Estrogen receptor subtype-selective ligands: Asymmetric synthesis and biological evaluation of cis- and trans-5,11-dialkyl-5,6,11,12- tetrahydrochrysenes

Article abstract of DOI:10.1021/jm990101b

We have recently reported that racemic 5,11-cis-diethyl-5,6,11,12- tetrahydrochrysene-2,8-diol (THC, rac-2b) acts as an agonist on estrogen receptor alpha (ERα) and as a complete antagonist on estrogen receptor beta (ERβ) (Sun et al. Endocrinology 1999, 1

Full text of DOI:10.1021/jm990101b

2456
J . Med. Chem. 1999, 42, 2456-2468
Estr ogen Recep tor Su btyp e-Selective Liga n d s: Asym m etr ic Syn th esis a n d
Biologica l Eva lu a tion of cis- a n d tr a n s-5,11-Dia lk yl-
5,6,11,12-tetr a h yd r och r ysen es
Marvin J . Meyers,† J un Sun,‡ Kathryn E. Carlson,† Benita S. Katzenellenbogen,‡,§ and
J ohn A. Katzenellenbogen*,†
Departments of Chemistry, Molecular and Integrative Physiology, and Cell and Structural Biology, University of Illinois,
Urbana, Illinois 61801
Received March 5, 1999
We have recently reported that racemic 5,11-cis-diethyl-5,6,11,12-tetrahydrochrysene-2,8-diol
(THC, rac-2b) acts as an agonist on estrogen receptor alpha (ERR) and as a complete antagonist
on estrogen receptor beta (ERâ) (Sun et al. Endocrinology 1999, 140, 800-804). To further
investigate this novel ER subtype-selective estrogenic activity, we have synthesized a series of
cis- and trans-dialkyl THCs. cis-Dimethyl, -diethyl, and -dipropyl THCs 2a -c were prepared
in a highly enantio- and diastereoselective manner by the acyloin condensation of enantio-
merically pure R-alkyl-â-arylpropionic esters, followed by a Lewis acid-mediated double cycli-
zation under conditions of minimal epimerization. ERR and ERâ binding affinity of both cis
and trans isomers of dimethyl, diethyl, and dipropyl THCs was determined in competitive
binding assays, and their transcriptional activity was determined in reporter gene assays in
mammalian cells. Nearly all THCs examined were found to be affinity-selective for ERâ. All
these THCs are agonists on ERR, and THCs with small substituents are agonists on both ERR
and ERâ. As substituent size was increased, ERâ-selective antagonism developed first in the
(R,R)-cis enantiomer series and finally in the trans diastereomer and (S,S)-cis enantiomer series.
The most potent and selective ligand was identified as (R,R)-cis-diethyl THC 2b, which
mimicked the ERâ-selective antagonist character of racemic cis-diethyl THC 2b. This study
illustrates that the antagonist character in THC ligands for ERâ depends in a progressive
way on the size and geometric disposition of substituent groups and suggests that the induction
of an antagonist conformation in ERâ can be achieved with these ligands with less steric
perturbation than in ERR. Furthermore, antagonists that are selectively effective on ERâ can
have structures that are very different from the typical antiestrogens tamoxifen and raloxifene,
which are antagonists on both ERR and ERâ.
In tr od u ction
extensive efforts are being made to develop ligands
which selectively antagonize undesirable estrogenic
effects, such as the stimulation of breast cancer, while
promoting positive estrogen effects in maintaining bone
and cardiovascular health.
The estrogen receptor (ER), a member of the nuclear
hormone receptor superfamily, mediates the activity of
estrogens in the regulation of a number of important
physiological processes, including the development and
function of the female reproductive system and the
maintenance of bone mineral density and cardiovascular
health. While the stimulation of processes in these
tissues has important health benefits, the stimulation
of other tissues, such as the breast and uterus, can
increase the risk of cancer at these sites. Intriguingly,
some pharmaceutical agents, such as tamoxifen and
raloxifene, act as antagonists in some tissues, such as
the breast and uterus, while acting as agonists in other
tissues, such as the liver and vasculature.1 This mixed
response has raised the interesting issue of tissue-, cell-,
and gene-specific activity of estrogens based on the
ligand, the receptor, and the effector site, which has
been termed “tripartite receptor pharmacology”.2 Today,
Until recently, it had been assumed that estrogen-
related events were mediated by only one estrogen
receptor. However, the recent discovery of a second
estrogen receptor (ERâ)3,4 has suggested that tissue and
cell selectivity of certain estrogens may be due, in part,
to their action through ERâ, separate from or in
conjunction with the classical estrogen receptor (ERR).
This possibility has been supported by the difference
in tissue distribution between ERR and ERâ.3,5-7 Fur-
thermore, it has been shown that the pharmacology of
traditional ER agonists and antagonists is reversed for
ERâ in the context of certain ER effector sites.8,9
Although the two ER subtypes are both activated by
binding estradiol (1, E2), the ligand-binding domain
(LBD) and activation function-2 (AF-2) region of the
proteins are only 56% conserved and the A/B domain/
activation function-1 (AF-1) is only 20% conserved.3,4
This suggests that ligands may be developed which have
different affinities, potencies, and agonist versus an-
tagonist behavior for the two ER subtypes. Indeed, some
* Address correspondence to: J ohn A. Katzenellenbogen, Depart-
ment of Chemistry, University of Illinois, 461 Roger Adams Laboratory,
Box 37-5, 600 S. Mathews Ave., Urbana, IL 61801. Tel: (217) 333-
6310. Fax: (217) 333-7325. E-mail: jkatzene@uiuc.edu.
† Department of Chemistry.
‡ Department of Molecular and Integrative Physiology.
§ Department of Cell and Structural Biology.
10.1021/jm990101b CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/29/1999

Estrogen Receptor Subtype-Selective Ligands
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 13 2457
Sch em e 2
known ligands have been shown to have subtype-
selective affinities and a degree of subtype-selective
agonist/antagonist character.5,10
Sch em e 1
In previous work, we have described the synthesis of
trans-THC isomers 3b,c and racemic cis isomers 2a ,b
by the same approach. However, we used racemic esters,
and we did not evaluate the extent of epimerization that
occurred when PPA was used as the cyclization re-
agent.13 By this earlier approach, the dimethyl THC was
obtained primarily as the cis isomer, but the diethyl and
dipropyl isomers were obtained as mixtures of cis and
trans isomers. Even though the cis- and trans-diethyl
THCs were separable by recrystallization (and their
stereochemistry confirmed by X-ray crystallographic
analysis), we desired a cis-selective route to each of the
dialkyl-substituted THCs. We surmised that if epimer-
ization could be avoided throughout the sequence, then
the asymmetric strategy presented in Scheme 1 should
be a feasible means for preparing the cis isomers
diastereoselectively.
Asym m etr ic Syn th esis of R-Alk yl Ester s 7a -c.
Although several methods are available for the asym-
metric R-alkylation of carboxylic acid derivatives, we
decided to use Myers’ pseudoephedrine chiral auxiliary
method because of its ease of use, high yields, and high
diastereoselectivities.14 The acylation of pseudoephe-
drine amines 4 with the appropriate anhydrides pro-
ceeded in excellent yield, to afford amides 5a -c (Scheme
2). Enolate formation from 5a -c with LDA and asym-
metric alkylation with m-methoxybenzyl bromide gave
amides 6a -c in excellent yields. In our hands, however,
many of the amides 5a -c and 6a -c were thick viscous
oils, which prevented the enrichment of diastereomeric
ratios by crystallization.
We have recently surveyed a number of compounds
previously synthesized in our laboratories and found
that the racemic cis-diethyltetrahydrochrysene (THC)
(rac-2b) is an agonist on ERR and a complete antagonist
on ERâ and has a 10-fold higher affinity for ERâ relative
to ERR.11 Furthermore, the (R,R)-enantiomer of cis-
diethyl THC was found to be a more potent antagonist
than rac-2b, while the (S,S)-enantiomer was an agonist
on both receptor subtypes.11 The work described herein
details the asymmetric synthesis of a series of dialkyl-
substituted cis-THCs and the evaluation of both the
relative binding affinity (RBA) and agonist/antagonist
selectivity of both cis-THCs (2a -c) and trans-THCs
(3a -c). From an analysis of the activity of the whole
series of compounds, it is clear that the (R,R)-enanti-
omers develop ERâ-selective antagonism with smaller
substituents than do the (S,S)-enantiomers and trans
isomers. This suggests that the induction of an antago-
nist conformation in ERâ can be achieved with these
ligands with less steric perturbation than in ERR.
The hydrolysis of amides 6a -c proved to be challeng-
ing (Scheme 3). Initial attempts to effect methanolysis
with sulfuric acid and methanol furnished ester (R)-7a
in good yields after 3 h at reflux; however, as the alkyl
substituent increased in steric bulk, the hydrolysis
became quite sluggish (e.g., ester (S)-7b was furnished
in only 20% yield after 5 h at reflux). Thus, two methods
described by Myers and co-workers14 were utilized to
prepare the crude carboxylic acids. To prepare esters
7b, a one-pot MsOH/LiBH4/n-Bu4NOH hydrolysis14 of
amides 6b provided the crude carboxylic acids, which,
after a simple workup, were then treated with excess
Resu lts a n d Discu ssion
Retr osyn th esis of cis-Dia lk yl THCs. Following a
strategy previously described,12,13 we reasoned that the
acyloin condensation of an enantiomerically pure ester,
followed by a double Friedel-Crafts cyclization under
conditions of minimal epimerization, would selectively
furnish the cis-THC isomer as a single enantiomer
(Scheme 1).

