(Received for publication, May 25, 1996, and in revised form, October 9, 1996)
From the Institut de Biologie Moléculaire des
Plantes-CNRS, Département d'Enzymologie Cellulaire et
Moléculaire, 28 rue Goethe, F-67083 Strasbourg Cedex, France,
¶ CNRS-URA 1386 Faculté de Pharmacie, Laboratoire de
Synthèse Bio-Organique, 74 route du Rhin,
F-67048 Strasbourg Cedex, France, ** Department of Pharmacology,
School of Medicine, Oregon Health Sciences University,
Portland, Oregon 97201-3098, and
Commissariat à l'Energie Atomique,
Service des Molécules Marquées,
F-91191 Gif-sur-Yvette Cedex, France
Incubation of Vicia sativa
microsomes, containing cytochrome P450-dependent
lauric acid -hydroxylase (
-LAH), with
[1-14C]11-dodecynoic acid (11-DDYA) generates a major
metabolite characterized as 1,12-dodecandioic acid. In addition to
time- and concentration-dependent inactivation of lauric acid and
11-DDYA oxidation, irreversible binding of 11-DDYA (200 pmol of 11-DDYA
bound/mg of microsomal protein) at a saturating concentration of
11-DDYA was observed. SDS-polyacrylamide gel electrophoresis analysis
showed that 30% of the label was associated with several protein bands
of about 53 kDa. The presence of
-mercaptoethanol in the incubate
reduces 1,12-dodecandioic acid formation and leads to a polar
metabolite resulting from the interaction of oxidized 11-DDYA with the
nucleophile. Although the alkylation of proteins was reduced, the
lauric acid
-hydroxylase activity was not restored, suggesting an
active site-directed inactivation mechanism. Similar results were
obtained when reconstituted mixtures of cytochrome P450 from family
CYP4A from rabbit liver were incubated with 11-DDYA. In contrast, both 11- and 10-DDYA resulted in covalent labeling of the cytochrome P450
2B4 protein and irreversible inhibition of activity. These results
demonstrate that acetylenic analogues of substrate are efficient
mechanism-based inhibitors and that a correlation between the position
of the acetylenic bond in the inhibitor and the regiochemistry of
cytochromes P450 oxygenation is essential for enzyme inactivation.
The biological roles of plant P450 fatty acid hydroxylases are
poorly understood. Evidence suggests that long-chain fatty acid
-hydroxylases such as the oleic acid
-hydroxylase (1, 2) may play
a role in plant cuticle biosynthesis. The terminal hydroxyl group
appears to be required for the polymerization of cutin monomers. In
view of the differential inhibition of
-LAH1 and oleic acid
-hydroxylase
activities from Vicia sativa by mechanism-based inhibitors
and aliphatic acyltriazole derivatives, it was suggested that both
reactions are catalyzed by different forms of P450 in V. sativa seedlings (2).
The enzymatic and structural characterization of plant P450s has been
hampered by the low specific contents of these enzymes and by the
difficulties encountered during their solubilization and purification
(3, 4). Unfortunately, our efforts to purify fatty acid hydroxylases
were unsuccessful, and only partly purified preparations2 exhibiting -LAH activity
after reconstitution have been obtained.
A remarkable property of the plant fatty acid hydroxylases is their inducibility by clofibrate (1, 5), a hypolipidemic drug that induces peroxisomal proliferation in mammals (6, 7) and plants (8).
Although several similarities exist between -hydroxylases from
plants and mammals (i.e. selective induction by arylphenoxy compounds and phthalate esters, similar substrate specificity, and
inactivation by terminal acetylene), antibodies raised against the rat
liver lauric acid
-hydroxylase (CYP4A1) neither inhibit
-LAH
activity from plants nor interact in Western blotting with microsomal
proteins from V. sativa.3
Mechanism-based inhibitors containing a terminal acetylene were
previously shown to be potent irreversible inhibitors of both plant (9,
10) and mammalian (11, 12) fatty acid -hydroxylases. Examples of
in vitro and in vivo inactivation of mammalian
fatty acid
-hydroxylases by these compounds have been published
(13, 14, 15). Inhibition of enzyme activity was usually associated with the
inhibition-dependent alkylation of the enzyme heme group and the apoprotein. Thus one goal of the current work was to develop compounds that might selectively label plant fatty acid hydroxylases as
an aid to monitor the purification of these proteins.
