Triarylpyrazoles with basic side chains: Development of pyrazole-based estrogen receptor antagonists

Article abstract of DOI:10.1016/S0968-0896(00)00226-1

Recently, we developed a novel triaryl-substituted pyrazole ligand system that has high affinity for the estrogen receptor (ER) (Fink, B. E.; Mortenson, D. S.; Stauffer, S. R.; Aron, Z. D.; Katzenellenbogen, J. A. Chem. Biol. 1999, 6, 205). Subsequent wor

Full text of DOI:10.1016/S0968-0896(00)00226-1

Bioorganic & Medicinal Chemistry 9 (2001) 151±161
Triarylpyrazoles with Basic Side Chains:
Development of Pyrazole-Based Estrogen Receptor Antagonists
Shaun R. Stau€er,a Ying R. Huang,a Zachary D. Aron,a Christopher J. Coletta,a
Jun Sun,b Benita S. Katzenellenbogenb,c and John A. Katzenellenbogena,*
aDepartment of Chemistry, University of Illinois and University of Illinois College of Medicine, Urbana, IL 61801, USA
bDepartment of Physiology, University of Illinois and University of Illinois College of Medicine, Urbana, IL 61801, USA
cDepartment of Cell and Structural Biology, University of Illinois and University of Illinois College of Medicine, Urbana, IL 61801, USA
Received 16 June 2000; accepted 15 August 2000
AbstractÐRecently, we developed a novel triaryl-substituted pyrazole ligand system that has high anity for the estrogen receptor
(ER) (Fink, B. E.; Mortenson, D. S.; Stau€er, S. R.; Aron, Z. D.; Katzenellenbogen, J. A. Chem. Biol. 1999, 6, 205). Subsequent
work has shown that some analogues in this series are very selective for the ERa subtype in terms of binding anity and agonist
potency (Stau€er, S. R.; Coletta, C. J.; Tedesco, R.; Sun, J.; Katzenellenbogen, J. A. J. Med. Chem. 2000, submitted). We now
investigate how this pyrazole ER agonist system might be converted into an antagonist or a selective estrogen receptor modi®er
(SERM) by incorporating a basic or polar side chain like those typically found in antiestrogens and known to be essential deter-
minants of their mixed agonist/antagonist character. We selected an N-piperidinyl-ethyl chain as a ®rst attempt, and introduced it at the
four possible sites of substitution on the pyrazole core structure to determine the orientation that the pyrazole might adopt in the ER
ligand binding pocket. Of these four, the C(5) piperidinyl-ethoxy-substituted pyrazole 5 had by far the highest anity. Also, it bound to
the ER subtype alpha (ERa) with 20-fold higher anity than to ERb. In cell-based transcription assays, pyrazole 5 was an antagonist on
both ERa and ERb, and it was also more potent on ERa. Based on structure-binding anity relationships and on molecular modeling
studies of these pyrazoles in a crystal structure of the ERa±raloxifene complex, we propose that pyrazoles having a basic substituent on
the C(5) phenyl group adopt a binding mode that is di€erent from that of the pyrazole agonists that lack this group. The most favorable
orientation appears to be one which places the N(1) phenol in the A-ring binding pocket so that the basic side chain can adopt an
orientation similar to that of the basic side chain of raloxifene. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
side chains of these ligands displace the C-terminal helix
(helix-12) from the position it normally occupies in ER-
LBD complexes with agonists. In this antagonist con-
formation, helix-12 moves into a position where it
occludes a hydrophobic groove on the surface of the
receptor that normally functions as a tethering site for the
interaction of ER with coactivator proteins, an interaction
that is important in mediating agonist activity.3,4
The term SERM, or selective estrogen receptor modi-
®er, is used to describe ligands for the estrogen receptor
(ER) that display an interesting balance of agonist and
antagonist activity that can vary from tissue to tissue.1
SERMs have particular utility in menopausal hormone
replacement and in the treatment and prevention of
breast cancer, although the pharmacological pro®les of
the currently available agents are not optimal.
It is clear, however, that more subtle changes in the
conformation of ER must underlie some of the phar-
macological di€erences among various SERMs.5,6
Recently, conformation-sensitive peptides have been
used to discriminate among these di€erent ligand-
induced ER conformations.7 9 Nevertheless, it is not
completely clear how the ligand structure of the various
SERMs determines ER conformation, at least at this
level of detail. In addition, di€erential interaction of
SERMs with the two ER subtypes, ERa and ERb,
might also underlie certain aspects of their tissue-selec-
tive activity.10 12
SERMs typically consist of a non-steroidal core structure
onto which is attached a side chain bearing a basic or
polar function. X-ray crystal structures of the ER ligand
binding domain (LBD) complexed with the SERMs
raloxifene2 or hydroxytamoxifen3 show that the large
*Corresponding author at Department of Chemistry, University of Illi-
nois, 600 South Mathews Avenue, Urbana, IL 61801, USA. Tel.: +1-
217-333-6310; fax: +1-217-333-7325; e-mail: jkatzene@uiuc.edu
0968-0896/01/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00226-1

152
S. R. Stau€er et al. / Bioorg. Med. Chem. 9 (2001) 151±161
In recent studies of novel, heterocyclic non-steroidal
estrogens amenable to combinatorial assembly and
library synthesis, we identi®ed tetrasubstituted pyr-
azoles as a system that a€orded high anity ligands for
ER.13 Subsequently, we used parallel synthesis, as well
as directed synthesis, to characterize structure±anity
relationships in this series.14,15 In the process, we found
several high anity compounds, and in particular, some
that show very high binding anity and transcriptional
ecacy for the ER alpha subtype (ERa).15 Although
some of these compounds were also agonists on the ER
beta subtype (ERb), their potency and binding anity
on this ER subtype was much lower.15 The high di€er-
ence in ER-subtype binding anities shown by some of
these pyrazoles raised an interesting prospect that this
system might be a favorable starting point for the devel-
opment of ERa potency-selective antagonists or SERMs.
This would depend, of course, on whether these ligands
could be converted to antagonists by appropriate sub-
stitution of basic or polar side chains.
In this report, we describe the preparation of several ana-
logues of an ERa high-anity pyrazole which embody the
type of basic or polar side chain that is typically found in
SERMs. We evaluate how this substitution is tolerated in
terms of ER binding anity, and in some selected cases,
we show it a€ects the transcriptional ecacy and potency
of these ligands. We also develop a model for how these
basic side chain-substituted pyrazoles are likely to be
orientated in the ligand-binding pocket of ER.
Figure 1. Pyrazole triol core structure (1) and basic side chain-con-
taining pyrazoles (2±5).
pyrazoles, which behave as agonists, bind in an orien-
tation in which the C(3) phenol plays a role analogous
to the critical A-ring of estradiol.15 We have illustrated
this binding orientation with the relative positions we
have chosen for pyrazole 1 and estradiol in Figure 1. The
orientation of the SERM raloxifene relative to estradiol,
known from X-ray crystallography,2 is also displayed in
Figure 1. In this orientation, the benzothiophene ring
system of raloxifene mimics the AB ring system of
estradiol, so that the basic side chain is directed roughly
in the estradiol 11b direction (Fig. 1), where it extends
outward to displace helix-12, as noted above.2
Results and Discussion
An example of a favorable pattern of substitution for high
anity ER pyrazoles ligands consists of three phenols at
the 1-, 3- and 5-positions and an alkyl group at the 4-
position, and is shown in Figure 1 (pyrazole 1).13 16 The
piperidinyl-ethyl substituent, which is widely repre-
sented in many SERM classes (see, for example, Ralox-
ifene, Figure 1),1,17 was chosen in this study as a typical
basic side chain that might be expected to confer the
characteristic mixed agonist/antagonist activity of these
compounds. The question at hand, however, is at which
position on a pyrazole such as 1 should this basic sub-
stituent be attached. Molecular modeling studies using
the LBD crystal structures provided a starting point
towards answering this question.