2458 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 13
Meyers et al.
Ta ble 1. Optimization of the Cyclization of Silyl Ether
Sch em e 3
(R,R)-8a
entry
reagents
PPA, neat
conditions
cis:transa
1
2
3
4
5
rt, 1 h
2:1
2:1
2:1
9:1
13:1
PPA, neat
rt, 4 min
0 °C, 2 h
rt, 23 h
rt, 3 h
10 equiv MsOH, CH2Cl2
10 equiv BF3‚OEt2, CH2Cl2
2 equiv TiCl4, CH2Cl2
a
Determined by integration of 1H NMR peaks.
Ta ble 2. Cyclization of Silyl Ethers 8a -c with TiCl4
yield,
%
crude
recrystallized
cis:transa
product
R
cis:transa
erb
(R,R)-9a
(S,S)-9a
(R,R)-9b
(S,S)-9b
(R,R)-9c
(S,S)-9c
Me
Me
Et
Et
Pr
50
42
71
65
69
74
13:1
NDc
15:1
64:1
66:1
>100:1
>100:1
NAd
98:2
>99:1
>99:1
>99:1
>99:1
>99:1
Sch em e 4
18:1
∼20:1
Pr
∼20:1
NA
Determined by integration of 1H NMR peaks. Determined
a
b
by CSP HPLC (R,R Whelk-O 1). c ND, not determined. NA, not
d
applicable.
occurs very quickly (entry 2). To find a reagent which
would avoid this epimerization, we surveyed a number
of nonprotic Lewis acids (AlCl3, BF3-OEt2, TiCl4, POCl3,
ZnCl2, ZnCl2/POCl3, LiClO4, and SnCl4) with silyl ether
rac-8b, prepared as previously described.13 TiCl4 and
BF3-etherate were found to be optimal reagents, in
terms of their reactivity, cleanliness of reaction, and
stereoselectivity, with the former being considerably
more efficient and consistent than the latter. Both
reagents effected the cyclization of (R,R)-8a to dimethyl
THC (R,R)-9a with high stereoselectivity (entries 4 and
5). It is notable that the ratio of 13:1 cis:trans for THC
(R,R)-9a amounts to only 3-4% overall epimerization
in the acyloin condensation and cyclization steps.
Cyclization of silyl ethers 8a -c with TiCl4 (2 equiv)
furnished optically active cis-THCs 9a -c in moderate
to excellent yields with minimal epimerization, as
evident from their high cis:trans ratios and enantiopu-
rities (Table 2). The stereochemical assignments and
ratios were determined by comparison of 1H NMR
spectra with known cis- and trans-THC isomers (see
below). The effect of epimerization, which produces the
trans isomer, could be reduced by careful recrystalliza-
tion. In the case of diethyl THCs 9b, all of the trans
isomer was removed by recrystallization, as ascertained
diazomethane to furnish esters 7b in moderate yields
and excellent enantiomeric ratios (er g 98:2). Alterna-
tively, hydrolysis of amides (R,R,S)-6a and 6c with
aqueous sulfuric acid in dioxane followed by methylation
with diazomethane also furnished esters (S)-7a and 7c
in moderate to good yields (53-77%). Presumably, the
yields in these hydrolyses could be improved by extend-
ing the hydrolysis times; however, Myers has shown
that this is likely to compromise the enantiopurity.14
The absolute stereochemistry of esters 7a -c was as-
signed by analogy to (R)-R-methylbenzenepropionic acid,
prepared by Myers by an analogous route.14
1
Acyloin Con d en sa tion . The acyloin coupling of
esters 7a -c proceeded in good yields to afford silyl
ethers 8a -c with minimal, if any, epimerization, as
by H NMR. Unfortunately, attempts to crystallize the
propyl analogues (9c) were not successful. Handling of
these compounds required care, as these THCs are
sensitive to light and air (see below).13
1
assessed by H NMR (Scheme 4). TMSCl was used to
trap the enediolate as the bis-silyl ether; it evidently
also functioned as a methoxide scavenger,15 thus mini-
mizing epimerization of the product. An attempt to
couple the aldehyde corresponding to ester (S)-7a using
Stetter chemistry to provide the acyloin16 was unsuc-
cessful, presumably because of steric hindrance by the
R-alkyl substituent.17 After minimal purification, the
silyl ethers were cyclized with the appropriate acid to
furnish THCs 9a -c (see below).
Dou ble Cycliza tion . Initial cyclization attempts of
silyl ether (R,R)-8a with PPA and with MsOH yielded
9a as a 2:1 mixture of cis:trans isomers (entries 1-3,
Table 1). When the cyclization was quenched at about
50% completion, the same cis:trans ratio was observed,
suggesting that, under these conditions, epimerization
Deprotection of methyl ethers 9b (100% cis) with
AlBr3 and EtSH furnished the desired phenols, but with
considerable epimerization (approximately 10-15%),
presumably because of the strong protic acid conditions.
The bis-phenol products obtained by this method were
also quite sensitive to oxygen and light. During workup
and purification by chromatography, the products rap-
idly turned yellow-orange, presumably due to oxidation.
However, the extent of this decomposition proved to be
minimal (less than 5%), and the decomposition products
could be removed by chromatography and recrystalli-
zation in the presence of trace levels of ascorbic acid
added as an antioxidant.
Fortunately, deprotection with boron tribromide proved
to be much more suitable, as no epimerization was

Estrogen Receptor Subtype-Selective Ligands
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 13 2459
Ta ble 3. Deprotection of THCs 9a -c with BBr3
racemic dimethylsilyl ether rac-8a with PPA was orig-
inally thought to give predominantly the trans-dimethyl
THC 3a ;13 this original report was incorrect. The 1H
NMR spectra of the major isomer produced in all of the
cyclizations presented in Table 1 match what had
previously been reported as the trans-dimethyl isomer
(Table 4, entries 1-3). However, it is now clear that the
major dimethyl THC isomer is cis, because the starting
ester was of high enantiopurity and the THC is formed
with minimal epimerization. In the major isomer, the
chemical shifts of the allylic protons are shifted upfield
(0.26 ppm), the methyl protons downfield (0.17 ppm),
and the meta ArH protons upfield (0.04 ppm) relative
to the minor isomer (entries 1 and 3). This matches the
pattern seen in the diethyl THC series for the cis isomer
relative to the trans isomer (entries 4-6, stereochem-
istry confirmed by X-ray crystallography).13 In addition,
the product (9a ) of the cyclization with TiCl4 was
recrystallized to enhance the isomer ratio to 64:1. This
crude
recrystallized
product
R
yield, % cis:transa
cis:transa
erb
(R,R)-2a
(S,S)-2a
(R,R)-2b
(S,S)-2b
(R,R)-2c
(S,S)-2c
Me
Me
Et
Et
Pr
100
88
78
100
93
79
NDc
ND
100:0
100:0
33:1
100:0
83:1
100:0
100:0
73:1
98:2
>99:1
>99:1
>99:1
>99:1
>99:1
Pr
23:1
67:1
a
b
Determined by integration of 1H NMR peaks. Determined
by CSP HPLC (ChiralPak AS). c ND, not determined.
Ta ble 4. Selected 1H NMR Chemical Shifts for cis- and
trans-THC Isomers
entry
THC
isomer
δ a, ppm
δ b, ppm
δ c, ppm
purified compound was found to be optically active
24
([R]D
-109° (c 0.98, CHCl3)), which also distin-
1
2
3
4
5
6
7
8
(R,R)-9a
dimethyl
trans-9a
(R,R)-9b
rac-9b
diethyl
(R,R)-9c
dipropyl
cis
2.92
2.90
3.19
2.58
2.59
2.94
2.72
2.98
1.07
1.05
0.90
0.96
0.97
0.79
0.93
0.78
7.31
7.30
7.35
7.23
7.23
7.33
7.27
7.34
cisa
guishes it from the trans isomer, a meso compound,
which is optically inactive. The cis-dipropyl THC 9c gave
spectra which were shifted in the same fashion from the
previously described trans-dipropyl THC,13 as described
above for 9a ,b (entries 7 and 8).
transb
cis
cisc
transc,d
cis
transd
The synthesis of trans-3a was accomplished by the
cyclization of (R,R)-8a with PPA to furnish a 1:2 ratio
of trans:cis isomers of 9a in 60% yield. Since this
isomeric mixture was inseparable by recrystallization,
the mixture was deprotected with BBr3 and recrystal-
lized twice from MeOH to provide exclusively trans-3a .
a
b
Reported as trans,13 actually cis. Minor isomer indicated in
Table 1. c Stereochemistry confirmed by X-ray crystallography.13
Data from compounds prepared previously.13
d
detected in any of the bis-phenolic products 2a -c
(Scheme 4 and Table 3). Also, at the bis-phenol stage,
the cis:trans ratio of the propyl analogues could be
improved to approximately 70:1 by recrystallization.
Curiously, decomposition problems of the bis-phenols
were much less significant when they were generated
by the BBr3 deprotection than by the AlBr3/EtSH
deprotection. Thus, both the (R,R)- and (S,S)-enanti-
omers of the cis-dimethyl, -diethyl, and -dipropyl THC
diols were prepared in high enantio- and diastereose-
lectivity (er g 98:2, dr > 98:2 cis:trans).
Estr ogen Recep tor Bin d in g Affin ities. THC diols
2a -c, 3a -c, and unsubstituted THC 10 (prepared as
previously described)13 were evaluated in competitive
radiometric binding assays to determine their affinities
for the ER (Table 5).18,19 Relative binding affinity (RBA)
values were determined with lamb uterine cytosol ER
preparations and with purified full-length human ERR
and ERâ, and they are reported relative to estradiol (E2),
which is set at 100%.
THC Ster eoch em istr y Con fir m a tion a n d Syn -
th esis of tr a n s-Dim eth yl THC 3a . The cyclization of
All of the cis-THCs examined show a binding prefer-
ence for ERâ relative to ERR (â/R ratio 3-16), with some
Ta ble 5. Relative Binding Affinitiesa of THCs for the Estrogen Receptors R and â
entry
ligand
R
uterine ER
hERR
100
3.0 ( 0.7
222 ( 18
24 ( 6
hERâ
100
6.5 ( 0.3
254 ( 57
76 ( 18
75 ( 12
432 ( 21
67 ( 5
144 ( 5
14 ( 4
92.3 ( 4.5
26 ( 10
5.1 ( 4.0
â/R ratio
1
2
3
4
5
6
7
8
9
estradiol
10b
H
H
100
8.1 ( 0.5
252 ( 25
13 ( 3
2
1
3
8
2
5
6
16
3
5
trans-3a
(R,R)- 2a
(S,S )-2a
trans-3bb
rac-2bb
(R,R)-2b
(S,S)-2b
trans-3cb
(R,R)-2c
(S,S)-2c
Me
Me
Me
Et
Et
Et
Et
Pr
Pr
Pr
3.2 ( 0.2
137 ( 16
14 ( 5
9.3 ( 2.7
221 ( 42
14 ( 3
17 ( 7
23 ( 11
1.5 ( 0.4
59 ( 16
4.5 ( 0.7
1.4 ( 0.3
0.9 ( 0.2
33.6 ( 2.8
5.2 ( 1.0
1.6 ( 0.4
10
11
12
3
a
Determined by a competitive radiometric binding assay with [3H]estradiol; cytosol preparations of lamb uterus or full-length human
ERR and ERâ (PanVera) were used; see Experimental Section.18,19 Values are reported as the mean ( SD (n > 2) or range (n ) 2) under
these conditions; the Kd for estradiol is 0.3 nM. Prepared as previously described.13
b

