Pharmacophore analysis of the nuclear oxysterol receptor LXRα

Article abstract of DOI:10.1021/jm0004749

A cell-free assay was developed for the orphan nuclear receptor LXRα that measures the ligand-dependent recruitment of a peptide from the steroid receptor coactivator 1 (SRC1) to the nuclear receptor. Using this ligand-sensing assay (LiSA), the structural requirements for activation of the receptor by oxysterols and related compounds were studied. The minimal pharmacophore for receptor activation was shown to be a sterol with a hydrogen bond acceptor at C24.24(S),25-Epoxycholesterol (1), which meets this criterion, is among the most efficacious of the oxysterols and is an attractive candidate as the LXRα natural hormone. Cholenic acid dimethylamide (14) showed increased efficacy compared to 1, whereas the unnatural oxysterol 22(S)-hydroxycholesterol (4) was shown to be an antagonist of 1 in the LiSA. The structural requirements for SRC1 recruitment in the LiSA correlated with the transcriptional activity of compounds in a cell-based reporter assay employing LXRα-GAL4 chimeric receptors. Site-directed mutagenesis identified Trp⁴⁴³ as an amino acid critical for activation of LXRα by oxysterol ligands. This information was combined with the structure-activity relationship developed from the LiSA to develop a 3D homology model of LXRα. This model may aid the design of synthetic drugs targeted at this transcriptional regulator of cholesterol homeostasis.

Full text of DOI:10.1021/jm0004749

886
J . Med. Chem. 2001, 44, 886-897
P h a r m a cop h or e An a lysis of th e Nu clea r Oxyster ol Recep tor LXRr
Thomas A. Spencer,† Dansu Li,† J onathon S. Russel,† J on L. Collins,| Randy K. Bledsoe,‡ Thomas G. Consler,‡
Linda B. Moore, Cristin M. Galardi, David D. McKee, J ohn T. Moore, Michael A. Watson, Derek J . Parks,#
Millard H. Lambert,§ and Timothy M. Willson*,|
Dartmouth College, Hanover, New Hampshire 03755, and Glaxo Wellcome Research and Development, Research Triangle
Park, North Carolina 27709
Received November 8, 2000
A cell-free assay was developed for the orphan nuclear receptor LXRR that measures the ligand-
dependent recruitment of a peptide from the steroid receptor coactivator 1 (SRC1) to the nuclear
receptor. Using this ligand-sensing assay (LiSA), the structural requirements for activation of
the receptor by oxysterols and related compounds were studied. The minimal pharmacophore
for receptor activation was shown to be a sterol with a hydrogen bond acceptor at C24. 24(S),25-
Epoxycholesterol (1), which meets this criterion, is among the most efficacious of the oxysterols
and is an attractive candidate as the LXRR natural hormone. Cholenic acid dimethylamide
(14) showed increased efficacy compared to 1, whereas the unnatural oxysterol 22(S)-
hydroxycholesterol (4) was shown to be an antagonist of 1 in the LiSA. The structural
requirements for SRC1 recruitment in the LiSA correlated with the transcriptional activity of
compounds in a cell-based reporter assay employing LXRR-GAL4 chimeric receptors. Site-
directed mutagenesis identified Trp443 as an amino acid critical for activation of LXRR by
oxysterol ligands. This information was combined with the structure-activity relationship
developed from the LiSA to develop a 3D homology model of LXRR. This model may aid the
design of synthetic drugs targeted at this transcriptional regulator of cholesterol homeostasis.
In tr od u ction
functions as a heterodimer with the 9-cis-retinoic acid
receptor RXR (NR2B) to regulate gene expression.4
Recently, several genes involved in cholesterol metabo-
lism and homeostasis have been shown to be regulated
by the LXRR/RXR heterodimer, including CYP7A,5,11
CETP,15 and ABCA1.16-19 On the heels of these exciting
results, LXRR has emerged as a promising target for
development of drugs for the treatment of cardiovascu-
lar disease.2,20
The steroid, retinoid, and thyroid hormones function
as chemical messengers through activation of nuclear
receptors within cells.1 These receptors are ligand-
activated transcription factors which couple the level
of their cognate hormones to changes in gene expression.
To date, the nuclear receptor gene family encodes
approximately 50 sequence-related proteins, the major-
ity of which were isolated as orphan receptors prior to
the identification of their cognate hormones.2 The liver
X receptors, LXRR (NR1H3) and LXRâ (NR1H2), to-
gether with the farnesoid X receptor, FXR (NR1H4),
form a subfamily of these orphan receptors.3,4 Recently,
we and others have reported that LXR and FXR are
hormone receptors for oxysterols5-7 and bile acids,8-10
respectively. Unlike the classical steroid hormone recep-
tors, LXR and FXR bind to their cognate hormones with
relatively low affinity. Among the oxysterols that acti-
vate LXRR most effectively is 24(S),25-epoxycholesterol
(1).5,11 The concentrations of this oxysterol in the
liver12-14 are consistent with its proposed role5 as an
endogenous activator of LXRR in that organ. LXRR
Initial studies of the structural requirements for acti-
vation of LXRR employed cell-based reporter assays.5-7
Data from these assays suggested that the presence of
an oxygen atom on the sterol side chain, which can
function as a hydrogen bond acceptor, was important
for receptor activation. Unfortunately, the low potency
of oxysterol ligands in cell-based assays has hampered
the generation of precise structure-activity relation-
ships (SARs). A subsequent study using a competition
binding assay confirmed the preference for a hydrogen
bond acceptor at C24 of the sterol side chain.21 Notably,
oxysterols bind to LXRR with higher affinity than was
anticipated from their potency in cell-based assays.
Nuclear receptors regulate gene transcription through
interaction with cellular cofactors, known as coactiva-
tors and corepressors.22,23 Agonist ligands promote
displacement of corepressors and recruitment of coac-
tivators, which can remodel chromatin and modulate
the assembly of the basal transcriptional machinery
within cells.22,23 Structural and biochemical studies have
revealed that a short R-helical LxxLL sequence on the
coactivator protein is critical for its interaction with the
C-terminal domain of the nuclear receptor. In the
present study, we developed a ligand-sensing assay
(LiSA) that measures ligand-dependent recruitment of
* Address correspondence to: Timothy M. Willson, GlaxoSmith-
Kline, Nuclear Receptor Discovery Research, NTH-M1421, Five Moore
Drive, Research Triangle Park, NC 27709-3398. Tel: (919) 483-9875.
Fax: (919) 315-0430. E-mail: tmw20653@gsk.com.
† Department of Chemistry, Dartmouth College.
| Department of Medicinal Chemistry, Glaxo Wellcome Research and
Development.
‡ Department of Molecular Sciences, Glaxo Wellcome Research and
Development.
Department of Molecular Endocrinology, Glaxo Wellcome Research
and Development.
#
Department of Molecular Biochemistry, Glaxo Wellcome Research
and Development.
§ Department of Structural Chemistry, Glaxo Wellcome Research
and Development.
10.1021/jm0004749 CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/03/2001

Pharmacophore Analysis of LXRR
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 6 887
Ch a r t 1. Structures of Oxysterols (A) and Related Compounds (B)a
a
All compounds contain the 3â-hydroxy-∆5-androstane skeleton with the respective side chains at C17. The carbon atom at the site of
the hydrogen bond donor substitution on the sterol side chain is numbered.
Sch em e 1a
a peptide from the steroid receptor coactivator 1 (SRC1)24
to LXRR by fluorescence resonance energy transfer. The
structural requirement for activation of LXRR was
evaluated using a set of oxysterols, which included all
the potential regio- and stereoisomers between C22 and
the end of the sterol side chain, and a set of nitrogen-
containing analogues with potentially excellent hydro-
gen bond acceptor capability at C24 (Chart 1). The
resulting SARs were used to develop a pharmacophore
model of LXRR that may aid the design of new ligands
for this orphan receptor.
Resu lts
Syn th esis of Oxyster ols a n d Rela ted An a logu es.
The prospective ligands that were not commercially
available were prepared as follows. 24(S),25-Epoxycho-
lesterol (1) and its 24(R)-epimer (2) were synthesized
as recently described.25 23(S)- and 23(R)-Hydroxycho-
lesterol (6 and 7) were prepared (Scheme 1) from
i-steroid iodide 21,26 which had been employed in the
synthesis of 1 and 2, via lithiation27 and reaction with
isovaleraldehyde. Treatment of the resulting product
mixture with aqueous acid gave 6 and 7 which were
separated chromatographically, as has previously been
accomplished.28,29 24(S)- and 24(R)-Hydroxycholesterol
(8 and 9) were prepared as we have recently described14
by modification of the approach used by Ourisson.30 27-
Hydroxycholesterol (11) was synthesized by the method
of Schroepfer.31 The unusual oxysterol 12 with an
unnatural, modified side chain structure was available
from a study of the synthesis of 1.32 24-Ketocholesterol
(13) was synthesized, as shown in Scheme 1, by alky-
a
Reagents: (i) t-BuLi, pentane; (ii) (CH3)2CHCH2CHO, -78 °C;
(iii) p-TsOH, H2O, dioxane; (iv) abs. EtOH, Na, (CH3)2CHCH2CO2Et;
(v) Al2O3, H2O, dioxane, ∆; (vi) LiCH2CON(CH3)2, THF, -78 to 0
°C; (vii) 1. t-Bu(CH3)2SiCl, imidazole, DMAP, 2. (COCl)2, CH2Cl2,
DMF, 3. CH3NH2.
lation of iodide 21 with the anion of ethyl isovalerate
to afford 22, which was converted to 13 by treatment
with Al2O3 in aqueous dioxane at reflux,33 followed by
treatment with aqueous acid.

