Coumarin metabolism by rat esophageal microsomes and cytochrome P450 2A3

Article abstract of DOI:10.1021/tx010065v

The rat esophagus is strikingly sensitive to tumor induction by nitrosamines, and it has been hypothesized that this tissue contains cytochrome P450 enzymes (P450s) which catalyze the metabolic activation of these carcinogens. The metabolic capacity of the esophagus is not well characterized. In the study described here, the products of ¹⁴C-coumarin metabolism by rat esophageal microsomes were identified and quantified. Metabolite characterization was by LC/MS/MS and GC/MS and comparison to standards, quantification was by radioflow HPLC. The coumarin metabolites formed by rat esophageal microsomes were compared to those formed by P450 2A3. The major metabolites formed by esophageal microsomes were 8-Hydroxycoumarin, o-Hydroxyphenylacetaldehyde (o-HPA), and o-hydroxyphenylacetic acid (o-HPAA). A smaller amount of 5-hydroxycoumarin, about one-third the 8-hydroxycoumarin, was also formed. o-HPA and o-HPAA are products of coumarin 3,4-epoxidation. The relative rates of coumarin 8-Hydroxylation and 3,4-epoxidation were similar. Coumarin 8-hydroxylation has not previously been reported as a major pathway in any tissue, and no P450s have yet been reported to catalyze this reaction. P450 2A3 catalyzed both the 7-hydroxylation and 3,4-epoxidation of coumarin. P450 2A3 was previously characterized as a coumarin 7-hydroxylase, however, in this study, we report that it catalyzes the formation of o-HPA more efficiently. The K_m and V_max were 1.3 ± 0.35 μM and 0.65 ± 0.06 nmol/min/nmol P450 for coumarin 7-hydroxylation and 1.4 ± 0.58 μM and 3.1 ± 0.46 nmol/min/nmol P450 for o-HPA formation.

Full text of DOI:10.1021/tx010065v

1
386
Chem. Res. Toxicol. 2001, 14, 1386-1392
Cou m a r in Meta bolism by Ra t Esop h a gea l Micr osom es
a n d Cytoch r om e P 450 2A3
Linda B. von Weymarn and Sharon E. Murphy*
Department of Biochemistry, Molecular Biology & Biophysics, University of Minnesota
Cancer Center, Minneapolis, Minnesota 55455
Received March 23, 2001
The rat esophagus is strikingly sensitive to tumor induction by nitrosamines, and it has
been hypothesized that this tissue contains cytochrome P450 enzymes (P450s) which catalyze
the metabolic activation of these carcinogens. The metabolic capacity of the esophagus is not
1
4
well characterized. In the study described here, the products of C-coumarin metabolism by
rat esophageal microsomes were identified and quantified. Metabolite characterization was
by LC/MS/MS and GC/MS and comparison to standards, quantification was by radioflow HPLC.
The coumarin metabolites formed by rat esophageal microsomes were compared to those formed
by P450 2A3. The major metabolites formed by esophageal microsomes were 8-hydroxycou-
marin, o-hydroxyphenylacetaldehyde (o-HPA), and o-hydroxyphenylacetic acid (o-HPAA). A
smaller amount of 5-hydroxycoumarin, about one-third the 8-hydroxycoumarin, was also
formed. o-HPA and o-HPAA are products of coumarin 3,4-epoxidation. The relative rates of
coumarin 8-hydroxylation and 3,4-epoxidation were similar. Coumarin 8-hydroxylation has
not previously been reported as a major pathway in any tissue, and no P450s have yet been
reported to catalyze this reaction. P450 2A3 catalyzed both the 7-hydroxylation and 3,4-
epoxidation of coumarin. P450 2A3 was previously characterized as a coumarin 7-hydroxylase,
however, in this study, we report that it catalyzes the formation of o-HPA more efficiently.
The K
m
and Vmax were 1.3 ( 0.35 µM and 0.65 ( 0.06 nmol/min/nmol P450 for coumarin
7
-hydroxylation and 1.4 ( 0.58 µM and 3.1 ( 0.46 nmol/min/nmol P450 for o-HPA formation.
In tr od u ction
The rat esophagus is uniquely sensitive to tumor
In the past, we have suggested that P450 2A3 may be
an important catalyst of nitrosamine metabolism in the
rat esophagus (8). However, our more recent data do not
support this conclusion (11, 12).
induction by nitrosamines (1, 2). Most nitrosamines are
metabolically activated by P450-catalyzed R-hydroxyla-
tion (1). It has been hypothesized by us, and others, that
either a unique or relatively abundant P450 in the rat
esophagus is responsible for this activation and therefore
contributes to the sensitivity of the esophagus to nitro-
samine carcinogenesis (3-6). It is an ongoing goal of our
laboratory to characterize the metabolic capacity of the
rat esophagus and specifically to identify P450s present
in this tissue.
P450 2A3 is an extrahepatic P450. cDNA for this
enzyme was originally generated from rat lung mRNA
(14). However, the enzyme is much more abundant in
rat nasal mucosa (8, 15). Very low levels of P450 2A3
mRNA, about 60-fold less than those in the lung have
been detected in the rat esophagus (8). In addition,
antibodies to the closely related mouse P450s, 2A4 and
2
A5, detected an immunoreactive protein of the correct
molecular weight in rat esophageal microsomes. The
amount of this protein was similar to the amount
detected in rat lung (8). This is in contrast to the very
different levels of P450 2A3 mRNA in the esophagus and
the lung (8). P450 2A3 is an efficient coumarin 7-hy-
droxylase (11). However, rat esophageal microsomes do
not catalyze detectable levels of coumarin 7-hydroxylation
P450 enzymes present in the rat esophagus include
A1, 2A3, and the recently cloned P450, 2B21 (7-9). P450
B21 protein has not yet been either purified from the
1
2
rat or expressed in a heterologous system (9). Therefore,
its catalytic activities are unknown. P450 1A1 is rela-
tively abundant in the esophagus but it does not catalyze
metabolism of the potent esophageal carcinogen N-
(11). Therefore, the level of P450 2A3 in the rat esopha-
1
nitrosobenzylmethylamine (NBzMA) (10). In contrast,
gus must be quite low.