Fatty acid acetylenic analogues have been used successfully to explore
the functional role and the reaction mechanisms of several mammalian
fatty acid -hydroxylases (15, 16). With the goal of selectively
inhibiting the plant
-LAH, we developed a series of radiolabeled
acetylenic analogues of lauric acid, [1-14C]8-, 9-, 10-, and 11-dodecynoic acids (DDYAs) (Fig. 1). We report here results of
microsomal incubations from clofibrate-treated V. sativa
seedlings, containing a lauric acid
-hydroxylase, with this series
of acetylenes. The mechanism by which the plant
-LAH was inactivated
by the radiolabeled acetylenic 11-dodecynoic acid was investigated,
metabolites generated were identified, and the chemical labeling of
microsomal proteins was demonstrated. For comparison, the effects of
10- and 11-DDYA on the oxidation of lauric acid by mammalian CYP4A and
2B4 were investigated.
Radiolabeled [1-14C]lauric acid (45 Ci/mol) was from Commissariat à l'Energie Atomique (Gif-sur-Yvette, France). Thin-layer plates (Silica gel G60 F254) were from Merck (Darmstadt, Germany). Clofibrate and NADPH were purchased from Sigma. Dilauroylphosphatidylglycerol was from Calbiochem, and silylating reagent (N,O-bis-(trimethylsilyl)trifluoroacetamide + trimethylchlorosilane) was from Pearce Europe (Oud-Beijerland, The Netherlands).
Synthesis of Acetylenic Analogues[1-14C]8-, 9-, 10-, and 11-dodecynoic acids were synthesized from 1-bromo-6-hexanol, 1-bromo-7-heptanol, and 1-bromo-9-nonanol by established procedures (17). [1-14C]Dodecynoic acids were prepared by Commissariat à l'Energie Atomique with a radiochemical purity greater than 98% and a specific activity of 42.3 Ci/mol.
Microsomal PreparationV. sativa L. (var.
Lolita) seedlings were purchased from S. A. Blondeau (Bersée,
France). Four-day-old etiolated seedlings were aged for 72 h in 1 mM clofibrate solution before isolation of microsomal
fractions as described by Salaün et al. (5). The
100,000 × g microsomal pellets were resuspended in
either 50 mM phosphate buffer (pH 7.4), 30% (v/v) glycerol
or in the same mixture containing 1.5 mM of
-mercaptoethanol. Microsomal proteins were quantified by a
microassay procedure from Bio-Rad using bovine serum albumin as a
standard. P450 was measured by the method of Omura and Sato (18).
Enzymatic activities were
measured as described previously (19, 20). The standard assay contained
in a final volume of 0.2 ml 0.19-0.43 mg of microsomal protein, 20 mM phosphate buffer (pH 7.4), and radiolabeled substrate.
-Hydroxylase activities were measured in the presence of 1 mM NADPH plus a regenerating system (consisting of final
concentration of 6.7 mM glucose-6-phosphate and 0.4 IU of
glucose-6-phosphate dehydrogenase), and 375 µM
-mercaptoethanol. The reaction was initiated with NADPH at 27 °C
and stopped with 0.2 ml of acetonitrile/acetic acid (99.8:0.2, v/v),
mixed, and aliquots (100 µl) were directly spotted onto silica plates
or injected into a RP-HPLC system. For metabolite isolation and
characterization, incubates were extracted twice with diethyl ether.
The solvent was dried, and the residue was dissolved in methanol for
TLC and RP-HPLC analysis. Products were quantified by liquid
scintillation.
