Delineating novel metabolic pathways of DPC 963, a non-nucleoside reverse transcriptase inhibitor, in rats. Characterization of glutathione conjugates of postulated oxirene and benzoquinone imine intermediates by LC/MS and LC/NMR

Article abstract of DOI:10.1021/tx010153f

The metabolic activation of (S)-5,6-difluoro-4-cyclopropylethynyl-4- trifluoromethyl-3,4-dihydro-2(1H)-quinazolinone, DPC 963, in rats was investigated by identifying and characterizing the GSH and mercapturic acid conjugates excreted in the bile and urine, respectively. The structures of these adducts, which were unequivocally elucidated by LC/MS/MS and NMR experiments, revealed the existence of at least three distinct metabolic pathways leading to these products. One of the pathways, which has been described previously, involves the activation of the acetylene group after an initial hydroxylation on the methine carbon of the cyclopropyl ring. Metabolite M1 was demonstrated to be formed via this pathway after an enzymatic addition of GSH across the triple bond of the substituted acetylene. The second pathway, also previously described, leads to diastereoisomeric GSH adducts M3 and M4 after the formation of a highly reactive oxirene intermediate. This postulated oxirene subsequently rearranges to an α, β-unsaturated cyclobutenyl ketone intermediate capable of undergoing a 1,4-Michael addition with a nucleophile such as GSH. In addition to these pathways, DPC 963 was found to undergo a metabolic activation previously undescribed for structural analogues of this compound. It is postulated that an oxidative defluorination mediated by cytochrome P450 leads to the formation of a putative benzoquinone imine intermediate which subsequently reacts with GSH to form two aromatic ring-substituted regioisomeric conjugates, M5 and M6. In addition to forming the GSH adducts, the benzoquinone imine was also found to be reduced to its unreactive hydroquinone metabolite, which was excreted as the glucuronide conjugate in rat bile. Studies with induced rat microsomes, cDNA-expressed rat P450 isozymes, and polyclonal antibodies against rat P450 clearly demonstrated that the rat P450s 3A1/3A2 were responsible for the formation of postulated oxirene and benzoquinone intermediates.

Full text of DOI:10.1021/tx010153f

388
Chem. Res. Toxicol. 2002, 15, 388-399
Delin ea tin g Novel Meta bolic P a th w a ys of DP C 963, a
Non -Nu cleosid e Rever se Tr a n scr ip ta se In h ibitor , in Ra ts.
Ch a r a cter iza tion of Glu ta th ion e Con ju ga tes of
P ostu la ted Oxir en e a n d Ben zoqu in on e Im in e
In ter m ed ia tes by LC/MS a n d LC/NMR
Hao Chen, J ohn Shockcor, Weiqi Chen, Robert Espina, Liang-Shang Gan, and
A. E. Mutlib*
Drug Metabolism and Pharmacokinetics Section, Bristol-Myers Squibb Pharma Company,
Stine-Haskell Research Center, P.O. Box 30, 1094 Elkton Road, Newark, Delaware 19714
Received September 20, 2001
The metabolic activation of (S)-5,6-difluoro-4-cyclopropylethynyl-4-trifluoromethyl-3,4-dihy-
dro-2(1H)-quinazolinone, DPC 963, in rats was investigated by identifying and characterizing
the GSH and mercapturic acid conjugates excreted in the bile and urine, respectively. The
structures of these adducts, which were unequivocally elucidated by LC/MS/MS and NMR
experiments, revealed the existence of at least three distinct metabolic pathways leading to
these products. One of the pathways, which has been described previously, involves the
activation of the acetylene group after an initial hydroxylation on the methine carbon of the
cyclopropyl ring. Metabolite M1 was demonstrated to be formed via this pathway after an
enzymatic addition of GSH across the triple bond of the substituted acetylene. The second
pathway, also previously described, leads to diastereoisomeric GSH adducts M3 and M4 after
the formation of a highly reactive oxirene intermediate. This postulated oxirene subsequently
rearranges to an R, â-unsaturated cyclobutenyl ketone intermediate capable of undergoing a
1,4-Michael addition with a nucleophile such as GSH. In addition to these pathways, DPC
963 was found to undergo a metabolic activation previously undescribed for structural analogues
of this compound. It is postulated that an oxidative defluorination mediated by cytochrome
P450 leads to the formation of a putative benzoquinone imine intermediate which subsequently
reacts with GSH to form two aromatic ring-substituted regioisomeric conjugates, M5 and M6.
In addition to forming the GSH adducts, the benzoquinone imine was also found to be reduced
to its unreactive hydroquinone metabolite, which was excreted as the glucuronide conjugate
in rat bile. Studies with induced rat microsomes, cDNA-expressed rat P450 isozymes, and
polyclonal antibodies against rat P450 clearly demonstrated that the rat P450s 3A1/3A2 were
responsible for the formation of postulated oxirene and benzoquinone intermediates.
In tr od u ction
One of the objectives in the field of metabolism-
mediated toxicity is to elucidate the mechanisms by
which xenobiotics undergo metabolic activation to pro-
duce reactive intermediates. Knowledge of the different
pathways by which xenobiotics can be metabolized to
reactive intermediates increases our understanding of the
relationship between the chemical nature of xenobiotics
and their effects on living systems. Furthermore, such
knowledge will be helpful in identifying chemical func-
tional groups in a given molecule, which predispose that
compound to a sequence of processes that might eventu-
ally lead to toxicity.
F igu r e 1. Chemical structures of DPC 963, DPC 961, and
efavirenz.
currently in the early stages of clinical trials. DPC 963
is a structural analogue of efavirenz and DPC 961, the
metabolism of which have been described previously (1-
5). DPC 963 differs from DPC 961 and efavirenz in
having two fluorine substituents on the aromatic ring
(Figure 1). The addition of fluorines to the ring decreases
the reactivity of the pi electrons (i.e., decreases the energy
level of the highest occupied molecular orbital in the
aromatic ring). This subsequently leads to deactivation
of the aromatic system toward P450-mediated metabo-
lism such as hydroxylation. However, substitution with
fluorine also predisposes the aromatic ring to a potential
P450-mediated dehalogenation reaction. There is ample
DPC 963 (Figure 1), (S)-5,6-difluoro-4-cyclopropyleth-
ynyl-4-trifluoromethyl-3,4-dihydro-2(1H)-quinazolinone,
is a specific non-nucleoside human immunodeficiency
virus-1 reverse transcriptase inhibitor (NNRTI) and is
* Correspondence should be addressed to this author at the Drug
Metabolism and Pharmacokinetics Section, Bristol-Myers Squibb
Pharma Co., P.O. Box 30, 1094 Elkton Rd., Newark, DE 19714. Tel:
(302) 451-4679, Fax: (302) 366-5769.
10.1021/tx010153f CCC: $22.00 © 2002 American Chemical Society
Published on Web 02/12/2002