2460 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 13
Meyers et al.
of the (S,S)-enantiomers showing high â/R selectivity.
The (R,R)-enantiomers, with lower â/R subtype selec-
tivities (â/R ratio 3-6), have up to 26-fold greater
absolute affinities for either ER subtype relative to their
(S,S)-enantiomers. The difference in relative affinities
between the (R,R)- and (S,S)-enantiomers is especially
apparent in the diethyl series, where (R,R)-2b has a 10-
fold greater affinity for ERâ and a 26-fold greater
affinity for ERR relative to (S,S)-2b. All of the trans-
THC isomers 3a -c have significantly greater affinities
for both receptor subtypes than the corresponding cis
isomers (∼4-16-fold). However, the trans isomers show
only minimal ERâ subtype selectivity (â/R ratio 1-3).
The unsubstituted THC 10 (Table 5, entry 2) also shows
minimal ERâ selectivity (∼2-fold). The RBAs from
uterine cytosol ER preparations match the RBA values
for full-length ERR, as expected, because uterine ER is
predominantly ERR.3,5
Tr a n scr ip tion a l Activa tion Assa ys. The transcrip-
tional activities of THCs 10, 2a -c, and 3a -c were
assayed in human endometrial cancer (HEC-1) cells
with ERR and ERâ and an estrogen-responsive reporter
gene construct, (ERE)3-pS2-CAT, containing these es-
trogen response elements, the estrogen-responsive pS2
promoter, and the chloramphenicol acetyltransferase
reporter gene. Activities are normalized to that of 10-8
M E2 which is set at 100, as previously described.11
Dose-response curves are presented in Figure 1. An-
tagonist activity of all of the THCs was determined with
1 nM E2 and 1 µM THC diol concentrations; those that
showed significant antagonist activity were subse-
quently assayed over a larger concentration range.
trans-dipropyl 3c is surprising, given the agonist char-
acter of the (S,S)-2b and trans-3b. To the best of our
knowledge, cis-diethyl (R,R)-2b, both cis-dipropyl THC
enantiomers 2c, and trans-dipropyl THC 3c are the first
compounds described as pure ERâ-selective antagonists.
Discu ssion
E n a n t ioselect ive Syn t h esis of cis-5,11-Dia lk yl-
5,6,11,12-tetr a h yd r och r ysen e-2,8-d iols. We have de-
scribed a method for the asymmetric synthesis of cis-
5,11-dialkyl-5,6,11,12-tetrahydrochrysene-2,8-diols in a
six-step sequence utilizing Myers’ pseudoephedrine
chiral auxiliary methodology.14 The crucial acyloin and
cyclization steps were achieved with minimal epimer-
ization by careful choice of reaction conditions and
reagents, allowing the enantio- and diastereoselective
formation of cis-dialkyl THCs with high enantiopurity.
These compounds, together with the corresponding
trans stereoisomers, provide an interesting set of ligands
and transcriptional modulators for the estrogen receptor
subtypes ERR and ERâ that includes compoundss
agonists on ERR but complete antagonists on ERâsthat
could prove to be useful as pharmacological probes for
determining the biological effects mediated by the
individual ER subtypes.
Str u ctu r e-Bin d in g Affin ity Rela tion sh ip s. The
size of the substituent at C-5 and C-11 in the THCs has
a major effect on ER binding affinity, but the optimum
size also depends on substituent stereochemistry and
ER subtype. On ERR, the highest affinities are obtained
with trans-disposed Me or Et groups; although the
binding affinities of the corresponding cis-THC diaster-
eomers were uniformly lower, Me and Et also were
preferred in the (R,R)-series and Me in the (S,S)-series.
Substituent-affinity trends on ERâ are similar: Again,
the trans isomers bind better than the cis isomers in
all cases, with an Et substituent preferred in the trans
and (R,R)-series and a Me in the (S,S)-series. The lower
affinity of the Pr-substituted THCs relative to the
smaller congeners in all cases suggests that substituent
tolerance may be exceeded by groups of this size.
Although the trans-THC isomers have higher affinity
for both ER subtypes than do their corresponding cis
isomers, they show no or only modest ERâ subtype
affinity selectivity (1-3-fold). By contrast, the cis ste-
reoisomers have a more distinct binding preference for
ERâ (3-16-fold), with the (S,S)-enantiomers generally
having greater preference. Interestingly, the (R,R)-
enantiomers usually have higher affinities for both ERR
and ERâ.
Unsubstituted THC diol 10 does not show pronounced
agonist/antagonist selectivity, exhibiting mixed agonist/
antagonist character for both ERR and ERâ (Figure
1A,E). The trans-dimethyl THC 3a is a full or nearly
full agonist on both ERR and ERâ (Figure 1B,F). The
cis-dimethyl THC enantiomers (R,R)-2a and (S,S)-2a
exhibit somewhat different agonist/antagonist character
(Figure 1B,F). (R,R)-2a is a full agonist on ERR, but it
is a very weak agonist on ERâ, with considerable
antagonistic character. (S,S)-2a shows a curious and
reproducible biphasic dose-response curve on both ER
subtypes: It is a full agonist at concentrations up to 10-7
M but shows antagonistic character at 10-6 M on both
ER subtypes. All of the cis-diethyl and cis-dipropyl THC
diols 2b,c are agonists on ERR in a manner similar to
racemic diethyl THC (rac)-2b (Figure 1C,D).11 However,
on ERâ, the diethyl (R,R)-2b and both cis-dipropyl THC
2c enantiomers exhibit no, or very low, transcriptional
activity, and they appear to very effectively antagonize
the effect of estradiol (Figure 1G,H). In contrast, diethyl
(S,S)-2b is an agonist on ERâ and begins to antagonize
the effect of estradiol weakly only at very high concen-
trations (Figure 1G). trans-Diethyl THC diol 3b is an
agonist with some weak antagonist character on ERâ
(Figure 1G). In contrast, trans-dipropyl THC diol 3c,
like the corresponding cis-dipropyl enantiomers, exhibits
only minimal transcriptional activity on ERâ and very
effectively antagonizes the effect of estradiol on ERâ
(Figure 1H). Although trans-dipropyl THC 3c is more
potent as an ERâ antagonist than cis-diethyl (R,R)-2b,
it has some minimal activity in the absence of E2. The
ERâ antagonist character of the dipropyl (S,S)-2c and
Others have reported on compounds that show dif-
ferences in their binding affinity for ERR and ERâ.5 ERR
affinity selectivities up to 5-fold have been found for
certain substituted steroidal estrogens, especially those
with 17R-substituents. The highest ERâ affinity selec-
tivities of about 7-fold have been reported for certain
nonsteroidal phytoestrogens, such as genistein. A num-
ber of the cis-dialkyl THCs that we have prepared show
equally high or greater ERâ affinity selectivities.
In our initial report on the biological activity of the
cis-diethyl THCs,11 we reported relative binding affini-
ties for the purified LBDs of ERR and ERâ. Subse-
quently, we have found that these ER LBD RBA values,
while exhibiting similar trends, are quantitatively dif-

F igu r e 1. Transcription activation by ERR (upper panels, A-D) and ERâ (lower panels, E-H) in response to THCs 2a -c, 3a -c, and 10. Human endometrial cancer (HEC-1) cells
were transfected with expression vectors for ERR or ERâ and the estrogen-responsive (ERE)3-pS2-CAT reporter gene and were treated with the indicated concentrations of estradiol
(E2) or ligand (2a -c, 3a -c, or 10) for 24 h (solid lines and symbols). The antagonist activity of 2a -c, 3a -c, and 10 on ERR and ERâ was assayed in the presence of 1 nM E2 (dashed
lines and open symbols). CAT activity was normalized for â-galactosidase activity from an internal control plasmid. Values are the mean ( SD for three or more separate experiments
and are expressed as a percent of the ERR or ERâ response with 10-8 M E2. For some values, error bars are too small to be visible.

2462 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 13
Meyers et al.
F igu r e 2. Maximum efficacy of THCs as a function of
substituent size and stereochemistry. Maximum efficacy is the
level of transcription activation by ERâ in response to THCs
2a -c, 3a -c, and 10 at a ligand concentration of 10-7 M and
is expressed as a percent of the transcriptional stimulation
with 10-8 M E2. Transfection assays were conducted in HEC-1
cells using the (ERE)3-pS2-CAT reporter gene construct as
described in Figure 1.
F igu r e 3. Ligands for the estrogen receptor: (A) numbering
scheme for 7R,11â-disubstituted E2 and 5,11-disubstituted
THCs and (B) antiestrogens hydroxytamoxifen and raloxifene.
but has lower affinity). However, overall, the major
difference between affinity and potency in this THC
series is that relative to estradiol; all of the THCs are
somewhat less potent as transcriptional modulators
than would be predicted from their affinities. The fact
that there is not always a direct correlation between
ligand binding affinity and transcriptional potency
suggests that factors beyond ligand-receptor interac-
tion, such as receptor-coactivator interactions or dif-
ferential utilization of AF-1 and AF-2, are likely to be
important determinants of transcriptional potency.2
Liga n d Bin d in g Affin ity a n d Estr ogen Recep tor
Str u ctu r e. The THCs that we have examined bear
substituents at C-5 and C-11, positions that correspond
to the C-7 and C-11 positions in estradiol (Figure 3A).
From an earlier analysis of the tolerance of ERR for
substituents on the steroidal ligand estradiol,20 we had
ascertained that there was considerable “preformed
pocket volume” in ERR that would accommodate sizable
substituents at the 7R- and 11â-positions; these posi-
tions are, respectively, below and above the plane of the
steroid skeleton. Recently reported X-ray crystal struc-
tures of ERR with the agonists estradiol21,22 and dieth-
ylstilbestrol (DES)23 have delineated these pockets in
greater detail and have shown that they have a com-
bined volume of nearly 200 Å3. In fact, in the ERR-
DES structure, the two ethyl groups of the ligand fill
the 7R- and 11â-pockets in preference to the peripheral
regions of the receptor that are normally occupied by
the B- and C-rings of steroidal ligands.23
ferent from the corresponding values for full-length
ERR/â RBAs, which we are reporting here. The differ-
ence is particularly marked in the (S,S)-2b affinity for
full-length ERâ, which is about 10-fold higher than its
affinity for the corresponding purified ERâ LBD.11 This
difference is reproducible, but at this point an explana-
tion is not apparent.
St r u ct u r e-Tr a n scr ip t ion a l E ffica cy a n d P o-
t en cy R ela t ion sh ip s. In cell-based reporter gene
transcription assays, all of the THCs are agonists
through ERR; those with small substituents are also
agonists through ERâ, but as substituent size grows,
they become antagonists. This relationship between the
efficacy (i.e., agonist and antagonist character of these
THC ligands) and the size and geometry of their alkyl
substituents is illustrated in a direct fashion in Figure
2. THCs with smaller substituents (H, Me) in all cases
but one are agonists on both receptors. As substituent
size is increased, ERâ-selective antagonism develops
first in the (R,R)-THC series, being clearly evident at
the Me substituent size and complete at the Et and Pr
sizes. Eventually, the ERâ-selective antagonism devel-
ops as well in the trans-THC and (S,S)-THC series,
being complete only at the Pr substituent size (Figure
2).
As has been noted by others, the binding affinity of a
ligand is not always perfectly reflected in its potency in
a receptor transcriptional activation assay.10,11 As far
as we have examined within the THC series, however,
there appears to be reasonable concordance between a
compound’s relative binding affinity and its relative
transcriptional potency. Thus, in general, those THCs
with the highest affinities (the trans isomers and the
(R,R)-enantiomers) were also the most potent as ago-
nists or antagonists, although this is not uniformly the
case (see, for example, (S,S)-dimethyl THC, which is
more potent as an agonist on ERR than the (R,R)-isomer
but has lower affinity, and trans-dipropyl THC, which
is a more potent ERâ antagonist than (R,R)-diethyl THC
The THC ligands that we have studied have a
tetracyclic structure that is closely related to steroidal
ligands such as estradiol, so one may reasonably pre-
sume that they fit into the binding site in ER in a
fashion similar to that of estradiol. In this orientation,
the two substituents in the trans-THCs, which project
below and above the plane of the ligand core, can both
be nicely accommodated in the 7R and 11â preformed
pockets in ERR (Figure 4), consistent with the high
binding affinity of the trans isomers. The disposition of
the substituents in the cis isomers, however, is less
clear, because both substituents project out from the