888 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 6
Spencer et al.
Ta ble 1. Activity of Compounds in the LXRR LiSAa
Sch em e 2a
compound
name
LXRR
no.
EC50 (nM) RE
1
2
3
4
5
6
7
8
9
24(S),25-epoxycholesterol
24(R),25-epoxycholesterol
20(S)-hydroxycholesterol
22(S)-hydroxycholesterol
22(R)-hydroxycholesterol
23(S)-hydroxycholesterol
23(R)-hydroxycholesterol
24(S)-hydroxycholesterol
24(R)-hydroxycholesterol
460 ( 80
670 ( 60
470 ( 20
1.0
1.0
1.0
ia
325 ( 60
5900 ( 470 0.70
0.85
ia
1.0
0.85
0.35
0.30
ia
130 ( 30
220 ( 50
1160 ( 20
250 ( 50
10 25-hydroxycholesterol
11 27-hydroxycholesterol
12 23(R)-CH2OH-25-bisnorergoster-
3â,24-diol
13 24-ketocholesterol
180 ( 20
170 ( 20
720 ( 80
1.3
1.4
0.55
ia
14 cholenic acid dimethylamide
15 cholenic acid monomethylamide
16 22-[N-(2′-methylpropionyl)amino]-23,24-
bisnorchol-5-en-3â-ol
17 22-[N-(N,N-dimethylformamido)amino]-
23,24-bisnorchol-5-en-3â-ol
18 cholenic acid
ia
3000 ( 1000 0.2
105 ( 5
19 cholenic acid methyl ester
20 cholesterol
1.1
ia
a
EC50 ) concentration of the test compound that leads to half-
maximal activity ( standard error, n ) 3; RE ) relative efficacy
of the test compound determined by the maximal increase in
relative fluorescence compared to 24(S),25-epoxycholesterol (1); ia
) RE < 0.1.
a
Reagents: (i) NaN3, benzene, DMF; (ii) LiAlH4, Et2O; (iii) 1.
(CH3)2CHCOCl, pyridine, Et2O, 2. p-TsOH, H2O, dioxane; (iv)
(CH3)2NCOCl, pyridine, Et2O; (v) p-TsOH, H2O, dioxane.
Dimethylamide 14 had previously been prepared from
cholenic acid (18),34,35 but since we had in hand25,32
substantial quantities of C22 iodide 21, we chose to
develop an alternate synthesis. Alkylation of 21 with
the anion of N,N-dimethylacetamide afforded 93% of
i-steroid 23, which upon treatment with aqueous acid
yielded 68% of 14, as shown in Scheme 1. Monomethyl-
amide 15 was prepared in 63% overall yield from
cholenic acid (18) by treatment successively with TB-
DMS chloride, oxalyl chloride, and methylamine without
isolation of intermediates. This procedure afforded 15
directly, and it was not investigated whether the TB-
DMS derivative of the C3 hydroxyl group of 18 had not
formed or was cleaved by HCl generated during the
reaction with oxalyl chloride. Iodide 21 was also used
to prepare inverse amide 16 and urea 17 (Scheme 2).
Treatment of 21 with NaN3 gave 93% of azide 24, which
was reduced with LiAlH4 to amine 25 in 79% yield.
Treatment of 25 with isobutyryl chloride, followed by
aqueous acid treatment of the crude product, afforded
the desired amide 16 in 44% overall yield. An analogous
reaction sequence with 25 using carbamyl chloride
afforded 22% of urea 17 and 37% of 3â-chloro compound
27. The latter presumably resulted from the presence
of chloride ion during the aqueous acid deprotection
step. When the intermediate i-steroid intermediate 26
was isolated and purified before conversion to 17, the
overall yield of 17 from 25 was 56%.
lated, and complexed with the streptavidin-labeled
fluorophore allophycocyanin. A peptide encompassing
the second LxxLL motif8 (amino acids 676-700) of
SRC124 was synthesized, biotinylated, and complexed
with streptavidin-labeled europium chelate. Time-
resolved fluorescence resonance energy transfer was
used to monitor the interaction between the SRC1
peptide and the LXRR ligand-binding domain. A basal
interaction was detected, which increased in the pres-
ence of increasing concentrations of known LXRR ago-
nists.
Activa tion of LXRr by Oxyster ols. The putative
LXRR natural ligand 24(S),25-epoxycholesterol (1)
showed a saturable dose-dependent increase in fluores-
cence in the LiSA with an EC50 of 460 nM (Table 1 and
Figure 1A). This value is in good agreement with the
reported Ki of 200 nM for binding to LXRR21 but is lower
than the estimated EC50 of 4-7 µM in cell-based
reporter assays.5,21 Notably, the increase in SRC1
recruitment is consistent with the agonist activity of 1
in these cell-based reporter assays. The unnatural
epimer 24(R),25-epoxycholesterol (2) exhibited a higher
EC50 but equivalent efficacy for recruitment of the SRC1
peptide to LXRR (Figure 1A). Both 1 and 2 were
previously shown to be full agonists in a cell-based
reporter assay,5 and the difference in EC50 for SRC1
recruitment may contribute to the lower potency of 2
compared to 1 in cells. The oxysterols 3-11 were
assayed to evaluate the effect of regio- and stereochem-
istry of hydroxyl substitution between C20-C27 of the
cholesterol side chain (Chart 1). At C20, oxysterol 3 with
the S-stereochemistry showed equivalent activity to 1.
At C22, only the R-isomer 5 was active, although it was
less efficacious than 1. At C23, only the S-isomer 6 was
active, although it was less efficacious and less potent
than 1. At C24, both isomers were active, but the
S-isomer 8 was more efficacious than its R-epimer 9.
Develop m en t of a Cell-fr ee Assa y for LXRr
Activa tion . To study the structural requirements for
LXRR activation by oxysterols, we chose to develop a
cell-free assay that monitors the ligand-dependent
recruitment of SRC1. A related cell-free LiSA has
recently been employed to study the structural require-
ments for activation of FXR by bile acids and synthetic
ligands.8,36 The purified human LXRR ligand-binding
domain was expressed from E. coli, purified, biotiny-

Pharmacophore Analysis of LXRR
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 6 889
raised by cholesterol feeding in rats,13,14 making 1 and
8 the leading candidates for physiological regulators of
LXRR activity in the liver.13,14
Activa tion of LXR r by 24-Ketoch olester ol (13)
a n d Rela ted Com p ou n d s. To further refine the phar-
macophore for LXRR activation, a series of side chain-
modified cholesterol derivatives was studied. All of the
analogues were designed to have a hydrogen bond
acceptor at C24 (Chart 1B). 24-Ketocholesterol (13)
showed comparable activity to 1 (Table 1), as was
previously shown using a cell-based assay.5 Among a
series of nitrogen-substituted analogues, dimethylamide
14 was notable for its increased efficacy for SRC1
recruitment compared to 1. The secondary amide 15
showed low efficacy, and the inverse amide 16 and urea
17 were inactive (Table 1 and Figure 1B). Thus, nitrogen
substitution appears to be tolerated at C25 but not at
C23 on the cholesterol side chain. Cholenic acid (18) was
weakly active, but its methyl ester 19 showed activity
comparable to 1. These results demonstrate that a
hydrogen bond acceptor at C24 fulfills the minimal
structural requirement for activation of LXRR by ste-
roidal ligands. The best synthetic activators are sub-
stituted with a carbonyl function at this position, which
may function as an isosteric replacement for the epoxide
in 1.
F igu r e 1. Dose-response analysis of selected oxysterols and
related compounds in the LXRR LiSA. (A) Oxysterols: 24(S),25-
epoxycholesterol (1; b), 24(R),25-epoxycholesterol (2; O), 23(S)-
hydroxycholesterol (6; ]), 23(R)-hydroxycholesterol (7; [), 25-
hydroxycholesterol (10; 2), and 27-hydroxycholesterol (11; 4).
(B) Nitrogen-containing analogues: cholenic acid dimethyl-
amide (14; O), cholenic acid monomethylamide (15; [), 22-[N-
(2′-methylpropionyl)amino]-23,24-bisnorchol-5-en-3â-ol (16; ]),
and 22-[N-(N,N-dimethylformamido)amino]-23,24-bisnorchol-
5-en-3â-ol (17; 2). 24(S),25-Epoxycholesterol (1; b) is shown
for comparison. Data are representative of 3 independent
experiments.
In cr ea sed Effica cy of LXRrActiva tion by Ch olen -
ic Acid Dim eth yla m id e (14). Since dimethylamide 14
showed increased recruitment of the SRC1 peptide in
the LXRR LiSA compared to 1, we decided to evaluate
its activity in an established cell-based reporter as-
say.5,21 CV-1 cells were transfected with a plasmid
expressing a chimera of the human LXRR ligand-
binding domain fused to the yeast transcription factor
GAL4 and a reporter plasmid containing the GAL4
upstream activation sequence (UAS) fused to a minimal
promoter and the gene for secreted alkaline phos-
phatase. It was previously established that oxysterols
show equivalent activity on either the chimeric LXRR-
GAL4 or full-length LXRR receptors.5,21 The induction
of alkaline phosphatase expression was measured by a
standard colorimetric assay in the presence and absence
of 10 µM amides 14-16 and 24(S),25-epoxycholesterol
(1) (Figure 2A). The dimethylamide 14 showed greater
activation of the reporter gene compared to 1, while the
secondary amides 15 and 16 were much less active.
Notably, the relative activities of 1 and 14-16 in cells
paralleled their efficacy for SRC1 recruitment in the
LXRR LiSA (Figure 1B). Full dose-response analysis
showed that the increased activity of 14 was due to its
greater efficacy on LXRR rather than increased potency
compared to 1 (Figure 2B). Thus, the increase in
recruitment of SRC1 by 14 measured in the LiSA
translates to increased efficacy for activation of LXRR
in cells.
Both the C25- and C27-substituted oxysterols 10 and
11 showed low efficacy in the LiSA. The weak activity
of 25-hydroxycholesterol (10) on LXRR contrasts with
its activity as a suppressor of SREBP signaling,37 which
suggests that these two regulators of cholesterol me-
tabolism may be under the control of different pools of
oxysterols.
The effect of the regio- and stereochemistry of hy-
droxyl substitution measured in the LiSA is in general
agreement with the SAR observed in cell-based
assays.5-7,21 However, the increased potency of oxys-
terols in the cell-free LiSA allows a more precise
measurement of relative potency and efficacy (Figure 1
and Table 1). The activity of 24(R)-hydroxycholesterol
(9) in the LXRR LiSA was surprising, since it did not
displace [3H]-1 in a competition binding assay.21 How-
ever, we note that 9 has been reported to activate LXRR
in a cell-based assay,5 which is consistent with the LiSA
data. The C20- and C24-substituted oxysterols (Chart
1) showed the best activity in the LiSA, with 1-3 and
8 demonstrating the most efficacy for recruitment of
SRC1. Of these oxysterols, 24(R),25-epoxycholesterol (2)
has not been detected as a naturally occurring com-
pound38 and 20(S)-hydroxycholesterol (3), like 22(R)-
hydroxycholesterol (5), is only found in the adrenals as
an intermediate in pregnenolone biosynthesis.39 By
contrast, both 24(S),25-epoxycholesterol (1) and 24(S)-
hydroxycholesterol (8) have been detected in liver
extracts at physiologically relevant concentrations.12,14,40
Importantly, the hepatic level of these oxysterols is
Id en tifica tion of a n LXRr An ta gon ist. All but one
of the oxysterols that had been reported to bind to
LXRR21 were also active in the LiSA. The exception was
22(S)-hydroxycholesterol (4), which binds to LXRR with
a Ki of 180 nM21 but did not recruit SRC1 in the LiSA
assay at concentrations up to 100 µM (Table 1). Since
this oxysterol binds but does not activate LXRR, we
determined whether it functioned as an antagonist in
the LiSA. As shown in Figure 3A, 100 µM 4 was able to