On the basis of our current model, we might have con-
sidered it unlikely that the pyrazole ligand could even
accommodate a basic side chain. When bound with the
ER-LBD in the agonist conformation, there is no position
on the pyrazole analogue where such a side chain could be
disposed so as to occupy a region of the ligand binding
pocket that is normally occupied by this group in other
SERMs (i.e. the estradiol 11b region, Figure 1).2,3 How-
ever, because of the near symmetry of these triaryl pyr-
azoles and the known ¯exibility of the ligand-binding
pocket of the ER,2,3 we were not convinced that the binding
orientation preferences of the smaller core pyrazole agonist
ligands would be retained. The additional bulk and polar-
ity of the basic side chain that we were adding to engen-
der antagonist activity might cause the ligand to adopt
an entirely di€erent orientation in the binding pocket.
From published crystal structures, it is known that the ER-
LBD adopts a di€erent conformation when complexed
with agonists (E2, DES) versus antagonists (hydroxy-
tamoxifen, raloxifene).2,3 In our initial molecular modeling
studies, we examined various binding modes for the pyra-
zole core structure (i.e. without a basic side chain) based on
the larger ligand binding pocket in the ER-LBD±ralox-
ifene (antagonist) complex.13 However, in a subsequent
publication,15 knowing now that these pyrazoles were
agonists, we reexamined our initial predictions, this time
using the ER-LBD±DES (agonist) structure. As a result of
this additional modeling work, as well as further struc-
ture±activity relationship studies we have done in this
series of compounds,15 we now believe that these core
Therefore, as an initial e€ort to develop pyrazole-based
SERMs, we prepared four pyrazoles in which the basic
piperidinyl-ethyl side chain is attached at four di€erent
positions (2±5, Fig. 1), and we determined the binding
anity of these derivatives for the ER. Each of these

S. R. Stau€er et al. / Bioorg. Med. Chem. 9 (2001) 151±161
153
isomers serves as a model to determine which position
on the heterocycle sca€old can best accommodate the
SERM-de®ning basic side chain in the ER-LBD. Once
identi®ed, such a site would obviously be a prime target
for further study through a series of pyrazoles in which
the structure of this basic substituent is varied.18
phenylhydrazine, we initially investigated the classical
diazotization±reduction route, starting from the corre-
sponding substituted aniline 6. However, we had diculty
isolating the desired hydrazine precursor. Fortunately, we
found the approach of Demers and Klaubert,19 involving
metallation of the corresponding aromatic bromide (7)
followed by reaction with an azodicarboxylate ester,
conveniently a€orded the di-Boc protected hydrazine 8.
Because isolation of the free hydrazine proved to be
problematic, we performed the Boc deprotection and
pyrazole cyclization in one-pot, which gave pyrazole 9 in
65% yield. Selective demethylation using either AlCl3±
EtSH or BBr3 a€orded pyrazole 2 in 55±75% yield.
The four basic side chain-containing pyrazoles (2±5)
shown in Figure 1 are designed to probe several poten-
tial ligand orientations within the binding pocket. For
example, pyrazole 2 was proposed on the basis of our
initial docking study of the N-aryl pyrazole 1 in the
ERa±raloxifene X-ray crystal structure13 (now revised,
as noted above15), which placed the C(5) phenol in the
subpocket corresponding to the A-ring of E2. In this
mode, the N(1) basic side chain group resides in what
would be the estradiol 11b direction. Pyrazole 3 was
envisioned to undergo a ligand ``¯ipping'' around the
bond between the pyrazole core and the C(3) phenol, to
place the C(4)-linked basic side chain in the estradiol
11b direction. Pyrazoles 4 and 5 are regioisomers that
have the basic side chain at the 3- and 5-positions,
respectively. Pyrazole 5, a C(5)-substituted analogue,
was predicted to have a potentially favorable interaction
with ER by way of a ca. 120ꢀ counterclockwise rotation
about an axis perpendicular to the pyrazole ring,
resulting in a binding mode that would make the N(1)
phenol the mimic of the A-ring in estradiol. Pyrazole 4,
a regioisomer of pyrazole 5, was considered unlikely to
have a favorable interaction with the ER, because the
phenols and the basic side chain groups are displayed in
a ``long'' orientation. This leaves the remaining two free
phenols too close to one another for proper interaction
with the A-ring and D-ring subpockets in ER. Never-
theless, it was prepared for the sake of completeness.
The synthesis of the C(4) basic side chain-containing
pyrazole 3 is shown in Scheme 2. Introduction of a basic
side chain into the pyrazole structure at the 4-position
was done early in the synthesis, starting with Freidel±
Crafts acylation of anisole with 6-bromohexanoyl chlo-
ride. For this transformation, the temperature had to be
carefully maintained at 10 ꢀC to avoid reaction of the
product ketone 10 with a second equivalent of anisole, a
side reaction that results in the formation of a 1,1-di-
aryl-alkene by-product upon dehydration. The bromo-
ketone 10 was then aminated with excess piperidine in
DMF to a€ord amino-ketone 11 in high yield. Con-
densation with the p-nitrophenyl benzoate furnished the
b-diketone 12, which was cyclized with phenyl hydra-
zine hydrochloride to give the desired protected pyra-
zole 13. Deprotection using AlCl3±EtSH yielded the
®nal pyrazole 3.
The ®nal pyrazoles 4 and 5 were initially prepared as an
approx. 1:1 mixture of regioisomers by the route shown
in Scheme 3. Starting from 4-hydroxybutryophenone,
we used a Mitsunobu reaction to prepare the piperidinyl-
ethoxy-substituted ketone 14. Claisen condensation
a€orded the requisite b-diketone 15. Treatment with
phenylhydrazine hydrochloride followed by BBr3
deprotection a€orded the desired pyrazoles 4 and 5 as a
Chemical syntheses
The synthesis of pyrazole 2 is shown in Scheme 1. For the
preparation of the required piperidinyl-ethoxy-substituted
Scheme 1. Synthesis of N(1) basic side chain-containing pyrazole 2.

154
S. R. Stau€er et al. / Bioorg. Med. Chem. 9 (2001) 151±161
Scheme 2. Synthesis of C(4) basic side chain-containing pyrazole 3.
Scheme 3. Synthesis of C(3) and C(5) basic side chain-containing pyrazoles 4 and 5.
1:1 mixture that could be separated by chromatography.
Pyrazole 4 was further puri®ed by recrystallization from
MeOH, and de®nitive regiochemical assignment of this
isomer was made on the basis of a single crystal X-ray
analysis. An ORTEP diagram of this compound is
shown in Figure 2.
Estrogen receptor binding anity of pyrazoles 2±5
The binding anity of pyrazoles 2±5 for ER was
assayed in a competitive radiometric binding assay as
described previously,20,21 and anities are expressed as
relative binding anity (RBA) values, where the anity
of estradiol is considered to be 100% (Table 1). The
anity of all compounds was tested both in a natural
ER preparation from lamb uterine cytosol,21 which is
predominantly ERa,10,11 as well as with puri®ed recom-
binant human ERa and ERb (see Experimental).20
Isolation of the last pyrazole (5) proved to be more
problematical because we found that the basic side
chain of these pyrazoles undergoes partial cleavage
when BBr3 is used as a deprotection reagent. This is a
signi®cant problem, because the triphenolic by-product
resulting from this cleavage was dicult to separate from
the desired C(5) basic side chain isomer by chromato-
graphy and it has distinctly di€erent biological activity.15
Later, we found that we could do this deprotection selec-
tively, without this side reaction, using the milder AlCl3±
EtSH reagent. Ultimately, however, we used a regiose-
lective route (described elsewhere in detail18) to produce
the desired pyrazole 5.
When uterine cytosol was used as a source of estrogen
receptor, pyrazole 2 had a relatively low binding anity,
just 1.5% that of estradiol, a value which is 10-fold lower
than that of the parent N(1) aryl pyrazole (1).13,15 Pyra-
zole 3 with the basic side chain on C(4) also displayed
very low anity for the receptor. Apparently, even with
the reorientation proposed above, this analogue, in
which the basic side chain is attached to the pyrazole

S. R. Stau€er et al. / Bioorg. Med. Chem. 9 (2001) 151±161
155
Figure 2. X-ray crystal structure for 4 (ORTEP; ellipsoids drawn at
the 35% probability level).