Estrogen Receptor Subtype-Selective Ligands
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 13 2463
F igu r e 4. Stereorenditions of (A) 7R,11â-diethyl E2,16 minimized with SYBYL 6.5 using the Tripos force field; (B) trans-5,11-
diethyl THC 3b; (C) (R,R)-5,11-diethyl THC 2b; (D) (S,S)-5,11-diethyl THC 2b. Structures in B-D were taken from X-ray crystal
structures of the trans- and racemic cis-THCs reported previously.13
same side of the ligand core structure. If one assumes
that the tetracyclic core of the cis-THC ligands is also
bound in the same manner as that of estradiol and the
trans-THCs, then the two substituents in the cis isomers
would both project above the plane in the cis (S,S)-series
or below the plane in the cis (R,R)-series (Figure 4).24
In the cis isomers, steric interactions between the two
substituents and the ring system also result in a flexing
of the central two rings, giving the tetracyclic core an
overall twist (Figure 4C,D).
Clearly, the binding affinity measurements indicate
that in this geometry, the cis-THC ligands are less well-
accommodated by both ERR and ERâ than are the trans
isomers. Interestingly, however, with all cis-THCs on
both ER subtypes (except the cis-dimethyl THCs on
ERâ), the (R,R)-enantiomer, with the two substituents
projecting downward, binds better than the (S,S)-
enantiomer. Again, all of the THCs we have studied,
regardless of substituent size and stereochemistry, bind
better to ERâ than to ERR. Thus, the two ER subtypes
have a significantly different tolerance to the size and
stereochemistry of substituents on the THC ring system
and to conformational distortions of the THC core ring
structure. What is most intriguing, however, is the very
different levels of transcriptional efficacy that some of
the THC ligands have on the two ER subtypes.
Str u ctu r a l Ba sis for th e ERâ-Selective An ta go-
n ist Activity of th e Tetr a h yd r och r ysen es. The THC
compounds that we have studied differ in activity from
well-known antiestrogenic compounds hydroxytamox-
ifen and raloxifene, which are full antagonists on ERâ
but usually have only minimal agonist activity on ERR
and substantially antagonize E2 agonism when assayed
in systems similar to those used here.5,10,25-27 For
example, in HEC-1 cells, hydroxytamoxifen has 25% the
agonist activity of estradiol through ERR but is a pure
antagonist through ERâ.26,27 In contrast, cis-diethyl
(R,R)-2b, both enantiomers of cis-dipropyl 2c, and trans-
dipropyl 3c are nearly full agonists on ERR.
These THC ligands are also quite different structur-
ally from hydroxytamoxifen and raloxifene, which have
bulky substituents that project outward from the ligand
core and contain basic tertiary amine side chains
(Figure 3B). From the recent ERR X-ray crystal struc-
21,23
tures,
these substituents were seen to project
upward from the ligand core toward helix 12 and result
in a change in the tertiary structure of the ERR LBD:
By steric interactions, the bulky substituent displaces
helix 12 from its normal docking site in the ERR-
agonist structures and repositions it into a groove in
which the amphipathic helix motif LXXLL, found in co-
activator proteins, is thought to interact. In this con-
formation, the basic amine function on the large sub-
stituent is stabilized by salt bridge formation with a
specific aspartic acid residue (D351).
The antagonist activity of raloxifene and hydroxy-
tamoxifen, then, is presumed to arise from the blocking
of co-activator LXXLL helix interaction by the reposi-
tioned helix 12. Because they are also antagonists on
ERâ, one would presume that hydroxytamoxifen and
raloxifene would induce an “antagonist conformation”
in ERâ by a similar repositioning of helix 12 into the
co-activator helix binding groove of this ER subtype.
But how is it that some of the THCs, (R,R)-diethyl
THC and all three dipropyl THCs, can be full or nearly
full agonists on ERR yet full antagonists on ERâ? None
of these ERâ-selective antagonists have substituents of
the size and character found in tamoxifen and ralox-
ifene. Also, with increase in substituent size, ERâ-
selective antagonism developed first in the (R,R)-cis-
THC series, in which the substituents project downward
from the ligand core, yet with two propyl substituents,
antagonism was induced in ERâ with all three of the
THCs. Thus, it would appear that an antagonist con-

2464 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 13
Meyers et al.
distilled over CaH2. LiCl was dried in vacuo at 130 °C
overnight and stored in a desiccator. Diazomethane was
prepared from N-methyl-N-nitrosourea as previously de-
scribed.28 All reactions were carried out under nitrogen or
argon, using oven- or flame-dried glassware, unless stated
otherwise. Reaction progress was monitored by analytical thin-
layer chromatography using 0.25-mm HLF silica plates with
UV254 indicator (Analtech), and visualization was achieved
by UV light (254 nm) or phosphomolybdic acid indicator.
Hexane was distilled prior to use in chromatography. Flash
chromatography was performed using Woelm 32-63-µm silica
gel packing. Radial preparative-layer chromatography was
performed on a Chromatotron instrument (Harrison Research,
Inc., Palo Alto, CA) using EM Science silica gel Kieselgel 60
PF254 as adsorbent.
Melting points were determined on a Thomas-Hoover Un-
imelt capillary apparatus and are uncorrected. 1H and 13C
NMR spectra were obtained with Varian Unity 400- and 500-
MHz spectrometers. Chemical shifts (δ) are reported in parts
per million downfield from tetramethylsilane and referenced
from solvent references. NMR coupling constants are reported
in hertz (Hz). 13C NMR spectra were determined using
attached proton test (APT) experiment. Low- and high-
resolution electron impact (EI) mass spectra were obtained on
a Micromass 70-VSE spectrometer. Low- and high-resolution
fast atom bombardment (FAB) spectra were obtained on
Micromass ZAB-SE and 70-SE-4F spectrometers, respectively.
Elemental analyses were performed by the Microanalytic
Service Laboratory of the University of Illinois.
Cis:trans isomer ratios were determined by integration of
1H NMR peaks with comparison to known compounds.13
Enantiomeric ratios were determined with one of three HPLC
columns from Regis Technologies, Inc. and Chiral Technolo-
gies: (A) S,S Whelk-O 1 (4.6 mm × 25 cm), (B) R,R Whelk-O
1 (4.6 mm × 25 cm), or (C) ChiralPak AS (4.6 mm × 25 cm).
Gen er a l Meth od for P seu d oep h ed r in e Acyla tion . Ac-
cording to the method of Myers,14 the appropriate anhydride
(1.2 equiv) was added slowly to a stirred 0.5 M solution of (+)-
or (-)-pseudoephedrine (1 equiv) and triethylamine (1.2 equiv)
in CH2Cl2 at room temperature. After having been stirred for
1 h, the reaction mixture was quenched with water. The
organic layer was washed with half-saturated NaHCO3 (aque-
ous), 1 M HCl, and brine. The extract was dried over Na2SO4
and concentrated to provide amides 5a -c.
(1S ,2S )-N -(2-H y d r o x y -1-m e t h y l-2-p h e n y le t h y l)-N -
m eth ylp r op ion a m id e ((S,S)-5a ). Amide (S,S)-5a was ob-
tained as a white solid (5.38 g, 96%): mp 112-113.5 °C (lit.14
mp 114-115 °C); spectroscopic data matched those previously
reported.14 Anal. (C13H19NO2) C, H, N.
(1R ,2R )-N -(2-H y d r o xy -1-m e t h yl-2-p h e n y le t h y l)-N -
m eth ylp r op ion a m id e ((R,R)-5a ). Recrystallization from hot
toluene provided amide (R,R)-5a as a white solid (3.67 g,
92%): mp 113-114 °C; spectroscopic data identical to those
for (S,S)-5a .14 Anal. (C13H19NO2) C, H, N.
(1S ,2S )-N -(2-H y d r o x y -1-m e t h y l-2-p h e n y le t h y l)-N -
m eth ylbu tyr a m id e ((S,S)-5b). Recrystallization from hot
toluene furnished amide (S,S)-5b as a white crystalline solid
(7.70 g, 93%): mp 82.5-84 °C; 1H NMR (3:1 rotamer ratio,
asterisk denotes minor rotamer peaks, 500 MHz, C6D6) δ 7.32
(d, 2H, J ) 7.2), 7.05-7.20 (m, 3H), 4.94 (br s, 1H), 4.53 (t,
1H, J ) 7.2), 4.22 (br, 1H), 4.19* (dd, 1H, J ) 8.8, 2.7), 3.78*
(quint, 1H, J ) 6.7), 2.83* (s, 3H), 2.44* (sextet, 1H, J ) 7.8),
2.18* (m, 1H), 2.12 (s, 3H), 1.72-1.85 (m, 2H), 1.58 (sextet,
2H, J ) 7.4), 0.97 (d, 3H, J ) 7.1), 0.94* (t, 3H, J ) 7.4), 0.81
(t, 3H, J ) 7.4), 0.60* (d, 3H, J ) 6.8); 13C NMR (asterisk
denotes minor rotamer peaks, 125 MHz, C6D6) δ 174.8, 174.0*,
144.2, 143.3*, 128.9, 128.7, 127.8, 127.8, 127.3, 76.8, 75.8*,
58.9, 36.5, 36.0*, 27.1, 19.4*, 19.0, 15.7*, 14.7, 14.7*, 14.4; MS
(FAB) m/z 236 (MH+). Anal. (C14H21NO2) C, H, N.
formation can be induced in ERâ by a change in tertiary
structure that may be different from the displacement
of helix 12 by a large, basic substituent on a ligand, as
has been observed in ERR with hydroxytamoxifen and
raloxifene.
The groove into which the LXXLL helix from co-
activators binds is made up of portions of helices 3, 4,
and 5, as well as helix 12.23 It is possible that ligand-
induced motions of one or more helicessother than helix
12scould distort this groove so that the co-activator
helix would no longer be able to bind. Such an alterna-
tive ligand-induced conformational change blocking co-
activator helix binding would, of course, need to be
induced by the appropriate THCs more easily in ERâ
than in ERR. Clearly, additional X-ray crystallographic
structural work, comparing the structure of the ERR
and ERâ LBDs complexed with ERâ-selective antago-
nists, is needed to settle this issue definitively.
In summary, we have developed an enantioselective
asymmetric synthesis of cis-dimethyl-, -diethyl-, and
-dipropyltetrahydrochrysenes and their trans congeners
which has provided us with a set of novel ligands for
the estrogen receptor that are interesting for studying
the tolerance of both ERR and ERâ to ligand substitu-
tion and stereochemistry and as probes for the confor-
mational basis of agonist/antagonist character. Both the
cis- and trans-THC series were found to have higher
binding affinities for ERâ relative to ERR. THCs with
small substituents (at the 5- and 11-ring positions) are
agonists on both ER subtypes; however, as substituent
size is increased, ERâ-selective antagonism develops
first in the (R,R)-enantiomer of the cis-THC series and
finally in the trans and (S,S)-enantiomer THC series.
Thus, cis-(R,R)-diethyl THC, trans-dipropyl THC 3c,
and both cis-dipropyl enantiomers 2c were full or nearly
full antagonists on ERâ yet are full or nearly full
agonists on ERR. These compounds are the first to be
described that are full agonists on ERR and full antago-
nists on ERâ.
These studies illustrate that the antagonist character
in THC ligands for ERâ depends in a progressive way
on the size and geometric disposition of substituent
groups. Furthermore, antagonists that are selectively
effective on ERâ can have structures that are very
different from the typical antiestrogens tamoxifen and
raloxifene, all of which have very bulky substituents and
are complete or nearly complete antagonists on both ER
subtypes. This suggests, at least for ERâ, that an
antagonist state of the receptor can be reached by
conformational change that is different from that which
has been demonstrated so far in antagonist complexes
with ERR. The subtype-selective efficacy of some of
these THCs should be useful in evaluating the biological
role of ERâ through studies in various in vitro and in
vivo test systems and in examining the conformation
of ERR and ERâ agonist/antagonist complexes by X-ray
crystallography.
Exp er im en ta l Section
Gen er a l. The synthesis of compounds 3b,c and 10 has been
described previously.13 Reagents and solvents were purchased
from Aldrich, Fisher, and Mallinckrodt. THF was distilled
immediately prior to use from sodium/benzophenone. CH2Cl2
and toluene were distilled from CaH2. n-Butyllithium was
titrated against N-pivaloyl-o-toluidine. Et3N and TMSCl were
(1R ,2R )-N -(2-H y d r o xy -1-m e t h yl-2-p h e n y le t h y l)-N -
m eth ylbu tyr a m id e ((R,R)-5b). Recrystallization from hot
toluene furnished amide (R,R)-5b as a white crystalline solid
(7.83 g, 95%): mp 85-86.5 °C; spectroscopic data identical to
those of (S,S)-5b. Anal. (C14H21NO2) C, H, N.