GSH Adducts of Oxirene and Benzoquinone Imine of DPC 963
Chem. Res. Toxicol., Vol. 15, No. 3, 2002 389
were purchased from Varian Sample Preparation Products
(Harbor City, CA). Waters Symmetry C18 columns (2.1 × 150
and 3.9 × 150 mm, 5 µM) were obtained from Waters Corp.
(Milford, MA). Beckman C18 semipreparative and analytical
C18 columns (10 × 300, 4.6 × 250 mm, respectively) were
obtained from Beckman Instruments Inc. (Fullerton, CA).
Control and various induced (dexamethasone, phenobarbital,
3-methylcholanthrene, â-naphthoflavone, Aroclor 1254) rat liver
microsomes were obtained from In Vitro Technologies (Balti-
more, MD). Rat P450 1A1, 1A2, 2A2, 2B1, 2C6, 2C11, 2C12,
2D1, 2D2, 2E1, 3A1, and 3A2 supersomes; polyclonal antibodies
against rat P450s 1A2, 2B1/2B2, and 2C11; and normal rabbit
serum were purchased from Gentest Corp. (Woburn, MA).
Monoclonal antibody against rat P450 3A1 and polyclonal
antibody against rat P450 3A1/3A2 were obtained from Xenotech
(Kansas City, KS). The chemical inhibitors sulfaphenazole
(human P450 2C9), troleandomycin (human P450 3A4), and
D609 (8, rat P450s 2B1/2) were obtained from Sigma-Aldrich
Chemical Co. (Milwaukee, WI) while furafylline (human P450
1A2) was purchased from Research Biochemicals International
(Natick, MA). Ketoconozole (P450 3A) was obtained from ICN
Biomedicals Inc. (Aurora, OH). â-Glucuronidase (E. coli, EC
3.2.1.31) was obtained from Sigma-Aldrich Chemical Co. (Mil-
waukee, WI). All solvents and reagents were of the highest grade
commercially available.
literature precedence for heme-mediated oxidative de-
halogenation, whereby fluorine is lost easily from a
compound. One of the aims of having two fluorine
substituents on the aromatic ring was to confer some
degree of metabolic stability, but at the same time, it was
realized that the potential for defluorination did exist.
Efavirenz and DPC 961, which had a chlorine substitu-
ent, were easily metabolized by aromatic ring hydroxyl-
ation. There was no evidence of any dechlorinated
metabolites formed from these compounds, consistent
with previous observations that chlorine is not as easily
lost as fluorine during metabolism of compounds.
Efavirenz and DPC 961 were found to be metabolized
extensively by rats, cynomolgus monkeys, and humans
with the major metabolite being a glucuronide conjugate
of the aromatic ring hydroxylated metabolite. There were
significant differences in the nature of the metabolites
that were excreted in the urine of all three species (1, 2,
5) after being dosed with either efavirenz or DPC 961.
LC/MS and NMR analyses of bile and urine samples from
rats dosed with efavirenz showed the presence of gluta-
thione-derived adducts formed by enzymatic addition of
glutathione across the triple bond of the acetylene moiety.
The formation of such conjugate has been linked to the
species-specific nephrotoxicity observed in rats (3, 4).
Further studies with the structural analogue, DPC 961,
were done in rats to investigate if a similar GSH adduct
was produced. However, the structures of the GSH
conjugates isolated from the bile of rats dosed with DPC
961 pointed to a totally different mechanism for the
formation of the GSH adducts with DPC 961 as compared
to that with efavirenz. The GSH conjugate of efavirenz
was formed by enzymatic addition of GSH across the
triple bond after the initial hydroxylation on the methine
carbon of the cyclopropyl ring (2). It was postulated that
the GSH adducts of DPC 961 found in rat bile were
produced as a result of oxidation of the triple bond
leading to an unstable oxirene intermediate which sub-
sequently rearranged to an R,â-unsaturated cyclobutenyl
compound (5). It was of considerable interest to us to
compare and contrast the metabolites formed from DPC
963 with its previously studied analogues, efavirenz and
DPC 961. The metabolism of DPC 963 was investigated
in rats, cynomolgus monkeys, and humans (6). It was
found that DPC 963 was metabolized extensively in these
species to produce 8-OH DPC 963 glucuronide and sulfate
conjugates as the major metabolites. However, significant
differences in the nature of metabolites that were ex-
creted in the urine of all three species were observed (6).
Liqu id Ch r om a togr a p h y/Ma ss Sp ectr om etr y (LC/MS).
LC/MS was carried out by coupling an HPLC system to a PE
Sciex API 300 mass spectrometer using a turbo ionspray
interface held at 400-450 °C. The electrospray needle was
maintained at 4500-5000 V with the declustering potential set
at 30-40 V. Ultrapure nitrogen was used as the nebulizer gas
set at 40 psi. The turbo ionspray gas flow rate was 6 L/min.
The mass spectrometer was operated either in the positive or
in the negative ion mode (both full scan and MS/MS mode) to
achieve optimum sensitivity for GSH adducts. MS/MS analysis
was carried out using nitrogen as the collision gas. The collision
energy was kept between 30 and 40 eV. The metabolites were
separated on a Waters Symmetry C18 column (2.1 × 150 mm)
by a gradient solvent system consisting of acetonitrile and 10
mM ammonium formate (pH 3.2). The percentage of acetonitrile
was increased from 25 to 80% over 18 min with the solvent flow
rate set at 0.4 mL/min. After 18 min, the column was washed
with 80% acetonitrile for 2 min before reequilibrating with the
initial mobile phase. Aliquots of bile and urine samples were
injected directly onto the HPLC column, and the eluent was
introduced into the source of the mass spectrometer. To detect
metabolites in the fractions from C18 cartridges and from a
semipreparative HPLC column, aliquots (60 µL) of samples were
introduced to the mass spectrometer using the flow injection
analysis method. The mobile phase consisted of a mixture of
acetonitrile and 10 mM ammonium formate (pH 3.2) (1:1 v/v)
delivered at a flow rate of 0.4 mL/min.
Liqu id Ch r om a togr a p h y/High -F ield NMR. All the spec-
tra were obtained using a Bruker Avance 500 MHz NMR
The present study was carried out to elucidate the
structures of all the GSH conjugates of DPC 963 in rats.
An attempt was made to delineate the possible routes of
metabolism that led to these metabolites. Furthermore,
we aimed to identify the rat P450 enzyme(s) responsible
for the metabolic activation of DPC 963 using various
techniques. To confirm the isozymes responsible for
forming the reactive intermediates, microsomal incuba-
tions were performed employing antibodies or selective
chemical inhibitors directed against specific P450. The
results were confirmed by carrying out incubations with
various rat cDNA-expressed P450 isozymes.
1
spectrometer equipped with a H/13C or 19F LC/NMR flow-probe
with a cell volume of 120 µL. Suppression of the residual water
and acetonitrile signals was carried out using the WET solvent
suppression method in all the LC/NMR experiments. Proton
chemical shifts were referenced to DMSO at δ 2.49 and to
acetonitrile at δ 2.0, and fluorine chemical shifts were referenced
to TFA at δ -78.5. Signal multiplicities are reported as
follows: s (singlet), d (doublet), t (triplet), q (quartet), dd (double
of doublets), and m (multiplet). An HP1100 LC system was used
with a variable-wavelength UV detector set at 240 nm. A Waters
Symmetry C18 column (3.9 × 150 mm, 5 µm) was used to
separate DPC 963 metabolites. A gradient from 25% aceto-
nitrile-d3 and 75% D2O to 80% acetonitrile-d3 over 20 min at a
flow rate of 0.8 mL/min was employed for separation. Both
solvents contained 0.05% TFA. The structures of DPC 963 GSH
adducts were determined from 1H and 19F NMR as well as
proton-proton (COSY) and total correlated spectroscopy
(TOCSY) 2-dimensional NMR.
Ma ter ia ls a n d Meth od s
Ch em ica ls a n d Su p p lies. DPC 963 was synthesized as
described before (7). Bond-Elut C18 cartridges (10 g/60 cm3)