P450 2A3, which is believed to be the ortholog of the
human P450 2A6, is an efficient catalyst of both NBzMA
and N′-nitrosonornicotine (NNN) metabolism (11, 12).
NNN is a tobacco specific nitrosamine considered a likely
causative agent for esophageal cancer in smokers (13).
Despite the lack of coumarin 7-hydroxylase activity in
the rat esophagus, coumarin inhibits NBzMA metabolism
by rat esophageal microsomes (11). This inhibition is
significantly increased when the enzyme is incubated
with coumarin and NADPH prior to the addition of
NBzMA. Therefore, a coumarin metabolite formed by rat
esophageal microsomes is contributing to the observed
inhibition. The P450 that catalyzes coumarin metabolism
in the rat esophagus may also contribute to the catalysis
of NBzMA metabolism. In addition, characterizing cou-
marin metabolism by rat esophageal microsomes will
*
To whom correspondence should be addressed.
Abbreviations: NBzMA, N-nitrosobenzylmethylamine; NNN, N-
1
nitrosonornicotine; o-HPA, o-hydroxyphenylacetaldehyde; o-HPAA,
o-hydroxyphenylacetic acid; o-HPPA, o-hydroxyphenylpropionic acid;
APCI, atmospheric pressure chemical ionization; BSTFA, N,O-bis-
trimethylsilyl)trifluoroacetamide; TMCS, trimethylchlorosilane; TMS,
trimethylsilyl.
(
1
0.1021/tx010065v CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/13/2001

Coumarin, P450 2A3, and Rat Esophageal Metabolism
Chem. Res. Toxicol., Vol. 14, No. 10, 2001 1387
Sch em e 1. Cou m a r in Meta bolism
help to establish what P450s are present in this tissue.
Coumarin can be metabolized via many different
pathways (ref 16 and Scheme 1). Six hydroxylated
coumarins as well as o-HPA and o-HPAA, generated by
lactone ring cleavage, have been identified as products
of coumarin metabolism in various tissues and animal
species (16, 17). o-HPA is formed nonenzymatically from
the unstable 3,4-epoxide of coumarin and o-HPA can be
further metabolized to o-HPAA (18). In all the species
and tissues studied to date, the predominate pathway of
coumarin metabolism is either 7-hydroxylation or lactone
ring cleavage, i.e., 3,4-epoxidation (16, 17, 19, 20). Hy-
droxylated coumarins other than 7-hydroxycoumarin
have been detected as minor metabolites in both in vivo
and in vitro studies of coumarin metabolism in a number
of species. Little is know about the enzymes involved in
the formation of these products or their biological rel-
evance.
the metabolites formed to the products of P450 2A3-
catalyzed coumarin metabolism.
Ma ter ia ls a n d Meth od s
Ch em ica ls. Esophagi from male Fisher 344 rats, age 7-13
weeks, were purchased from Harlan Bioproducts (Indianapolis,
1
4
IN). [U- C-Benzyl]Coumarin (58 mCi/mmol) was purchased
from Amersham Pharmacia Biotech (Piscataway, NJ ). 4-Hy-
droxycoumarin, 2-hydroxyphenyl acetic acid, o-coumaric acid,
and 6,7-dihydroxycoumarin were purchased from Fluka Chemi-
cal Co. (Ronkonoma, NY), 3-hydroxycoumarin was from Indofine
Chemical Co. (Sommerville, NJ ), o-HPA was a gift from Dr. Lois
D. Lehman-McKeeman of Procter & Gamble (Cincinnati, OH),
and 5-hydroxycoumarin was a gift from Dr. Takashi Harayama
(Okayama University, J apan). BSTFA/TMCS (1%) was pur-
chased from Pierce (Rockford, IL). All other chemicals were
purchased from Sigma Chemical Co. (St. Louis, MO) or Aldrich
Chemical Co. (Milwaukee, WI).
In the rat, the predominant pathway of coumarin
metabolism is 3,4-epoxidation. The major urinary cou-
marin metabolite is o-HPAA, and the major hepatic
microsomal metabolite is o-HPA (17, 21). Both 3,4-
epoxycoumarin and o-HPA contribute to the hepatotox-
icity that is observed in rats administered coumarin (22).
We hypothesize that one of these metabolites may
contribute to the NADPH-dependent coumarin inhibition
of NBzMA metabolism by rat esophageal microsomes. In
this study, we characterized the complete metabolism of
coumarin by rat esophageal microsomes and compared
Syn th esis of 6-Hyd r oxycou m a r in a n d 8-Hyd r oxycou -
m a r in . 6-Hydroxy- and 8-hydroxycoumarin were synthesized
from 2,5-dihydroxybenzaldehyde or 2,3-dihydroxybenzaldehyde,
respectively, by the method of Harayama et al. (23). The
dihydroxybenzaldehyde (552 mg) was dissolved in 40 mL of
diethylaniline. Carbethoxymethylenetriphenylphosphoran (1.67
g) was added, and the solution was refluxed for 40 min. Upon
cooling, 40 mL of a 5% HCl solution was added and the mixture
was extracted with diethyl ether (three times). The ether layer
was extracted with a 5% NaOH solution. The aqueous layer was
acidified with 5% HCl and again extracted with diethyl ether
(three times). The ether was evaporated and the residue was

1
388 Chem. Res. Toxicol., Vol. 14, No. 10, 2001
von Weymarn and Murphy
dissolved in ethyl acetate. The crude mixture was purified by
preparative TLC on silica plates (Uniplate, Analtech, Newark,
DE) with a mobile phase of 50% hexane:50% ethyl acetate. The
products were further purified by HPLC system I. The identities
of the products were confirmed by LC/MS and 1H NMR (600
MHz). 8-hydroxycoumarin, APCI/MS [m/z, (relative abundance)]:
al. (24). The mobile phase consisted of 21% tetrahydrofuran,
0.2% formic acid in water (A) and 0.2% formic acid in water
(B). The metabolites were eluted with 32% A:68% B, for 5 min,
followed by a linear gradient to 46% A:54% B in 15 min and
then to 100% A over 10 min. The flow rate was 1 mL/min. In
system III, the metabolites were eluted isocratically with 79.2%
H O:19.8% methanol:1% acetic acid. Detection was by UV
2
absorbance at 254 nm radioactivity, quantified on a â-ram
radioflow detector (IN/US Systems, Inc., Tampa, FL).
