Design, synthesis, and biological evaluation of potent dual agonists of nuclear and membrane bile acid receptors

Article abstract of DOI:10.1021/jm401873d

Bile acids exert genomic and nongenomic effects by interacting with membrane G-protein-coupled receptors, including the bile acid receptor GP-BAR1, and nuclear receptors, such as the farnesoid X receptor (FXR). These receptors regulate overlapping metabolic functions; thus, GP-BAR1/FXR dual agonists, by enhancing the biological response, represent an innovative strategy for the treatment of enteroendocrine disorders. Here, we report the design, total synthesis, and in vitro/in vivo pharmacological evaluation of a new generation of dual bile acid receptor agonists, with the most potent compound, 19, showing promising pharmacological profiles. We show that compound 19 activates GP-BAR1, FXR, and FXR regulated genes in the liver, increases the intracellular concentration of cAMP, and stimulates the release of the potent insulinotropic hormone GLP-1, resulting in a promising drug candidate for the treatment of metabolic disorders. We also elucidate the binding mode of the most potent dual agonists in the two receptors through a series of computations providing the molecular basis for dual GP-BAR1/FXR agonism.

Full text of DOI:10.1021/jm401873d

Article
pubs.acs.org/jmc
Design, Synthesis, and Biological Evaluation of Potent Dual Agonists
of Nuclear and Membrane Bile Acid Receptors
Claudio D’Amore,§,∥ Francesco Saverio Di Leva,†,∥ Valentina Sepe,‡,∥ Barbara Renga,§
Chiara Del Gaudio,‡ Maria Valeria D’Auria,‡ Angela Zampella,*,‡ Stefano Fiorucci,§
and Vittorio Limongelli*,‡
†Drug Discovery and Development, Istituto Italiano di Tecnologia, Via Morego, 30, 16163 Genoa, Italy
‡Dipartimento di Farmacia, Universita
̀
di Napoli “Federico II”, Via D. Montesano, 49, I-80131 Napoli, Italy
§Dipartimento di Medicina Clinica e Sperimentale, Nuova Facolta
1-06132 Perugia, Italy
̀ ̀
di Medicina, Universita degli Studi di Perugia, Via Gambuli,
S
* Supporting Information
ABSTRACT: Bile acids exert genomic and nongenomic effects
by interacting with membrane G-protein-coupled receptors,
including the bile acid receptor GP-BAR1, and nuclear
receptors, such as the farnesoid X receptor (FXR). These
receptors regulate overlapping metabolic functions; thus, GP-
BAR1/FXR dual agonists, by enhancing the biological response,
represent an innovative strategy for the treatment of enter-
oendocrine disorders. Here, we report the design, total
synthesis, and in vitro/in vivo pharmacological evaluation of a
new generation of dual bile acid receptor agonists, with the
most potent compound, 19, showing promising pharmaco-
logical profiles. We show that compound 19 activates GP-
BAR1, FXR, and FXR regulated genes in the liver, increases the
intracellular concentration of cAMP, and stimulates the release of the potent insulinotropic hormone GLP-1, resulting in a
promising drug candidate for the treatment of metabolic disorders. We also elucidate the binding mode of the most potent dual
agonists in the two receptors through a series of computations providing the molecular basis for dual GP-BAR1/FXR agonism.
In addition to FXR, four nuclear receptors (PXR, CAR, VDR,
and LXRs)9−11 and G-protein-coupled receptors are known
targets for primary or secondary bile acids. The bile acid
receptor TGR5 (M-BAR) is a member of the rhodopsin-like
superfamily of G-protein-coupled receptor (GPCR)12 chris-
tened as G-protein-coupled bile acid receptor 1, GP-BAR1. GP-
BAR1 is highly expressed on the plasma membrane of liver,
small intestine, colon, adipose tissue, and skeletal muscle cells
as well as in macrophages/monocytes and is activated by
secondary bile acids, lithocolic acid (LCA) and tauro-LCA
(TLCA).13,14 Once activated by secondary bile acids, GP-BAR1
regulates multiple metabolic functions and increases energy
expenditure.15 In enteroendocrine L cells, GP-BAR1 activation
stimulates the secretion of glucagon-like peptide (GLP) 1, an
insulinotropic factor that enhances insulin release, thus
regulating glucose blood levels, gastrointestinal motility, and
appetite.16 In addition, GP-BAR1 has been recently identified
as a negative regulator of NF-kB. Indeed, in animal models of
liver inflammation, GP-BAR1 activation by 23(S)-methyl-
CDCA, a highly selective GP-BAR1 semisynthetic bile acid
INTRODUCTION
■
Bile acids (BAs), the end products of cholesterol catabolism,
are the main component of bile. Because of their amphipathic
nature, bile acids play an essential role in nutrient metabolism
and absorption.1 In the past decade bile acids have been
identified as signaling molecules endowed with genomic and
nongenomic effects. In 1999 it was demonstrated that primary
bile acids act as endogenous ligands for the farnesoid X
receptor (FXR), a member of the nuclear receptor super-
family2,3 and that chenodeoxycholic acid (CDCA) and its
conjugated forms are the most potent endogenous ligands.
Acting as an obligatory heterodimer with the retinoid X
receptor (RXR), FXR binds to specific responsive elements on
promoters of target genes, thus regulating the transcription of
enzymes/proteins involved in bile acid synthesis, transport,
conjugation, and excretion. Thus, the canonical physiological
role for FXR is to function as a bile acid sensor in enterohepatic
tissues to maintain bile acid homeostasis. FXR is also involved
in regulating critical check points in lipid4 and glucose
homeostasis.5 Furthermore, FXR ligands exert anti-inflamma-
tory6 and antifibrotic effects, making this nuclear receptor7 an
appealing pharmacological target in the treatment of common
human diseases ranging from metabolic syndrome to cancer.8
Received: October 26, 2013
Published: January 4, 2014
© 2014 American Chemical Society
937
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
agonist,17 attenuates the NF-kB activation induced by LPS
through the reduction of IkBα phosphorylation, thus
abrogating NF-kB translocation into nucleus.18 In rodent
models of colitis, GP-BAR1 expression increases in inflamed
tissues while its absence correlates with an increased intestinal
permeability and enhanced susceptibility to develop colitis.19
Moreover, GP-BAR1 was recently demonstrated to be essential
in maintaining gastric and intestinal mucosal integrity with a
protective effect against gastrointestinal injury caused by
nonsteroidal anti-inflammatory drugs.20
Because the two receptors might have overlapping activities,
the development of ligands endowed with dual activity toward
GP-BAR1 and FXR appears to be an intriguing strategy to
target enterohepatic and metabolic disorders.21,22 Thus, the 6α-
ethyl-3α,7α-dihydroxy-24-nor-5β-cholan-23-yl-23-triethylam-
monium sulfate, also named INT-767 (1, Figure 1), a
chenodeoxycholic acid derivative acting on both FXR and
GP-BAR1, attenuates signs and symptoms of liver injuries in
rodent models of metabolic syndrome.23−26
(7, quantitative yield). The large coupling constant observed
for H-6 proton signal at δH 2.83 (dd, J = 13.0, 5.5 Hz)
demonstrates its axial disposition and therefore the α-
orientation of the ethyl group at C-6. As previously reported
for the total synthesis of solomonsterol B, a C23 sulfate steroid
of marine origin endowed with PXR agonistic activity,29
Fisher’s esterification with formic acid and acetic anhydride30
generated the performate derivative (8) which was subjected to
“Beckmann rearrangement” by treatment with sodium nitrite in
a trifluoroacetic anhydride/trifluoroacetic acid mixture.31
Prolonged alkaline hydrolysis of 23-nitrile derivative 9
afforded 6α-ethyl-24-nor-7-ketolithocholic acid (10) in a 69%
three-step chemical yield. Sodium borohydride reduction of the
C7-ketone in a mixture of THF/H2O followed by methanol/p-
toluenesulfonic acid treatment afforded C23 methyl ester
1
derivative 11 (89% isolated yield in a two-step sequence). H
NMR spectrum analysis revealed that the sodium borohydride
reduction proceeded in a stereoselective manner, affording the
exclusive formation of 3α,7α-dihydroxy-6α-ethyl-24-nor-5β-
cholan-23-oic acid (see Experimental Section) intermediate,
as judged by the shape of H-7 (3.62, br s) which is consistent
with an equatorial disposition for this proton and therefore with
the axial α-orientation of the hydroxyl group.
Finally, the introduction of the sulfate functional group at
C23 was achieved by a three-step sequence including
protection of alcoholic functions at C3 and C7, reduction of
the methyl ester at C23 with lithium borohydride, and
subsequent sulfation with triethylammonium−sulfur trioxide
complex at 95 °C (72% three-step chemical yield). Removal of
hydroxyl protecting groups through treatment with methanol
and a few drops of HCl and purification by reversed-phase solid
extraction on a C18 cartridge afforded compound 1 as the
triethylammonium salt. HR ESIMS, negative ion mode m/z
471.2795 (calculated 471.2780 for C25H43O6S), and NMR data
(Table S1 in Supporting Information) secured the chemical
structure as reported in 1.
Figure 1. Previously reported dual GP-BAR1/FXR agonists.
Here, using compound 1 as a template for dual GP-BAR1/
FXR agonism, we present a new generation of dual bile acid
receptor agonists. This strategy resulted in the discovery of
compound 19, the most potent dual agonist of bile acid
receptors so far identified, which shows a promising activity for
the treatment of diabetes. The binding mode of the most
potent GP-BAR1/FXR dual agonists in the two receptors was
elucidated through a series of computations using the
homology model of GP-BAR1 and the X-ray structure of
FXR. Our findings provide the molecular basis for GP-BAR1/
FXR dual modulation and are valuable for further investigations
on the functional mechanism of bile acid receptors.
The above synthesis was completed in a total of 15 steps
starting from commercially available CDCA (3) and was found
to proceed with overall yield of 20%.
Derivative 11 was also used as starting material to obtain 13,
a compound with three sulfate groups, respectively, at C3 and
C7 on rings A and B and at C23 on the side chain (Scheme 1).
In order to increase the chemical space and investigate the
structural requisites for dual GP-BAR1/FXR agonism, we
developed a series of bile acid derivatives, modifying bile acid
chemical scaffold at different levels: (i) the side chain length
(C23 vs C24 and C26), (ii) the presence of sulfate groups on
the side chain and on rings A and B of the tetracyclic nucleus,
(iii) the configuration at C7, and (iv) the alkylation at C6.
Chenodeoxycholan Sulfate Derivatives. Methyl ester of
chenodeoxycholic acid (14) was reduced (LiBH4, MeOH/
THF) obtaining the triol 16 in high chemical yield (96%,
Scheme 2).
Exhaustive Sulfation (Et3N·SO3) and HPLC purification
afforded pure chenodeoxycholan sulfate 18 and the correspond-
ing trisulfate derivative 20.
The same synthetic protocol was applied on the methyl ester
of 6-ECDCA (15) to produce 6-ethylchenodeoxycholan sulfate
19 and trisulfate derivative 21 (Scheme 2).
RESULTS AND DISCUSSION
■
Medicinal Chemistry. Even if compound 1 has been
investigated as a dual GP-BAR1/FXR agonist,24,25 to date, no
efficient synthetic procedures have been published. Thus, to
have a positive control in biological assays, our research work
started with the total synthesis of 1. From a structural point of
view, the most intriguing feature of 1 is the presence of a
truncated C23 side chain including a sulfate end-terminus
group.
As depicted in the Scheme 1, the synthesis of compound 1
started with the commercially available chenodeoxycholic acid
(CDCA, 3), which was regioselectively oxidized to the C7
hydroxyl group27 and subsequently methylated at C24 carboxyl
acid to give methyl ester of 7-ketolithocholic acid (4) in a 65%
two-step chemical yield. Aldolic addition to a silyl enol ether
intermediate28 generated the methyl 3α-hydroxy-6-ethylidene-
7-keto-5β-cholan-24-oate 5 in a 77% yield. Exocyclic double
bond hydrogenation (H2 on Pd/C) and alkaline hydrolysis of
methyl ester functionality on the side chain afforded the useful
intermediate acid 3α-hydroxy-6α-ethyl-7-keto-5β-cholan-24-oic
tert-Butylsilyl protection on C3 and C7 hydroxyl groups of
CDCA methyl ester 14 followed by LiBH4 reduction afforded
alcohol 22 (Scheme 3). Swern oxidation32/C2 Horner33
homologation gave the conjugate bis-homo ethyl ester 23
938
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
a
Scheme 1. Total Synthesis of 1 and Its Trisulfate Derivative 13
a
Reagents and conditions: (a) NaBr, TBABr, NaClO 10%, MeOH/AcOH/H2O/AcOEt 3:1:0.25:6.5; (b) p-TsOH, MeOH dry, 65% over two steps;
(c) DIPA, n-BuLi, TMSCl, TEA dry, THF dry, −78 °C; (d) acetaldehyde, BF3(OEt)2, CH2Cl2, −60 °C, 77% over two steps; (e) H2, Pd/C, THF/
MeOH 1:1, quantitative yield; (f) NaOH 5% in MeOH/H2O 1:1 v/v, quantitative yield; (g) HCOOH, HClO4, 81%; (h) TFA, trifluoroacetic
anhydride, NaNO2; (i) KOH 30% in MeOH/H2O 1:1 v/v, 86% over two steps; (j) NaBH4, THF/H2O 4:1 v/v; (k) p-TsOH, MeOH dry, 89% over
two steps; (l) 2,6-lutidine, tert-butyldimethylsilyl trifluoromethanesulfonate, CH2Cl2, 0 °C, 78%; (m) LiBH4, MeOH dry, THF, 0 °C, 95%; (n) Et3N·
SO3, DMF, 95 °C; (o) HCl 37%, MeOH, 87% over two steps; (p) LiBH4, MeOH dry, THF; (q) Et3N·SO3, DMF, 95 °C, 72% over two steps.
a
Scheme 2. Chenodeoxycholan Sulfate Derivatives
a
Reagents and conditions: (a) LiBH4, MeOH dry, THF, 0 °C; (b) Et3N·SO3, DMF, 95 °C.
whose double bond reduction and silyl deprotection proceeded
smoothly affording ethyl ester 24 in 61% yield over six steps.
Once again LiBH4 reduction, sulfation, and HPLC separation
gave pure 26 and 27 derivatives (Scheme 3).
efficiently separated by HPLC (Scheme 4). Bis-homo
derivatives (36, 37, 38) were obtained following the synthetic
protocol already described for the corresponding chenodeox-
ycholan derivatives and depicted in Scheme 4.
Ursodeoxycholan Sulfate Derivatives. Among BAs,
ursodeoxycholic acid (UDCA) is structurally unique, differing
in the β-orientation of the hydroxyl group at the C7 position
with respect to CDCA. Considering the inactivity of UDCA
and its C24 alcohol derivative on FXR,2,3,34 structural
modification on a 7β-hydroxylcholanoid scaffold could be
instrumental in generating selective GP-BAR1 agonists.
First, silyl protection on UDCA methyl ester 28 followed by
reduction, sulfation at C24, and deprotection gave ursodeox-
ycholan sulfate 30 (55% chemical yield in four steps, Scheme 4,
steps a−d). Second, deprotection on derivative 29 and
exhaustive sulfation produced a mixture of ursodeoxycholan
derivatives (31, 32, 33) differing in the sulfation pattern and
Pharmacological Evaluation on the Library. To assess
transcriptional activity, all the synthetic derivatives obtained in
this study were tested in the luciferase reporter assays on
HepG2 cells transfected with human FXR and on HEK-293T
cells transfected with human GP-BAR1.
