Plant sterol biosynthesis: Novel potent and selective inhibitors of cytochrome P450-dependent obtusifoliol 14α-methyl demethylase

ARCHIVES

OF BIOCHEMISTRY

Vol. 297, No.

1,

AND

BIOPHYSICS

August 15, pp. 123-131, 1992

Plant Sterol Biosynthesis: Novel Potent and Selective Inhibitors of Cytochrome P450-Dependent Obtusifoliol 14a-Methyl Demethylase’ Florence Institut Institut

Salmon,

Maryse

Taton,

Pierre

Benveniste,

and Alain

de Biologie Mol&ulaire des Plantes, Dbpartement d %nzymologie de Botaniyue, 28 rue Goethe, 67083, Strasbourg Cbdex, France

Received February

Rahier’

Cellulaire et Moltkulaire,

CNRS- UPR 406,

3, 1992, and in revised form April 3, 1992

The I?-(-) isomer of methyl 1-(2,2-dimethylindan-lyl)imidazole-5-carboxylate (CGA 214372; 2) strongly inhibited P450-dependent obtusifoliol 14cr-demethylase (P4500BT.14DM) (Iso = 8 X lo-‘M, &,/Km = 5 X lo-‘) in a maize (Zea mays) microsomal preparation. Kinetic studies indicated uncompetitive inhibition with respect to obtusifoliol. The corresponding S-(+) isomer was a 20fold weaker inhibitor for P4500BT.14DM. The molecular features of a variety of analogues of 2 were related to their potency as inhibitors of P4500BT.14DM in vitro, allowing delineation of the key structural requirements governing inhibition of the demethylase. CGA 214372 proved to have a high degree of selectivity for P4500BT.14DM. This allowed easy distinction of this activity from other P450-dependent activities present in the maize microsomal preparation and gave strong evidence that P4500BT.14DM is a herbicidal target. Microsomal maize P4500sT.14DM and yeast P450LAN.14DM, the only known examples of P450-dependent enzymes carrying out an identical metabolic function in different eukaryotes, showed distinct inhibition patterns with CGA 214372 and ketoconazole, a substituted imidazole antimycotic. Q 1992 Academic Press, Inc.

Cytochrome P450 (P450) enzymes are membranebound hemoproteins that are implicated in the oxidative metabolism of a variety of endogenous and exogenous compounds. In plant tissues, P450-linked enzymes catalyze reactions involved in biosynthetic pathways leading to the synthesis of phytoalexins, lignin phenolics, ter1 This investigation was supported by the Ciba-Geigy Ltd. Company, Agricultural Division, Basel, Switzerland. F.S. was the recipient of a Ciba-Geigy Ltd. do&oral fellowship. * To whom correspondence should be addressed.

0003.9861/92 $5.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form

penoids, and sterols [for recent reviews see (l-3)]. They are also suspected of participating in a number of reactions associated with xenobiotic metabolism (4) in a manner similar to those of the numerous mammalian liver P45Os involved in detoxification processes (5, 6). The function and specificity of P45Os are much less well known in plants than in animals and microorganisms, probably because of the low content and supposed lability of P450 in cellfree plant systems. The plant P450-dependent obtusifoliol14a-methyl demethylase (P4500aT.1*DM)catalyzes the removal of the 14~ methyl group of obtusifoliol (12), which is metabolized to the corresponding 4a-methyl-5ol-ergosta-8,14,24(28)trien-3P-ol(13), (7). The demethylation requires NADPH and oxygen and is photoreversibly inhibited by carbon monoxide (7) (Fig. 1). In addition, obtusifoliol demethylation has been shown to occur in the endoplasmic reticulum fraction of corn (Zea muys) embryos (8). Unlike other P45Os which are involved in detoxication processes and possess a broad substrate range (5,6), P4500BT.14DM, which plays a biosynthetic role, displays a high degree of substrate and product specificity (8). A variety of important agrochemicals such as fungicides and plant growth regulators have been shown to inhibit the 14-demethylase reaction (9, lo), a key step in plant sterol biosynthesis. This property has recently led to the development of potent herbicides having P4500aT,140M as their primary site of action (11). The inhibition of plant growth by such herbicides has been ascribed to the damaging effects of excessive amounts of As-14a-methyl-sterols and (or) depletion of typical A5-sterols in the plant cell membranes. This mode of action is similar to that proposed for azole containing fungicides (11). Therefore, P4500aT.140M is of interest because of its role in plant metabolism and because it is a target for agronomically important compounds. 123

Inc. reserved.

124

ET AL.

SALMON

(12) FIG.

(14)

(13) 1.

Pathway

for the conversion

of obtusifoliol

(12) to 24-methylenelophenol

Recently, a novel class of herbicidally active imidazole carbonic acid esters has been found (12). In particular, CGA 201029 (1) (Fig. 2) has been shown recently and during the course of the present work to be a potent in uiuo inhibitor of sterol14a-demethylation in plants (13). In this paper, using a microsomal preparation isolated from corn (2. muys) seedlings, we report investigations that demonstrate first the in vitro inhibition of P4500sr.i4nM by 2 and its analogues and second the nature of structural and stereochemical features governing the extremely powerful inhibition of P4500BT,14DMby this series of novel imidazole derivatives. In addition, clear-cut

CGA 201029

(15) (15) by Zea mays microsomes.

differential inhibition by these compounds of a variety of P450-linked monooxygenase activities in this plant microsomal preparation was demonstrated. This indicates the high selectivity of these derivatives for P4500BT.14DM among the multiple distinct forms of P450 that were shown to be present in this plant material. In addition, comparative inhibition of the microsomal plant and yeast 14a-demethylases by 2 and ketoconazole 11 was studied. EXPERIMENTAL PROCEDURES Plant Materials, Reagents, Substrate, and Inhibitors Maize (variety LGll) seedlings were allowed to germinate in moist vermiculite for 60 h at 25°C. The seedlings were harvested and the

CGA 214372

CGA 214373

CGA 212193

2

3

4

6

7

6

1

5

0 7’ ix

Br

N

OMe

0 ti

0

11

10

ketoconazole

FIG. 2.