Figure 3. Transcription activation pro®le for pyrazole 5 with ERa and
ERb. Human endometrial cancer (HEC-1) cells were transfected with
expression vectors for ERa or ERb and an (ERE)3-pS2-CAT reporter
gene and were treated with indicated concentrations of pyrazole 5
alone (solid lines) or pyrazole 5 in the presence of 10 9 M estradiol
(E2) (dashed lines) for 24 h. CAT activity was normalized for b-galac-
tosidase activity from an internal control plasmid. Values are the mean
ÆSD for three separate experiments, and are expressed as a percent of
the ERa or ERb response with 10 9 M E2, which is set at 100%. For
some values, error bars are too small to be visible.
core solely through aliphatic linkages, does not ®t well
in the ligand binding pocket of ER. More intriguing are
the two regioisomeric pyrazoles 4 and 5; they display
very distinctive preferences for binding to the ER, pyr-
azole 5 having a 65- to 90-fold higher anity than pyr-
azole 4 in all three test systems. As noted previously, the
higher anity of isomer 5 versus 4 was expected.
Pyrazole 5 has the highest anity of all basic side chain
pyrazoles investigated thus far, and, in fact, is among
the highest anity of all pyrazoles we have reported as
ER ligands.13 15 When we evaluated the binding anity
of these pyrazoles with pure human ERa and ERb (Table
1),20 pyrazole 5 showed a distinctive preference for bind-
ing to ERa (ca. 20-fold), a characteristic that it shares with
the parent pyrazole triol 1 (ca. 230-fold).15 Pyrazole 2
shows only a 5-fold anity preference for ERa. Pyrazoles
3 and 4 had very low anity for both ERa and ERb, and
they were not investigated further either for binding
anity or transcriptional activity.
assayed at di€erent concentrations of ligand in the pre-
9
sence of 10 M E2.
Pyrazole 5 showed no stimulation of transcriptional
activity of ERa or ERb, and it was a full antagonist on
ERa and ERb, completely antagonizing activity stimu-
lated by estradiol (Fig. 3). Thus, introduction of the
type of basic side chain typically found in the mixed
agonist/antagonist SERMs at the proper site on the
parent pyrazole 1, which is an agonist, converts this
pyrazole system into an antagonist (pyrazole 5). The
relatively good ERa binding anity selectivity of pyra-
zole 5 (ca. 20-fold; cf. Table 1) is preserved to a con-
siderable degree in its potency as a transcriptional
antagonist. Its IC50 value is ca. 20 nM on ERa and ca.
160 nM on ERb, which gives an ERa antagonist potency
selectivity of ca. 10-fold (Fig. 3).
Transcription activity of pyrazole 5
Based on the promising ERa binding selectivity of pyr-
azole 5 and because of its piperidine side chain, we were
interested in whether this compound had antagonistic
properties, and in particular, whether it might function
as an antagonist on one or both of the ER subtypes in a
potency selective manner. Figure 3 shows the transcrip-
tional activity pro®les of pyrazole 5 assayed in HEC-1
cells with either ERa or ERb. Cells were transfected with
an expression plasmid for ERa or ERb, together with an
estrogen-responsive reporter gene construct ((ERE)3-
pS2-CAT), and reporter gene expression was determined
at di€erent concentrations of ligand. Antagonism was
In work to be presented elsewhere, we have prepared
additional analogues of pyrazole 5 having other basic
and polar side chains as substituents on the C(5) phenyl
group.18 Some of these have very high anity selectivity
for ERa over ERb, and they act as ERa potency selec-
tive antagonists (J. Sun et al., in preparation).
Binding models for pyrazoles 2 and 5 in the ERꢀ±
raloxifene crystal structure
Table 1. Estrogen receptor binding anity of pyrazoles 1±5
As previously mentioned, the low binding anity of
pyrazole 4 was not surprising. The low anity of pyra-
zole 3, however, was unexpected, because, as mentioned
earlier, when the pyrazole core is ¯ipped, the basic side
chain attached to C(4) appears to be orientated properly
for this substituent to project through the 11b channel.
Compound Site of basic
side chain
RBA lamb
uterine cytosola
RBA
hERaa
RBA
hERba
1
2
3
4
5
[none]
N(1)
C(4)
C(3)
C(5)
20.3Æ3
35.7Æ6
0.15Æ0.01
1.5Æ0.4
0.65Æ0.13 0.13Æ0.03
0.017Æ0.006
0.29Æ0.26
24.5Æ1
0.05b
0.13b
0.019b
0.01b
0.65Æ0.02
11.5Æ1
Pyrazole 2, in which the basic side chain is substituted on
the N(1) phenyl group, was initially believed to be the
most favorable candidate for interacting with ER. How-
ever, this turned out not to be the case; pyrazole 5 has the
aRelative binding anity (RBA) where estradiol=100%. Values are
the average of repeat determinationsÆSD (nꢁ3) or Ærange (n=2). For
details of the assay procedure, see Experimental.
bSingle determinations.

156
S. R. Stau€er et al. / Bioorg. Med. Chem. 9 (2001) 151±161
crystal structure.2 Both ligands were docked in the
ERa±raloxifene crystal structure in a similar manner,
and they were subjected to a three-step minimization
protocol to generate the ®nal receptor±ligand model
complexes shown (see Experimental, Molecular Model-
ing for details). Residue side chains that were close to
each ligand in its initial position were allowed to
undergo torsional movement during the docking opera-
tion, in order to obtain an optimal ®t, prior to mini-
mization. A crossed-stereo view of pyrazoles 2 and 5 is
shown in Figure 5(A) and (B). The ®nal models have
similar energies, which at least at this level of re®ne-
ment, simply indicates that they have no bad contacts or
unfavorable conformations.
Figure 4. Pyrazoles 2 and 5 oriented with corresponding aryl rings
oriented in a congruent fashion.
highest anity. These two pyrazoles have somewhat
related structures (cf. Fig. 1); the three aryl groups (two
unsubstituted and one side chain-substituted) can be
superimposed in a congruent fashion, as illustrated in
Figure 4. When oriented in this fashion, the only di€er-
ence between these two isomers is the position of the
two ring nitrogen atoms and the C(4) ethyl substituent.
In both of the views shown in Figure 5, the three aryl
groups displayed peripherally from the pyrazole core
are accommodated comfortably and in a very similar
manner within the ligand binding pocket. This was
achieved with only minimal and energetically favorable
movement of a few residues that were initially in close
contact with the ligand.
In an attempt to understand the 16-fold higher anity of
pyrazole 5 than pyrazole 2, we performed some docking
studies of these two ligands with the ERa±raloxifene
Figure 5. Crossed-stereo views of pyrazole 2 (top) and 5 (bottom) docked and minimized in ERa±raloxifene structure ligand binding pocket showing
residues within 2.5 A of ligand. The pyrazoles were initially positioned with respect to certain reference atoms of the ligand raloxifene, and then
minimized together with the same contact residues in the protein using the Flexidock routine within SYBYL. For details, see Experimental. The
yellow surface is a cutaway of the Connolly surface of the protein ligand-binding pocket.

S. R. Stau€er et al. / Bioorg. Med. Chem. 9 (2001) 151±161
157
The most signi®cant di€erence between the complexes
Experimental
with pyrazoles 2 and 5 is the position of the ethyl group
(cf. Fig. 4). In Figure 5(B), the C(4) ethyl group of pyr-
azole 5 projects into a region of ER that accommodates
the 18-methyl of estradiol in the ER±E2 complex.2 In this
subpocket, the favorable van der Waals interaction that
normally occurs between Leu525 and the 18-methyl group
of E2 is conveniently replaced by a similar interaction
with the ethyl substituent of pyrazole 5 (Fig. 5(B)).