For carbon monoxide inhibition studies, the incubates were equilibrated with a stream of a CO/O2/N2 (2:2:8, v/v/v) mixture at 4 °C for 20 min. After equilibration at 27 °C for 5 min, reactions were initiated with NADPH and carried out either in darkness or under a white light from a 150-watt heat-filtered quartz lamp located 15 cm from the sample.
Monoclonal antibodies directed against NADPH-cytochrome P450 reductase from Jerusalem artichoke tubers (21) were preincubated with microsomes at 20 °C for 15 min before enzyme activity measurement as described previously (1).
Inactivation with Mechanism-based InhibitorsThe procedure of Salaün and co-workers (10) was followed. Microsomes were preincubated at 27 °C with 1 mM NADPH plus a regenerating system in the presence of variable concentrations of [1-14C]11-DDYA. At different times, 100 µl of preincubated microsomes was added to the incubation medium, which contained 100 µM [1-14C]lauric acid, 1 mM NADPH, and the regenerating system in a total volume of 200 µl. Incubations were performed for 5 min at 27 °C and stopped as described above.
Chromatographic AnalysisMetabolites were resolved by silica thin-layer chromatography with a mixture of diethyl ether/light petroleum (boiling point, 40-60 °C)/formic acid (70:30:1, v/v/v). The areas corresponding to 12-hydroxylauric acid (RF = 0.45) and 1,12-dodecandioic acid (RF = 0.6) were scraped directly into counting vials or eluted with ether and subjected to RP-HPLC analysis using a mixture of water/acetonitrile/acetic acid (25:75:0.2) as described previously (20, 22).
Radioactivity of RP-HPLC effluents was monitored with a computerized on-line solid scintillation counter (Ramona-D; Raytest, Germany).
Chemical Hydrogenation of MetabolitesCompounds isolated by RP-HPLC were chemically hydrogenated (H2 and Pd/charcoal) as described by Weissbart et al. (19) and subjected again to RP-HPLC analysis before mass spectra analysis.
Gas Chromatography-Mass SpectrometryGas chromatography and electron impact (70 eV) ionization mass spectrometry studies were performed as described elsewhere (19). The mass spectrum of the methyl ester derivative of the enzymatic product formed from [1-14C]11-DDYA showed predominant ions at m/z 98 (base peak), m/z 227 (M-31, 28%), m/z 185 (M-73, 30%), and m/z 153 (32%). This spectrum is similar to that obtained with the methyl ester of authentic 1,12-dodecandioic acid.
Metabolites generated from [1-14C]8-, 9-, and 10-DDYA by
the enzyme were hydrogenated, methylated, and silylated as described by
Weissbart et al. (19) and subjected to GC-MS analysis. The mass spectrum of the major metabolite generated from 10-DDYA (Fig. 2C, peak 5) was similar to that obtained with the authentic
12-hydroxylaurate methyl ester trimethylsilyl ether derivative (MeTMS).
Similar mass spectra were obtained for metabolites generated from 8- and 9-DDYA (Fig. 2, A, peak 1, and B, peak
3).
Reconstitution of Activity from Partly Purified P450 Preparation
A partially purified mixture of CYP4A5, -4A6, and -4A7 (CYP4As) was obtained from rabbit kidney during the preparation of P4502CAA (23). CYP2B4 was isolated as described by Coon et al. (24), and purified rabbit NADPH cytochrome P450 reductase and rabbit cytochrome b5 were obtained as described by Laethem et al. (23). A purified NADPH cytochrome P450 reductase from Helianthus tuberosus (21) showing similar efficiency as the rabbit reductase was also used to reconstitute solubilized P450s from both plant and mammalian sources.
Reconstitution of P450 activities was performed by the procedure of
Laethem and co-workers (25). Typical incubations contain 50 pmol of
P450, 50 pmol of NADPH cytochrome P450 reductase, 200 pmol of
cytochrome b5, 15 µg of
dilauroylphosphatidylglycerol, and 100 µM substrate.
Reactions were started by adding NADPH (1 mM final
concentration) with the regenerating system noted above. The specific
activity of the CYP4As with lauric acid was 8 nmol/min/nmol of P450.