As reported in Figure 2A, in addition to 1 and 6-ECDCA (2),
already reported as potent FXR agonists, compounds 19 and
26, 6-ethylchenodeoxycholan sulfate and bis-homochenodeox-
ycholan sulfate, respectively, were potent inducers of FXR
transactivation. Specifically, compound 19 transactivates FXR
with the same potency of 1 and 2. None of the tested
compounds turned out to be an FXR antagonist (Figure 2B).
939
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
a
Scheme 3. Bis-homochenodeoxycholan Sulfate Derivatives
a
Reagents and conditions: (a) 2,6-lutidine, tert-butyldimethylsilyl trifluoromethanesulfonate, CH2Cl2, 0 °C, 83%; (b) LiBH4, MeOH dry, THF, 0 °C,
quantitative yield; (c) DMSO, oxalyl chloride, TEA dry, CH2Cl2, −78 °C; (d) LiOH, TEPA, THF dry, reflux, 91% over two steps; (e) H2, Pd(OH)2,
degussa type, THF/MeOH 1:1, quantitative yield; (f) HCl 37%, EtOH, 81%; (g) LiBH4, MeOH dry, THF, 77%; (h) Et3N·SO3, DMF, 95 °C.
a
Scheme 4. Ursodeoxycholan Sulfate and Bis-homoursodeoxycholan Sulfate Derivatives
a
Reagents and conditions: (a) 2,6-lutidine, tert-butyldimethylsilyl trifluoromethanesulfonate, CH2Cl2, 0 °C; (b) LiBH4, MeOH dry, THF, 0 °C, 80%
over two steps; (c) Et3N·SO3, DMF, 95 °C; (d) HCl 37%, MeOH, 69% over two steps; (e) HCl 37%, MeOH, 96%; (f) Et3N·SO3, DMF, 95 °C; (g)
DMSO, oxalyl chloride, TEA dry, CH2Cl2, −78 °C then LiOH, TEPA, THF dry, reflux, 40% over two steps; (h) H2, Pd/C, THF/MeOH 1:1,
quantitative yield; (i) LiBH4, MeOH dry, THF, 88%; (j) Et3N·SO3, DMF, 95 °C; (k) HCl 37%, MeOH, 75% over two steps; (l) Et3N·SO3, DMF, 95
°C.
Similar to what was observed for FXR, results of trans-
activations of CREB-responsive elements in HEK-293T
transiently transfected with the human membrane bile acid
receptor GP-BAR1 revealed that compounds 19 and 26 were
potent inducers of cAMP-luciferase reporter gene (Figure 3).
Moreover, none of tested compounds were able to induce
cAMP-luciferase reporter gene in the absence of GP-BAR1
(Figure S1 in Supporting Information), thus indicating that the
above induction is GP-BAR1 mediated. The potency of
compound 19 in transactivating CREB responsive element
was similar to that of TLCA, the most potent endogenous GP-
BAR1 agonist (Figure 3). Forskolin, which increases cAMP in a
receptor-independent manner, was used as positive control in
these experiments.
Pharmacological Evaluation on Derivative 19, the
Most Potent GP-BAR1/FXR Dual Agonist So Far
Identified. Because 19 transactivates both FXR and GP-
BAR1, we have then investigated the concentration−response
curve for this compound. As shown in Figure 4, compound 19
transactivates FXR with an EC50 of ∼1 μM (panels A and B)
and induces GP-BAR1 activity with an EC50 of ∼0.2 μM
(panels C and D) and both values were comparable to those
previously reported for the reference dual agonist 1.25 In
addition, the efficacy of 19 was 852% and 112% compared to
CDCA and TLCA, respectively, thus indicating that compound
19 is a potent GP-BAR1/FXR dual agonist. To further examine
the effect of compound 19 on FXR-regulated activities, we have
carried out gene expression analysis assessing its activity on
three FXR regulated genes, i.e., small heterodimer partner
940
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
Figure 2. Transactivation assays on FXR. (A) HepG2 cells were transfected with pSG5-FXR, pSG5-RXR, pCMV-βgal, and p(hsp27)TKLUC
vectors. Cells were stimulated with 10 μM all tested compounds and with 1 μM compound 1. CDCA (10 μM) was used as a positive control. Results
are expressed as the mean standard error: (∗) p < 0.05 vs nontreated cells (NT). (B) HepG2 cells were transfected with pSG5-FXR, pSG5-RXR,
pCMV-βgal, and p(hsp27)TKLUC vectors. Cells were stimulated with CDCA, 10 μM, in combination with 50 μM tested compounds. Results are
expressed as the mean standard error: (∗) p < 0.05 vs nontreated cells (NT); (#) p < 0.05 vs CDCA.
(SHP), organic solute transporter (OST) α, and bile salt export
pump (BSEP).
increases intracellular concentration of cAMP in GLUTAg cells
and is significantly more potent that TLCA, the physiological
ligand of GP-BAR1, and 6-ECDCA (2). Interestingly, the three
compounds caused a robust increase of GLP-1 concentrations
in cell supernatants (panel E).
It is worth mentioning that GLP-1 is an appealing
pharmacological target in the treatment of diabetes. GLP-1
receptor agonists, like incretins, are currently used in the
treatment of type 1 diabetes, although concerns on the safety of
these ligands have emerged recently.38−42
For these purposes, HepG2 cells were used. Results shown in
Figure 5, panels A−C, confirmed data reported in Figure 2 and
demonstrate that compound 19 is a potent FXR agonist that
increases the expression of the three FXR target genes more
efficiently than CDCA (3), the FXR endogenous ligand, and to
a similar extent to 6-ECDCA (2), the most potent FXR steroid
ligand so far available35 and endowed with promising
pharmacological properties.36
We have then examined the effects of 19 on GP-BAR1 on
GLUTAg cells. GLUTAg cells37 are an intestinal enter-
oendocrine cell line, known for their ability to release
glucagon-like peptide 1 (GLP-1) in response to GB-BAR1
agonists. This cell line expresses high levels of GP-BAR1 and
reacts to receptor binding by a robust generation of cAMP.
Therefore, GLUTAg cells are considered a suitable model to
evaluate the interactions between GP-BAR1 and its ligands. As
illustrated in Figure 5, panel D, compound 19 effectively
Bile acids are potent inducers of GLP-1 release and can
therefore hold utility in the treatment of diabetes. The fact that
19 releases GLP-1 from the intestine, without altering its half-
life, could result in beneficial effects without increasing the risk
of long-acting GLP-1 derivatives.
In addition, the specificity of these interactions in the ability
of 19 to modulate FXR was further investigated using FXR−/−
cells. As shown in Figure 5, panel F, while compounds 2 and 19
increased BSEP expression in primary hepatocytes isolated
941
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
Figure 3. Activation of GP-BAR1. HEK-293T cells were co-transfected with human GP-BAR1 and a reporter gene containing a cAMP responsive
element in front of the luciferase gene. At 24 h after transfection, cells were stimulated with 10 μM compounds 1−3, 13, 18−21, 26−27, 30−33,
36−38. Luciferase activity served as a measure of the rise in intracellular cAMP following activation of GP-BAR1. TLCA (T, 10 μM) stimulates
cAMP production in a GP-BAR1 dependent manner. Forskolin (F, 10 μM) stimulated cAMP production independently of GP-BAR1. Results are
expressed as the mean standard error: (∗) p < 0.05 vs nontreated cells (NT).
Figure 4. Concentration−response curves for compound 19. (A, B) HepG2 cells were trasfected with human FXR and stimulated with increasing
concentrations of 19 (0.01, 0.1, 1, and 10 μM). (C, D) HEK-293T were transfected with human GP-BAR1 and stimulated with increasing
concentration of 19 (0.1, 1, and 10 μM). CDCA (10 μM) was used as a positive control to evaluate the FXR transactivation. TLCA (10 μM) was
used as a positive control to evaluate the GP-BAR1 activity. Results are expressed as the mean standard error: (∗) p < 0.05 vs nontreated cells
(NT).
from wild type mice, this effect was lost when hepatocytes were
prepared from FXR−/− mice.
Computational Studies. To disclose the molecular basis of
dual GP-BAR1/FXR agonism, we investigated the binding
942
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
Figure 5. Pharmacological evaluation on derivative 19, the most potent GP-BAR1/FXR dual agonist so far identified. (A−C) Real-time PCR analysis
of mRNA expression of FXR target genes (A) SHP, (B) OSTα, and (C) BSEP in HepG2 cells primed with compounds 2, 3, and 19. Values are
normalized relative to GAPDH mRNA and are expressed relative to those of nontreated cells (NT), which are arbitrarily set to 1: (∗) p < 0.05 vs
NT. (D, E) Effect of 2 and 19 on intracellular generation of cAMP and GLP-1 release in GLUTAg cells. The data are the mean SE of four to five
experiments: (∗) p < 0.05 vs NT. (F) Compounds 2 and 19 stimulate BSEP expression in wild type hepatocytes but lose their efficacy in hepatocytes
obtained from FXR−/− mice. Data are the mean SE of four experiments: (∗) p < 0.05 vs control and (∗∗) p > 0.05 vs wild type cells.
mechanism of the most potent dual agonists, compounds 1 and
19, to the two receptors, performing a series of computations.
While the X-ray structure of FXR has been resolved allowing
the discovery of several synthetic and natural ligands,21,43 the
tridimensional structure of GP-BAR1 has not been determined
yet. Therefore, we first decided to build the tridimensional
model of this receptor using homology modeling techniques
(Figure 6A and Figure 6B). As template structure, we choose
the agonist conformation of the adenosine A2A receptor (PDB
code 2ydo),44 which shows one of the highest sequence identity
and similarity values, 20% and 55%, respectively (see
Supporting Information for details and Figure S2). This
model, together with the crystal structure of FXR, was used
in the subsequent docking study to disclose the binding mode
of compounds 1 and 19 in the two receptors and then to define
the structural requirements for dual GP-BAR1/FXR agonism.
Binding Mode of 1 in GP-BAR1 and FXR. To take into
account the receptor flexibility, 10 different GP-BAR1
conformations, obtained from the MD simulation on the apo
form of the receptor, were considered for the docking
calculations of compound 1 (see Supporting Information).
The AutoDock4.2 docking program45 found similar binding
results in the different receptor conformations, with the most
populated clusters showing the ligand occupying the region
underneath the extracellular loop II. However, most of the
docking solutions were not fully in agreement with the
requisites necessary for the binding of bile acid derivatives to
GP-BAR1 reported in literature.46 These include (i) the
interaction of the 3α and/or 7α hydroxyl groups with a
negatively charged region in the binding site, (ii) the insertion
of the side chain at C17 position into a rather large and polar
cavity, (iii) the insertion of the alkyl substituent at C6 position
into an ancillary hydrophobic pocket. The inaccuracy of
docking calculations was probably due to the conformations
of the side chains of the binding site residues, which in the apo
form are different with respect to the agonist-bound state. To
overcome this limitation, we had to step up the computational
strategy and take into account the full receptor flexibility. In
such a way, even difficult ligand/protein binding mechanisms
can be elucidated.47−49 In this case, molecular dynamics (MD)
simulations on the 1/GP-BAR1 complex embedded in a POPC
bilayer surrounded by explicit water were carried out. In this
calculation, the docking pose that fulfilled most of the
aforementioned structural requirements for the binding was
selected as starting conformation. Then we performed steered
MD simulations to refine the ligand binding conformation
optimizing the distance and orientation of the ligand−protein
interactions. The obtained complex was equilibrated and
underwent an over-100 ns MD simulation during which the
whole system was fully flexible. The ligand binding con-
formation, depicted in Figure 6C, was stable throughout the
whole simulation, as shown by the low rmsd value reported in
Figure 6D, with all the ligand−protein interactions conserved.
In particular, the ligand 3α-hydroxyl group H-bonds with
Glu169 and interacts via a water molecule with Tyr240 side
chain, while the 7α-hydroxyl group is involved in a H-bond
with Asn93 and water-mediated interactions with Tyr89 and
Glu169. On the other side, the sulfate side chain points toward
943
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
Figure 6. Tridimensional model of GP-BAR1 and binding mode of 1. Side (A) and top (B) views of the tridimensional model of the GP-BAR1
receptor embedded in the POPC bilayer. The receptor is shown as gray cartoons. Lipids are depicted as yellow sticks. The ligand binding cavity (B)
is shown as transparent orange surface. Residue numbers were assigned according to the amino acid sequence of hGP-BAR1. (C) Binding mode of 1
obtained through over 100 ns long MD simulation. The ligand and interacting residues are represented as yellow and gray licorice, respectively, while
the GP-BAR1 receptor is shown as gray cartoons. Water molecules are displayed as spheres, while all the hydrogens are omitted for clarity reasons.
(D) The rmsd plot of the ligand heavy atoms during the production MD run. Prior to the rmsd calculations, trajectory frames were aligned using the
coordinates of the Cα carbons of the transmembrane helices.
helices I, II, and VII where the ligand can establish water-
mediated interactions with the side chains of Ser21 and Ser267
and with the backbone CO groups of Leu68 and Leu71.
The steroidal scaffold engages several hydrophobic inter-
actions with Leu71, Phe83, Leu174, and Trp237, further
stabilizing the ligand binding mode.
The 6-ethyl group of 1 deepens in a hydrophobic pocket
formed by residues Leu97, Phe138, Leu173, and Leu174. The
peculiar shape and lipophilicity of this pocket suggest the
importance of having short and hydrophobic chains at the C-6
position on the steroidal scaffold to achieve optimal GP-BAR1
agonist activity. The proposed ligand binding model is in full
agreement with the structural features required for GP-BAR1
binding46 and is further substantiated by recent mutagenesis
data,50 reporting a decrease of the binding affinity for ligands
similar to 1 in Asn93Ala, Glu169Ala, and Tyr240Ala mutant
forms of GP-BAR1. These findings and the stability of the
ligand−protein interactions along the MD simulation support
the reliability of our receptor model and the proposed binding
mode.
performed using the crystal structure of the FXR-LBD from
Rattus norvegicus (rFXR) in complex with 6-ethylchenodeox-
ycholic acid (6-ECDCA, 2), a semisynthetic GP-BAR1/FXR
dual agonist bile acid derivative with a FXR binding affinity that
is significantly greater than the endogenus ligand CDCA,35 and
the GRIP-1 coactivator peptide NID-3 (PDB code 1osv).51
rFXR-LBD shares the 95% of homology with that of the human
FXR-LBD (hFXR-LBD), with all of the residues in the ligand
binding pocket conserved among the two species. For docking
studies we again used the AutoDock4.2 software (AD4.2),
which has been successfully used in our previous computational
studies on NRs.52 The best scored and most occurring ligand
docking conformation shows 1 bound to the cavity formed by
helices H3, H5, H7, H11, and H12 (Figure 7).
Here, the steroidal scaffold establishes favorable hydrophobic
interactions with the side chains of Met262, Met287, Ala288,
and Trp466, while the sulfate group salt-bridges with the
Arg328 guanidinium group. In addition, the 3α-OH H-bonds
with the side chains of Tyr358 and the protonated His444,
while the 7α-OH H-bonds with the Ser329 and Tyr366
hydroxyl groups. The latter residue together with Tyr358,
Ile359, and Phe363 also engages hydrophobic contacts with the
6-ethyl substituent, further stabilizing the binding pose. Overall,
the binding mode is very similar to that experimentally found
for 6-ECDCA.51 In fact, like 6-ECDCA, 1 interacts through its
A movie showing the stability of the ligand binding mode
during the 100 ns MD simulation can be found as Supporting
Information.