890 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 6
Spencer et al.
F igu r e 2. Activity of selected compounds on LXRR-GAL4
chimeric receptors in a cell-based reporter assay. (A) 24(S),25-
Epoxycholesterol (1) and the amides 14-16 were assayed at
a concentration of 10 µM; (-) vehicle control. Data are the
mean of 3 independent experiments ( standard error. (B)
Dose-response analysis of 24(S),25-epoxycholesterol (1; b) and
cholenic acid dimethylamide (14; 0). Data are representative
of 2 independent experiments.
F igu r e 3. Antagonist activity of 22(S)-hydroxycholesterol (4).
The concentrations of 1 and 4 are indicated; (-) vehicle control.
(A) LXRR LiSA. Data are the mean of 3 independent experi-
ments ( standard error. (B) LXRR-GAL4 cell-based assay.
Data are the mean of 8 independent experiments ( standard
error.
partially block the recruitment of the SRC1 peptide
induced by 500 nM 24(S),25-epoxycholesterol (1). To
confirm these results, 4 was tested in the LXRR-GAL4
cell-based assay (Figure 3B). A 50 µM dose of 4 did not
induce the reporter gene but was able to block the
activity of 1. Thus, 22(S)-hydroxycholesterol (4) is a
functional antagonist in the LXRR LiSA and LXRR-
GAL4 assay.
To test whether Trp443 in LXRR could play a role in
activation of the receptor by oxysterols, it was changed
to Phe, Tyr, Leu, Cys, Ala, or Thr by site-directed
mutagenesis of the LXRR-GAL4 expression plasmid.
The resulting wild-type and mutant LXRR-GAL4 chi-
meras were transiently expressed in CV-1 cells. Western
blot analysis of the transfected cell extracts, using an
antibody to GAL4, showed that all of the receptors were
expressed at comparable levels (Figure 4B). Therefore,
point mutation of Trp443 did not appear to affect the
stability or expression level of the protein. The activa-
tion of the wild-type and mutant receptors was exam-
ined using the GAL4-responsive reporter. As shown in
Figure 4C, the wild-type LXRR-GAL4 chimera activated
the reporter gene by 15-20-fold in the presence of 10
µM 24(S),25-epoxycholesterol (1). By contrast, none of
the mutant LXRR-GAL4 chimeras were activated by (1).
These results demonstrate that Trp443 plays a critical
role in ligand-mediated activation of LXRR and suggests
that this residue may be the hydrogen bond donor that
interacts with oxysterol ligands.
Con str u ction of a Hom ology Mod el of LXRr. A
model of the LXRR ligand-binding domain was built
from the crystal structure of the retinoic acid receptor
γ (RARγ)44 (Figure 5). 24(S),25-Epoxycholesterol (1) was
docked manually in the ligand-binding site in an
orientation which was consistent with the SAR gener-
ated in the LXRR LiSA and the results of the pharma-
cophore analysis. In the resulting model (Figure 5A,B),
the side chain of 24(S),25-epoxycholesterol (1) adopts a
Tr p 473 Is Requ ir ed for LXRr Activa tion . The data
from the LiSA and cell-based assays are consistent with
a minimum pharmacophore in which the oxysterol side
chain acts as a hydrogen bond acceptor from the LXRR
protein. The C-termini of all ligand-activated nuclear
receptors contain a region known as the activation
function 2 (AF2) that is required for interaction with
SRC1. The X-ray crystal structures of nuclear receptor
ligand-binding domains have shown that AF2 forms an
R-helix.41 With the aid of these structures we generated
a sequence alignment of several nuclear receptors,
including LXRR, LXRâ, and FXR (Figure 4A). The
alignment shows a conserved Glu in the center of the
AF2 helix for each receptor. The PPARγ crystal struc-
ture42,43 revealed that this amino acid (Glu471) forms half
of a “charge clamp” that directs the interaction of SRC1
with PPARγ. The crystal structure also revealed that
Tyr473 in the PPARγ AF2 helix forms a hydrogen bond
with the agonist rosiglitazone, which serves to stabilize
the charge clamp interaction between the receptor and
SRC1.42,43 Notably, our sequence alignment (Figure 4A)
suggests that LXRR, LXRâ, and FXR contain a Trp at
the analogous position in the AF2 helix.

Pharmacophore Analysis of LXRR
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 6 891
groups of the active 22(R)-hydroxycholesterol (5), 23(S)-
hydroxycholesterol (6), and 24(S)-hydroxycholesterol (8)
would all lie between 2.0-2.5 Å from Trp443 (Figure 5B).
Only the activity of 20(S)-hydroxycholesterol (3) is not
explained by the current model. The hydroxyl group of
3 would lie 4 Å from Trp443, which requires that either
the ligand or protein be shifted slightly to form a strong
hydrogen bond. Notably, the relative lack of stereose-
lectivity with C24-substituted sterols (1 vs 2 and 8 vs
9) is explained by this model, since both pairs of
diastereomers place a hydrogen bond acceptor within
2.5 Å of Trp443 (Figure 5B). The model also provides a
possible explanation for the LXRR antagonist activity
of 22(S)-hydroxycholesterol (4). As shown in Figure 5C,
the C22 hydroxyl group of 4 can accept a hydrogen bond
from His421, which lies on helix 10 of the ligand-binding
domain. This interaction could allow 4 to bind to the
receptor without stabilizing the AF2 helix. Thus, the
homology model is generally consistent with the SAR
generated in the LiSA by both LXRR agonists and an
antagonist.
Discu ssion
The orphan nuclear receptor LXRR is a potential new
target for the development of drugs for the treatment
of cardiovascular disease.2,20 In the absence of a crystal
structure for this nuclear receptor, we opted to develop
a homology model of the ligand-binding domain to aid
the design of LXRR agonists. Use of the cell-free LiSA
has allowed us to extend the SARs for oxysterol activa-
tion of LXRR that had been generated using cell-based
reporter assays.5-7 The cell-free LiSA has several
advantages over other techniques for studying the
structural requirements for activation of nuclear recep-
tors. Compared to a competition binding assay,21 the
LiSA assay does not require the synthesis of a radio-
ligand. The LiSA also generates information about the
relative efficacy of a test ligand, which cannot be derived
from a binding assay. Notably, the LiSA can be run in
a 384-well format, which is particularly useful in high-
throughput screening for synthetic nonsteroidal ligands.
Cell-based reporter assays5-7 can be affected by the
ability of the ligand to cross cell membranes as well the
metabolism of the ligand within the cell. This may
explain the lower potency of oxysterols in cell-based
compared to cell-free assays. The data from cell-based
assays can also vary due the nature of the host cell and
the structure of the reporter gene. For example, our data
indicate that cholenic acid dimethylamide (14) is more
efficacious than 24(S),25-epoxycholesterol (1) in the cell-
free LiSA (Figure 1B) and in CV-1 cells employing a
LXRR-GAL4 chimeric receptor (Figure 2). A recent
report also described dimethylamide 14 as one of the
most potent and efficacious activators, compared to
natural and unnatural cholanoic acids, of the full-length
LXRR/RXRR heterodimer in HEK-293 cells using a c-fos
gene reporter.45 By contrast, it was reported that
dimethylamide 14 was less efficacious than 24(S),25-
epoxycholesterol (1) in CV-1 cells employing full-length
LXRR and a CYP7A gene reporter.21 The different
results generated with various cell-based assays under-
score the value of the LXRR LiSA as an independent
method for studying the structural requirements for
receptor activation.
F igu r e 4. Pharmacophore analysis of LXRR. (A) Sequence
alignment of the C-termini of nuclear receptor ligand-binding
domains. The AF2 helix and part of helix 10 are boxed; (b)
indicates the C-terminus of a receptor. Within the AF2 helix,
conserved acidic amino acids are highlighted in red and
hydrophobic amino acids in green. The black arrow indicates
the position of the polar amino acids (highlighted in yellow)
that are proposed to bind to agonist ligands in the NR1
subfamily. (B) Western blot analysis of extracts of CV-1 cell
transfected with expression plasmids containing wild-type and
mutant LXRR-GAL4 receptors. The black arrow indicates the
LXRR protein. (C) Activation of wild-type and mutant LXRR-
GAL4 chimeric receptors by 24(S),25-epoxycholesterol (1) in
CV-1 cells. Data are the mean of 3 independent experiments
( standard error.
low-energy extended conformation that permits a hy-
drogen bond interaction between Trp443 and the epoxide
oxygen. The model also identifies Arg305 as the amino
acid that may interact with the C3 hydroxyl group of
the sterol A-ring at the other end of the ligand-binding
pocket. The orientation of the sterol side chain is
consistent with the diastereoselectivity observed with
the C22- and C23-substituted sterols 4-7. The hydroxyl