Metabolites generated were identified as - and (
-1)-hydroxylauric acids in a molar ratio of 17:1, respectively. CYP2B4 generated a
mixture of laurate metabolites hydroxylated at the
,
-1, and
-2 positions (1:5:0.7, respectively) with a specific activity of 800 pmol/min/nmol of P450.
Microsomal incubates containing
[1-14C]DDYA were extracted twice with 1 ml of cold
acetone to remove unreacted substrate and metabolites.
Acetone-insoluble protein (200-400 µg) was collected by
centrifugation (15 min at 3000 × g), dissolved in 40 µl of Tris buffer (180 mM, pH 6.8) containing 5% (v/v)
SDS, 5% (v/v) -mercaptoethanol, 30% (v/v) glycerol, and 0.025%
(v/v) bromphenol blue, and heated for 1 min in boiling water. Aliquots
of this SDS solution (10 µl) were counted by liquid scintillation and analyzed by SDS-PAGE (26) using a 10% acrylamide gel. After Coomassie
Blue staining, gels were soaked for 20 min in Amplify and dried.
Radioactivity was detected on dry gels after a 1-month exposure of
Kodak x-ray films (Hyper Film-
max, Amersham Corp.) or 8-15 days of
exposure using a bioimaging analyzer (Fugix Bas 2000).
Incubations of microsomes from clofibrate-treated V. sativa seedlings with [1-14C]8- and 9-dodecynoic acids generated two metabolites (Fig. 2, A and B). The metabolites generated either from 8-DDYA or 9-DDYA show similar RF values of 0.35 (Fig. 2, A, peak 1, and B, peak 3) and 0.45 (Fig. 2, A, peak 2, and B, peak 4) on TLC and similar retention times of 11 and 12.5 min by RP-HPLC analysis (not shown). Product formation was NADPH dependent and was inhibited by CO and by antibodies against a NADPH cytochrome P450 reductase isolated from Jerusalem artichoke (Table I). After catalytic hydrogenation and conversion to the corresponding MeTMS derivatives, fractions 1 and 3 (Fig. 2, A and B) generated from either 8- or 9-DDYA showed an MS fragmentation pattern identical to that of the authentic 12-hydroxylaurate MeTMS derivative. Based on HPLC retention times and MS fragmentation analysis, metabolites 1 and 3 were identified as 12-hydroxy-8-dodecynoic and 12-hydroxy-9-dodecynoic acids, respectively (Fig. 1). The structures of metabolites 2 and 4 remain to be determined.
When [1-14C]10-DDYA was incubated with microsomes only one metabolite was generated. The reaction product showed a RF of 0.4 on TLC (Fig. 2C, peak 5) and a retention time of 18 min by RP-HPLC analysis. After RP-HPLC isolation and chemical hydrogenation, this product showed a retention time in RP-HPLC similar to authentic 12-hydroxylauric acid (32 min). The identity of the hydrogenated metabolite was further confirmed by GC-MS analysis after conversion to a MeTMS derivative. The mass spectrum was identical to that of authentic 12-hydroxylauric acid (MeTMS). From these results metabolite 5 (Fig. 2C) was identified as 12-hydroxy-10-dodecynoic acid (Fig. 1).
As for 8- and 9-DDYA, metabolism of 10-DDYA was inhibited with both carbon monoxide and antireductase antibodies (Table I), suggesting the involvement of P450 in oxidation of 10-DDYA.
As shown in Table II, the Km values for the acetylenic derivatives were similar (25-110 µM) to that obtained for lauric acid. However, marked differences were noted in Vmax. The Vmax increased as the triple bond in the substrate moved away from the methyl end.
|
The
efficiency of 11-DDYA to irreversibly inhibit the microsomal -LAH
activity from V. sativa seedlings was previously
demonstrated (10). When microsomes were incubated with
[1-14C]11-DDYA, a polar metabolite (Fig.