General
Melting points were determined on a Thomas-Hoover
UniMelt capillary apparatus and are uncorrected. All
reagents and solvents were obtained from Aldrich,
Fisher or Mallinckrodt. Tetrahydrofuran was freshly
distilled from sodium/benzophenone. Dimethylforma-
mide was vacuum distilled prior to use, and stored over
4 A molecular sieves. n-Butyllithium was titrated with
N-pivaloyl-o-toluidine. Et3N was stirred with phenyliso-
cyanate, ®ltered, distilled, and stored over 4 A molecular
sieves. All reactions were performed under a dry N2
atmosphere unless otherwise speci®ed. Reaction progress
was monitored by analytical thin-layer chromatography
using GF silica plates purchased from Analtech. Visua-
lization was achieved under illumination with short wave
UV light (254 nm) or with a potassium permanganate
indicator spray. Radial preparative-layer chromato-
graphy was performed on a Chromatotron instrument
(Harrison Research, Inc., Palo Alto, CA) using EM
Science silica gel Kieselgel 60 PF254 as adsorbent. Flash
column chromatography was performed using Woelm
32±63 mm silica gel packing. The preparation of pyra-
zole 5 by a regioselective route is described elsewhere.18
By contrast, in the complex with pyrazole 2, the ethyl
substituent is oriented in a di€erent position (down-
ward), where it no longer bene®ts as much from a favor-
able hydrophobic±hydrophobic interaction, as was the
case in the complex with pyrazole 5. Leu525 is some-
what closer to the pyrazole core, where it interacts with
the N(1) phenyl ring. The closest residue for interaction
with the C(4) ethyl of pyrazole 2 (Fig. 5(A)) is Met388.
This residue is a somewhat ``softer'' partner for interac-
tion than Leu525, and may be less stabilizing. Other
di€erences between the ER complexes with pyrazoles 2
and 5 which could account for their di€ering binding
anity may have to do with electronic e€ects and are
not considered in these models. As noted earlier (Fig. 4),
the positions of the nitrogen atoms are di€erent between
the two binding modes, so the electronic molecular sur-
faces that are presented to the receptor are predicted to
be quite di€erent.
1H and 13C NMR spectra were recorded on either a Var-
ian Unity 400 or 500 MHz spectrometers using CDCl3 or
MeOD-d4 as solvent. Chemical shifts were reported as
parts per million down®eld from an internal tetra-
methylsilane standard (d 0.0 for 1H) or relative to solvent
peaks. NMR coupling constants are reported in Hertz.
13C NMR spectra were determined using either the
Attached Proton Test (APT) or standard 13C pulse
sequence parameters. Low-resolution and high-resolu-
tion electron impact mass spectra were obtained on Fin-
nigan MAT CH-5 or 70-VSE spectrometers. Elemental
analyses were performed by the Microanalytical Service
Laboratory of the University of Illinois. Single crystal X-
ray analysis on pyrazole 4 was performed to determine
connectivity. A rigorous analysis was not performed
because of the presence of a disordered solvent molecule
in the unit cell. However, for the purposes of regio-
chemical assignment, the analysis was satisfactory.
Conclusions
We have prepared a series of basic side chain-sub-
stituted pyrazoles as potential selective antagonists for
ER subtypes. The initial four analogues were designed
to explore the four possible orientations that the pyra-
zole core structure might adopt in the ER ligand bind-
ing pocket, and thereby to identify which site on the
pyrazole could best accommodate the rather large, basic
side chain substituent. Of the four analogues, the C(5)
piperidinyl-ethoxy-substituted pyrazole 5 was found to
have the highest anity for ERa, with an anity among
the highest observed for all ER ligands of the pyrazole
type. It also shows considerable anity selectivity for ERa
versus ERb. Through cell-based transfection assays, we
found that pyrazole 5 was an antagonist on both ERa and
ERb, and that its potency on ERa was somewhat
higher, re¯ecting its ERa anity selectivity. Modeling
studies using the crystal structure of the ERa±raloxifene
complex were used to compare potential binding modes
for pyrazoles 2 and 5. The binding modes appear to be
quite similar, and the origin of the higher anity for
pyrazole 5 versus 2 may result from a more favorable
positioning of the ethyl group in the binding pocket in
the complex with pyrazole 5. In work to be presented
elsewhere, we have prepared additional analogues of
pyrazole 5 having other basic and polar side chains as
substituents on the C(5) phenyl group.18 Some of these
have very high anity selectivity for ERa over ERb,
and they act as ERa potency selective antagonists (J.
Sun et al., in preparation). This work is the ®rst e€ort
towards developing a novel SERM based on a pyrazole
core structure.
Biological procedures
Relative binding anity assay. Ligand binding anities
(RBAs) using lamb uterine cytosol as a receptor source
were determined by a competitive radiometric binding
assay using 10 nM [3H]estradiol as tracer and dextran-
coated charcoal as an adsorbent for free ligand, as pre-
viously described.21 Binding anities with puri®ed
recombinant human ERa and ERb were determined by
a competitive radiometric binding assay using 10 nM
[3H]estradiol as tracer, commercially available ERa and
ERb preparations (PanVera Inc. Madison, WI), and
hydroxylapatite (HAP) to adsorb bound receptor±ligand
complex.20 HAP was prepared following the recom-
mendations of Williams and Gorski.22 All incubations
were done at 0 ꢀC for 18±24 h. Binding anities are
expressed relative to estradiol (RBA=100%) and are
reproducible with a coecient of variation of 0.3.

158
S. R. Stau€er et al. / Bioorg. Med. Chem. 9 (2001) 151±161
Transcriptional activation assay. Human endometrial
cancer (HEC-1) cells were maintained in culture and
transfected as described previously.23,24 Transfection of
HEC-1 cells in 60 mm dishes used 0.4 mL of a calcium
phosphate precipitate containing 0.5 mg of pCMVbGal
as internal control, 2 mg of the reporter gene plasmid,
100 ng of the ER expression vector, and carrier DNA to
a total of 5 mg DNA. Compounds were added to the cell
culture media as ethanol solutions so as to yield a ®nal
ethanol concentration of 0.1%. CAT activity, normal-
ized for the internal control b-galactosidase activity,
was assayed as previously described.23,24
of the side chain residues within 8 A of the ligand while
holding the backbone and residues Glu353 and Arg394
®xed. A ®nal, third minimization of both the ligand and
receptor was conducted using the Anneal function (hot
radius 8 A, interesting radius 16 A), again holding resi-
dues Glu353 and Arg394 ®xed. The result of this was
considered to be the ®nal model. Minimizations were
done using the TRIPOS Force®eld (MAXIMIN) with
the Powell gradient method and default settings (®nal
.
RMS <0.05 kcal/mol A). The total energy for the ERa
complex with pyrazole 2 was 806 kcal/mol and for the
complex with pyrazole 5 was 868 kcal/mol. At this level
of re®nement, these numbers cannot be used to predict
relative binding anities. However, they indicate that
there are no serious errors or unfavorable contacts in
the ER-LBD ligand models.
Molecular modeling
Small molecule modeling. The starting structures of pyr-
azoles 2 and 5 used for the ER-LBD docking studies
were generated from random conformational searches
using the TRIPOS force ®eld as implemented in Sybyl
6.5.2. Low energy conformers then underwent full
minimization until a convergence criterion of 0.001 kcal/
mol was met using the AM1 semi-empirical force ®eld.
Charge calculations were done using Gasteiger±Huckel
method and molecular surface properties displayed
using MOLCAD module in Sybyl 6.5.2.