Chemical

structure

of compounds utilized

in this investigation.

Numbers are those used in the text.

INHIBITION

OF MAIZE

STEROL

embryos separated from the caryopses. Saccharomyces cereuisioe (wild type) FL 530 was a gift from Dr. J. Demontigny (Strasbourg). The following reagents and chemicals were purchased from Sigma: NADPH (tetrasodium salt), glucose-6-phosphate dehydrogenase (type XV from baker’s yeast), GSSG, Trisma base, and glucose 6-phosphate. Ohtusifoliol was synthetized as previously described (8). The chemical structures of the imidazole derivatives used in the present work are shown in Fig. 1. The identification numbers are the following: CGA 201029/R 69020 [methyl 1-(2,2-dimethylindan-l-yl)-imidazole-5 carboxylate] (1); CGA 214372 [methyl l-(2,2-dimethylindan(lR)-yl)-imidazole+carboxylate] (2); CGA 214373 [methyl l-(2,2-dimethylindan-(1s).yl)-imidazole-5-carboxylate] (3); (2,2dimethylindanl-yl)-imidazole-5-carboxylate (4); [(2,2-dimethylindan-l-yl)-imidazole] (5); [methyl 1-(2,2-dimethylindan-l-yl)-pyrrole-2-carboxylate] (6); [methyl 1-(1,2,3,4-tetrahydnaphtalhene-l-yl)-imi~zole-5-carboxylate] (7); [methyl l-(l-phenyl-l-(2’-pyridine)-methane-l-yl)-imidazole-5carboxylate] (as nitrate derivative) (8); [methyl l-(2,2-dimethyl-5-chloroindan-l-yl)-imidazole&carboxylate] (9); [methyl 1-(‘l-bromochromane-4-yl)-imidazole-5carboxylatel (as nitrate derivative) (10). These compounds were synthetized by W. Lutz, H. P. Fischer, H. R. Waespe, and H. Rempfer of Ciba-Geigy Ltd. and G. Van Lommen and V. Sipido of Janssen Pharmaceuticals.

Standard Inhibition

Assay of P4500BT,14DM

Microsomes (0.75 ml = 4 mg protein) from 2. mays were prepared and incubated aerobically as described previously (7) in the presence of obtusifoliol (12) (50-150 PM), 1 mM NADPH, 0.5 unit of glucose-6phosphate dehydrogenase, 10 mM glucose 6-phosphate, 0.1% (w/v) Tween 80, and inhibitors (0.001-100 FM) as indicated in tables and figures. Incubations were performed at 30°C for 4 h. The reaction mixture contained 1 mM cyanide to prevent 4a-demethylation of obtusifoliol and its metabolites (39). The extraction, purification, and GLC analysis of the mixture of substrate and demethylation products were performed as previously described (7, 8), allowing the conversion ratio to be determined and corrected from the value obtained in the boiled assay. The reaction rate was deduced from this conversion ratio and the amount of substrate added. In the case of inhibition assays, microsomes (0.75 ml) were incubated as previously described with the addition of increasing concentrations of the inhibitor (six to eight different concentrations). In order to compare the Is, values obtained with different microsomal preparations, a control, for which the Z5,,value of compound 1 was measured, was used as standard since the &,avalues measured varied slightly with the concentration of the microsomal preparation used (4.0-6.0 mg * ml-‘). The Iso values obtained were normalized to that obtained for the standard compound (1), which permitted a more accurate comparison of the relative affinities for compounds that could not be assayed in the same experiment or in the case of repeated experiments.

Enzyme Assays Yeast cytochrome P450-dependent lanosterol14wmethyl-demethylaase (40). A loo-ml sample from a 24-h culture of S. cereuisiae (wild type) was grown semiaerobically in Y.P.G. medium (1% yeast extract, 1% bactopeptone, 2% glucose) at 30°C in the dark; thereafter the yeast was transferred into 2 liters of the same freshly prepared medium and allowed to grow for 12 h under the same conditions. They were harvested by centrifugation at 4000g for 10 min. Cells were washed with 0.1 M TrisHCl buffer, pH 7.5, containing 1 mM dithiothreitol, 1 mM PMSF: and 0.5 M saccharose (20 ml/100 ml starting culture), harvested by centrif-

3 Abbreviations used: PMSF, phenylmethylsulfonyl fluoride; DMSO, dimethyl sulfoxide; BSA, bovine serum albumin; CA4H, cinnamic acid 4.hydroxylase; ECOD, 7-ethoxycoumarin-0-deethylase; LAH, lauric acid hydroxylase.