3A, peak b) was generated. RP-HPLC
analysis of the material in peak b showed that it contained two labeled metabolites present in a molar ratio of 1:2.7 (Fig. 3B, peaks 10 and 11). The metabolite in
peak 11 was methylated than subjected to mass spectral
analysis. The fragmentation pattern of the methyl derivative of
peak 11 was similar to that of authentic 1,12-dodecandioic
acid dimethyl ester. Ions fragments were at m/z 74 (95%),
m/z 87 (41%), m/z 112 (31%), m/z 185 (M-73, 30%), m/z 227 (M-31, 28%), and a base peak at
m/z 98. As expected, ion fragments derived from the
14C isotope were increased by 2 atomic mass units and
detected at m/z 187 and 229. The metabolite in peak
10 was also isolated by RP-HPLC and subjected to GC-MS analysis.
Despite the presence of 14C, we were unable to determine
the structure of this product. Howewer, acidic hydrolysis of this
metabolite led to 1,12-dodecandioic acid (Fig. 1), suggesting that an
unknown nucleophile present in the incubation medium reacted with a
reactive intermediate. That this metabolite represents the initial
product comes from incubations with [1-14C]11-DDYA as a
function of time. The concentration of the unknown metabolite followed
the formation of dicarboxylic acid (Fig. 4, A
and B). The Kmm(app) for
11-DDYA was 25 µM, and the Vmax(app) was 20 pmol/min/mg, 10 times lower
than with laurate (200 pmol/min/mg) as a substrate (Table II).
As previously reported, 11-DDYA inhibited lauric acid hydroxylation by
V. sativa microsomes. The decrease in -LAH activity was
correlated with the formation of 11-DDYA metabolites (Fig. 5A). When [1-14C]lauric acid,
an alternate substrate, and [1-14C]11-DDYA were incubated
together, 12-hydroxylauric acid and 1,12-dodecandioic acid were
generated in a ratio of about 5:1, respectively. Metabolic activity of
the microsomal preparation toward both substrates (lauric acid and
11-DDYA) decreased with time, but the ratio of metabolites remained
constant, suggesting that both compounds were metabolized by the same
or very similar P450 isoforms (results not shown). Increasing
concentrations of lauric acid in the presence of 50 µM
[1-14C]11-DDYA resulted in a decrease in
1,12-dodecandioic acid formation and in a parallel increase of the
-LAH activity, indicating that an alternate substrate protects the
plant P450 enzyme against inactivation. Finaly, when microsomal
incubation containing 100 µM lauric acid and NADPH was
supplemented with increasing concentrations of 1,12-dodecandioic acid
(2-50 µM final concentrations), no inhibition of the
-LAH activity was measured (results not shown). These experiments
clearly indicate that the diacid generated from 11-DDYA was not
involved in the inhibition of
-LAH activity.
The possibility that enzyme inactivation was the result of heme destruction was investigated. The P450-CO spectra was measured before and after incubation with various 11-DDYA concentrations, and no appreciable changes were observed.
To investigate whether the inhibition was due to covalent modification of the protein, fractions were extracted twice with 1 ml of cold acetone to completely remove residual substrate and metabolites and to precipitate proteins. After solubilization, total radioactivity bound was determined, and an aliquot was analyzed by SDS-PAGE. All four isomeric DDYAs were examined, and only the [1-14C]11-DDYA covalently labeled protein. Compared with tissues from untreated plants, covalent binding was greater in microsomes from clofibrate-treated seedlings, required NADPH, and was inhibited by CO and monoclonal antibodies to NADPH cytochrome P450 reductase (results not shown). Covalent binding was correlated with the dependent formation of 1,12-dodecandioic acid (Fig. 5B).
Autoradiography after SDS-PAGE separation of the inactivated microsomal
proteins showed that 70% of the labeling was associated with low
molecular weight (LMW) components at the front of the gel. The
remaining 30% was associated with proteins with subunit molecular weights of about 53,000 (Fig. 6,
lane 2).