390 Chem. Res. Toxicol., Vol. 15, No. 3, 2002
Chen et al.
Sch em e 1. P r op osed Meta bolic P a th w a ys Lea d in g to th e F or m a tion of DP C 963 Glu ta th ion e Ad d u cts in
Ra ts
An im a l Stu d ies. Bile duct-cannulated Sprague-Dawley
male rats (300-350 g; Charles River Laboratories, Wilmington,
MA) housed in metabolic cages were given an oral DPC 963
suspension twice daily at a dose of 30 mg/kg for 3 days before
treatment with a higher dose of DPC 963 at 300 mg/kg/day for
the next 3 days. The rats were housed individually in suspended,
stainless-steel, wire-mesh cages equipped with an automatic
watering system. The study room was environmentally con-
trolled for temperature (72 ( 4 °F), relative humidity (40-70%),
and light (a 12 h light/dark cycle). Rats had free access to water
and were given a specific amount of certified Purina rodent chow
each day. Bile and urine samples were collected over ice on a
daily basis and stored at -20 °C until assayed. Aliquots of bile
and urine samples were analyzed by LC/MS/MS as described
above.
Isola tion of GSH-Rela ted Ad d u cts of DP C 963 fr om Ra t
Bile a n d Ur in e. Bile or urine samples collected from several
rats were pooled, diluted 1:1 with distilled water (for bile only),
and loaded on a Bond-Elut C18 cartridge (10 g/60 cm3), which
was preconditioned with methanol and water. The sample was
allowed to elute under gravity at a rate of less than 1 mL/min.
After the sample was loaded, the column bed was washed with
25 mL of deionized water followed by eluting with 25 mL of
different percentages of methanol in water (5-100%). The
fractions containing the GSH adducts (5-30% methanol frac-
tion) were dried and reextracted on C18 cartridges using
different proportions of methanol in 0.2% acetic acid as the
eluent. The samples from the second extraction on C18 car-
tridges (60-70% methanol fraction) were concentrated. The
presence of GSH adducts was confirmed by LC/MS analyses.
The resultant crude mixture of GSH conjugates was purified
on a C18 semipreparative column (Beckman, 10 × 250 mm)
using a mixture of either acetonitrile/0.1% acetic acid or
acetonitrile/10 mM ammonium formate buffer (pH 3.4) as the
mobile phases. The percentage of acetonitrile was increased from
20% to 60% in 20 min with the solvent delivered at a flow rate
of 4.7 mL/min. The eluent was monitored at 240 nm. After
pooling the fractions containing the GSH adducts, the samples
were dried under vacuum and submitted for NMR analyses.
Syn th esis (En zym a tic) of Meta bolite M7. The glucuronide
conjugate of defluorinated metabolite M7 (Scheme 1) was
isolated from the bile of rats by using the similar method as
described above. Briefly, the glucuronide conjugate was ex-
tracted from the bile samples using C18 cartridges. The crude
bile extract was further purified on a Symmetry C18 column
(3.9 × 150 mm) using a mixture of acetonitrile/0.05% trifluoro-
acetic acid as the mobile phases. The percentage of acetonitrile
was increased from 10% to 70% in 18 min with the solvent
delivered at a flow rate of 1.0 mL/min. The peak eluting at 7
min (tR) was collected from several injections, pooled, and dried
under vacuum to give a white powder (approximately 3.4 mg).
The sample was analyzed by LC/MS/MS and shown to have
[M-H]- at m/z 489 with characteristic fragment ions at m/z 175
(glucuronic acid), 243, and 313 (- glucuronic acid). 1H NMR
clearly showed the presence of glucuronic acid and two aromatic
proton signals. The existence of an intact cyclopropyl moiety
was also confirmed by NMR data. The glucuronide conjugate
was dissolved in 1 mL of potassium phosphate buffer (0.1 M,
pH 6.5) containing â-glucuronidase (1000 units). The mixture
was incubated at 37 °C, and after 3 h, the reaction mixture was
analyzed by HPLC using a Symmetry C18 column (3.9 × 150
mm). The column was eluted with a mixture of acetonitrile/
0.05% trifluoroacetic acid (40:60, v/v) delivered at a flow rate of
1.0 mL/min. The peak eluting at 3.5 min was collected from
several injections and analyzed by LC/MS/MS. The fractions
containing metabolite M7 were pooled and samples dried under
vacuum to give approximately 2.6 mg of compound. 1H NMR
(DMSO-d6) δ 0.71 (2H, m, cyclopropyl-CH2), 0.85 (2H, m,

GSH Adducts of Oxirene and Benzoquinone Imine of DPC 963
Chem. Res. Toxicol., Vol. 15, No. 3, 2002 391
cyclopropyl-CH2), 1.35 (1H, m, cyclopropyl-CH), 6.71 (1H, d, J
) 8.5 Hz, aromatic 8-H), 7.34 (1H, dd, J ) 8.4, 8.5 Hz, aromatic
7-H); ESI-MS m/z 313 ([M-H]-).
Ch a r a cter iza tion of th e Ra t P 450 Resp on sible for th e
F or m a tion of GSH Ad d u cts M3-M6. The formation of
metabolites M3-M6 was investigated by incubating DPC 963
with either noninduced or induced rat liver microsomes and with
cDNA-expressed rat P450 enzymes. All incubations were per-
formed either in the presence or in the absence of GSH. To
further characterize the enzyme(s) involved in a particular
reaction, the in vitro formation of the respective GSH adducts
was studied using antibodies against the rat P450 isozymes and/
or by employing selective chemical inhibitors.
Liver Micr osom a l In cu ba tion s. Unless stated otherwise,
all microsomal incubations were carried out using the following
protocol: rat liver control and induced microsomes (100 µg),
(NADPH (0.1 mM), (GSH (1 mM), MgCl2 (3 mM), DPC 963
(50 µM), with the final volume adjusted to 0.5 mL with 0.1 M
phosphate buffer (pH 7.4). Incubations were carried out for 40
min and the reactions terminated by the addition of 1 mL of
ice-cold acetonitrile. The samples were vortexed and centrifuged
at 1600g for 10 min. The supernatant was separated and dried
under nitrogen. The dried samples were reconstituted in 200
µL of the HPLC mobile phase and aliquots analyzed by LC/MS
operated in the selected ion monitoring mode (SIM).
cDNA-Exp r essed P 450 Su p er som e In cu ba tion s. To probe
which rat isozyme was responsible for the formation of GSH
adducts M3-M6, incubations were carried out with com-
mercially available rat P450 supersomes including 1A1, 1A2,
2A2, 2B1, 2C6, 2C11, 2D1, 2D2, 3A1, and 3A2. Each of these
supersomes consisted of the cytochrome P450, the P450 reduc-
tase, and cytochrome b5. The incubation mixtures consisted of
the supersomes (50 pmol of P450), NADPH (0.1 mM), DPC 963
(50 µM), MgCl2 (3 mM), and 0.1 M phosphate buffer. Incubations
were performed in the presence or absence of GSH (1 mM). The
final volume of incubation was adjusted to 0.5 mL. The mixtures
were incubated for 40 min, after which the reactions were
terminated, extracted, and analyzed as described above.
F igu r e 2. LC-ESI/MS/MS spectra of the glutathione adduct
M1 (top) isolated from rat bile and the cysteinylglycine conjugate
M2 (bottom) isolated from rat urine.
dried under nitrogen. The residues were reconstituted and
analyzed by LC/MS as described above.
Resu lts
In h ibitor y Stu d ies w ith An tibod ies. Dexamethasone-
induced rat liver microsomes were used in immunoinhibition
experiments. The incubations consisted of microsomes (10 µg),
anti-rat CYP enzyme polyclonal antiserum or anti-rat CYP
antibody (100 µg), MgCl2 (3 mM), NADPH (1 mM), DPC 963
(50 µM), and 0.1 M phosphate buffer (pH 7.4) in the presence
of GSH (1 mM) in a final volume of 0.5 mL. Microsomes were
preincubated with individual antibodies for 15 min at room
temperature, followed by addition of other components. The
incubations were carried out for an additional 40 min, after
which the reaction was terminated by the addition of 1 mL of
ice-cold acetonitrile. The samples were analyzed by LC/MS SIM
as described above.
Ch a r a cter iza tion of GSH-Der ived Con ju ga tes M1
a n d M2. Metabolite M1 was found to be a GSH-sulfate
diconjugate of dihydroxylated DPC 963. LC/MS/MS analy-
ses (Figure 2) of M1 showed a characteristic loss of a
sulfate group (-80 amu) from the parent ion ([M-H]-
at m/z 734). The total addition of 419 amu to the molecule
suggested that this metabolite was formed by an initial
hydroxylation with subsequent phase II conjugation
reactions. Because of the presence of a sulfate group, the
mass difference (419 - 80 ) 339) suggested the addition
of two oxygen atoms (due to phase I oxidation) as well
as a GSH moiety (339 - 32 ) 307). The ion at m/z 525,
a characteristic fragment by a loss of glutamate (129
amu) from the GSH (10), confirmed M1 as a GSH
conjugate. The MS/MS spectrum (Figure 2) of metabolite
M1 suggested that the GSH moiety was attached to the
cyclopropyl ethynyl side chain as previously observed
with efavirenz (2). An ion at m/z 381 indicated a fragment
produced by cleavage of the cysteinyl C-S bond with the
In h ibitor y Stu d ies w ith Ch em ica l In h ibitior s. Inhibition
of M3-M6 GSH adduct formation was assessed by known
chemical inhibitors for P450s which have been well character-
ized for human P450s (9). We attempted to use these chemical
inhibitors so that we could gain information about the rat P450
that might be involved in the production of metabolites M3-
M6. The incubation consisted of dexamethasone-induced rat
liver microsomes (100 µg), NADPH (0.1 mM), MgCl2 (3 mM),
GSH (1 mM), and various chemical inhibitors used at two
concentrations. The following selective inhibitors were examined
for their effect: furafylline (10 and 25 µM), D609 (10 and 25
µM), sulfaphenazole (10 and 25 µM), troleandomycin (10 and
25 µM), and ketoconazole (1 and 10 µM). Furafylline, D609, and
troleandomycin were preincubated for 20 min with the mi-
crosomes and NADPH in the presence or absence of GSH (1
mM) before adding DPC 963 (50 µM). The volume of each
incubation was adjusted to 0.5 mL with 0.1 M phosphate buffer
(pH 7.4). The incubations were carried out at 37 °C for 40 min.
To terminate the reaction, 1 mL of ice-cold acetonitrile was
added to the incubation mixtures, and the samples were
vortexed and centrifuged. The supernatants were separated and
1
retention of sulfur by the drug molecule (Figure 2). H-
LC/NMR of metabolite M1 present in rat bile provided
conclusive evidence of a GSH adduct and the possible site
1
of conjugation. The H-LC/NMR showed signals for the
aromatic proton (δ 7.40, 1H, dd, J ) 7.2, 10.3 Hz) which
was coupled to the two fluorines (positions 5 and 6). A
singlet at δ 6.40 was assigned to the alkene proton, and
multiplets at δ 1.01 and 1.05 were assigned to the
cyclopropyl protons (Figure 3). The characteristic mul-
1
tiplet for the cyclopropyl methine proton seen in H NMR
(at approximately δ 1.3) was distinctly absent as was
demonstrated by total correlated spectroscopy (TOCSY,