GC/MS An a lysis. Rat esophageal microsomes samples to be
analyzed by GC/MS were prepared as described for the LC/MS
analyses. To obtain sufficient material to purify coumarin
metabolites by HPLC prior to GC/MS analysis, microsomes from
four different preparations were pooled (16 esophagi) and
incubated for 1 h with 100 µM coumarin in the presence of an
NADPH generating system. P450 2A3 (20 pmol/mL) was
incubated with 40 µM coumarin for 20 min as described above.
Both sample and control reactions were injected on HPLC
system I. Two fractions of the eluant were collected from all
reactions in duplicate. Fraction one contained o-HPA and
o-HPAA (9-13 min), and fraction two contained 7-hydroxy and
8-hydroxycoumarin (13-17 min). Methanol was removed from
each fraction by evaporation under a stream of nitrogen and
the coumarin metabolites were extracted with ethyl acetate
+
+
1
63 ([M + H] , 100), 135 ([M + H-CO] , 17), 119 ([M +
+
+
+
1
H-CO
2
] , 10), 107 ([C
): δ 10.14 (s, 1H, OH); 7.96 (d, J ) 9.3, 1H, 4-H);
.12-7.04 (m, 3H, 5,6,7-H); 6.41 (d, J ) 9.3, 1H, 3-H). 6-Hy-
7
H
7
O] , 60), 91 ([C
7
H
7
] , 23). H NMR
(
7
DMSO-d
6
droxycoumarin: APCI/MS/MS [m/z, (relative abundance)]: 163
[M + H] , 57), 135 ([M + H-CO] , 20), 119 ([M + H -CO
0), 107 ([C H O] , 100. H NMR (DMSO-d ): δ 9.69 (s, 1H,
+
+
+
(
1
2
] ,
+
1
7
7
6
OH); 7.93 (d, J ) 9, 1H, 4-H); 7.19 (d, J ) 9.6, 1H, 5-H); 7.00-
.95 (s + d, 2H, 7,8-H); 6.39 (d, J ) 9.6, 1H, 3-H). Proton
6
assignment was confirmed by heteronuclear multiple quantum
correlation (HMQC) experiments.
Cou m a r in Meta bolism by P 450 2A3 a n d Ra t Esop h a gea l
Micr osom es. Baculovirus (BV) expressed P450 2A3 was pro-
vided by Xinxin Ding (Wadsworth Center, Albany, NY) as a cell
lysate, and coumarin metabolism was analyzed as previously
described (11). Reactions were run under conditions where
product formation was linear with time and protein. The
1
4
products formed by the metabolism of [U- C benzyl]coumarin
0.5-10 µM) by P450 2A3 (2-10 pmol/mL) were analyzed by
(
2 4
(three times). The organic phase was dried with Na SO , and
one of three HPLC systems. Detection was by radioflow analysis,
and comparison to co-injected metabolite standards detected by
UV absorbance at 254 nm. To determine the kinetic parameters
of P450 2A3-catalyzed coumarin 7-hydroxylation and 3,4-
epoxidation (o-HPA formation), duplicate reactions were run at
five [U- C benzyl]coumarin concentrations (0.5, 1, 3, 5, and 10
µM). The concentration of P450 2A3 was 2 pmol/mL and the
incubation time 10 min. The metabolites were analyzed on
HPLC system I with radioflow detection. Two independent
determinations of the kinetic parameters were carried out.
the ethyl acetate was evaporated under nitrogen. The dry
sample was dissolved in 20 µL of BSTFA:1% TMCS and heated
at 65 °C for 30 min to generate TMS-derivatized metabolites.
The recovery of standards (listed below), carried through the
same procedure, was >95%. The derivatized samples were
1
4
analyzed on a Hewlett-Packard (Wilmington, DE) GC/MS
system consisting of a model G1530A gas chromatograph, with
a model 6890 autosampler interfaced with a model 5973 mass
selective detector operating in the electron impact ionization
mode. A 30 m × 0.25 mm DB-5ms column (J & W Scientific,
Folsom, CA) with 0.25 µm film thickness was used for separation
of the metabolites. The inlet temperature was 280 °C. The GC
oven temperature was held at 80 °C for 1 min, then increased
by 5 °C/min to 300 °C. The mass spectrometer (MS quad
temperature 150 °C, MS source temperature, 230 °C) was tuned
against pentafluorobenzoic acid and operated in the full scan
mode (m/z 100-400). The retention times of the derivatized
standards under these conditions were o-HPA (12.11 min),
o-HPAA (17.81 min), o-HPPA (20.19 min), 8-hydroxy- (20.69
min), 6-hydroxy- (23.16 min), 4-hydroxy- (23.39 min), 7-hydroxy-
coumarin (23.68 min), and 6,7-dihydroxycoumarin (28.52 min).
Rat esophageal microsomes were prepared as described
previously (3). Four rat esophagus were used for each prepara-
tion of microsomes. Coumarin metabolism was analyzed by
1
4
incubating microsomes (0.5-2 mg/mL) with [U- C benzyl]
coumarin (3-20 µM) for 30-60 min at 37 °C in 200 µL of 100
mM Tris buffer (pH 7.4) in the presence of an NADPH generat-
ing system (11). The reaction was terminated by the addition
of 20 µL 15% trichloroacetic acid. Samples, co-injected with
standards, were analyzed by reverse-phase HPLC (system I,
II, and III) with radioflow detection. Coumarin metabolism by
rat esophageal microsomes was analyzed for seven separate
microsomal preparations.