125

14-DEMETHYLASE

ugation, and washed again with the same buffer (1 ml/100 ml starting culture). After centrifugation (SOOOg,10 min), cells were disrupted by glass bead homogenization (0.25 mm diameter) for a total of 15 min (cooled in ice) (2 min disruption at I-min cooling intervals) in a glass tube (1 ml of beads for 100 ml starting culture) using a Vortex homogenizer. Cell debris, mitochondria, and nuclei were removed by centrifugation (13,000g for 15 min). Microsomes were isolated by centrifugation at 140,OOOgfor 120 min of the 13,000g supernatant and resuspended with an Elvejehm potter in 0.1 M phosphate buffer, pH 7.5, containing 1 mM reduced glutathione, 1 mM EDTA, 4 mM MgCl,, and 20% (v/v) glycerol. Microsomes (1 ml = 2-4 mg protein) were incubated aerobically in the presence of exogenous lanosterol (20 PM), 1 mM NADPH, 0.5 unit of glucose-6-phosphate dehydrogenase, 10 mM glucose 6-phosphate, 1 mM cyanide, 0.1% (w/v) Tween 80, and inhibitor. Incubation, extraction, purification, and GLC analysis of the mixture of substrate and demethylation products were performed as previously described for P4500sr.i40M with the exception that it was performed on the 4,4-dimethyl-sterol fraction instead of the 4a-methyl sterol fraction. The conversion ratio and the reaction rate were determined in a manner similar to that described above for plant P4500sr.14nM. The Iso value were also measured in the way described for inhibition of the plant demethylase. 7-Ethoxycoumarin-0-deethylase (14). Maize microsomes (20-200 ~1; buffer, pH 8.6, to 0.1-l mg protein) were diluted with 0.05 M Tris-HCl a final volume of 2 ml. The standard assay contained 250 pM 7-ethoxycoumarin (stock solution in dimethyl sulfoxide), 1 mM glucose 6phosphate, 1 unit of glucose&phosphate dehydrogenase, and 50 pM NADPH. Incubations were performed in a thermostated cuvette at 30°C after preequilibration for 2 min before starting the assay by addition of NADPH. Fluorescence changes at 460 nm were directly recorded as a function of time, using an excitation wavelength of 380 nm, with a Shimadzu RF 5000 spectrofluorometer. The assay was calibrated by addition of known amounts of 7-hydroxycoumarin to the incubation medium. Under these conditions, rates of 7-hydroxycoumarin formation were linear for 10 min. Cinnamic acid 4-hydroxylose (27). Microsomes (200 ~1, 1 mg protein) resuspended in 0.1 M phosphate buffer, pH 7.4, were incubated 10 min at 25°C in the presence of trans-cinnamic acid (300 PM), trans-[3“C]cinnamic acid (0.25 PCi, 58 &i/pmol), 100 PM NADPH, 2 mM glucose 6-phosphate and 1 unit of glucose-6-phosphate dehydrogenase. Inhibitor, when present, was added in DMSO. The reaction was stopped by addition of 20 ~1 4 M HCl. Carrier cinnamic acid and p-coumaric acid (0.1 mg) were added to the incubaticn mixture after the end of the reaction. An aliquot of this mixture (100 ~1) was analyzed by TLC on silica gel with toluene/acetic acid (6/7, v/v) saturated with H,O as migrating solvent. The amount of radioactivity associated with residual cinnamic acid (R, = 0.85) and produced p-coumaric acid (R, = 0.75) was determined by mean of a Berthold LB 520 TLC radioscanner. Laurie acid hydroxylase [adapted from (28)]. Microsomes (100 ~1; 500 pg protein) were incubated with 50 @M sodium laurate, 3 ).LM [l“C I lauric acid (28.8 Ci/mol), 6.7 mM glucose 6-phosphate, 1U glucose6-phosphate dehydrogenase, and 1 mM NADPH in a 20 mM phosphate buffer (final volume 200 ~1). Incubations were started by addition of NADPH, run for 30 min at 25°C and stopped by addition of 200 pl acetonitrile containing 0.2% acetic acid. An aliquot (100 ~1) was then analyzed by TLC on silica gel plates with diethylether/hexane/formic acid (70/30/l, v/v) as migrating solvent. The amount of radioactivity associated with residual lauric acid (RI = 0.78) and produced 9-hydroxy lauric acid (R, = 0.52) was determined by means of a Berthold LB 520 TLC radioscanner.

In Vitro Binding

of Azole Inhibitors

to Microsomal

P450

Binding spectra with P450 were determined by differential spectroscopy using a Shimadzu MPS-2000 spectrophotometer according to the method described by Jefcoate (21). Diluted microsomes (1 ml = 2

126

SALMON

mg proteins) were placed in both sample and reference cuvette (1 cm path length) of the spectrophotometer using a slitwidth of 1 nm, and the baseline was recorded. Tested compounds dissolved in dimethyl sulfoxide (1 mg/ml) were directly added to the sample cuvette, an identical volume of dimethyl sulfoxide being added in the reference cuvette. The contents were mixed and the spectrum between 380 and 500 nm was recorded after 1 min. The absorbance change between the type II peak (425-435 nm) and the corresponding through (400-410 nm) was related to the concentration of azole to permit the determination of & values.

