Oxidative metabolism of the mammalian lignans enterolactone and enterodiol by rat, pig, and human liver microsomes

Article abstract of DOI:10.1021/jf9809176

Hepatic microsomes from aroclor-treated male Wistar rats biotransform enterolactone to 12 metabolites, six of which carry an additional hydroxy group at the aromatic and six at the aliphatic moiety according to HPLC/MS and GC/MS analysis. The aromatic hyd

Full text of DOI:10.1021/jf9809176

J. Agric. Food Chem. 1999, 47, 1071−1077
1071
Oxid a tive Meta bolism of th e Ma m m a lia n Lign a n s En ter ola cton e a n d
En ter od iol by Ra t, P ig, a n d Hu m a n Liver Micr osom es
Eric J acobs and Manfred Metzler*
Institute of Food Chemistry, Department of Chemistry, University of Karlsruhe,
GY-76128 Karlsruhe, Germany
Hepatic microsomes from aroclor-treated male Wistar rats biotransform enterolactone to 12
metabolites, six of which carry an additional hydroxy group at the aromatic and six at the aliphatic
moiety according to HPLC/MS and GC/MS analysis. The aromatic hydroxylation products were
identified with the help of synthesized reference compounds as enterolactone monohydroxylated in
the para position and in both ortho positions of the original phenolic hydroxy group of either aromatic
ring. The synthesis of the reference compounds and their spectroscopic characterization is described.
Enterodiol is metabolized by hepatic microsomes from aroclor-treated male rats to three aromatic
and four aliphatic monohydroxylated metabolites. Aromatic hydroxylation occurs in the para position
and the two ortho positions of the original phenolic hydroxy group. Most of the metabolites of
enterolactone and enterodiol were also formed with microsomes from uninduced rat, pig, and human
liver, suggesting that oxidative metabolism is a common feature in the disposition of these lignans
in the mammalian organism.
Keyw or d s: Mammalian lignans; enterolactone; enterodiol; microsomal metabolism
INTRODUCTION
The mammalian lignans enterolactone (ENL) and
enterodiol (END) are weakly estrogenic compounds
formed by intestinal bacteria from the plant lignans
matairesinol and secoisolariciresinol (Axelson and Set-
chell, 1981; Setchell et al., 1982), which are constituents
of flaxseed, whole-grain products, vegetables, and fruit
(Thompson et al., 1991). Once generated in the intestine,
ENL and END are absorbed and reach the liver where
F igu r e 1. Synthetic route for enterolactone (ENL) and
derivatives monohydroxylated at the aromatic rings.
they are conjugated mainly with glucuronic acid or
sulfate (Adlercreutz et al., 1995a) before entering the
circulation (Axelson and Setchell, 1981). ENL and END
effects of their metabolites. We have therefore studied
appear in various body fluids in humans and animals
the oxidative metabolism of ENL and END with rat,
(Morton et al., 1994; Atkinson et al., 1993; Adlercreutz
pig, and human liver microsomes.
et al., 1995b) and may reach up to 10 000-fold the
concentration of normal circulating endogenous estro-
gens (Rickard and Thompson, 1997).
MATERIALS AND METHODS
Epidemiological studies suggest that high levels of
lignans may protect against hormone-dependent can-
cers, especially breast cancer (Rose, 1992; Adlercreutz,
1995). This is supported by data from in vivo and in
vitro studies. Flaxseed was found to inhibit mammary
and colon carcinogenesis in rats (Serraino and Thomp-
son, 1992; J enab and Thompson, 1996). The lignans also
inhibited the growth of different human tumor cells in
culture (Sung et al., 1996). In addition, lignans compete
with estradiol for the type II estrogen receptor and
stimulate the synthesis of sex hormone binding globulin
(Adlercreutz et al., 1992). ENL inhibits placental aro-
matase (Adlercreutz et al., 1993; Wang et al., 1994).
These data imply that lignans may play an important
role in the prevention of cancer. However, little is known
about the biotransformation of these compounds and the
Syn t h esis of E n t er ola ct on e, E n t er od iol, a n d St a n d -
a r d s for Meta bolites. ENL was synthesized as depicted in
Figure 1 by a modification of the tandem addition procedure
described by Pelter et al. (1983). The starting materials were
obtained by standard methods (Enders et al., 1979; Seebach
and Corey, 1974; Na¨sman and Pensar, 1985; Feringa et al.,
1994). Briefly, 5.6 mmol of n-butyllithium (3.5 mL of a 1.6 M
solution in hexane) was added to a solution of 5 mmol of
3-benzyloxybenzaldehyde bis(phenylthio)acetal (2.07 g) in 10
mL of dry tetrahydrofurane (THF) at -40 °C. The mixture was
stirred for 1.5 h at this temperature and then cooled to -85
°C, and a solution of 5 mmol of 2-buten-4-olide (0.42 g) in 1.5
mL of dry THF was added. The mixture was stirred for another
2.5 h at -85 °C and thereafter treated dropwise with a solution
of 5 mmol of 3-benzyloxybenzylbromide (1.39 g) in 2.5 mL of
dry THF and 0.6 mL of N,N-dimethyl-2-imidazolidinone. The
reaction mixture was kept at -30 °C overnight and then
allowed to warm to room temperature. Subsequently, water
(20 mL) was added, the aqueous phase was extracted with
ethyl acetate (3 × 20 mL), and the extract was washed with
water and dried over magnesium sulfate. Chromatography on
silica gel with methylene chloride yielded 2.1 mmol (1.46 g,
* Corresponding author. Phone: +49-721-608-2132. Fax:
+49-721-608-7255. E-mail: Manfred.Metzler@chemie.uni-
karlsruhe.de.
10.1021/jf9809176 CCC: $18.00 © 1999 American Chemical Society
Published on Web 02/03/1999

1072 J. Agric. Food Chem., Vol. 47, No. 3, 1999
Jacobs and Metzler
42%) of an orange gum. The product was refluxed with Raney
nickel (made of 13.6 g of nickel-aluminum alloy) in absolute
ethanol (55 mL) for 5 h to yield 625 mg of trans-3,4-bis-(3-
hydroxybenzyl)-γ-butyrolactone (ENL; quantitative yield). Af-
ter crystallization from chloroform, the purity of ENL was
>99% according to GC/MS.