Kin etics a n d Sta tistica l An a lysis. K
m
and Vmax for cou-
marin metabolism by P450 2A3 were determined using the EZ-
fit 5 kinetics program from Perrella Scientific [Amherst, NH
LC/MS a n d LC/MS/MS An a lysis. Two esophageal microso-
mal preparations (4 esophagi each) were pooled for each
experiment. The analysis was run in duplicate on two different
pooled preparations. Coumarin metabolite samples were pre-
pared as described for the radioflow analysis experiments except
that nonradioactive coumarin was used at a concentration of
(25)]. This program uses a nonlinear regression method of curve
fitting and the Runs test of residuals to determine statistically
whether experimental data are randomly distributed around the
curve with 95% confidence. K
using the Z-test.
m
and Vmax values were compared
1
00 µM. Control reactions included reactions without NADPH,
reactions without coumarin, and reactions without NADPH that
were spiked with either 5-hydroxycoumarin or 8-hydroxycou-
marin. At the termination of the reactions, the samples were
centrifuged and 100 µL of the supernatant was analyzed directly
by either LC/MS or LC/MS/MS. HPLC system I was used. LC/
MS and LC/MS/MS was carried out on a Finnigan MAT model
LCQ Deca instrument, operated in the positive ion APCI mode,
interfaced with a Waters Alliance HPLC system. The APCI
settings were capillary temperature, 200 °C; vaporizer temper-
Resu lts
To characterize any coumarin metabolites formed by
rat esophageal microsomes and P450 2A3, 6-hydroxy and
8-hydroxycoumarin, which were not available, were
synthesized. Two HPLC systems (I and III) were devel-
oped to separate all coumarin metabolites. System I was
essentially that which we have used previously to analyze
2
ature, 600 °C. The auxiliary gas (N ) was applied at a pressure
of 39.4 psi. For the analysis of samples, the instrument was
tuned to 8-hydroxycoumarin.
7
-hydroxycoumarin (11). The acetic acid concentration
was increased from 0.2 to 1.0%, and the methanol
concentration was decreased to 29.7%. These relatively
minor, but critical, changes allowed most coumarin
metabolites to be separated by reverse-phase HPLC
Figure 1A). 8-Hydroxycoumarin eluted just prior to
7-hydroxycoumarin, and 6-hydroxycoumarin eluted be-
fore 8-hydroxycoumarin. However, 5-hydroxycoumarin
HP LC System s. Systems I and II used a Phenomenex
(
Torrance CA) Bondclone C18 column (3.9 × 300 mm, 10 µm),
and system III used a Phenomenex Luna C18 column (4.6 × 250
(
mm, 5 µm). In system I, the metabolites were eluted isocratically
with 69.3% H
2
O:29.7% methanol:1% acetic acid, and the flow
rate was 1 mL/min. System II was similar to that of Peters et

Coumarin, P450 2A3, and Rat Esophageal Metabolism
Chem. Res. Toxicol., Vol. 14, No. 10, 2001 1389
min coeluted with 7-hydroxycoumarin. The same two
metabolites were detected with two other HPLC systems,
system II used previously by Peters et al. (24) and system
III which separates 5-hydroxycoumarin from coumarin.
5
-Hydroxycoumarin was not detected as a product of
P450 2A3-catalyzed coumarin metabolism. The kinetic
parameters of P450 2A3-catalyzed coumarin metabolism
were determined. P450 2A3 catalyzed the formation of
o-HPA at five times the rate of coumarin 7-hydroxylation.
The K
nmol/min/nmol P450 for coumarin 7-hydroxylation and
.4 ( 0.58 µM and 3.1 ( 0.46 nmol/min/nmol P450 for
o-HPA formation. These two K values do not differ (z )
0.124, p ) 0.901).
m
and Vmax were 1.3 ( 0.35 µM and 0.65 ( 0.06
1
m
-
1
4
The products of [U- C benzyl]coumarin metabolism by
rat esophageal microsomes were also analyzed with
HPLC system I. The metabolites formed by esophageal
microsomes were quite different then those formed by
P450 2A3 (Figure 1, panels B and C). The major metabo-
R
lite peak coeluted with 8-hydroxycoumarin (t , 14 min).
Little, if any, 7-hydroxycoumarin was detected. Two
additional metabolite peaks, coeluting with o-HPAA (10
min) and o-HPA (11 min), were also detected. The ratio
of 8-hydroxycoumarin to the products of the 3,4-epoxide
pathway, o-HPA and o-HPAA, was 1.3 for the sample
illustrated in Figure 1C. Within the same microsomal
preparation, this ratio did not differ between duplicates
or change with coumarin concentration (1-20 µM).
However, this ratio did vary between esophageal mi-
crosomal preparations, ranging from 1:1 to 2:1. In cou-
marin metabolite analyses using HPLC system III, a
radioactive peak that coeluted with 5-hydroxycoumarin
was also identified as a product of rat esophageal
metabolism. In several independent analyses 5-hydroxy-
coumarin was formed at approximately one-third the rate
of 8-hydroxycoumarin. The rate of coumarin metabolism
increased 4-fold from 1-20 µM (data not shown). Kinetic
parameters were not determined since, unlike P450 2A3,
the enzyme(s) did not appear to saturate by 20 µM
coumarin.
F igu r e 1. HPLC analysis of coumarin and coumarin metabo-
lites. (A) Standards. (B) Products of P450 2A3 (2.1 pmol)
1
4
catalyzed [U- C benzyl]coumarin (3 µM) metabolism (incubation
time, 30 min). (C) Products of rat esophageal microsomes
1
4
catalyzed [U- C benzyl]coumarin (20 µM) metabolism. (Incuba-
To further characterize the hydroxylated coumarin
metabolites formed by rat esophageal microsomes, LC/
MS and LC/MS/MS analysis conditions were developed.