892 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 6
Spencer et al.
F igu r e 5. Homology model of the LXRR ligand-binding domain. (A) Protein backbone is shown as a yellow worm, with the
amino acids Trp443 and Arg305 highlighted. 24(S),25-Epoxycholesterol (1) is shown as a van der Waals space-filling representation
with carbon atoms colored green and oxygen atoms red. The C-terminal AF2 helix is labeled. (B) View of the ligand-binding
pocket showing the proposed hydrogen bonds between 24(S),25-epoxycholesterol (1) and Trp443 and Arg305. The hydrogens of C20-
C24 are highlighted to show that one face of the sterol side chain is pointed toward Trp443. (C) View of the ligand-binding pocket
showing the proposed hydrogen bonds between 22(S)-hydroxycholesterol (4) and His421 and Arg305
.
On the basis of the SARs from the LiSA, we have
developed a homology model of the LXRR ligand-binding
domain that incorporates the critical role Trp443 plays
in activation of the receptor. Since the other members
of the NR1 subfamily, LXRâ and FXR, also contain a
Trp at the analogous position in AF2 (Figure 4A), it will
be intriguing to see if they use a similar mechanism of
activation by their cognate ligands. Trp443 in LXRR lies
in an analogous position to Tyr473 of PPARγ (Figure 4A).
Like LXRR, PPARγ binds to DNA as a heterodimer with
the 9-cis-retinoic acid receptor RXR.4 Notably, the
PPARγ/RXR and LXRR/RXR heterodimers can each be
activated by the binding of 9-cis-retinoic acid to the RXR
side of the dimer.46,47 The molecular basis of this
phenomenon, known as “permissive activation”, was
recently explained by analysis of the X-ray crystal
structure of the PPARγ/RXRR heterodimer.43 A salt
bridge, which is formed between Lys431 of RXRR and
the C-terminal carboxylate of PPARγ, provides a mech-
anism by which 9-cis-retinoic acid can promote recruit-
ment of SRC1 to its heterodimeric partner. The se-
quence alignment in Figure 4A shows that the C-
terminus of LXRR also lies at the end of its AF2 helix.
Thus, the LXRR AF2 helix, like PPARγ, may be held in
a quasi-active state by a salt bridge with RXR. Our
model (Figure 5B) proposes that agonists bind to LXRR
in an orientation that generates a hydrogen bond
between the ligand and Trp443. This hydrogen bond
provides a mechanism by which oxysterols can further
stabilize the AF2 helix and increase the level of SRC1
recruitment to the LXRR/RXR heterodimer.
22(S)-Hydroxycholesterol (4) failed to recruit SRC1
in the LiSA but was able to block the activity of
24(S),25-epoxycholesterol (1) (Figure 3A). Oxysterol 4
also blocked the activity of 1 in the cell-based assay
employing the LXRR-GAL4 chimeric receptor (Figure
3B). Notably, this oxysterol has the unnatural stereo-
chemistry at C22, and there is no evidence that it is
formed in vivo. Consistent with the antagonist activity
of 4 in these assays, our model (Figure 5C) shows that
the hydroxyl group of 4 is on the wrong face of the sterol
side chain to form a hydrogen bond with Trp443. Instead,
4 may form a hydrogen bond with His421 that results in
ligand binding without promoting recruitment of SRC1.

Pharmacophore Analysis of LXRR
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 6 893
Epoxycholesterol (1),25 24(R),25-epoxycholesterol (2),25 24(S)-
hydroxycholesterol (8),14 24(R)-hydroxycholesterol (9),14 27-
hydroxycholesterol (11),31 and 23(R)-hydroxymethyl-26,27-
dinorergost-5-en-3â-ol (12)32 were prepared as previously
described.
Syn th esis. Gen er a l. NMR spectra were run in CDCl3,
unless otherwise indicated, on a 300-MHz Varian spectrom-
eter. The chemical shifts are reported in units of δ. Melting
points (mp) were determined using a Thomas-Hoover ap-
paratus in capillary tubes and are uncorrected. Thin-layer
chromatography (TLC) was carried out on EM plastic sheets
precoated with silica gel 60 F-254 (Whatman). Visualization
was obtained by exposure to 5% phosphomolybdic acid in
2-propanol. Flash chromatography was carried out on EM
Reagent silica gel 60 (230-400 mesh), unless otherwise noted.
MgSO4 was used to dry all organic layers.
Interestingly, a modulator of PPARγ (GW0072) has been
identified that binds to this receptor without interacting
with Tyr473 and fails to recruit SRC1.48 GW0072 is an
antagonist of the PPARγ-GAL4 chimera and of adipo-
cyte differention but displays tissue-specific PPARγ
agonist activity in animal models of type 2 diabetes.48,49
Since oxysterol 4 binds to LXRR without stabilizing AF2
or recruiting SRC1, it may also profile as a tissue-
specific modulator of receptor activity. Unfortunately,
the physical properties of this oxysterol have not allowed
us to test this hypothesis in vivo. However, the observa-
tion that the structure and the position of the hydrogen
bond acceptor affect SRC1 recruitment and receptor
activation augurs well for the use of rational design to
develop drugs with modified LXRR agonist activity.
Our model of the LXRR ligand-binding domain re-
quires a functional group on one face of the cholesterol
side chain to accept a hydrogen bond from Trp443 to
effect activation of the receptor (Figure 5B). The LiSA
demonstrated that placement of a carbonyl group at C24
was sufficient to activate the receptor. For example, the
ketone 13 was a full agonist, and the dimethylamide
14 was even more efficacious. This suggests that place-
ment of a hydroxyl group on the sterol side chain, which
can both donate and accept a hydrogen bond, offers no
additional advantage for activation of LXRR. Notably,
24(S),25-epoxycholesterol (1), which has only a hydrogen
bond acceptor in the sterol side chain, is among the most
effective oxysterols at promoting recruitment of the
SRC1 peptide (Figure 1A). Thus, the requirement for
an agonist to accept, but not donate, a hydrogen bond
is consistent with the structure of the efficacious natural
(e.g. 1) and synthetic (e.g. 14) LXRR ligands. As men-
tioned earlier, 1 is found in the liver at sufficient
concentration to activate LXRR12-14 and (along with 8)
increases in concentration in the nuclei of rat liver cells
after administration of an atherogenic diet.13,14 Of
possible relevance also is the fact that, among all the
oxysterols, 24(S),25-epoxycholesterol (1) is formed by a
unique biosynthetic pathway involving the squalene
dioxide shunt.13 In concert with our pharmacophore
analysis, these observations suggest that LXRR may
have evolved as an epoxycholesterol receptor5 rather
than a generalized oxysterol sensor.6
CH2Cl2 was distilled from calcium hydride. Tetrahydrofuran
and ether were distilled from sodium/benzophenone. Pyridine
was distilled from calcium hydride onto 3 Å molecular sieves.
Benzene was dried over 4 Å molecular sieves for 6 h and then
distilled onto 4 Å molecular sieves. All reagents, unless
otherwise noted, were obtained from Aldrich Chemical Co.
6â-Meth oxy-3r,5-cyclo-23-ca r boxyeth yl-5r-ch olest-24-
on e (22). To a solution of 23 mg (222 mmol) of sodium in 10
mL of absolute ethanol was added dropwise 35 mg (222 mmol)
of ethyl isobutyrylacetate with stirring at room temperature
under nitrogen. This mixture was then added dropwise to a
suspension of 78 mg (171 mmol) of 21, prepared by the method
of Partridge26 as modified previously,25 in 10 mL of absolute
ethanol at 45 °C and the mixture was heated at reflux (oil bath
temperature 78-80 °C) with stirring for 24 h (21 only becomes
soluble at reflux). The ethanol was evaporated and the residue
was dissolved in 25 mL of ether and washed with 25 mL of
10% HCl solution, 25 mL of saturated NaHCO3 solution, and
25 mL of brine. The ether layer was dried, filtered, and
evaporated to afford 99 mg of residue that was chromato-
graphed (1:19 EtOAc/hexanes) to yield 36 mg (43%) of yellow
oily 22: 1H NMR 4.21-4.09 (m, 2H), 3.72-3.66 (m, 1H), 3.30
(s, 3H), 2.80-2.74 (m, 1H), 1.04-1.02 (d, J ) 7.2 Hz, 6H),
1.04-1.02 (d, J ) 7.2 Hz, 3H), 0.95 (s, 3H), 0.62 (s, 3H); EI-
HRMS (M+) calcd for C31H50O4 486.3709, found 486.3709.
3â-Hyd r oxych olest-5-en -24-on e (13). According to a pro-
cedure by Greene33 a mixture of 36 mg (74 mmol) of 22 and
600 mg of Al2O3 in 2 mL of 1% aqueous dioxane was heated at
reflux (oil bath temperature 100-105 °C) with stirring under
nitrogen for 96 h. Vacuum filtration through a pad of Celite
followed by evaporation of the solvent afforded 30 mg of crude
decarboxylated i-steroid, which was mixed with 1.4 mg (9.00
mmol) of p-toluenesulfonic acid monohydrate in 1 mL of 25%
aqueous dioxane and heated (oil bath temperature 78-80 °C)
with stirring for 5 h. The mixture was cooled to room
temperature, quenched with 10 mL of saturated NaHCO3
solution, and extracted with 3 × 10 mL of ether. The ether
was dried, filtered, and evaporated to afford 20 mg of crude
residue that was chromatographed (1:4 EtOAc/hexanes) to
yield 10 mg (34%) of colorless 13: mp 134-137 °C (lit.30,50 mp
137-138 °C); 1H NMR 5.28 (m, 1H), 3.52-3.39 (m, 1H), 1.04-
1.02 (d, J ) 7.2 Hz, 6H), 1.04-1.02 (d, J ) 7.2 Hz, 3H), 0.95
(s, 3H), 0.62 (s, 3H); 13C NMR (CDCl3) 215.8, 141.0, 121.7, 71.8,
56.9, 56.0, 50.2, 42.5, 42.4, 41.0, 39.9, 37.4, 36.7, 35.6, 32.0,
31.7, 30.0, 29.9, 28.3, 24.4, 21.2, 19.6, 18.7, 18.5, 18.5, 12.0,
1.2.
Con clu sion
We have developed a homology model for the nuclear
oxysterol receptor LXRR. The model incorporates the
structural requirements for activation of the receptor
by oxysterols and analogues, and the critical role that
Trp443 plays in this process. The model explains the
LXRR agonist activity of 24(S),25-epoxycholesterol (1)
as well as the antagonist activity of 22(S)-hydroxycho-
lesterol (4). Since LXRR plays an important role in the
regulation of cholesterol homeostasis, this model may
be useful in the design of new drugs for the treatment
of cardiovascular disease.2,20 The recent disclosure of a
nonsteroid LXRR agonist18 provides a template for
development of these analogues.
23(S)- a n d 23(R)-Ch olest-5-en -3â,23-d iol (6 a n d 7). Ac-
cording to a procedure by Bailey27 to a solution of 106 mg (0.23
mmol) of 21 in 3 mL of dry ether was added dropwise 0.37
mL (0.50 mmol) of 1.36 M t-BuLi in pentane at -78 °C with
stirring under nitrogen. After stirring for 5 min, the mixture
was allowed to warm slowly and stand at room temperature
over 1 h. The mixture was cooled to -78 °C and treated
dropwise with 40 mg (0.46 mmol) of isovaleraldehyde, stirred
for 3 h, slowly warmed to room temperature, and stirred for
12 h. The mixture was quenched with saturated NH4Cl
solution, extracted with 2 × 10 mL of ether, and the ether was
washed with 20 mL of NaHCO3, 20 mL of brine, dried, filtered,
Exp er im en ta l Section
Oxyster ols a n d Rela ted Com p ou n d s. 22(S)-Hydroxy-
cholesterol (4), 22(R)-hydroxycholesterol (5), 25-hydroxycho-
lesterol (10), cholenic acid (18), cholenic acid methyl ester (19),
and cholesterol (20) were purchased from Sigma. 24(S)-

