Discovery and optimization of non-steroidal FXR agonists from natural product-like libraries.

Article abstract of DOI:10.1039/b300525a

The efficient regulation of cholesterol biosynthesis, metabolism, acquisition, and transport is an essential component of lipid homeostasis. The farnesoid X receptor (FXR) is a transcriptional sensor for bile acids, the primary product of cholesterol metabolism. Accordingly, the development of potent, selective, small molecule agonists, partial agonists, and antagonists of FXR would be an important step in further deconvoluting FXR physiology. Herein, we describe the development of four novel classes of potent FXR activators originating from natural product-like libraries. Initial screening of a 10,000-membered, diversity-orientated library of benzopyran containing small molecules for FXR activation utilizing a cell-based reporter assay led to the identification of several lead compounds possessing low micromolar activity (EC50's = 5-10 microM). These compounds were systematically optimized employing parallel solution-phase synthesis and solid-phase synthesis to provide four classes of compounds that potently activate FXR. Two series of compounds, bearing stilbene or biaryl moieties, contain members that are the most potent FXR agonists reported to date in cell-based assays. These compounds may find future utility as chemical tools in studies aimed at further defining the physiological role of FXR and discovering potential therapeutic agents for the treatment of diseases linked to cholesterol and bile acid metabolism and homeostasis.

Full text of DOI:10.1039/b300525a

View Online / Journal Homepage / Table of Contents for this issue
Discovery and optimization of non-steroidal FXR agonists from
natural product-like libraries†
K. C. Nicolaou,*a,b Ronald M. Evans,c A. J. Roecker,a Robert Hughes,a Michael Downesc and
Jeffery A. Pfefferkorna
a Department of Chemistry and The Skaggs Institute for Chemical Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, 92037, California
b Department of Chemistry and Biochemistry, University of California San Diego,
9500 Gilman Drive, La Jolla, 92093, California
c Howard Hughes Medical Institute, The Salk Institute of Biological Sciences,
10010 North Torrey Pines Road, La Jolla, 92037, California
Received 15th January 2003, Accepted 4th February 2003
First published as an Advance Article on the web 27th February 2003
The efficient regulation of cholesterol biosynthesis, metabolism, acquisition, and transport is an essential component
of lipid homeostasis. The farnesoid X receptor (FXR) is a transcriptional sensor for bile acids, the primary product
of cholesterol metabolism. Accordingly, the development of potent, selective, small molecule agonists, partial
agonists, and antagonists of FXR would be an important step in further deconvoluting FXR physiology. Herein, we
describe the development of four novel classes of potent FXR activators originating from natural product-like
libraries. Initial screening of a 10000-membered, diversity-orientated library of benzopyran containing small
molecules for FXR activation utilizing a cell-based reporter assay led to the identification of several lead compounds
possessing low micromolar activity (EC50’s = 5–10 µM). These compounds were systematically optimized employing
parallel solution-phase synthesis and solid-phase synthesis to provide four classes of compounds that potently
activate FXR. Two series of compounds, bearing stilbene or biaryl moieties, contain members that are the most
potent FXR agonists reported to date in cell-based assays. These compounds may find future utility as chemical tools
in studies aimed at further defining the physiological role of FXR and discovering potential therapeutic agents for the
treatment of diseases linked to cholesterol and bile acid metabolism and homeostasis.
Introduction
The efficient regulation of cholesterol biosynthesis, metabolism,
acquisition and transport is an essential function of mam-
malian cells. High levels of cholesterol are associated with
atherosclerosis, a leading cause of death in the western world
and a major risk factor correlated with the occurrence of
coronary heart disease and stroke. Until recently, recommenda-
tions for the treatment of hypercholestemia were focused on
the use of statins, which inhibit the de novo biosynthesis of
cholesterol, and the use of bile acid sequestering agents.1 While
statin-based agents are still in widespread use as cholesterol-
lowering drugs, an understanding of the mechanisms control-
ling cholesterol homeostasis is evolving, and this has led to new
molecular targets as candidates for therapeutic intervention.