Chemical synthesis
1-[4-(2-N-Piperidinyl-ethoxy)-phenyl]-3,5-di-(4-hydroxy-
phenyl)-4-ethyl-pyrazole (2). To a stirred solution of the
protected pyrazole 9 (100mg, 0.19mmol) in CH2Cl2
(1.0 mL) at 0 ꢀC was added BBr3 (1.0M CH2Cl2, 5 equiv,
0.95mL) dropwise. The reaction was kept at 0 ꢀC for 4 h
and then quenched with water. The solution was neu-
tralized with satd NaHCO3 and then repeatedly extracted
with EtOAc. The combined organic layers were washed
with brine, dried over Na2SO4 and concentrated to
a€ord a tan residue. Puri®cation by ¯ash chromato-
graphy (5% TEA and 5% MeOH in CH2Cl2) a€orded 2
as an o€-white powder (75%): mp 125±130 ꢀC (decom-
posed); 1H NMR (CDCl3, 400 MHz) d 0.98 (t, 3H,
J=7.3), 1.51 (br s, 2H), 1.62 (quint, 4H, J=5.5), 2.58 (q,
2H, 7.5), 2.67 (br s, 4H), 2.89 (t, 2H, J=5.5), 4.13 (t, 2H,
J=5.5), 6.76 (d, 2H, J=8.5), 6.87 (d, overlapping, 2H,
J=8.5), 6.88 (d, overlapping, 2H, J=9), 7.03 (d, 2H,
J=8.5), 7.16 (d, 2H, J=9.0), 7.49 (d, 2H, J=8.5); APT
13C NMR (CDCl3, 100 MHz) d 9.9 (CH3), 18.5 (CH2),
24.8 (CH2), 26.2 (CH2), 54.1 (CH2), 56.1 (CH2), 58.8
(CH2), 116.1 (CH), 116.8 (CH), 116.9 (CH), 121.4 (C),
123.1 (C), 126.8 (C), 128.5 (CH), 130.9 (CH), 133.0
(CH), 135.2 (C), 144.0 (C), 152.6 (C), 159.1 (C), 159.4
(C), 159.5 (C); MS (EI, 70 eV) m/z 483.3 (M); HRMS
(EI) calcd for C30H33N3O2 (M+ HBr): 483.2522.
Found: 481.2365 (M+ HBr 2).
Receptor docking studies. The lowest energy conformers
for pyrazoles 2 and 5, generated as indicated above,
were used for docking studies. Prior to docking, a +1
formal charge and a proton were added to each piper-
idinyl nitrogen. Ligands were then pre-positioned in the
crystal structure of the ERa±raloxifene complex,2 using
a least squares multi®tting of the most congruent atoms
of the pyrazole and raloxifene. In raloxifene this inclu-
ded the A-ring carbon atoms and the 1,4-carbon atoms
in both the 2- and 3-substituted phenyl groups. The
same ®tting strategy was used in the pyrazoles, all six
carbon atoms in the ring corresponding to the A-ring
and the 1,4-carbon atoms of the other two aryl rings.
Once pre-positioned, the raloxifene structure was
deleted and then the pyrazole ligand was optimally
docked in the ERa binding pocket using the TRIPOS
Flexidock routine. Hydrogen bond donors (Arg394,
His524) and hydrogen bond acceptors (Glu353, His524,
and Asp351) in ER were indicated in Flexidock, as well
as the hydrogen bond donors and acceptors in the ligand.
Side chains that were allowed to rotate during docking
of the pyrazoles 2 and 5 included Met543, Leu539,
Leu536, Leu354, Asp351, Trp383, all of which are near
the basic side chain group; Phe404, Leu387, and Met388,
which are near the A-ring mimic; Leu384, Leu346,
Leu428, and Trp383, which are near the B/C ring
region; and Ile424, His524, and Leu525, which are near
the D-ring subpocket. Arg394 and Glu353, the hydro-
gen bonding partners of the critical A-ring phenol
mimic, were kept ®xed during the entire docking and
minimization.
4-(4-N-Piperdinyl-butyl)-3,5-di-(4-hydroxyphenyl)-1-phe-
nylpyrazole hydrochloride (3). To a stirred CH2Cl2
(1.0 mL) solution of the protected pyrazole 13 (72 mg,
0.15 mmol) was added AlCl3 (116 mg, 0.87 mmol). The
mixture was cooled to 0 ꢀC and EtSH (55 mL, 0.73 mmol)
added dropwise. The mixture was then allowed to reach
rt and stirred for an additional 0.5 h. At this time the
reaction was concentrated under a stream of nitrogen in
a fume hood, brought up in CH2Cl2 and concentrated
once again. The crude mixture was re-dissolved in
CH2Cl2 and a mixture of THF (1 mL), water (1 mL),
and 6 M HCl (0.25 mL) added. The resulting precipitate
was isolated by vacuum ®ltration and was washed with
four 5 mL portions of water followed by four 5 mL
portions of ether to a€ord pyrazole 3 as a white solid
The best docked ligand ER-LBD complex from the
Flexidock routine then underwent a three-step mini-
mization: First, non-ring torsional bonds of the ligand
were minimized in the context of the receptor, using the
torsmin command. This was followed by minimization
1
(37 mg, 54%): H NMR (MeOD-d4, 500 MHz) d 1.30±
1.60 (m, 6H), 1.75 (br s, 6H), 2.73 (t, 2H, J=7.2), 2.78

S. R. Stau€er et al. / Bioorg. Med. Chem. 9 (2001) 151±161
159
(t, 4H, J=8.1), 6.90 (XX0 of AA0XX0, 2H, JAX=8.8,
an addition funnel. After 1 h at 0 ꢀC the reaction was
warmed to rt and stirred for 6 h. The crude mixture was
then concentrated in vacuo and the Ph3PO byproduct
removed by vacuum ®ltration. The ®ltrate was again con-
centrated and the product puri®ed by gradient ¯ash chro-
matography using 5% TEA and 30% EtOAc in hexanes
up to 5% TEA and 50% EtOAc in hexanes to a€ord 7 as a
clear oil (85%):25 1H NMR (CDCl3, 500 MHz) d 1.41 (m,
2H), 1.58 (m, 4H), 2.47 (br s, 4H), 2.72 (d, 2H, J=7.2),
4.05 (d, 2H, J=7.6), 6.79 (d, 2H, 8.3), 7.82 (d, 2H, J=8.1);
13C NMR (CDCl3, 125 MHz) d 22.1, 38.0, 67.5, 113.0,
116.5, 121.8, 123.8, 132.3, 136.6, 149.6, 158.0, 158.3; MS
(EI, 70 eV) m/z 284 (M+).
0
JXX =2.5), 7.07 (AA of AA0XX0,0 2H, JAX=8.8,
JAA =2.5), 7.23±7.36 (m, 5H), 7.32 (XX of AA XX , 2H,
00
0
0
0
JAX=8.6, JXX =2.4), 7.52 (AA of AA0XX0, 2H,
JAX=8.6, JAA =2.2); 13C NMR (MeOD-d4, 125 MHz) d
0
0
0
21.5, 22.5, 22.3, 23.0, 27.0, 52.8, 56.6, 115.1, 115.2, 117.4,
121.1, 124.8, 124.9, 127.0, 127.3, 128.5, 131.2, 139.8, 142.4,
151.5, 157.5, 157.9; MS (EI, 70eV) m/z (relative intensity,
%): 467 (M+ HCl, 21); HRMS (EI, M+) calcd for
C30H33N3O2: 467.2573. Found: 467.2579.
1-(4-Hydroxyphenyl)-3-[4-(2-N-piperidinyl-ethoxy)-
phenyl]-4-ethyl-5-(4-hydroxy-phenyl)-1H-pyrazole (4). A
CH2Cl2 solution (10 mL) of the crude protected pyra-
zole isomers (300 mg, 0.60 mmol) prepared from the 1,3-
dione 15 was cooled to 0 ꢀC and treated dropwise with
BBr3 (1 M in hexane, 5 equiv, 3 mmol). The reaction
was allowed to reach temperature and stir overnight.