Using HPLC system I, LC/MS/MS analysis of all six
hydroxylated coumarins was carried out. The relative
abundance of the product ions obtained for m/z 163, the
tion time, 60 min). Abbreviations: 4-, 6-, 7-, 8-OH, 4-, 6-, 7-,
8
-hydroxycoumarin; 6,7-diol, 6,7-dihydroxycoumarin; o-CA, o-
coumarin acid; o-HPPA, o-hydroxyphenylpropionic acid, o-HPA,
o-hydroxyphenylacetaldehyde; o-HPAA, o-hydroxyphenylacetic
acid. HPLC system I was used, details of the analysis and
enzyme assays are as described in the Materials and Methods.
All analyses were carried out in duplicate. Note: the retention
times are slightly different in the three chromatograms due to
the age of the HPLC column.
+
(M + H) ion for each hydroxylated coumarin is presented
in Table 1. The major fragment for both 6- and 8-
hydroxycoumarin was m/z 135 loss of CO from the
molecular ion, and the second most abundant fragment
partially coeluted with coumarin in system I. To separate
all the hydroxylated coumarin metabolites from one
another and 5-hydroxycoumarin from coumarin, a second
isocratic system (III) was used in which the methanol
concentration was decreased to 19.8%. Good resolution
of o-HPA, o-HPAA, and 6,7-dihydroxycoumarin was not
obtained with system III.
was m/z 119 loss of CO
obtained for 5-hydroxy- and 7-hydroxycoumarin was m/z
19. Essentially no m/z 135 fragment was detected.
2
. In contrast, the major fragment
1
Neither m/z 135 nor m/z 119 was present in the product
ion spectra of either 3 or 4-hydroxycoumarin, the major
fragment for these two hydroxycoumarins was m/z 121.
Confirmation that the major coumarin metabolite
formed by rat esophageal microsomes was 8-hydroxycou-
marin was obtained by LC/MS and LC/MS/MS analysis.
When the products of coumarin metabolism by esoph-
ageal microsomes were analyzed by LC/MS, a single
metabolite peak with a molecular ion of m/z 163 was
detected. The retention time and fragmentation pattern
were identical to 8-hydroxycoumarin (data not shown).
This metabolite was further characterized by LC/MS/MS.
Rat esophageal microsomes were incubated with cou-
We previously identified P450 2A3 as an efficient
coumarin 7-hydroxylase. The analysis used was HPLC
with fluorescence detection (11). In the present study, the
total metabolism of coumarin by P450 2A3 was deter-
1
4
mined using [U- C benzyl]coumarin. The products of the
reaction were analyzed by radioflow HPLC (Figure 1B).
Two major metabolite peaks were detected. The larger
one coeluted with o-HPA (11.5 min), a product of cou-
marin 3,4-epoxidation. No o-HPAA, the oxidation product
of o-HPA, was detected. The metabolite eluting at 15.1

1
390 Chem. Res. Toxicol., Vol. 14, No. 10, 2001
von Weymarn and Murphy
Ta ble 1. Ma jor F r a gm en ts P r esen t in P r od u ct Ion Sp ectr a of m /z 163 th e M + H Ion of Hyd r oxyla ted Cou m a r in s
m/z (relative abundance) ion composition
1
35
121
(C7H5O2)H
119
(C8H6O)H
107
C7H7O
95
91
C7H7
+
+
+
+
+
+
metabolite
(C8H6O2)H
(C6H6O)H
3
4
5
6
7
8
-hydroxy
-hydroxy
-hydroxy
-hydroxy
-hydroxy
-hydroxy
100
100
15
20
100
60
100
95
20
100
100
14
28
7
55
marin, and the metabolites formed were analyzed by
HPLC with MS/MS detection selecting for m/z 163 the
molecular ion + 1 for monohydroxylated coumarins. The
chromatogram illustrated in Figure 2A was obtained by
monitoring only m/z 163 from the total daughter ion
spectra obtained. Two peaks were detected with retention
times of 13.2 and 17.9 min. The peak at 13.2 min coelutes
with the 8-hydroxycoumarin standard. The product ion
spectrum of m/z 163 for this metabolite (Figure 2B) is
essentially identical to that of 8-hydroxycoumarin (Table
much larger coumarin peak obscured this product. Simi-
larly, its coelution with coumarin significantly increased
the limit of detection for 5-hydroxycoumarin by LC/MS
under the conditions used. As noted above, when the
products of coumarin metabolism by rat esophageal
microsomes were analyzed with a different HPLC system
(III), a peak coeluting with 5-hydroxycoumarin was
detected by radioflow HPLC. It was formed at about one-
third the rate of 8-hydroxycoumarin (data not shown).
o-HPA and o-HPAA were not detectable by LC/MS or LC/
MS/MS with the conditions used to analyze the hydroxy-
lated coumarins.
GC/MS analysis was used to characterize o-HPA and
o-HPAA, as well as 7-hydroxy- and 8-hydroxycoumarin
formed by P450 2A3 and esophageal microsomes. Mass
spectral data and retention times of the silylated cou-
marin metabolite standards are listed in Table 2. The
major fragment for all the hydroxylated coumarins was
1
).
The retention time and product ion spectrum of the
metabolite peak at 17.9 min (Figure 2) is identical to
-hydroxycoumarin (Figure 2C, Table 1). The major
fragment present in the product ion spectra for this
metabolite was m/z 119 generated by the loss of CO
-Hydroxycoumarin was initially not detected as a prod-
5
2
.