Cholesterol metabolism is controlled through a complex
feedback loop involving cholesterol itself and bile acids, which
are cholesterol’s primary oxidation products and, through
secretion in the gut, the single most critical regulators of choles-
terol absorption. The nuclear receptors LXR (liver X receptor)
and FXR (farnesoid X receptor) are the specialized sensors
for cholesterol and bile acids, respectively, that control tran-
scription of networks encoding key metabolic enzymes.2–5 For
example, activation of LXR by oxysterols (i.e. mono-oxygen-
ated cholesterol metabolites) leads to the up-regulation of
CYP7A1, the enzyme that catalyzes the rate limiting step in the
conversion of cholesterol to bile acids. In turn, bile acids6–8
such as chenodeoxycholic acid (CDCA, 1, Fig. 1) are ligands for
FXR whose activation leads to a down-regulation of CYP7A1
Fig. 1 Natural and synthetic agonists of FXR (farnesoid X receptor).
aCell-based assay.
† Electronic supplementary information (ESI) available: schemes
describing the synthesis of compounds in Fig. 2, 4, 5, 6 and 7. All
leading to the completion of the feedback circuit. In this circuit
FXR induces the expression of a transcriptional repressor SHP
(small heterodimer partner) which, in turn, binds to LRH-1
final compounds were characterized by 1H NMR spectroscopy and
HRMS are available on request. See http://www.rsc.org/suppdata/ob/
b3/b300525a/
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
908

View Online
(liver receptor homolog), required in CYP7A1 activation.9,10
Additionally, both LXR and FXR are implicated in the regu-
lation of several other gene products involved in cholesterol
absorption, metabolism and transport.11–17
screened in the cell-based assay in order to evaluate the
structural requirements of each region of the molecule for
potent FXR agonism. At this point we chose to utilize parallel
solution-phase chemistry for the construction of additional
focused libraries. This shift away from solid-phase chemistry
provided us with maximum flexibility as we sought to rapidly
and systematically optimize each region of the lead molecules
using smaller designed libraries.
Thus, potent, selective, small molecule FXR agonists, partial
agonists and antagonists would be powerful tools and would
have many potential applications.18 First, such compounds
would facilitate the analysis of FXR physiology in vivo. Second,
such compounds in conjunction with DNA arraying tech-
nology might allow for the discovery of new gene products
under the control of FXR. Third, FXR modulators might find
potential utility in the treatment of cholestasis and other dis-
ease states associated with aberrant levels, flow and release of
bile acids. Fourth, in the absence of a crystal structure of FXR,
a thorough SAR study of ligands that modulate the activity of
FXR would allow for the delineation of the structural require-
ments for ligand binding and might aid in the design of future
ligands and potential therapeutics.
As there is only one class of high affinity, non-steroidal
agonists for FXR, exemplified by GW406419,20 (3, Fig. 1),
our strategy for discovering similarly potent compounds was
initiated with the screening of a 10000-membered library
constructed around the privileged 2,2-dimethylbenzopyran
scaffold. Such privileged structures are attractive starting points
for lead discovery, particularly when little or no structural
information exists regarding the target, as they show good bind-
ing affinity toward a wide variety of enzymes and receptors.21–23
Indeed, we have been able to use this library as a starting point
for the successful discovery of potent inhibitors of NADH:
Ubiquinone oxidoreductase24 and for the identification and
optimization of novel antibacterial agents.25 The initial hits dis-
covered from screening this library for FXR activation could be
further optimized for potency and pharmacological properties
suitable for the applications mentioned above. Herein, we
describe the implementation of such a strategy culminating in
the discovery of four classes of potent and selective activators
of FXR.
Evaluation of benzopyran Region I SAR
Most of the FXR agonists reported to date including
CDCA (1), TTNPB (2) and GW4064 (3) (see Fig. 1) contain a
carboxylic acid moiety. We, thus, reasoned that incorporation
of an acid unit within either Region I, II or III of structure 26
(Fig. 3) would confer increased potency upon this rather weak
ligand (5–10 µM) identified via high throughput screening
(HTS). Guided by this reasoning, we evaluated the SAR of
Region I. Several compounds displaying the acid unit in various
positions were synthesized (e.g. compounds 28, 36, 52, 54 and
56, Fig. 4) and tested. None of these compounds, however,
showed improved activation of FXR. Interestingly, compound
29, bearing a meta methyl acrylate moiety, was a substantially
better activator of FXR than compound 26. The preparation
of compound 29 is representative of the methods employed
to construct these compounds and is described in Scheme 1
(see Supplementary Schemes 2–4† for further experimental
procedures). Thus, aldehyde 59 was selectively methylated29
(NaH, MeI), alkylated (2-methyl-3-butyn-2-ol, TFAA, DBU,
CuCl2), reduced (Lindlar, H2) and thermally cyclized to yield
benzopyran 60. Reductive amination of aldehyde 60 with
3-bromoaniline (NaCNBH3) followed by acylation with cyclo-
propanecarbonyl chloride (C3H5COCl, Et3N) and palladium-
mediated Heck coupling (Pd2(dba)3, P(o-tol)3, Et3N) with
methyl acrylate provided 29.