392 Chem. Res. Toxicol., Vol. 15, No. 3, 2002
Chen et al.
F igu r e 3. 1H-LC/NMR of the glutathione conjugate M1 isolated from rat bile.
data not shown). The GSH protons were assigned by
TOCSY experiments to be at δ 4.10 (Cys R), 3.98 (Glu
R), 3.85 (Gly R), 3.50 (Cys â), 2.94 (Cys â′), 2.35 (Glu γ),
and 2.12 (Glu â). The 1H-LC/NMR of the isolated M1
clearly showed the presence of GSH, the loss of the
aromatic proton on C8, and the appearance of a new
singlet at 6.40 ppm (Figure 3). The proton at 6.40 ppm
1
was assigned to position 12 (Figure 3) based on H-19F
long-range couplings. The absence of the cyclopropyl
methine proton, which was replaced by a hydroxyl group,
was confirmed by TOCSY (data not shown). The hy-
droxylation at the cyclopropyl methine is consistent with
the slight downfield shift of the cyclopropyl methylene
protons. Comparisons with reference metabolite (1H NMR
of glutathione-sulfate diconjugate of dihydroxylated
efavirenz, 2) suggested that the hydrogen sulfate was at
the C8 position and not on C14. The geometry of the
substituents on the double bond was not determined.
Metabolite M2, isolated from rat urine, was identified
as a cysteinylglycine-sulfate diconjugate with [M-H]- at
m/z 605. It showed a similar fragmentation pattern
(Figure 2) as metabolite M1. However, M2 showed
[M-H]- which was 129 amu less than that of metabolite
M1. Fragment ions at m/z 381 and 267 confirmed that
this was an S-linked hydroxylated metabolite probably
formed by the hydrolysis of glutamate (-129 amu) from
metabolite M1. The structure of metabolite M2 was
F igu r e 4. LC-ESI/MS/MS spectra of the glutathione adducts
M3 (top) and M4 (bottom) isolated from rat bile.
confirmed by isolating sufficient quantities from rat urine
and carrying out one-dimensional proton NMR. The
chemical shifts of protons on positions 7, 12, 15, and 16
(Figure 3) were similar to the corresponding protons in
at m/z 640 with identical fragment ions at m/z 565, 511,
494, 408, and 333 (the base peak). The ions at m/z 565
and 511 were characteristic fragments produced by a loss
of glycine (75 amu) and glutamate (129 amu), respectively
(10). The ion at m/z 408, a characteristic fragment
produced from the internal cleavage of aliphatic thioether
conjugates (10), suggested that the GSH moiety was
attached to the cyclopropyl ethynyl side chain. Further
tandem mass spectrometry studies of the ion at m/z 333
(MS/MS/MS, data not shown) showed that the quinazoli-
none ring of adduct M3 or M4 was intact. The structures
of these two GSH adducts were elucidated by 1H and
TOCSY NMR experiments. The NMR data for one of the
adducts (M3) are shown in Figures 5 and 6. The 1H NMR
spectrum clearly showed the characteristic signals for the
1
metabolite M1. The long-range H-19F coupling/decou-
pling experiments confirmed the alkene proton on C12.
1
The major difference between the H NMR spectra (data
not shown) of M1 and M2 was the absence of any
glutamic acid proton signals for M2.
Ch a r a cter iza tion of GSH Ad d u cts (M3 a n d M4)
Der ived fr om th e Cyclobu ten yl Keton e In ter m ed i-
a te. Metabolites M3 and M4 were separable on HPLC
and confirmed as stereoisomers derived from a putative
cyclobutenyl ketone intermediate. The structures of these
two GSH adducts isolated from rat bile are shown in
Scheme 1. The MS/MS analyses (Figure 4) of these
adducts exhibited the same pseudomolecular ion ([M+H]+)