The mixture was re-cooled to 0 ꢀC and quenched with
5 mL of MeOH and the solvent removed under reduced
pressure. To the crude mixture 8 mL of H2O was added
and the aqueous layer neutralized with a satd aqueous
NaHCO3 solution and extracted repeatedly with EtOAc
(5Â20mL). The organic layers were washed with brine,
dried over Na2SO4 and concentrated in vacuo to a€ord
a brown residue. Pyrazole 4 was cleanly separated from
the regioisomer 5 and other byproducts by ¯ash chro-
matography (7% Et3N and 7% MeOH in CH2Cl2). It
was further puri®ed by recrystallization from MeOH to
a€ord the C(3) isomer (4) as small white crystals whose
structure was veri®ed by X-ray analysis (23%). Pyrazole
N,N-di-tert-Butoxycarbonyl-N-[4-(2-N-piperidinyl-ethoxy)
phenyl]hydrazine (8). To a THF solution at 70 ꢀC con-
taining the substituted bromobenzene 7 (10.56 mmol)
was added n-BuLi (1.3 equiv) dropwise over 10 min.
After 30 min, tert-butoxycarbonyl-azodicarboxylate (15.84
mmol) was added as a solid in one portion. The mixture
was then slowly warmed to rt and 1 equiv of dilute
AcOH added. After stirring the quenched mixture for
30 min, the reaction was partitioned between water
(30 mL) and Et2O (50 mL). The layers were separated
and the aqueous layer extracted twice more with Et2O
(2Â50 mL). The combined organic layers were washed
with brine and dried over Na2SO4. Subsequent solvent
removal a€orded 6.8 g of a crude red oil. Flash chro-
matography (5% TEA and 25% EtOAc in hexanes)
produced 8 as a white solid (75%): mp 50±55 ꢀC; (note:
spectrum complicated by Boc rotamers); 1H NMR
(CDCl3, 400 MHz) d 1.38 (s, 9H), 1.40 (overlapping s,
11H), 1.61 (m, 4H), 2.45 (m, 4H), 2.73 (t, 2H, J=5.6),
4.05 (t, 2H, J=5.3), 6.82 (d, 2H, J=7.8), 7.25 (br s, 2H);
13C NMR (CDCl3, 100 MHz) d 24.2 (CH2), 24.8 (CH2),
28.2 (CH3), 28.3 (CH3), 36.8 (CH2), 46.0 (CH2), 46.7
(CH2), 55.1 (CH2), 57.8 (CH2), 60.4 (CH2), 65.7 (CH2),
66.0 (CH2), 81.0 (CH2), 81.8 (CH2), 113.9 (CH), 114.6
(CH), 120.7 (CH), 129.4 (CH), 156.7 (C), 158.7 (C),
171.2 (C); MS (EI, 70 eV) m/z 435.4 (M, 35%). Anal.
(C23H37N3O5): C, 63.42; H, 8.56; N, 9.65. Found: C,
63.20; H, 8.72; N, 9.47.
4: mp 151±153 ꢀC; H NMR ((CD3)2SO, 400 MHz) d
1
1.01 (t, 3H, J=7.6), 1.43 (m, 2H), 1.55 (m, 4H), 2.49 (br
s, 4H), 2.59 (q, 2H, J=7.6), 2.72 (t, 2H, J=5.6), 4.15 (t,
2H, J=6.0), 6.74 (d, 2H, J=8.4), 6.83 (d, 2H, J=8.4),
7.08 (apparent triplet due to three overlapping doublets,
6H, J=7.6), 7.67 (d, 2H, J=8.4), 9.71 (br s, 2H, OH
exchangeable with D2O); APT 13C NMR ((CD3)2SO,
100 MHz) d 15.4 (CH3), 16.8 (CH2), 23.9 (CH2), 25.6
(CH2), 54.5 (CH2), 57.5 (CH2), 65.6 (CH2), 114.5 (CH),
115.1 (CH), 115.4 (CH), 118.7 (C), 120.9 (C), 126.1
(CH), 126.5 (C), 128.5 (CH), 131.2 (CH), 131.8 (C),
141.2 (C), 148.5 (C), 156.2 (C), 157.4 (C), 157.9 (C);
HRMS (EI, M+) calcd for C30H33N3O3: 483.2521.
Found: 483.2516.
1-[4-(2-N-Piperidinyl-ethoxy)-phenyl]-3,5-bis-(4-methoxy-
phenyl)-4-ethyl-1H-pyrazole (9). To a stirred solution of
DMF (14mL), THF (6 mL), and 1,3-bis-(4-methoxy-
phenyl)-2-ethyl-propane-1,3-dione (500 mg, 1.52 mmol)15
was added hydrazine 8 in 3-fold excess (800mg, 4.59
mmol). The mixture was brought to re¯ux for 5 h and
then recooled to rt. The THF was removed under
reduced pressure and the remaining residue diluted with
H2O (30 mL) and extracted repeatedly with EtOAc. The
organic layers were washed with brine, dried over
Na2SO4 and concentrated to produce a crude red±
orange oil. Flash chromatography using 10% TEA and
20% EtOAc in hexanes a€orded the pure product 9 as
As was noted in the text, isolation of the other regio-
isomer (pyrazole 5) from this reaction mixture was
complicated by the fact that it could not be separated
e€ectively from the triphenolic byproduct that resulted
from partial cleavage of the basic side chain during the
BBr3 deprotection. This compound was prepared in a
regioselective fashion by another route that is described
elsewhere.18
1-Bromo-4-(2-N-piperidinyl-ethoxy)benzene (7). In
a
1
250 mL ¯ask charged with Ph3P (35 mmol) in 60 mL of
THF at 0 ꢀC was added di-iso-propyl-azodicarboxylate
(DIAD, 35 mmol) reagent dropwise. The solution was
kept at 0 ꢀC and stirred for 50 min. At this time N-(2-
hydroxyethyl)-piperidine was added dropwise via syr-
inge. After 10 min, p-bromophenol (35 mmol) and TEA
(86.7 mmol) in 20 mL THF were added dropwise via
an orange oil (65%): H NMR (CDCl3, 400 MHz) d
1.03 (t, 3H, J=8.1), 1.42 (m, 2H), 1.58 (quin, 4H, J=6);
2.47 (br s, 2H), 2.61 (q, 2H, J=6.2), 2.72 (t, 2H, J=6),
3.79 (s, 3H), 3.82 (s, 3H), 4.04 (t, 2H, J=6.2), 6.76 (d,
2H, J=9.0), 6.86 (d, 2H, J=9.0), 6.96 (d, 2H, J=9.1),
7.13 (d, 2H, J=9.1), 7.16 (d, 2H, J=9.1), 7.69 (d, 2H,
J=9.0); 13C NMR (CDCl3, 100 MHz) d 15.6 (CH3),

160
S. R. Stau€er et al. / Bioorg. Med. Chem. 9 (2001) 151±161
17.2 (CH2), 24.3 (CH2), 26.0 (CH2), 55.1 (CH3), 55.2
(CH3), 57.9 (CH2), 66.2 (CH2), 113.8 (CH), 113.9 (CH),
114.5 (CH), 119.9 (C), 123.3 (C), 126.0 (CH), 126.9 (C),
129.0 (CH), 131.3 (CH), 133.6 (C), 141.0 (C), 150.1 (C),
157.4 (C), 159.1 (C), 159.4 (C); HRMS (EI, M+) calcd
for C32H37N3O2: 511.2834. Found: 511.2825.
DIC and DMAP coupling conditions) in THF (5 mL)
was added dropwise. The resulting solution was allowed
to stir for 15 h at rt and 4 h at re¯ux (oil bath tempera-
ture of 70 ꢀC). At this time the mixture was allowed to
come to rt and quenched by the addition of H2O (10 mL).
