettes contained the appropriate amount of DMSO. Assay of cytochrm P-J&O reductase activity. The reduction of cytochrome c by NADPH-cytochrome P-450 reductase was assayed in BNF-, PB-, and control microsomes according to the method of Strobe1 and Dignam (23). However, quercetin was found to rapidly reduce cytochrome c directly, and therefore the artificial electron acceptor, DCPIP (23), had to be used instead as a substrate for the reductase in order to determine the effects of quercetin on reductase activity. DCPIP was only very slowly reduced by quercetin relative to the enzymatic reduction. Assay of cytochrm P-MO reduction Methods similar to those already described were used (7, 24). Buffers were degassed by applying a water aspirator vacuum and all solutions were saturated with carbon monoxide (99.5% min, Matheson). To a total volume of 1.0 ml were added microsomes (2 nmol of cytochrome P-450), buffer, and quercetin (2-50 pM for BNF-microsomes, 40-300 pM for PB-microsomes). When present, the following were also added at this point: an oxygen-scavenging system (60 mM glucose, 0.1 mg/ml glucose oxidase, 3000 units catalase/ml) and/or substrate (0.4 pM ethoxyresorufin in BNFmierosomes, 150 PM benzphetamine in PB-microsomes). The above concentrations were chosen so that the quercetin:P-450 ratios in these experiments were exactly the same as in the kinetic experiments, with substrate (when present) at V,, concentrations. The buffers were the same as those used for the kinetic experiments.

MARLETTA The incubation was gently bubbled with carbon monoxide for approximately 5 min, and then transferred to an anaerobic euvette that was subsequently flushed with CO for 3 min. The reaction was initiated with 10 ~1 of 50 mM NADPH in buffer, saturated with CO. To accomplish rapid mixing, the NADPH solution was added from a syringe through a septum covering the cuvette, and was rapidly stirred with a Teflon plunger inserted tightly into the septum. The reduction of cytoehrome P-450 was monitored at room temperature as an increase in absorbance at 448 (or 450) nm on a Perkin-Elmer Model 553 spectrophotometer. Full scale absorbance was 0.2, with a chart speed of 20 cm/min. Stoichiometry experiments. Separate incubations were performed in parallel to determine the rates of resorufin formation, NADPH oxidation, oxygen consumption, and H202 formation. All incubations contained, in a total volume of 4.0 ml, 50 mM potassium phosphate buffer, pH 7.8, and BNF-microsomes (1 nmol P-450/ml incubation). Some incubations additionally contained 1 pM ethoxyresorufin and/or 600 nM quercetin. Reactions were initiated with 0.15 mM NADPH. Resorufin formation was monitored fluorometrically as described earlier, except that the instrument was calibrated to 1 +M resorufin. At 600 nM quercetin, resorufin formation is inhibited by 80%. To measure NADPH oxidation, the reaction was scaled down to 1.0 ml. To inhibit pyrophosphatase, 2 mM 5’-AMP was included in some incubations (25). Even in the absence of 5’-AMP, the ratio of NADPH to 02 utilized was never less than 1.0; apparently pyrophosphatase activity was not interfering with the assay. NADPH was added and an aliquot of the reaction was immediately removed, diluted l/5, and measured at 340 nm. After 5-10 min, another aliquot was removed and treated similarly. Turbidity of the mixture prevented the measurement of A%,, directly; the time between aliquot removal and A%,, measurement was 3-5 s. Oxygen consumption was measured over a 5- to lo-min period with a YSI Model 53 oxygen monitor equipped with a Clark oxygen electrode, calibrated to 258 pM Oz (26, 27). HzO, was determined by the difference in rates of oxygen consumption in the presence and absence of 5 mM azide. Separate fluorometric experiments showed that the azide did not inhibit product formation. HsOs was also determined by the scopoletin method (28). Both methods gave similar results. With no substrate present, for example, the value was 2.50 nmol Hz02 mini’ nmol P-450-l by the scopoletin method (Table III), and 2.49 nmol Hz02 mini’ nmol P-450-l on the oxygen electrode. Spectral experiments. Quereetin absorbance was repeatedly scanned around its X,,, (379 nm) at approximately l-min intervals for 10 min on a Perkin-Elmer Model 553 spectrophotometer. Typically, both cuvettes contained ‘70 mM sodium phosphate buffer, pH 7.8, 4 ~1 DMSO vehicle, 0.3 mM

P-450

INHIBITION

BY

1

349

FLAVONOIDS

ETHOWAESOR”FIN]-‘,

[A]-’

FIG. 2. Inhibition of NADPH-supported ethoxyresorufin O-deethylation by quercetin. Experiments were performed as under Materials and Methods. The data are plotted on a Lineweaver-Burk plot, and each point represents the mean f SD of at least three determinations. Quercetin concentrations: 0, control (no quercetin); 0,lO nM; I, 25 nM; D, 50 AM; A, 100 nM; A, 250 nM.