Estrogen Receptor Subtype-Selective Ligands
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 13 2465
(1S ,2S )-N -(2-H y d r o x y -1-m e t h y l-2-p h e n y le t h y l)-N -
m eth ylp en ta n a m id e ((S,S)-5c). Flash chromatography (20%
acetone/benzene) provided (S,S)-5c as a clear viscous oil (3.85
g, 85%): 1H NMR (3:1 rotamer ratio, asterisk denotes minor
rotamer peaks, 500 MHz, C6D6) δ 7.33 (d, 2H, J ) 7.4), 7.05-
7.21 (m, 3H), 4.99 (br s, 1H), 4.54 (d, 1H, J ) 7.2), 4.21 (br,
1H), 4.18* (d, 1H, J ) 8.7), 3.80* (quint, 1H, J ) 7.0), 2.83*
(s, 3H), 2.49* (m, 1H), 2.26* (m, 1H), 2.14 (s, 3H), 1.85 (ABt,
2H, νA ) 1.87 ppm, νB ) 1.83 ppm, J AB ) 15.5, J ) 7.5), 1.74-
1.81* (m, 2H), 1.55 (quint, 2H, J ) 7.4), 1.29-1.39* (m, 2H),
1.20 (sextet, 2H, J ) 7.5), 0.98 (d, 3H, J ) 7.0), 0.90* (t, 3H,
J ) 7.4), 0.83 (t, 3H, J ) 7.3), 0.60* (d, 3H, J ) 6.8); 13C NMR
(asterisk denotes minor rotamer peaks, 100 MHz, C6D6) δ
175.1, 174.2*, 144.2, 143.0*, 129.0, 128.8, 127.8, 127.7, 127.2,
76.9, 75.9*, 58.9, 34.4, 33.9*, 28.2*, 27.7, 23.4*, 23.1, 14.8, 14.5;
MS (FAB) m/z 250 (MH+). Anal. (C15H23NO2) C, H, N.
minor rotamer peaks, 125 MHz, C6D6) δ 177.2, 160.2, 143.5,
142.2, 129.5, 129.5*, 128.6*, 128.4, 128.3*, 127.5, 127.3*, 126.9,
122.0*, 121.6, 115.5*, 115.2, 112.0*, 111.8, 76.6, 75.3*, 65.7,
54.6, 46.4, 39.8, 26.6, 14.4, 11.9; MS (FAB) m/z 356 (MH+);
HRMS (FAB) calcd for C22H30NO3 356.2226, found 356.2227.
(1′R,2′R,2S)-N-(2′-H yd r oxy-1′-m et h yl-2′-p h en ylet h yl)-
N-m eth yl-2-(3-m eth oxyben zyl)bu tyr a m id e ((R,R,S)-6b).
Flash chromatography (20% acetone/benzene) furnished amide
(R,R,S)-6b as a clear viscous oil (11.8 g, quantitative): spec-
troscopic data identical to those for (S,S,R)-6b; HRMS (FAB)
calcd for C22H30NO3 356.2226, found 356.2224.
(1′S,2′S,2R)-N-(2′-Hyd r oxy-1′-m eth yl-2′-p h en yleth yl)-N-
m et h yl-2-(3-m et h oxyb en zyl)p en t a n a m id e ((S,S,R)-6c).
Flash chromatography (35-50% EtOAc/hexane) furnished
amide (S,S,R)-6c as a yellow viscous oil (3.79 g, 61%): 1H NMR
(4:1 rotamer ratio, asterisk denotes minor rotamer peaks, 400
MHz, C6D6) δ 7.27 (d, 2H, J ) 7.1), 6.98-7.19 (m, 4H), 6.84
(dd, 1H, J ) 2.3, 1.7), 6.74 (d, 1H, J ) 7.4), 6.70* (ddd, 1H J
) 8.1, 2.5, 0.8), 6.64 (ddd, 1H J ) 8.1, 2.5, 0.8), 4.58 (br s,
1H), 4.44 (t, 1H, J ) 6.2), 4.31 (br, 1H), 4.15* (dd, 1H, J ) 8.7,
2.6), 4.01* (quintet, 1H, J ) 6.8), 3.41* (s, 3H), 3.32 (s, 3H),
3.09-3.16* (m, 1H), 2.99 (dd, 1H, J ) 12.9, 9.5), 2.81* (dd,
1H, J ) 13.2, 7.5), 2.77* (s, 3H), 2.66-2.75 (m, 1H), 2.57 (dd,
1H, J ) 12.9, 5.1), 2.17 (s, 3H), 1.79-1.92 (m, 1H), 1.38-1.47*
(m, 1H), 1.23-1.37 (m, 2H), 1.05-1.17 (m, 1H), 0.98-1.17*
(m, 1H), 0.82 (t, 3H, J ) 7.1), 0.78 (d, 3H, J ) 6.7), 0.76* (t,
3H, J ) 7.2), 0.66* (d, 3H, J ) 6.7); 13C NMR (asterisk denotes
minor rotamer peaks, 100 MHz, C6D6) δ 177.6, 160.5, 143.8,
143.3*, 143.0*, 142.6, 130.1*, 129.9, 129.0*, 128.8, 127.9,
127.7*, 127.4, 122.4*, 122.0, 115.9*, 115.6, 112.4*, 112.2, 76.8,
75.6*, 58.9, 55.1, 45.1, 40.4, 36.2, 21.2, 14.9, 14.7; MS (FAB)
m/z 370 (MH+); HRMS (FAB) calcd for C23H32NO3 370.2382,
found 370.2383. Anal. (C23H31NO3) C, H, N.
(1′R,2′R,2S)-N-(2′-H yd r oxy-1′-m et h yl-2′-p h en ylet h yl)-
N-m eth yl-2-(3-m eth oxyben zyl)p en ta n a m id e ((R,R,S)-6c).
Flash chromatography (35-60% EtOAc/hexane) furnished
amide (R,R,S)-6c as a clear viscous oil (5.81 g, 95%): spectro-
scopic data identical to those for (S,S,R)-6c; HRMS (FAB) calcd
for C23H32NO3 370.2382, found 370.2383. Anal. (C23H31NO3)
C, H, N.
Meth yl (2R)-3-(3-Meth oxyp h en yl)-2-m eth ylp r op ion a te
((R)-7a ). A solution of amide (S,S,R)-6a (1.16 g, 3.40 mmol)
in 1:3 concentrated H2SO4/MeOH (10 mL) was refluxed for 4
h. The reaction was neutralized with 1 M NaOH and extracted
with EtOAc. The organic layers were dried over Na2SO4 and
concentrated. Flash chromatography (20% acetone/benzene)
yielded (R)-7a as a clear oil (475 mg, 67% yield, 77% corrected
for starting material consumed): [R]58928 -19.0° (c 1.0, CHCl3);
1H NMR (400 MHz, CDCl3) δ 7.19 (t, 1H, J ) 7.8), 6.75 (dd,
2H, J ) 7.9, 2.1), 6.71 (t, 1H, J ) 2.0), 3.79 (s, 3H), 3.65 (s,
3H), 3.01 (dd, 1H, J ) 13.3, 6.7), 2.74 (sextet, 1H, J ) 7.0),
2.63 (dd, 1H, J ) 13.4, 7.8), 1.15 (d, 3H, J ) 6.8); 13C NMR
(100 MHz, CDCl3) δ 176.5, 159.5, 140.9, 129.3, 121.3, 114.6,
111.6, 55.0, 51.6, 41.3, 39.6, 16.7; MS (EI, 70 eV) m/z 208 (M+,
73), 148 (M - CO2Me, 58). Anal. (C12H16O3) C, H.
(1R ,2R )-N -(2-H y d r o x y -1-m e t h yl-2-p h e n y le t h yl)-N -
m eth ylp en ta n a m id e ((R,R)-5c). Flash chromatography
(20% acetone/benzene) provided amide (R,R)-5c as a clear
viscous oil (4.32 g, 96%): spectroscopic data identical to those
of (S,S)-5c. Anal. (C15H23NO2) C, H, N.
Gen er a l Meth od for Asym m etr ic Alk yla tion . According
to the method of Myers,14 n-butyllithium (2.08 equiv) was
added to a mixture of LiCl (g6 equiv) and diisopropylamine
(2.25 equiv) in THF (∼1.5 M) at -78 °C. The flask was warmed
to 0 °C for 15 min before recooling to -78 °C. An ice-cooled
solution of amides 5a -c (1 equiv) in THF (∼0.33 M) was added
via cannula. The resulting suspension was stirred at -78 °C
for 1 h, 0 °C for 15 min, and room temperature for 5 min. After
the mixture was recooled to 0 °C, 3-methoxybenzyl bromide
(1.5 equiv) was added and stirring was continued at 0 °C until
the reaction was complete as indicated by TLC (1-2 h). The
reaction was quenched with saturated NH4Cl (aqueous) and
EtOAc. The aqueous layer was extracted twice with EtOAc.
The combined organic layers were washed with brine, dried
over Na2SO4, and concentrated to provide amides 6a -c.
(1′S,2′S,2R)-N-(2′-Hyd r oxy-1′-m eth yl-2′-p h en yleth yl)-N-
m eth yl-2-(3-m eth oxyben zyl)p r op ion a m id e ((S,S,R)-6a ).
Flash chromatography (20% acetone/benzene) provided amide
(S,S,R)-6a as a light-yellow highly viscous oil (645 mg, 1.89
mmol, 84%): 1H NMR (3:1 rotamer ratio, asterisk denotes
minor rotamer peaks, 500 MHz, C6D6) δ 7.24 (d, 2H, J ) 7.6),
7.01-7.18 (m, 4H), 6.82 (br, 1H), 6.72 (d, 1H, J ) 7.1), 6.66
(dd, 1H J ) 8.2, 2.5), 4.47 (t, 1H, J ) 6.4), 4.41 (br s, 1H), 4.34
(br, 1H), 4.07* (dd, 1H, J ) 8.3, 2.9), 3.85* (quintet, 1H, J )
7.3), 3.40* (s, 3H), 3.33 (s, 3H), 3.09* (m, 1H), 3.04 (dd, 1H, J
) 13.1, 8.2), 2.77* (m, 1H), 2.74* (s, 3H), 2.62 (sextet, 1H, J )
6.8), 2.51 (dd, 1H, J ) 13.1, 6.1), 2.13 (s, 3H), 1.09* (d, 3H, J
) 6.8), 1.04 (d, 3H, J ) 6.8), 0.85 (d, 3H, J ) 6.3), 0.62* (d,
3H, J ) 6.6); 13C NMR (asterisk denotes minor rotamer peaks,
125 MHz, C6D6) δ 177.5, 176.9*, 160.3*, 160.2, 143.5, 142.9*,
142.8*, 142.3, 129.5*, 129.5, 128.6, 128.5, 128.3, 127.4, 127.3*,
127.0, 122.1*, 121.7, 115.4*, 115.2, 112.0*, 111.8, 76.3, 75.4*,
58.2, 54.7, 40.7, 40.5*, 38.9, 38.2*, 17.7, 17.6*, 15.4*, 14.2; MS
(FAB) m/z 342 (MH+). Anal. (C21H27NO3) C, H, N.
(1′R,2′R,2S)-N-(2′-H yd r oxy-1′-m et h yl-2′-p h en ylet h yl)-
N-m et h yl-2-(3-m et h oxyb en zyl)p r op ion a m id e ((R,R,S)-
6a ). Flash chromatography (35-50% EtOAc/hexane) furnished
amide (R,R,S)-6a as a clear viscous oil (5.37 g, 99%): spec-
Meth yl (2S)-3-(3-Meth oxyp h en yl)-2-m eth ylp r op ion a te
((S)-7a ). A solution of amide (R,R,S)-6a (5.21 g, 15.2 mmol)
in 1:1 18 N aqueous H2SO4/dioxane (52 mL) was refluxed for
2 h. After cooling, the solution was concentrated and parti-
tioned between water (200 mL) and EtOAc (200 mL). The
organic extract was washed with 3 N HCl and brine and was
dried over Na2SO4. Concentration gave the crude carboxylic
acid as a golden brown oil. The crude acid was dissolved in
diethyl ether (80 mL) and treated with excess diazomethane
at 0 °C. The reaction was quenched with a few drops of glacial
acetic acid. Concentration and bulb-to-bulb distillation (∼85-
100 °C at 0.2 Torr) gave (S)-7a as a clear liquid (2.44 g, 77%):
troscopic data identical to those for (S,S,R)-6a . Anal. (C21H27
-
NO3) C, H, N.
(1′S,2′S,2R)-N-(2′-Hyd r oxy-1′-m eth yl-2′-p h en yleth yl)-N-
m eth yl-2-(3-m eth oxyben zyl)bu tyr am ide ((S,S,R)-6b). Flash
chromatography (20% acetone/benzene) furnished amide (S,S,R)-
6b as a clear viscous oil (9.48 g, 88%): 1H NMR (5:1 rotamer
ratio, asterisk denotes minor rotamer peaks, 500 MHz, C6D6)
δ 7.24 (d, 2H, J ) 7.1), 7.01-7.13 (m, 4H), 6.82 (t, 1H, J )
1.8), 6.72 (d, 1H, J ) 7.5), 6.65 (dd, 1H J ) 8.1, 2.0), 4.47 (br
s, 1H), 4.43 (t, 1H, J ) 7.0), 4.21 (br, 1H), 4.05* (dd, 1H, J )
8.7, 3.7), 3.93* (quintet, 1H, J ) 7.0), 3.40* (s, 3H), 3.33 (s,
3H), 2.99 (dd, 1H, J ) 12.3, 8.5), 2.78* (dd, 1H, J ) 13.1, 6.7),
2.75* (s, 3H), 2.55-2.64 (m, 2H), 2.14 (s, 3H), 1.83-1.87 (m,
1H), 1.36-1.40 (m, 1H), 0.83 (t, 3H, J ) 7.4), 0.76-0.81 (m,
3H + 3H*), 0.63* (d, 3H, J ) 6.8); 13C NMR (asterisk denotes
28
[R]589 +28.8° (c 0.82, CHCl3); spectroscopic data identical to
those of ester (R)-7a . Anal. (C12H16O3) C, H.
Met h yl (2R)-2-(3-Met h oxyb en zyl)b u t yr a t e ((R)-7b ).
Methanesulfonic acid (2.5 mL, 38.4 mmol) was added to a
solution of amide (S,S,R)-6b (9.10 g, 25.6 mmol) in THF (100
mL) and refluxed for 3 h. The solution was cooled to 0 °C, and
LiBH4 (2.0 M in THF, 19.2 mL, 38.4 mmol) was added slowly.