894 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 6
Spencer et al.
and evaporated to give 200 mg of crude residue. According to
a procedure by Partridge,26 this residue (200 mg) was combined
with 13 mg (0.068 mmol) of p-toluenesulfonic acid in 4 mL of
25% aqueous dioxane and heated (oil bath temperature 79-
80 °C) for 5 h. The mixture was cooled to room temperature,
quenched with 10 mL of saturated NaHCO3 solution, and
extracted with 3 × 25 mL of CH2Cl2. The CH2Cl2 was dried,
filtered, and evaporated to afford 160 mg of crude residue that
was chromatographed (gradient: 1:9, 1:4, 1:2 EtOAc/hexanes)
to yield 44 mg (48% from 21) of a ca. 1:1 epimeric mixture of
6 and 7. Pure samples were obtained by HPLC on a 5-µm
chiral phase Bakerbond DNBPG column (4.6 × 25 mm,
Beckman) using 96:4 hexanes:2-propanol. 23(R)-Hydroxycho-
lesterol (more polar29) (7): mp 168.1-168.7 (lit.29 mp 163-
166 °C); 1H NMR 5.32 (m, 1H), 3.79-3.70 (m, 1H), 3.56-3.45
(m, 1H), 0.99 (s, 3H), 0.92 (d, J ) 7.5 Hz, 6H), 0.67 (s, 3H).
23(S)-Hydroxycholesterol (6): mp 160.8-161.4 (lit.29 mp 160-
162 °C); 1H NMR (CDCl3) 5.32 (m, 1H), 3.81-3.70 (m, 1H),
3.56-3.44 (m, 1H), 0.99 (s, 3H), 0.89 (d, J ) 6.6 Hz, 6H), 0.69
(s, 3H).
mixture was stirred at -78 °C overnight, then at room
temperature for 2 h, diluted with 20 mL of H2O, and extracted
with 2 × 20 mL of Et2O, and 3 × 20 mL of CH2Cl2. The
combined organic layers were dried, filtered and evaporated
to give 124 mg of solid which was chromatographed (3:1
EtOAc:hexane), then recrystallized from acetone to give 63.8
mg (63% from 18) of 15: mp 223-224 °C; IR νmax 3312, 1642,
1
1555 cm-1; H NMR 5.37 (m, 2H), 3.53 (m, 1H), 2.81 (d, J )
4.8 Hz, 3H), 1.01 (s, 3H), 0.93 (d, J ) 6.3 Hz, 3H), 0.69 (s,
3H); 13C NMR (CD3OD) 177.5, 142.4, 12.6, 72.6, 58.3, 57.4,
51.9, 43.7, 43.2, 41.3, 38.7, 37.8, 37.0, 34.1, 33.5, 33.4, 33.2,
32.4, 29.3, 26.5, 25.5, 22.3, 20.0, 19.0, 12.5; HRMS (MH+) calcd
for C25H42NO2 388.3216, found 388.3204. Anal. (C25H41NO2)
C, H, N.
22-Azid o -6â-m e t h o x y -3r,5-c y c lo -5r-23,24-b is n o r -
ch ola n e (24). According to a modification of a procedure by
Koziara,54 to a solution of 201 mg (0.441 mmol) of 21 in 2 mL
of DMF and 2 mL of benzene was added 74.8 mg (1.15 mmol)
of NaN3. The mixture was heated at reflux overnight under
N2, diluted with 10 mL of H2O, and extracted with 3 × 10 mL
of benzene. The combined organic layers were dried, filtered,
and evaporated to give 213.5 mg of a colorless oil, which was
chromatographed (3:97 EtOAc:hexane) to give 155.4 mg (95%)
of oily 24: IR νmax 2096 cm-1; 1H NMR 3.35 (dd, J ) 3.3, 11.7
Hz, 1H), 3.31 (m, 3H), 3.03 (dd, J ) 7.2 Hz, J ) 11.7 Hz, 1H),
2.76 (m, 1H), 1.03 (d, J ) 6.6 Hz, 3H), 1.01 (s, 3H), 0.71 (s,
3H); 13C NMR 82.6, 58.3, 56.8, 56.5, 53.6, 48.2, 43.6, 43.2, 40.3,
37.2, 35.5, 33.6, 30.8, 28.2, 25.2, 24.5, 23.0, 21.7, 19.5, 18.0,
13.4, 12.6; Anal. (C23H37N3O) C, H, N.
22-Am in o -6â-m e t h o x y -3r,5-c y c lo -5r-23,24-b is n o r -
ch ola n e (25). According to a procedure by Bose,55 to a solution
of 119 mg (0.320 mmol) of azide 24 in 5 mL of dry ether was
added 86.7 mg (2.29 mmol) of LiAlH4. The mixture was heated
at reflux for 3 h under N2, diluted with 10 mL of wet ether
and 10 mL of H2O at 0 °C, and extracted with 3 × 20 mL of
ether. The combined organic layers were dried, filtered, and
evaporated to give 122 mg of yellow oil which was chromato-
graphed (88:12:2 EtOAc:MeOH:concd NH4OH) to give 86.8 mg
(79%) of colorless oily 25: IR νmax 3384 cm-1; 1H NMR 3.28 (s,
3H), 2.73 (m, 1H), 2.70 (m, 1H), 2.36 (dd, J ) 12.6 Hz, 7.2 Hz,
1H), 0.98 (s, 3H), 0.94 (d, J ) 6.6 Hz, 3H), 0.69 (s, 3H); 13C
NMR 82.5, 56.7, 56.5, 53.8, 48.1, 47.7, 43.5, 42.9, 40.3, 39.1,
35.4, 35.2, 33.5, 30.6, 28.0, 25.1, 24.4, 22.9, 21.6, 19.4, 17.0,
13.2, 12.5; EI-HRMS (M+) calcd for C23H39NO 417.3032, found
417.3030.
22-[N-(2′-Meth ylp r op ion yl)a m in o]-23,24-bisn or ch ol-5-
en -3â-ol (16). To a solution of 39.7 mg (0.115 mmol) of 25 in
4 mL of dry ether were added dropwise 0.3 mL (3.75 mmol) of
dry pyridine and 0.8 mL (7.6 mmol) of isobutyryl chloride in
-78 °C under N2. The mixture was stirred at -78 °C for 1 h,
then at room temperature overnight, diluted with 10 mL of
H2O, and extracted with 3 × 20 mL of ether. The combined
organic layers were washed with 10 mL of brine, dried, filtered,
and evaporated to give 53.0 mg of an oil, to which was added
23.7 mg (0.125 mmol) of p-toluenesulfonic acid monohydrate
in 3 mL of dioxane and 1 mL of H2O. The resulting mixture
was heated at reflux for 5 h, diluted with 10 mL of H2O, and
extracted with 3 × 20 mL of CH2Cl2. The combined organic
layers were washed with 10 mL of saturated NaHCO3, 10 mL
of H2O, dried, filtered, and evaporated to give 47.0 mg of
residue which was chromatographed (3:2 EtOAc:hexane) to
give 21.1 mg (44% from 25) of 16: mp 180.5-181.5 °C; 1H
NMR 5.46 (m, 1H), 5.34 (m, 1H), 3.52 (m, 1H), 3.34 (m, 1H),
2.95 (m, 1H), 2.30 (m, 3H), 1.15 (d, J ) 6.6 Hz, 6H), 1.01 (s,
3H), 0.96 (d, J ) 6.6 Hz, 3H), 0.69 (s, 3H); 13C NMR (CD3OD)
177.2, 140.1, 121.8, 71.9, 56.7, 54.3, 50.3, 45.0, 42.7, 42.5, 39.8,
37.5, 36.8, 36.7, 36.1, 32.1, 32.0, 31.8, 28.1, 24.6, 21.2, 20.0,
19.9, 19.6, 17.6, 12.1; FAB-HRMS (MH+) calcd for C26H44NO2
402.3372, found 402.3379. Anal. (C26H43NO2) C, H, N.
N ,N -Dim e t h yl-6â-m e t h oxy-3r,5-cyclo-5r-ch ola n a m -
id e (23). To a solution of 0.15 mL (1.07 mmol) of diisopropyl-
amine in 1 mL of dry THF at -78 °C was added 0.55 mL (1.10
mmol) of 2.0 M n-BuLi under N2. After the solution was stirred
at -78 °C for 30 min, a solution of 98.3 mg (1.13 mmol) of
N,N-dimethylacetamide in 1 mL of THF was added. The
resulting mixture was stirred at -78 °C for 20 min, then at 0
°C for 20 min, cooled to -78 °C and a solution of 201 mg (0.440
mmol) of 21 in 2 mL of THF was added dropwise. The reaction
mixture was stirred at -78 °C for 6 h, then at 0 °C overnight,
diluted with 25 mL of saturated NH4Cl, and extracted with 3
× 15 mL of CH2Cl2. The combined organic layers were washed
with 10 mL of brine, dried, filtered, and evaporated to give
225 mg of residue which was chromatographed (1:4 EtOAc:
hexane) to give 170.2 (93%) of oily 23: 1H NMR 3.28 (s, 3H),
2.97 (s, 3H), 2.89 (s, 3H), 2.73 (t, J ) 2.7 Hz, 1H), 0.98 (s, 3H),
0.90 (d, J ) 6.3 Hz, 3H), 0.68 (s, 3H); 13C NMR 173.8, 82.6,
56.7, 56.3, 48.2, 43.6, 43.0, 40.5, 37.5, 35.9, 35.6, 35.5, 35.2,
33.6, 31.4, 30.7, 30.6, 28.5, 25.2, 24.4, 23.0, 21.7, 19.5, 18.7,
13.3, 12.5; EI-HMRS (M+) calcd for C27H45NO2 415.3450, found
415.3445.
N,N-Dim eth yl-3â-h yd r oxych olen a m id e (14). To a solu-
tion of 132.7 mg (0.320 mmol) of 23 in 3.5 mL of dioxane and
1.0 mL of H2O was added 29.5 mg of p-toluenesulfonic acid
monohydrate. The resulting mixture was heated at reflux for
5 h, diluted with 10 mL of H2O, and extracted with 3 × 20 mL
of CH2Cl2. The combined organic layers were washed with 10
mL of saturated NaHCO3, 10 mL of H2O, dried, filtered, and
evaporated to give 140 mg of residue which was chromato-
graphed (3:17 EtOAc:hexane) to give 86.8 mg (68%) of 14: mp
187.5-189.5 °C (lit.34 mp 189.5-190 °C); 1H NMR 5.34 (m,
1H), 3.52 (m, 1H), 3.00 (s, 3H), 2.94 (s, 3H), 1.00 (s, 3H), 0.94
(d, J ) 6.6 Hz, 3H), 0.68 (s, 3H); 13C NMR 174.0, 141.1, 121.9,
72.0, 57.0, 56.2, 50.4, 42.6, 42.5, 40.0, 37.6, 37.5, 36.8, 36.0,
35.7, 32.2, 31.9, 31.5, 30.7, 28.5, 24.6, 21.4, 19.7, 18.8, 12.2.
N-Meth yl-3â-h yd r oxych olen a m id e (15). According to a
procedure by Okamoto,51 to a solution of 98.3 mg (0.262 mmol)
of 18 (Sigma) in 2 mL of dry DMF were added 136.3 mg (0.899
mmol) of t-BuMe2SiCl, 146.0 mg (2.17 mmol) of imidazole and
71.0 mg (0.585 mmol) of 4-(dimethylamino)pyridine at 0 °C.
The mixture was stirred at room temperature for 6 h under
N2, diluted with 10 mL of H2O, and extracted with 3 × 10 mL
of EtOAc. The combined organic layers were washed with 10
mL of H2O, 10 mL of brine, dried, filtered, and evaporated to
give 144 mg of residue. According to procedures by Firestone52
and Cope,53 to a solution of 140 mg of this material in 2 mL of
dry CH2Cl2 were added slowly 0.2 mL (2 mmol) of (COCl)2 and
1 drop of dry DMF. The resulting mixture was stirred at room
temperature under N2 for 2 h and the solvent was evaporated
to give a yellow solid residue, to which was added ca. 120 mmol
of CH3NH2, generated using the method of Russel32 by dripping
10 mL of 40 wt % aqueous CH3NH2 onto solid KOH pellets
and passing the resultant gas through a KOH drying tube
before being condensed with a coldfinger at -78 °C. The
22-[N-(N,N-Dim eth ylfor m a m id o)a m in o]-23,24-bisn or -
ch ol-5-en -3â-ol (17). To a solution of 67.7 mg (0.196 mmol)
25 in 3 mL of ether and 0.3 mL (4 mmol) of dry pyridine at
-78 °C was added dropwise 0.5 mL (6 mmol) of dimethylcar-
bamyl chloride under N2. The resulting mixture was stirred

