Equol, a natural estrogenic metabolite from soy isoflavones: Convenient preparation and resolution of R- and S-equols and their differing binding and biological activity through estrogen receptors alpha and beta

Article abstract of DOI:10.1016/j.bmc.2003.11.035

Equol is a metabolite produced in vivo from the soy phytoestrogen daidzein by the action of gut microflora. It is known to be estrogenic, so human exposure to equol could have significant biological effects. Equol is a chiral molecule that can exist as the enantiomers R-equol and S-equol. To study the biological activity of racemic (±)-equol, as well as that of its pure enantiomers, we developed an efficient and convenient method to prepare (±)-equol from available isoflavanoid precursors. Furthermore, we optimized a method to separate the enantiomers of equol by chiral HPLC, and we studied for the first time, the activities of the enantiomers on the two estrogen receptors, ERα and ERβ. In binding assays, S-equol has a high binding affinity, preferential for ERβ (K_i[ERβ]=16 nM; β/α=13 fold), that is comparable to that of genistein (K _i[ERβ]=6.7 nM; β/α=16), whereas R-equol binds more weakly and with a preference for ERα (K_i[ERα]=50 nM; β/α=0.29). All equol isomers have higher affinity for both ERs than does the biosynthetic precursor daidzein. The availability and the in vitro characterization of the equol enantiomers should enable their biological effects to be studied in detail.

Full text of DOI:10.1016/j.bmc.2003.11.035

Bioorganic & Medicinal Chemistry 12 (2004) 1559–1567
Equol, a natural estrogenic metabolite from soy isoflavones:
convenient preparation and resolution of R- and S-equols and their
differing binding and biological activity through estrogen receptors
alpha and beta
a
b
c
d
Rajeev S. Muthyala, Young H. Ju, Shubin Sheng, Lee D. Williams,
d
c
b,
Daniel R. Doerge, Benita S. Katzenellenbogen, William G. Helferich *
a,
and John A. Katzenellenbogen *
aDepartment of Chemistry, University of Illinois, 600 South Mathews Avenue, Urbana, IL 61801, USA
cDepartment of Molecular and Integrative Physiology, University of Illinois, Urbana, IL 61801, USA
bDepartment of Food Science and Human Nutrition, University of Illinois, 905 South Goodwin Avenue, Urbana, IL 61801, USA
dNational Center for Toxicological Research, Jefferson, AR 72079, USA
Received 28 July 2003; accepted 19 November 2003
Abstract—Equol is a metabolite produced in vivo from the soy phytoestrogen daidzein by the action of gut microflora. It is known
to be estrogenic, so human exposure to equol could have significant biological effects. Equol is a chiral molecule that can exist as the
enantiomers R-equol and S-equol. To study the biological activity of racemic (Æ)-equol, as well as that of its pure enantiomers, we
developed an efficient and convenient method to prepare (Æ)-equol from available isoflavanoid precursors. Furthermore, we opti-
mized a method to separate the enantiomers of equol by chiral HPLC, and we studied for the first time, the activities of the enan-
tiomers on the two estrogen receptors, ERa and ERb. In binding assays, S-equol has a high binding affinity, preferential for ERb
(K
i i
[ERb]=16 nM; b/a=13 fold), that is comparable to that of genistein (K [ERb]=6.7 nM; b/a=16), whereas R-equol binds more
weakly and with a preference for ERa (K [ERa]=50 nM; b/a=0.29). All equol isomers have higher affinity for both ERs than does
i
the biosynthetic precursor daidzein. The availability and the in vitro characterization of the equol enantiomers should enable their
biological effects to be studied in detail.
#
2004 Elsevier Ltd. All rights reserved.
3
1. Introduction
therapy using antiestrogens such as tamoxifen. Thus, it
is unclear whether the biological effect of these phyto-
estrogens will be a health benefit or a risk to humans. It
is also possible that their potential for benefit to one
tissue or one health condition could be compromised by
their potential risk to another tissue or condition.
Isoflavanoids found in soy, such as genistein and daid-
zein, have attracted great interest as dietary phytoestro-
gens that might be effective for menopausal hormone
1
replacement therapy. Suggestions have been made that
these isoflavones might also be useful in the prevention
2
or treatment of breast cancer. Recent studies in ani-
In examining the impact of the estrogenic activity of soy
phytoestrogens, one needs to consider not only the iso-
flavones and their conjugates that are ingested, but also
biologically active metabolites that might be generated
from them in vivo. Daidzein, a prominent isoflavone in
soy, is converted to the corresponding chroman, S-(À)-
equol, a compound whose estrogenic activity exceeds
mals, however, raise the possibility that the mammary
cell growth-promoting effects of these compounds
might, in fact, increase breast cancer risk and might also
compromise the effectiveness of breast cancer hormone
4
,5
Abbreviations: ER, Estrogen Receptor; HPLC, High Performance
Liquid Chromatography; RBA, Relative Binding Affinity; HEC-1,
Human Endometrial Carcinoma Cells-1.
that of daidzein. This reductive metabolic conversion
6
appears to result from the action of gut microflora.
Equol was first isolated from pregnant mares’ urine in
*
Corresponding authors. Tel.: +1-217-333-6310; fax: +1-217-333-7325
W.H.); tel.: +1-217-244-5414 (J.A.K.); e-mail addresses: helferic@
staff.uiuc.edu; jkatzene@uiuc.edu
7
1932 and was subsequently identified in the plasma of
sheep (presumably derived from formononetin in red
(
0
968-0896/$ - see front matter # 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2003.11.035

1560
R. S. Muthyala et al. / Bioorg. Med. Chem. 12 (2004) 1559–1567
clover)8 and in human urine (from daidzein).9 Cur-
iously, the extent of conversion of isoflavones to equol
varies greatly among humans, presumably because of
differences in the composition of gut microflora.
Although typical laboratory animal species (mouse, rat,
monkey) consistently produce high levels of equol, only
about 35% of humans are high equol producers, and
variations in urinary equol levels as great as 600-fold
palladium on carbon as catalyst. There are two major
disadvantages with these particular procedures, how-
ever: first, the catalyst had to be activated prior to use
by stirring under oxygen for 3 days and second, large
amounts of catalyst were used. For example, 30 g of
10% palladium on carbon was used to reduce 6.7 g of
daidzein diacetate. Given the cost of the palladium
catalyst ($60 for 10 g), it is clear that scaling up the
reduction to give, for example, 100 g of equol would be
prohibitive.