The organic layer was isolated and washed with H2O
(3Â15 mL), dried over Na2SO4, and concentrated under
reduced pressure to a€ord a yellow solid. The remaining
ester was removed by warming the crude mixture in
15% ethyl acetate/hexanes and ®ltering o€ the excess
crystalline ester. The ®ltrate was concentrated under
reduced pressure and puri®ed by ¯ash chromatography
(10% TEA and 55% EtOAc in hexanes) to a€ord 12 as
6-Bromo-1-(4-methoxy-phenyl)-hexan-1-one (10). To a
stirred solution of AlCl3 (12.40 g, 92.75 mmol) in 1,2-
dichloroethane (30 mL) at 0 ꢀC was added 6-bromohex-
anoyl chloride (23.70 g, 17 mL, 111.30 mmol) dropwise
over 10 min. The resulting solution was stirred at rt for
0.5 h, cooled to 15 ꢀC and a solution of anisole
(10.00 g, 10.05 mL, 92.75 mmol) in 1,2-dichloroethane
(10 mL) was added dropwise over 20 min. The reaction
mixture was allowed to stir at 15 ꢀC for 1 h, then
quenched with H2O (50 mL). The aqueous layer was
isolated and extracted with CH2Cl2 (2Â50 mL), and the
organic layers were combined and washed with H2O
(2Â50 mL), saturated NaHCO3 (2Â50 mL), and brine
(2Â50 mL). Subsequent drying over Na2SO4 and solvent
removal in vacuo a€orded a crude orange oil. Upon
standing at rt, large crystals formed after 48 h. The
crystals were isolated and rinsed with cold hexane to
a
yellow oil (227 mg, 51%): 1H NMR (CDCl3,
400 MHz) d 1.22±1.48 (m, 4H), 1.50±1.75 (m, 6H), 2.10
(q, 2H, J=7.5), 2.34 (t, 2H, J=9.8), 2.42 (br s, 4H), 3.83
(s, 6H), 5.05 (t, 1H, J=6.6), 6.89 (d, 4H, J=8.9), 7.94 (d,
4H, J=8.9); 13C NMR (CDCl3, 100 MHz) d 24.0, 25.3,
26.1, 26.3, 29.3, 54.3, 55.4, 57.0, 58.7, 113.9, 128.9,
128.9, 130.8, 130.8, 163.6, 194.6.
3,5-Bis-(4-methoxyphenyl)-4-(4-N-piperidinyl-butyl)-1-
phenyl-1H-pyrazole (13). To a stirred solution of DMF
(45 mL), THF (20 mL), and b-diketone 12 (170 mg,
0.40 mmol) was added phenylhydrazine hydrochloride
(289 mg, 2.00 mmol). The solution was brought to re¯ux
(oil bath temperature between 110 and 120 ꢀC) for 22 h.
The reaction mixture was allowed to cool to rt, and the
THF was evaporated under reduced pressure. The
remaining mixture was diluted with H2O (40 mL) and
extracted with EtOAc (3Â40 mL). The organic layers
were combined and washed with saturated LiCl
(3Â40 mL), saturated NaHCO3 (2Â40 mL), H2O
(2Â40 mL), and brine (2Â40 mL). The organic layer was
dried over MgSO4 and concentrated under reduced
pressure to a€ord a crude brown oil. Puri®cation by
¯ash chromatography (10% TEA and 55% EtOAc in
hexanes) a€orded 13 as a reddish oil (79 mg, 40%).
Compound 13 was veri®ed by 1H NMR and used in the
next step of the reaction scheme without further char-
acterization.
a€ord 10 (23 g, 87%): mp 50±52 ꢀC; H NMR (CDCl3,
1
400 MHz) d 1.52 (quint, 2H, J=7.6), 1.76 (quint, 2H,
J=7.5), 1.91 (quint, 2H, J=7.1), 2.94 (t, 2H, J=7.3),
3.42 (t, 2H, J=6.7), 3.87 (s, 3H), 6.93 (d, 2H, J=8.8),
7.94 (d, 2H, J=9.0); 13C NMR (CDCl3, 100 MHz) d
23.39, 27.79, 32.51, 33.59, 37.80, 55.34, 113.56, 129.88,
130.13, 163.24, 198.48; HRMS (EI, M+) calcd for
C13H17O2Br: 286.0391. Found: 286.0394.
1-(4-Methoxy-phenyl)-6-(N-piperidinyl)-hexan-1-one (11).
To a stirred solution of ketone 10 (248 mg, 1.00 mmol)
in DMF (30 mL) at 0 ꢀC was added piperidine (170 mg,
0.20 mL, 2.00 mmol) dropwise over 5 min. The solution
was placed in a 70 ꢀC oil bath for 4 h and then cooled to
rt. The crude mixture was concentrated under reduced
pressure to remove a majority of the DMF and excess
piperidine to a€ord an orange residue. The residue was
taken up in CH2Cl2 (30 mL), washed with brine
(2Â20 mL), dried over MgSO4 and concentrated under
reduced pressure to a€ord 11 as a white solid (227 mg,
1-[4-(2-N-Piperidinyl-ethoxy)-phenyl]-butan-1-one (14).
To a solution of PPh3 (10.05 g, 38.31 mmol) in THF
(150 mL) at 0 ꢀC was added diisopropyl azodicarbox-
ylate (7.5 mL, 38.31 mmol) dropwise. After stirring at
0 ꢀC for 50 min, N-piperidinylethanol (5.09 mL, 38.31
mmol) was added followed 10 min later by addition of a
solution of 4-hydroxy-butyrophenone (5.24 g, 31.9
mmol) in Et3N (13.4 mL, 95.8 mmol) and THF (30 mL).
The mixture was allowed to warm to rt and stirred
overnight and then concentrated in vacuo. To the crude
residue was added Et2O (150 mL) and the insoluble
Ph3PO by-product removed by ®ltration. The ®ltrate
was concentrated once again and the residue puri®ed by
¯ash chromatography (10% TEA and 25% EtOAc in
hexanes) to a€ord an oil. This material was dissolved in
Et2O (60 mL) and acidi®ed with HCl (4 M in dioxane,
6.3 mL). The precipitate that formed was collected by
vacuum ®ltration and partitioned between saturated
NaHCO3 (35 mL) and ether (50 mL). The organic layer
was washed with brine, dried over Na2SO4 and con-
centrated under reduced pressure to give 14 as pale-yellow
79%): mp 163±165 ꢀC; H NMR (CDCl3, 500 MHz) d
1
1.35±1.45 (m, 2H), 1.65±2.00 (m, 8H), 2.26 (br q, 2H,
J=8), 2.59 (br q, 2H, J=7), 2.88±2.96 (m, 4H), 3.51 (br
d, 2H), 3.88 (s, 3H), 6.93 (d, 2H, J=8.9), 7.93 (d, 2H,
J=8.9); 13C NMR (CDCl3, 100 MHz) d 22.1, 22.3, 23.3,
23.4, 26.4, 37.5, 53.1, 55.5, 57.2, 113.8, 129.9, 130.3,
163.5, 198.5; MS (EI, 70 eV) m/z (relative intensity, %):
289 (M+, 3); HRMS (EI, M+) calcd for C19H27NO2:
289.2042. Found: 289.2060.
1,3-Bis-(4-methoxy-phenyl)-2-(4-N-piperidinyl-butyl)-pro-
pane-1,3-dione (12). To a solution of ketone 11 (350 mg,
1.21 mmol) and TMEDA (140.6 mg, 0.18 mL, 1.21
mmol) in THF (15 mL) at 0 ꢀC was added a 1.0 M
solution of LiHMDS (3.02 mL, 3.02 mmol) dropwise.
The solution was allowed to stir at rt for 0.5 h, then was
cooled back to 0 ꢀC. At this time a solution of 4-nitro-
phenyl 4-methoxybenzoate (990 mg, 3.63mmol; prepared
from p-nitrophenol and 4-methoxybenzoic acid using

S. R. Stau€er et al. / Bioorg. Med. Chem. 9 (2001) 151±161
161
crystals (3.05 g, 35%). An analytically pure sample was
References and Notes
obtained by recrystallization from hexane: mp 48±50 ꢀC;
1H NMR (CDCl3, 400 MHz) d 0.99 (t, 3H, J=7.5),
1.43±1.49 (m, 2H), 1.64 (quint, 4H, J=5.9), 1.75 (sext,
2H, J=7.6), 2.55 (br s, 4H), 2.87 (t, 2H, J=5.9), 2.90 (t,
2H, J=7.2), 4.19 (t, 2H, J=6.0), 6.93 (XX0 of AA0XX0,
1. Grese, T. A.; Dodge, J. A. Curr. Pharm. Design 1998, 4, 71.
2. Brzozowski, A. M.; Pike, A. C.; Dauter, Z.; Hubbard, R.
E.; Bonn, T.; Engstrom, O.; Ohman, L.; Greene, G. L.; Gus-
tafsson, J.-A.; Carlquist, M. Nature 1997, 389, 753.