Effects of Thiol Compounds
The most plausible mechanism for
generation of 1,12-dodecandioic acid, the final reaction product, from
11-DDYA is via a reactive ketene intermediate. Alkylation by the
electrophile could occur in or out of the enzyme active site. We
therefore incubated microsomes with 11-DDYA in the presence of
nucleophilic compounds such as nucleophile amino acids, GSH, and
-mercaptoethanol. The two latter compounds strongly reduced diacid
formation. Cysteine was the only amino acid to produce similar
inhibition. RP-HPLC or TLC analysis of metabolites generated by
incubates containing
-mercaptoethanol showed the formation of
an additional metabolite (Fig. 3A, peak a). The
formation of this metabolite increased with increasing concentrations of
-mercaptoethanol and paralleled the decrease in
diacid formation (results not shown). The metabolite was isolated by
RP-HPLC, and after acid hydrolysis (4 N HCl, 12 h)
1,12-dodecandioic acid was released. Similar results were obtained
using both cysteine and glutathione. SDS-PAGE analysis also showed that
the labeled protein was greatly reduced (Fig. 6, line 4) when compared
to incubation without
-mercaptoethanol (Fig. 6, lane 2).
At least two protein bands remained labeled, but the nonspecific
labeling disappeared. This suggests that the reactive intermediate that mainly diffuses outside of the enzyme site interacts more efficiently with exogenous nucleophile compounds than with protein or water.
The
mixture of CYP4A isoforms generated mainly 12-hydroxylauric acid and to
a lesser extent 11-hydroxylauric acid. The ratio of /
-1 oxidation
was 17:1, respectively. CYP2B4 produced a mixture of 12-, 11-, and
10-hydroxylated lauric acids in a molar ratio of 1:5:0.7,
respectively.
Incubation of CYP4As with 11-DDYA resulted in a NADPH- and
time-dependent loss of lauric acid hydroxylation that
occurred with a half-life of 12.8 min. Under the same incubation
conditions, 100 µM 10-DDYA did not inactive the CYP4A
preparation, although this acetylenic analogue was efficiently
converted to 12-hydroxy-10-DDYA (results not shown). No significant
loss of spectrally detected P450 was observed. Autoradiography of CYP4A
after SDS-PAGE demonstrated that only incubation with the terminal
acetylene 11-DDYA resulted in the covalent labeling of proteins when
NADPH was present (Fig. 7A, lane
4). A great deal of the radioactivity was detected associated with
low molecular weight proteins present in the partially purified preparation.
Both 10-DDYA and 11-DDYA inhibited lauric acid hydroxylation by reconstituted CYP2B4 in a NADPH- and time-dependent manner. Enzyme activity was inhibited by 50% with 10-DDYA and 11-DDYA in 15 and 9.8 min, respectively. Autoradiography of inactivated CYP2B4 after SDS-PAGE demonstrated that both mechanism-based inhibitors bind covalently to the hemoprotein (Fig. 7B, lanes 6 and 8). These results strongly contrast with those obtained with P450s from the CYP4A family and show that CYP2B4 produces a metabolite of 10-DDYA that is able to covalently bind to the hemoprotein.
Although mammalian fatty acid -hydroxylases have been
extensively studied, and several of the corresponding genes have been cloned, the catalytic mechanism of these oxygenases is not well established (27, 28, 29). Less is known about the catalytic mechanism and
substrate specificity of
-hydroxylases from plants, enzymes that
have not been successfully purified. Among the increasing number of
plant P450 cDNAs sequenced to date, none shows greater than 40%
similarity to P450s from families 2, 4, 52, and 102, which catalyze
fatty acid oxidation in animals, fungi, and bacteria, respectively.
This is intriguing from a phylogenetic point of view, since not only
substrate specificity but also selective induction by peroxisome
proliferators appear similar between plant and animal
-hydroxylases.
One aim of the present work was to probe a better understanding of the
mechanism of suicidal inactivation by acetylenes and to determine the
relation between the regioselectivity of different fatty acid
hydroxylase isoforms and the potency of the external versus
internal position of the triple bond in the inhibitor molecule.