NADPH, and rnicrosomes (0.98, 0.50, and 0.25 nmol P-450 from P:B-, BNF-, and control microsomes, respectively). The sample cuvette additionally contained 12.5 @M Iquercetin, delivered in 4 ~1 DMSO. RESULTS

Inhibition of NADPH-dependent ethoxyresorz&il;! 0-deethylation by quercetin. Quercetin is a potent inhibitor of ethoxyresorufin O-deethylation in BNF-microsomes; concentrations as iow as 10 nM caused 15% inhibition, while 250 nM quer-

cetin caused 80% inhibition. Our findings were unusual in that the pattern of inhibition was dependent on inhibitor concentration (Fig. 2). The two lowest quercetin concentrations, 10 and 25 nM, exhibited competitive inhibition. The three highest quercetin concentrations examined, 50, 100, and 250 nM, showed a mixed inhibition pattern (Table I). Changes in the order of addition of substrate and inhibitor to the reaction mixture did not affect these results.

350

SOUSA

AND

MARLETTA

TABLE KINETIC

PARAMETERS

Ethoxyresorufin deethylation, BNF-microsomes Control 10 25 50 100 250

nM nM nM nM nM

Q Q Q Q Q

ASSOCIATED

I

WITH QUERCETIN

Inhibition

Kl (PM)

Competitive Competitive Mixed Mixed Mixed

0.14 0.14 0.20 0.33 0.60 0.75

INHIBITION

(nmol

Inhibition of p-nitroanisole Gdemethylation by quercetin. The turnover numbers (2.6 and 2.3 min-‘) and K, values (78 and 100 PM) for pnitroanisole demethylation were similar in PB- and BNF-microsomes, respectively, and in both cases quercetin exhibited a mixed inhibition pattern. However, quercetin was a more potent inhibitor with BNF-microsomes; 0.25-5.0 PM gave 26-‘78% inhibition, while in PB-

% of

OF QUERCf3lN

and Methods. were determined

microsomes 3.0-12.5 30-63% inhibition.

K (PM)

P-450-l)

0.034 0.034 0.133 0.133 0.133 Km and V,,,,, values as in Segel (31).

PM

quercetin

II

ON T% PAlTERN IN BNF-YICRWOMS

OF B(a)P

METABOUTES

no quercetin _W 10 nM Q __.____ 0 100 nM Q.....etd 250 nM Q-e....0 10 uM Q _______ q

30

TOTAL METABOLITES

20

9,lO dial

435 dial

7,6 dial

DETECTABLE

Note. Microsomal incubations contained quercetin. B(a)P metabolites were separated

were

gave

Analysis of B(a)P metabolites. In the fluorescence assay, quercetin inhibited the formation of B(a)P phenolic metabolites by BNF-microsomes from 32% at 0.1 pM quercetin to 95% at 500 pM quercetin. The rate for the uninhibited reaction was 90 pm01 min-l nmol P-450-l. The effects of quercetin on the metabolite profile of [3H]B(a)P were examined by HPLC separation and subsequent quantification of metabolites and parent compound by liquid scintillation counting. The HPLC assay detected the B(a)P metabolites shown in Table II, which were TABLE

INFLUENCE

V man nmol

REACTIONS

1.0 0.99 0.90 0.78 0.40 0.30

Note. Experiments were performed as described under Materials calculated from Lineweaver-Burk plots shown in Fig. 2. K, values

Control and PB-microsomes exhibited negligible turnover of ethoxyresorufin (data not shown), illustrating the specificity of this substrate for the BNF-induced form(s) of the enzyme as previously reported (18).

min-’