Results and discussion
Previous screening technologies for identifying small molecule
activators of FXR utilized a fluorescence resonance energy
transfer (FRET) assay to detect the ligand-dependent recruit-
ment of the coactivator SRC-1 to FXR.26 The association of
FXR with a coactivator is a necessary event for transcriptional
activation. In the present investigations, however, we employed
a cell-based transcription assay in which an FXR responsive
promoter is linked to a luciferase reporter as our primary
screen. In addition to ensuring that only cell permeable com-
pounds were selected for further optimization, this approach
allows for the detection of FXR activation in a natural system
(i.e. correct folding of the protein and in the presence of a
complete compliment of coactivators and corepressors).27,28
Initial screening of a 10000-membered combinatorial library
of benzopyran-based small molecules in this high-throughput,
cell-based assay for FXR activation produced several lead
compounds whose structures are listed in Fig. 2a (4–15).
Guided by the preliminary SAR gained from the evaluation of
this initial library we designed and synthesized a follow-up
focused library of ca. 200 benzopyran-based compounds on
solid support (see Supplementary Scheme 1†). A selection of
the most active compounds, possessing EC50 values between 5
and 10 µM, discovered from this second round of screening is
shown in Fig. 2b (16–27). Compounds 26 and 27 proved to be
among the most active at this stage and were the subject of
further optimization as described below.
Scheme 1 Representative procedure for the preparation of Region I
modified compounds: synthesis of methyl acrylate 29. Reagents and
conditions: (a) see ref. 28; (b) 1.5 equiv. 2-methyl-3-butyn-2-ol, 1.5 equiv.
DBU, 1.7 equiv. trifluoroacetic anhydride, 0.1 equiv. CuCl2, CH3CN, 0
25 ЊC, 12 h, 75%; (c) N,N-diethylaniline, 190 ЊC, 0.5 h, 90%; (d) 1.5
equiv. 3-bromoaniline, THF, 70 ЊC, 4 h; then 2.0 equiv. NaCNBH3, 10%
MeOH, 70 Њ C, 4 h, 83%; (e) 1.3 equiv. cyclopropanecarbonyl chloride,
1.3 equiv. Et3N, 0.1 equiv. 4-DMAP, CH2Cl2, 25 ЊC, 12 h, 90%; (f ) 4.0
equiv. methyl acrylate, 0.2 equiv. Pd2(dba)3, 0.5 equiv. P(o-tol)3, 5.0
equiv. Et3N, DMF, 90 ЊC, 24 h, 80%. DBU
=
1,8-
diazabicyclo[5.4.0]undec-7-ene, 4-DMAP = 4-dimethylaminopyridine,
Pd2(dba)3 = tris(dibenzylideneacetone)dipalladium(0).
In further refining the SAR of Region I, an important
observation was that the location of the methyl acrylate moiety
at the meta position was crucial for potent activation of FXR as
compound 53 (Fig. 4) bearing a para methyl acrylate does not
activate FXR. In order to further examine what functionality
was tolerable at the meta position, we synthesized the addi-
tional compounds shown in Fig. 4. From biological screening
With initial lead compounds identified and validated, the
stage was set for the systematic optimization of the three
regions of the lead structure shown in Fig. 3. As detailed in the
following sections, focused libraries were synthesized and
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
909

View Online
Fig. 2 a) Selected hits from a high throughput screen for FXR agonism of a 10,000-membered benzopyran-based natural product-like library (EC50
= 5–10 µM). b) Selected low affinity FXR agonists from follow-up solid phase benzopyran library (EC50 = 5–10 µM). See Supplementary Scheme 1†
for details of the focused library synthesis. The boxed compounds represent the most potent FXR agonists.
of these compounds it became clear that the length and rigidity
of the tether between the aromatic core and the interacting
functionality (either methyl ester or methyl ether) are important
for FXR agonism. For instance, compounds 41 and 45 appear
to possess either too short or too long of a tether for potent
activity; compounds 35 and 46–49 presumably cannot adopt
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
910

View Online
stitution patterns in this region of the molecule (see Supple-
mentary Scheme 5† for preparation of these compounds). Only
compounds 65 (EC50 = 358 nM) and 68 (EC50 = ca. 1 µM)
were more effective than compound 29 in activating FXR.