1
0À13
have been noted.
Transfer hydrogenation is an attractive alternative to
conventional hydrogenation, and a popular reagent for
1
9
this reaction is ammonium formate. Ram and Spicer
noted that ketones and aldehydes undergo reduction to
methylene groups using ammonium formate, provided
2
0
that acetic acid is used as the solvent. To test whether
this methodology could be extended to the vinylogous
ester carbonyl functionality of isoflavones, we per-
formed exploratory experiments. Using formononetin in
a model reaction, we found that with ammonium for-
mate in acetic acid solvent and 10% palladium on car-
bon as the catalyst, we obtained monomethyl equol in
11% yield. The rest of the material consisted of pro-
ducts arising from partial reduction of either the double
bond or the keto group. Longer reaction times did not
lead to any improvement in yield. In further investig-
ations in alternative solvents, such as tetrahydrofuran
or ethanol, we observed only partial reduction of the
double bond of isoflavanoids, as noted in the litera-
Equol is a chiral molecule that can exist in two
enantiomeric forms, and the enantiomer produced by
metabolic reduction from isoflavones is known to be S-
(
À)-equol. Equol is uterotrophic in mice, having a
potency similar to that of genistein, and its binding
affinity for the estrogen receptor from sheep uterus is
reported to be about 1/250th that of estradiol.1 (Æ)-
Equol produced by synthesis has been shown to induce
breast cancer cell proliferation in culture at concentra-
4
1
5
tions as low as 100 nM.
As part of an effort aimed at investigating the role that
equol might play in breast cancer and the potential for
differential activity of the two equol enantiomers
2
1
ture. Attempts to use alternative transfer hydrogena-
tion sources such as cyclohexene were also not
successful. By varying the catalyst, we found that Raney
Ni also led only to a mixture of partially reduced pro-
ducts. We were gratified to find, however, that Pearl-
1
6
through the two estrogen receptors, ERa and ERb, we
encountered the need to prepare equol on a large scale
and to obtain pure samples of S- and R-equol. In this
report, we describe a convenient preparation of (Æ)-
equol based on transfer hydrogenation that begins from
readily available isoflavone precursors, as well as a bio-
mimetic synthesis of (Æ)-equol. We also effect a robust
chromatographic resolution of (Æ)-equol into R- and S-
equol. We then evaluate, for the first time, the ERa and
ERb binding affinity and transcriptional activity of (Æ)-,
R- and S-equol, and compare them with those of the
biosynthetic precursor daidzein and the other abundant
soy isoflavone genistein. Curiously, the equol enantio-
mers show differential behavior on the two ER sub-
types, with S-equol having a preference for ERb and R-
equol having a preference for ERa.
2
2
man’s catalyst, which is 20% Pd(OH) on carbon, was
2
highly effective in this reduction.
Using this catalyst, we could reduce formononetin to
monomethyl equol in 68% yield (Scheme 1). The methyl
protecting group of the latter compound could be sub-
sequently removed using aluminum trichloride in the
2
3
presence of ethane thiol to give equol in 84% yield.
The overall yield for this two-step process is 57%. It is
more convenient, however, to prepare equol directly
from daidzein, and this direct reduction can be done in
61% yield (Scheme 1). We have used the latter one-step
process to obtain >100 g of (Æ)-equol from daidzein.
In addition to methods involving hydrogenation, we
also examined the possibility of using reagents such as
NaBH /CH CO H or Et SiH/CF CO H to reduce for-
2. Results
4
3
2
3
3
2
2
.1. Synthesis
mononetin or its protected derivatives. But all such
attempts led only to complete recovery of the starting
material. Since these reagents are known not to reduce
ester groups, the negative results presumably indicate
that the carbonyl group of formononetin behaves more
like an ester carbonyl rather than a ketone functionality.
Thus, we anticipated that more reactive hydride donors,
such as those derived from dihydroaromatics, could
be used to accomplish the reduction. A good example of
this class of hydride donors is dihydroacridine, a
compound that has, in fact, been reported to reduce
Our starting materials for the production of equol were
the isoflavanoids formononetin and daidzein, which are
readily available in large quantities from red clover and
soy, respectively. The key step in the conversion of iso-
flavanoids to equol involves reduction of a vinylogous
ester to an ether functional group. Most examples of
such reductions involve either multiple steps or toxic
1
7,18
reagents, with the exception of two reports, both of
which utilized hydrogenation in acetic acid with 10%

R. S. Muthyala et al. / Bioorg. Med. Chem. 12 (2004) 1559–1567
1561
Scheme 1.
xanthones, which bear structural and functional simila-
2
The CD spectra obtained from the isolated peaks are
shown in Figure 2. Peak 1 (17.5 min retention time)
showed negative ellipticity at 280 nm, whereas Peak 2
(19.4 min retention time) showed positive ellipticity,
with the two CD spectra being essentially mirror images
of one another, properties that are characteristic of an
enantiomeric pair. Based on prior structural assignment
of (À)-equol as S-equol (Beilstein registry number
87752), Peak 1 in Figure 1 is S-(À)-equol and Peak 2 is
R-(+)-equol. As noted above, S-(À)-equol is the natu-
4
rities to isoflavones. In our attempts to extend this
reaction to isoflavones, we found that using tri-
2
5
fluoroacetic acid as the solvent and dihydroacridine as
the reductant, formononetin could indeed be converted
to monomethyl equol (Scheme 2), in moderate yield.
We believe that the above conversion is ‘biomimetic’ in
nature, since dihydroacridine closely mimics the func-
tion of NADH, which is the most likely reducing agent
in a biosynthetic context. An interesting sidelight of this
reaction is that its progress can be followed by the
increase in intense fluorescence of acridine as it is being
generated.
2
6,27
rally occurring enantiomer.