2466 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 13
Meyers et al.
Water (55 mL) was added cautiously, followed by tetrabutyl-
ammonium hydroxide (40% w/w in H2O, 85 mL, 128 mmol).
The heterogeneous mixture was stirred for 1 h. The solution
was acidified to pH < 2 with 3 N HCl, extracted with ether,
and concentrated to give 4.5 g of the crude acid. A solution of
the crude acid in ether was treated with excess diazomethane
at 0 °C. After the mixture was quenched with glacial acetic
acid and concentrated, bulb-to-bulb distillation (105-110 °C
at 0.45 Torr) gave ester (R)-7b as a clear liquid (3.45 g, 15.5
mmol, 61%): er g 98:2 (HPLC column A, 0.25% IPA/hexanes,
recrystallizations from diethyl ether provided (R,R)-9a in a
cis:trans ratio of 64:1: mp 196.5-197.5 °C; er ) 98:2 (HPLC
24
column B, 0.5% IPA/hexanes, tR ) 13.1 min); [R]589 -109° (c
1
0.98, CHCl3); H NMR (500 MHz, CDCl3) δ 7.31 (d, 2H, J )
8.3), 6.76-6.79 (m, 4H), 3.84 (s, 6H), 3.14 (dd, 2H, J ) 15.1,
6.1), 2.92 (quint.d, 2H, J ) 6.2, 1.5), 2.59 (dd, 2H, J ) 15.2,
1.5), 1.07 (d, 6H, J ) 6.8); 13C NMR (100 MHz, CDCl3) δ 158.2,
136.0, 131.6, 127.9, 122.8, 114.7, 111.0, 55.2, 36.4, 28.9, 17.4;
MS (EI, 70 eV) m/z 320 (M+, 100), 305 (M - Me, 39). Anal.
(C22H24O2) C, H.
24
1
tR ) 20.5 min); [R]589 -30.3° (c 1.02, CHCl3); H NMR (500
MHz, CDCl3) δ 7.19 (t, 1H, J ) 7.9), 6.74 (dd, 2H, J ) 7.9,
2.1), 6.70 (s, 1H), 3.79 (s, 3H), 3.62 (s, 3H), 2.92 (dd, 1H, J )
13.7, 8.2), 2.71 (dd, 1H, J ) 13.5, 6.8), 2.60 (tdd, 1H, J ) 8.4,
6.8, 5.1), 1.63 (ddq, 1H, J ) 13.6, 8.7, 7.5), 1.58 (dqd, 1H, J )
13.7, 7.5, 5.2), 0.91 (t, 3H, J ) 7.5); 13C NMR (125 MHz, CDCl3)
δ 176.0, 159.6, 141.1, 129.3, 121.2, 114.5, 111.6, 55.1, 51.4, 49.0,
38.1, 25.1, 11.7; MS (EI, 70 eV) m/z 222 (M+, 62), 162 (M -
CO2Me, 41). Anal. (C13H18O3) C, H.
Meth yl (2S)-2-(3-Meth oxyben zyl)bu tyr a te ((S)-7b). Es-
ter (S)-7b was prepared by the same method described for (R)-
7b from amide (R,R,S)-6b (9.32 g, 26.2 mmol). Isolation of the
second distillate from bulb-to-bulb distillation (∼90-100 °C
at 0.3 Torr) furnished ester (S)-7b as a clear liquid (3.36 g,
(5S,11S)-2,8-Dim eth oxy-5,11-dim eth yl-5,6,11,12-tetr ah y-
d r och r ysen e ((S,S)-9a ). Flash chromatography (10% ether/
hexanes) provided silyl ether (S,S)-8a as a clear oil (72%)
which cyclized upon treatment with TiCl4. Recrystallization
from ether provided THC (S,S)-9a as a white solid (487 mg,
42% yield, 33:1 cis:trans). Further recrystallization from
EtOAc furnished (S,S)-9a in a 66:1 cis:trans ratio: mp 195-
196.5 °C; er g 99:1 (HPLC column B, 0.5% IPA/hexanes, tR
)
28
10.9 min); [R]589 +106° (c 1.0, CHCl3); spectroscopic data
identical to those for (R,R)-9a . Anal. (C22H24O2) C, H.
(5R,11R)-2,8-Dim eth oxy-5,11-d ieth yl-5,6,11,12-tetr a h y-
d r och r ysen e ((R,R)-9b). Flash chromatography (10% ether/
hexanes) provided silyl ether (R,R)-8b as a clear oil (72%)
which cyclized upon treatment with TiCl4. Flash chromatog-
raphy (10% ether/hexanes) provided THC (R,R)-9b as a white
solid (1.16 g, 71% yield, 15:1 cis:trans). Repeated recrystalli-
zations from ether/hexanes furnished (R,R)-9b (100% cis
isomer): mp 132.5-134 °C; er g 99:1 (HPLC column B, 0.5%
58%): er ) 98:2 (HPLC column A, 0.25% IPA/hexanes, tR
)
24
18.9 min); [R]589 +32.6° (c 1.0, CHCl3); spectroscopic data
identical to those of ester (R)-7b. Anal. (C13H18O3) C, H.
Meth yl (2R)-2-(3-Meth oxyben zyl)p en ta n oa te ((R)-7c).
Ester (R)-7c was prepared by the same method as ester (S)-
7a from amide (S,S,R)-6c (3.66 g, 9.91 mmol). Bulb-to-bulb
distillation provided (R)-7c as a clear liquid (1.48 g, 63%):
24
IPA/hexanes, tR ) 11.8 min); [R]589 -209° (c 0.97, CHCl3);
1H NMR (500 MHz, CDCl3) δ 7.23 (d, 2H, J ) 8.0), 6.74-6.77
(m, 4H), 3.84 (s, 6H), 2.94-3.06 (br m, 2H), 2.82 (dd, 2H, J )
15.6, 1.4), 2.59 (br m, 2H), 1.49 (dqd, 2H, J ) 15.0, 7.4, 3.4),
1.33 (ddq, 2H, J ) 14.7, 10.6, 7.3), 0.97 (t, 6H, J ) 7.4); 13C
NMR (125 MHz, CDCl3) δ 158.2, 136.2, 131.7, 128.3, 122.9,
114.7, 110.9, 55.2, 35.8, 32.1, 23.2, 12.0; MS (EI, 70 eV) m/z
348 (M+, 100), 319 (M - Et, 64); HRMS (EI, 70 eV) calcd for
26
1
[R]589 -24.4° (c 1.19, CHCl3); H NMR (400 MHz, CDCl3) δ
7.19 (t, 1H, J ) 7.8), 6.74 (dd, 2H, J ) 7.9, 2.1), 6.70 (t, 1H, J
) 1.9), 3.79 (s, 3H), 3.61 (s, 3H), 2.87-2.96 (m, 1H), 2.62-
2.74 (m, 2H), 1.57-1.69 (m, 1H), 1.42-1.52 (m, 1H), 1.21-
1.40 (m, 2H), 0.89 (t, 3H, J ) 7.3); 13C NMR (100 MHz, CDCl3)
δ 176.2, 159.6, 141.1, 129.3, 121.1, 114.5, 111.5, 55.1, 51.5, 47.4,
38.5, 34.2, 20.6, 13.9; MS (EI, 70 eV) m/z 236 (M+, 67), 176 (M
- CO2Me, 41). Anal. (C14H20O3) C, H.
C
24H28O2 348.2089, found 348.2084. Anal. (C24H28O2) C, H.
(5S,11S)-2,8-Dim eth oxy-5,11-d ieth yl-5,6,11,12-tetr a h y-
d r och r ysen e ((S,S)-9b). Flash chromatography (10% ether/
hexanes) provided silyl ether (S,S)-8b as a clear oil (67%)
which cyclized upon treatment with TiCl4. Flash chromatog-
raphy (10% ether/hexanes) provided THC (S,S)-9b as a white
solid (966 mg, 65% yield, 17.6:1 cis:trans). Repeated recrys-
tallizations from ether/hexanes furnished (S,S)-9b (100% cis
isomer): mp 131.5-133 °C; er g 99:1 (HPLC column B, 0.5%
Meth yl (2S)-2-(3-Meth oxyben zyl)p en ta n oa te ((S)-7c).
Ester (S)-7c was prepared by the same method as ester (S)-
7a from amide (R,R,S)-6c (4.85 g, 13.1 mmol). Bulb-to-bulb
26
distillation gave (S)-7c as a clear liquid (1.65 g, 53%): [R]589
+19.9° (c 1.08, CHCl3); spectroscopic data identical to that of
ester (R)-7c. Anal. (C14H20O3) C, H.
24
IPA/hexanes, tR ) 9.5 min); [R]589 +192° (c 0.99, CHCl3);
Gen er a l Meth od for Acyloin Con d en sa tion a n d Lew is
Acid Cycliza tion . In a three-necked round-bottom flask,
sodium metal (3 equiv) was melted in refluxing toluene (∼0.5
M) under an Ar atmosphere. A solution of the appropriate
esters 7a -c (1 equiv) and TMSCl (6 equiv) in toluene (∼1.5
M) was added dropwise via an addition funnel, and reflux was
continued for 18-24 h. After the heterogeneous purple mixture
was cooled to room temperature, the mixture was filtered to
remove sodium metal and salts. The precipitate was washed
with anhydrous ether. The combined organic solutions were
washed with 1 M HCl, washed with brine (2 times), and dried
over MgSO4. Concentration and flash chromatography (10%
ether/hexanes) provided silyl ethers 8a -c as clear oils (67-
88% yield). Solutions of silyl ethers 8a -c (1 equiv) in CH2Cl2
(0.15 M) were cooled to -78 °C and treated with TiCl4 (1.0 M
in CH2Cl2, 2 equiv). The reaction was protected from light with
aluminum foil, allowed to warm to room temperature, and
stirred for 3-4 h. The reaction was quenched with half-
saturated NaHCO3 (aqueous) and partitioned with ether. The
organic layer was washed with water and brine and dried over
MgSO4. Concentration provided crude tetrahydrochrysenes
9a -c.
spectroscopic data was identical to those for (R,R)-9b; HRMS
(EI, 70 eV) calcd for C24H28O2 348.2089, found 348.2086. Anal.
(C24H28O2) C, H.
(5R,11R)-2,8-Dim eth oxy-5,11-dipr opyl-5,6,11,12-tetr ah y-
d r och r ysen e ((R,R)-9c). Flash chromatography (5% ether/
hexanes) provided silyl ether (R,R)-8c as a clear oil (71%)
which cyclized upon treatment with TiCl4. Flash chromatog-
raphy (5% ether/hexanes) provided THC (R,R)-9c as an oil (495
mg, 69% yield, ∼20:1 cis:trans). Flash chromatography (20%
CH2Cl2/hexanes) and crystallization attempts were unsuccess-
ful in improving the cis:trans ratio: er g 99:1 (HPLC column
28
B, 0.5% IPA/hexanes, tR ) 11.9 min); [R]589 -118° (c 0.5,
1
CHCl3); H NMR (500 MHz, CDCl3) δ 7.27 (d, 2H, J ) 8.4),
6.78 (dd, 2H, J ) 8.3, 2.6), 6.76 (d, 2H, J ) 2.4), 3.85 (s, 6H),
3.01 (dd, 2H, J ) 15.6, 6.0), 2.81 (dd, 2H, J ) 15.6, 1.4), 2.70-
2.75 (br m, 2H), 1.47-1.58 (m, 2H), 1.33-1.42 (m, 6H), 0.93
(t, 6H, J ) 6.8); 13C NMR (125 MHz, CDCl3) δ 158.2, 136.2,
131.7, 128.3, 122.8, 114.8, 110.8, 55.2, 33.8, 32.4, 32.3, 20.6,
14.2; MS (EI, 70 eV) m/z 376 (M+, 100), 333 (M - Pr, 91);
HRMS (EI, 70 eV) calcd for C26H32O2 376.2402, found 376.2397.
(5S,11S)-2,8-Dim eth oxy-5,11-dipr opyl-5,6,11,12-tetr ah y-
d r och r ysen e ((S,S)-9c). Flash chromatography (5% ether/
hexanes) provided silyl ether (S,S)-8c as a clear oil (71%) which
cyclized upon treatment with TiCl4. Flash chromatography (5%
ether/hexanes) provided THC (S,S)-9c as an oil (585 mg, 74%
yield, ∼20:1 cis:trans). Flash chromatography (20% CH2Cl2/
hexanes) and crystallization attempts were unsuccessful in
improving the cis:trans ratio: er g 99:1 (HPLC column B, 0.5%
(5R,11R)-2,8-Dim eth oxy-5,11-dim eth yl-5,6,11,12-tetr ah y-
d r och r ysen e ((R,R)-9a ). Flash chromatography (10% ether/
hexanes) provided silyl ether (R,R)-8a as a clear oil (88%)
which cyclized upon treatment with TiCl4 to provide the crude
THC (R,R)-9a as a white solid in a ratio of 12.5:1 of the cis:
trans isomers. Recrystallization from ether furnished (R,R)-
9a as a white solid (175 mg, 50%, 33:1 cis:trans). Repeated