GSH Adducts of Oxirene and Benzoquinone Imine of DPC 963
Chem. Res. Toxicol., Vol. 15, No. 3, 2002 393
F igu r e 5. 1H-LC/NMR of the glutathione adduct isomer M3 isolated from rat bile.
acids showed the same pseudomolecular ion ([M+H]+) at
m/z 496 with a mass spectral fragmentation pattern
similar to those of the GSH adducts (data not shown).
1
The H NMR spectra (data not shown) were similar to
those of GSH adducts except for the distinct absence of
any proton signals from the glutamic acid and glycine.
The following chemical shifts were observed for the
protons of one of the isolated N-acetylcysteine adducts:
δ 1.85 (2H, m, cyclobutyl), δ 2.10 (2H, m, cyclobutyl), δ
2.24 (2H, m, cyclobutyl), δ 2.65 (1H, m, cysteine â), δ 2.87
(1H, m, cysteine â′), δ 3.55 (1H, q, cyclobutyl), δ 3.62 (1H,
q, cyclobutyl), δ 4.35 (1H, m, cysteine R), δ 6.80 (1H, d,
Ar-8), δ 7.39 (1H, q, Ar-7). The singlet for the methyl
group of the N-acetyl appeared at 1.90 ppm. The other
N-acetylcysteine adduct isolated from rat urine was not
further characterized.
Ch a r a cter iza tion of GSH Ad d u cts (M5 a n d M6)
Der ived fr om th e Ben zoqu in on e Im in e In ter m ed i-
a te. LC/MS/MS analyses of rat bile showed two well-
resolved GSH adducts (M5 and M6) with the same
pseudomolecular ion ([M+H]+ at m/z 620). The MS/MS
analysis of these adducts demonstrated slightly different
fragment ions for M5 and M6. The major fragment ions
produced from M5 were at m/z 545, 491, 474, 425, 408,
388 (the base peak), and 345; in contrast, the major
fragment ions observed with M6 were at m/z 602, 557,
545, 491, 371 (the base peak), and 345 (Figure 7). The
ions at m/z 545 and 491 were characteristic fragments
produced by a loss of glycine (75 amu) and glutamate (129
amu), respectively. The presence of the ion at m/z 425
(relatively in high abundance in the M5 spectrum but
fairly low in M6) produced by losing the glutamate (129
amu) and the cyclopropyl acetylene group (66 amu)
indicated that the conjugation occurred on the aromatic
ring. The diagnostic ion at m/z 345 was derived from
cleavage of the cysteinyl C-S bond with the retention of
the sulfur on the aromatic ring. Compared to the molec-
ular weight of DPC 963, the addition of 29 amu to the
ion at m/z 345 (316 - 19 + 16 + 32 ) 345) suggested
that a fluorine was replaced by an oxygen atom. The
origin of major fragment ions for both M5 and M6 is
F igu r e 6. Total correlated spectroscopy (TOCSY) of the gluta-
thione adduct isomer M3 isolated from rat bile.
GSH protons at δ 4.35 (Cys R), 3.94 (Glu R), 3.92 (Gly
R), 2.82 (Cys â), 2.58 (Cys â′), 2.48 (Glu γ), and 2.12 (Glu
â). The presence of the aromatic protons on C7 and C8
1
was clearly demonstrated by H NMR (Figure 5). How-
ever, the characteristic multiplet for the cyclopropyl
protons seen in 1H NMR of the parent compound was
absent. Instead, characteristic signals for a cyclobutane
ring were present (Figure 5). The protons of the cyclo-
butane ring were further demonstrated to be part of the
same spin system by TOCSY (Figure 6). The 1H NMR
analyses of the other GSH adduct M4 exhibited a similar
spectrum with slight differences in the chemical shifts
and the split patterns of the cyclobutyl ring protons (data
not shown) as compared to M3.
The structures of two corresponding N-acetylcysteine
conjugates, present in rat urine, were confirmed by mass
spectral and NMR analyses. Both of these mercapturic

394 Chem. Res. Toxicol., Vol. 15, No. 3, 2002
Chen et al.
1
comparison, the H-LC/NMR spectrum of the other GSH
adduct (M6) is shown in Figure 10. The presence of a
signal as a doublet (δ 7.26, d, J ) 7.3 Hz) for the proton
on C7 is consistent with the expected coupling to the
fluorine at C5.
The structures of two corresponding N-acetylcysteine
conjugates formed from M5 and M6 were confirmed by
mass spectral and NMR analyses. Both mercapturic
acids, found in rat urine, showed the same pseudo-
molecular ion ([M+H]+) at m/z 476 with the mass spectral
fragmentation patterns similar to the corresponding GSH
adduct. The 1H NMR spectra were similar to those of
GSH adducts except for the distinct absence of glutamate
and glycine proton signals at δ 2.08 (2H, Glu â), 2.42 (2H,
Glu γ), 3.84 (1H, Glu R), and 3.74 (2H, Gly). The signals
for the protons of N-acetylcysteine adducts matched those
of the corresponding GSH adduct (see Figures 3 and 5
for the assignment for signals). The following chemical
shifts were observed for the protons of one of the isolated
N-acetylcysteine adducts: δ 0.27 (2H, m, cyclopropyl),
0.46 (2H, m, cyclopropyl), 1.29 (1H, m, cyclopropyl), 3.22
(1H, m, cysteine â), 3.45 (1H, m, cysteine â′), 4.42 (1H,
m, cysteine R), 6.64 (1H, s, Ar-H). The singlet for the
methyl group of the N-acetyl appeared at 1.85 ppm. In
addition, 19F NMR clearly showed the presence of signals
for the fluorine at C-5 and the absence of fluorine at C-6.
The other N-acetylcysteine adduct isolated from rat urine
1
was not characterized by H and 19F NMR.
F igu r e 7. LC-ESI/MS/MS spectra of the glutathione adducts
M5 (top) and M6 (bottom) isolated from rat bile.
LC/MS analyses of bile samples from rats dosed with
DPC 963 showed the presence of a number of GSH
conjugates including M1 and M3-M6 (Scheme 1). How-
ever, the glucuronide and sulfate conjugates of the
aromatic ring hydroxylated metabolite were the major
drug-related components in bile and urine.
Ch a r a cter iza tion of th e Ra t P 450 Resp on sible for
th e F or m a tion of GSH Ad d u cts M3-M6. Several
approaches were used to identify the rat P450 enzyme(s)
responsible for the formation of GSH adducts M3-M6.
The studies using control and induced rat liver mi-
crosomes demonstrated a marked difference in the for-
mation of respective GSH conjugates. It was found that
greater quantities of the GSH adducts M3-M6 were
produced in the presence of dexamethasone-induced rat
microsomes as compared to control (saline-treated) mi-
postulated in Figure 7. The structures of these two GSH
adducts were further elucidated by 1H and 19F NMR
experiments. The NMR data for one of these adducts
(M5) are shown in Figures 8 and 9. The 1H-LC/NMR
spectrum clearly showed the characteristic proton signals
for the cyclopropyl group as well as for the GSH moiety
(Figure 8). The GSH protons were assigned at δ 4.42 (Cys
R), 3.85 (Glu R), 3.74 (Gly R), 3.29 (Cys â), 3.14 (Cys â′),
2.48 (Glu γ), and 2.05 (Glu â). In addition, the presence
of the proton at C8 was unequivocally demonstrated as
a singlet at δ 6.62 while the proton at C7 was distinctly
absent. Further 19F-LC/NMR characterization clearly
showed the absence of fluorine at C6 and retention of
the fluorine at C5 (δ 134.7, d) as compared to the fluorine
signals of DPC 963 (see the inset in Figure 9). As a
F igu r e 8. 1H-LC/NMR of the glutathione adduct isomer M5 isolated from rat bile.

GSH Adducts of Oxirene and Benzoquinone Imine of DPC 963
Chem. Res. Toxicol., Vol. 15, No. 3, 2002 395
F igu r e 9. 19F NMR of the glutathione adduct isomer M5 isolated from rat bile.
F igu r e 10. 1H-LC/NMR of the glutathione adduct isomer M6 isolated from rat bile.
crosomes. In addition to M5 and M6, dexamethasone-
induced rat microsomes produced much higher levels of
metabolite M7 as compared to microsomes induced with
other chemicals. The cDNA-expressed rat P450 enzymes
produced similar GSH adducts as microsomes. The rat
P450 3A1 and 3A2 produced the highest levels of me-
tabolites M3-M7 as compared to other P450s, hence
confirming the significant role of these isozymes in
producing reactive metabolic intermediates of DPC 963.
The rat P450 responsible for the formation of M3-M7
was further examined by the inhibitory effects of the
mono- and polyclonal antibodies to rat P450 enzymes.
The anti-rat P450 3A1/3A2 antibody inhibited production
of M3-M6 by about 85-90% as compared to the control.
Studies carried out with the antibody against P450 1A2,
2B1/2, 2C11, and 3A1 showed much less extent of
inhibition on the formation of these conjugates. The
reduction in the formation of GSH adducts in the pres-
ence of other anti-rat P450 antibodies could possibly be
due to cross-reaction with P450 3A1/3A2 isozymes.
Nonetheless, the contribution from other P450 isozymes
capable of catalyzing the same metabolic reactions cannot
be ruled out. Further evidence for the rat P450 3A1/3A2
being responsible for the production of M3-M6 was
obtained by inhibiting microsomes with chemical inhibi-
tors. Studies with chemical inhibitors specific for human
P450 isozymes (9) supported the involvement of rat P450
3A1/3A2 in forming reactive metabolic intermediates
from DPC 961, a close analogue of DPC 963 (5). Incuba-
tion of microsomes in the presence of TAO or ketoconazole
resulted in marked inhibition (>70%) in the formation
of each DPC 963 GSH adduct M3-M6. In contrast, the
inhibitors for other P450 enzymes [furafylline (P450 1A2,
10 µM), D609 (P450 2B1/2, 10 µM), and sulfaphenazole