The mechanism of inactivation by terminal acetylenes has been extensively studied. It has been reported that the occurrence of heme alkylation versus apoprotein binding is related to oxidative attack at the internal versus terminal carbon of a terminal triple bond (15, 30).
Substrate analogues containing an internal acetylene were
-hydroxylated by
-LAH from V. sativa without any
measurable enzyme inactivation, corroborating previous results with
internal ethylenic laurate analogues (19). These studies had shown that
only the methyl end of substrates is accessible to oxidation by
V. sativa
-LAH, and that internal double or triple bonds
did not affect the regioselectivity but only kinetic parameters of the
hydroxylase. In the present study, two unknown polar metabolites were
also generated from 8- and 9-DDYA in addition to the 12-hydroxy
derivatives. Inhibition experiments with both CO and antireductase
antibodies suggest that these products may be generated by
-LAH.
In contrast, oxidation of the terminal acetylene (11-DDYA) resulted
mainly in the formation of 1,12-dodecandioic acid, which could be
formed by the addition of water to a putative ketene intermediate. An
unknown minor metabolite was also generated, probably by the same
reaction of an endogenous nucleophile other than water with the ketene,
since acidic hydrolysis of the metabolite produced the diacid. The
time- and concentration-dependent chemical labeling of
proteins by 11-DDYA was correlated with the formation of the
metabolites, suggesting a partition of the reactive intermediate between diacid formation, alternate product formation, and protein binding causing inactivation of the enzyme (Fig. 8).
In the presence of -mercaptoethanol, cysteine, or glutathione the
nonspecific labeling of microsomal proteins and diacid formation were
strongly decreased, whereas the kinetics of
-LAH inhibition were not
altered. In addition, there was persistent chemical labeling of the
Mr 53,000 protein in the presence of thiol
reagents. These findings are consistent with a reactive intermediate
that inactivates by reaction within the active site, whereas
nonspecific reactions with water or proximal protein residues occur
after diffusing outside the active site (Fig. 8). The fact that the
formation of 1,12-dodecandioic acid was not completely suppressed by
nucleophilic reagents suggests that this intermediate might also
interact in the active site with a molecule of water generated during
catalysis (Fig. 8). In addition, the lack of protection against
inactivation of
-LAH and the persistent chemical labeling of the
Mr 53,000 protein despite the presence of thiol reagents suggest that these nucleophiles cannot access the active site.
Similar effects of external nucleophiles have been recently reported by
Lopez-Garcia and co-workers (31) for inactivation of CYP2C9 by
thienilic acid. In contrast, Shirane and co-workers (32) reported that
inactivation of P450-BM3 (CYP102) by mechanism-based inhibitors (monothioesters of fatty acids) was completely prevented by
glutathione, indicating that the enzyme is inactivated by metabolites that diffuse out of the active site.
The partition ratio between covalent binding of proteins and metabolite
formation decreased from about 10 in the absence to 2.5 in the presence
of -mercaptoethanol. Under these conditions, the distribution of the
labeling was 30% associated with Mr
52,000-53,000 proteins and 70% with LMW compounds. This is close to
the ratio of about 2 that was found for inactivation of CYP4A1 by
11-DDYA (15). If the LMW compounds comprise heme adducts, a point that remains to be clarified, this would indicate that
oxidation of
acetylenes may involve heme alkylation, in contrast with the proposal
by Ortiz de Montellano and co-workers (15) that addition of activated
oxygen to the terminal carbon of acetylene involves protein rather than
heme modification.
The effects of internal and terminal acetylenic analogues on V. sativa -LAH with the effects on reconstituted activities of
purified mammalian forms that oxidize laurate at different positions
were compared. Our results show that the internal acetylenes exert an
unexpected and highly destructive effect on CYP2B4. In contrast, with
the CYP4s, only the terminal acetylene produced covalent binding and
irreversible inhibition, in agreement with the results obtained with
the V. sativa enzyme. Using a wheat microsome system, we
have recently demonstrated that cis-9-octadecen-16-ynoic acid (33) and 10-DDYA (34) irreversibly inhibit the oleic acid and
lauric acid (
-1)-hydroxylases, respectively. In addition, the
in vivo inhibition of lauric acid (
-1)-hydroxylase from
wheat by undec-9-yne-1-sulfonic acid confirms the efficacy of suicide substrate inhibitors containing an internal triple bond between the two
carbons preferentially attacked by the P450 enzyme (34).