Estrogen Receptor Subtype-Selective Ligands
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 13 2467
28
IPA/hexanes, tR ) 10.1 min); [R]589 +100° (c 0.5, CHCl3);
2b; HRMS (EI, 70 eV) calcd for C22H24O2 320.1776, found
320.1772. Anal. (C22H24O2‚0.2H2O) C, H.
spectroscopic data identical to those for (R,R)-9c; HRMS (EI,
70 eV) calcd for C26H32O2 376.2402, found 376.2393.
(5R,11R)-5,11-Dip r op yl-5,6,11,12-tetr a h yd r och r ysen e-
2,8-d iol ((R,R)-2c). Flash chromatography (25% EtOAc/hex-
anes) provided diol (R,R)-2c as a pink foam (202 mg, 93% yield,
33:1 cis:trans). Recrystallization from CHCl3 provided (R,R)-
2c as off-white crystals (73:1 cis:trans): mp 182-184.5 °C; er
g 99:1 (HPLC column C, 15% IPA/hexanes, tR ) 9.3 min); 1H
NMR (500 MHz, acetone-d6) δ 8.21 (s, 2H), 7.17 (d, 2H, J )
7.9), 6.66-6.69 (m, 4H), 2.88 (dd, 2H, J ) 15.5, 5.7), 2.78 (dd,
2H, J ) 15.5, 1.7), 2.65-2.70 (br m, 2H), 1.45-1.56 (m, 2H),
1.26-1.42 (m, 6H), 0.88 (t, 6H, J ) 7.1); 13C NMR (100 MHz,
CDCl3) δ 154.0, 136.5, 131.6, 128.3, 123.0, 115.8, 112.7, 33.8,
32.3, 32.3, 20.6, 14.2; MS (EI, 70 eV) m/z 348 (M+, 100), 305
(M - Pr, 100); HRMS (EI, 70 eV) calcd for C24H28O2 348.2089,
found 348.2099. Anal. (C24H28O2‚0.3H2O) C, H.
(5S,11S)-5,11-Dip r op yl-5,6,11,12-tetr a h yd r och r ysen e-
2,8-d iol ((S,S)-2c). Flash chromatography (25% EtOAc/hex-
anes) provided diol (S,S)-2c as a yellow foam (192 mg, 79%
yield, 23:1 cis:trans). Recrystallization from CHCl3 provided
(S,S)-2c as off-white crystals (67:1 cis:trans): mp 182-184.5
°C; er g 99:1 (HPLC column C, 15% IPA/hexanes, tR ) 6.5
min); spectroscopic data identical to those for (R,R)-2c; HRMS
(EI, 70 eV) calcd for C24H28O2 348.2089, found 348.2099. Anal.
(C24H28O2‚0.7H2O) C, H.
5,11-tr a n s-Dim eth yl-5,6,11,12-tetr a h yd r och r ysen e-2,8-
d iol (3a ). A 1:2 mixture of trans-9a and (R,R)-9a (25 mg, 0.078
mmol) was deprotected with BBr3 as described above to yield
3a as a 1:2 trans:cis mixture (25 mg, quantitative). Recrys-
tallization twice from MeOH furnished exclusively the trans
isomer 3a as a white powder (2.7 mg): 1H NMR (500 MHz,
acetone-d6) δ 8.23 (s, 2H), 7.30 (d, 2H, J ) 8.2), 6.68-6.71 (m,
4H), 3.20 (quint d, 2H, J ) 7.0, 1.4), 2.97 (dd, 2H, J ) 15.4,
6.8), 2.64 (dd, 2H, J ) 15.4, 1.4), 0.84 (d, 6H, J ) 7.1); 13C
NMR (125 MHz, acetone-d6) δ 157.1, 137.1, 132.1, 126.7, 124.7,
116.9, 113.6, 36.7, 26.8, 18.2; MS (EI, 70 eV) m/z 292 (M+, 44),
277 (M - Me, 23); HRMS (EI, 70 eV) calcd for C20H20O2
292.1463, found 292.1470. Anal. (C20H20O2‚0.85H2O) C; H:
calcd, 7.11; found, 6.63.
2,8-Dim et h oxy-5,11-tr a n s-d im et h yl-5,6,11,12-t et r a h y-
d r och r ysen e (tr a n s-9a ). A solution of (R,R)-8a (200 mg, 0.4
mmol) in PPA (1.8 g) was stirred with a mechanical stirrer
for 75 min. Water and EtOAc were added, and the mixture
was stirred until all of the material had dissolved. The organic
layer was washed with saturated NaHCO3 and brine and dried
over Na2SO4. Concentration provided crude trans-9a :(R,R)-9a
as a 1:2 mixture. Recrystallization from EtOAc furnished the
same ratio of isomers as white crystals (77 mg, 60%). Further
recrystallizations did not remove (R,R)-9a from the product
mixture. trans-9a : 1H NMR (500 MHz, CDCl3) δ 7.36 (d, 2H,
J ) 9.4), 6.75-6.79 (m, 4H), 3.84 (s, 6H), 3.20 (quint d, 2H, J
) 6.9, 1.2), 3.10 (dd, 2H, J ) 15.3, 6.8), 2.68 (dd, 2H, J ) 15.4,
1.2), 0.91 (d, 6H, J ) 6.9); 13C NMR (125 MHz, CDCl3) δ 158.3,
136.6, 131.7, 127.0, 123.7, 114.9, 110.9, 55.2, 36.3, 26.1, 17.9;
MS (EI, 70 eV) m/z 320 (M+, 100), 305 (M - Me, 46); HRMS
(EI, 70 eV) calcd for C22H24O2 320.1776, found 320.1777.
Gen er a l Meth od for Meth yl Eth er Dep r otection . BBr3
(1.0 M in CH2Cl2, 3 equiv) was added to a solution of methyl
ethers 9a -c (1 equiv) in CH2Cl2 (∼0.1 M) at -78 °C. The flask
was wrapped in foil and allowed to warm to room temperature
for 6-12 h. The reaction was quenched with water and
partitioned between water and EtOAc, which was acidified
with 3 N HCl. The organic layer was washed with brine and
dried over Na2SO4. Concentration provided diols 2a -c.
(5R,11R)-5,11-Dim eth yl-5,6,11,12-tetr a h yd r och r ysen e-
2,8-d iol ((R,R)-2a ). Recrystallization from EtOAc/hexanes
provided diol (R,R)-2a as yellow-tinted crystals (42 mg,
quantitative yield, 54:1 cis:trans). Further recrystallization
from CHCl3/acetone/hexanes furnished (R,R)-2a as a white
solid (100% cis isomer): mp 200.5-202.5 °C; er ) 98:2 (HPLC
column C, 15% IPA/hexanes, tR ) 13.2 min); 1H NMR (400
MHz, CDCl3) δ 8.18 (s, 2H), 7.21 (d, 2H, J ) 9.1), 6.66-6.71
(m, 4H), 3.01 (dd, 2H, J ) 15.2, 6.2), 2.86 (quint.d, 2H, J )
6.4, 1.6), 2.55 (dd, 2H, J ) 15.2, 1.6), 1.00 (d, 6H, J ) 6.9); 13
C
NMR (100 MHz, acetone-d6) δ 156.9, 136.5, 131.8, 127.5, 123.7,
116.6, 113.6, 36.8, 29.6, 17.7; MS (EI, 70 eV) m/z 292 (M+, 48),
277 (M - Me, 24); HRMS (EI, 70 eV) calcd for C20H20O2
292.1463, found 292.1464. Anal. (C20H20O2‚0.8H2O) C, H.
Biologica l P r oced u r es. 1. R ela t ive Bin d in g Affin it y
Assa y. Relative binding affinities were determined by com-
petitive radiometric binding assays using 10 nM [3H]E2 as
tracer as previously described,18,19 using either lamb uterine
cytosol diluted to approximately 1.5 nM of receptor or purified
full-length human ERR and ERâ purchased from Pan Vera.
Free ligand was removed by adsorption to dextran-coated
charcoal for cytosol RBAs.18 Hydroxyapatite was used to absorb
the purified receptor-ligand complexes.19 Incubations were
done at 0 °C for 18-24 h.
2. Tr a n scr ip tion a l Activa tion Assa y. Human endome-
trial cancer (HEC-1) cells were maintained in culture and
transfected as described previously.25,26 Transfection of HEC-1
cells in 60-mm dishes used 0.4 mL of a calcium phosphate
precipitate containing 0.5 µg of pCMVâGal as internal control,
2 µg of the reporter gene plasmid, 100 ng of ER expression
vector, and carrier DNA to a total of 5 µg of DNA. CAT activity,
normalized for the internal control â-galactosidase activity,
was assayed as previously described.25,26
(5S,11S)-5,11-Dim eth yl-5,6,11,12-tetr a h yd r och r ysen e-
2,8-d iol ((S,S)-2a ). Recrystallization from EtOAc/hexanes
furnished diol (S,S)-2a as yellow crystals (81 mg, 88% yield,
82.6:1 cis:trans): mp 199.5-202 °C; er g 99:1 (HPLC column
C, 15% IPA/hexanes, tR ) 10.7 min); spectroscopic data
identical to those for diol (R,R)-2a ; MS (EI, 70 eV) m/z 292
(M+, 100), 277 (M - Me, 50); HRMS (EI, 70 eV) calcd for
20H20O2 292.1463, found 292.1465. Anal. (C20H20O2‚0.7H2O)
C, H.
(5R,11R)-5,11-Diet h yl-5,6,11,12-t et r a h yd r och r ysen e-
C
2,8-d iol ((R,R)-2b). Flash chromatography (25% EtOAc/
hexanes) provided diol (R,R)-2b as an off-white powder (145
mg, 78%). Recrystallization from 25% EtOAc/hexanes provided
(R,R)-2b as off-white crystals (100% cis isomer): mp 241-243
°C; er g 99:1 (HPLC column C, 15% IPA/hexanes, tR ) 10.7
1
min); H NMR (400 MHz, acetone-d6) δ 8.23 (s, 2H), 7.16 (d,
2H, J ) 8.2), 6.65-6.72 (m, 4H), 2.89 (dd, 2H, J ) 15.6, 5.5),
2.80 (dd, 2H, J ) 15.6, 1.7), 2.52-2.60 (br m, 2H), 1.43 (dqd,
2H, J ) 14.9, 7.6, 3.2), 1.27 (ddq, 2H, J ) 14.4, 10.4, 7.4), 0.95
(t, 6H, J ) 7.4); 13C NMR (100 MHz, acetone-d6) δ 157.0, 136.7,
132.0, 128.0, 123.9, 116.7, 113.7, 36.7, 32.6, 24.0, 12.3; MS (EI,
70 eV) m/z 320 (M+, 75), 291 (M - Et, 96); HRMS (EI, 70 eV)
calcd for C22H24O2 320.1776, found 320.1771. Anal. (C22H24O2)
C, H.
Ack n ow led gm en t. This work has been supported
through grants from the National Institutes of Health
(PHS 5R37 DK15556 and PHS 5R37 CA18119).
Su p p or tin g In for m a tion Ava ila ble: Proton NMR spec-
tra for 2a -c and 3a and chiral stationary-phase HPLC traces
for 2a -c. This material is available free of charge via the
Internet at http://pubs.acs.org.
(5S,11S)-5,11-Diet h yl-5,6,11,12-t et r a h yd r och r ysen e-
2,8-d iol ((S,S)-2b). Flash chromatography (25% EtOAc/hex-
anes) provided diol (S,S)-2b as a yellow powder (96 mg,
quantitative). Recrystallization from 25% EtOAc/hexanes pro-
vided (S,S)-2b as off-white crystals (100% cis isomer): mp
241-243 °C dec; er g 99:1 (HPLC column C, 15% IPA/hexanes,
tR ) 7.4 min); spectroscopic data identical to those for (R,R)-
Refer en ces
(1) Katzenellenbogen, B. S. Estrogen Receptors: Bioactivities and
Interactions with Cell Signaling Pathways. Biol. Reprod. 1996,
54, 287-293.