Measurement of P450 Content and Other Analytical Procedures P450 content in microsomes was spectrophotometrically determined from its dithionite (Na&04) reduced CO difference spectrum using extinction coefficient difference between 450 and 490 nm of 91 mM-’ * cm- i (41). Membrane protein was determined as described by Schacterle and Pollack (42), with BSA as a standard (fraction V, Sigma Chemical Co). GLC was carried out on a capillary column [WCOT; 25 m long X 0.25 mm internal diameter) coated with OV17 (H, flow 2 ml/ min), with a GLC instrument equipped with a flame-ionization detector. The temperature program used included a fast rise from 60 to 24O’C (30”C/min), then a slow rise from 240 to 280°C (Z”C/min).

ET AL.

mg protein corresponding to a total P450 concentration of approximatively 0.6 @M was determined in our microsomal preparation. Therefore, the complete inhibition of P45Oosr.i4nM activity, achieved with a 0.1 PM concentration of 2, would suggest binding interactions in the range of stoichiometric concentrations of enzyme and compound 2. The exceptional avidity of P45Oosr,i4nM for compound 2 could have the attractive potential of achieving virtually irreversible inhibition (tight binding) of this enzyme without the need of forming a covalent bond between the inhibitor and the enzyme, an advantage for agrochemical applications. Kinetics of Demethylation

Inhibition

The effect of varying concentrations of compound 2 on obtusifoliol demethylation is shown in Fig. 3. Lineweaver-

r--* -1

l/V

nmol;‘min

I mg protein

RESULTS AND DISCUSSION Potent Inhibition of Maize Microsomal P4500BT.14DMby methyl l-(2,2-dimethylindan-(1 R)-yl)-imidazole-5carboxylate (CGA 214372; 2) The corn seedling microsomal preparation used in this study was incubated with obtusifoliol(l2) in the presence of a range of concentrations of compound 2, the enantiomerit pure R form of CGA 201029 (l), and the resulting demethylation products, 4a-methyl-5a-methyl-ergosta8,24-dien-3fl-ol (14) and 24-methylene-lophenol (15), were quantified by GLC after sterol extraction and separation (7,8) (see Experimental Procedures). A dose-response curve was then obtained and the ISo value (concentration producing 50% inhibition) was determined. The results showed that 2 was an extremely potent inhibitor of P4500er.i4n~ with an I60 value of 8 X lo-’ M, the most potent inhibitor of the plant demethylase described so far. Moreover, the relative affinity of compound 2 when compared to that of the physiological substrate (12), estimated by the ratio lJKm = 5 X 10e5, is one of the highest among the compounds known to inhibit P450dependent reactions. Information on the in vitro effects of azole derivatives on other plant P450-dependent reactions is relatively scarce. However, literature data, when available, indicate that concentrations of the best known inhibitors needed to block these reactions are indeed significantly higher; for instance, ethoxyresorufin-O-deethylase and Tetcyclacis (& = 7.5 X lOWEM, 15,,/Km = 0.2) (14), (ZS, 3S)-Paclobutrazol and ent-kaurene oxidase (Is,, = 2.3 X lO-sM, &JKm = 5 X 10m2) (15), ketoconazole and (I,, = 7 X 1O-8 M, &c/Km = 10-2) (16yeast P450LAN.14DM 18). An average total P450 content of 0.12 + 0.04 nmol/

lO’/S

l/V

nmol-‘.

min x mg

PM-’

protein

B

S= 50pM

180.

*’

120.

0

10

50

111 nM FIG. 3. Kinetics of the inhibition compound 2. Obtusifoliol was used as plot. The concentrations of inhibitor nM; *, 5 nM; 0, 10 nM; v, 20 nM. (B)

of microsomal P4500Br.iln~ by a substrate. (A) Lineweaver-Burk used were: 0, no inhibitor; Zr, 1 Dixon plot.

INHIBITION TABLE

OF MAIZE

I

Inhibition of Microsomal P4500m.14~~ and Binding to the Total Microsomal Fraction by CGA 214372 (2) and Analogues

Comparative

Compound 1 2 3 4 5 6 7 8 9 10

ho (FM) 0.030 0.008 0.170 1.50 0.155

f 0.005 f 0.002 2 0.030 f 0.30 + 0.030 n.i. 0.033 f 0.005 0.043 * 0.005 0.075 * 0.005 0.062 f 0.005

Kd

(PM)

10 -cl 6.5 f 1 12.5 3~0.5 n.b. 5.5 f 1 n.b. 8.5 f 1 24.5 _t 1 n.d. n.d.

AAma1390-425 per milligram protein (lo3 X ADO X mg prott’) 1.25 1.21 0.91 n.b. 5.32 n.b. 2.13 0.72

Note. Results are means of two to five experiments. n.i., not inhibitory, i.e., <5% inhibition at 100 pM; n.b., not bound; n.d., not determined.