396 Chem. Res. Toxicol., Vol. 15, No. 3, 2002
Chen et al.
(P450 2C9, 10 µM)] had no effect on the production of
these GSH conjugates. In the absence of enzyme kinetic
data and the inability to quantitate each of the GSH
adducts accurately, it was impossible to obtain the exact
contributions each P450 made toward the formation of
the reactive intermediates. However, the studies con-
ducted with cDNA-expressed P450s, chemical inhibitors,
and P450 antibodies suggested that P450 3A1/3A2 played
a dominant role in forming these metabolites.
intermediate, which was responsible for the production
of two isomeric GSH conjugates, M3 and M4 (Scheme
1). Such a formation of an oxirene intermediate has been
postulated and reported before with the close analogue
DPC 961 (5). As shown in Scheme 1, it is postulated that
an initial oxidation of the triple bond, catalyzed by
cytochrome P450, results in the formation of a putative
oxirene intermediate which undergoes rapid rearrange-
ment with the adjacent cyclopropyl ring to form a reactive
cyclobutenyl ketone intermediate. The rearrangement to
the cyclobutenyl ketone could take place via the keto-
carbene a or through the intermediate b (5). The GSH
conjugation with the intermediate occurs via 1,4-Michael
addition to produce two isomeric GSH adducts. Such a
ring expansion of the cyclopropane via metabolic activa-
tion of an adjacent triple bond was observed with DPC
961 and was absent in efavirenz-dosed animals. The
ability of DPC 963 to undergo an oxidation of the triple
bond perhaps depends on how it is docked at the active
site on P450. The hydrogen bonding provided by the
cyclourea moiety of DPC 961 and DPC 963 may make it
favorable for the triple bond to undergo oxidation.
Protein-substrate hydrogen bonding can dictate a cata-
lytically favorable orientation of substrate within the
P450 active site (11). As discussed previously (5, and
references cited therein), the postulated occurrence of
oxirene as an intermediate during chemical and enzy-
matic oxidation of alkynes has been previously reported.
However, prior to these studies there was no direct
evidence for the existence of such oxirenes mainly due
to the high reactivity of these intermediates. By intro-
ducing a cyclopropyl substituent, which has been widely
used to probe enzymatic mechanisms (12), and by using
Discu ssion
The structural differences between DPC 963 and its
analogues, DPC 961 and efavirenz, are shown in Figure
1. The metabolism of DPC 961 and efavirenz has been
described previously (1, 2, 5). Compared to efavirenz,
which has a cyclocarbamate fused to a chlorobenzene,
DPC 963 consists of a cyclourea coupled to a difluoroben-
zene. DPC 963 and DPC 961 are both cycloureas, except
the latter is fused to a chlorobenzene (see Figure 1). In
our previous studies with the structural analogues
efavirenz and DPC 961 (Figure 1), a number of unusual
GSH adducts were found in the bile of rats (2, 5). Due to
slight structural differences between DPC 963 and these
two compounds, it was of interest to see if DPC 963 was
metabolized to produce similar GSH conjugates.
DPC 963 was metabolized via similar metabolic path-
ways as those described previously for DPC 961 and
efavirenz. The glucuronide and sulfate conjugates of the
aromatic ring hydroxylated metabolite were the major
metabolites found in the urine of various species (6). In
addition to the glucuronide and sulfate conjugates, a
number of glutathione-derived adducts were also found
in the bile and urine of rats dosed with DPC 963. The
structures of these GSH adducts, isolated from rat bile
and urine, were determined by LC/MS/MS and LC/NMR.
Characterization of these GSH conjugates, which pro-
vided valuable insight into the identity of the reactive
intermediates from which they were derived, revealed at
least three distinct metabolic pathways leading to these
adducts.
Analysis of bile from rats dosed with DPC 963 showed
that one of the GSH adducts was similar to that produced
by efavirenz. Consistent with the observations made with
efavirenz glutathione adduct (2), this was truly a dicon-
jugate possessing both a sulfate and GSH moieties.
Significant quantities of the corresponding cysteinyl-
glycine conjugate were also present in the urine of rats
dosed with DPC 963. The formation of this GSH conju-
gate (M1) of dihydroxylated DPC 963 appeared to follow
a similar metabolic pathway to efavirenz (Scheme 1). The
compound was hydroxylated on the methine carbon of
the cyclopropyl ring before glutathione S-transferase-
catalyzed addition of GSH across the triple bond could
take place, although the exact configuration of the
substituted product remains unknown. As with efavirenz,
the addition of GSH to the triple bond of the cyclo-
propanol derivative of DPC 963 was catalyzed by a
glutathione transferase, and occurred in a species-specific
manner only in rats and did not take place in cynomolgus
monkeys and humans (6). The structure of metabolite M1
indicated the existence of multiple steps that involved
both phase I and phase II biotransforming enzymes prior
to the formation of the final product.
a
13C analogue, the transient formation of the oxirene
intermediate was demonstrated via 13C NMR studies and
by identifying GSH adducts that were derived from the
rearranged intermediate, a cyclobutenyl ketone (5).
In addition to these GSH adducts formed by two
distinct metabolic pathways, structures of additional
GSH conjugates, M5 and M6 (Scheme 1), found in rat
bile pointed to the existence of yet another metabolic
pathway. These were characterized as being ring-
substituted regioisomers with only one fluorine remain-
ing at the aromatic C5 position. It was also found that
glucuronide and sulfate conjugates of the 6-hydroxy
derivative (formed after defluorination at this position)
were present in the bile of rats. It appears that the
defluorination of DPC 963 took place preferentially at the
aromatic C6 position. It is postulated that these GSH
adducts (M5 and M6) are a consequence of a putative
p-benzoquinone imine intermediate produced via P450-
mediated oxidative defluorination of DPC 963 (Scheme
1). Halogen-containing compounds are considered rela-
tively harmless because of their chemical stability.
However, removal of halogen substituents is thought to
be a crucial step in the bioactivation to products ulti-
mately responsible for the toxic effects. Various P450
enzymes play a pivotal role in this metabolic activation
(13-16). DPC 963 was found to be defluorinated both in
vivo and in vitro, resulting in products that could
potentially lead to toxicities. This observation has im-
plications in the search for new chemical entities for
pharmaceutical development, since incorporation of a
fluorine substituent in a drug has traditionally been
viewed as a means of slowing down the metabolism or
preventing bioactivation of compounds (17-23). More
It appears that the oxidation of the DPC 963 triple
bond took place to form an unstable oxirene metabolic