Substituted aromatic amide derivatives such as 69–77 were all
found to be less active than the parent compound 68. Alkyl
derivatives 78 and 79 were inactive as were sulfonamide 82,
thiourea 84, and thioamide 83 suggesting the importance of
acylation at this position. The sum of these results pointed to
Region II requiring moderately bulky cycloalkyl amide moieties
for optimal activity.
Fig. 3 Selected regions of interest for SAR evaluation of lead
compound 26. Region I: right-hand aromatic system; Region II: acyl
group region; Region III: left-hand benzopyran ring system.
the correct orientation for potent activation; and compounds
30, 31, 34, 38, 39, 40, and 50 do not apparently present the
correct interacting functionality to the receptor as they are
inactive. Indeed, of all the analogs designed to probe the
SAR of Region I, only compounds 29 and 33 are capable of
activating FXR to a significant extent. Due to relative ease
of synthesis of compound 29 we chose to use this analog as a
starting point for the optimization of Region II.
Evaluation of benzopyran Region III SAR
Having thoroughly examined Regions I and II, we then turned
our attention to the optimization of Region III (see Fig. 6 for
structures and Supplementary Schemes 6 and 7† for preparation
of compounds). Incorporation of polar H-bond donating
functional groups such as those that adorn compounds 86,
93, 94, 98 and 100 did not improve the activity of the analogs.
Nor did the addition of H-bond acceptors such as in 89, 90,
95, 99 and 101 improve the ability of the parent compound
68 to activate FXR. Finally, the addition of bulky lipophilic
groups to the benzopyran moiety afforded compounds that
Evaluation of benzopyran Region II SAR
As shown in Fig. 5, we examined the effect of numerous sub-
Fig. 4 Examination of Region I SAR. See Scheme 1 and Supplementary Schemes 2–4† for a description of the synthesis of these compounds.
a Benzopyran double bond is also saturated in this compound. Boxed compounds represent the most potent FXR agonists.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
911

View Online
Fig. 5 Examination of the acyl group (Region II) SAR. See Supplementary Scheme 5† for a description of these compounds. Boxed compounds are
the most active FXR agonists.
Fig. 6 Examination of the benzopyran (Region III) SAR. See Scheme 2 and Supplementary Schemes 6 and 7† for a description of the synthesis of
these compounds. Boxed compounds are the most active FXR agonists.
only weakly activated FXR. However, replacement of the
double bond in the benzopyran unit by a dichlorocyclopropane
unit provided analog 102 (EC50 = 333 nM). This potent
compound was synthesized as depicted in Scheme 2. Thus,
benzopyran 103 was cyclopropanated under phase transfer
conditions (adogen 464 (cat), NaOH, CHCl3) and converted to
the corresponding cinnamate via a Heck coupling (Pd2(dba)3,
P(o-tol)3, Et3N) with methyl acrylate to yield 102. Replacement
of the benzoyl group in Region II of compound 102 with the
cyclohexanecarbonyl moiety afforded the even more potent
compound 149 (EC50 = 188 nM) (Table 1c).
significant improvement in potency over compound 65 (EC50 =
358 nM), it became clear that further possibilities for optimiza-
tion of this class of compounds were limited. Therefore, we
chose to examine the effect of replacing the benzopyran moiety
with other ring systems.
Fig. 7 presents a set of compounds in which the benzopyran
moiety was replaced with certain groups of varying molecular
diversity (see Supplementary Schemes 8 and 9† for preparation).