OF P-450

1.6 dime BENZOfA)PYRENE

3,6 dime

6,12 dime

9-OH

3-OH

METABOLITES

10 PM B(a)P and the indicated concentrations of by HPLC as described under Materials and Methods.

P-450

INHIBITION

present in the indicated proportions (expressed as the percentage of total metabolites). The quercetin concentrations were chosen based on their effects in either the ethoxyresorufin or p-nitroanisole assays, or in the binding assays (see below). Table II shows that these concentrations, ranging lOOO-fold from 10 nM to 10 pM, did not significantly alter the B(a)P metabolite pattern; the amount of each metabolite relative to one another remained approximately the same as controls. However, as also shown in the fluorescence assay, the total metabolism was decreased by quercetin, which supports earlier studies (3, 5). Binding spectra of quercetin with microsomes. The difference spectra resulting from the interaction of quercetin with BNF-, PB-, and control microsomes were measured. Quercetin exhibits a type I binding spec.trum with all types of microsomes. The binding of quercetin to PB-, BNF-, and control microsomes appeared biphasic, when plotted as I/AA vs l/ [quercetin] or as a Scatchard plot (22). Therefore, high- and low-affinity spectral binding constants (KS) were calculated: PB-microsomes, KS = 8 pM (high affinity), KS = 288 PM: (low affinity); BNF-microsomes, KS = 20 PM (high affinity), KS = 513 pM (low affinity); control microsomes, K, = 28 pM (high affinity), KS = 182 pM (low affinity). The same measurements were made on cholate-solubilized microsomes, in an attempt to separate the binding to microsomal membrane versus binding to the enzyme. Biphasic binding was again observed for all types of microsomes, and KS values ar’e similar to those before solubilization. Efect of quercetin on cytochrome P-450 reductase activity. Quercetin had only a slight, if any,, inhibitory effect on NADPHcytochrome P-450 reductase when using DCPIP as electron acceptor. For example, 125 pM quercetin inhibited DCPIP reduction by 20 and 40% in control and BNFmicrosomes, respectively, but had no effect in PB-microsomes. The inhibition occurs at quercetin concentrations much higher than those used in the kinetic experiments.

BY

FLAVONOIDS

351

Efect of quercetin on cytochrome P-450 reduction, Cytochrome P-450 is the natural acceptor of electrons from NADPH-cytochrome P-450 reductase. The reduction of cytochrome P-450 was monitored as described under Materials and Methods. The presence of the oxygen-scavenging system in BNF- or PB-microsomes caused no change in the initial rate of reaction, but in BNF-microsomes the final AAdd was increased to approximately 0.12 from 0.06. We observed a fast and a slow phase of reduction as previously reported (29, 30). We found that quercetin, at all concentrations tested, did not inhibit either phase of cytochrome P-450 reduction in BNFor in PB-microsomes, and this held true whether or not the oxygen-scavenging system and/or substrate were present. Both the initial rate and the final AA were unchanged from controls in the presence of quercetin: 7.5 nmol P-450 reduced/min, final AAd = 0.12 for BNFmicrosomes (67% reduction); 8.7 nmol P450 reduced/min, final AAdS = 0.10 for PB-microsomes (57% reduction). Inhibition of cumene hydroperoxide-dependent ethoxyresorujk 0-deethylation by quercetin. To determine whether quercetin can inhibit a step in the reaction pathway after cytochrome P-450 reduction, CHP was employed in the ethoxyresorufin assay in order to bypass the steps requiring NADPH, P-450 reductase, and oxygen. As shown in Fig. 3, 50-200 nM quercetin inhibited this reaction, with the lower concentrations, 50 and 100 nM, exhibiting a competitive inhibition pattern and the higher concentrations, 150 and 200 nM, showing mixed inhibition. Kinetic parameters are listed in Fig. 3. This biphasic inhibition pattern is similar to that observed in the NADPH-supported ethoxyresorufin reactions. To ensure that the inhibition by quercetin was not simply the result of inhibition of CHP binding to the enzyme, control experiments monitored the formation of a spectral complex at 440 nm between CHP and the enzyme (19, 20), and the rate of formation and final extent of this complex were not affected in the presence of quercetin (not shown).