(t, 7′), 38.71 (t, 7), 42.46 (d, 8), 45.75 (d, 8′), 72.32 (t, 9), 113.86
(d), 115.01 (d), 115.92 (d), 116.31 (d), 118.07 (d), 120.43 (d),
125.72 (s), 130.03 (d), 140.70 (s), 149.03 (s), 150.11 (s), 157.43
(s), 181.18 (s, 9′).
trans-3-(2,3-Dihydroxybenzyl)-2-(3-hydroxybenzyl)-γ-butyro-
lactone. Mass spectrum (EI, 70 eV) m/z (% relative abun-
dance): 530 (100, M+), 515 (25), 351 (4), 335 (16), 268 (45,
McLafferty rearrangement product), 179 (12, benzylic frag-
mentation), 73 (9). 1H NMR (MeOD) δ (ppm): 2.46-2.98 (6H,
m, 7, 7′, 8, 8′), 3.92 (1H, dd, 9A), 4.07 (1H, dd, 9B), 6.48-7.02
(7H, m). 13C NMR (MeOD) δ (ppm): 33.95 (t, 7′), 35.04 (t, 7),
42.78 (d, 8), 46.82 (d, 8′), 71.92 (t, 9), 112.88 (d), 114.63 (d),
120.15 (d), 120.34 (d), 122.40 (d), 125.35 (d), 130.46 (d), 145.00
(s), 146.37 (s), 153.02 (s), 153.44 (s), 157.21 (s), 180.36 (s, 9′).
trans-3-(3,4-Dihydroxybenzyl)-2-(3-hydroxybenzyl)-γ-butyro-
lactone. Mass spectrum (EI, 70 eV) m/z (% relative abun-
dance): 530 (85, M+), 515 (4), 293 (21), 268 (100, McLafferty
rearrangement product), 179 (16, benzylic fragmentation), 73
(31). 1H NMR (MeOD) δ (ppm): 2.24-2.70 (4H, m, 7, 8, 8′),
2.79-2,99 (2H, m, 7′), 3.87 (1H, dd, 9A), 4.06 (1H, dd, 9B),
6.37-7.14 (7H, m). 13C NMR (MeOD) δ (ppm): 35.81 (t, 7′),
38.61 (t, 7), 42.90 (d, 8), 47.54 (d, 8′), 72.84 (t, 9), 114.76 (d),
116.49 (d), 116.84 (d), 117.31 (d), 121.06 (d), 121.76 (d), 130.62
(d), 131.45 (s), 140.94 (s), 145.01 (s), 146.32 (s), 158.71 (s),
181.44 (s, 9′).
trans-3-(2,5-Dihydroxybenzyl)-2-(3-hydroxybenzyl)-γ-butyro-
lactone. Mass spectrum (EI, 70 eV) m/z (% relative abun-
dance): 530 (100, M+), 515 (5), 335 (63), 293 (7), 268 (64,
McLafferty rearrangement product), 180 (10, McLafferty re-
arrangement product), 73 (9). 1H NMR (CDCl3/MeOD) δ
(ppm): 2.55-2.95 (6H, m, 7, 7′, 8, 8′), 3.97 (1H, dd, 9B), 4.11
(1H, dd, 9A), 6.43-6.84 (6H, m), 7.10-7.14 (1H, m). 13C NMR
(CDCl3/MeOD) δ (ppm): 32.98 (t, 7′), 35.08 (t, 7), 40.54 (d, 8),
47.08 (d, 8′), 72.48 (t, 9), 114.31 (d), 114.79 (d), 116.42 (d),
116.97 (d), 117.74 (d), 121.63 (d), 126.97 (s), 130.14 (d), 139.86
(s), 148.70 (s), 150.09 (s), 157.34 (s), 180.99 (s, 9′).
For the synthesis of 2,3-bis-(3-hydroxybenzyl)-butane-1,4-
diol (END) according to Groen and Leemhuis (1980), 0.77
mmol (230 mg) of ENL dissolved in 5 mL of dry THF was
added to 5.7 mmol (216 mg) of lithium aluminum hydride in
30 mL of dry THF and stirred for 1 h at room temperature
before the mixture was refluxed for 2.5 h. The reduction gave
END in quantitative yield with a purity >99% according to
GC/MS.
The same procedure was used at a 1 mg scale for the
chemical reduction of trans-2-(2,3-dihydroxybenzyl)-3-(3-hy-
droxybenzyl)-γ-butyrolactone, trans-3-(3,4-dihydroxybenzyl)-
2-(3-hydroxybenzyl)-γ-butyrolactone, and trans-2-(2,5-dihy-
droxybenzyl)-3-(3-hydroxybenzyl)-γ-butyrolactone to obtain
2-(2,3-dihydroxybenzyl)-3-(3-hydroxybenzyl)-butane-1,4-diol,
2-(3,4-dihydroxybenzyl)-3-(3-hydroxybenzyl)-butane-1,4-diol, and
2-(2,5-dihydroxybenzyl)-3-(3-hydroxybenzyl)-butane-1,4-diol, re-
spectively.
Seven monohydroxylation products of ENL, viz., trans-2-
(2,3-dihydroxybenzyl)-3-(3-hydroxybenzyl)-γ-butyrolactone (yield
15%, purity 97%), trans-2-(3,4-dihydroxybenzyl)-3-(3-hydroxy-
benzyl)-γ-butyrolactone (yield 21%, purity 96%), trans-2-(3,5-
dihydroxybenzyl)-3-(3-hydroxybenzyl)-γ-butyrolactone (yield
15%, purity 97%), trans-2-(2,5-dihydroxybenzyl)-3-(3-hydroxy-
benzyl)-γ-butyrolactone (yield 29%, purity 98%), trans-3-(2,3-
dihydroxybenzyl)-2-(3-hydroxybenzyl)-γ-butyrolactone (yield
7%, purity 58%), trans-3-(3,4-dihydroxybenzyl)-2-(3-hydroxy-
benzyl)-γ-butyrolactone (yield 23%, purity 96%), and trans-3-
(2,5-dihydroxybenzyl)-2-(3-hydroxybenzyl)-γ-butyrolactone (yield
21%, purity 94%) were synthesized in the same manner
(Figure 1) from the respective bis(phenylthio)acetals and
benzyl bromides, which were prepared from the respective
benzylated hydroxybenzaldehydes purchased from Fluka (Dei-
senhofen, Germany), Lancaster (Mu¨hlheim, Germany), or
Aldrich (Steinheim, Germany). These and the other chemicals
used were of the highest purity available.