Burk kinetics were found with linear l/V against 1/S plots. In the presence of different concentrations of 2, the best fit was obtained with parallel straight-line plots (Fig. 3A). In addition, the best fit for plotting inhibition kinetics according to the method of Dixon (19) was also obtained with parallel straight-line plots (Fig. 3B). The plots obtained with both of these types of representation were apparently characteristic of uncompetitive inhibition. These results mean. first, that compound 2 is not competing for the obtusifoliol (12) binding site and secondly that substrate (12) and inhibitor (2) combine with different enzyme states. If purely uncompetitive, 2 would combine only with the enzyme-substrate complex and not with the free enzyme (20). In the case of P45Oosr.i4nM, this inhibition pattern could reflect binding of compound 2 exclusively to the particular spin state of the cytochrome induced upon binding of the substrate (21). Structural Requirements for Inhibition or Binding to Microsomal P450

STEROL

127

14-DEMETHYLASE

the imidazole ring. As indicated in Table I, P45Oosr,i4nM was not inhibited by 6. Inhibition of P450-dependent reaction by azole derivatives has been generally ascribed to the formation of a coordinate bond between the nonbonded electron pair on the N3 of the azole ring with the ferric heme iron (16, 22). Although it does not prove it, this result is in accordance with such a manner of inhibition of P45Oosr.i4nM by 2. Moreover, the absence of induced type II spectral change of the total P450 fraction present in the microsomes by compound 6, whereas 2 led to a clearly marked type II difference spectrum (Fig. 4), indicated that 2 is binding to one or more P450 species present in the microsomes in this way. As 2 possesses an asymmetric carbon, it was of interest to examine the inhibitory effects of the R, S form, CGA 201029 (l), and its separate enantiomers R(-), compound 2, and S(+), compound 3. By far, the strongest inhibition of P4500BT.14nM was exerted by the R enantiomer 2 and the weakest by the S enantiomer 3, with the racemate 1 taking an intermediate position. This data thus revealed a remarkable enantiomeric selectivity of the R enantiomer with an enantiomeric (R/S) activity ratio >20. Because of their potential capacity to inhibit both P45Oosr.i4nM and other P450-dependent reactions in plants, chirality within this series of compounds is of paramount importance. Indeed, various diastereoisomers of paclobutrazol and triadimenol, two widely used agrochemical triazole derivatives, have recently been studied with respect to their relative inhibitions of the plant gibberellin pathway versus the fungal sterol pathway P450-dependent processes, and respective plant growth regulating and antifungal activities. Results indicated that the configuration about the chiral carbon bearing a hydroxyl group determined which pathway was inhibited (15). However, great

l/‘hA

of P4500BT.14DM

With the aim of providing insight into the mechanism of inhibition of P4500sr.i4nM by compound 2, we compared the inhibition of microsomal P45Oosr.i4oM by a series of 10 analogues differing in a single structural or stereochemical feature. Table I summarizes the I60 with compounds in which (a) the N3 atom of the imidazole ring is replaced by a carbon atom, (b) the absolute configuration of the Cl carbon atom of the dimethylindan moiety is opposite, (c) the 5carboxylic ester is altered, (d) or the dimethylindan hydrophobic structure is modified. We first examined the inhibition of P4500Br,i4nM by compound 6, which lacks the presence of the N3 atom of

0.05 1/[1]

0.1 PM-'

FIG. 4. Spectral titration of total maize microsomal P450 with compound 2 (O), compound 3 (O), compound 1 (A), and ketoconazole @I. The inset shows the difference spectra obtained following addition of 50 pM CGA 214372 (2). AA represents the absorbance difference between 425 and 390 nm.

128

SALMON

care should be taken in extrapolating in vitro activities to physiological effects and agrochemical applications, since enantioselectivity might also occur during uptake and transport for instance. We next examined the inhibitory properties of a series of analogs of 2 with structural modifications of the hydrophobic N-substituent. I50 values from Table I show that narrow variations of the dimethylindan hydrophobic moiety were well accommodated (compounds 1,7,9,10) by P4500BT.14D~and that additional bulk (e.g., compound 8) appears to be tolerated. In addition the planar sixcarbon aromatic ring possessed by all these derivatives may also play an important role for the binding of this hydrophobic moiety. Indeed, it has been recently shown in the case of yeast P450LAN.14DMthat a discrete binding site that can accommodate planar aromatic ligands exists for this enzyme (23). Since 2 did not show competitive kinetics, a hydrophobic pocket accommodating this N-substituent moiety distinct from that binding the sterol backbone has to be postulated. This is unlike that shown in a number of models developed for the inhibition of fungal P450LAN.14DM by azoles, where the hydrophobic part of the azole inhibitor is postulated to overlap the sterol binding site (15, 24). This last proposal should be made with caution when it does not rely upon kinetic evidence. Indeed, in the rare examples, to our knowledge, where the kinetics of in vitro inhibition of 14-demethylation in fungi or rat liver by other azole derivatives were studied, noncompetitive (mixed) inhibition with respect to lanosterol was indicated (25, 26). Another structure-activity pattern apparent in Table I is that the presence of the carboxyl methyl ester is of importance for inhibition. Indeed, elimination of this substituent (compound 5) led to a 5-fold poorer inhibitor. Moreover, the free carboxylic acid derivative (4) was 50-fold less inhibitory, which suggests that the ester substituent is directly involved in the inhibition, probably by interacting with a nonbasic area of the binding site. However, the poor affinity of 4 could also result from a much smaller partition of this derivative into the membrane phase owing to its acidic function. In conclusion, based on the compounds evaluated in this study, compound 2 must be anchored to P4500nr.i4nM by two functional groups: (i) the netu N atom of the imidazole ring and (ii) the carboxylic ester function (probably through the carbonyl moiety). The rigid hydrophobic and relatively planar, dimethylindan part of the inhibitor tolerated narrow variations but had to be in optimal orientation with the imidazole part of the molecule to provide the maximal additional hydrophobic binding. Moreover, this strict positioning could be of importance for selective inhibition of P4500BT,14DM.

ET AL.