To elucidate the binding mode of 1 to FXR ligand binding
domain (FXR-LBD), molecular docking calculations were
944
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
although GP-BAR1 and FXR are structurally very different, the
active ligands show common structural features. In particular,
side chain elongation led to derivative 19, which is the most
potent compound of the series toward both targets. Compound
19 occupies the GP-BAR1 binding site similarly to 1 (Figure
8A). However, different from 1, the 19 side chain can engage
direct (i.e., not water-mediated) H-bonds with the Ser21,
Ser267, and Ser270 residues on helices TM1 and TM7, thus
explaining the improved agonist profile of 19 in comparison
with 1.
In the FXR site, the sulfate group of 19 engages a salt-bridge
interaction with the side chain of Arg328 in a way similar to 1.
At variance with 1, compound 19 is also able to H-bond with
the backbone NH of Met262 in the loop connecting helices H1
and H2, and the sulfate group is close enough to interact with
the Arg261 side chain (Figure 8B). The latter interactions
further stabilize the binding mode of this ligand, thus explaining
the higher activity of 19 if compared with that of 1. It is
interesting to note that docking calculations show that even
shorter side chains, as in the case of 2, are allowed, since the
ligand is able to interact with at least one serine residue in GP-
BAR1 (Figure S5A in Supporting Information) and with
Arg328 in FXR (Figure S5B in Supporting Information).
Furthermore, the presence of the sulfate group on the side
chain is preferable to the carboxylate, since these groups are
placed in solvent accessible regions in both receptors, where the
sulfate is able to establish a larger number of interactions with
water molecules and surrounding residues. Finally, docking
studies would suggest that a negatively charged group is not
strictly necessary; however, a polar group, able to form H-bond
interactions, is desirable to activate GP-BAR1. On the other
hand, a negatively charged group is required to activate FXR
interacting with either Arg328 or Arg261 or both residues in
the FXR-LBD.
On the other side, the presence of the ethyl group and
hydroxyl group at C6 and C7, respectively, and their
configuration are important for the dual activity. In fact, the
most potent agonists of the series, compounds 1, 2, and 19,
have both these groups in the α configuration. In GP-BAR1,
while the ligand 7α-OH group H-bonds with Asn93, the 6α-
ethyl deepens into a hydrophobic pocket formed by Leu97,
Phe138, and Leu173. These nonpolar interactions are
important to stabilize the ligand binding conformation. In
fact, compounds that do not present the 6α-ethyl group show a
lower activity (e.g., compare 2 vs 3 and 18 vs 19). In FXR,
Figure 7. Binding mode of 1 in FXR. The ligand is depicted as yellow
sticks, respectively. FXR is shown as orange (helices H3, H4, and
H12) and gray cartoons. Amino acids involved in ligand binding are
shown as orange sticks. Hydrogens are omitted for clarity.
3α-OH group with Tyr358 on helix H11 and with His444 on
helix H12, contributing to stabilization of the cation−π
interaction between His444 and Trp466 in the AF-2 domain.
This interaction favors helices H3, H4, H11, and particularly
H12 to adopt a conformation competent for the binding of
coactivator peptides, thus inducing the transcription of target
genes. Moreover, similar to the carboxylate group of 6-ECDCA,
the sulfate moiety of 1 salt-bridges with the Arg328 side chain.
This interaction, as well as the aforementioned cation−π
between Tyr358 on H11 and His444 on H12, is important for
the activation of FXR by bile acid derivatives.51
Insight on Structural Requisites for Dual GP-BAR1/
FXR Agonism. Docking calculations in the GP-BAR1 and
FXR binding sites were performed on the most potent dual
agonist of the series, compound 19. For these calculations the
GP-BAR1 conformation obtained from the MD simulation on
the 1/GP-BAR1 complex was used. The quality of this
structure was assessed using the MolProbity server (http://
molprobity.biochem.duke.edu) showing 100% of all residues in
allowed conformational regions according to the Ramachan-
dran analysis (Figure S3). For this ligand, docking calculations
found, among the best scored solutions, a binding mode very
similar to those of 1 in the two receptors (Figure 8). Using
these results, we were able to define the structural requisites to
have mutually active ligands on the two targets. In fact,
Figure 8. Binding mode of 19 in GP-BAR1 and in FXR. (A) Conformation obtained from MD simulations on the 1/GP-BAR1 complex. (B)
Conformation obtained from MD simulations within rFXR-LBD (PDB code 1osv). Compound 19 is represented as cyan sticks. The receptors are
shown as gray and orange (helices H3, H4, and H12 in FXR) cartoons. Amino acids involved in ligand binding are depicted as gray (in GP-BAR1)
and orange (in FXR) sticks. Hydrogens are omitted for clarity.
945
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
similar to what was observed for 1, the 7α-OH of 2 and 19 H-
bonds with the Ser329 and Tyr366 hydroxyl groups while the
6α-ethyl substituent inserts into the hydrophobic pocket
formed by Tyr358, Ile359, Phe363, and Tyr366. As for GP-
BAR1, also in FXR the absence of the 6α-ethyl group leads to
less active compounds (compare 2 vs 3 and 18 vs 19). It is
interesting to note that despite the lack of the 6α-ethyl group,
26 is still able to activate, albeit to a minor extent than the more
potent derivatives 1, 2, and 19, both GP-BAR1 and FXR in
transactivation assays (Figure 2A and 3). This can be ascribed
to the presence of a longer side chain allowing the ligand to
tightly interact with the serine residues on helices TM1 and
TM7 in GP-BAR1 and with the Arg261 and Arg328 side chains
in FXR.
Furthermore, as shown in Figures 2 and 3, all ursodeox-
ycholan sulfate (30−33) and bis-homoursodeoxycholan sulfate
derivatives (36−38) were inactive toward both receptors.
These findings demonstrate that despite the presence of one or
more sulfate groups on the UDCA scaffold, the configuration at
C7 once more plays a fundamental role to properly
accommodate the molecules in the binding sites of the
receptors. In fact, inverted configuration does not allow the
7-OH group to interact with Tyr89 and Asn93 on TM3 in GP-
BAR1 and with Ser329 and Tyr366 in the FXR-LBD. The loss
of these interactions, together with the lack of the 6α-ethyl
substituent on the UDCA scaffold, can explain the inability of
these derivatives to activate either GP-BAR1 or FXR in
transactivation assays.
In conclusion, this study is relevant for further investigations
on the functional mechanism of these two receptors and for the
design of novel dual GP-BAR1/FXR agonists, providing new
opportunities for the treatment of enterohepatic and metabolic
disorders.
EXPERIMENTAL SECTION
■
Chemistry. Specific rotations were measured on a Jasco P-2000
polarimeter. High-resolution ESI mass spectra were performed with a
Micromass Q-TOF mass spectrometer. NMR spectra were obtained
on Varian Inova 400 and Varian Inova 700 NMR spectrometers (1H at
400 and 700 MHz, 13C at 100 MHz) equipped with a Sun hardware
and recorded in CDCl3 (δH = 7.26 ppm and δC = 77.0 ppm) and
CD3OD (δH = 3.30 ppm and δC = 49.0 ppm). J values are in hertz, and
chemical shifts (δ) are reported in ppm and referenced to CHCl3 and
CHD2OD as internal standards. HPLC was performed using a Waters
model 510 pump equipped with Waters Rheodine injector and a
differential refractometer, model 401. Reaction progress was
monitored via thin-layer chromatography (TLC) on Alugram silica
gel G/UV254 plates. Silica gel MN Kieselgel 60 (70−230 mesh) from
Macherey-Nagel Company was used for column chromatography. All
chemicals were obtained from Sigma-Aldrich, Inc. Solvents and
reagents were used as supplied from commercial sources with the
following exceptions. Tetrahydrofuran, dichloromethane, diisopropyl-
amine, and triethylamine were distilled from calcium hydride
immediately prior to use. Methanol was dried from magnesium
methoxide as follows. Magnesium turnings (5 g) and iodine (0.5 g) are
refluxed in a small (50−100 mL) quantity of methanol until all of the
magnesium has reacted. The mixture is diluted (up to 1 L) with
reagent grade methanol, refluxed for 2−3 h, and then distilled under
nitrogen. All reactions were carried out under argon atmosphere using
flame-dried glassware. The purity of all of the intermediates, checked
Finally, independent from the side chain length and the
alkylation at C6, the introduction of additional sulfate groups
on rings A and B produces a loss in the activity of cholane
derivatives toward both receptors (compare 1 vs 13, 26 vs 27,
and 19 vs 21 in Figures 2 and 3). This effect might be
addressed to the bulkiness and the highly negative charge of
these groups, which can play a negative role in the ligand/
receptor binding process in both GP-BAR1 and FXR.
1
by H NMR, was greater than 95%. The purity of tested compounds
was determined to be always greater than 95% by analytical HPLC
analysis as reported for each compound.
Compounds 4 and 5 were prepared as previously reported.27,28
Methyl 6α-Ethyl-3α-hydroxy-7-keto-5β-cholan-24-oate (6).
A solution of methyl 3α-hydroxy-6-ethyliden-7-keto-5β-cholan-24-oate
5 (1.3 g, 3.06 mmol) in dry THF/dry MeOH (50 mL, 1:1 v/v) was
hydrogenated in the presence of palladium, 5 wt %, on activated
carbon (50 mg). The flask was evacuated and flushed first with argon
and then with hydrogen. The mixture was stirred at room temperature
under H2 for 8 h. The catalyst was filtered through Celite, and the
recovered filtrate was concentrated under vacuum to give 6 (1.13 g,
quantitative yield). An analytic sample was obtained by silica gel
chromatography, eluting with hexane/EtOAc 8:2 and 0.5% of
triethylamine. [α]2D5 −3.7 (c 0.5, CHCl3). Selected 1H NMR (400
MHz, CDCl3): δ 3.67 (3H, s), 3.57 (1H, m), 2.57 (1H, t, J = 11.5 Hz),
2.37 (1H, m), 2.24 (1H, dd, J = 6.6, 9.6 Hz), 2.20 (1H, m), 1.22 (3H,
s), 0.93 (3H, d, J = 6.2 Hz), 0.85 (3H, t, J = 7.3 Hz), 0.67 (3H, s). 13C
NMR (100 MHz, CDCl3): δ 215.1, 178.4, 71.6, 56.3, 51.3, 50.6, 50.4,
47.5, 46.4, 44.4, 43.8, 40.4, 38.2, 36.6, 36.3, 35.2, 32.4, 32.0, 30.6, 29.3,
25.8, 23.5 (2C), 22.8, 18.8, 12.5, 12.0. HRMS-ESI m/z 433.3323 [M +
H+], C27H45O4 requires 433.3318.
CONCLUSIONS
■
In the present study, bile acid derivatives have been developed
leading to the identification of compound 19, which is the most
potent GP-BAR1/FXR dual agonist reported so far. In
particular, 19 transactivates FXR and increases the expression
of FXR-regulated genes in the liver. On the other hand, 19
increases CREB transactivation, a measure of its GP-BAR1
agonism in HEK-293T cells transfected with the membrane
receptor. Furthermore, the activity of 19 on GP-BAR1 was also
confirmed on cell lines known to respond to GP-BAR1 ligands,
GLUTAg cells.53 In vivo and in vitro studies have shown that
19 increases the intracellular concentration of cAMP and
stimulates the release of the potent insulinotropic hormone
GLP-1,37 thus representing a promising lead for the treatment
of insulin-dependent conditions including diabetes. Further-
more, the binding mode of the most potent dual agonist to GP-
BAR1 and FXR was elucidated through a series of computa-
tional studies. These simulations also provided the molecular
bases to achieve potent GP-BAR1/FXR dual agonism. Despite
its potent GP-BAR1 activity, 19 maintains the ability to
stimulate FXR and FXR-target genes including BSEP. This
remains a potential drawback because FXR activation worsens
liver injury in rodent model of cholestasis, and therefore, the
use of dual GP-BAR1/FXR agonist might have similar side
effects of selective FXR ligands.54
6α-Ethyl-3α-hydroxy-7-keto-5β-cholan-24-oic Acid (7). Com-
pound 6 (1.1 g, 2.5 mmol) was hydrolyzed with a methanol solution of
sodium hydroxide (5%, 30 mL) in H2O (6 mL) overnight under reflux.
The resulting solution was then concentrated under vacuum, diluted
with water, acidified with HCl, 6 N, and extracted with ethyl acetate (3
× 50 mL). The collected organic phases were washed with brine, dried
over anhydrous Na2SO4, and evaporated under reduced pressure to
give 7 in quantitative yield (1.1 g). An analytic sample was obtained by
silica gel chromatography, eluting with CH2Cl2/MeOH 95:5. [α]D25
1
−21.5 (c 0.35, CH3OH). Selected H NMR (400 MHz, CD3OD): δ
3.46 (1H, m), 2.83 (1H, dd, J = 13.0, 5.5 Hz), 2.50 (1H, t, J = 11.2
Hz), 2.34 (1H, m), 2.20 (1H, m), 1.22 (3H, s), 0.96 (3H, d, J = 6.6
Hz), 0.81 (3H, t, J = 7.3 Hz), 0.71 (3H, s). 13C NMR (100 MHz,
CD3OD): δ 215.5, 187.0, 71.7, 56.4, 53.3, 52.2, 51.2, 50.5, 45.4, 43.8,
40.4, 36.8, 36.6, 35.3, 32.6, 32.3, 32.0, 30.6, 29.3, 25.6, 23.9, 23.0, 20.8,
946
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
18.8, 12.5, 12.3. HRMS-ESI m/z 419.3164 [M + H+], C26H43O4
requires 419.3161.
dried with Na2SO4, and evaporated to give 11 as an amorphous solid
(645 mg, 96%). An analytic sample was obtained by silica gel
chromatography, eluting with hexane/EtOAc 6:4. [α]D25 + 0.95(c 0.96,
CHCl3). Selected 1H NMR (400 MHz, CD3OD): δ 3.61 (3H, s), 3.62
(1H, s ovl), 3.28 (1H, m), 0.95 (3H, d, J = 6.9 Hz), 0.88 (3H, s), 0.87
(3H, t, J = 7.2 Hz), 0.69 (3H, s). 13C NMR (100 MHz, CD3OD): δ
175.7, 73.2, 71.2, 57.8, 51.8, 51.6, 46.9, 43.9, 43.8, 43.1, 42.3, 41.5,
41.0,36.7, 35.2, 35.1, 34.4, 31.2, 29.3, 24.5, 23.8, 23.5, 21.9, 19.6, 12.3,
12.0. HRMS-ESI m/z 421.3322 [M + H+], C26H45O4 requires
421.3318.
6α-Ethyl-3α-formyloxy-7-keto-5β-cholan-24-oic Acid (8). A
solution of 7 (1.0 g, 2.6 mmol) in 30 mL of 90% formic acid
containing 90 μL of 70% perchloric acid was stirred at 47−50 °C for 6
h. The temperature of the heating bath was lowered to 40 °C. Then 24
mL of acetic anhydride was added over 10 min, and the mixture was
stirred for 10 min more. The solution was cooled to room
temperature, poured into 50 mL of water, and extracted with diethyl
ether. The organic layers were washed with water to neutrality, dried
over Na2SO4, and evaporated to give 940 mg of 8 (81%). An analytic
sample was obtained by silica gel chromatography, eluting with
CH2Cl2/MeOH 95:5. [α]D25 −25.9 (c 0.56, CH3OH). Selected 1H
NMR (400 MHz, CDCl3): δ 7.99 (1H, s), 4.79 (1H, m), 2.71 (1H, dd,
J = 5.9,12.8 Hz), 1.29 (3H, s), 0.93 (3H, d, J = 6.3 Hz), 0.80 (3H, t, J =
7.2 Hz), 0.66 (3H, s). 13C NMR (100 MHz CDCl3): δ 212.5, 180.0,
160.6, 73.2, 54.8, 51.9, 50.6, 49.9, 48.9, 43.6, 42.6, 38.9, 35.7, 35.1,
33.8, 30.9, 30.7, 30.3,28.2, 27.6, 25.9, 24.5, 23.4, 21.8, 18.8, 12.0, 11.9.