1
The synthesis and characterization of the materials by H
and 13C NMR spectroscopy and by GC/MS (after derivatization
with N,O-bis(trimethylsilyl)acetamide, BSA) are described in
detail elsewhere (J acobs, 1998). A brief account of the spec-
troscopic data of the monohydroxylated ENL compounds is
given below. NMR signals for ENL were assigned according
to Cooley et al. (1984); see Figure 4 for the numbering of
positions.
trans-3,4-Bis(3-hydroxybenzyl)-γ-butyrolactone (ENL). Mass
spectrum (EI, 70 eV) m/z (% relative abundance): 442 (67, M+),
263 (10), 217 (8), 205 (5), 180 (100, McLafferty rearrangement
product), 165 (6), 73 (7). 1H NMR (CDCl3/MeOD) δ (ppm):
2.40-2.65 (4H, m, 7, 8, 8′), 2.90 (1H, dd, 7′A), 3.00 (1H, dd,
7′B), 3.85 (1H, dd, 9A), 4.15 (1H, dd, 9B), 6.45-6.76 (6H, m),
7.11-7.20 (2H, m). 13C NMR (CDCl3/MeOD) δ (ppm): 35.35
(t, 7′), 38.73 (t, 7), 41.85 (d, 8), 46.94 (d, 8′), 72.20 (t, 9), 114.12
(d), 114.39 (d), 116.05 (d), 116.66 (d), 120.48 (d), 121.20 (d),
130.22 (2d), 139.85 (s), 140.36 (s), 157.72 (2s), 180.58 (s, 9′).
trans-2-(2,3-Dihydroxybenzyl)-3-(3-hydroxybenzyl)-γ-butyro-
lactone. Mass spectrum (EI, 70 eV) m/z (% relative abun-
dance): 530 (15, M+), 515 (5), 351 (100), 335 (5), 179 (12,
benzylic fragmentation), 73 (6). 1H NMR (MeOD) δ (ppm):
2.40-2.87 (6H, m, 7, 7′,8, 8′), 3.47 (1H, dd, 9A), 3.87 (1H, dd,
9B), 6.43-7.12 (7H, m).
trans-2-(3,4-Dihydroxybenzyl)-3-(3-hydroxybenzyl)-γ-butyro-
lactone. Mass spectrum (EI, 70 eV) m/z (% relative abun-
dance): 530 (100, M+), 515 (5), 267 (46, benzylic fragmenta-
tion), 205 (6), 179 (15, benzylic fragmentation), 73 (6). 1H NMR
(MeOD) δ (ppm): 2.27-2.79 (6H, m, 7, 7′, 8, 8′), 3.75 (1H, dd,
9A), 3.92 (1H, dd, 9B), 6.38-7.01 (7H, m). 13C NMR (MeOD) δ
(ppm): 39.14 (t, 7′), 39.52 (t, 7), 42.40 (d, 8), 47.74 (d, 8′), 72.75
(t, 9), 113.25 (d), 114.54 (d), 116.61 (d), 117.52 (d), 120.94 (d),
121.94 (d), 130.75 (d), 141.46 (s), 145.12 (s), 146.32 (s), 158.61
(s), 158.69 (s), 181.52 (s, 9′).
trans-2-(3,5-Dihydroxybenzyl)-3-(3-hydroxybenzyl)-γ-butyro-
lactone. Mass spectrum (EI, 70 eV) m/z (% relative abun-
dance): 530 (51, M+), 515 (8), 293 (3), 268 (100, McLafferty
rearrangement product), 253 (3), 73 (3). 1H NMR (MeOD) δ
(ppm): 2.34-2.98 (6H, m, 7, 7′, 8, 8′), 3.88 (1H, dd, 9A), 4.55
(1H, dd, 9B), 6.17-7.10 (7H, m). 13C NMR (MeOD) δ (ppm):
35.96 (t, 7′), 39.16 (t, 7), 42.70 (d, 8), 47.41 (d, 8′), 72.75 (t, 9),
102.09 (d), 108.91 (d), 108.91 (d), 114.52 (d), 116.57 (d), 120.96
(d), 130.69 (d), 140.91 (d), 141.50 (s), 141.57(s), 158.64 (s),
159.76(s), 181.35 (s, 9′).
trans-2-(2,5-Dihydroxybenzyl)-3-(3-hydroxybenzyl)-γ-butyro-
lactone. Mass spectrum (EI, 70 eV) m/z (% relative abun-
dance): 530 (71, M+), 515 (5), 351 (100), 335 (16), 267 (6,
benzylic fragmentation), 73 (6). 1H NMR (CDCl3/MeOD) δ
(ppm): 2.34 (1H, dd, 7′A), 2.61-2.82 (4H, m, 7, 8, 8′), 3.24 (1H,
dd, 7′B), 3.89 (1H, dd, 9B), 4.13 (1H, dd, 9A), 6.52-6.70 (6H,
m), 7.05-7.08 (1H, m). 13C NMR (CDCl3/MeOD) δ (ppm): 30.22
2,3-Bis(3-hydroxybenzyl)-butane-1,4-diol (END). Mass spec-
trum (EI, 70 eV) m/z (% relative abundance): 500 (27, M+
-
TMSOH), 410 (100, M+ - 2TMSOH), 231 (25), 217 (22), 205
(4), 194 (10), 180 (48, McLafferty rearrangement product), 73
(11). 1H NMR (DMSO-d6) δ (ppm): 1.90 (2H, m, 8), 2.37-2.63
(4H, m, 7), 3.37 (4H, m, 9), 6.54-6.57 (6H, m), 6.99-7.06 (2H,
m), 9.20 (2H, s). 13C NMR (DMSO-d6) δ (ppm): 33.86 (t, 7),
42.43 (d, 8), 59.99 (t, 9), 112.38 (d), 115.78 (d), 119.49 (d),
128.80 (d), 142.92 (s), 156.99 (s).