Biological assays showed that replacement of the benzopyran
with a small aromatic unit generally had a detrimental effect
on activity. For instance, compounds 110 and 112–117 (see
Table 1c and Fig. 7) were inactive, while compounds 111 and
Although compound 149 (EC50 = 188 nM) represents a
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
912

View Online
Fig. 7 Examination of the benzopyran replacement (Region III) SAR. See Scheme 3 and Supplementary Schemes 8, 9 and 10† for a description of
the synthesis of these compounds.
were all potent activators of FXR in the cell-based reporter
assay. The synthesis of compound 105 is outlined in Scheme 3
(see Supplementary Scheme 14† for the preparation of 121–
129). Thus, acylation of 3-bromoaniline (C6H11COCl, Et3N)
gave cyclohexylamide 131. Subsequent reaction of 131 under
Heck coupling conditions (Pd2(dba)3, P(o-tol)3, Et3N) with
methyl acrylate gave 132. Finally, alkylation (4-bromobenzyl
bromide, NaH) of cinnamate 132 followed by a second Heck
coupling (Pd2(dba)3, P(o-tol)3, Et3N) with tert-butyl acrylate
gave 105.
This initial survey of the three regions of SAR outlined in
Fig. 3 led to the identification of four classes of potent FXR
agonists (based on the differences in Region III substitution)
for further evaluation. The first class is represented by benzo-
pyran-derived dichlorocyclopropane 149 (EC50 = 188 nM).
Compound 105 (EC50 = 127 nM) is the prototypical member
of the bis-cinnamate series. Finally, compounds 121 (EC50
=
36 nM) and 124 (EC50 = 69 nM) are members of the stilbene
and biaryl series, respectively. Based on data presented in
Figs 4, 5 and 6, compound 149 (Table 1c) appeared to represent
the optimal potency that could be readily obtained in the
benzopyran-derived series. However, the bis-cinnamate, biaryl,
and stilbene series were thought to still possess considerable
potential for further development and rigorous SAR analysis.
Below we detail the results of such investigations which indeed
led to further enhancement of biological activity.
Scheme 2 Synthesis of compound 102. Exploration of Region III
SAR. Reagents and conditions: (a) CHCl3–2.0 M NaOH (7 : 1), adogen
464 (cat.), 25 ЊC, 6 h, 85%; (b) 2.0 equiv. methyl acrylate, 0.2 equiv.
Pd2(dba)3, 0.6 equiv. P(o-tol)3, 5.0 equiv. Et3N, DMF, 90 ЊC, 24 h, 75%.
DMF = dimethylformamide.
118 showed only moderate activation of FXR (EC50 = 680 nM
and 606 nM, respectively). However, replacement of the benzo-
pyran with an aromatic ring bearing substituents at the para
position produced compounds with improved activity. For
example, 4-tert-butyl cinnamate 105 (EC50 = 127 nM), stilbenes
121 and 122 (EC50 = 36 and 208 nM, respectively), biaryls 124–
127 (EC50 = 510, 69, 77, 227 nM, respectively) and aryl thio-
phenes 128 and 129 (EC50 = 206 and 256 nM, respectively)
Examination of the bis-cinnamate series
Similar to the results described above, the meta substituted
methyl cinnamate moiety on the “right-hand” region of the
molecule remained a necessary component for optimal activity
in the bis-cinnamate series (see Table 1a and Supplementary
Scheme 11†). Replacement of this methyl acrylate unit with
either a methyl or ethyl allylic ether (136 and 137) caused only a
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
913

View Online
Table 1 Panel a) Highlights of Region I SAR. Panel b) Highlights of Region II SAR in the bis-cinnamate series. Panel c) Effects of benzopyran
substitution. Panel d) Highlights of Region III SAR including the bis-cinnamate, styryl and biaryl series. Values represent the mean of at least four
experiments. RE = relative efficacy of the indicated compound at 1 µM to 100 µM CDCA.
slight decrease in activity (EC50 = 243 and 220 nM, respec-
tively). A marked decline in potency accompanied substitution
of the methyl acrylate by more sterically bulky ethers or esters
(133 and 134) or amides (135). Interestingly, saturation of
the acrylate olefin (139) afforded only a two-fold decrease in
potency, EC50 = 274 nM, which supports the notion that con-
formational rigidity is a factor contributing to, but not essential
for, high affinity ligands. Importantly, compound 139 suggests
that the methyl acrylate moiety is not simply functioning as a
latent electrophile.
Region II also closely mirrored the preceding data as cyclo-
alkyl amides remained the optimal substituents (105 and
140–142: EC50 = 127–250 nM) in the bis-cinnamate series
(see Table 1b and Supplementary Scheme 12†). Aromatic and
heterocyclic amides as well as alkyl ureas led to moderate
potency (143–145: EC50 = 205–236 nM) whereas incorporation
of bulky ureas such as in compound 146 rendered compounds
of only marginal efficacy.