HRMS-ESI m/z 447.3116 [M + H+], C27H43O5 requires 447.3110.
6α-Ethyl-3α-formyloxy-7-keto-24-nor-5β-cholan-23-nitrile
(9). Crude 8 (930 mg, 2.08 mmol), 6.7 mL of cold trifluoroacetic acid,
and 1.8 mL (15.6 mmol) of trifluoroacetic anhydride were stirred at
0−5 °C until dissolution. Sodium nitrite (435 mg, 6.3 mmol) was
added in small portions. After the addition was complete, the reaction
mixture was stirred first at 0−5 °C for 1 h, then at 38−40 °C for 2 h.
On completion, the mixture was neutralized with NaOH, 2 N, and
then the product was extracted with 50 mL of diethyl ether (3 × 50
mL), followed by washing with brine and dried over anhydrous
Na2SO4. The ether was removed under reduced pressure to afford 860
mg of 9 in quantitative yield, which was subjected to the next step
without any purification.
Methyl 3α,7α-Di(tert-butyldimethylsilyloxy)-6α-ethyl-24-
nor-5β-cholan-23-oate (12). To a solution of 11 (600 mg, 1.4
mmol) in 20 mL of CH2Cl2 at 0 °C were added 2,6-lutidine (14.3
mmol, 1.6 mL) and tert-butyl dimethylsilyltrifluoromethanesulfonate
(4.2 mmol, 960 μL). After the mixture was stirred for 24 h at 0 °C, the
reaction was quenched by addition of aqueous NaHSO4 (1 M, 50 mL).
The layers were separated, and the aqueous phase was extracted with
CH2Cl2 (3 × 50 mL). The combined organic layers were washed with
NaHSO4, water, saturated aqueous NaHCO3, and brine. Purification
by flash chromatography on silica gel using hexane/ethyl acetate, 9:1,
and 0.5% of triethylamine as eluent gave 12 (722 mg, 78%) as a clear,
colorless oil. [α]D25 +2.44 (c 0.62, CHCl3). Selected H NMR (400
1
MHz, CDCl3): δ 3.69 (1H, br s), 3.48 (3H, s), 3.40 (1H, m), 1.02
(3H, d, J = 6.3 Hz), 0.90 (3H, t, J = 7.3 Hz), 0.89 (3H, s), 0.86 (18H,
s), 0.70 (3H, s), 0.042 (12H, s). 13C NMR (100 MHz, CDCl3): δ
176.3, 74.1, 72.9, 56.0, 51.6, 51.6, 49.4, 45.1, 43.2, 42.4, 41.9, 41.5,
40.8, 35.7, 34.9, 34.1, 33.9, 31.8, 31.3, 26.7, 26.6, 25.7 (3C), 25.6 (3C),
24.4, 24.0, 23.0, 19.3, 18.8, 12.0, 11.9, −3.0 (2C), −2.8 (2C). HRMS-
ESI m/z 641.5051 [M + H+], C38H73O4Si2 requires 641.5047.
6α-Ethyl-3α,7α-dihydroxy-24-nor-5β-cholan-23yl-23-triethy-
lammonium Sulfate 1. Dry methanol (133 μL, 3.3 mmol) and
LiBH4 (1.7 mL, 2 M in THF, 3.3 mmol) were added to a solution of
12 (722 mg, 1.1 mmol) in dry THF (15 mL) at 0 °C under argon, and
the resulting mixture was stirred for 8 h at 0 °C. The mixture was
quenched by addition of NaOH (1 M, 2.2 mL) and then allowed to
warm to room temperature. Ethyl acetate was added, and the separated
aqueous phase was extracted with ethyl acetate (3 × 30 mL). The
combined organic phases were washed with water, dried (Na2SO4),
and concentrated. Purification by silica gel (n-hexane/ethyl acetate
8:2) gave C23 alcohol derivative as a colorless oil (650 mg, 95%).
6α-Ethyl-3α-hydroxy-7-keto-24-nor-5β-cholan-23-oic Acid
(10). Crude compound 9 (860 mg, 2.08 mmol) was refluxed with
30% KOH in ∼50 mL of methanol/water, 1:1. After the mixture was
stirred for 48 h, the basic aqueous solution was neutralized with HCl, 6
N. Then methanol was evaporated and the residue was extracted with
AcOEt (3 × 50 mL). The combined organic layers were washed with
brine, dried, and evaporated to dryness to give white solid residue 10
(723 mg, 86%). An analytic sample was obtained by silica gel
chromatography, eluting with CH2Cl2/MeOH 95:5. [α]D25 −25.2 (c
[α]D25 −0.25 (c 0.5, CH3OH). Selected H NMR (400 MHz, CDCl3):
1
1
0.22, CH3OH). Selected H NMR (400 MHz, CD3OD): δ 3.46 (1H,
δ 3.70 (1H, m), 3.67 (1H, br s), 3.60 (1H, m), 3.37 (1H, m), 0.94
(3H, d, J = 6.4 Hz), 0.87 (3H, s), 0.87 (3H, t, ovl), 0.86 (18H, s), 0.66
(3H, s), 0.042 (12H, s). 13C NMR (100 MHz, CDCl3): δ 73.6, 71.2,
61.0, 56.6, 50.8, 45.6, 43.0, 41.5, 40.3, 39.9, 36.0, 35.8, 34.4, 33.5, 33.2
(2C), 31.3, 28.6, 26.2 (7C), 23.9, 23.4, 22.5, 21.0, 19.0, 12.0, 11.8,
−4.3 (2C), −4.4 (2C). HRMS-ESI m/z 621.5093 [M + H+],
C37H73O3Si2 requires 621.5098.
The triethylamine−sulfur trioxide complex (906 mg, 5 mmol) was
added to a solution of C23 alcohol (620 mg, 1 mmol) in DMF dry (7
mL) under an argon atmosphere, and the mixture was stirred at 95 °C
for 24 h. The solution was then concentrated under vacuum. To the
solid dissolved in methanol (30 mL) was added three drops of HCl,
37% v/v, and the mixture was stirred for 5 h at room temperature. At
the end of reaction, silver carbonate was added to precipitate chloride.
Then the reaction mixture was centrifuged and the supernatant was
concentrated in vacuo. The residue was poured over a RP18 column.
The fraction eluted with H2O/MeOH (1:1) gave compound 1 as a
white solid (500 mg, 87%). A sample (20 mg) was subjected to
purification by HPLC on a Nucleodur 100-5 C18 (5 μm, 10 mm i.d. ×
250 mm) with MeOH/H2O (65:35) as eluent (flow rate 3 mL/min
(tR = 24.8 min)). [α]2D5 +2.9 (c 0.17, CH3OH). 1H and 13C NMR data
in CD3OD are given in Table S1. HR ESIMS m/z 471.2775 [M −
Na], C25H43O6S requires 471.2780.
m), 2.82 (1H, dd, J = 6.4, 12.6 Hz), 1.25 (3H, s), 1.25 (3H, d ovl),
0.81 (3H, t, J = 7.3 Hz), 0.74 (3H, s). 13C NMR (100 MHz, CD3OD):
δ 215.7, 177.6, 71.7, 56.3, 53.3, 51.2, 50.6, 50.5, 45.3, 43.8, 42.5, 40.4,
36.8, 35.3, 35.0, 32.6, 30.5, 29.4, 25.6, 23.9, 22.9, 20.9, 19.6, 12.5, 12.3.
HRMS-ESI m/z 405.3008 [M + H+], C25H41O4 requires 405.3005.
Methyl 6α-Ethyl-3α, 7α-dihydroxy-24-nor-5β-cholan-23-
oate (11). Compound 10 (715 mg, 1.7 mmol) was dissolved in a
solution of tetrahydrofuran/water (50 mL, 4/1 v/v) and treated at 0
°C with NaBH4 (320 mg, 8.5 mmol). After 1 h, water and MeOH were
added dropwise during a period of 15 min at 0 °C with effervescence
being observed. Then after evaporation of the solvents, the residue was
diluted with water, acidified with HCl, 1 N, and extracted with AcOEt
(3 × 50 mL). The combined organic phases were washed with brine,
dried over anhydrous Na2SO4, and evaporated under reduced pressure.
The crude residue was purified by flash chromatography on silica gel,
using dichloromethane/methanol, 9:1, as eluent, to afford 640 mg of
3α,7α-dihydroxy-6α-ethyl-24-nor-5β-cholan-23-oic acid (93% yield).
[α]D25 +3.7 (c 1.48, CH3OH). Selected 1H NMR (400 MHz, CD3OD):
δ 3.62 (1H, br s), 3.30 (1H, m ovl), 0.97 (3H, d, J = 7.9 Hz), 0.88 (3H,
s), 0.87 (3H, t ovl), 0.70 (3H, s). 13C NMR (100 MHz, CD3OD): δ
177.8, 73.2, 71.2, 57.5, 51.7, 46.9, 43.9, 43.8, 42.5, 41.6,41.0, 36.7, 35.1,
34.5, 34.4, 31.3, 29.4, 24.6, 23.8, 23.5, 21.9, 19.6, 12.3, 12.0. HRMS-
ESI m/z 407.3166 [M + H+], C25H43O4 requires 407.3161.
The above compound (640 mg, 1.6 mmol) was dissolved in 50 mL
of dry methanol and treated with p-toluenesulfonic acid (1.5 g, 8.0
mmol). The solution was left to stand at room temperature overnight.
The mixture was quenched by addition of NaHCO3 solution until
neutrality. Most of the solvent was evaporated, and the residue was
extracted with EtOAc. The combined extract was washed with brine,
Compound 13. To a solution of 11 (30 mg, 0.07 mmol) in dry
THF (5 mL) at 0 °C, dry methanol (20 μL, 0.5 mmol) and LiBH4
(250 μL, 2 M in THF, 0.5 mmol) were added. The resulting mixture
was stirred for 8 h at 0 °C. The mixture was quenched by addition of
NaOH (1 M, 240 μL), and then ethyl acetate was added. The
separated aqueous phase was extracted with ethyl acetate (3 × 30 mL).
947
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
The combined organic phases were washed with water, dried
(Na2SO4), and concentrated. Purification by silica gel (CH2Cl2/
MeOH 9:1) gave triol derivative as a white solid (27 mg, quantitative
5β-Cholan-3α,7α,24-tryl-3,7,24-sodium Trisulfate (20). The
triethylamine−sulfur trioxide complex (91 mg, 0.5 mmol) was added
to a solution of triol 16 (40 mg, 0.1 mmol) in dry DMF (3 mL) under
an argon atmosphere, and the mixture was stirred at 95 °C for 3 h.
Most of the solvent was evaporated, and the residue was poured over a
RP18 column to remove excess SO3·NEt3. The fraction eluted with
H2O/MeOH, 9:1, gave a mixture that was further purified by HPLC
on a Nucleodur 100-5 C18 (5 μm, 4.6 mm i.d. × 250 mm) with
MeOH/H2O (35:65) as eluent (flow rate 1 mL/min) to give 20 (44
1
yield). [α]D25 +1.97 (c 0.5, CH3OH). Selected H NMR (400 MHz,
CDCl3): δ 3.67 (1H, br s), 3.60 (2H, m), 3.43 (1H, m), 0.93 (3H, d, J
= 6.6 Hz), 0.88 (3H, s), 0.87 (3H, t, J = 7.1 Hz), 0.67 (3H, s). 13C
NMR (100 MHz, CDCl3): δ 73.2, 71.2, 60.8, 57.9, 51.7, 51.6, 46.9,
43.9, 43.1, 41.5, 41.0, 39.8, 36.7, 36.6, 34.5, 34.4, 34.2, 31.2, 29.3, 24.5,
23.7, 23.5, 21.9, 19.5, 19.3, 12.2, 12.0. HRMS-ESI m/z 393.3367 [M +
H+], C25H45O3 requires 393.3369.
mg, tR = 8 min, 65%). [α]2D5 +4.78 (c 0.12, CH3OH). H NMR (400
1
Triethylamine−sulfur trioxide complex (127 mg, 0.7 mmol) was
added to triol (27 mg, 0.07 mmol) in DMF dry (10 mL) under an
argon atmosphere, and the mixture was stirred at 95 °C for 48 h. The
reaction mixture was quenched with water (1.6 mL), and the solution
was poured over a C18 silica gel column to remove excess Et3N·SO3.
The fraction eluted with H2O/MeOH (9:1) gave a mixture, which was
further purified by HPLC on a Nucleodur 100-5 C18 (5 μm, 10 mm
i.d. × 250 mm) with MeOH/H2O (35:65) as eluent (flow rate 3 mL/
min) to give compound 13 as a white solid (47 mg, 72%). [α]D25 −2.3
MHz, CD3OD): δ 4.45 (1H, br s), 4.15 (1H, m), 3.95 (2H, t, J = 6.5
Hz), 2.35 (1H, d, J = 12.5 Hz), 2.3 (1H, d, J = 14.0 Hz), 0.97 (3H, d
ovl), 0.97 (3H, s), 0.70 (3H, s). HR ESIMS m/z 661.1395 [M − Na],
C24H39Na2O12S3 requires 661.1399.
The same sequence of reactions and purification procedure was
carried out on 6α-ethylchenodeoxycholic acid (15) to give a mixture of
compounds 19 and 21. Purification by HPLC on a Nucleodur 100-5
C18 (5 μm, 4.6 mm i.d. × 250 mm) with MeOH/H2O (65:35) as
eluent (flow rate 1 mL/min) gave compound 19 (tR = 22.0 min) as a
white solid (15 mg). Purification by HPLC on a Nucleodur 100-5 C18
(5 μm, 4.6 mm i.d. × 250 mm) with MeOH/H2O (35:65) as eluent
(flow rate 1 mL/min) gave compound 21 (tR = 24.0 min) as a white
solid (27 mg).
1
(c 0.24, CH3OH). Selected H NMR (400 MHz, CD3OD): δ 4.62
(1H, br s), 4.11 (1H, m), 4.03 (2H, m), 3.19 (18H, q, J = 7.2 Hz), 1.30
(27H, t, J = 7.2 Hz), 0.99 (3H, d, J = 6.5 Hz), 0.96 (3H, s), 0.93 (3H,
t, J = 7.4 Hz), 0.71 (3H, s). HR ESIMS m/z 833.4322 [M − Et3NH],
C37H73N2O12S3 requires 833.4326.