Interaction of 2 and Analogs with Total P450 Contained in Maize Microsomes The addition of compound 2 to maize microsomes containing P45O (0.12 nmol/mg protein) led to a clearly marked Type II difference spectra (Fig. 4) expected for binding of the sp2 N3 atom of the inhibitor to the ferric heme iron of one or more P450 species present in the preparation (16,22). From the measurement of the change in absorbance difference between 409 and 425 nm in the difference spectra with a range of inhibitor concentrations, Lineweaver-Burk double-reciprocal plots were obtained, which allowed apparent Kd values to be determined. A value of Kd = 6.5 + 0.5 X lop6 M was found, which is about 3 orders of magnitude higher than the 15,,value obtained for inhibition of the microsomal demethylase activity. The binding affinities of the above-mentioned series of analogues of 2 to the total P450 fraction present in the microsomes were also investigated and the corresponding Kd values determined (Table I). When the spectral binding and demethylase inhibition are viewed together, the data show that the obtained Kd values did not correlate with the corresponding I50 values. A possible reason for this discrepancy between kinetics and spectroscopic studies is the presence of other constituents in the microsomal system used, which could mask the specific spectral differences induced by the binding of 2 to P4500Br,14DM. In that case the results would suggest that P4500BT.14DMis a minor proportion of the total amount of microsomal P450 isozymes. However, one cannot exclude a different mode of binding of 2 and analogues to P4500BT.14DMthan to other P450 species present in the microsomes, for which binding of 2 induced measurable type II difference spectra. Whatever the reason for this discrepancy, these results show that spectroscopic analyses performed with maize microsomes are not suitable for the determination of the specific inhibition of P4500Br,14nM by azoles. Selectivity

of Action of 2 in Maize

Having shown that P4500BT.14DMwas strongly inhibited by 2, the degree of selectivity of this inhibitor was investigated by comparing P4500BT.14DMinhibition with that of other cytochrome P450-dependent monooxygenases whose activities could be measured in our microsomal preparation. Thus, compound 2 and its enantiomer 3 were assayed in parallel with three other P450-dependent monooxygenases present in the same microsomal preparation, cinnamic acid 4-hydroxylase (CA4H) (27), 7ethoxycoumarin-0-deethylase (ECOD) (14), and lauric acid hydroxylase (LAH) (28). Since the three enzymatic activities mentioned here have been measured in vitro in the same maize microsomal preparation, it is relevant to with those for compare the Iso values obtained P4500BT.14DM.

INHIBITION

OF MAIZE

STEROL

TABLE Comparative

Inhibition

II

by CGA 214372 (2) and CGA 214373 in the Same Maize Microsomal

activity

Obtusifoliol 14wdemetbylase Cinnamic acid 4-hydroxylase 7 Ethoxycoumarine-0-deethylase Laurie acid hydroxylase Binding constant of (2) to the total P450 fraction

0.170 ?I 0.030

>lOO 95 f 5 >lOO

>lOO

(PM)

>lOO

160 2 161

,100

l-5

present in the microsomes, Kd = 6.5 + 0.5

Sterol 14a-demethylaseis the only known example of a P450-dependent reaction which occurs widely in eukaryotes and which has a unique metabolic role. We recently clearly demonstrated that plant P4500BT.i40M has a high degree of substrate specificity which is different from that of the yeast and mammalian enzymes, indicating clearly that the plant demethylase is a distinct enzyme

(8). Thus we were also interested in comparing the in vitro inhibition of these two enzymes by compound 2, and by another substituted imidazole, widely used clinically as antimycotic ketoconazole (11). This compound has been shown particularly to be a potent inhibitor of P450-dependent lanosterol demethylase (P450LAN.i40M) in yeast and fungal microsomes (16). Results from Table III indicate that yeast and plant demethylases are both highly sensitive to these two imidazole inhibitors. However, maize P4500B~.14DMappeared to be 30-fold more susceptible to inhibition by compound 2 than yeast P450LAN.i4nM. In contrast, yeast P450LAN.i4nM was 2.5-fold more strongly inhibited by ketoconazole (11) than maize P4500BTXDM

*

These data could reflect differential adsorption of the inhibitor by the protein or other constituents in the two distinct and anisotropic microsomal preparations used. However, one cannot exclude the possibility that this important discrepancy would be inherent to sensitivity differences between yeast and plant 14-demethylases toward 2 and 11. To give support to one of these hypotheses, purification of yeast and plant 14-demethylases would be necessary, a task which is currently investigated in this laboratory in the case of the plant enzyme.

TABLE of Plant

Reactions

K,,, for substrate

0.008 + 0.002

Comparative Inhibition of Plant and Yeast Microsomal 14a-Methyl Demethylases by 2 and Ketoconazole 11

Inhibition

P450-Dependent

zm for (3) (NM)

(2)

From results presented in Table II, it appears clear that compounds 2 and 3 have a high degree of selectivity for P45OosT,i.+oM. Indeed the affinity of 2 for this enzyme is at least lO,OOO-fold higher than it is for CA4H, ECOD, or LAH. These results clearly show that at least two independent cytochrome P450 isoforms are involved in the reactions carried out respectively by P45OosT.i4nM and the three enzymes CA4H, ECOD, and LAH. Since CA4H and LAH have been shown recently to be distinct P450 isoforms (29), there is no doubt that P4500BT.14DM, CA4H, and LAH are catalyzed by distinct isoforms of P450 in our microsomal preparation. Moreover, the data demonstrate the narrow spectrum of susceptible isozymes to the class of P450 inhibitors studied herein.