2.3. Estrogen receptor binding affinity of R- and S-equol
and the related soy isoflavones, daidzein and genistein
The binding affinities of (Æ)-, R- and S-equols to ERa
and ERb are shown in the first two sections of Table 1
2.2. Chromatographic resolution of R- and S-equol and
stereochemical assignment
(
RBA values and K values), together with those of
i
Equol isomers were initially separated using an analy-
tical scale (2 mm diameter) b-cyclodextrin stationary
phase liquid chromatography (LC) column under
reversed phase conditions. Figure 1 shows the excellent
separation of the two enantiomeric forms of racemic
equol under these analytical scale conditions. The
separation was then scaled up to a 10 mm diameter LC
column to effect a semi-preparative isolation of
approximately 10 mg total from each peak; good
separations were obtained with injection of up to 400 mg
of the racemate.
daidzein and genistein for comparison. Binding affinities
were determined by a competitive radiometric binding
assay using purified full length human ERa and ERb,
with [ H]estradiol as tracer. In this assay, the affinities
are obtained as Relative Binding Affinity (RBA) values,
where estradiol has an affinity of 100.
3
While RBA values can be used to make comparisons
between the binding affinities of different compounds on
the same ER subtype, they are not appropriate for
comparisons between different subtypes (e.g., ERa and
ERb), because estradiol (the reference compound
and tracer in the competitive binding assays) binds to
ERa with an affinity ca. 2.5 times higher than to ERb
The individual collected fractions were analyzed using
HPLC with electrospray mass spectrometric (ES–MS/
MS) detection in the product ion mode (m/z 80–250
from m/z 243), and the mass spectra of the individual
enantiomers were identical to that observed for racemic
equol (product ions m/z 133, 123, 107, 105; data not
shown). The purified fractions contained <0.3% of the
other isomer, based on LC-ES/MS/MS analysis; fur-
thermore, no other chemical contaminants were
observed in either fraction by using LC-UV or LC-ES/
MS/MS analysis.
2
8
(K =0.2 nM for ERa versus K =0.5 nM for ER).
d
d
Therefore, for each compound on each ER subtype, we
have calculated the corresponding equilibrium binding
inhibition constant or K values. The K values of the
i
i
compounds represent their absolute affinities for both
ER subtypes, and they are also the appropriate indices
for comparison with the potencies of these compound
(EC50 values) determined in the transcription assays (see
below).
The equol enantiomers have distinctly different binding
affinities: The absolute binding affinities (K values) of
i
the naturally occurring equol enantiomer (S-equol)
and of the unnatural enantiomer (R-equol) on ERa and
ERb are nearly reversed from one another, with S- equol
having a strong preference for ERb (b/a=13) and
Scheme 2.

1562
R. S. Muthyala et al. / Bioorg. Med. Chem. 12 (2004) 1559–1567
Figure 1. Separation of equol enantiomers by chiral phase liquid chromatography. Separations were done on an analytical column with a b-cyclo-
dextrin-based chiral stationary phase, under reversed phase conditions (see Experimental Methods). Panel A. Chromatogram of equol racemate.
LP Ca n. el B. Chromatogram of S-(À)-equol obtained by preparative chiral LC. Panel C. Chromatogram of R-(+)-equol obtained by preparative chiral
R-equol having a moderate preference for ERa (b/
a=0.29). Significantly, the binding characteristics of S-
equol, particularly its high ERb preference, are most
comparable to those of the most estrogenic soy iso-
flavone genistein, although R-equol binds better to ERa
than does genistein. Also of note is the fact that all of
the equols have much higher affinities for both ERs
than does the abundant soy isoflavone daidzein, the
biosynthetic precursor. As expected, the binding affinity
of the racemate, (Æ)-equol, is almost exactly the aver-
age of that of the individual equol enantiomers. The
much higher binding affinity of genistein relative to
daidzein on ERb is likely due to the decrease in polarity
of the vinylogous ester functionality that occurs as a
result of the intramolecular hydrogen bond between the
carbonyl group and the proximal –OH function in gen-
istein, an interaction that is absent in daidzein.2
9
2.4. Transcriptional activities of the equols and iso-
flavone compounds
Figure 2. Circular dichroism spectra of equol enantiomers. The curve
showing the negative ellipticity is from material in the first peak to
elute (corresponding to S-(À)-equol, see Fig. 1A), and the curve
showing the positive ellipticity is from material in the second peak to
elute (corresponding to R-(+)-equol, see Fig. 1A). For details, see
Experimental Methods.
The equols and the related isoflavones were assayed for
transcriptional activity through both estrogen receptor
subtypes. These cotransfection assays were conducted in

R. S. Muthyala et al. / Bioorg. Med. Chem. 12 (2004) 1559–1567
1563
human endometrial carcinoma (HEC-1) cells, using
expression plasmids for either full-length human ERa or
ERb and an estrogen-responsive luciferase reporter
shown in Figure 3, and the EC50 values (which are
comparable to the K values for binding affinity, not the
RBA values) are summarized in the rightmost portion
of Table 1.
i
3
0
gene. In all cases, transcriptional activity of the var-
ious compounds through either ERa or ERb is pre-
sented relative to that obtained with 10 M estradiol,
which is set at 100%. The dose–response curves are
À9
From the transcription activation curves, it is apparent
that all of the compounds are agonists on both ERa and
Table 1. Binding affinitiesa,b and Transcriptional potencies of isoflavones and equol for ERa and ERb
c
a
d
b
i
d
c
d
RBA (%)
b/a
K
(nM)
hERb
0.5
b/a
EC50 (nM)
b/a
Ligand
hERa
hERb
hERa
0.2
hERa
0.021
250
hERb
Estradiol
Daidzein
Genistein
(Æ)Equol
R(+)Equol
100
100
0.040Æ0.001
1
4
440
8
0.4
1.5
0.11
100
6.6
0.19
2.5
0.010Æ0.006
0.017Æ0.003
0.20Æ0.02
0.40Æ0.04
0.10Æ0.01
2000
1300
6.7
31
170
16
7.4Æ0.5
1200
100
50
180
80
200
66
12
1.60Æ0.04
0.30Æ0.02
3.20Æ0.06
3.2
0.29
74
330
65
2.7
0.20
1.3
0.7
32
S(À)Equol
200
13
85
a
d
RBA=Relative Binding Affinity (as a%) measured in a competitive binding assay (see Experimental). The K values of estradiol for ERa and ERb
are 0.2 nM and 0.5 nM, respectively.28
b
c
K
Transcriptional activity measured using a cotransfection assay in HEC-1 cells (see Experimental and Fig. 3). Transcriptional potency=EC50
i
=equilibrium binding competition constant, calculated from the K
d
of estradiol on ERa or ERb: (K
d
/RBA)Â100.