6α-Ethyl-3α,7α-dihydroxy-5β-cholan-24-yl-24-sodium Sul-
fate (19). [α]D25 +3.18 (c 0.51, CH3OH). Selected H NMR (400
1
Chenodeoxycholan Sulfate Derivative Synthesis. Methyl
3α,7α-Dihydroxy-5β-cholan-24-oate (14). A mixture of CDCA
(3, 100 mg, 0.25 mmol) and p-toluenesulfonic acid (237 mg, 1.25
mmol) in dry methanol (10 mL) was left to stand at room
temperature for 1 h. The mixture was quenched by addition of
NaHCO3 solution until neutrality. Most of the solvent was evaporated,
and the residue was extracted with EtOAc. The combined extract was
washed with brine, dried with Na2SO4, and evaporated to give the
desired methyl ester as colorless amorphous solids (102 mg,
quantitative yield). An analytic sample was obtained by silica gel
chromatography, eluting with CH2Cl2/MeOH 95:5. [α]D25 +7.81 (c
MHz, CD3OD): δ 3.96 (2H, t, J = 6.3 Hz), 3.65 (1H, br s), 3.30 (1H,
m ovl), 0.97 (3H, d, J = 6.5 Hz), 0.92 (3H, s), 0.90 (3H, t, J = 7.0 Hz),
0.70 (3H, s). 13C NMR (100 MHz, CD3OD): δ 72.2, 70.9, 69.6, 56.4,
50.5, 45.9, 42.5, 41.1, 39.9, 39.6, 35.7, 35.6 (2C), 35.4, 33.4, 32.0, 31.9,
29.3, 28.2, 23.6 (2C), 23.1, 20.5, 18.9, 12.2, 11.9. HR ESIMS m/z
485.2935 [M − Na], C26H45O6S requires 485.2937.
6α-Ethyl-5β-cholan-3α,7α,24-tryl-3,7, 24-sodium Trisulfate
(21). [α]D25 −1.9 (c 0.23, CH3OH). Selected H NMR (400 MHz,
1
CD3OD): δ 4.55 (1H, br s), 4.09 (1H, m), 3.96 (2H, t, J = 6.0 Hz),
0.97 (3H, d, J = 6.4 Hz), 0.92 (3H, s), 0.90 (3H, t, J = 7.0 Hz), 0.70
(3H, s). HR ESIMS m/z 689.1715 [M − Na], C26H43Na2O12S3
requires 689.1712.
1
2.33, CHCl3). Selected H NMR (400 MHz, CDCl3): δ 3.73 (1H, br
s), 3.57 (3H, s), 3.30 (1H, m ovl), 2.26 (1H, m), 2.13 (1H, m), 0.84
(3H, d, J = 6.0 Hz), 0.82 (3H, s), 0.60 (3H, s). 13C NMR (100 MHz,
CDCl3): δ 174.4, 71.2, 67.6, 55.3, 51.0, 49.8, 42.0, 41.1, 39.2, 38.8
(2C), 34.9, 34.8, 34.5, 34.2, 32.3, 30.5, 30.4, 29.9, 27.7, 23.0, 22.3, 20.1,
17.7, 11.2. HRMS-ESI m/z 407.3165 [M + H+], C25H43O4 requires
407.3161.
Bis-homochenodeoxycholan Sulfate Derivatives. 3α,7α-Di-
(tert-butyldimethylsilyloxy)-5β-cholan-24-ol (22). 2,6-Lutidine
(290 μL, 2.5 mmol) and tert-butyl dimethylsilyltrifluoromethane-
sulfonate (171 μL, 0.75 mmol) were added at 0 °C to a solution of
methyl ester 14 (100 mg, 0.25 mmol) in 10 mL of CH2Cl2. After the
mixture was stirred for 2 h at 0 °C, the reaction was quenched by
addition of aqueous NaHSO4 (1 M, 50 mL). The layers were
separated, and the aqueous phase was extracted with CH2Cl2 (3 × 50
mL). The combined organic layers were washed with NaHSO4, water,
saturated aqueous NaHCO3, and brine. Purification by flash
chromatography on silica gel using hexane/ethyl acetate, 99:1, and
0.5% of triethylamine as eluent gave the corresponding methyl ester
(131 mg, 83%) as a clear, colorless oil. To a solution of methyl ester
(100 mg, 0.16 mmol) in dry THF (5 mL) at 0 °C were added dry
methanol (45 μL, 1.12 mmol) and LiBH4 (560 μL, 2 M in THF, 1.12
mmol). The resulting mixture was stirred for 3 h at 0 °C. The mixture
was quenched by addition of 1 M NaOH (320 μL) and then ethyl
acetate. The organic phase was washed with water, dried (Na2SO4),
and concentrated. Purification by silica gel (hexane/ethyl acetate 8:2)
gave compound 22 as a white solid (97 mg, quantitative yield). [α]2D5
+7.47 (c 0.12, CHCl3). Selected 1H NMR (400 MHz, CDCl3): δ 3.82
(1H, br s), 3.47 (2H, t, J = 6.5 Hz), 3.41 (1H, m), 0.93 (3H, d, J = 6.0
Hz), 0.88 (9H, s), 0.88 (3H, s ovl), 0.86 (9H, s), 0.65 (3H, s), 0.08
(6H, s), 0.04 (6H, s). 13C NMR (100 MHz, CDCl3): δ 72.8, 69.7,
63.4, 56.3, 50.1, 42.5, 42.0, 40.7, 40.6, 39.8, 36.0, 35.8, 35.2, 34.8, 32.4,
32.2, 31.2, 29.7 (3C), 28.3, 26.3 (3C), 26.0 (3C), 24.1, 23.0, 20.7, 18.8,
12.0, −2.2, −4.4, −4.5, −5.4. HRMS-ESI m/z 607.4946 [M + H+],
C36H71O3Si2 requires 607.4942.
3α,7α,24-5β-Cholantriol (16). Methyl ester 14 (100 mg, 0.25
mmol) was dissolved in dry THF (5 mL) at 0 °C in the presence of
dry methanol (70 μL, 0.84 mmol) and LiBH4 (875 μL, 2 M in THF,
1.75 mmol). After 8 h, a solution of 1 M NaOH (500 μL) was added
and then allowed to warm to room temperature. Ethyl acetate was
added, and the separated aqueous phase was extracted with ethyl
acetate (3 × 30 mL). The combined organic phases were washed with
water, dried (Na2SO4), and concentrated. Purification by silica gel
(CH2Cl2/MeOH 95:5) gave triol derivative 16 as a colorless oil (90
mg, 96%). [α]D25 +0.82 (c 2.07, CH3OH). Selected H NMR (400
1
MHz, CDCl3): δ 3.80 (1H, br s), 3.56 (2H, m), 3.40 (1H, m), 0.91
(3H, d, J = 6.0 Hz), 0.87 (3H, s), 0.63 (3H, s). 13C NMR (100 MHz,
CDCl3): δ 71.8, 68.4, 63.2, 56.0, 50.3, 42.5, 41.5, 39.6, 39.3 (2C), 35.5,
35.3, 34.9, 34.5, 32.8, 30.5 (2C), 29.3, 28.2, 23.6, 22.7, 20.5, 18.6, 11.7.
HRMS-ESI m/z379.3216 [M + H+], C24H43O3 requires 379.3212.
3α,7α-Dihydroxy-5β-cholan-24-yl-24-sodium Sulfate (18). At
a solution of triol 16 (40 mg, 0.1 mmol) in DMF dry (3 mL) was
added triethylamine−sulfur trioxide complex (36 mg, 0.2 mmol) under
an argon atmosphere, and the mixture was stirred at 95 °C for 1 h.
Most of the solvent was evaporated, and the residue was poured over a
RP18 column to remove excess SO3·NEt3. The fraction eluted with
H2O/MeOH 1:1 gave a mixture that was further purified by HPLC on
a Nucleodur 100-5 C18 (5 μm, 4.6 mm i.d. × 250 mm) with MeOH/
H2O (65:35) as eluent (flow rate 1 mL/min) to give 35 mg (73%) of
compound 18 (tR = 11.2 min). [α]D25 +4.93 (c 0.05, CH3OH). Selected
1H NMR (400 MHz, CD3OD): δ 3.95 (2H, t, J = 6.4 Hz), 3.78 (1H,
br s), 3.30 (1H, m ovl), 0.96 (3H, d ovl), 0.95 (3H, s), 0.70 (3H, s).
HR ESIMS m/z 457.2627 [M − Na], C24H41O6S requires 457.2624.
Ethyl 3α, 7α-Di(tert-butyldimethylsilyloxy)-25,26-bis-homo-
5β-chol-24-en-26-oate (23). DMSO (4.18 mL, 0.75 mmol) was
added dropwise for 5 min to a solution of oxalyl chloride (14.7 mL,
0.37 mmol) in dry dichloromethane (5 mL) at −78 °C under argon
atmosphere. After 30 min, a solution of 22 (90 mg, 0.15 mmol) in dry
948
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
CH2Cl2 (2 mL) was added dropwise and the mixture was stirred at
−78 °C. After 30 min, Et3N dry (6.83 mL, 0.75 mmol) was added
dropwise to the solution and the mixture was allowed to warm to
room temperature. After 1 h, the reaction was quenched by addition of
aqueous NaHSO4 (1 M, 50 mL). The layers were separated, and the
aqueous phase was extracted with CH2Cl2 (3 × 50 mL). The
combined organic layers were washed with saturated aqueous
NaHSO4, saturated aqueous NaHCO3, and brine. The organic phase
was then dried over Na2SO4 and concentrated to give the
corresponding aldehyde (90 mg, quantitative yield) as a colorless oil,
which was used without any further purification. To a solution of
aldehyde (0.15 mmol) in THF dry (10 mL) was added LiOH (250
mg, 10.5 mmol) and TEPA (triethylphosphonoacetate, 2.07 mL, 10.5
mmol). The reaction mixture was stirred for 45 min at room
temperature and then quenched with water (10 mL). The mixture was
then extracted with EtOAc (3 × 30 mL), and the organic phase was
concentrated in vacuo. Flash chromatography (hexane/EtOAc, 99:1)
afforded pure 23 (92 mg, 91% over two steps). [α]D25 +3.31 (c 0.5,
triol 25 (28 mg, 0.07 mmol) in DMF dry (3 mL) under an argon
atmosphere, and the mixture was stirred at 95 °C for 3 h. Most of the
solvent was evaporated, and the residue was poured over a RP18
column to remove excess SO3·NEt3. The fraction eluted with H2O/
MeOH 1:1 gave a mixture that was further purified by HPLC on a
Nucleodur 100-5 C18 (5 μm, 4.6 mm i.d. × 250 mm) with MeOH/
H2O (65:35) as eluent (flow rate 1 mL/min) to give 4 mg of
compound 26.
The fraction eluted with H2O/MeOH (9:1) gave a mixture that was
further purified by HPLC (Nucleodur 100-5 C18, 5 μm, 4.6 mm i.d. ×
250 mm), with MeOH/H2O (35:65) as eluent (flow rate 1 mL/min),
to give 33 mg of compound 27.
25,26-Bis-homo-3α,7α-dihydroxy-5β-cholan-26-yl-26-so-
dium Sulfate (26). [α]2D5 +9.13 (c 0.05, CH3OH). Selected 1H NMR
(400 MHz, CD3OD): δ 3.98 (2H, t, J = 6.5 Hz), 3.79 (1H, m), 3.31
(1H, m ovl), 0.95 (3H, d, J = 6.8 Hz), 0.93 (3H, s), 0.63 (3H, s). HR
ESIMS m/z 485.2939 [M − Na], C26H45O6S requires 485.2937.
25,26-Bis-homo-5β-cholan-3α,7α,26-tryl-3,7,26-sodium Tri-
sulfate (27). [α]D25 +1.35 (c 1.13, CH3OH). Selected 1H NMR
(400 MHz, CD3OD): δ 4.45 (1H, br s), 4.13 (1H, m), 3.98 (2H, t, J =
6.5 Hz), 0.95 (3H, d, J = 6.6 Hz), 0.93 (3H, s), 0.69 (3H, s). HR
ESIMS m/z 689.1716 [M − Na], C26H43Na2O12S3requires 689.1712.
Ursodeoxycholan Sulfate and Bis-homoursodeoxycholan
Sulfate Derivatives. 3α,7β-Di(tert-butyldimethylsilyloxy)-5β-
cholan-24-ol (29). To a solution of methyl 3α,7β-dihydroxy-5β-
cholan-24-oate (2 g, 5 mmol) in CH2Cl2 dry (50 mL) was added at 0
°C 2,6-lutidine (49 mmol, 5.72 mL) and tert-butyl dimethylsilyl-
trifluoromethanesulfonate (15 mmol, 3.43 mL). After 2 h, the reaction
was quenched by addition of aqueous NaHSO4, 1 M. The layers were
separated, and the aqueous phase was extracted with CH2Cl2 (3 × 50
mL). The combined organic layers were washed with NaHSO4,
saturated aqueous NaHCO3, and finally water. The organic phase was
then dried over Na2SO4 and concentrated to give the corresponding
methyl ester 28 (3.10 g, quantitative yield) as a colorless oil, which was
used without any further purification. Dry methanol (600 μL, 15
mmol) and LiBH4 (7.50 mL, 2 M in THF, 15 mmol) were added to a
solution of methyl ester (3.10 g, 5 mmol) in dry THF (15 mL) at 0 °C
under argon, and the resulting mixture was stirred for 8 h at 0 °C. The
mixture was quenched by addition of NaOH (1 M, 10 mL) and then
allowed to warm to room temperature. Ethyl acetate was added, and
the separated aqueous phase was extracted with ethyl acetate (3 × 30
mL). The combined organic phases were washed with water, dried
(Na2SO4), and concentrated. Purification by silica gel (n-hexane/ethyl
acetate 99:1) gave compound 29 as a colorless oil (2.40 g, 80%). [α]2D5
+21.7 (c 0.49, CHCl3). Selected 1H NMR (400 MHz, CDCl3): δ 3.61
(2H, m), 3.50 (2H, m), 0.88 (3H, d, ovl), 0.88 (3H, s), 0.88 (9H, s),
0.83 (9H, s), 0.60 (3H, s), 0.003 (6H, s), −0.01 (6H, s). 13C NMR
(100 MHz, CDCl3): δ 72.7, 72.5, 61.0, 55.5, 55.4, 44.0, 43.8, 42.8,
40.0, 39.0, 38.8, 38.0, 37.8, 35.4, 35.1, 34.1, 32.8, 31.8 (2C), 30.8, 28.7,
27.3, 26.4 (3C), 25.9 (3C), 23.5, 21.2, 19.0, 12.1, −2.8 (2C), −4.6
(2C). HRMS-ESI m/z 607.4947 [M + H+], C36H70O3Si2 requires
607.4942.
1
CHCl3). Selected H NMR (400 MHz, CDCl3): δ 6.93 (1H, dt, J =
5.1, 15.6 Hz), 5.78 (1H, d, J = 15.6 Hz), 4.1 (2H, m), 3.77 (1H, br s),
3.38 (1H, m), 0.91 (3H, d ovl), 0.90 (9H, s), 0.90 (3H, s ovl), 0.87
(9H, s), 0.61 (3H, s), 0.06 (6H, s), 0.03 (6H, s). 13C NMR (100 MHz,
CDCl3): δ 178.5, 149.8, 120.8, 72.5, 69.5, 59.9, 55.8, 49.9, 42.3, 41.8,
40.5, 40.4, 39.5, 35.6 (2C), 35.0, 34.6, 34.3, 32.2, 30.9, 29.6, 29.1 (2C),
28.0, 26.0 (3C), 25.8 (3C), 23.8, 22.7, 20.5, 18.4, 14.2, 12.0, −2.4,
−4.5, −4.6, −5.6. HRMS-ESI m/z 675.5207 [M + H+], C40H75O4Si2
requires 675.5204.