Comparative

(3) of Distinct Preparation

(PM)

150 for

Enzymatic

129

14-DEMETHYLASE

III

and Yeast Microsomal 14a-Methyl Demethylases and the Fungicide Ketoconazole (11)

by the Herbicide

CGA 214372

(2)

K,,, apparent Enzyme

&II (PM)

(PM)

~m/Km

Kdn (PM)

Maize P4500ar.r4nM Yeast P450un.r4o~ Maize P4500sr.mm

0.008 f 0.002 0.30 k 0.05

160 6b

5 x w5 5 x w2

10

0.16

160

Drug CGA 214372 (2) Ketoconazole

(11)

Yeast

P450LAN.14DM

a Measured with the total microsomal *Taken from Ref. (43). ’ Taken from Ref. (26). d Not determined.

P450 fraction.

& 0.02

0.07'

6*

1 x w3 1.2 x 1o-3

6.5 k 0.5 f 0.5

10 f 0.5 n.d.d

130

SALMON

In conclusion, of wider importance in this paper is, firstly, the in vitro demonstration of the feasibility of inhibiting selectively and strongly a plant P450 with an imidazole derivative, and, second, that in the particular this selective inhibition is connected W& Of P4500BT.14DM9 with herbicidal properties. Indeed, the very low doses of 2 (or 1) required to inhibit P4500er,14nM both in vitro (this work) and in uivo (13) together with the high selectivity of 2 for P4500sr.1JnM strongly suggest that is the primary target of 1 and 2. Our data P4500BT.l4DM suggest that this inhibition, leading to the replacement of typical A5-sterols by As-14a-methyl-sterols, is probably responsible for the observed phytotoxic effects. In this respect, we have observed a very good correlation between the 150values measured in vitro for compounds l-4 and both the replacement of A5-sterols by As-14a-methyl-sterols and the growth inhibition of maize treated with an identical concentration of these four derivatives (unpublished results). Moreover, taken together, our results concerning the structural requirements for tight binding to P4500sr.14nM among this series of imidazole derivatives correspond well with the putative herbicidal pharmacophore substructure proposed, based on a set of analogues of compound 2 (12). Once a new herbicide is discovered and its target enzyme has been identified, this information can be used to aid the rational development of herbicide-tolerant varieties of crop plants. From this point of view, recent work from our laboratory showed the feasibility of selecting, following uv mutagenesis of tobacco protoplasts, calli resistant to a triazole experimental fungicide which was shown to inhibit plant P4500Br.14nM (3, 35). In the case of one resistant callus, it has been suggested that an increase of sterol synthesis of mutational origin would be responsible for triazole resistance. Finally, the question is raised of the structure of the substrate and inhibitor binding domains of P4500er,14nM from different organisms, namely, plants, fungi, and animals. To discover this, the knowledge of the primary sequence of these domains is required. Strong variation of substrate and inhibitor specificity may result from limited amino acid replacements insofar as they are located in the immediate vicinity of the binding site. Indeed, recent work has shown the feasibility of altering P450 substrate specificity, regioselectivity, or inhibitor susceptibility via mutation of a single or a few amino acid residues (30,32). an altered form (P450sol) of yeast Furthermore, has been recently isolated (33). In particular, P450LAN.14DM P450s01 showed a single amino acid substitution (Gly 310 replaced by Asp) when compared to P450~N~4nM. Interestingly, among other consequences, this alteration led to a fundamentally different mode of interaction with diniconazole, an azole inhibitor of P450LAN,14nM (33). Along these lines, it has been recently shown that P450dependent 14a-demethylases from different species of

ET AL.

yeast were immunologically different, although they may have had a few common antigenic sites (34). So far, the lanosterol-14cu-demethylase genes from S. cerevisiae and from Candida tropicalis have been cloned and sequenced (33, 36). The coding region of these genes contains sequences highly preserved in many other P450 (37) and in particular in the bacterial P450-dependent d-camphor oxydase whose X-ray crystal structure is known (38). The success in purifying plant P4500nr.14nM and the cloning of its gene could then allow more detailed information about the molecular basis for substrate and inhibitor recognition of this unique and highly selective monooxygenase. ACKNOWLEDGMENT We are indebted to B. Bastian for kindly

typing the manuscript.

REFERENCES 1. Durst, F. (1991) in Microbial and Plant Cytochrome P-450: Biochemical Characteristics, Genetic Engineering and Practical Implications (Ruckpaul, K., and Rein, H., Eds.), pp. 191-232, Akademie Verlag, Berlin. 2. Donalson, R. P., and Luster, D. G. (1991) Plant Physiol. 96, 669674. 3. Mihaliak, C., Karp, F., and Croteau, R. (1991) in Methods in Plant Biochemistry (Lea, P. I., Ed.), Vol. 9, Academic Press, London. 4. Frear, D. S., Swanson, H. R., and Tanaka, F. S. (1969) Phytochemistry 8,2157-2169. 5. Lu, A. Y. H., and West, S. B. (1978) Pharmacol. Ther 2, 337-358. 6. Guengerich, F. P. (1991) J. Biol. Chem. 266(16), 10019-10023. 7. Rahier, A., and Taton, M. (1986) Biochem. Biophys. Res. Commun.