.
d
For each index, the b/a ratio is calculated such that the ratio is >1 for compounds having higher affinity or greater potency on ERb than on ERa.
Figure 3. Dose–response curve for transcriptional activation by the indicated compounds. In each case, the solid circles are for ERa and the solid
triangles are for ERb. (A) estradiol (E
); (B) genistein; (C) daidzein; (D) (Æ)-equol; (E) S-equol; (F) R-equol. Human endometrial carcinoma (HEC-
) cells were transfected with expression vectors for ERa or ERb and an (ERE) -pS2-luc reporter gene and were treated with estradiol or the spe-
2
1
2
cified compound for 24 h, at the concentrations indicated. Luciferase activity was expressed relative to b-galactosidase activity from an internal
control plasmid. The maximal activity with 1 nM estradiol was set at 100% (A). Values are the meanÆSD from three separate experiments.

1564
R. S. Muthyala et al. / Bioorg. Med. Chem. 12 (2004) 1559–1567
ERb, giving maximal efficacies that are comparable to
that of estradiol. Genistein appears to be somewhat of a
superagonist on ERa, a characteristic that has been
noted by others, and the maximum efficacy of R-equol
might be slightly less than that of estradiol.
commercially available isoflavone precursors. It is not,
however, an enantioselective synthesis, although enantio-
selective syntheses have been described for both optical
antipodes of equol dimethyl ether starting from sub-
stituted phenylacetic acids, using a stereoselective alkyl-
3
1
5
1
ation and cyclization strategy. Rather than attempt to
replicate this multistep procedure, we chose simply to
resolve R- and S-equol by chiral LC. Using a chiral
b-cyclodextrin column, we obtained an excellent separa-
tion that could be scaled up readily to produce multi-
milligram amounts of the pure equol stereoisomers. The
identity of the eluted material was confirmed by LC-ES/
MS/MS, and the assignment of stereoisomers was done
by CD; the first eluting peak gave a CD spectrum hav-
ing a long-wavelength peak with a negative ellipticity,
which is known to correspond to S-equol, the naturally
occurring enantiomer, and the second peak, with a
When the potencies are compared, it is evident that all
of the compounds are less potent than estradiol. Over-
all, the potencies (EC50 values) of the equols and the
two isoflavones are comparable (ca. within a factor of
1
0) to their binding affinities (K values); the same is true
i
for their ERa versus ERb potencies (b/a values). The
only significant difference between affinity and potency
is that S-equol has a high binding affinity preference for
ERb (b/a=13), but essentially no ER subtype pref-
erence in terms of transcriptional potency (b/a=1.3).
This is in contrast to the ERa preference of R-equol and
the ERb preference of genistein, evident in their binding
affinities (b/a=0.29 and 180, respectively), that are
largely maintained in their transcriptional potencies
2
6,27
positive CD corresponding to R-equol.
The two equol enantiomers showed very different beha-
vior in terms of their binding affinities and transcrip-
tional potencies with ERa and ERb. In competitive
binding affinity assays, we found that the binding affi-
nity of the natural enantiomer, S-equol, is quite high
and shows a 13-fold preference for ERb, being similar
in these respects to genistein (b/a=180), the most
estrogenic soy isoflavone, whereas R-equol shows a
preference for ERa (b/a=0.29). Daidzein, the iso-
flavone precursor of equol, binds more weakly and
shows little ER subtype selectivity. In cotransfection
transcription assays, the potency of the equol stereo-
isomers and of genistein nicely reflect their binding affi-
nities for ERa and ERb, although the strong ERb
preference of S-equol in terms of binding affinity is
considerably muted and the transcriptional potency of
daidzein is higher than expected from its affinity. Thus,
genistein is the only compound that stands out as hav-
ing an accentuated potency for ERb in transcription
assays, whereas both S-equol and genistein have a pref-
erence for ERb in terms of binding. Whether these dif-
ferences in ER subtype selectivity between binding and
transcription assays are due to cellular metabolism or
differential interaction with cellular coregulators is an
(
b/a=0.20 and 12, respectively).
3. Discussion
Legumes (including soy, sprouts, and red clover) are
rich sources of isoflavones that have been shown to be
cancer-preventive.
soybeans and soy products are genistin and daidzin (b-
glycosides), and their aglycone forms, genistein and
daidzein. Biochanin A and formononetin are proestro-
genic isoflavones also found in legumes. After inges-
tion, the b-glycosides undergo further metabolism by
3
2,33
The major isoflavones found in
3
4
3
5,36
intestinal microflora,
and through the action of
bacterial glucosidases, they are converted to aglycones
genistein and daidzein).3
population can convert daidzein to equol.
amount of urinary equol excretion is correlated with the
7À39
In humans, 30–40% of the
(
4
0,41
The
1
0,42
consumption of soy products
4
and a reduced risk of
3,44
breast cancer.
equol might be more estrogenic than daidzein.
In vitro experiments suggest that
,35,41
4
The
actions of estrogens, including those of the phytoestro-
gens such as daidzein (and its equol metabolite) and
genistein, the principal phytoestrogens in soy, are medi-
ated through the estrogen receptors (ERs), of which
there are two subtypes, ERa and ERb. These proteins
are ligand-regulated transcription factors, and their tis-
sue distribution, biological functions, and response to
structurally diverse ligands are quite different.
The phytoestrogens are often considered to be ERb-
selective ligands.
5
2À55
issue that needs to be evaluated further.