352

SOUSA

AND

MARLETTA

FIG. 3. Inhibition of CHP-supported ethoxyresorufin 0-deethylation by quercetin. Experiments were performed as under Materials and Methods. The data are plotted on a Lineweaver-Burk plot, and each point represents the mean t SD of at least two determinations. Quercetin concentrations: 0, control (no quercetin); 0, 50 nM; n , 100 nM; 0, 150 nM; A, 200 nM. Km = 0.62 avg. K, (competitive) = 0.22 PM. PM, vrnax = 0.06 nmol min-’ nmol P-450-‘,

Stoichiometry experiments. The results presented in Table III show that the proportions of NADPH oxidized, oxygen consumed, and H202 and product formed were close to the expected 1:l:l ratio. Although 600 nM quercetin alone did not appreciably alter the rates from the case where no substrates at all were present, 1 PM ethTABLE

oxyresorufin stimulated NADPH and oxygen consumption. Importantly, when 600 nM quercetin was included in the incubation with 1 PM ethoxyresorufin, 80% inhibition of resorufin formation occurred along with an almost twofold increase in the rate of H202 formation, indicating that quercetin is uncoupling the reaction III

STOICHIOMETRYOF ETHOXYRESORUFINDEETHYLATIONIN BNF-MICROSOMES nmol Condition No substrate 1 pM Ethoxyresorufin ~FM Ethoxyresorufin + 600 nM quercetin 600 nM Quercetin

-NADPH

min-’

nmol

-02

P-450-i HA

Resorufin

Ratio NADPH/Oa/HaOa + Resorufin

3.08 31 0.42 4.02 + 0.37

3.06 f 0.23 3.53 f 0.15

2.50 +- 0.08 2.54 + 0.09

0 1.26 f 0.08

1.01/1.00/0.82 1.14/1.00/1.08

4.10 f 0.35 3.08 f 0.53

4.16 f 0.26 3.00 ?I 0.00

4.05 -c 0.10 2.91 k 0.18

0.26 f 0.00 0

0.98/1.00/1.04 1.03/1.00/0.97

Note. All incubations contained 50 mM potassium phosphate P-450/ml incubation). Rates were measured as described under by the scopoletin method. The means f SD of at least two ratio, we let 0, = 1.00.

buffer, pH 7.8, and BNF-microsomes (1 nmol Materials and Methods. Hz02 was determined determinations are shown. To calculate the

P-450

INHIBITION

Spectral experiments. Figure 4 shows that a time-dependent decrease in absorbance was seen during the repeated scanning of a range encompassing the X,,, of quercetin (379 nm). This absorbance change did not occur if either microsomes, NADPH, or O2 were omitted, or if the system was saturated with CO. Furthermore, the extent of the decrease in absorbance was dependent on both quercetin and microsome concentration. Figure 4 illustrates the effect in PB-microsomes, where a 40% decrease in quercetin absorbance was seen over a lo-min period. Similar results were observed with BNFand control microsomes (not shown). Structure/activity experiments. Table IV illustrates thle results of structure/activity studies with other naturally occurring llavonoids. These experiments were carried out with 13NF-microsomes using the ethoxyresorufin assay. Dihydroquercetin was equipotent to quercetin at inhibiting ethoxyresoruhn deethylation, and kaempferol and ga.langin were stronger inhibitors than quercetin. Morin displayed the weakest inhibition. The order of inhibitory potency, from strongest to weakest, is galangin > kaempferol > quercetin - dihydroquercetin > morin. DISCUSSION

The results show that the naturally occurring flavonoid, quercetin, is a potent inhibitor of at least three cytochrome P450-catalyzed reactions in rat liver microsomes in vitro: ethoxyresorufin O-deethylation, p-nitroanisole 0-demethylation, and B(a)P hydroxylation. The kinetic data indicate that inhibition is complex, with the type of inhibition depending upon inhibitor concentrations, and the potency depending on the pretreatment of the animals. In an attempt to unravel this kinetic complexity, it was necessary to understand the effect of quercetin on the various steps in the proposed mechanism for P-450-mediated hydroxylation (Fig. 5). In Table I, we interpret the altered K, values in the presence of quercetin as evidence that quercet.in interferes with substrate

BY

353

FLAVONOIDS

340

380 Wavelength

420

460 (nm)