2-(2,3-Dihydroxybenzyl)-3-(3-hydroxybenzyl)-buta ne-1,4-
diol. Mass spectrum (EI, 70 eV) m/z (% relative abundance):
678 (3, M+), 588 (22, M+ - TMSOH), 498 (95, M+ - 2TMSOH),
268 (26), 231 (87), 180 (100), 73 (10).
2-(3,4-Dihydroxybenzyl)-3-(3-hydroxybenzyl)-buta ne-1,4-
diol. Mass spectrum (EI, 70 eV) m/z (% relative abundance):
678 (11, M+), 588 (72, M+ - TMSOH), 498 (70, M+
2TMSOH), 268 (90), 231 (41), 180 (100), 73 (4).
-
2-(2,5-Dihydroxybenzyl)-3-(3-hydroxybenzyl)-buta ne-1,4-
diol. Mass spectrum (EI, 70 eV) m/z (% relative abundance):
678 (41, M+), 588 (100, M+ - TMSOH), 498 (55, M+
2TMSOH), 268 (14), 231 (16), 180 (6), 73 (15).
-

Microsomal Metabolism of Mammalian Lignans
J. Agric. Food Chem., Vol. 47, No. 3, 1999 1073
Micr osom a l P r ep a r a tion s a n d In cu ba tion s. Microsomes
were prepared from the liver of untreated or aroclor-treated
male Wistar rats and untreated pigs. Treatment with Aroclor
1254 (one intraperitonal injection on day 1 of a dose of 500
mg/kg body weight dissolved in sesame oil at 100 mg/mL,
sacrifice on day 6) was carried out to induce hepatic cyto-
chrome P-450-dependent monooxygenases. Microsomes of hu-
man liver were kindly provided by Dr. W. Dekant, University
of Wu¨rzburg. All microsomes were prepared according to the
method of Lake (1987). Protein concentration was determined
by the protein assay of Pierce based on copper ions and
bicinchoninic acid (Smith et al., 1985). The concentration of
cytochrome P-450 was measured by the method of Omura and
Sato (1964). All chemicals used for microsomal preparations
and incubations were of the highest purity available.
The microsomal incubations were carried out in a final
volume of 2 mL containing 1 mg of microsomal protein, 25 µM
test compound dissolved in DMSO (final concentration 2.5%),
and a NADPH-generating system (0.9 U isocitrate dehydro-
genase, 9.4 mM isocitrate, 1.21 mM NADP+, and 4.3 mM
MgCl2) in phosphate buffer (0.1 M, pH 7.4). After a preincu-
bation period of 2 min at 37 °C, the reaction was initiated by
the addition of 70 µL of the NADPH-generating system and
terminated after 30 min at 37 °C by extraction with 3 × 2 mL
of ethyl acetate. This extract was evaporated to dryness and
dissolved in 100 µL of methanol for HPLC analysis.
F igu r e 2. HPLC profile of the enterolactone (ENL) metabo-
lites from aroclor-induced rat liver microsomes.
An a lytica l Meth od s. HPLC analysis was carried out on a
Polygosil RP-18 column (250 × 4 mm; 5 µm) with a linear
gradient program from 25% A (methanol)/75% B (16% methanol/
84% water, adjusted to pH 2.8 with formic acid) to 30% A/70%
B in 55 min. The flow rate was 1 mL/min; the UV absorbance
was monitored at 275 nm or recorded with a diode array
detector.
For HPLC/MS analysis, the same HPLC conditions were
used, but the gradient started at 15% A/85% B and ended at
20% A/80% B. Two ion source types were used: (i) atmospheric
pressure chemical ionization (APCI) with the following condi-
tions: corona current, 49.8 µA; corona voltage, 6.232 kV;
vaporizer temperature, 544 °C; 0.1 mL/min H2O/25% aqueous
NH3 (1/1, v/v) was added to the eluate prior to the ion source;
(ii) electrospray ionization (ESI) with conditions as follows:
spray voltage, 2.72 kV; spray current, 99.9 µA; capillary
temperature, 200 °C; 0.1 mL/min H2O was added to the eluate.
For GC/MS analysis, the sample was evaporated to dryness
under reduced pressure, and the residue was trimethylsilyl-
ated with 30 µL of BSA (1:10 in heptane). Conditions were as
follows: SE-30 fused silica column (30 m × 0.25 mm; 0.25 µm);
flow of He, 40 cm/s; column temperature, from 70 to 250 °C at
30 °C/min, from 250 to 270 °C at 1 °C/min; injector, 275 °C;
transfer line, 275 °C; ion source, 150 °C; ionization energy, 70
eV (ion trap mass detector); injection volume, 1 µL.
F igu r e 3. GC profile of the enterolactone (ENL) metabolites
from aroclor-induced rat liver microsomes.
Ta ble 1. Ch r om a togr a p h ic a n d Sp ectr oscop ic Da ta of
En ter ola cton e (ENL) a n d Its Meta bolites
[M - H]-
(HPLC/MS)
RT in
HPLC
(min)
UV
maxima
(nm)
RESULTS
HPLC
peak
corresponding
GC peak
M+
(GC/MS)
ESI APCI
Micr osom a l Meta bolism of En ter ola cton e. Pilot
studies with hepatic microsomes from untreated and
aroclor-treated rats, untreated pigs, and untreated
humans showed that liver microsomes of aroclor-treated
rats metabolized ENL to the greatest extent. The extract
from this microsomal incubation gave rise to nine
metabolite peaks in HPLC analysis using an UV detec-
tor (Figure 2, peaks A-I). When the HPLC eluate
containing all the metabolites was trimethylsilylated
and subject to GC/MS, 12 compounds were found
(Figure 3, peaks 1-12) that were absent in the respec-
tive eluate obtained from control incubations without
ENL. ENL itself coelutes with peak 5 in GC/MS analysis
and therefore had to be separated from the metabolites
by HPLC. By analyzing each HPLC metabolite peak by
GC/MS, the 12 ENL metabolite peaks of the GC profile
were correlated with the nine metabolite peaks observed
in HPLC. This correlation and the molecular ions of the
A
B/C
16.4
280
295 295
1, 2 (dp)a
6
7
11
440
530
530
530
530
440
440
440
440
440
530
530
442
19.4/20.4 284, 294 313 313
D
28.2
279
313 313
12
E
F
G
H
I
32.4
35.5
38.6
41.6
47.0
nmb
313 295
5 (dp)
4, 5 (dp)
4, 10 (dp)
10 (dp)
3, 4, 5 (dp)
8
279, 244 313 295
280, 242 313 295
286, 244 295 295
nm
ndc nd
9
ENL 51.4
277, 283 297 297
a
b
dp, daughter product. nm, not measurable (intensity too low
for UV spectrum). c nd, not detected.