Ethyl 25,26-Bis-homo-3α,7α-dihydroxy-5β-cholan-26-oate
(24). Compound 23 (90 mg, 0.13 mmol) and THF dry (15 mL)
were mixed and deoxygenated with flowing nitrogen for 5 min. The
catalyst Pd(OH)2 (5 mg, mmol), 20 wt % on carbon (Degussa type),
was added. The mixture was transferred to a standard PARR apparatus
and flushed with nitrogen and then with hydrogen several times. The
apparatus was shacked under 50 psi of H2. After 2 h, the reaction was
complete. The catalyst was filtered through Celite, and the recovered
filtrate was concentrated under vacuum to afford 92 mg of ethyl ester
in quantitative yield, which was subjected to the next step without any
purification. To the ethyl ester dissolved in ethanol (30 mL) was
added 1 mL of HCl, 37% v/v, and the mixture was stirred for 5 h at
room temperature. At the end of reaction, silver carbonate was added
to precipitate chloride. Then the reaction mixture was centrifuged and
the supernatant was concentrated in vacuo to give the desired ethyl
ester 24 (50 mg, 81% over two steps) as a colorless amorphous solid.
An analytic sample was obtained by silica gel chromatography, eluting
with CH2Cl2/MeOH 95:5. [α]D25 +6.72 (c 0.25, CHCl3). Selected H
1
NMR (400 MHz, CDCl3): δ 4.1 (2H, q, J = 7.0 Hz), 3.8 (1H, br s),
3.38 (1H, m), 2.29 (2H, t, J = 6.7 Hz), 1.24 (3H, t, J = 7.0 Hz), 0.94
(3H, d, J = 6.6 Hz), 0.92 (3H, s), 0.68 (3H, s). 13C NMR (100 MHz,
CDCl3): δ 175.4, 72.8, 69.0, 61.4, 57.4, 51.5, 43.6, 43.1, 41.0, 40.7,
40.3, 37.0, 36.7, 36.5, 35.8, 35.2, 34.8, 34.0, 31.3, 29.3, 26.6, 26.5, 24.6,
23.5, 21.8, 19.3, 14.6, 12.3. HRMS-ESI m/z 449.3635 [M + H+],
C28H49O4 requires 449.3631.
3α,7β-Dihydroxy-5β-cholan-24-yl-24-sodium Sulfate (30).
To a solution of compound 29 (100 mg, 0.16 mmol) in DMF dry
(7 mL) was added 58 mg of triethylamine−sulfur trioxide complex
(0.32 mmol). The mixture was stirred at 95 °C for 4 h. Then the
solution was concentrated under vacuum. To the solid dissolved in
methanol (30 mL) was added three drops of HCl, 37% v/v, and the
mixture was stirred for 3 h at room temperature. At the end of
reaction, silver carbonate was added to precipitate chloride. Then the
reaction mixture was centrifuged and the supernatant was concen-
trated in vacuo. The residue was poured over a RP18 column. The
fraction eluted with H2O/MeOH (7:3) gave 53 mg (69% over two
25,26-Bis-homo-3α,7α,26-5β-cholantriol (25). Ethyl ester 24
(40 mg, 0.09 mmol) was reduced with LiBH4 (2 M in THF dry, 135
μL, 0.27 mmol) and dry MeOH (11 μL, 0.27 mmol) in dry THF at 0
°C for 5 h. The mixture was quenched by addition of 1 M NaOH
solution (180 μL), and then ethyl acetate was added. The organic
phase was washed with water, dried (Na2SO4), and concentrated.
Purification by silica gel (CH2Cl2/MeOH 9:1) gave compound 25 as a
white solid (28 mg, 77%). [α]D25 +15.0 (c 0.32, CH3OH). Selected H
1
NMR (400 MHz, CDCl3): δ 3.82 (1H, br s), 3.61 (2H, t, J = 6.6 Hz),
3.43 (1H, m), 0.89 (3H, d, J = 6.0 Hz), 0.88 (3H, s), 0.64 (3H, s). 13C
NMR (100 MHz, CDCl3): δ 71.8, 68.4, 62.8, 56.0, 50.4, 42.5, 41.4,
39.6 (2C), 39.4, 35.8, 35.6, 35.2, 35.0, 34.5, 32.8 (2C), 30.5, 28.2, 26.1,
25.8, 23.6, 22.7, 20.5, 18.6, 11.6. HRMS-ESI m/z 407.3523 [M + H+],
C26H47O4 requires 407.3525.
25,26-Bis-homo-3α,7α-dihydroxy-5β-cholan-26-yl-26-so-
dium Sulfate (26) and 25,26-Bis-homo-5β-cholan-3α, 7α, 26-
tryl-3,7,26-sodium Trisulfate (27). The triethylamine−sulfur
trioxide complex (132 mg, 0.7 mmol) was added to a solution of
steps) of compound 30. [α]2D5 +15.3 (c 0.97, CH3OH). Selected H
1
NMR (400 MHz, CD3OD): δ 3.96 (2H, t, J = 6.6 Hz), 3.48 (2H, m),
0.97 (3H, d, J = 6.4 Hz), 0.96 (3H, s), 0.71 (3H, s). 13C NMR (100
MHz, CD3OD): δ 72.3 (2C), 69.6, 57.0, 56.4, 44.8, 43.0 (2C), 41.2,
40.7, 37.3 (2C), 36.7, 36.0, 35.5, 32.7, 30.6, 29.8, 27.8, 26.5, 23.3, 22.6,
20.0, 12.0. HR ESIMS m/z 457.2621 [M − Na], C24H41O6S requires
457.2624.
949
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
5β-Cholan-3α,7β,24-tryl-3,7,24-sodium Trisulfate (31), 7β-
Hydroxy-5β-cholan-3α,24-diyl-3,24-sodium Disulfate (32), and
7β,24-Dihydroxy-5β-cholan-3α-yl-3-sodium Sulfate (33). To
compound 29 (100 mg, 0.16 mmol) dissolved in methanol (30 mL)
was added 500 μL of HCl, 37% v/v, and the mixture was stirred for 2 h
at room temperature. At the end of reaction, silver carbonate was
added to precipitate chloride. Then the reaction mixture was
centrifuged and the supernatant was concentrated in vacuo to give
60 mg (96%) of the desired triol as colorless amorphous solids. An
analytic sample was obtained by silica gel chromatography, eluting
with CH2Cl2/MeOH 95:5.
s), 0.05 (6H, s), 0.04 (6H, s). 13C NMR (100 MHz, CDCl3): δ 72.7,
72.5, 63.2, 55.5, 55.1, 43.9, 43.7, 42.8, 40.0, 38.8, 38.0, 37.8, 35.9 (2C),
35.6, 35.2, 34.1, 30.9 (2C), 30.8, 30.3, 29.7, 28.6, 27.3, 26.4 (3C), 25.9
(3C), 23.5, 21.2, 18.8, 12.1, −2.8, −3.3, −4.6 (2C). HRMS-ESI m/z
635.5257 [M + H+], C38H75O3Si2 requires 635.5255.
3α,7β-Dihydroxy-25,26-bis-homo-5β-cholan-26-yl-26-so-
dium Sulfate (36). Compound 36 (180 mg, 75% over two steps) was
synthesized, starting from compound 35 (300 mg, 0.47 mmol), by
procedures analogous to those detailed above for compound 30.
[α]D25+14.7 (c 0.17, CH3OH). Selected 1H NMR (400 MHz, CD3OD):
δ 3.98 (2H, t, J = 6.5 Hz), 3.47 (2H, m), 0.96 (3H, s), 0.95 (3H, d, J =
6.0 Hz), 0.71 (3H, s). HR ESIMS m/z 485.2935 [M − Na],
C26H45O6S requires 485.2937.
25,26-Bis-homo-5β-cholan-3α,7β,26-tryl-3,7,26-sodium Tri-
sulfate (37) and 25,26-Bis-homo-7β-hydroxy-5β-cholan-3α,26-
diyl-3,26-sodium disulfate (38). The triethylamine−sulfur trioxide
complex (214 mg, 1.2 mmol) was added to a solution of compound 36
(60 mg, 0.12 mmol) in DMF dry (3 mL). The mixture was stirred at
95 °C for 5 h. Then the solution was concentrated under vacuum. The
residue was poured over a RP18 column. The fraction eluted with
H2O/MeOH (7:3) gave a mixture that was further purified by HPLC,
on a Nucleodur 100-5 C18 (5 μm, 10 mm i.d. × 250 mm) with
MeOH/H2O (35:65) as eluent (flow rate 3 mL/min), to give 37 and
38, 8 and 5 mg, respectively.
The triethylamine−sulfur trioxide complex (118 mg, 0.65 mmol)
was added to a solution of triol (50 mg, 0.13 mmol) in DMF dry (3
mL). The mixture was stirred at 95 °C for 2 h. Then the solution was
concentrated under vacuum. The residue was poured over a RP18
column. The fraction eluted with H2O/MeOH (99:1) gave a mixture
that was further purified by HPLC on a Nucleodur 100-5 C18 (5 μm,
10 mm i.d. × 250 mm) with MeOH/H2O (35:65) as eluent (flow rate
3 mL/min) to give 13 mg of 31 and 8 mg of 32. The fraction eluted
with H2O/MeOH (1:1) gave a mixture that, purified by HPLC on a
Nucleodur 100-5 C18 (5 μm, 10 mm i.d. × 250 mm) with MeOH/
H2O (65:35) as eluent (flow rate 3 mL/min), afforded 14 mg of 33.
5β-Cholan-3α,7β,24-tryl-3,7,24-sodium Trisulfate (31). [α]2D5
1
+20.6 (c 0.08, CH3OH). H NMR (400 MHz, CD3OD): δ 4.28 (1H,
m), 4.21 (1H, br s), 3.96 (2H, t, J = 6.7 Hz), 0.98 (3H, s), 0.97 (3H, d,
J = 6.8 Hz), 0.71 (3H, s). HR ESIMS m/z 661.1395 [M − Na],
C24H39Na2O12S3 requires 661.1399.
25,26-Bis-homo-5β-cholan-3α,7β,26-tryl-3,7,26-sodium Tri-
sulfate (37). [α]D25 +3.34 (c 0.34, CH3OH). Selected 1H NMR
(400 MHz, CD3OD): δ 4.30 (1H, m), 4.23 (1H, m), 3.98 (2H, t, J =
6.6 Hz), 0.98 (3H, s), 0.95 (3H, d, J = 6.1 Hz), 0.70 (3H, s). HR
ESIMS m/z 689.1715 [M − Na], C26H43Na2O12S3 requires 689.1712.
25,26-Bis-homo-7β-hydroxy-5β-cholan-3α,26-diyl-3,26-so-
7β-Hydroxy-5β-cholan-3α,24-diyl-3,24-sodium Disulfate
(32). [α]D25 +35.7 (c 0.11, CH3OH). H NMR (400 MHz, CD3OD):
1
δ 4.23 (1H, m), 3.96 (2H, t, J = 6.7 Hz), 3.48 (1H, m), 0.98 (3H, s),
0.97 (3H, d, J = 6.8 Hz), 0.71 (3H, s). HR ESIMS m/z 559.2015 [M −
Na], C24H40NaO9S2 requires 559.2011.
7β,24-Dihydroxy-5β-cholan-3α-yl-3-sodium Sulfate (33).
[α]D25 +37.4 (c 0.08, CH3OH). 1H NMR (400 MHz, CD3OD): δ
4.23 (1H, m), 3.50 (2H, m), 3.48 (1H, m), 0.97 (3H, s), 0.96 (3H, d,
ovl), 0.71 (3H, s). HR ESIMS m/z 457.2627 [M − Na], C24H41O6S
requires 457.2624.
1
dium Disulfate (38). [α]2D5 +24.2 (c 0.11, CH3OH). H NMR (400
MHz, CD3OD): δ 4.23 (1H, m), 3.96 (2H, t, J = 6.3 Hz), 3.48 (1H,
m), 0.97 (3H, s), 0.95 (3H, d, J = 6.6 Hz), 0.70 (3H, s). HR ESIMS
m/z 587.2329 [M − Na], C26H44NaO9S2 requires 587.2324.
Transactivation. HepG2 cells were cultured at 37 °C in minimum
essential medium with Earl’s salts containing 10% fetal bovine serum
(FBS), 1% L-glutamine, and 1% penicillin/streptomycin. The trans-
fection experiments were performed using Fugene HD (Promega)
according to manufactured specifications. Cells were plated in a 24-
well plate at 5 × 104 cells/well.
For FXR mediated transactivation, HepG2 cells were transfected
with 75 ng of pSG5-FXR, 75 ng of pSG5-RXR, 100 ng of pCMV-β-
galactosidase, and 250 ng of the reporter vector p(hsp27)-TK-LUC
containing the FXR response element IR1 cloned from the promoter
of heat shock protein 27 (hsp27).
For GP-BAR1 mediated transactivation, HEK-293T cells were
transfected with 200 ng of pGL4.29 (Promega), a reporter vector
containing a cAMP response element (CRE) that drives the
transcription of the luciferase reporter gene luc2P, with 100 ng of
pCMVSPORT6-human GP-BAR1, and with 100 ng of pGL4.70
(Promega), a vector encoding the human Renilla gene.
In control experiments HEK-293T cells were transfected only with
vectors pGL4.29 and pGL4.70 to exclude any possibility that
compounds could activate the CRE in a GP-BAR1 independent
manner. At 24 h post-transfection, cells were stimulated 18 h with
CDCA (3) at 10 μM, 1 (1 and 10 μM), and compounds 2, 3, 13, 18−
21, 26, 27, 30−33, and 36−38, 10 μM; in another experimental
setting, cells were primed with the combination of CDCA (3) at 10
μM and compounds 13, 18, 20, 21, 27, 30−33, and 36−38 (50 μM).
After treatments, cells were lysed in 100 μL of lysis buffer (25 mM
Tris-phosphate, pH 7.8; 2 mM DTT; 10% glycerol; 1% Triton X-100),
and 20 μL of cellular lysate was assayed for luciferase activity using the
luciferase assay system (Promega). Luminescence was measured using
Glomax 20/20 luminometer (Promega). For FXR mediated trans-
activation, luciferase activities were normalized for transfection
efficiencies by dividing the luciferase relative light units by β-
galactosidase activity expressed from cells co-transfected with
pCMVβgal. For GP-BAR1 mediated transactivation, luciferase
activities were normalized with Renilla activities.
Ethyl 3α,7β-Di(tert-butyldimethylsilyloxy)-25,26-bis-homo-
5β-chol-24-en-26-oate (34). Compound 34 (450 mg, 40% over
two steps) was synthesized, starting from compound 29 (1.0 g, 1.6
mmol), by analogous procedures to those detailed above for
compound 23. [α]D25 +11.7 (c 0.58, CHCl3). H NMR (400 MHz,
1
CDCl3): δ 6.9 (1H, dt, J = 7.3, 15.5 Hz), 5.78 (1H, d, J = 15.5 Hz), 4.1
(2H, q, J = 7.3 Hz), 3.64 (1H, m), 3.50 (1H, br s), 1.26 (3H, t, J = 7.3
Hz), 0.91 (3H, d, ovl), 0.90 (3H, s), 0.86 (18H, s), 0.62 (3H, s), 0.04
(6H, s), 0.03 (6H, s). 13C NMR (100 MHz, CDCl3): δ 167.0, 150.0,
120.9, 72.7, 72.5, 60.0, 55.5, 54.9, 43.9, 43.7, 42.8, 39.9, 38.8, 37.9,
37.8, 35.3 (2C), 35.1, 34.3, 34.1, 30.9 (2C), 30.8, 28.9, 27.3, 26.4 (3C),
25.9 (3C), 23.5, 21.2, 18.6, 14.3, 12.1, −2.8, −3.4, −4.6 (2C). HRMS-
ESI m/z 675.5207 [M + H+], C40H75O4Si2 requires 675.5204.