FIG. 4. Interaction of quercetin with eytochrome P-450. Reference and sample cuvettes contained buffer, PB-microsomes, and NADPH in an aerobic atmosphere, as described under Materials and Methods. The sample cuvette additionally contained 12.5 pM quercetin. The first scan is numbered 1 and the last scan (approximately 10 min later) is numbered 10.

binding (Step 1, Fig. 5). Alternatively, an altered K, might reflect a change in &, the rate for product formation, if rapid equilibrium conditions, which are difficult to verify, are not met (31). However, our first suggestion is supported by the additional finding that quercetin itself exhibits a type I binding spectrum and may therefore compete directly for the substrate binding site. In fact, our data revealed the presence of two type I binding sites for quercetin; this is not affected by cholate solubilization and therefore interaction directly with the protein is the most important factor for both the high- and low-affinity binding. However, the optically determined K, values for quercetin are higher than the quereetin concentrations which give inhibition. Another step along the catalytic pathway that could be susceptible to quercetin inhibition is Step 2 (Fig. 5), the reduction of cytochrome P-450 by the reductase. Inhibition could result from binding to the reductase or to P-450, and/or from

354

SOUSA

AND

MARLETTA

TABLE FLAVONOID

IV STUDIES

STRUCTURE/ACTIVITY

Concentration Compound

tested

(nM)

12.5

25

100

Quercetin

n.d.”

20

60

Morin

n.d.

4

33

Dihydroquercetin

n.d.

18

59

20

48

82

42

58

91

OH

OH

0

OH

OH

0 OH

OH

0

OH

Kaempferol OH

0

Note. Ethoxyresorufin assays were performed as described the percentage of inhibition at 500 nM ethoxyresorufin (V,.,). ‘Not done.

disturbance of the lipid environment such that electron flow to cytochrome P-450 is altered. Experiments using the artificial electron acceptor, DCPIP, in an assay for the reductase indicated that quercetin is not an effective inhibitor, at least not at concentrations that were relevant to our kinetic experiments. Although Buening et al. (5) did find quercetin inhibition of the reductase in human liver microsomes using cytochrome c as electron acceptor, we

under

Materials

and Methods.

Values

indicate

found that quercetin directly reduced cytochrome c, preventing its use as an assay. Evidence for the need to repeat this assay using the natural electron acceptor, cytochrome P-450, was provided by the findings of Huang et al. (7) that ANF stimulated cytochrome P-450 reduction but not cytochrome c reduction by reductase in rabbit liver microsomes. We performed experiments to measure cytochrome P-450 reduction under anaerobic,

P-450

FIG.

reductase;

5. Proposed reaction R, substrate; XOOH,

INHIBITION

pathway organic

for P-450 hydroperoxide.

carbon monoxide-saturated conditions and found that quercetin did not cause inhibition of the reductase in either BNF- or PB-microsomes, whether or not a substrate was also present. Therefore, it cannot account for the inhibition observed. Use of CHP enables one to bypass the steps involving NADPH, cytochrome P450 reductase, and oxygen (Fig. 5). We carried out ethoxyresorufin experiments using CHP and found that quercetin does inhibit this reaction (Fig. 3). Control experiments established that quercetin does not inhibit the reaction simply by preventing CHP binding to the enzyme: the spectral complex at 440 nm formed between this alkyl peroxide and the enzyme was not affected in the presence of quercetin. Hence, it seems likely that quercetin acts on the P-450-Fe III-peroxide anion complex and/or a step beyond it (Fig. 5). The stoichiometry experiments further reveal at least one mechanism by which quercetin acts. These experiments establish that quercetin has a pronounced uncoupling effect on the ethoxyresorufin reaction. The almost twofold increase in the rate of H,O, formation may arise from increased dissociation of the peroxide anion complex, although in general it is thought that HzOz arises from dissociation of superoxide from the one-electron-reduced P-450-oxygen complex, with subsequent dismutation to HzOz (32). Our experiments do not discriminate between these possibilities.

BY

355

FLAVONOIDS

oxidations.