12 ENL metabolites are summarized in Table 1 together
with the HPLC/MS data and the UV maxima as
obtained with a diode array detector.

1074 J. Agric. Food Chem., Vol. 47, No. 3, 1999
Jacobs and Metzler
Ta ble 2. Id en tifica tion of ENL Meta bolites Mon oh yd r oxyla ted a t th e Ar om a tic Rin gs
GC peak
position of hydroxylationa
chemical name of metabolite
6
7
8
9
11
12
para in ring B
para in ring A
ortho in ring B (2,3-catechol)
ortho in ring A (2,3-catechol)
ortho in ring B (3,4-catechol)
ortho in ring A (3,4-catechol)
trans-2-(2,5-dihydroxybenzyl)-3-(3-hydroxybenzyl)-γ-butyrolactone
trans-3-(2,5-dihydroxybenzyl)-2-(3-hydroxybenzyl)-γ-butyrolactone
trans-2-(2,3-dihydroxybenzyl)-3-(3-hydroxybenzyl)-γ-butyrolactone
trans-3-(2,3-dihydroxybenzyl)-2-(3- hydroxybenzyl)-γ-butyrolactone
trans-2-(3,4-dihydroxybenzyl)-3-(3-hydroxybenzyl)-γ-butyrolactone
trans-3-(3,4-dihydroxybenzyl)-2-(3-hydroxybenzyl)-γ-butyrolactone
a
Relative to existing hydroxy groups in ENL.
F igu r e 4. Possible positions of enterolactone (ENL) and
enterodiol (END) for aliphatic and aromatic hydroxylation.
Positions are numbered according to Cooley et al. (1984).
F igu r e 5. Microsomal metabolites of enterolactone (ENL)
formed through aromatic hydroxylation. The GC peak numbers
refer to Figure 3.
The mass spectrum of the parent compound ENL in
its trimethylsilylated form exhibited a strong molecular
ion at m/z 442 (17%, data not shown). Six of the 12 ENL
metabolites analyzed by GC/MS (viz., peaks 6-9, 11,
and 12) had their molecular ions at m/z 530, which
implied one additional hydroxy group. The mass spectra
of the other six GC peaks (i.e., peaks 1-5 and 10)
exhibited their highest m/z at 440, which suggested the
introduction of one additional double bond into the ENL
molecule. This double bond might have been formed
through elimination of a trimethylsilylated hydroxy
group during GC/MS. As aromatic and aliphatic hy-
droxylation can occur at various sites of the ENL
molecule (Figure 4) it was assumed that the metabolites
with m/z 530 represent products of aromatic hydroxyl-
ation whereas those with m/z 440 were less stable
products of aliphatic hydroxylation.
This proposition was corroborated by the mass spectra
obtained by HPLC/MS (Table 1). ENL exhibited a [M
- H] ion at m/z 297 both with electrospray ionization
(ESI) and atmospheric pressure chemical ionization
(APCI). The HPLC peaks B-D, corresponding to GC/
MS peaks 6, 7, 11, and 12 and thought to represent
metabolites hydroxylated at the aromatic rings, dis-
played [M - H] ions at m/z 313 in HPLC/MS with both
ESI and APCI. In contrast, HPLC peaks E-G, corre-
sponding to GC/MS peaks 4, 5, and 10, exhibited [M -
H] ions at m/z 313 only with the more gentle ESI but
gave rise to [M - H] ions at m/z 295 with APCI,
indicating the elimination of water. It is suggested that
HPLC peaks E-G are aliphatic hydroxylation products
of ENL sufficiently stable for ESI but not for APCI.
HPLC peaks A and H displayed [M - H] ions at m/z
295 with both EPI and APCI and may therefore repre-
sent highly unstable aliphatic hydroxylation products
eliminating water readily with any ionization method.
Elimination from some of the aliphatic positions can
lead to multiple olefinic daughter products (Figure 4).
For further identification of the aromatic hydroxy
metabolites, six derivatives of ENL with an additional
hydroxy group in ring A or B in para position or in the
two ortho positions of the existing phenolic hydroxy
group were synthesized. In addition, an ENL derivative
hydroxylated in the meta position of ring B was pre-
pared by chemical synthesis. With these reference
compounds, the ENL metabolites represented by GC
peaks 6-9, 11, and 12 were unequivocally identified as
shown in Table 2 by cochromatography in HPLC and
GC as well as identical UV and mass spectrum. The
complete pattern of these metabolites is depicted in
Figure 5.
Micr osom a l Meta bolism of En ter od iol. When
incubations of END with hepatic microsomes from
untreated and aroclor-treated rats, untreated pigs, and
untreated humans were analyzed by HPLC with UV
detector, the highest turnover was observed with aro-
clor-treated rat liver microsomes, but the number and
amounts of metabolites of END were markedly lower
than with ENL. Six metabolite peaks of END were
detectable by HPLC analysis (Figure 6, peaks A-F). The
eluate containing the metabolites was then trimethyl-
silylated and analyzed by GC/MS, which gave rise to
seven metabolite peaks as shown in Figure 7. Peaks 4a
and 4b could only be separated on a new GC column.