140,1064-1072. 8. Taton, M., and Rahier, A. (1991) Biochem. J. 277, 483-492. 9. Taton, M., Ullmann, P., Benveniste P., and Rahier, A. (1988) Pestic. Biochem. Physiol. 30, 178-189. 10. Haughan, P. A., Lenton, J. R., and Goad, L. J. (1988) Phytochemistry 27(8), 2491-2500. 11. Burden, R. S., James, C. S., Cooke, D. T., and Anderson, N. H. (1987) Proc. 1987 Br. Crop Protection Conf. Weeds 3B-4,171-178. 12. Van Lommen, G., Lutz, W., Sipido, V., Huxley, P., Van Gestel, J., and Fischer, H. P. (1990) ZUPAC Meet. 169, Abstract OlD-03. 13. Streit, L., Moreau, M., Gaudin, J., Ebert, E., and Vanden Bossche, H., (1991) Pestic. Biochem. Physiol. 40, 162-168. 14. Werck-Reichhart, D., Gabriac, B., Teutsch, H., and Durst, F. (1990) Biochem. J. 2’70, 729-735. 15. Burden, R. S., Carter, G. A., Clark, T., Cooke, D. T., Croker, S. J., Deas, A. H. B., Hedden, P., James, C. S., and Lenton, J. R. (1987) Pestic. Sci. 21, 253-267. 16. Vanden Bossche, H., Marichal, P., Gorrens, J., Bellens, D., Verkooven, H., Coene, M. L., Lauwers, W., and Janssen, P. A. (1987) Pestic. Sci. 21, 289-306. 17. Aoyama, Y., Yoshida, Y., Sonoda, Y., and Sato, Y. (1991) Biochim. Biophys. Acta 1081, 262-266. 18. Yoshida, Y., and Aoyama, Y. (1987) Biochem. Pharmacol. 36(2), 229-235. 19. Dixon, M. (1953) Biochem. J. 55, 170. 20. Ault, A. (1974) J. Chem. Educ. 51,381-386. 21. Jefcoate, C. R. (1978) in Methods in Enzymology (Fleischer, S., and Packer, L., Eds.), Vol. 52, p. 258, Academic Press, New York.

INHIBITION

OF MAIZE

22. Schenkman, J. B., Remmer, H., and Estabrook, R. W. (1967) Mol. Phurmacol. 3, 113-123. 23. Wright, G. D., Parent, T., and Honek, J. F. (1990) Biochim. Biophys. Actu 1040,95-101. 24. Keller, W. (1987) Pestic. Sci. 18, 129-147. 25. Barrett-Bee, K., Lees, J., Pinder, P., Campbell, J., and Newboult, L. (1988) Ann. N. Y. Acad. Sci. 544, 231-244. 26. Lavrijsen, K., Van Houdt, J., Thijs, D., Meuldermans, W., and Heykants, J. (1987) Xenobiotica 17(l), 45-57. 27. Russell, D. W. (1971) J. Biol. Chem. 246(12), 3870-3878. 28. Salaun, J. P., Benveniste, I., Reichhart, D., and Durst, F. (1978) Eur. J. Biochem. 90, 155-159. D., Teutsch, H., andDurst, F. (1991) 29. Gabriac, B., Werck-Reichhart, Arch. Biochem. Biophys. 288(l), 302-309. 30. Atkins, W. M., and Sligar, S. G. (1989) J. Am. Chem. Sot. 111, 2715-2717. 31. Wells, J. A., Cunningham, B. C., Graycar, T. P., and Estell, D. A. (1987) Proc. N&l. Acud. Sci. USA 84, 5167-5171. 32. Lindberg, R. L. P., and Negishi, M. (1989) Nature (London) 339, 632-634. 33. Ishida, N., Aoyama, Y., Hatanaka, R., Oyama, Y., Imajo, S., Ishiguro, M., Oshima, T., Nakazato, H., Noguchi, T., Maitra, U. S., Mohan,

STEROL

131

14-DEMETHYLASE

V. P., Sprinson, D. B., and Yoshida, Y. (1988) Biochem. Biophys. Res. Commun. 155(l), 317-323. 34. Kaergel, E., Aoyama, Y., Schunk, W.-H., Mueller, ida, Y. (1990) Yeast 6,61-67. 35. Maillot-Vernier, P., Schaller, H., Benveniste, (1990) Plant Physiol. 93, 1190-1195.

H. G., and Yosh-

P., and Belliard,

G.

36. Chen, C., Kalb, V. F., Turi, T. G., and Loper, J. C. (1988) DNA 7, 617-626. 37. Kalb, V. F., and Loper, J. C. (1988) Proc. Natl. Acud. Sci. USA 85, 7221-7225. 38. Poulos, T. L., Perez, M., and Wagner, G. C. (1982) J. Biol. Chem. 257,10427-10429. 39. Pascal, S., Taton, M., and Rahier, A. (1990) Biochem. Biophys. Res. Commun. 172(l), 98-106. 40. Aoyama, Y., and Yoshida, Y. (1978) Biochem. Biophys. Res. Commun. 85(l), 28-34. 41. Omura, T., and Sato, R. (1964) J. Biol. Chem. 239, 2370-2378. 42. Schacterle, G. R., and Pollack, R. L. (1973) Anal. Biochem. 51,654655. 43. Aoyama, Y., Yoshida, Y., Sonoda, Y., and Sato, Y. (1987) J. Biol. Chem. 262, 1239-1243.