GSH Adducts of Oxirene and Benzoquinone Imine of DPC 963
Chem. Res. Toxicol., Vol. 15, No. 3, 2002 397
Sch em e 2. P ostu la ted Mech a n ism s of Cytoch r om e P 450-Med ia ted Oxid a tive Deflu or in a tion of DP C 963
Lea d in g to th e F or m a tion of p-Ben zoqu in on e Im in e In ter m ed ia te
importantly, benzoquinone imine, being a “soft” electro-
phile with high reactivity, may covalently interact with
cellular macromolecules and could elicit oxidative stress
through redox cycling.
can be explained due to the possible reduction of benzo-
quinone imine with the two electrons coming from
NADPH via P450 reductase. Since P450 can effectively
transfer electrons from its reductase to many substrates
with high enough oxidation potentials (24-26), it is not
unreasonable to expect that the P450 reaction intermedi-
ate might also accept electrons in this fashion. The
benzoquinone imine intermediate is an oxidant and a
highly reactive electrophile. Nevertheless, in the presence
of NAD(P)H or other reducing equivalents, the benzo-
quinone imine can be reduced to its nonreactive hydro-
quinone form.
One of the postulated mechanisms of DPC 963 de-
fluorination involves an abstraction of a single electron
by the active ferryl species of cytochrome P450, leading
to a nitrogen-centered radical intermediate (Scheme 2).
Delocalization of the unpaired electron on the nitrogen
to the aromatic carbons affords the C6-centered radical.
Recombination of the carbon radical with the iron-bound
hydroxyl leads to an unstable “σ-complex” that loses a
proton and fluoride anion, giving rise to the benzoquinone
imine. A similar mechanism had been previously postu-
lated for bioactivation of para-fluorinated anilines (16).
The location of fluorine at the C6 position (para to the
nitrogen of the cyclourea) on DPC 963 bears resemblance
to these para-fluorinated anilines. Recent studies on the
P450-catalyzed conversion of para-fluorinated anilines or
phenols have indicated the direct formation of benzo-
quinone imines or benzoquinones, accompanied by loss
It appears that the p-benzoquinone imine derivative
represents a likely candidate for the reactive intermedi-
ate, based on the structure of the GSH conjugates. The
possibility of cytochrome P450-mediated formation of the
p-benzoquinone imine metabolic intermediate from DPC
963 was demonstrated by rat liver microsomal incuba-
tions. Trapping the reactive intermediate with mi-
crosomes fortified with GSH produced two GSH conju-
gates identical to those found in vivo. The proposed
metabolic pathway leading to the formation of these GSH
adducts is shown in Scheme 1. Rat P450 3A1/3A2
enzymes catalyze the oxidative defluorination of DPC 963
to directly form the reactive p-benzoquinone imine in-
termediate. GSH adds to this intermediate via Michael
addition, giving rise to two regioisomeric adducts, M5 and
M6. Metabolite M5 was expected to form in greater
quantities, since the anion resulting from GSH addition
enjoys greater resonance stabilization. Indeed, it was
found that M5 was present in relatively greater quanti-
ties than M6, both in vivo and in vitro. It was also found
that the addition of GSH to the benzoquinone imine
intermediate took place nonenzymatically since both
conjugates were formed with cDNA-expressed enzymes
that apparently did not have any glutathione transferase
activities. The presence of a 6-hydroxy derivative (M7)

398 Chem. Res. Toxicol., Vol. 15, No. 3, 2002
Chen et al.
of a halogen anion upon the oxidation of these compounds
(16, 27, 28). Alternatively, the P450 (FeO)3+, a high-
valent iron-oxo intermediate, adds to the pi electron
system of the aromatic ring at C6 directly and forms a σ
adduct with a delocalized unpaired electron at C7 and
other positions (Scheme 2). Electron transfer within this
σ adduct from the aromatic ring to the porphyrin-iron
converts it to a σ cation adduct with a positive charge
delocalized in the ring (see Scheme 2). Rearrangements
of the bonds in this σ cation adduct may create the
benzoquinone and the halogen anion. This postulated
mechanism is similar to the one proposed previously for
the metabolism of hexahalogenated benzenes (15).
Recently, a GSH conjugate of a COX-2 inhibitor pos-
sessing a fluorobenzene ring was described (29). The
origin of this GSH conjugate was postulated to occur via
an epoxide, which led to the formation of only one GSH
adduct. However, with DPC 963, two GSH adducts were
formed, indicating the involvement of a benzoquinone
imine rather than an epoxide intermediate in the de-
fluorination of the aromatic ring.
The P450 involved in the formation of oxirene and
benzoquinone imine intermediate was investigated by
using differently induced microsomes, anti-rat P450
antibodies, and selective chemical inhibitors, and by
employing cDNA-expressed isozymes. It was demon-
strated that dexamethasone-induced microsomes pro-
duced much higher levels of GSH conjugates M3-M6
than the noninduced microsomes. These results sug-
gested the involvement of P450 3A enzymes in catalyzing
the metabolic reactions capable of producing the GSH
adducts. It is known that CYP 3A2 is constitutively
expressed in control male rat liver microsomes, making
up approximately 5% of the total hepatic P450, while
cytochrome CYP 3A1 is not constitutively expressed in
either male or female rat liver microsomes (30). A
significant induction of these enzymes is observed in rats
treated with dexamethasone (30). The higher levels of
GSH adducts produced by dexamethasone-induced rat
liver microsomes further supported the involvement of
CYP3A1/3A2 in forming reactive intermediates from DPC
963. Results from studies conducted with anti-rat P450
antibodies, chemical inhibitors, and cDNA-expressed
P450s confirmed that CYP 3A1/3A2 played a major role
in forming these reactive precursors to the GSH adducts
M3-M6.
Overall, the metabolism of DPC 963 was similar to that
of efavirenz in being principally converted to the glu-
curonide or sulfate conjugate of the ring-hydroxylated
metabolite. However, changes in the fused cyclic ring of
DPC 963, whereby the cyclocarbamate fused with a
chlorobenzene was replaced with a cyclourea coupled
with a difluorobenzene, led to significant differences in
the metabolic activation between these two compounds.
The most notable difference was the nature of GSH
adducts produced by the two compounds. In rats, an
initial hydroxylation on the cyclopropyl ring of efavirenz
was found to be a prerequisite for enzymatic addition of
GSH across the triple bond. However, in addition to the
GSH conjugate described above, rats were able to directly
oxidize the triple bond of DPC 963, resulting in an
intermediate that was capable of 1,4-Michael addition
with GSH. Another novel metabolic reaction observed in
rats was the oxidative defluorination which produced a
chemically reactive p-benzoquinone imine intermediate
capable of reacting with a nucleophile such as GSH. The
direct triple bond oxidation and oxidative defluorination
of DPC 963 was found to be specifically catalyzed by rat
P450 3A enzymes.
Refer en ces
(1) Mutlib, A. E., Chen, H., Nemeth, G., Markwalder, J ., Seitz, S.,
Gan, L.-S., and Christ, D. D. (1999) Identification and charac-
terization of efavirenz metabolites by LC/MS and high field
NMR: Species differences in the metabolism of efavirenz. Drug
Metab. Dispos. 27, 1319-1333.
(2) Mutlib, A. E., Chen, H., Nemeth, G., Gan, L.-S., and Christ, D.
D. (1999) LC/MS and high field NMR characterization of novel
mixed diconjugates of the nonnucleoside HIV-1 reverse tran-
scriptase inhibitor, efavirenz. Drug Metab. Dispos. 27, 1045-1056.
(3) Gerson, R. J ., Mutlib, A. E., Meunier, P. C., Haley, P. J ., Gan,
L.-S., Chen, H., Davis, M. H., Gemzik, B., Christ, D. D., Krahn,
D. F., Markwalder, J . A., Seitz, S. P., Miwa, G. T., and Robertson,
R. T. (1999) Species-specific nephrotoxicity induced by glutathione
conjugation of efavirenz in rats. Toxicol. Sci. 48 (1-S), 1833
(Society of Toxicology Meeting, New Orleans, LA).
(4) Mutlib, A. E., Gerson, R. J ., Meunier, P. C., Haley, P. J ., Gan,
L.-S., Chen, H., Davis, M. H., Gemzik, B., Christ, D. D., Krahn,
D. F., Markwalder, J . A., Seitz, S. P., Robertson, R. T., and Miwa,
G. T. (2000) Species-dependent metabolism of efavirenz produces
a nephrotoxic glutathione conjugate in rats. Toxicol. Appl. Phar-
macol. 169, 102-113.
(5) Mutlib, A. E., Chen, H., Shockcor, J ., Espina, R., Chen, Sy., Cao,
K., Du, A., Nemeth, G., Prakash, S., and Gan, L.-S. (2000)
Characterization of novel glutathione adducts of a non-nucleoside
reverse transcriptase inhibitor, (S)-6-chloro-4-(cyclopropyleth-
ynyl)-4-(trifluoromethyl)-3,4-dihydro-2-(1H)-quinazolinone (DPC
961), in rats. Possible formation of an oxirene metabolic inter-
mediate from a disubstituted alkyne. Chem. Res. Toxicol. 13, 775-
784.
(6) Chen, H., Shockcor, J ., Espina, R., Gan, L.-S., and Mutlib, A. E.
(1999) Formation of novel glutathione adducts of DPC 963 in rats
via multiple metabolic pathways. ISSX Proc. 15, 105.
(7) Corbett, J . W., and Ko, S. S. (1998) Preparation of 4,4-disubsti-
tuted-3,4-dihydro-2-(1H)-quinazolinones as HIV reverse tran-
scriptase inhibitors. Chem. Abstr. 129, 290148.
(8) Yanev, S., Kent, U. M., Pandova, B., and Hollenberg, P. (1999)
Selective mechanism-based inactivation of cytochrome P450 2B1
and P450 2B6 by a series of xanthates. Drug Metab. Dispos. 27,
600-604.
(9) Newton, D. J ., Wang, R. W., and Lu, A. Y. H. (1994) Cytochrome
P450 inhibitors. Evaluation of specificities in the vitro metabolism
of therapeutic agents by human liver microsomes. Drug Metab.
Dispos. 23, 154-158.
(10) Baillie, T. A., and Davis, M. R. (1993) Mass spectrometry in the
analysis of glutathione conjugates. Biol. Mass Spectrom. 22, 319-
325.
(11) Raag, R., and Poulos, T. L. (1991) Crystal structures of cytochrome
P-450cam complexed with camphane, thiocamphor and adaman-
tane: factors controlling P-450 substrate hydroxylation. Biochem-
istry 30, 2674-2684.
(12) Suckling, C. J . (1988) The cyclopropyl group in studies of enzyme
mechanism and inhibition. Angew. Chem., Int. Ed. Engl. 27, 537-
552.
(13) MacDonald, T. L. (1983) Chemical mechanisms of halocarbon
metabolism. CRC Crit. Rev. Toxicol. 11, 85-120.
(14) Den Besten, C., Van Bladeran, P. J ., Duizer, E., Vervoort, J ., and
Rietjens, I. M. C. M. (1993) Cytochrome P450-mediated oxidation
of pentafluorophenol to tetrafluorobenzoquinone as the primary
reaction product. Chem. Res. Toxicol. 6, 674-680.
(15) Rietjens, I. M. C. M., and Vervoort, J . (1992) A new hypothesis
for the mechanism for cytochrome P450 dependent aerobic
conversion of hexahalogenated benzenes to pentahalogenated
phenols. Chem. Res. Toxicol. 5, 10-19.
(16) Rietjens, I. M. C. M., and Vervoot J . (1991) Bioactivation of
4-fluorinated anilines to benzoquinoneimines as primary reaction
products. Chem.-Biol. Interact. 77, 263-281.
(17) Hecht, S. S., LaVoie, E. J ., Bedenko, V., Pingaro, L., Katayama,
S., Hoffmann, D., Sardella, D. J ., Boger, E., and Lehr, R. E. (1981)
Reduction of tumorigenicity and of dihydrodiol formation by
fluorine substitution in the angular rings of dibenzo(a, i)pyrene.
Cancer Res. 41, 4341-4345.
(18) Hey, M. M., Haaf, H., McLachlan, J . A., and Metzler, M. (1986)
Indirect evidence for the metabolic dehalogenation of tetra-
fluorodiethylstilbestrol by rat and hamster liver and kidney
microsomes. Biochem. Pharmacol. 34, 2135-2139.