The correlation of HPLC and GC peaks, the UV maxima

Microsomal Metabolism of Mammalian Lignans
J. Agric. Food Chem., Vol. 47, No. 3, 1999 1075
Ta ble 3. Ch r om a togr a p h ic a n d Sp ectr oscop ic Da ta of
En ter od iol (END) a n d Its Meta bolites
RT in
UV
[M - H]-
(HPLC) corresponding
HPLC HPLC maxima
M+
peak
(min)
(nm)
nma
nm
246, 276
249, 279
APCI
GC peak
(GC/MS)
A
B
C
D/E
7.8
8.9
10.6
14.9
299
299
299
299
317
317
301
(dp)b,c
(dp)c
3 (dp)
6 (dp)
4
588
588
588
588
678
678
590
F
18.0
33.9
279
276, 283
5
END
a
b
nm, not measurable (intensity too low for UV spectrum). dp,
daughter product. c GC peaks 1 and 2 could not be unambiguously
correlated with the HPLC peaks due to the small amounts of
metabolites.
F igu r e 6. HPLC profile of the enterodiol (END) metabolites
from aroclor-induced rat liver microsomes.
F igu r e 8. Microsomal metabolites of enterodiol (END) formed
through aromatic hydroxylation. The GC peak numbers refer
to Figure 7.
synthesized and compared with the END metabolites
in HPLC and GC by cochromatography. Identical reten-
tion times as well as identical UV and mass spectra
clearly showed that GC peak 4a is 2-(2,5-dihydroxy-
benzyl)-3-(3-hydroxybenzyl)-butane-1,4-diol (END hy-
droxylated at the para position), GC peak 4b is 2-(2,3-
dihydroxybenzyl)-3-(3-hydroxybenzyl)-butane-1,4-diol
(END hydroxylated at the more sterically hindered
ortho position), and GC peak 5 is 2-(3,4-dihydroxy-
benzyl)-3-(3-hydroxybenzyl)-butane-1,4-diol (END hy-
droxylated at the less sterically hindered ortho position).
A metabolic scheme is depicted in Figure 8.
P a tter n of ENL a n d END Meta bolites w ith Mi-
cr osom es fr om Differ en t Sp ecies. Incubations of
ENL with hepatic microsomes of untreated and aroclor-
treated male rats, untreated pigs, and untreated hu-
mans showed qualitatively similar elution profiles in
HPLC and GC/MS but quantitative differences in the
relative amounts of the various metabolites (Table 4).
Metabolites 3, 8, and 9 could only be found with
microsomes from aroclor-treated rats in very low quan-
tities (see also Figure 3) and are omitted from Table 4.
Metabolites 11 and 12 were not observed with human
microsomes and only in minute amounts with porcine
microsomes. The other metabolites, i.e., GC peaks 1, 2,
4-7, and 10 were formed with microsomes from all
three species, although in different quantities. With
F igu r e 7. GC profile of the enterodiol (END) metabolites from
aroclor-induced rat liver microsomes.
data, and the MS data of the END metabolites are
summarized in Table 3.
In GC/MS, trimethylsilylated END does not exhibit
a molecular ion at m/z 590 but a [M - 90] ion at m/z
500, arising probably through the loss of TMSOH. The
molecular ions of GC peaks 4a, 4b, and 5 were at m/z
678, indicating an additional hydroxy group that is not
eliminated. In contrast, GC peaks 1-3 and 6 displayed
their highest m/z at 588, which suggests the elimination
of an additional hydroxy group. Thus, the microsomal
metabolites of END appear to consist of three stable
aromatic hydroxylated products with molecular ion at
m/z 678 and four less stable aliphatic hydroxylated
products with molecular ion at m/z 588. This assignment
is supported by the data from HPLC/MS with APCI
(Table 3).
As the two aromatic rings of END are equivalent,
there are only four possible aromatic monohydroxylation
products (Figure 4). Three of them were chemically

1076 J. Agric. Food Chem., Vol. 47, No. 3, 1999
Jacobs and Metzler
Ta ble 4. P a tter n of ENL Meta bolites w ith Micr osom es
fr om Va r iou s Sp ecies (Exp r essed in P er cen t of Tota l
Meta bolitesa )
we could detect six ENL metabolites with aliphatic
hydroxylation (viz., HPLC peaks A and E-I) at least
three aliphatic positions must have been hydroxylated
by liver microsomes. The ultimate identification of the
hydroxylated positions requires the synthesis of ap-
propriate reference compounds, which is in progress in
our laboratory. Aromatic hydroxylations of ENL can
yield eight different metabolites as there are four
positions at each of the two rings that are not equivalent
(Figure 4). With aroclor-induced rat liver microsomes,
hydroxylation takes place at six aromatic positions as
could be shown by GC/MS analysis. All six metabolites
were identified by means of synthetic reference sub-
stances as products of a ring hydroxylation in the para
position and the two ortho positions of the existing
phenolic hydroxy groups in both rings. The complete
pattern of these metabolites is depicted in Figure 5. No
evidence was obtained for the presence of ENL metabo-
lites hydroxylated in the meta position, for which one
reference compound was available; the failure to find
such metabolites is not surprising as meta hydroxyl-
ation of aromatic compounds is an uncommon metabolic
reaction.
aroclor-
metabolite
type of
treated
untreated
pig
human
(GC/MS)b hydroxylationc rat (1.71)d rat (0.59)d (0.54)d (0.39)d
1
2
4
5
6
aliphatic
aliphatic
aliphatic
aliphatic
aromatic
aromatic
aliphatic
aromatic
aromatic
1
2
8
19
24
13
14
3
5
4
26
7
10
10
29
3
6
13
13
11
26
9
20
1
1
3
5
19
17
7
7
6
10
11
12
43
nde
nd
16
6
a
Calculated from the peak areas of the total ion current
chromatogram obtained by GC/MS analysis (e.g., Figure 3 for
b
aroclor-induced rat liver microsomes). According to Figure 3.
d
c According to tentative identification (see text). Cytochrome
P-450 content (nmol/mg of microsomal protein). e nd, not detected,
<0.2%.
human microsomes, the generation of aliphatic hy-
droxylation products (peaks 4, 5, and 10) is preferred
at the expense of aromatic hydroxylation products
(peaks 6 and 7).
In contrast to ENL, which was readily metabolized
by aroclor-induced rat liver microsomes, the number and
amounts of END metabolites formed by the same
microsomes were much less. Aliphatic hydroxylation of
END can theoretically take place at two different
positions leading to four diastereomers (Figure 4). As
four such metabolites (HPLC peaks A-D/E) were found,
both aliphatic positions of END must be assumed to
undergo metabolic hydroxylation. In addition, three of
the four possible aromatic positions were hydroxylated
(GC peaks 4a, 4b, and 5), although to a very small
extent. These were identified with the help of synthetic
reference compounds as the para and both ortho posi-
tions of END (Figure 8).