3α,7β-Di(tert-butyldimethylsilyloxy)-25,26-bis-homo-5β-
cholan-26-ol (35). A solution of compound 34 (400 mg, 0.6 mmol)
in THF dry/MeOH dry (10 mL/10 mL, v/v) was hydrogenated in the
presence of palladium, 5 wt %, on activated carbon (10 mg). The flask
was evacuated and flushed first with argon and then with hydrogen.
After 2 h, the reaction was complete. The catalyst was filtered through
Celite, and the recovered filtrate was concentrated under vacuum to
give the ethyl ester, which was used without any further purification.
Dry methanol (73 μL, 1.8 mmol) and LiBH4 (900 μL, 2 M in THF,
1.8 mmol) were added to a solution of ethyl ester (400 mg, 0.6 mmol)
in dry THF (15 mL) at 0 °C under argon, and the resulting mixture
was stirred for 8 h at 0 °C. The mixture was quenched by addition of 1
M NaOH (1.2 mL) and then allowed to warm to room temperature.
Ethyl acetate was added, and the separated aqueous phase was
extracted with ethyl acetate (3 × 30 mL). The combined organic
phases were washed with water, dried (Na2SO4), and concentrated.
Purification by silica gel (n-hexane/ethyl acetate 99:1) gave compound
35 as a colorless oil (334 mg, 88% over two steps). [α]D25 +9.3 (c 0.4,
CHCl3). 1H NMR (400 MHz, CDCl3): δ 3.75 (2H, m), 3.65 (1H, m),
3.52 (1H, m), 0.92 (3H, s), 0.89 (3H, d, ovl), 0.89 (18H, s), 0.64 (3H,
950
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
To realize dose−response curves for compound 19, cells were
transfected as described previously and then treated with increasing
concentrations of 19 (0.01, 0.1, 1, and 10 μM). Luciferase activities
were assayed and normalized as described above in order to calculate
EC50.
GLUTAg cells were kindly donated by Dr. D. J. Drucker, Banting
and Best Diabetes Centre, University of Toronto, Toronto General
Hospital, 200 Elizabeth Street MBRW-4R402, Toronto, Canada M5G
2C4.
Docking conformations were clustered on the basis of their rmsd
(tolerance = 2.0 Å) and were ranked based on the AutoDock scoring
function.45
Molecular Dynamics. The same computational protocol was used
to set up both the apo GP-BAR1-membrane and the GP-BAR1/1-
membrane complexes. A 94 Å × 94 Å (in x and y axes) pre-
equilibrated POPC phospholipid bilayer was initially generated using
the membrane-builder tool of CHARMM-GUI.org (http://www.
charmm-gui.org). In order to place the receptor into the bilayer, a hole
was generated, and all lipids in close contact (<1 Å distance from any
protein atoms) were deleted. Each complex was solvated and
neutralized using the solvation and autoionize modules of VMD
1.9.1.64 The TIP3 water model65 was used to treat the solvent. The
ff 99SBildn,66,67 gaff,68 and lipid1169 Amber force fields were used to
parametrize the protein, the ligand, and the lipids, respectively. All the
simulations were performed with the NAMD 2.8 MD code.70 A 10 Å
cutoff (switched at 8.0 Å) was used for atom−pair interactions. The
long-range electrostatic interactions were computed by means of the
particle mesh Ewald (PME) method71 using a 1.0 Å grid spacing in
periodic boundary conditions. The SHAKE algorithm was applied to
constrain bonds involving hydrogen atoms, and thus an integration 2 fs
time step interval could be used. Amber charges were applied to the
proteins and water molecules, whereas the ligand charges were
computed using the restrained electrostatic potential (RESP) fitting
procedure.72 The ESP was first calculated by means of the Gaussian 03
package73 using a 6-31G* basis set at Hartree−Fock level of theory,
and then the RESP charges were obtained by a two-stage fitting
procedure using Antechamber.74 The system was equilibrated in the
NPT ensemble using a target temperature and pressure of 300 K and 1
atm, respectively. Harmonic constraints were applied to the protein
and the ligand, which were gradually released along the equilibration
process.
FXR−/− mice were originally donated by Dr. F. Gonzalez,
Laboratory of Metabolism, Center for Cancer Research, National
Cancer Institute, National Institutes of Health, Bethesda, Maryland
20892, U.S., as previously reported.55
Cells were plated in a 24-well plate at 5 × 104 cells/well, and the
transfection experiments were performed using Fugene HD
(Promega) according to manufactured specifications.
Real-Time PCR. PCR was performed using Primers, and
experimental conditions were as described previously.52
cAMP and GLP-1 Measurement. cAMP and GLP-1 release in
GLUTAg cells were measured as described previously.20
Homology Modeling. The crystal structure of human adenosine
A2A receptor (PDB code 2ydo)44 was used as template to build the
homology model of human GP-BAR1. First a multiple alignment of
the sequence of different GPCRs (adenosine receptors, histamine H2
receptor, sphingosine 1-phosphate receptors, cannabinoid receptors,
muscarinic acetylcholine receptors, dopamine receptor, rhodopsin,
etc.) was performed using ClustalW56,57 to identify the most
conserved regions along the GP-BAR1 sequence (Swissprot ID code
q8tdu6). Then we chose as template the agonist conformation of the
adenosine A2A receptor that shows one of the highest sequence
identity and similarity values, 20% and 55%, respectively (Figure S2).
The alignment between the model (GP-BAR1) and the template
(A(2A)R-GL31) primary sequence was refined considering highly
conserved amino acid residues among GPCRs. In particular, in the
extracellular region, where the ligand binding site is located, the
presence of the disulfide bond between Cys85 and Cys155 of the
extracellular loop II (EL2) and the transmembrane helix III (TM3),
respectively, was considered. Finally, using Modeller, version 9.11,58
we built the tridimensional structure of the GP-BAR1 receptor.
The homology model of GP-BAR1 was embedded in a POPC
phospholipids bilayer to mimic the physiological environment and
submitted first to minimization and then to a 100 ns long MD
simulation to check its stability (see section on molecular dynamics
methods). A geometric validation of the tridimensional structure of the
receptor used in the docking calculations was carried out using the
MolProbity server (http://molprobity.biochem.duke.edu). The model
shows ∼94% of all residues in favored conformational regions and
100% in allowed ones according to the Ramachandran analysis and 0%
of backbone bonds and angles with bad values (Figure S4).
Molecular Docking. Molecular docking calculations in the rFXR-
LBD X-ray structure (PDB code 1osv)51 and in the three-dimensional
model of hGP-BAR1 were carried out using the AutoDock4.2 software
package.45 Ligand tridimensional structures were generated with the
Maestro build panel.59 For each ligand, an extensive ring conforma-
tional sampling was performed with the OPLS-AA force field60 and a
2.0 Å rmsd cutoff using MacroModel (version 9.9)61 as implemented
in Maestro 9.3.59 All conformers were then refined using LigPrep62 as
implemented in Maestro 9.3.59 Protonation states at pH 7.0 were
assigned using Epik.63 Protein structure was prepared through the
Protein Preparation Wizard through the graphical user interface of
Maestro 9.3.59 Water molecules were removed, and hydrogen atoms
were added and minimized using the OPLS-2005 force field.60 Ligands
and receptor structures were converted to AD4 format files using
ADT, and the Gasteiger−Marsili partial charges were then assigned.
Grid points of 70 × 70 × 70 for FXR and of 65 × 80 × 55 for GP-
BAR1 with a 0.375 Å spacing were calculated around the binding
cavity using AutoGrid4.2. Thus, 100 separate docking calculations
were performed for each run. Each docking run consisted of 10 million
energy evaluations using the Lamarckian genetic algorithm local search
(GALS) method. Otherwise default docking parameters were applied.
Production runs were performed in NPT conditions at 1 atm and
300 K.
All the residue labels were taken from crystal structure of rFXR-
LBD with PDB code 1osv and the wild-type amino acidic sequence of
human GP-BAR1.
All figures were rendered using PyMOL (http://www.pymol.org).
The tridimensional model of GP-BAR1 is available upon request.
ASSOCIATED CONTENT
* Supporting Information
■
S
Tabulated NMR data for 1; activation of cAMP responsive gene
in absence of GP-BAR1 (Figure S1); alignment GP-BAR1/
template A(2A)R-GL31 (Figure S2); protein backbone rmsd of
GP-BAR1 model (Figure S3); Ramachandran plot of the GP-
BAR1 structure (Figure S4); binding modes of 6-ECDCA in
GP-BAR1 and FXR (Figure S5); tridimensional movie showing
the stability of the binding mode of 1 to GP-BAR1; NMR
proton spectra for all tested compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Authors
*A.Z.: phone, +39 081 678525; fax, +39 081 678552; e-mail,
angela.zampella@unina.it.
■
*V.L.: phone, +39 081 678 641; e-mail, vittoriolimongelli@
gmail.com.
Author Contributions
∥C.D., F.S.D.L., and V.S. contributed equally to this work.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
This work was supported by grants from MAREX-Exploring
Marine Resources for Bioactive Compounds: From Discovery
■
951
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
pregnane X receptor and constitutive androstane receptor. Biochem.
Pharmacol. 2009, 77, 689−699.
(12) Takeda, S.; Kadowaki, S.; Haga, T.; Takaesu, H.; Mitaku, S.
Identification of G protein-coupled receptor genes from the human
genome sequence. FEBS Lett. 2002, 520, 97−101.
(13) Maruyama, T.; Miyamoto, Y.; Nakamura, T.; Tamai, Y.; Okada,
H.; Sugiyama, E.; Nakamura, T.; Itadani, H.; Tanaka, K. Identification
of membrane type receptor for bile acids (M-BAR). Biochem. Biophys.
Res. Commun. 2002, 298, 714−719.
(14) Kawamata, Y.; Fujii, R.; Hosoya, M.; Harada, M.; Yoshida, H.;
Miwa, M.; Fukusumi, S.; Habata, Y.; Itoh, T.; Shintani, Y.; Hinuma, S.;
Fujisawa, Y.; Fujino, M. A G protein-coupled receptor responsive to
bile acids. J. Biol. Chem. 2003, 278, 9435−9440.
(15) Watanabe, M.; Houten, S. M.; Mataki, C.; Christoffolete, M. A.;
Kim, B. W.; Sato, H.; Messaddeq, N.; Harney, J. W.; Ezaki, O.;
Kodama, T.; Schoonjans, K.; Bianco, A. C.; Auwerx, J. Bile acids
induce energy expenditure by promoting intracellular thyroid hormone
activation. Nature 2006, 439, 484−489.
(16) Thomas, C.; Gioiello, A.; Noriega, L.; Strehle, A.; Oury, J.;
Rizzo, G.; Macchiarulo, A.; Yamamoto, H.; Mataki, C.; Pruzanski, M.;
Pellicciari, R.; Auwerx, J.; Schoonjans, K. TGR5-mediated bile acid
sensing controls glucose homeostasis. Cell Metab. 2009, 10, 167−177.
(17) Pellicciari, R.; Sato, H.; Gioiello, A.; Costantino, G.;
Macchiarulo, A.; Sadeghpour, B. M.; Giorgi, G.; Schoonjans, K.;
Auwerx, J. Nongenomic actions of bile acids. Synthesis and preliminary
characterization of 23- and 6,23-alkyl-substituted bile acid derivatives
as selective modulators for the G-protein coupled receptor TGR5. J.
Med. Chem. 2007, 50, 4265−4268.
(18) Wang, Y. D.; Chen, W. D.; Yu, D.; Forman, B. M.; Huang, W.
The G-protein-coupled bile acid receptor, Gpbar1 (TGR5), negatively
regulates hepatic inflammatory response through antagonizing nuclear
factor kappa light-chain enhancer of activated B cells (NF-κB) in mice.
Hepatology 2011, 54, 1421−1432.
(19) Cipriani, S.; Mencarelli, A.; Chini, M. G.; Distrutti, E.; Renga,
B.; Bifulco, G.; Baldelli, F.; Donini, A.; Fiorucci, S. The bile acid
receptor GPBAR-1 (TGR5) modulates integrity of intestinal barrier
and immune response to experimental colitis. PLoS One 2011, 6,
e25637.
to Sustainable Production and Industrial Applications (Call
FP7-KBBE-2009-3, Project No. 245137), MIUR-PRIN 2010/
2011 (Grant E61J12000210001), Swiss National Supercomput-
ing Centre (CSCS, Project s386), and from PSC Partners
foundation.
ABBREVIATIONS USED
■
AD4.2, AutoDock4.2; BA, bile acid; BSEP, bile salt export
pump; CAR, constitutive androstane receptor; CDCA,
chenodeoxycholic acid; 6-ECDCA, 6-ethylchenodeoxycholic
acid; FXR, farnesoid X receptor; GLP-1, glucagon-like peptide
1; GPCR, G-protein-coupled receptor; GP-BAR-1, G-protein-
coupled bile acid receptor 1; HepG2, human hepatoma cell
line; LBD, ligand binding domain; LCA, lithocholic acid; LXR,
liver X receptor; M-BAR/TGR5, membrane-type bile acid
receptor/Takeda G-protein-coupled receptor; MD, molecular
dynamics; NR, nuclear receptor; OSTα, organic solute
transporter α; PXR, pregnane X receptor; RXR, retinoid X
receptor; RT-PCR, real-time polymerase chain reaction; SHP,
small heterodimer partner; TLCA, taurolithocolic acid; UDCA,
ursodeoxycholic acid; VDR, vitamin D receptor; HEK-293T,
human embryonic kidney 293T
REFERENCES
■
(1) de Aguiar Vallim, T. Q.; Tarling, E. J.; Edwards, P. A. Pleiotropic
roles of bile acids in metabolism. Cell Metab. 2013, 17, 657−669.
(2) Makishima, M.; Okamoto, A. Y.; Repa, J. J .; Tu, H.; Learned, R.
M.; Luk, A.; Hull, M. V.; Lustig, K. D.; Mangelsdorf, D. J.; Shan, B.
Identification of a nuclear receptor for bile acids. Science 1999, 284,
1362−1365.
(3) 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.
(4) Zhang, Y.; Lee, F. Y.; Barrera, G.; Lee, H.; Vales, C.; Gonzalez, F.
J.; Willson, T. M.; Edwards, P. A. Activation of the nuclear receptor
FXR improves hyperglycemia and hyperlipidemia in diabetic mice.
Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 1006−1011.
(5) Renga, B.; Mencarelli, A.; D’Amore, C.; Cipriani, S.; Baldelli, F.;
Zampella, A.; Distrutti, E.; Fiorucci, S. Glucocorticoid receptor
mediates the gluconeogenic activity of the farnesoid X receptor in
the fasting condition. FASEB J. 2012, 26, 3021−3031.
(6) Wang, Y. D.; Chen, W. D.; Wang, M.; Yu, D.; Forman, B. M.;
Huang, W. Farnesoid X receptor antagonizes nuclear factor κB in
hepatic inflammatory response. Hepatology 2008, 48, 1632−1643.
(7) Fiorucci, S.; Rizzo, G.; Antonelli, E.; Renga, B.; Mencarelli, A.;
Riccardi, L.; Orlandi, S.; Pruzanski, M.; Morelli, A.; Pellicciari, R. A
farnesoid X receptor-small heterodimer partner regulatory cascade
modulates tissue metalloproteinase inhibitor-1 and matrix metal-
loprotease expression in hepatic stellate cells and promotes resolution
of liver fibrosis. J. Pharmacol. Exp. Ther. 2005, 314, 584−595.