A plausible mechanism, based on the formation of a putative unstable acetylene epoxide would involve: (i) the migration of an alkyl chain resulting from the rearrangement of the acetylene epoxide; and (ii) the formation of a diacid (10-methyl-undecane-1,11-dioic acid) resulting from water addition to an exoketene. In fact, this seems not to be the case, since we could not measure the formation of any diacid from 8-, 9-, and 10-DDYA either with 2B4 or with plant in-chain hydroxylases or P450-BM3 from Bacillus megaterium.
Our results on inactivation of -LAH from V. sativa and
CYP4A by the terminal acetylene corroborate the mechanism postulated by
Ortiz de Montellano and co-workers (15), except for the high degree of
labeling of low molecular weight compounds observed with both plant and
mammal systems (Figs. 6 and 7). However, following this postulate,
inactivation of CYP2B4 by 11-DDYA would mainly result from porphyrin
N-alkylation rather than binding to protein. Our results
demonstrate that incubation of both internal (10-DDYA) and terminal
acetylene (11-DDYA) results in a potent and irreversible inhibition of
CYP2B4 and labeling of the enzyme protein. Neither diacid formation nor
labeled LMW compounds were observed. The absence of diacid formation
contrasts with the results of Roberts and co-workers (35), who
demonstrated the formation of 2-naphthylacetic acid and the labeling of
CYP2B4 when incubated with 2-ethynylnaphthalene. Moreover, Miller and
White (36) recently demonstrated that water is the principal active
site nucleophile of CYP2B4. In the present study, the absence of diacid
formation suggests that the sequence of events leading to enzyme
inactivation could be different from those already described. One
alternative is for a direct interaction of an oxirene intermediate with
heme or protein. A second possibility is for a very fast and exclusive
interaction of a putative ketene with a residue from the active site of
CYP2B4. In both cases, either the reactive intermediate is short lived,
or the presence of reactive nucleophiles nearby does not allow
interaction with water. Finally, covalent binding to heme cannot be
excluded, even though no decrease of the spectrometrically P450-CO
complex was observed in inactivated CYP2B4 preparations. Our results
resemble most those already described for the inactivation of
P450SCC by acetylenic steroids. Nagahisa and co-workers
(37) did not observed the formation of electrophilic intermediates
under conditions that completely inactivated the enzyme.
Finally, in view of the natural presence of acetylenic compounds in
several plant species (38, 39, 40), it was of great interest to demonstrate
that compounds containing an internal acetylenic function might be
metabolized by both plant and mammalian P450-dependent
fatty acid -hydroxylases. The results show that inactivation of P450
occurs when the acetylenic function is adequately positioned even with
isoforms that perform in-chain oxidation.
A second goal of the present work was to devise a means to selectively tag fatty acid hydroxylases as a first step for further isolation and cloning. Very recently, a protein labeled with 11-DDYA was isolated from V. sativa microsomes by successive SDS-PAGE containing increasing concentration of polyacrylamide. The alkylated protein was submitted to in-gel proteolysis, and the resulting peptides were sequenced. By several criteria such as NH2-terminal and heme binding amino acid sequences, we concluded that the alkylated protein corresponded to a cytochrome P450 isozyme. To obtain insight into the substrate specificity, isolation of the corresponding cDNA and its expression in yeast are currently under way.
We are grateful to Prof. J. Capdevila for critically reading the manuscript. We thank Dr. C. Larroque (INSERM U-128) for the generous gift of purified cytochrome b5 from rabbit and Dr. A. Lesot for preparing monoclonal antibodies against NADPH cytochrome P450 reductase from H. tuberosus. We also thank M. F. Castaldi for excellent technical assistance and L. Thiriet for typing the manuscript.