FP,

NADPH-cytochrome

P-450

We have also considered the possibility that quercetin might be a substrate for P-450; other flavonoids such as kaempferol and galangin are metabolized by this enzyme system (unpublished results). Quercetin does exhibit a type I binding spectrum, and spectral experiments indicated that the loss of quercetin absorbance obeyed the requirements of a P-450 reaction: omission of NADPH, microsomes, or oxygen, or saturation with CO, completely inhibited the absorbance change (Fig. 4). However, we were unable to detect any metabolites of quercetin using HPLC analysis (not shown). In addition, unlike other substrates, 600 nM quercetin did not stimulate NADPH or oxygen uptake in microsomes, nor did it affect HzOz formation (Table III). We are therefore examining the alternative possibility that quercetin acts as a pseudosubstrate (32) and that oxidation of quercetin to the semiquinone or quinone could not only result in the loss of absorbance at 379 nm (quercetin) seen in the spectral experiments, but could also cause reductive destruction of the heme-oxo complex, thereby preventing substrate hydroxylation. We have not detected the quinone (X,, 510 nm) during our spectral scanning experiments, but this may be due to its instability. Attempts are currently underway to trap the quinone with various nucleophiles. This speculative redox activity of quercetin cannot account for all the inhibition

356

SOUSA

AND

observed however, since structure/activity studies showed that galangin and kaempferol, two tlavonoids lacking the catechol moiety in the B ring, are at least as potent as quercetin (Table IV). These compounds may be even more efficient uncouplers. In the stoichiometry experiments, the quercetin concentrations used are so low that spectral changes at 379 nm would be impossible to detect; however, even at higher quercetin concentrations the rate of quercetin disappearance is very slow relative to rates of NADPH and 02 consumption and therefore its loss would be negligible. Our kinetic experiments show inhibition patterns that depend on quercetin concentration. It is possible that the P-450 isozymes responsible for substrate hydroxylation have different sensitivities toward quercetin inhibition. To determine whether specific isozyme inhibition could account for the biphasic kinetics, we performed experiments with the substrate, B(a)P. Quercetin inhibited the hydroxylation of B(a)P in BNF-microsomes, but quercetin concentrations varying lOOO-fold did not alter the metabolite pattern of this substrate in our system. Fujino et al. (33) observed that monoclonal antibodies to various P-450 isoenzymes inhibited the formation of individual B(a)P metabolites to different extents, suggesting the presence of various isoenzymes in their microsomal preparation, each with different positional specificities for substrate. With quercetin as an inhibitor, we were unable to show a similar effect, suggesting that at least with this substrate, quercetin does not show isozyme-selective inhibition. This issue will have to be examined further with purified P-450 in a reconstituted system. In summary, quercetin is a potent inhibitor of P-450 reactions. Quercetin interferes with substrate binding and exerts a strong uncoupling effect. Inhibition of P-450 reduction does not occur. The biphasic kinetics underscore the complexity of this inhibition, which, in the case of the substrate B(a)P, does not appear to be isozyme-specific.

MARLETTA

ACKNOWLEDGMENTS We

thank

Alon

Coppens

for

carrying

out

the

DCPIP experiments. During the course of this work M.A.M. held the Mitsui Career Development Chair in Problems in Contemporary Technology and he gratefully Group,

acknowledges Tokyo.

the

support

of

the

Mitsui

REFERENCES 1. SOUSA,

Proc.

R. L., AND 48, 1476.

MARLETTA,

M.

Fed

A. (1984)

2. HERRMANN, K. (19’76) J. Food Z’echnd 11, 433448. 3. WIEBEL, F. J., GELBOIN, H. V., Buu-HOI, N. P., STOUT, M. G., AND BURNHAM, W. S. (1974) in Chemical DiPaolo, York. 4. HAUGEN, (1976)

Carcinogenesis J. A., eds.),

(Ts’O, pp. 249-270,

D. A., COON, J. J., AND NEBERT, J. Biol. Chem. 251, 1817-1827.

5. BUENING, M. K., CHANG, FORTNER, J. G., WOOD, A. H. (1981) 6. HUANG, M-T.,

(1980)

P. 0. P., and Dekker, New

R. L., HUANG, M-T., A. W., AND CONNEY,

Res. 41, 67-72. R. L., AND CONNEY,

Cancer CHANG,

FecL Proe.

A. H.

39,2053.