1
6
It is of note that the range of ERa binding affinity
values reported for soy phytoestrogens is remarkably
wide, with values for genistein spanning nearly a 100-
fold range (RBA values 0.010–0.88) and for daidzein an
1
6,45À49
5
6
8-fold range (RBA values 0.01–0.08). The ERa bind-
ing affinity values we have obtained for both of these
compounds (Table 1) are clearly within these ranges,
but we note that it proved far more difficult to obtain
consistent binding affinity values for these two com-
pounds than for a very wide range of other estrogens
that we have evaluated in the same binding assay over a
period of more than 30 years. Furthermore, in our
hands genistein has a higher binding affinity for the ER
in a preparation from lamb uterus, giving an RBA value
of 0.69, a value that is similar to that obtained in rat
3
1,50
Equol is a chiral molecule, capable
of existing in two enantiomeric forms with potentially
different biological activities, with natural equol, being
7
,26,27
the S-(À) enantiomer.
The relative binding affinity
and transcriptional potency of the equol enantiomers
for the two ER subtypes have not been reported
previously.
To facilitate studies on the receptor interactions and
estrogenic activity of equol and its enantiomers, as well
as experiments in animal models of human diseases, we
have developed, as reported here, a convenient method
to prepare (Æ)-equol in multigram quantities from
5
7
uterine ER by Sheehan (RBA 0.45). One would expect
the ER from these uterine sources to be predominantly
1
6,58
ERa.

R. S. Muthyala et al. / Bioorg. Med. Chem. 12 (2004) 1559–1567
1565
The different tissue distribution and distinct biological
roles of ERa and ERb and their potential for mediating
divergent effects of estrogens on target tissue stimula-
tion and proliferation highlight the importance of char-
acterizing carefully the efficacy and potency of various
estrogens through these two receptors. ERa is found in
most estrogen target tissues, in some cases nearly by
itself or accompanied by lower levels of ERb. ERb, on
the other hand, is found by itself or in higher levels than
ERa only in a few regions of the brain, in the ovary,
prostate and testis.1
was purchased from Indofine Chemical (Belle Mead, NJ).
Ammonium formate, ethanethiol and 20% Pd(OH) on
carbon were purchased from Aldrich Chemical Company.
2
5.2. Chemical synthesis, separation and spectroscopic
methods
5.2.1. Reduction of formononetin. Formononetin (5.0 g,
18.6 mmol) was suspended in acetic acid (25 mL) toge-
ther with ammonium formate (11.75 g, 186 mmol) and
6,58,59
20% Pd(OH) on carbon (4.8 g). The reaction mixture
2
was then refluxed for 1 h and allowed to cool to room
temperature. After diluting with 50 mL of EtOAc, the
reaction mixture was filtered through a 2 cm pad of
Celite that was subsequently washed with an additional
150 mL of EtOAc. The filtrate was evaporated in vacuo
to give a light brown oily residue which was purified on
silica gel using 15% EtOAc/hexane as the solvent. Tri-
turation of this oil with 125 mL of water gave a white
fluffy solid that was filtered and dried to give 3.2 g
Although at this point not fully understood, the biolo-
gical functions of ERa and ERb have been studied both
6
0,61
by the generation of gene knockout mice
and by the
6
2
use of ER subtype-selective ligands. One interesting
hypothesis that has been advanced, which is consistent
with the generally greater transcriptional effectiveness of
ERa vs ERb, is that estrogens acting through ERa
effect a strong stimulatory or proliferative drive in tar-
get tissues (or estrogen-responsive tumors) and that
1
(68%) of monomethyl equol. H NMR (500 MHz, ace-
6
3
estrogens acting through ERb moderate this drive.
tone-d ): d 7.05 (d, 8.4 Hz, 2H), 6.84 (d, 8.1 Hz, 1H),
6
Thus, both the relative levels of ERa and ERb, as well
as the potency or efficacy selectivity of the particular
estrogen that is acting through these two subtypes,
could determine the ultimate magnitude of the stimula-
tory effect. Consistent with this hypothesis are the
reports that as breast and prostate tumors become pro-
gressively more malignant, there is a general increase in
6.74 (d, 8.4 Hz, 2H), 6.30 (dd, 2.36 Hz, 8.1 Hz, 1H), 6.23
(d, 2.36 Hz, 1H), 4.15 (m, 1H), 3.87 (t, 10.72 Hz, 1H),
3.67 (s, 3H), 3.00 (m, 1H), 2.81 (m, 2H). Mp 156–
ꢀ
67
ꢀ
158 C; reported: 158 C.
5.2.2. Synthesis of (Æ)-equol from monomethyl equol.
Aluminum trichloride (2.10 g, 15.7 mmol) was mixed
with ethane thiol (12 mL) together with 5 mL of anhy-
6
3À65
as well as the reports of
the ratio of ERa to ERb,
hyperplasia in certain target tissues in ERb knockout
mice, although the observations of all laboratories are
ꢀ
drous CH Cl at 0 C under nitrogen. The light yellow
2
2
6
3
solution was allowed to stir for 5 min. after which
monomethyl equol (1.0 g, 3.9 mmol) was added in one
6
6
not consistent on this latter issue. In light of our study
on the effects of equol and its enantiomers on the two
estrogen receptor subtypes, it would appear prudent to
evaluate carefully, in vivo, the biological effects of not
only the isoflavones but also their metabolites and their
stereoisomers. Such investigations would greatly help in
evaluating the potential effects of the ingestion of soy
isoflavones on human health and disease.
ꢀ
quenched by addition of water and the mixture was
warmed to 40 C and left to stir at that temperature,
portion. After stirring for 1 h at 0 C, the reaction was
ꢀ
with the flask open to air, until all traces of ethane thiol
were completely removed. A white solid which pre-
cipitated out was dissolved in EtOAc (20 mL), washed
with 10% HCl and dried on MgSO . The organic sol-
4
vent was removed in vacuo to give a light brown oil
which was chromatographed on silica gel using 35%
EtOAc/hexane as the eluent to give 0.78g (84%) of (Æ)-
4. Conclusion
1
equol. H NMR (500 MHz, acetone-d ): d 6.93 (d, 8.58
6
We have developed a simple synthetic route to equol from
readily available isoflavanoid starting materials. Chiral
HPLC provided ready access to the two enantiomers of
equol and allowed us to examine, for the first time, their
differential effects on the two estrogen receptor sub-
types. S-equol shows a strong binding preference in favor
of ERb that is greater than that of its daidzein precursor
and comparable to that of genistein, whereas the R iso-
mer exhibits a binding selectivity for ERa. In vitro
transfection assays show that the S isomer, however, has
little ERb selectivity in terms of transcriptional potency.