5
uct of coumarin metabolism by rat esophageal mi-
crosomes using either radioflow HPLC or LC/MS analy-
sis. This metabolite coelutes with coumarin in HPLC
system I. Therefore, in the case of radioflow analysis, the
3
m/z 219, which corresponds to the loss of CH from the
TMS derivatized molecular ion, m/z 234. The molecular
ion m/z 208 was the most abundant ion in the mass
F igu r e 2. LC/MS/MS analysis of rat esophageal microsomes-catalyzed coumarin metabolism. Coumarin (100 µM) was incubated
0 min with rat esophageal microsomes and the products were analyzed by LC/MS/MS with selection for m/z 163, the protonated
6
molecular ion of the hydroxylated coumarins. (A) Chromatogram monitoring m/z 163. (B) Product ion spectra of the metabolite peak
eluting at 13.2 min. (C) Product ion spectra of the metabolite peak eluting at 17.9 min.

Coumarin, P450 2A3, and Rat Esophageal Metabolism
Chem. Res. Toxicol., Vol. 14, No. 10, 2001 1391
Ta ble 2. Ma jor F r a gm en ts P r esen t in EI Sp ectr a fr om
GC/MS An a lysis of TMS Der iva tized Cou m a r in
Meta bolite Sta n d a r d s
lation of coumarin. The mouse P450 2G1, like P450 2A3,
catalyzes both the formation of o-HPA and 7-hydroxy-
coumarin with a similar difference in efficiency for the
two reactions (19).
Rat liver microsomes catalyze the 3,4-epoxidation of
coumarin (21). However, what rat P450 catalyzes this
reaction has not been reported. Rat nasal mucosa mi-
crosomes metabolize coumarin to both 7-hydroxycou-
retention
m/z (percent relative
metabolite
time (min)
abundance)
4
6
7
8
-hydroxycoumarin
-hydroxycoumarin
-hydroxycoumarin
-hydroxycoumarin
23.39
23.16
23.68
20.69
12.11
17.81
234(45), 219(100), 206(45)
234(55), 219(100), 191(10)
234(90), 219(100), 191(20)
234(20), 219(100), 191(45)
208(100), 193(40), 175(30)
296(60), 253(80), 164(52),
marin and o-HPA (19). The K
lation by rat nasal microsomes is 1.8 µM, similar to the
m
K of P450 2A3 (11, 29). The ratio of o-HPA to 7-hy-
m
of coumarin 7-hydroxy-
o-HPA
o-HPAA
1
47(100)
droxycoumarin formation in the nasal mucosa was re-
ported to be about 2:1 (19), less than the 5:1 ratio for
P450 2A3. P450 2A3 is abundant in the nasal mucosa
(8, 15) and most likely contributes to coumarin metabo-
lism in this tissue.
spectra of silylated o-HPA. The fragment ions detected
were m/z 193 and m/z 165 corresponding to the sequential
loss of CH and CO. o-HPPA is derivatized twice and the
3
major fragment detected was m/z 147 most likely a TMS
dimer. A fragment ion of a TMS dimer is often detected
in GC/MS analysis of compounds which are silylated at
multiple sites (26). More unique ions in the o-HPAA
spectrum included the molecular ion m/z 296 and a
We previously reported that in contrast to both rat
nasal microsomes and P450 2A3, rat esophageal mi-
crosomes do not catalyze the 7-hydroxylation of coumarin
(11). In the present study, we confirmed this and identi-
fied the major coumarin metabolite formed by esophageal
microsomes as 8-hydroxycoumarin. This is the first report
of 8-hydroxycoumarin as a major metabolite of coumarin
in any tissue. However, 8-hydroxycoumarin may be one
of several unidentified hydroxylated coumarins formed
during coumarin metabolism by rat nasal microsomes
(19). All the hydroxylated coumarins in the rat nasal
mucosa were formed to a lesser extent than was o-HPA.
Rat esophageal microsomes catalyzed the formation of
o-HPA and its oxidation product, o-HPAA, i.e., coumarin
3,4-epoxidation, to almost the same extent as the 8-hy-
droxylation of coumarin. However, the relative rates of
3,4-epoxidation and 8-hydroxylation varied between dif-
ferent preparations of esophageal microsomes. There are
two possible explanations for the observed variation. One
is that different enzymes catalyze these reactions and
that one of these is denatured to varying extents during
the isolation of microsomes. We and others have previ-
ously reported significant variation in the metabolic
capacity of rat esophageal microsomal preparations (3,
30). The second possibility is that that some of the 3,4
epoxide which is formed reacts with protein instead of
decomposing to o-HPA and o-HPAA and that this is
variable between preparations of microsomes. This would
also result in the inaccurate quantitation of the 3,4-
epoxidation pathway. In preliminary experiments, we
were able to trap the epoxide with GSH and therefore to
measure 3,4-epoxidation as the formation of a GSH
adduct instead of the sum of o-HPA and o-HPAA (data
not shown). In those experiments, the rate of 3,4-
epoxidation rate did not increase to greater than the rate
of 8-hydroxylation.
3
fragment ion generated by loss of CH and CO, m/z 253.
To analyze the coumarin metabolites formed by P450
A3 and rat esophageal microsomes with GC/MS, two
2
fractions from each sample were collected following
HPLC analysis. The first fraction, eluting between 9 and
1
3 min (Figure 2), should contain o-HPA and o-HPPA,
and the second fraction, eluting from 9 to 13 min (Figure
), 7-hydroxy and 8-hydroxy coumarin. Samples were
2
extracted, derivatized with BSTFA/TMCS, then analyzed
by GC/MS and compared to standards. The two coumarin
metabolite fractions collected from P450 2A3 reactions
contained o-HPA and 7-hydroxycoumarin, respectively.
Whereas the first fraction collected from rat esophageal
microsomal samples contained both o-HPA and o-HPAA
and the second fraction 8-hydroxycoumarin. The spectra
for each of these metabolites were identical to standards.