Pharmacophore Analysis of LXRR
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 6 895
at -78 °C for 1 h, then at room temperature for 17 h, diluted
with 10 mL of H2O, and extracted with 3 × 20 mL of CH2Cl2.
The combined organic layers were washed with 3 × 10 mL of
10% KOH, 10 mL of brine, dried, filtered, and evaporated to
give 135 mg of red-yellow residue, which was chromatographed
(1:1 EtOAc:hexane) to give 57.4 mg (70%) of oily 26: 1H NMR
4.39 (m, 1H), 3.35 (m, 1H), 3.30 (s, 3H), 2.89 (s, 6H), 2.83 (m,
1H), 2.75 (t, J ) 2.7 Hz, 1H), 1.00 (s, 3H), 0.95 (d, J ) 6.6 Hz,
3H), 0.72 (s, 3H); 13C NMR 158.8, 82.5, 56.7, 56.5, 54.5, 48.2,
46.6, 43.5, 43.1, 40.3, 37.2, 36.4, 35.4, 35.2, 33.5, 30.6, 28.2,
25.1, 24.4, 22.9, 21.7, 19.4, 17.6, 13.2, 12.5; EI-HRMS (M+)
calcd for C26H44N2O2 416.3403, found 416.3406. To 52.1 mg
(0.125 mmol) of this 26 was added 25.5 mg (0.134 mmol) of
p-toluenesulfonic acid monohydrate and 3 mL of dioxane and
1 mL of H2O. The resulting mixture was refluxed overnight,
diluted with 10 mL of H2O, and extracted with 3 × 15 mL of
CH2Cl2. The combined organic layers were washed with 10 mL
of saturated NaHCO3, 10 mL of H2O and 10 mL of brine, dried,
filtered, and evaporated to give 63 mg of residue which was
crystallized twice from CH2Cl2-hexane to give 40.4 mg (80%)
nin (APC, Molecular Probes). At the same time, a biotinylated
peptide comprising amino acids 675-699 of SRC1 (CPSSH-
SSLTERHKILHRLLQEGSPS-CONH2) at a concentration of
20 nM was incubated in assay buffer with an equimolar
amount of streptavidin-labeled europium (Wallac) for 20-25
min. After the initial incubations were completed, a 20 molar
excess (400 nM) of biotin was added to each of the solutions
to block the unattached streptavidin reagents. After 20 min
at room temperature, the solutions were mixed, yielding a
concentration of 10 nM for the dye-labeled LXRR protein and
SRC1 peptide. 49 µL of the protein/peptide mixture was added
to each well of an assay plate containing 1 µL of test compound.
The final volume in each well was 0.05 mL, and the concentra-
tion in the well for the dye-labeled protein and peptide was
10 nM. The final test compound concentrations were between
1 nM and 100 µM. The plates were incubated at room
temperature for 2-4 h and then counted on a Wallac Victor
fluorescent plate reader in a time-resolved mode. The relative
fluorescence was measured at 665 nm.
P la sm id s. Expression vectors were generated by insertion
of cDNA encoding human LXRR (amino acids 162-447) fused
in-frame with the GAL4 DNA binding domain in the pSG5
vector (Stratagene, La J olla, CA). The Trp443 codon (TGG
nucleotide sequence) of LXRR was altered to a Phe (TCC), Tyr
(TAC), Leu (CTG), Cys (TGC), Ala (GCG), or Thr (ACG) codon
by Quick-Change Mutagenesis (Stratagene, La J olla, CA).
Cell-Ba sed Rep or ter Assa y. Cell-based assays for LXRR
activity were performed as previously described5 using CV-1
cells transfected with a chimeric LXRR-GAL4 receptor expres-
sion plasmid and a (UAS)5-tk-SPAP reporter plasmid. Trans-
fections were performed using Lipofectamine (Life Technolo-
gies,Grand Island,NY)accordingtomanufacturer’s instructions.
A â-galactosidase expression plasmid was included in each
transfection for use as a normalization control.
Wester n Blottin g. Equal amounts of total cellular protein
were electrophoresed on 12% SDS-PAGE and transferred to
Trans-Blot nitrocellulose membranes (Bio-Rad, Hercules, CA).
LXRR-GAL4 chimeric proteins were visualized using an ECL
detection system (Amersham Pharmacia Biotech, Piscataway,
NY) after incubation with mouse monoclonal anti-GAL4 DNA
binding domain antibody (Santa Cruz Biotechnology, Santa
Cruz, CA) and a horseradish peroxidase-linked goat anti-
mouse secondary antibody (Southern Biotech Associates,
Birmingham, AL).
Hom ology Mod el. A model for the LXRR ligand-binding
domain was built from the crystal structure of RARγ bound
to all-trans-retinoic acid44 using the MVP program.56 Compared
with RARγ, LXRR has one single-residue insertion, located
between the predicted helices 6 and 7, and two single-residue
deletions, located in the loops leading into helix 3 and the AF2
helix. The insertion and deletions all lie 10 Å or more from
the proposed ligand-binding site, and no special effort was
made to model their effects. The MVP buildup procedure was
used to obtain low-energy conformations for mutated side
chains in the binding site region. This involved a search of 90
side chain torsional coordinates, with torsional coordinate
energy minimization during the buildup process and Cartesian
coordinate energy minimization after the buildup was com-
plete. The all-trans-retinoic acid molecule from RARγ was
included in these calculations to bias side chains away from
the ligand-binding site. 24(S),25-Expoxycholesterol (1) was
manipulated into the LXRR binding site graphically, consider-
ing several alternative conformations of the sterol side chain
and several orientations of the steroid ring system. The initial
conformation of 1 was taken from subunit A of the small-
molecule crystal structure.25
1
of 17: mp 203-204 °C; H NMR 5.35 (m, 1H), 4.38 (m, 1H),
3.52 (m, 1H), 3.38 (m, 1H), 2.91 (s, 6H), 2.85 (m, 1H), 1.00 (s,
3H), 0.97 (d, J ) 6.6 Hz, 3H), 0.70 (s, 3H); 13C NMR 158.8,
141.1, 121.8, 71.9, 56.8, 54.3, 50.3, 46.6, 42.7, 42.5, 39.9, 37.5,
37.2, 36.7, 36.4, 32.1, 31.8, 29.9, 28.1, 24.6, 21.3, 19.6, 17.6,
12.2; EI-HRMS (M+) calcd for C25H42N2O2 402.3246, found
402.3247.
22-[N-(N,N-Dim eth ylfor m am ido)am in o]-3â-ch lor o-23,24-
bisn or ch ol-5-en e (27). To a solution of 99.8 mg (0.289 mmol)
of 25 in 4 mL of ether and 0.4 mL (5 mmol) of dry pyridine at
-78 °C was added dropwise 0.7 mL (8 mmol) of dimethylcar-
bamyl chloride under N2. The resulting mixture was stirred
at -78 °C for 1 h, then at room temperature for 17 h, diluted
with 10 mL of H2O, and extracted with 3 × 20 mL of CH2Cl2.
The combined organic layers were washed with 10 mL of
saturated NaHCO3, 10 mL of brine, dried, filtered, and
evaporated to give 331 mg of red-yellow residue, to which was
added 38.8 mg (0.204 mmol) of p-toluenesulfonic acid mono-
hydrate, 3 mL of dioxane, and 1 mL of H2O. The resulting
mixture was refluxed overnight, diluted with 10 mL of H2O,
and extracted with 3 × 15 mL of CH2Cl2. The combined organic
layers were washed with 10 mL of saturated NaHCO3, 10 mL
of H2O and 10 mL of brine, dried, filtered, and evaporated to
give 125 mg of residue which was chromatographed (3:1
EtOAc:hexane) to give 25.4 mg (22%) of 17 and 45.4 mg (37%)
1
of 27: mp 229-231 °C; H NMR 5.37 (m, 1H), 4.38 (m, 1H),
3.76 (m, 1H), 3.39 (ddd, J ) 3.6 Hz, 5.4 Hz, 13.2 Hz, 1H), 2.91
(s, 6H), 2.85 (m, 1H), 1.03 (s, 3H), 0.97 (d, J ) 6.6 Hz, 3H),
0.70 (s, 3H); 13C NMR 158.8, 140.9, 122.6, 60.5, 56.6, 54.3, 50.2,
46.6, 43.6, 42.7, 39.7, 39.3, 37.2, 36.5, 36.4, 33.5, 32.0, 31.9,
28.1, 24.5, 21.1, 19.4, 17.6, 12.1. Anal. (C25H41ClN2O) C, H, N.
LiSA. A modified polyhistidine tag (MKKGHHHHHHG)
was fused in frame of the human LXRR ligand-binding domain
(amino acids 183-447 of GenBank accession number U22662,
with the 14th amino acid corrected to A from R). The LXRR
fusion protein was expressed in E. coli and purified as
previously described.8,21 The purified protein was diluted to
approximately 10 µM in PBS and a 5-fold molar excess of NHS-
LC-Biotin (Pierce) was added in a minimal volume of PBS.
This solution was incubated with gentle mixing for 30 min at
ambient room temperature. The biotinylation reaction was
stopped by the addition of 2000-fold molar excess of Tris-HCl,
pH 8. The modified LXRR protein was dialyzed against 4 buffer
changes, each of at least 50 volumes, with PBS containing 5
mM DTT, 2 mM EDTA and 2% sucrose. The biotinylated LXRR
protein was subjected to mass spectrometric analysis to reveal
the extent of modification by the biotinylation reagent. In
general, approximately 95% of the protein had at least a single
site of biotinylation; the overall extent of biotinylation followed
a normal distribution of multiple sites, ranging from 1 to 9.
Ack n ow led gm en t. The research at Dartmouth Col-
lege was supported by NIH Grant HL52069.
Refer en ces
The biotinylated protein was incubated for 20-25 min at a
concentration of 20 nM in assay buffer (50 mM NaF, 50 mM
MOPS, pH 7.5, 0.1 mM CHAPS, 0.1 mg/mL FAF-BSA, 10 mM
DTT) with equimolar amounts of streptavidin-AlloPhycoCya-
(1) Mangelsdorf, D. J .; Thummel, C.; Beato, M.; Herrlich, P.;
Schuetz, G.; Umesono, K.; Blumberg, B.; Kastner, P.; Mark, M.;
Chambon, P.; Evans, R. M. The nuclear receptor superfamily:
The second decade. Cell 1995, 83, 835-839.