(8) Baptissart, M.; Vega, A.; Maqdasy, S.; Caira, F.; Baron, S.;
Lobaccaro, J. M.; Volle, D. H. Bile acids: from digestion to cancers.
Biochimie 2013, 95, 504−517.
(9) Staudinger, J. L.; Goodwin, B.; Jones, S. A.; Hawkins-Brown, D.;
MacKenzie, K. I.; LaTour, A.; Liu, Y.; Klaassen, C. D.; Brown, K. K.;
Reinhard, J.; Willson, T. M.; Koller, B. H.; Kliewer, S. A. The nuclear
receptor PXR is a lithocholic acid sensor that protects against liver
toxicity. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 3369−3374.
(10) Makishima, M.; Lu, T. T.; Xie, W.; Whitfield, G. K.; Domoto,
H.; Evans, R. M.; Haussler, M. R.; Mangelsdorf, D. J. Vitamin D
receptor as an intestinal bile acid sensor. Science 2002, 296, 1313−
1316.
(20) Cipriani, S.; Mencarelli, A.; Bruno, A.; Renga, B.; Distrutti, E.;
Santucci, L.; Baldelli, F.; Fiorucci, S. Activation of the bile acid receptor
GPBAR1 protects against gastrointestinal injury caused by non-
steroidal anti-inflammatory drugs and aspirin in mice. Br. J. Pharmacol.
2013, 168, 225−237.
(21) Fiorucci, S.; Mencarelli, A.; Palladino, G.; Cipriani, S. Bile-acid-
activated receptors: targeting TGR5 and farnesoid-X-receptor in lipid
and glucose disorders. Trends Pharmacol. Sci. 2009, 30, 570−580.
(22) Fiorucci, S.; Cipriani, S.; Baldelli, F.; Mencarelli, A. Bile acid-
activated receptors in the treatment of dyslipidemia and related
disorders. Prog. Lipid Res. 2010, 49, 171−185.
(23) Cipriani, S.; Mencarelli, A.; Palladino, G.; Fiorucci, S. FXR
activation reverses insulin resistance and lipid abnormalities and
protects against liver steatosis in Zucker (fa/fa) obese rats. J. Lipid Res.
2010, 51, 771−784.
(24) Baghdasaryan, A.; Claudel, T.; Gumhold, J.; Silbert, D.; Adorini,
L.; Roda, A.; Vecchiotti, S.; Gonzalez, F. J.; Schoonjans, K.;
Strazzabosco, M.; Fickert, P.; Trauner, M. Dual farnesoid X
receptor/TGR5 agonist INT-767 reduces liver injury in the Mdr2−/
− (Abcb4−/−) mouse cholangiopathy model by promoting biliary
HCO− output. Hepatology 2011, 54, 1303−1312.
3
(25) Rizzo, G.; Passeri, D.; De Franco, F.; Ciaccioli, G.; Donadio, L.;
Rizzo, G.; Orlandi, S.; Sadeghpour, B.; Wang, X. X.; Jiang, T.; Levi, M.;
Pruzanski, M.; Adorini, L. Functional characterization of the
semisynthetic bile acid derivative INT-767, a dual farnesoid X receptor
and TGR5 agonist. Mol. Pharmacol. 2010, 78, 617−630.
(26) Pellicciari, R.; Fiorucci, S.; Pruzanski, M. Preparation of Bile
Acid Derivatives as FXR Ligands for the Prevention or Treatment of
FXR-Mediated Diseases or Conditions. PCT Int. WO 2008002573 A2
20080103, 2008.
(11) Kohle, C.; Bock, K. W. Coordinate regulation of human drug-
metabolizing enzymes, and conjugate transporters by the Ah receptor,
952
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
(27) Sepe, V.; Ummarino, R.; D’Auria, M. V.; Chini, M. G.; Bifulco,
G.; Renga, B.; D’Amore, C.; Debitus, C.; Fiorucci, S.; Zampella, A.
Conicasterol E, a small heterodimer partner sparing farnesoid X
receptor modulator endowed with a pregnane X receptor agonistic
activity, from the marine sponge Theonella swinhoei. J. Med. Chem.
2012, 55, 84−93.
(28) Gioiello, A.; Macchiarulo, A.; Carotti, A.; Filipponi, P.;
Costantino, G.; Rizzo, G.; Adorini, L.; Pellicciari, R. Extending SAR
of bile acids as FXR ligands: discovery of 23-N-(carbocinnamyloxy)-
3α,7α-dihydroxy-6α-ethyl-24-nor-5β-cholan-23-amine. Bioorg. Med.
Chem. 2011, 19, 2650−2658.
(29) Sepe, V.; Ummarino, R.; D’Auria, M. V.; Renga, B.; Fiorucci, S.;
Zampella, A. The first total synthesis of solomonsterol B, a marine
pregnane X receptor agonist. Eur. J. Org. Chem. 2012, 5187−5194.
(30) Tserng, K. Y.; Klein, P. D. Formylated bile acids: improved
synthesis, properties, and partial deformylation. Steroids 1977, 29,
635−648.
(31) Schteingart, C. D.; Hofmann, A. F. Synthesis of 24-nor-5 beta-
cholan-23-oic acid derivatives: a convenient and efficient one-carbon
degradation of the side chain of natural bile acids. J. Lipid Res. 1988,
29, 1387−1395.
(32) Mancuso, A. J.; Swern, D. Activated dimethyl sulfoxide: useful
reagents for synthesis. Synthesis 1981, 3, 165−185.
(33) Nagoka, H.; Kishi, Y. Further synthetic studies on rifamycin S.
Tetrahedron 1981, 37, 3873−3888.
(34) Fujino, T.; Une, M.; Imanaka, T.; Inoue, K.; Nishimaki-Mogami,
T. Structure−activity relationship of bile acids and bile acid analogs in
regard to FXR activation. J. Lipid Res. 2004, 45, 132−138.
(35) Pellicciari, R.; Fiorucci, S.; Camaioni, E.; Clerici, C.; Costantino,
G.; Maloney, P. R.; Morelli, A.; Parks, D. J.; Willson, T. M. 6-α-Ethyl-
chenodeoxycholic acid (6-ECDCA), a potent and selective FXR
agonist endowed with anticholestatic activity. J. Med. Chem. 2002, 45,
3569−3572.
(36) Fiorucci, S.; Cipriani, S.; Mencarelli, A.; Baldelli, F.; Bifulco, G.;
Zampella, A. Farnesoid X receptor agonist for the treatment of liver
and metabolic disorders: focus on 6-ethyl-CDCA. Mini-Rev. Med.
Chem. 2011, 11, 753−762.
(37) Parker, H. E.; Wallis, K.; le Roux, C. W.; Wong, K. Y.; Reimann,
F.; Gribble, F. M. Molecular mechanisms underlying bile acid-
stimulated glucagon-like peptide-1 secretion. Br. J. Pharmacol. 2012,
165, 414−423.
(38) Petersen, A. B.; Christensen, M. Clinical potential of lixisenatide
once daily treatment for type 2 diabetes mellitus. Diabetes, Metab.
Syndr. Obes. 2013, 6, 217−231.
(39) Nauck, M. A.; Baranov, O.; Ritzel, R. A.; Meier, J. J. Do current
incretin mimetics exploit the full therapeutic potential inherent in
GLP-1 receptor stimulation? Diabetologia 2013, 56, 1878−1883.
(40) Scheen, A. J. Cardiovascular effects of dipeptidyl peptidase-4
inhibitors: from risk factors to clinical outcomes. Postgrad. Med. 2013,
125, 7−20.
(41) Cohen, D. Reports of pancreatitis are 20−30 times more likely
with GLP-1 drugs, analysis finds. Br. Med. J. 2013, 346, f2607.
(42) Butler, A. E.; Campbell-Thompson, M.; Gurlo, T.; Dawson, D.
W.; Atkinson, M.; Butler, P. C. Marked expansion of exocrine and
endocrine pancreas with incretin therapy in humans with increased
exocrine pancreas dysplasia and the potential for glucagon-producing
neuroendocrine tumors. Diabetes 2013, 62, 2595−2604.
(43) Fiorucci, S.; Distrutti, E.; Bifulco, G.; D’Auria, M. V.; Zampella,
A. Marine sponge steroids as nuclear receptor ligands. Trends
Pharmacol. Sci. 2012, 33, 591−601.
(44) Lebon, G.; Warne, T.; Edwards, P. C.; Bennett, K.; Langmead,
C. J.; Leslie, A. G.; Tate, C. G. Agonist-bound adenosine A2A receptor
structures reveal common features of GPCR activation. Nature 2011,
474, 521−525.
(45) Huey, R.; Morris, G. M.; Olson, A. J.; Goodsell, D. S. A
semiempirical free energy force field with charge-based desolvation. J.
Comput. Chem. 2007, 28, 1145−1152.
(46) Tiwari, A.; Maiti, P. TGR5: an emerging bile acid G-protein-
coupled receptor target for the potential treatment of metabolic
disorders. Drug Discovery Today 2009, 14, 523−530.
(47) Grazioso, G.; Limongelli, V.; Branduardi, D.; Novellino, E.; De
Micheli, C.; Cavalli, A.; Parrinello, M. Investigating the mechanism of
substrate uptake and release in the glutamate transporter homologue
GltPh through metadynamics simulations. J. Am. Chem. Soc. 2012, 134,
453−463.
(48) Limongelli, V.; Bonomi, M.; Parrinello, M. Funnel metady-
namics as accurate binding free-energy method. Proc. Natl. Acad. Sci.
U.S.A. 2013, 110, 6358−6363.
(49) Limongelli, V.; Marinelli, L.; Cosconati, S.; La Motta, C.; Sartini,
S.; Mugnaini, L.; Da Settimo, F.; Novellino, E.; Parrinello, M.
Sampling protein motion and solvent effect during ligand binding.
Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 1467−1472.
(50) Macchiarulo, A.; Gioiello, A.; Thomas, C.; Pols, T. W. H.; Nuti,
R.; Ferrari, C.; Giacche, N.; De Franco, F.; Pruzanski, M.; Auwerx, J.;
Schoonjans, K.; Pellicciari, R. Probing the binding site of bile acids in
TGR5. ACS Med. Chem. Lett. 2013, 4, 1158−1162.
(51) Mi, L. Z.; Devarakonda, S.; Harp, J. M.; Han, Q.; Pellicciari, R.;
Willson, T. M.; Khorasanizadeh, S.; Rastinejad, F. Structural basis for
bile acid binding and activation of the nuclear receptor FXR. Mol. Cell
2003, 11, 1093−1100.
(52) Di Leva, F. S.; Festa, C.; D’Amore, C.; De Marino, S.; Renga, B.;
D’Auria, M. V.; Novellino, E.; Limongelli, V.; Zampella, A.; Fiorucci, S.
Binding mechanism of the farnesoid X receptor marine antagonist
suvanine reveals a strategy to forestall drug modulation on nuclear
receptors. Design, synthesis, and biological evaluation of novel ligands.
J. Med. Chem. 2013, 56, 4701−4717.
(53) Brubaker, P. L.; Schloos, J.; Drucker, D. J. Regulation of
glucagon-like peptide-1 synthesis and secretion in the GLUTag
enteroendocrine cell line. Endocrinology 1998, 139, 4108−4114.
(54) Fiorucci, S.; Baldelli, F. Farnesoid X receptor agonists in biliary
tract disease. Curr. Opin. Gastroenterol. 2009, 25, 252−259.
(55) Mencarelli, A.; Renga, B.; Migliorati, M.; Cipriani, S.; Distrutti,
E.; Santucci, L.; Fiorucci, S. The bile acid sensor farnesoid X receptor
is a modulator of liver immunity in a rodent model of acute hepatitis. J.
Immunol. 2009, 183, 6657−6666.
(56) Chenna, R.; Sugawara, H.; Koike, T.; Lopez, R.; Gibson, T. J.;
Higgins, D. G.; Thompson, J. D. Multiple sequence alignment with the
Clustal series of programs. Nucleic Acids Res. 2003, 31, 3497−3500.
(57) Larkin, M. A.; Blackshields, G.; Brown, N. P.; Chenna, R.;
McGettigan, P. A.; McWilliam, H.; Valentin, F.; Wallace, I. M.; Wilm,
A.; Lopez, R.; Thompson, J. D.; Gibson, T. J.; Higgins, D. G. ClustalW
and ClustalX version 2. Bioinformatics 2007, 23, 2947−2948.
(58) Sali, A.; Blundell, T. L. Comparative protein modelling by
satisfaction of spatial restraints. J. Mol. Biol. 1993, 234, 779−815.
(59) Maestro, version 9.3; Schrodinger, LLC: New York, NY, 2012.
̈
(60) Jorgensen, W. L.; Maxwell, D. S.; Tirado-Rives, J. Development
and testing of the OPLS all-atom force field on conformational
energetics and properties of organic liquids. J. Am. Chem. Soc. 1996,
118, 11225−11236.
(61) MacroModel, version 9.9; Schrodinger, LLC: New York, NY,
̈
2012.
(62) LigPrep, version 2.5; Schrodinger, LLC: New York, NY, 2012.
̈
(63) Epik, version 2.3; Schrodinger, LLC: New York, NY, 2012.
̈
(64) Humphrey, W.; Dalke, A.; Schulten, K. VMD: visual molecular
dynamics. J. Mol. Graphics 1996, 14, 33−38.
(65) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R.
W.; Klein, M. L. Comparison of simple potential functions for
simulating liquid water. J. Chem. Phys. 1983, 79, 926−935.
(66) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K.
M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W;
Kollman, P. A. A second generation force field for the simulation of
proteins, nucleic acids, and organic molecules. J. Am. Chem. Soc. 1995,
117, 5179−5197.
(67) Lindorff-Larsen, K.; Piana, S.; Palmo, K.; Maragakis, P.; Klepeis,
J. L.; Dror, R. O.; Shaw, D. E. Improved side-chain torsion potentials
953
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954

Journal of Medicinal Chemistry
Article
for the Amber ff99SB protein force field. Proteins 2010, 78, 1950−
1958.
(68) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D.
A. Development and testing of a general amber force field. J. Comput.
Chem. 2004, 25, 1157−1174.
(69) Skjevik, Å. A.; Madej, B. D.; Walker, R. C.; Teigen, K. LIPID11:
a modular framework for lipid simulations using Amber. J. Phys. Chem.
B 2012, 116, 11124−11136.
(70) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid,
́
E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. Scalable
molecular dynamics with NAMD. J. Comput. Chem. 2005, 26, 1781−
1802.
(71) Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald: an N·
log(N) method for Ewald sums in large systems. J. Chem. Phys. 1993,
98, 10089−10092.
(72) Bayly, C. I.; Cieplak, P.; Cornell, W. D.; Kollman, P. A. A well-
behaved electrostatic potential based method using charge restraints
for determining atom-centered charges: the RESP model. J. Phys.
Chem. 1993, 97, 10269−10280.
(73) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.;
Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.;
Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.;
Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.;
Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao,
O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J.
B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R.
E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.;
Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J.
J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.;
Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman,
J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.;
Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W; Johnson, B.; Chen,
W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03; Gaussian,
Inc.: Wallingford, CT, 2004.
(74) Wang, J.; Wang, W.; Kollman, P. A.; Case, D. A. Automatic
atom type and bond type perception in molecular mechanics. J. Mol.
Graphics Modell. 2006, 25, 247−260.
954
dx.doi.org/10.1021/jm401873d | J. Med. Chem. 2014, 57, 937−954