Discu ssion
The efficiency of coumarin 7-hydroxylation by the rat
P450 2A3 is comparable to the closely related human and
mouse enzymes P450 2A6 and 2A5 (19, 27, 28). However,
clearly P450 2A3 does not have the specificity of these
enzymes (19). We report here, that P450 2A3 catalyzes
the 3,4-epoxidation of coumarin five times more ef-
ficiently than it does coumarin 7-hydroxylation. o-HPA,
the decomposition product of coumarin 3,4-epoxide (18)
is the major product of P450 2A3-catalyzed coumarin
metabolism (Figure 1B). Previously, we characterized
P450 2A3 as a coumarin 7-hydroxylase. In that study,
HPLC with fluorescence detection was used to quantify
7
-hydroxycoumarin (11). In the present study, the use
1
4
of C-coumarin allowed all the products of metabolism
to be detected. The kinetic parameters for coumarin
A single P450 may or may not catalyzes the 8-hydroxy-
lation and 3,4 epoxidation of coumarin by rat esophageal
microsomes. P450 1A1 is present in the rat esophagus
and may contribute to coumarin 3,4-epoxidation in this
tissue (7, 19). The small amount of P450 2A3 present in
the esophagus (8) could also contribute to this pathway.
Neither P450 1A1 nor 2A3 catalyzes the 8-hydroxylation
of coumarin (ref 19, Figure 1). Therefore, a second
enzyme would have to catalyze the 8-hydroxylation of
coumarin. One candidate for this reaction is the recently
cloned esophageal P450 2B21 (9). We are in the process
of expressing this enzyme in a baculovirus/Sf9 system
and will in the future determine if it is a catalyst of either
coumarin 8-hydroxylation or 3,4-epoxidation. P450 2B21
7
-hydroxylation from the two studies agree. The previ-
ously reported K was 1.7 ( 0.41 µM and Vmax, 1.7 (
.008 nmol/min/nmol P450 (11) and the values reported
m
0
here are 1.3 ( 0.35 µM and 0.65 ( 0.06 nmol/min/nmol
P450.
This is the first report of a P450 2A enzyme catalyzing
the 3,4-epoxidation of coumarin. However, several human
P450s have been reported to catalyze this reaction.
P450s, 1A1, 1A2, 2E1, and 3A4 all catalyze the formation
of o-HPA with reasonable efficiencies (19). The K values
m
range from 12 to 54 µM and Vmax from 2 to 7 nmol/min/
nmol P450. None of these P450s catalyzed the 7-hydroxy-

1
392 Chem. Res. Toxicol., Vol. 14, No. 10, 2001
von Weymarn and Murphy
is present in rat liver and a small amount of 8-hydroxy-
coumarin may be formed by rat liver slices (20).
The possibility exists that one enzyme may catalyze
both the 8-hydroxylation and 3,4-epoxidation of cou-
marin, analogous to the catalysis of both coumarin
(12) Murphy, S. E., Isaac, I. S., Ding, X., and McIntee, E. J . (2000)
Specificity of cytochrome P450 2A3-catalyzed R-hydroxylation of
N′-nitrosonornicotine enantiomers. Drug Metab. Dispos. 28,
1
263-1266.
(13) Hecht, S. S. (1998) Biochemistry, biology, and carcinogenicity of
tobacco-specific N-nitrosamines. Chem. Res. Toxicol. 11, 559-603.
(
14) Kimura, S., Kozak, C. A., and Gonzalez, F. J . (1989) Identification
of a novel P450 expressed in rat lung: cDNA cloning and
sequence, chromosome mapping and induction by 3-methyl-
cholanthrene. Biochemistry 28, 3798-3803.
7
-hydroxylation and 3,4-epoxidation by P450 2A3. How-
ever, the variation we observed in the ratio of 8-hydroxy-
lation to 3,4-epoxidation of coumarin between individual
esophageal preparations suggests that two distinct en-
zymes catalyzing this reaction in the esophagus. Future
studies will determine the identity of the enzyme or
enzymes that catalyze the 3,4-epoxidation and the 8-hy-
droxylation of coumarin in the esophagus and whether
these same enzymes are catalysts of nitrosamine me-
tabolism in this tissue.
(
15) Su, T., Sheng, J . S., Lipinska, T. W., and Ding, X. (1996)
Expression of CYP2A genes in rodent and human nasal mucosa.
Drug Metab. Dispos. 24, 884-890.
16) Pelkonen, O., Raunio, H., Rautio, A., Pasanen, M., and Lang, M.A.
(1997) The metabolism of coumarin. In Coumarins: Biology,
Applications and Mode of Action (O’Kennedy, R., and Thornes,
R. D., Eds.) pp 67-92, J ohn Wiley & Sons Ltd, New York.
17) Fentem, J . H., and Fry, J . R. (1993) Species differences in the
metabolism and hepatotoxicity of coumarin. Comp. Biochem.
Physiol. 104C, 1-8.
18) Born, S. L., Rodriguez, P. A., Eddy, C. L., and Lehmann-
McKeeman, L. D. (1997) Synthesis and reactivity of coumarin 3,4-
epoxide. Drug Metab. Dispos. 25, 1318-1323.
19) Zhuo, X., Gu, J ., Zhang, Q. Y., Spink, D. C., Kaminsky, L. S., and
Ding, X. (1999) Biotransformation of coumarin by rodent and
human cytochrome P-450: Metabolic basis of tissue-selective
toxicity in olfactory mucosa of rats and mice. J . Pharmacol. Exp.
Ther. 288, 463-471.
(
(
(
(
Ack n ow led gm en t. We thank Dr. Lois D. Lehman-
McKeeman for o-HPA, Takashi Harayama for 5-hydroxy-
coumarin, Xinxin Ding for baculovirus expressed P450
2
A3, Peter Villalta of the Cancer Center Analytical
Chemistry and Biomarker Core for LC/MS/MS analysis
and Ed McIntee for the NMR analyses. This research was
supported by a grant from the NCI, CA-74913.
(
20) Steensma, A., Beamand, J . A., Walters, D. G., Price, R. J ., and
Lake, B. G. (1994) Metabolism of coumarin and 7-ethoxycoumarin
by rat, mouse, guinea pig, cynomolgus monkey, and human
precision-cut liver slices. Xenobiotica 24, 893-907.