896 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 6
Spencer et al.
(2) Kliewer, S. A.; Lehmann, J . M.; Willson, T. M. Orphan nuclear
receptors: shifting endocrinology into reverse. Science 1999, 284,
757-760.
(3) Nuclear Receptors Committee. A unified nomenclature system
for the nuclear receptor superfamily. Cell 1999, 97, 161-163.
(4) Mangelsdorf, D. J .; Evans, R. M. The RXR heterodimers and
orphan receptors. Cell 1995, 83, 841-850.
(29) Hirano, Y.; Eguchi, T.; Ishiguro, M.; Ikekawa, N. Configuration
at the C-23 position of 23-hydroxy- and 23,25-dihydroxycholes-
terols. Chem. Pharm. Bull. 1983, 31, 394-400.
(30) Koch, P.; Nakatani, Y.; Luu, B.; Ourisson, G. A stereoselective
synthesis and a convenient synthesis of optically pure (24R)- and
(24S)-24-hydroxycholesterols. Bull. Soc. Chim. Fr. 1983, 189-
194.
(5) Lehmann, J . M.; Kliewer, S. A.; Moore, L. B.; Smith-Oliver, T.
A.; Blanchard, D. E.; Spencer, T. A.; Willson, T. M. Activation
of the nuclear receptor LXR by oxysterols defines a new hormone
response pathway. J . Biol. Chem. 1997, 272, 3137-3140.
(6) J anowski, B. A.; Willy, P. J .; Devi, T. R.; Falck, J . R.; Mangels-
dorf, D. J . An oxysterol signaling pathway mediated by the
nuclear receptor LXRR. Nature 1996, 383, 728-731.
(7) Forman, B. M.; Ruan, B.; Chen, J .; Schroepfer, G. J ., J r.; Evans,
R. M. The orphan nuclear receptor LXRR is positively and
negatively regulated by distinct products of mevalonate metabo-
lism. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 10588-10593.
(8) Parks, D. J .; Blanchard, S. G.; Bledsoe, R. K.; Chandra, G.;
Consler, T. G.; Kliewer, S. A.; Stimmel, J . B.; Willson, T. M.;
Zavacki, A. M.; Moore, D. D.; Lehmann, J . M. Bile acids: Natural
ligands for an orphan nuclear receptor. Science 1999, 284, 1365-
1368.
(9) Makishima, M.; Okamoto, A. Y.; Repa, J . J .; Tu, H.; Learned,
R. M.; Luk, A.; Hull, M. V.; Lustig, K. D.; Mangelsdorf, D. J .;
Shanz, B. Identification of a nuclear receptor for bile acids.
Science 1999, 284, 1362-1365.
(10) Wang, H.; Chen, J .; Hollister, K.; Sowers, L. C.; Forman, B. M.
Endogenous bile acids are ligands for the nuclear receptor FXR/
BAR. Mol. Cell 1999, 3, 543-553.
(11) Peet, D. J .; Turley, S. D.; Ma, W.; J anowski, B. A.; Lobaccaro,
J .-M. A.; Hammer, R. E.; Mangelsdorf, D. J . Cholesterol and bile
acid metabolism are impaired in mice lacking the nuclear
oxysterol receptor LXRR. Cell 1998, 93, 693-704.
(12) Spencer, T. A.; Gayen, A. K.; Phirwa, S.; Nelson, J . A.; Taylor,
F. R.; Kandutsch, A. A.; Erickson, S. K. 24(S),25-Epoxycholes-
terol. Evidence consistent with a role in the regulation of hepatic
cholesterogenesis. J . Biol. Chem. 1985, 260, 13391-13394.
(13) Spencer, T. A. The squalene dioxide pathway of steroid biosyn-
thesis. Acc. Chem. Res. 1994, 27, 83-90.
(14) Zhang, Z.; Li, D.; Blanchard, D. E.; Lear, S. R.; Erickson, S. K.;
Spencer, T. A. Key regulatory oxysterols in liver: Analysis as
∆4-3-ketone derivatives by high performance liquid chromatog-
raphy and response to physiological perturbations. J . Lipid Res.,
in press.
(15) Luo, Y.; Tall, A. R. Sterol upregulation of human CETP expres-
sion in vitro and in transgenic mice by an LXR element. J . Clin.
Invest. 2000, 105, 513-520.
(31) Kim, H. S.; Wilson, W. K.; Needleman, D. H.; Pinkerton, F. D.;
Wilson, D. K.; Quiocho, F. A.; Schroepfer, G. J ., J r. Inhibitors of
sterol synthesis. Chemical synthesis, structure, and biological
activities of (25R)-3â,26-dihydroxy-5R-cholest-8(14)-en-15-one, a
metabolite of 3â-hydroxy-5R-cholest-8(14)-en-15-one. J . Lipid
Res. 1989, 30, 247-261.
(32) Spencer, T. A.; Li, D.; Russel, J . S.; Tomkinson, N. C. O.; Willson,
T. M. Further studies on the synthesis of 24(S),25-epoxycholes-
terol. A new, efficient preparation of desmosterol. J . Org. Chem.
2000, 65, 1919-1923.
(33) Greene, A. E.; Cruz, A.; Crabbe, P. Decarbalkoxylation of â-keto
esters by a new mild procedure. Tetrahedron Lett. 1976, 2707-
2708.
(34) Louw, D. F.; Strating, J .; Backer, H. J . ∆5-Steroids and provi-
tamins
D with branched side chains, III. Preparation and
reduction of some ∆5-steroid-ω-amides. Recl. Trav. Chim. Pays-
Bas 1954, 73, 667-677.
(35) Corey, E. J .; Grogan, M. J . Stereocontrolled syntheses of
24(S),25-epoxycholesterol and related oxysterols for studies on
the activation of LXR receptors. Tetrahedron Lett. 1998, 39,
9351-9354.
(36) Maloney, P. R.; Parks, D. J .; Haffner, C. D.; Fivush, A. M.;
Chandra, G.; Plunket, K. D.; Creech, K. L.; Moore, L. B.; Wilson,
J . G.; Lewis, M. C.; J ones, S. A.; Willson, T. M. Identification of
a chemical tool for the orphan nuclear receptor FXR. J . Med.
Chem. 2000, 43, 2971-2974.
(37) Brown, M. S.; Goldstein, J . L. The SREBP pathway: Regulation
of cholesterol metabolism by proteolysis of a membrane-bound
transcription factor. Cell 1997, 89, 331-340.
(38) Taylor, F. R.; Kandutsch, A. A.; Gayen, A. K.; Nelson, J . A.;
Nelson, S. S.; Phirwa, S.; Spencer, T. A. 24,25-Epoxysterol
metabolism in cultured mammalian cells and repression of
3-hydroxy-3-methylglutaryl-CoA reductase. J . Biol. Chem. 1986,
261, 15039-15044.
(39) Burstein, S.; Gut, M. Biosynthesis of pregnenolone. Recent Prog.
Horm. Res. 1971, 27, 303-349.
(40) Saucier, S. E.; Kandutsch, A. A.; Gayen, A. K.; Swahn, D. K.;
Spencer, T. A. Oxysterol regulators of 3-hydroxy-3-methylglu-
taryl-CoA reductase in liver. Effect of dietary cholesterol. J . Biol.
Chem. 1989, 264, 6863-6869.
(16) Schwartz, K.; Lawn, R. M.; Wade, D. P. ABC1 Gene expression
and ApoA-I-mediated cholesterol efflux are regulated by LXR.
Biochem. Biophys. Res. Commun. 2000, 274, 794-802.
(17) Costet, P.; Luo, Y.; Wang, N.; Tall, A. R. Sterol-dependent
transactivation of the ABC1 promoter by the liver X receptor/
retinoid X receptor. J . Biol. Chem. 2000, 275, 28240-28245.
(18) Repa, J . J .; Turley, S. D.; Lobaccaro, J . M. A.; Medina, J .; Li, L.;
Lustig, K.; Shan, B.; Heyman, R. A.; Dietschy, J . M.; Mangels-
dorf, D. J . Regulation of absorption and ABC1-mediated efflux
of cholesterol by RXR heterodimers. Science 2000, 289, 1524-
1529.
(19) Venkateswaran, A.; Laffitte, B. A.; J oseph, S. B.; Mak, P. A.;
Wilpitz, D. C.; Edwards, P. A.; Tontonoz, P. Control of cellular
cholesterol efflux by the nuclear oxysterol receptor LXRR. Proc.
Natl. Acad. Sci. U.S.A. 2000, 97, 12097-12102.
(20) Chawla, A.; Saez, E.; Evans, R. M. Don’t know much bile-ology.
Cell 2000, 103, 1-4.
(41) Weatherman, R. V.; Fletterick, R. J .; Scanlan, T. S. Nuclear-
receptor ligands and ligand-binding domains. Annu. Rev. Bio-
chem. 1999, 68, 559-581.
(42) Nolte, R. T.; Wisely, G. B.; Westin, S.; Cobb, J . E.; Lambert, M.
H.; Kurokawa, R.; Rosenfeld, M. G.; Willson, T. M.; Glass, C.
K.; Milburn, M. V. Ligand binding and co-activator assembly of
the peroxisome proliferator-activated receptor-γ. Nature 1998,
395, 137-143.
(43) Gampe, R. T., J r.; Montana, V. G.; Lambert, M. H.; Miller, A.
B.; Bledsoe, R. K.; Milburn, M. V.; Kliewer, S. A.; Willson, T.
M.; Xu, H. E. Asymmetry in the PPARγ/RXRR crystal structure
reveals the molecular basis of heterodimerization among nuclear
receptors. Mol. Cell 2000, 5, 545-555.
(44) Renaud, J .-P.; Rochel, N.; Ruff, M.; Vivat, V.; Chambon, P.;
Gronemeyer, H.; Moras, D. Crystal structure of the RAR-γ
ligand-binding domain bound to all-trans retinoic acid. Nature
1995, 378, 681-689.
(45) Song, C.; Hiipakka, R. A.; Liao, S. Selective activation of liver X
receptor R by 6R-hydroxy bile acids and analogues. Steroids
2000, 65, 423-427.
(46) Kliewer, S. A.; Umesono, K.; Noonan, D. J .; Heyman, R. A.;
Evans, R. M. Convergence of 9-cis retinoic acid and peroxisome
proliferator signaling pathways through heterodimer formation
of their receptors. Nature 1992, 358, 771-774.
(47) Willy, P. J .; Umesono, K.; Ong, E. S.; Evans, R. M.; Heyman, R.
A.; Mangelsdorf, D. J . LXR, a nuclear receptor that defines a
distinct retinoid response pathway. Genes Dev. 1995, 9, 1033-
1045.
(48) Oberfield, J . L.; Collins, J . L.; Holmes, C. P.; Goreham, D. M.;
Cooper, J . P.; Cobb, J . E.; Lenhard, J . M.; Hull-Ryde, E. A.; Mohr,
C. P.; Blanchard, S. G.; Parks, D. J .; Moore, L. B.; Lehmann, J .
M.; Plunket, K.; Miller, A. B.; Milburn, M. V.; Kliewer, S. A.;
Willson, T. M. A peroxisome proliferator-activated receptor γ
ligand inhibits adipocyte differentiation. Proc. Natl. Acad. Sci.
U.S.A. 1999, 96, 6102-6106.
(21) J anowski, B. A.; Grogan, M. J .; J ones, S. A.; Wisely, G. B.;
Kliewer, S. A.; Corey, E. J .; Mangelsdorf, D. J . Structural
requirements of ligands for the oxysterol liver X receptors LXRR
and LXRâ. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 266-271.
(22) Xu, L.; Glass, C. K.; Rosenfeld, M. G. Coactivator and corepressor
complexes in nuclear receptor function. Curr. Opin. Genet. Dev.
1999, 9, 140-147.
(23) Freedman, L. P. Increasing the complexity of coactivation in
nuclear receptor signaling. Cell 1999, 97, 5-8.
(24) Onate, S. A.; Tsai, S. Y.; Tsai, M.-J .; O’Malley, B. W. Sequence
and characterization of a coactivator for the steroid hormone
receptor superfamily. Science 1995, 270, 1354-1357.
(25) Tomkinson, N. C. O.; Willson, T. M.; Russel, J . S.; Spencer, T.
A. Efficient, stereoselective synthesis of 24(S),25-epoxycholes-
terol. J . Org. Chem. 1998, 63, 9919-9923.
(26) Partridge, J . J .; Faber, S.; Uskokovic, M. R. Vitamin D3
metabolites. I. Synthesis of 25-hydroxycholesterol. Helv. Chim.
Acta 1974, 57, 764-771.
(27) Bailey, W. F.; Punzalan, E. R. Convenient general method for
the preparation of primary alkyllithiums by lithium-iodine
exchange. J . Org. Chem. 1990, 55, 5404-5406.
(28) Lier, J . E. V.; Smith, L. L. Sterol metabolism. X. Epimeric 23-
hydroxycholesterols. J . Pharm. Sci. 1970, 59, 719-721.
(49) Harrington, W. W.; Brown, K. K. Unpublished results.
(50) Riegel, B.; Kaye, I. A. Preparation of 24-keto- and 24-hydroxy-
cholesterol and certain derivatives. J . Am. Chem. Soc. 1944, 66,
723-724.