As mentioned above, replacement of the benzopyran moiety
with a benzyl group bearing a tert-butyl acrylate moiety in
the para-position yielded compound 105 with dramatically
increased efficacy (EC50 = 127 nM). Interestingly, placement
of the same tert-butyl acrylate group in either the meta or
ortho positions of the aromatic ring (107 and 109, Fig. 7 and
Supplementary Scheme 9† for synthesis) in Region III led to
only micromolar potency. Further investigation of the “left-
hand” region in this series of compounds demonstrated that a
decrease in ester group size yielded a corresponding decrease
in efficacy (EC50 of t-butyl > i-propyl > ethyl > methyl; com-
pounds 105, 150–152, Table 1d and Supplementary Scheme
10†). Similarly, substitution of the ester with either carboxylic
acid or amide functionality provided less effective compounds
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
914

View Online
receptor site for potent activation. Hence, stilbenes in which the
aromatic nucleus is removed two carbon atoms further away
from the core of the molecule should be adorned with small
substituents while the biaryl compounds should be substituted
with larger functionality for optimal activity. Cleavage of
resins 172 and 173 with NaOMe yielded methyl acrylates
121,125,126 and 174–264. Analysis of the library by LCMS
after purification (PTLC) showed the average purity of these
compounds to be >95%.
Screening of this compound library in the cell-based assay
led to some intriguing results as summarized in Table 2. Thus, it
was found that in both the stilbene and biaryl series, analogs
bearing the cyclohexyl amide moiety are generally the most
potent followed by those bearing the isopropyl amide or
isopropyl urea units. As expected, stilbenes bearing smaller
substituents were more potent than those carrying larger
functionality. For instance, unsubstituted stilbene 121 and
mono-fluoro stilbenes 192, 200, and 203 were among the
most active, while mono-methyl derivative 174 and tri-methyl
derivative 195 were among the least active. Also of interest were
the heterocyclic compounds 207 and 210 which retained good
potency (EC50 = 309 and 227 nM, respectively) and may possess
improved pharmacological properties. In the biaryl series, com-
pounds which present more bulky substituents at the terminus
of the structure were more active. In this series, compounds 259
(EC50 = 25 nM) and 244 (EC50 = 38 nM) were particularly
active. Overall, most of the compounds synthesized in this
follow-up study were efficient activators of FXR, providing
further support for our working hypothesis for the FXR
binding pocket, as described above. This model may provide
a solid basis for further development of FXR activators.
A summary of the molecular requirements for potent FXR
activation is shown in Fig. 8. Thus, in Region I the presence of
the meta methyl acrylate unit is important for potent activation
as only a few modifications retain good activity. The most
potent compounds possess a cycloalkylamide in Region II.
Finally, Region III is the most tolerant and several structural
elements were found to provide a good fit within the pocket of
the receptor.
Scheme 3 Preparation of bis-cinnamate 105. Reagents and conditions:
(a) 1.1 equiv. C6H11COCl, 1.3 equiv. Et3N, 0.05 equiv. 4-DMAP,
CH2Cl2, 25 ЊC, 3 h, 95%; (b) 4.0 equiv. methyl acrylate, 5.0 equiv. Et3N,
0.2 equiv. Pd2(dba)3, 0.6 equiv. P(o-tol)3, DMF, 90 ЊC, 12 h, 80%; (c) 1.1
equiv. NaH, THF, 0 ЊC, 30 min; then 1.3 equiv. 4-bromobenzyl
bromide, THF, 0 ЊC, 2 h, 90%; (d) 4.1 equiv. acrylate, 5.0 equiv.
Pd2(dba)3, 0.15 equiv. P(o-tol)3, DMF, 90 ЊC, 12 h, 75%.
with EC50 values in the micromolar range. Substitution of the
tert-butyl acrylate moiety with a methyl or ethyl allylic ether
(156 and 157) retained considerable potency (EC50 = 233 and
198 nM, respectively). However, the more bulky phenyl allylic
ether 158 possessed only micromolar activity. In addition,
saturation of the acrylate moiety (159) showed a two-fold
decrease in potency from the parent compound (105). Finally,
substitution of the ortho position of the aromatic ring of the
tert-butyl acrylate series with oxygenated functionality (161–
167, see Table 1d and Supplementary Scheme 13† for synthesis)
afforded compounds with very low biological activity.