Hz, 2H), 6.75 (d, 8.15 Hz, 1H), 6.69 (d, 8.58 Hz, 2H),
6.27 (dd, 2.36 Hz, 8.15 Hz, 1H), 6.22(d, 2.36 Hz, 1H),
4.11 (m, 1H), 3.79 (t, 10.72 Hz, 1H), 2.97 (m, 1H), 2.75
1
3
(m, 2H). C NMR (125 MHz, acetone-d ): d 155.8,
6
154.5, 132.1, 129.8, 127.9, 115.2, 113.1, 107.8, 102.7,
ꢀ
ꢀ
7
70.7, 37.5, 31.5. Mp 184–187 C; reported 187–189 C.
5.2.3. Conversion of daidzein to (Æ)-equol. Daidzein (5.0
g, 18.6 mmol) was suspended in acetic acid (25 mL)
together with ammonium formate (11.75 g, 186 mmol)
and 20% Pd(OH) on carbon (4.8 g). The reaction mix-
2
ture was then refluxed for 1 h and allowed to cool to
room temperature. After diluting with 50 mL of EtOAc,
the reaction mixture was filtered through a 2 cm pad of
Celite that was subsequently washed with an additional
150 mL of EtOAc. The filtrate was evaporated in vacuo
to give a light brown oily residue that was purified on
silica gel using 30% EtOAc/hexane as the solvent. Tri-
turation of this oil with 125 mL of water gave a white
5. Experimental
5
.1. Materials
Formononetin was a gift from Dr. Muralee Nair,
Michigan State University, East Lansing, MI). Daidzein

1566
R. S. Muthyala et al. / Bioorg. Med. Chem. 12 (2004) 1559–1567
fluffy solid which was filtered and dried to give 2.9 g
1
conditions, the K of estradiol for ERa is ca. 0.2 nM,
and for ERb 0.5 nM. The determination of these RBA
values is reproducible in separate experiments with a
CV of 0.3, and the values shown represent the aver-
ageÆrange or SD of 2 or more separate determinations,
respectively.
d
28
(
61%) of (Æ)-equol. H NMR (500 MHz, acetone-d ): d
6
6
8
2
.93 (d, 8.58 Hz, 2H), 6.75 (d, 8.15 Hz, 1H), 6.69 (d,
.58 Hz, 2H), 6.27 (dd, 2.36 Hz, 8.15 Hz, 1H), 6.22(d,
.36 Hz, 1H), 4.11 (m, 1H), 3.79 (t, 10.72 Hz, 1H), 2.97
(
m, 1H), 2.75 (m, 2H). 13C NMR (125MHz, acetone-d6):
d 155.8, 154.5, 132.1, 129.8, 127.9, 115.2, 113.1, 107.8,
ꢀ
7
1
02.7, 70.7, 37.5, 31.5. Mp 184–187 C; reported: 187–
ꢀ
5.4. Cell culture and transcription activation assays
1
89 C.
Human endometrial carcinoma (HEC-1) cells were
3
0
5.2.4. Resolution of equol enantiomers by chiral liquid
chromatography (LC). Initial LC separation was per-
maintained in culture as described. Transfection of
HEC-1 cells in 24-well plates used a mixture of 0.35 mL
of serum-free Improved Minimal Essential medium
(IMEM) and 0.15 mL of Hank’s Balanced Salt Solution
(HBSS) containing 5 mL of lipofectin (Life Technolo-
gies, Rockville, MD), 1.6 mg of transferrin (Sigma, St.
Louis, MO), 0.5 mg of pCMVb-galactosidase as internal
control, 1 mg of the reporter gene plasmid, 100 ng of ER
expression vector, and carrier DNA to a total of 3 mg
formed using an HP1050 quaternary pumping system
(
Hewlett Packard, Palo Alto, CA) and a Dionex
AS3500 autosampler (Dionex, Sunnyvale, CA) attached
to a PC1000 data handling system (Thermo Separation
Products, San Jose, CA). UV detection was provided by
a Spectraphysics MWD (Thermo Separation Products,
San Jose, CA) at a wavelength of 280 nm. Enantiomeric
separation was achieved using a Cyclobond I 2000 RSP
b-cyclodextrin-R,S-hydroxypropyl ether chiral sta-
tionary phase (analytical column 250Â2.0 mm, 5 mm
particle size; semi-preparative column 250Â10 mm, 5
mm, particle size, Astec, NJ, USA) and an isocratic
mobile phase consisting of 65% 0.1% formic acid (aq)
and 35% of acetonitrile. Respective flow rates for the
analytical and semi-preparative columns were 0.25 and
ꢀ
DNA per well. The cells were incubated at 37 C in a
5% CO containing incubator for 6 h. The medium was
2
then replaced with fresh medium containing 5% char-
coal-dextran treated calf serum and the desired con-
centrations of ligands. Reporter gene activity was
assayed at 24 h after ligand addition. Luciferase activ-
ity, normalized for the internal control b-galactosidase
3
0
activity, was assayed as described.
5
mL/min. Injection volumes ranged from 5–100 mL,
and all separations were performed at ambient temper-
ature.
Acknowledgements
5
.2.5. Mass spectrometry. LC with tandem mass spec-
trometry (LC-MS/MS) was conducted using a Waters
790 liquid handling system (Waters Assoc., Milford,
This work was supported by National Institutes of
Health grants: PHS 5R37 DK15556 (to J. A. K.), 5R01
CA19118 (to B. S. K.), and 5R01 CA77355 (to W. G.
H.), and by a grant from the Breast Cancer Research
Foundation (to B. S. K.). We thank Kathryn Carlson
for performing the binding assays.
2
MA) and a Quattro Ultima triple quadrupole mass
spectrometer equipped with an electrospray (ES) inter-
face (Micromass, Manchester, UK). The entire column
effluent (0.25 mL/min) was directed into the ion source
held at 120 C for acquisition of positive ions in the
product ion or multiple reaction monitoring modes
ꢀ
References and notes
6
8
(
MRM), as previously described.
1
. Adlercreutz, H.; Mousavi, Y.; Clark, J.; Hockerstedt, K.;
Hamalainen, E.; Wahala, K.; Makela, T.; Hase, T. J.