Refer en ces
(
1) Preussmann, R., and Stewart, B.W. (1984) N-nitroso carcinogens.
In Chemical Carcinogenesis (Searle, C. E., Ed.) pp 643-828, Am.
Chem. Soc. Monogr., Washington, DC.
(21) Lake, B. G., Osborne, D. J ., Walters, D. G., and Price, R. J . (1992)
Identification of o-hydroxyphenylacetaldehyde as a major me-
tabolite of coumarin in rat hepatic microsomes. Food Chem.
Toxicol. 30, 99-104.
(
2) Lijinsky, W. (1992) The Chemistry and Biology of N-Nitroso
Compounds, Cambridge University Press, Cambridge, England.
3) Murphy, S. E., and Spina, D. A. (1994) Evidence for a high affinity
enzyme in rat esophageal microsomes which R-hydroxylates N′-
nitrosonornicotine. Carcinogenesis 15, 2709-2713.
(
22) Born, S. L., Hu, J . K., and Lehman-McKeeman, L. D. (2000)
o-hydroxyphenylacetaldehyde is a hepatotoxic metabolite of cou-
marin. Drug Metab. Dispos. 28, 218-223.
(
(
23) Harayama, T., Katsuno, K., Nishioka, H., Fujii, M., Nishita, Y.,
Ishii, H., and Kanebo, Y. (1994) A convenient synthesis of a simple
coumarin from salicylaldyde, and wittig reagent (I): a synthesis
of methoxy- and hydroxycoumarins. Heterocycles 39, 613-622.
24) Peters, M. M., Walters, D. G., van, O. B., van, B. J ., and Lake, B.
G. (1991) Effect of inducers of cytochrome P-450 on the metabo-
(
4) Labuc, G. E., and Archer, C. (1982) Esophageal and hepatic
microsomal metabolism of N-nitrosomethylbenzylamine and N-
nitrosodimethylamine in the rat. Cancer Res. 42, 3181-3186.
5) Bartsch, H., Margison, G. P., Malaveille, C., Camus, A. M., Brun,
G., Margison, J . M., Kolar, G. F., and Wiessler, M. (1977) Some
aspects of metabolic activation of chemical carcinogens in relation
to their organ specificity. Arch. Toxicol. 39, 51-63.
(
(
14
lism of [3- C]coumarin by rat hepatic microsomes. Xenobiotica
2
1, 499-514.
(
6) von Hofe, E., Schmerold, I., Lijinsky, W., J eltsch, W., and
Kleihues, P. (1987) DNA methylation in rat tissues by a series of
homologous aliphatic nitrosamines ranging from N-nitrosodi-
methylamine to N-nitrosomethyldodecylamine. Carcinogenesis 8,
(
25) Perrella, F. W. (1988) EZ-FIT: a practical curve fitting micro-
computer program for the analysis of enzyme kinetics on IBM-
PC computers. Anal. Biochem. 174, 437-447.
(
26) Halket, J . M. (1997) Derivatives for Gas Chromatography-Mass
1
337-1341.
Spectrometry. In Handbook of Derivatives for Chromatography
(
(
(
7) Ahn, D., Putt, D., Kresty, L., Stoner, G. D., Fromm, D., and
Hollenberg, P. F. (1996) The effects of dietary ellagic acid on rat
hepatic and esophageal mucosal cytochromes P450 and phase II
enzymes. Carcinogenesis 17, 821-828.
8) Gopalakrishnan, G., Morse, M. A., Lu, J ., Weghorst, C. M.,
Sabourin, C. L. K., Stoner, G. D., and Murphy, S. E. (1999)
Expression of cytochrome P450 2A3 in rat esophagus: relevance
to N-nitrosobenzylmethylamine. Carcinogenesis 20, 885-891.
9) Brookman-Amissah, N., Mackay, A. G., and Swann, P. F. (2001)
Isolation and sequencing of the cDNA of a novel cytochrome P450
from rat oesophagus. Carcinogenesis 22, 155-160.
(
Blau, K., and Halket, J . M., Eds.) pp 305-306, J ohn Wiley &
Sons, New York.
(
27) Yamano, S., Tatsuno, J ., and Gonzalez, F. J . (1990) The CYP2A3
gene product catalyzes coumarin 7-hydroxylation in human liver
microsomes. Biochemistry 29, 1322-1329.
28) Felicia, N. D., Rekha, G. K., and Murphy, S. E. (2000) Charac-
terization of cytochrome P450 2A4 and 2A5 catalyzed 4-(meth-
ylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) metabolism. Arch.
Biochem. Biophys. 384, 418-424.
(
(
29) Patten, C. J ., Peterson, L. A., and Murphy, S. E. (1998) Evidence
for metabolic activation for N′-nitrosonornicotine and N-ni-
trosobenzylmethylamine by a rat nasal coumarin hydroxylase.
Drug Metab. Dispos. 26, 177-180.
(
10) Morse, M. A., Lu, J ., Stoner, G. D., Murphy, S. E., and Peterson,
L. A. (1999) Metabolism of N-nitrosobenzylmethylamine by hu-
man cytochrome P-450 enzymes. J . Toxicol. Environ. Health A
(
30) Huang, Q., Stoner, G., Resau, J ., Nickols, J ., and Mirvish, S. S.
5
8, 397-411.
(
1992) Metabolism of N-nitrosomethyl-n-amylamine by mi-
crosomes from human and rat esophagus. Cancer Res. 52,
547-3551.
(11) von Weymarn, L. B., Felicia, N. D., Ding, X., and Murphy, S. E.
(
1999) N-Nitrosobenzylmethylamine R-hydroxylation and cou-
3
marin 7-hydroxylation by rat esophageal microsomes and cyto-
chrome P450 2A3 and 2A6 enzymes. Chem. Res. Toxicol. 12,
1
254-1261.
TX010065V