Incubations of END with liver microsomes from
untreated rats, pigs, and humans generated only very
small quantities of metabolites and showed a similar
pattern as observed with aroclor-treated rat microsomes
(Figure 7): the predominating END metabolites in all
microsomes were GC peaks 3 and 6, although the ratio
varied between species: GC peak 3 accounted for 8% of
total metabolites with untreated rat, 39% with porcine,
and 13% with human hepatic microsomes, whereas GC
peak 6 represented 92% in untreated rat, 58% in porcine
and 87% in human microsomes. GC peaks 1, 2, and 4
were not detected in untreated rat, porcine, or human
microsomes.
DISCUSSION
A comparison of the oxidative biotransformation of
ENL and END by microsomes of untreated and aroclor-
treated rats, untreated pigs, and untreated humans
showed similar qualitative but different quantitative
patterns of metabolites. In all three uninduced species,
the major metabolic route is the hydroxylation at
aliphatic positions; the predominance of aliphatic over
aromatic hydroxylation is most striking in human
microsomes. Pretreatment of rats with aroclor, which
is a mixture of chlorinated biphenyls and is known to
induce several classes of cytochrome P450 isoenzymes
(Ryan et al., 1982), appears to shift the metabolic
pattern towards aromatic hydroxylation.
In conclusion, this study has shown that ENL and
END undergo aliphatic and aromatic hydroxylation at
various positions upon incubation with mammalian
hepatic microsomes. We have recently observed that
several of these novel metabolites of ENL and END are
also excreted in the urine of human subjects fed a diet
containing flaxseed (J acobs et al., 1999). All six products
of aromatic monohydroxylation of ENL (Figure 5) and
all three aromatic monohydroxylated metabolites of
END (Figure 8), but surprisingly none of the aliphatic
hydroxylation products, could be detected in human
urine. Studies are now in progress to synthesize further
reference compounds for the identification of all me-
tabolites of ENL and END and the investigation of their
genotoxic and estrogenic potentials.
The mammalian lignans ENL and END may reach
considerable concentrations in body fluids after inges-
tion of whole-grain products, in particular flaxseed, and
certain vegetables. As epidemiological and biochemical
studies suggest an anticarcinogenic potential for ENL
and END, these compounds are of considerable interest
for cancer prevention by appropriate nutrition. How-
ever, little is known about possible health risks. We
have recently reported that ENL and END as well as
their plant precursors matairesinol and secoisolarici-
resinol lack activity at various endpoints for genotoxicity
in vitro, e.g., the induction of micronuclei and gene
mutations in Chinese hamster V79 cells or the interfer-
ence with cell-free microtubule assembly (Kulling et al.,
1998). However, there are no reports in the literature
to date about the oxidative metabolism of these lignans
and the biological effects of their metabolites.
The results of the present study on the metabolism
of ENL and END in liver microsomes from untreated
and aroclor-treated rats, untreated pigs, and untreated
humans clearly show that both ENL and END undergo
oxidative biotransformation. ENL is metabolized by
microsomes of aroclor-treated rats to products carrying
one additional hydroxy group at different aliphatic and
aromatic positions. In theory, five aliphatic positions are
available for the introduction of a hydroxy function into
the ENL molecule (Figure 4). For racemic trans-ENL,
each hydroxylation can lead to two diastereomers. As

Microsomal Metabolism of Mammalian Lignans
J. Agric. Food Chem., Vol. 47, No. 3, 1999 1077
ABBREVIATIONS USED
J acobs, E. Synthese, Metabolismus und Genotoxizita¨t der
Lignane Enterolacton und Enterodiol. Ph.D. Dissertation,
University of Kaiserslautern, Germany, 1998.
J acobs, E.; Kulling, S. E.; Metzler, M. Novel metabolites of
the mammalian lignans enterolactone and enterodiol in
human urine. J . Steroid Biochem. Mol. Biol. 1999, in press.
J enab, M.; Thompson, L. U. The influence of flaxseed and
lignans on colon carcinogenesis and â-glucuronidase activity.
Carcinogenesis 1996, 17, 1343-1348.
Kulling, S. E.; J acobs, E.; Pfeiffer, E.; Metzler, M. Lack of
genotoxicity of the major mammalian lignans enterolactone
and enterodiol and their metabolic precursors at various
endpoints in vitro. Mutat. Res. 1998, 416, 115-124.
Lake, B. G. Preparation and characterisation of microsomal
fractions for studies on xenobiotic metabolism. In Biochemi-
cal Toxicology; Snell, K., Mullock, B., Eds.; IRL Press:
Oxford, 1987; pp 183-215.
Morton, M. S.; Wilcox, G.; Wahlqvist M, L.; Griffiths, K.
Determination of lignans and isoflavonoids in human female
plasma following dietary supplementation. J . Endocrinol.
1994, 142, 251-259.
APCI, atmospheric pressure chemical ionization; BSA,
N,O-bis(trimethylsilyl)acetamide; DMSO, dimethyl sulf-
oxide; EI, electron impact; END, enterodiol, 2,3-bis(3-
hydroxybenzyl)butane-1,4-diol; ENL, enterolactone, trans-
2,3-bis(3-hydroxybenzyl)-γ-butyrolactone; ESI, electro-
spray ionization; GC, gas chromatography; GC/MS, gas
chromatography/mass spectrometry; HPLC, high-per-
formance liquid chromatography; HPLC/MS, high-
performance liquid chromatography/mass spectrometry;
THF, tetrahydrofurane; TMS, trimethylsilyl.
ACKNOWLEDGMENT
We thank Mr. en-Naser and Dr. Hege (Knoll AG,
Ludwigshafen) for the HPLC/MS analysis of the me-
tabolites.
LITERATURE CITED
Na¨sman, J .-A. H.; Pensar, K. G. An improved one-pot prepara-
tion of 2-furanones. Synthesis Commun. 1985, 786-787.
Omura, T.; Sato, R. The carbon monoxide-binding pigment of
liver microsomes: I. Evidence for its hemoprotein nature.