3. Shiau, A. K.; Barstad, D.; Loria, P. M.; Cheng, L.; Kush-
ner, P. J.; Agard, D. A.; Greene, G. L. Cell 1998, 95, 927.
4. Darimont, B. D.; Wagner, R. L.; Apriletti, J. W.; Stallcup,
M. R.; Kushner, P. J.; Baxter, J. D.; Fletterick, R. J.; Yama-
moto, K. R. Genes Develop. 1998, 12, 3343.
0
0
0
0
2H, JAX=8.9, JXX =2.5), 7.93 (AA of AA XX , 2H,
0
JAX=9.0, JAA =2.5); 13C NMR d 14.1, 18.2, 24.2, 26.0,
40.4, 55.3, 57.9, 66.3, 114.4, 130.5, 162.7, 199.2; HRMS
calcd for C17H26NO2: 276.1963. Found: 276.1964.
5. Katzenellenbogen, J. A.; O'Malley, B. W.; Katzenellenbo-
gen, B. S. Mol. Endocrinol. 1996, 10, 119.
2-Ethyl-1-(4-methoxyphenyl)-3-[4-(2-N-piperidinyl-
a THF
6. McDonnell, D. P.; Clemm, D. L.; Hermann, T.; Goldman,
M. E.; Pike, J. W. Mol. Endocrinol. 1995, 9, 659.
7. Paige, L. A.; Christensen, D. J.; Gron, H.; Norris, J. D.;
Gottlin, E. B.; Padilla, K. M.; Chang, C.; Ballas, L. M.;
Hamilton, P. T.; McDonnell, D. P.; Fowlkes, D. M. Proc.
Natl. Acad. Sci. USA 1999, 96, 3999.
8. Norris, J. D.; Paige, L. A.; Christensen, D. J.; Chang, C.
Y.; Huacani, M. R.; Fan, D.; Hamilton, P. T.; Fowlkes, D.
M.; McDonnell, D. P. Science 1999, 285, 744.
9. Chang, C. Y.; Norris, J. D.; Gron, H.; Paige, L. A.;
Hamilton, P. T.; Kenan, D. J.; Fowlkes, D.; McDonnell, D. P.
Mol. Cell Biol. 1999, 19, 8226.
10. Kuiper, G.; Shughrue, P. J.; Merchenthaler, I.; Gus-
tafsson, J. A. Frontier Neuroendocrinol. 1998, 19, 253.
11. Kuiper, G. G. J. M.; Gustafsson, J.-A. FEBS Lett. 1997,
410, 87.
12. Kuiper, G. G. J. M.; Carlsson, B.; Grandien, K.; Enmark,
E.; Haggblad, J.; Nilsson, S.; Gustafsson, J. A. Endocrinology
1997, 138, 863.
13. Fink, B. E.; Mortensen, D. S.; Stau€er, S. R.; Aron, Z. D.;
Katzenellenbogen, J. A. Chem. Biol. 1999, 6, 205.
14. Stau€er, S. R.; Katzenellenbogen, J. A. J. Combinat.
Chem. 2000, 2, 318.
15. Stau€er, S. R.; Coletta, C. J.; Tedesco, R.; Sun, J.; Kat-
zenellenbogen, B. S.; Katzenellenbogen, J. A. J. Med. Chem.
2000, submitted.
ethoxy)-phenyl]-propane-1,3-dione (15). To
(25 mL) solution of ketone 14 (500 mg, 1.82 mmol) and
4-nitrophenyl 4-methoxybenzoate (745.3 mg, 2.73
mmol) at 0 ꢀC was added LiHMDS (1 M solution in
hexane, 4.55 mL, 4.55 mmol) dropwise. Upon complete
addition of LiHMDS the solution was allowed to warm
to rt and stir overnight. The reaction mixture was con-
centrated to near dryness and EtOAc (35 mL) added.
The crude mixture was washed sequentially with satu-
rated NaHCO3 and brine. The solvent was dried over
Na2SO4, removed in vacuo and the resulting oil puri®ed
via ¯ash chromatography (10% TEA and 50% EtOAc
in hexanes) to give 15 as a viscous red oil (465 mg,
1
63%): H NMR (CDCl3, 400 MHz) d 1.03 (t, J=7.31),
1.49 (br s, 2H), 1.70 (br s, 4H), 2.16 (quint, J=7.29 Hz,
2H), 2.63 (br s, 4H), 2.90 (br s, 2H), 3.85 (s, 3H), 4.25
(br s, 2H), 4.96 (t, J=6.76 Hz, 1H), 6.91 (d, J=9.05 Hz,
4H), 7.95 (d, J=8.96 Hz, 2H), 7.96 (d, J=8.85 Hz, 2H);
13C NMR (CDCl3, 125 MHz) d 12.8, 23.1, 24.0, 25.8,
55.0, 55.4, 57.6, 59.0, 66.2, 113.9, 114.5, 116.5, 129.2,
129.3, 130.9, 131.0, 163.0, 163.7, 194.9; HRMS (EI,
M+) calcd for C25H31NO4: 409.2253. Found: 407.2096
(M+ 2).
16. Stau€er, S. R.; Huang, Y.; Coletta, C. J.; Katzenellenbo-
gen, J. A. Bioorg. Med. Chem. 2000, 9, 141.
17. Magarian, R. A.; Overacre, L. B.; Singh, S.; Meyer, K. L.
Curr. Med. Chem. 1994, 1, 61.
Acknowledgements
18. Huang, Y.; Katzenellenbogen, J. A. Org. Lett. 2000, 2,
2833.
19. Demers, J. P.; Klaubert, D. H. Tetrahedron Lett. 1987, 42,
4933.
We are grateful for support of this research through
grants from the US Army Breast Cancer Research Pro-
gram (DAMD17-97-1-7076) and the National Institutes
of Health (PHS 5R37 DK15556 (to J. A. K.), PHS 5R37
CA18119 (to B. S. K.), PHS T32 CA09067 (traineeship
for Y. R. H.)). We thank Kathryn E. Carlson for per-
forming binding assays and Rosanna Tedesco for help
with molecular graphics. NMR spectra were obtained in
the Varian Oxford Instrument Center for Excellence
NMR Laboratory. Funding for this instrumentation
was provided in part from the W. M. Keck Foundation
and the National Science Foundation (NSF CHE 96-
10502). Mass spectra were obtained on instruments
supported by grants from the National Institute of
General Medical Sciences (GM 27019), the National
Institute of Health (RR 01575), and the National
Science Foundation (PCM 8121494).
20. Carlson, K. E.; Choi, I.; Gee, A.; Katzenellenbogen, B. S.;
Katzenellenbogen, J. A. Biochemistry 1997, 36, 14897.
21. Katzenellenbogen, J. A.; Johnson, H. J. Jr; Myers, H. N.
Biochemistry 1973, 12, 4085.
22. Williams, D.; Gorski, J. Biochemistry 1974, 13, 5537.
23. McInerney, E. M.; Weis, K. E.; Sun, J.; Mosselman, S.;
Katzenellenbogen, B. S. Endocrinology 1998, 139, 4513.
24. Sun, J.; Meyers, M. J.; Fink, B. E.; Rajendran, R.; Kat-
zenellenbogen, J. A.; Katzenellenbogen, B. S. Endocrinology
1999, 140, 800.
25. Grese, T. A.; Pennington, L. D.; Sluka, J. P.; Adrian, M.
D.; Cole, H. W.; Fuson, T. R.; Magee, D. E.; Phillips, D. L.;
Rowley, E. R.; Shetler, P. K.; Short, L. L.; Venugopalan, M.;
Yang, N. N.; Sato, M.; Glasebrook, A. L.; Bryant, H. U. J.
Med. Chem. 1998, 41, 1272.