Pharmacophore Analysis of LXRR
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 6 897
(51) Okamoto, M.; Tabe, M.; Fujii, T.; Tanaka, T. Asymmetric iso-
propylation of steroidal 24-aldehydes for the synthesis of 24(R)-
hydroxycholesterol. Tetrahedron: Asymmetry 1995, 6, 767-778.
(52) Firestone, R. A.; Pisano, J . M.; Falck, J . R.; McPhaul, M. M.;
Krieger, M. Selective delivery of cytotoxic compounds to cells
by the LDL pathway. J . Med. Chem. 1984, 27, 1037-1043.
(53) Cope, A. C.; Ciganek, E. Organic Syntheses; Wiley: New York,
1963; Collect. Vol. IV, pp 339-342.
(55) Bose, A. K.; Kistner, J . F.; Farber, L. A convenient synthesis of
axial amines. J . Org. Chem. 1962, 27, 2925-2927.
(56) Lambert, M. H. Docking conformationally flexible molecules into
protein binding sites. In Practical Application of Computer-Aided
Drug Design; Charifson, P. S., Ed.; Marcel Dekker: New York,
1997; pp 243-303.
(54) Koziara, A.; Zwierzak, A. Optimized procedures for one-pot
conversion of alkyl bromides into amines via the Staudinger
reaction. Synthesis 1992, 1063-1065.
J M0004749