J . Biol. Chem. 1964, 239, 2370-2377.
Pelter, A.; Ward, R. S.; Satyanarayana, P.; Collins, P. Syn-
thesis of lignan lactones by conjugate addition of thioacetal
carbanions to butenolide. J . Chem. Soc., Perkin Trans. 1983,
I, 643-647.
Rickard, S. E.; Thompson, L. U. Phytoestrogens and lignans:
effects on reproduction and chronic disease. In Antinutrients
and Phytochemicals in Food; Shahidi, F., Ed.; American
Chemical Society: Washington, DC, 1997; pp 273-296.
Rose, D. P. Dietary fiber, phytoestrogens, and breast cancer.
Nutrition 1992, 8, 47-51.
Ryan, D. E.; Thomas, P. E.; Levin, W. Purification and
characterization of a minor form of hepatic microsomal
cytochrome P-450 from rats treated with polychlorinated
biphenyls. Arch. Biochem. Biophys. 1982, 216, 272-288.
Seebach, D.; Corey, E. J . Generation and synthetic applications
of 2-lithio-1,3-dithianes. J . Org. Chem. 1974, 40, 231-237.
Serraino, M.; Thompson, L. U. The effect of flaxseed supple-
mentation on the initiation and promotional stages of
mammary tumorigenesis. Nutr. Cancer 1992, 17, 153-159.
Setchell, K. D. R.; Axelson, M.; Sjo¨vall, J .; Gustafsson, B. E.
Origin of lignans in mammals and identification of a
precursor from plants. Nature 1982, 298, 659-660.
Smith, P. K.; Krohn, R. J .; Olson, B. K., Klenk, D. C.
Measurement of protein using bicinchoninic acid. Anal.
Biochem. 1985, 150, 76-85.
Adlercreutz, H. Phytoestrogens: epidemiology and a possible
role in cancer protection. Environ. Health Perspect. 1995,
103 (Suppl. 7), 103-112.
Adlercreutz, H.; Mousavi, Y.; Clark, J .; Hockerstedt, K.;
Ha¨ma¨lainen, E.; Wa¨ha¨la¨, K.; Ma¨kela¨, T.; Hase, T. Dietary
phytoestrogens and cancer: in vitro and in vivo studies. J .
Steroid Biochem. Mol. Biol. 1992, 41, 331-337.
Adlercreutz, H.; Bannwart, C.; Wa¨ha¨la¨, K.; Ma¨kela¨, T.;
Brunow, G.; Hase, T.; Arosemena, P. J .; Kellis, J ., J r.;
Vickery, L. E. Inhibition of human aromatase by mam-
malian lignans and isoflavonoid phytoestrogens. J . Steroid
Biochem. Mol. Biol. 1993, 44, 147-153.
Adlercreutz, H.; van der Wildt, J .; Kinzel, J .; Attalla, H.;
Wa¨ha¨la¨, K.; Ma¨kela¨, T.; Hase, T.; Fotsis, T., Lignan and
isoflavonoid conjugates in human urine. J . Steroid Biochem.
Mol. Biol. 1995a , 52, 97-103.
Adlercreutz, H.; Musey, P. I.; Gould, K. G.; Collins, D. C.;
Fotsis, T.; Bannwart, C.; Ma¨kela¨, T.; Wa¨ha¨la¨, K.; Brunow,
G.; Hase, T., Effect of diet on lignans and isoflavonoid
phytoestrogens in chimpanzees. Life Sci. 1995b, 57, 655-
664.
Atkinson, D. A.; Hill, H. H.; Shultz, T. D. Quantification of
mammalian lignans in biological fluids using gas chroma-
tography with ion mobility detection. J . Chromatogr. 1993,
617, 173-179.
Axelson, M.; Setchell, K. D. R. The excretion of lignans in ratss
evidence for an intestinal bacterial source for this new group
of compounds. FEBS Lett. 1981, 123, 337-343.
Cooley, G.; Farrant, R. D.; Kirk D. N.; Patel, S.; Wynn, S.;
Buckingham, M. J .; Hawkes, G. E.; Hursthouse, M. B.;
Galas, A. M. R.; Lawson, A. M.; Setchell, K. D. R. Structural
analysis of the urinary lignan 2,3-bis-(3-hydroxybenzyl)-
butan-4-olide (‘enterolactone’). A 400 MHz nuclear magnetic
resonance study for the solution state and X-ray study for
the crystal state. J . Chem. Soc., Perkin Trans. 1984, II, 489-
497.
Enders, D.; Eichenauer, H.; Pieter, R. Enantioselektive Syn-
these von (-)-(R)- und (+)-(S)-(6)-Gingerol-Gewu¨rzprinzip
des Ingwers. Chem. Ber. 1979, 112, 3703-3714.
Feringa, B. L.; van Oeveren, A.; J ansen, J . F. G. A. Enantio-
selective synthesis of natural dibenzylbutyrolactone lignans
(-)-enterolactone, (-)-hinokinin, (-)-pluviatolide, (-)-en-
terodiol, and furofurane lignan (-)-eudesmin via tandem
conjugate addition to γ-alkoxybutenolides. J . Org. Chem.
1994, 59, 5999-6007.
Sung, M.-K.; Lautens, M.; Thompson, L. U. Mammalian
lignans inhibit the growth of estrogen-independent human
colon tumor cells. Proc. Am. Assoc. Cancer Res. 1996, 37,
279.
Thompson, L. U.; Robb, P.; Serraino, M.; Cheung, F. Mam-
malian lignan production from various foods. Nutr. Cancer
1991, 16, 43-52.
Wang, C.; Ma¨kela¨, T.; Hase, T.; Adlercreutz, H.; Kurzer, M.
S. Lignans and flavonoids inhibit aromatase enzyme in
human. J . Steroid Biochem. Mol. Biol. 1994, 50, 205-212.
Received for review August 18, 1998. Revised manuscript
received November 30, 1998. Accepted December 4, 1998. This
study has been supported by the Deutsche Forschungsgemein-
schaft (Grant Me 574/9-2).
Groen, M. B.; Leemhuis, J . Synthesis of compound X, a
nonsteroidal constituent of female urine, and congeners.
Tetrahedron Lett. 1980, 21, 5043-5046.
J F9809176