7. HUANG, M-T., CHANG, R. L., FORTNER, CONNEY, A. H. (1981) J. Biol Chem. 6836. 8. HUANG, M-T., U., KOOP,

D. W.

J. G., AND

256,6829-

JOHNSON, E. F., M~JLLER-EBERHARD, D. R., COON, M. J., AND CONNEY,

A. H. (1981) J. Biol. Chem. 256, 10897-10901. 9. LASKER, J. M., HIJANG, M-T., AND CONNEY, A. H. (1984) .I Pharm. Exp. Therap. 229, 162-170. 10. LASKER, (1982)

J. M., HUANG, M-T., Science (Washington,

AND

CONNEY,

A. H.

D. C.) 216, 1419-

1421. 11. 12.

KUHNAU, J. (1976) 117-191. KELLIS, J. T., JR.,

Science 13.

14.

16. 17.

18.

AND

(Washington,

Rev. VICKERY,

Nutr. L.

Diet. E.

24,

(1984)

D. C.) 225.1032-1034.

GUENGERICH, F. P., DANNAN, G. A., WRIGHT, S. T., MARTIN, M. V., AND KAMINSKY, L. S. (1982) Biochemistry 21,6019-6030. RYAN, D., Lu, A. Y. H., AND LEVIN, W. (1978) in Methods Packer, Press,

15.

World

in Enzymology L., eds.), Vol. New York.

52,

(Fleischer, p. 118,

S., and Academic

LOWRY, 0. H., ROSEBROUGH, N. J., FARR, A. L., AND RANDALL, R. J. (1951) J. Biol. Chem. 193, 265-275. OMURA, T., AND SATO, R. (1964) J. Bid. Chem. 239, 2370-2378. WEST, S. B., HUANG, M-T., MIWA, G. T., AND Lu, A. Y. H. (1979) Arch. Biochem. Biophys. 193, 42-50. BURKE, M. D., AND MAYER, R. T. (1974) Drug

Metub.

Disp.

2, 583-588.

P-450

INHIBITION

BF FLAVONOIDS

357

19. HRYCAY, E. G., GUSTAFSSON,J-A., INGELMAN- 25 BUENING,M. K., ANDFRANKLIN,M. R. (1974) Mol. SUNDEIERC, M., AND ERNSTER,L. (1976) Eur. PharmacoL 10, 999-1003. 26. COOPER, J. Bioehewr.61, 43-52. T. G. (1977) in The Tools of Biochemistry, p. 32, Wiley, New York. 20. RAHIMTULA,A. D., O’BRIEN, P. J., HRYCAY, 27. MISRA, H. P., AND FRIDOVICH,I. (1976) Anal. E. G., PETERSON, J. A., ANDESTABROOK, R. W. Bioch.em. 70, 632-634. (1974) B&hem. Biophys. Res. Commun 60, 28. HILDEBRANDT, A. G., ROOTS, I., TJOE, M., AND 695-702. HEINEMEYER,G. (1978) in Methods in Enzy21. ZANNONI,V. G. (1979) in Fundamentals of Drug mology (Fleischer, S.. and Packer, L., eds.), Metabolism and Drug Disposition (La Du, Vol. 52, p. 342-350, Academic Press, New B. N., Mandel, H. G., and Way. E. L., eds.), York. pp. 566-569, Robt. E. Krieger, Huntington, 29. PETERSON,J. A., EBEL, R. E., O’KEEFE, D. H., New York. MATSUBARA,T., ANDESTABROOK, R. W. (1976) 22. JEFCOATE:, C. R. (1978) in. Methods in Enzymology J. Biol. Chem 251,4010-4016. (Fleischer, S., and Packer, L., eds.), Vol. 52, 30. OPRIAN,D. D., VATSIS, K. P., AND COON, M. J. pp. 258-278, Academic Press, New York. (1979) J. Biol. Chem 254, 8895-8902. 23. STROBEL,H. W., AND DIGNAM, J. D. (1978) in 31. SEGEL,I. H. (1976) Biochemical Calculations, Ch. Methods in Enzymology (Fleischer, S., and 4, Wiley, New York. Packer, L., eds.), Vol. 52, pp. 89-96, Academic 32. KUTHAN, H., AND ULLRICH,V. (1982) Eur. J. Press, New York. B&hem. 126, 583-588. 33. FUJINO,T., PARK, S. S., WEST, D., AND GELBOIN, 24. GIGON, F’. L., GRAM, J. E., AND GILLETTE,J. R. H. V. (1982) Proc. Natl. Acad Sci. USA 79, (1968) Biochem. Biophys. Res. Cwnmun. 31, 558-562. 3682-3686.