2468 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 13
Meyers et al.
(2) Katzenellenbogen, J . A.; O’Malley, B. W.; Katzenellenbogen, B.
S. Tripartite Steroid Hormone Receptor Pharmacology: Interac-
tion with Multiple Effector Sites as a Basis for the Cell- and
Promoter-Specific Action of These Hormones. Mol. Endocrinol.
1996, 10, 119-131.
(3) Mosselman, S.; Polman, J .; Dijkema, R. ERâ: Identification and
Characterization of a Novel Human Estrogen Receptor. FEBS
Lett. 1996, 392, 49-53.
(4) Kuiper, G. G. J . M.; Enmark, E.; Pelto-Huikko, M.; Nilsson, S.;
Gustafsson, J .-Å. Cloning of a Novel Estrogen Receptor Ex-
pressed in Rat Prostate and Ovary. Proc. Natl. Acad. Sci. U.S.A.
1996, 93, 5925-5930.
(5) Kuiper, G. G. J . M.; Carlsson, B.; Grandien, K.; Enmark, E.;
Ha¨ggblad, J .; Nilsson, S.; Gustafsson, J .-Å. Comparison of the
Ligand Binding Specificity and Transcript Tissue Distribution
of Estrogen Receptors R and â. Endocrinology 1997, 138, 863-
870.
(6) Saunders, P. T. K.; Maguire, S. M.; Gaughan, J .; Millar, M. R.
Expression of Oestrogen Receptor Beta (ERâ) in Multiple Rat
Tissues Visualised by Immunohistochemistry. J . Endocrinol.
1997, 154, R13-R16.
(7) Register, T. C.; Adams, M. R. Coronary Artery and Cultured
Aortic Smooth Muscle Cells Express mRNA for Both the Clas-
sical Estrogen Receptor and the Newly Described Estrogen
Receptor Beta. J . Steroid Biochem. Mol. Biol. 1998, 64, 187-
191.
(8) Paech, K.; Webb, P.; Kuiper, G. G. J . M.; Nilsson, S.; Gustafsson,
J .-Å.; Kushner, P. J .; Scanlan, T. S. Differential Ligand Activa-
tion of Estrogen Receptors ERR and ERâ at AP1 Sites. Science
1997, 277, 1508-1510.
(9) Montano, M. M.; J aiswal, A. K.; Kaztenellenbogen, B. S.
Transcriptional Regulation of the Human Quinone Reductase
Gene by Antiestrogen-Liganded Estrogen Receptor-R and Es-
trogen Receptor-â. J . Biol. Chem. 1998, 273, 25443-25449.
(10) Barkhem, T.; Carlsson, B.; Nilsson, Y.; Enmark, E.; Gustafsson,
J .-Å.; Nilsson, S. Differential Response of Estrogen Receptor R
and Estrogen Receptor â to Partial Estrogen Agonists/Antago-
nists. Mol. Pharmacol. 1998, 54, 105-112.
(11) Sun, J .; Meyers, M. J .; Fink, B. E.; Rajendran, R.; Katzenellen-
bogen, J . A.; Katzenellenbogen, B. S. Novel Ligands that
Function as Selective Estrogens or Antiestrogens for Estrogen
Receptor-R or Estrogen Receptor-â. Endocrinology 1999, 140,
800-804.
(12) Collins, D. J .; Matthews, W. A. The Structure and Function of
Oestrogens. I. Synthesis of 4b-Methyl-trans-4b,5,6,10b,11,12-
hexahydrochrysene-2,8-diol. Aust. J . Chem. 1979, 32, 1093-
1106.
(13) Hwang, K.-J .; O’Neil, J . P.; Katzenellenbogen, J . A. 5,6,11,12-
Tetrahydrochrysenes: Synthesis of Rigid Stilbene Systems
Designed To Be Fluorescent Ligands for the Estrogen Receptor.
J . Org. Chem. 1992, 57, 1262-1271.
(14) Myers, A. G.; Yang, B. H.; Chen, H.; McKinstry, L.; Kopecky,
D. J .; Gleason, J . L. Pseudoephedrine as a Practical Chiral
Auxiliary for the Synthesis of Highly Enantiomerically Enriched
Carboxylic Acids, Alcohols, Aldehydes, and Ketones. J . Am.
Chem. Soc. 1997, 119, 6496-6511.
(15) Ru¨hlmann, K. Die Unsetzung von Carbonsa¨ureestern mit
Natrium in Gegenwart von Trimethylchlorsilan. Synthesis 1971,
236-253.
(16) Tedesco, R.; Katzenellenbogen, J . A.; Napolitano, E. 7-R,11-â-
Disubstituted Estrogens: Probes for the Shape of the Ligand
Binding Pocket in the Estrogen Receptor. Bioorg. Med. Chem.
Lett. 1997, 7, 2919-2924.
(17) Stetter, H.; Ra¨msch, R. Y. Uber die pra¨parative Nutzung der
1,3-Thiazoliumsalz-Katalysierten Acyloin - und Benzoin-Bil-
dung; IV. Herstellung von Acyloinen mit Funktionellen Gruppen.
Synthesis 1981, 477-478.
(18) Katzenellenbogen, J . A.; J ohnson, H. J ., J r.; Myers, H. N.
Photoaffinity Labels for Estrogen Binding Proteins of Rat
Uterus. Biochemistry 1973, 12, 4085-4092.
(19) Carlson, K. E.; Choi, I.; Gee, A.; Katzenellenbogen, B. S.;
Katzenellenbogen, J . A. Altered Ligand Binding Properties and
Enhanced Stability of a Constitutively Active Estrogen Recep-
tor: Evidence That an Open-Pocket Conformation Is Required
for Ligand Interaction. Biochemistry 1997, 36, 14897-14905.
(20) Anstead, G. M.; Carlson, K. E.; Katzenellenbogen, J . A. The
Estradiol Pharmacophore: Ligand Structure-Estrogen Receptor
Binding Affinity Relationships and a Model for the Receptor
Binding Site. Steroids 1997, 62, 268-303.
(21) Brzozowski, A. M.; Pike, A. C. W.; Dauter, Z.; Hubbard, R. E.;
Bonn, T.; Engstro¨m, O.; O¨ hman, L.; Greene, G. L.; Gustafsson,
J .-Å.; Carlquist, M. Molecular Basis of Agonism and Antagonism
in the Oestrogen Receptor. Nature 1997, 389, 753-758.
(22) Tanenbaum, D. M.; Wang, Y.; Williams, S. P.; Sigler, P. B.
Crystallographic Comparison of the Estrogen and Progesterone
Receptor’s Ligand Binding Domains. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 5998-6003.
(23) Shiau, A. K.; Barstad, D.; Loria, P. M.; Cheng, L.; Kushner, P.
J .; Agard, D. A.; Greene, G. L. The Structural Basis of Estrogen
Receptor/Coactivator Recognition and the Antagonism of This
Interaction by Tamoxifen. Cell 1998, 95, 927-937.
(24) If one of the cis-THC enantiomers bound with the ligand core
flipped 180° about the C-2/C-8 axis, the substituents from both
enantiomers would project in the same direction. In the flipped
orientation, however, the central two rings of the THC occupy
quite a different spatial position.
(25) Montano, M. M.; Mu¨ller, V.; Trobaugh, A.; Katzenellenbogen,
B. S. The Carboxyl-Terminal F Domain of the Human Estrogen
Receptor: Role in the Transcriptional Activity of the Receptor
and the Effectiveness of Antiestrogens as Estrogen Antagonists.
Mol. Endocrinol. 1995, 9, 814-825.
(26) McInerney, E. M.; Katzenellenbogen, B. S. Different Regions in
Activation Function-1 of the Human Estrogen Receptor Required
for Antiestrogen- and Estradiol-Dependent Transcription Activa-
tion. J . Biol. Chem. 1996, 271, 24172-24178.
(27) McInerney, E. M.; Weis, K. E.; Sun, J .; Mosselman, S.; Katzenel-
lenbogen, B. S. Transcription Activation by the Human Estrogen
Receptor Subtype â (ERâ) Studied with ERâ and ERR Receptor
Chimeras. Endocrinology 1998, 139, 4513-4522.
(28) Arndt, F. Diazomethane. Organic Syntheses; Wiley: New York,
1943; Collect. Vol. 2, pp 165-167.
J M990101B