Steroid. Biochem. Mol. Biol. 1992, 41, 331.
5
.2.6. Circular dichroism (CD) spectrophotometry. CD
spectra of separated equol enantiomers were obtained at
ambient temperature by using a Jasco 500A spectro-
polarimeter equipped with a quartz cell (1 cm path
length) containing equol solutions in 50% aqueous
methanol that produced 1 AU in a spectrophotometer.
The spectropolarimeter was calibrated using (1S)-(+)-
2
3
. Lamartiniere, C. A. Am. J. Clin. Nutr. 2000, 71, 1705S.
. Ju, Y. H.; Doerge, D. R.; Allred, K. F.; Allred, C. D.;
Helferich, W. G. Cancer Res. 2002, 62, 2474.
. Morito, K.; Hirose, T.; Kinjo, J.; Hirakawa, T.; Okawa,
M.; Nohara, T.; Ogawa, S.; Inoue, S.; Muramatsu, M.;
Masamune, Y. Biol. Pharm. Bull. 2001, 24, 351.
. Setchell, K. D.; Brown, N. M.; Lydeking-Olsen, E. J.
Nutr. 2002, 132, 3577.
4
5
1
0-camphorsulfonic acid before use.
6
7
. Chang, Y.-C.; Nair, M. G. J. Nat. Prod. 1995, 58, 1892.
. Marrian, G. F.; Haslewood, G. A. D. Biochem. J. 1932,
26, 1227.
. Shutt, D. A.; Braden, A. W. H. Aust. J. Agric. Res. 1968,
19, 545.
5
.3. Estrogen receptor binding assays
Relative binding affinities were determined by competi-
3
8
tive radiometric binding assays using 10 nM [ H]E as
2
2
8,49,69
tracer, using methods previously described.
source of ER was purified full-length human ERa and
ERb purchased from Pan Vera (Madison, WI).
Incubations were done at 0 C for 18–24 h, and hydro-
The
9
. Axelson, M.; Kirk, D. N.; Farrant, R. D.; Cooley, G.;
Lawson, A. M.; Setchell, K. D. Biochem. J. 1982, 201, 353.
2
8,49,69
1
1
0. Rowland, I. R.; Wiseman, H.; Sanders, T. A.; Adler-
creutz, H.; Bowey, E. A. Nutr. Cancer 2000, 36, 27.
1. Xu, X.; Harris, K. S.; Wang, H. J.; Murphy, P. A.; Hen-
drich, S. J. Nutr. 1995, 125, 2307.
ꢀ
xylapatite was used to absorb the purified receptor-
2
8,49,69
ligand complexes (human ERs).
The binding affi-
nities are expressed as relative binding affinity (RBA)
values, where the RBA of estradiol is 100%; under these
12. Lampe, J. W.; Karr, S. C.; Hutchins, A. M.; Slavin, J. L.
Proc. Soc. Exp. Biol. Med. 1998, 217, 335.

R. S. Muthyala et al. / Bioorg. Med. Chem. 12 (2004) 1559–1567
1567
1
3. Setchell, K. D.; Brown, N. M.; Desai, P.; Zimmer-
Nechemias, L.; Wolfe, B. E.; Brashear, W. T.; Kirschner,
A. S.; Cassidy, A.; Heubi, J. E. J. Nutr. 2001, 131, 1362S.
4. Shutt, D. A.; Cox, R. I. J. Endocr. 1972, 52, 299.
45. Lin, X.; Huebner, V. Curr. Opin. Drug Disc. 2000, 3, 383.
46. Sun, J.; Baudry, J.; Katzenellenbogen, J. A.; Katze-
nellenbogen, B. S. Mol. Endocrinol. 2003, 17, 247.
47. Meyers, M. J.; Sun, J.; Carlson, K. E.; Katzenellenbogen,
B. S.; Katzenellenbogen, J. A. J. Med. Chem. 1999, 42, 2456.
48. Meyers, M. J.; Sun, J.; Carlson, K. E.; Marriner, G. A.;
Katzenellenbogen, B. S.; Katzenellenbogen, J. A. J. Med.
Chem. 2001, 44, 4230.
1
1
5. Schmitt, E.; Dekant, W.; Stopper, H. Toxicol. In Vitro
2
001, 15, 433.
6. Pettersson, K.; Gustafsson, J. A. Annu. Rev. Physiol.
001, 63, 165.
˚
1
2
1
1
7. Wessely, F.; Prillinger, F. Berichte 1939, 72B, 629.
8. Lamberton, J. A.; Suares, H.; Watson, K. G. Australian J.
Chem. 1978, 31, 455.
49. Stauffer, S. R.; Coletta, C. J.; Tedesco, R.; Sun, J.; Kat-
zenellenbogen, B. S.; Katzenellenbogen, J. A. J. Med.
Chem. 2000, 43, 4934.
19. Ram, S.; Ehrenkaufer, R. E. Synthesis 1988, 91.
20. Ram, S.; Spicer, L. D. Tetrahedron Lett. 1988, 29, 3741.
21. Wahala, K.; Hase, T. A. Heterocycles 1989, 28, 183.
22. Pearlman, W. M. Tetrahedron Lett. 1967, 1663.
23. Node, M.; Nishide, K.; Fuji, K.; Fujita, E. J. Org. Chem.
50. Kuiper, G. G. J. M.; Carlsson, B.; Grandien, K.; Enmark,
˚
¨
E.; Haggblad, J.; Nilsson, S.; Gustafsson, J.-A. Endocri-
nology 1997, 138, 863.
51. Versteeg, M.; Bezuidenhoudt, B. C. B.; Ferreira, D. Tet-
rahedron 1999, 55, 3365.
1
980, 45, 4275.
52. Katzenellenbogen, J. A.; O’Malley, B. W.; Katze-
nellenbogen, B. S. Mol. Endocrinol. 1996, 10, 119.
53. McKenna, N. J.; O’Malley, B. W. J. Steroid Biochem.
Mol. Biol. 2000, 74, 351.
2
2
4. Yunnikova, L. P.; Tigina, O. V. Zhurnal Organicheskoi
Khimii 1993, 29, 651.
5. Srikrishna, A.; Reddy, T. J.; Viswajanani, R. Tetrahedron
1
996, 52, 1631.