GSH Adducts of Oxirene and Benzoquinone Imine of DPC 963
Chem. Res. Toxicol., Vol. 15, No. 3, 2002 399
(19) Liehr, J . G. (1984) Modulation of estrogen-induced carcinogenesis
by chemical modification. Arch. Toxicol. 55, 119-122.
(20) Liehr, J . G. (1983) 2-Fluoroestradiol: separation of estrogenicity
from carcinogenicity. Mol. Pharmacol. 23, 278-281.
(26) Zbaida, S., Brewer, C. F., and Levine, W. G. (1994) Hepatic
microsomal azoreductase activity-reactivity of azo dye substrates
is determined by their electron densities and redox potentials.
Drug Metab. Dispos. 22, 412-418.
(21) Morgan, P., Maggs, J . L., Page, P. C. B., and Park, B. K. (1992)
Oxidative dehalogenation of 2-fluoro-17R-ethynyloestrogenicity in
vivo. Biochem. Pharmacol. 44, 1717-1724.
(22) Oravec, C. T., Daniel, F. B., and Wong, L. K. (1983) Comparative
metabolism of 7, 12-dimethylbenz[a]anthracene and its non-
carcinogenic 2-flouro analogue by syrian hamster embryo cells.
Cancer Lett. 21, 43-55.
(23) Scribner, J . D., Scribner, N. K., and Koponen, G. (1982) Metabo-
lism and nucleic acid binding of 7-fluoro-2-acetamidofluorene in
rats: oxidative defluorination and apparent dissociation from
hepatocarcinogenesis of 8-(N-arylamide)guanine adducts on DNA.
Chem.-Biol Interact. 40, 27-43.
(27) Rietjens, I. M. C. M., Tyrakowska, B., Veeger, C., and Vervoort,
J . (1990) Reaction Pathways for biodehalogenation of fluorinated
anilines. Eur. J . Biochem. 194, 945-954.
(28) den Besten, C., van Bladeren, P. J ., Duizer, E., Vervoort, J ., and
Rietjens, I. M. C. M. (1993) Cytochrome P450-mediated oxidation
of pentafluorophenol to tetrafluorobenzoquinone as the primary
reaction product. Chem. Res. Toxicol. 6, 674-680.
(29) Yergey, J . A., Trimble, L. A., Silva, J ., Chauret, N., Li, C., Therien,
M., Grimm, E., and Nicoll-Griffith, D. A. (2001) In vitro metabo-
lism of the cox-2 inhibitor DFU, including a novel glutathione
adduct rearomatization. Drug Metab. Dispos. 29, 638-644.
(24) Hanzlik, R. P. (1981) Reactivity and toxicity among halogenated
methanes and related compounds: a physicochemical correlate
with predictive value. Biochem. Pharmacol. 30, 3027-3030.
(25) Goeptar, A. R., Te Koppele, J . M., Van Maanen, J . M. S.,
Zoetemelk, C. E. M., and Vermeulen, N. P. E. (1992) One-electron
reductive bioactivation of 2,3,5,6-tetramethylbenzoquinone by
cytochrome P450. Biochem. Pharmacol. 43, 1133-1134.
(30) Cooper, K. O., Reik, L. M., J ayyosi, Z., Bandiera, S., Kelley, M.,
Ryan, D. E., Danniel, R., McCluskey, S. A., Levin, W., and
Thomas, P. E. (1993) Regulation of two members of the steroid-
inducible cytochrome P450 subfamily (3A) in rats. Arch Biochem.
Biophys. 301, 345-354.
TX010153F