54. McKenna, N. J.; O’Malley, B. W. Endocrinology 2002,
143, 2461.
55. McKenna, N. J.; O’Malley, B. W. Cell 2002, 108, 465.
56. Commitee on Toxicity: Phytoestrogens and Health The
Food Standards Agency; Aviation House: London, 2003;
pp 154–164..
2
2
6. Verbit, L.; Clark-Lewis, J. W. Tetrahedron 1968, 24, 5519.
7. Kurosawa, K.; Ollis, W. D.; Redman, B. T.; Sutherland,
I. O. Chem. Commun. (London) 1968, 20, 1265.
8. Carlson, K. E.; Choi, I.; Gee, A.; Katzenellenbogen, B. S.;
Katzenellenbogen, J. A. Biochemistry 1997, 36, 14897.
9. Pike, A. C.; Brzozowski, A. M.; Hubbard, R. E.; Bonn,
T.; Thorsell, A. G.; Engstrom, O.; Ljunggren, J.; Gus-
2
2
57. Fang, H.; Tong, W.; Shi, L. M.; Blair, R.; Perkins, R.;
Branham, W.; Hass, B. S.; Xie, Q.; Dial, S. L.; Moland,
C. L.; Sheehan, D. M. Chem. Res. Toxicol. 2001, 14, 280.
58. Kuiper, G. G.; Enmark, E.; Pelto-Huikko, M.; Nilsson,
˚
tafsson, J. A.; Carlquist, M. EMBO J. 1999, 18, 4608.
3
3
0. Sun, J.; Meyers, M. J.; Fink, B. E.; Rajendran, R.; Kat-
zenellenbogen, J. A.; Katzenellenbogen, B. S. Endocrinol-
ogy 1999, 140, 800.
˚
S.; Gustafsson, J. A. Proc. Natl. Acad. Sci. U.S.A. 1996,
93, 5925.
1. Kuiper, G. G.; Lemmen, J. G.; Carlsson, B.; Corton, J. C.;
Safe, S. H.; van der Saag, P. T.; van der Burg, B.; Gus-
59. Mosselman, S.; Polman, J.; Dijkema, R. FEBS Lett. 1996,
392, 49.
60. Couse, J. F.; Korach, K. S. Endocr. Rev. 1999, 20, 358.
61. Couse, J. F.; Curtis Hewitt, S.; Korach, K. S. J. Steroid
Biochem. Mol. Biol. 2000, 74, 287.
˚
tafsson, J. A. Endocrinology 1998, 139, 4252.
3
3
2. Mazur, W. Baillieres Clin. Endocrinol. Metab. 1998, 12, 729.
3. Horn-Ross, P. L.; Lee, M.; John, E. M.; Koo, J. Cancer
Causes Control 2000, 11, 299.
62. Harris, H. A.; Katzenellenbogen, J. A.; Katzenellenbo-
gen, B. S. Endocrinology 2002, 143, 4172.
63. Lindberg, M. K.; Moverare, S.; Skrtic, S.; Gao, H.;
3
4. Tham, D. M.; Gardner, C. D.; Haskell, W. L. J. Clin.
Endocrinol. Metab. 1998, 83, 2223.
˚
3
3
5. Chang, Y. C.; Nair, M. G. J. Nat. Prod. 1995, 58, 1892.
6. Chang, Y. C.; Nair, M. G.; Nitiss, J. L. J. Nat. Prod.
Dahlman-Wright, K.; Gustafsson, J.A.; Ohlsson, C. Mol.
Endocrinol. 2003, 17, 203.
1
995, 58, 1901.
64. Jensen, E. V.; Cheng, G.; Palmieri, C.; Saji, S.; Makela,
S.; Van Noorden, S.; Wahlstrom, T.; Warner, M.;
3
3
3
7. Hawksworth, G.; Drasar, B. S.; Hill, M. J. J. Med.
Microbiol. 1971, 4, 451.
˚
Coombes, R. C.; Gustafsson, J.A. Proc. Natl. Acad. Sci.
8. Xu, X.; Duncan, A. M.; Merz, B. E.; Kurzer, M. S. Can-
cer Epidemiol. Biomarkers Prev. 1998, 7, 1101.
9. Rowland, I.; Faughnan, M.; Hoey, L.; Wahala, K.; Wil-
liamson, G.; Cassidy, A. Br. J. Nutr. 2003, 89 (Suppl. 1),
U.S.A. 2001, 98, 15197.
65. Palmieri, C.; Cheng, G. J.; Saji, S.; Zelada-Hedman, M.;
Warri, A.; Weihua, Z.; Van Noorden, S.; Wahlstrom, T.;
˚
Coombes, R. C.; Warner, M.; Gustafsson, J.A. Endocr.
Relat. Cancer 2002, 9, 1.
838.
4
0. Setchell, K. D.; Borriello, S. P.; Hulme, P.; Kirk, D. N.;
Axelson, M. Am. J. Clin. Nutr. 1984, 40, 569.
1. Sathyamoorthy, N.; Wang, T. T. Eur. J. Cancer 1997, 33,
66. Prins, G. S.; Birch, L.; Couse, J. F.; Choi, I.; Katze-
nellenbogen, B.; Korach, K. S. Cancer Res. 2001, 61,
6089.
4
2
384.
67. Dudley, K. H.; Miller, H. W.; Corley, R. C.; Wall, M. E.
J. Org. Chem. 1967, 32, 2317.
68. Twaddle, N. C.; Churchwell, M. I.; Doerge, D. R. J.
Chromatogr. B Analyt. Technol. Biomed. Life Sci. 2002,
777, 139.
4
4
2. Lampe, J. W. J. Nutr. 2003, 133 (Suppl. 3.), 956S.
3. Duncan, A. M.; Merz-DemLow, B. E.; Xu, X.; Phipps,
W. R.; Kurzer, M. S. Cancer Epidemiol. Biomarkers Prev.
2
000, 9, 581.
4. Ingram, D.; Sanders, K.; Kolybaba, M.; Lopez, D. Lancet
997, 350, 990.
4
69. Katzenellenbogen, J. A.; Johnson, H. J., Jr.; Myers, H. N.
Biochemistry 1973, 12, 4085.
1