Construction of biaryl and stilbene containing focused libraries
In an effort to further optimize the biaryl and stilbene series,
a 93-membered library of such compounds was constructed
employing a split-and-pool solid phase strategy. Individual
library members were identified via radiofrequency encoding
using IRORI tags and MacroKan technologies.21–23 As
shown in Scheme 4, Boc protected cinnamic acid 168 was
immobilized on Merrifield resin (Cs2CO3) to afford resin 169.
The Boc group of this resin was removed by treatment with
20% TFA in CH2Cl2 and the resultant resin-bound amine was
reductively alkylated with 4-bromobenzaldehyde (NaCNBH3)
to yield amino resin 170. Resin 170 was acylated with one of
three acyl groups to give amide or urea resins 171. The acylated
resins (171) were subjected to either Heck coupling (Pd2(dba)3,
P(o-tol)3, Et3N) with thirteen substituted styrenes or Suzuki
coupling (Pd(PPh3)4, Cs2CO3) with eighteen boronic acids to
yield stilbene resins 172 and biaryl resins 173, respectively.
In selecting appropriate styrenes and boronic acids for inputs
into this combinatorial library we were guided by initial com-
parisons of tert-butyl stilbene (123, EC50 = >1000 nM) to the
unsubstituted stilbene 102 (EC50 = 36 nM), and biaryl com-
pound 124 (EC50 = 510 nM) to 125 (EC50 = 69 nM) as shown
in Table 1. We reasoned that both the stilbene and the biaryl
ligands needed to fit into the same region of space within the
Fig.
8
Summary of structural requirements for potent FXR
activation. Region I: methyl acrylate or allylic methyl ether necessary
for optimum activity. In some instances, when other areas were
optimized, olefin can be removed while retaining some potency. Region
II: amide or urea essential for maximum activity. Alkyl or cycloalkyl
amide or urea affords most potent compounds. Region III: must have
para-position functionalized for activity. Steric bulk and length seem to
be the most important factors which govern potency. This region is
tolerant of many different structural motifs.
In order to determine how selectively the compounds above
activated FXR we screened some of the most active compounds
against a panel of nuclear receptors and the full details of
this cross-activation screening will be reported elsewhere.30 It
suffices to note here that most of these compounds were found
to be highly selective, activating only FXR. Notably, however,
compound 149 also potently activated SXR (FXR: EC50
=
188 nM, SXR: EC50 = 77 nM). Far from being discouraging,
this result may lead to compounds which may find utility in the
treatment of diseases linked to the accumulation of toxic bile
acids.31,32
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
915

View Online
Scheme 4 Solid phase synthesis of focused libraries of biaryl and stilbene cinnamates. Reagents and conditions: (a) 2.0 equiv. 168, 1.0 equiv.
Merrifield resin (0.91 mmol gϪ1), 2.0 equiv. Cs2CO3, 0.5 equiv. TBAI, DMF, 55 ЊC, 24 h; (b) 20% TFA in CH2Cl2, 25 ЊC, 1 h; (c) 10.0 equiv.
4-bromobenzaldehyde, 0.05 equiv. AcOH, THF–MeOH (2 : 1), 25 ЊC, 1 h; then, 8.0 equiv. NaCNBH3, THF–MeOH (2 : 1), 25 ЊC, 2 h; (d) for
R1C(O)Cl: 30.0 equiv. i-PrC(O)Cl or C6H11C(O)Cl, 40.0 equiv. Et3N, 1.0 equiv. 4-DMAP, CH2Cl2, 25 ЊC, 12 h; for R1NCO, 30.0 equiv. i-PrNCO, 40.0
equiv. Et3N, 1.0 equiv. 4-DMAP, DMF, 65 ЊC, 60 h; (e) 8.0 equiv. styrene, 10.0 equiv. Et3N, 0.5 equiv. Pd2(dba)3, 1.5 equiv. P(o-tol)3, DMF, 90 ЊC,
48 h; (f ) 5.0 equiv. boronic acid, 3.0 equiv. Cs2CO3, 0.5 equiv. Pd(PPh3)4, DMF, 90 ЊC, 24 h; (g) 10.0 equiv. NaOMe, Et2O–MeOH (10 : 1), 25 ЊC,
20 min. TBAI = tetrabutylammonium iodide, TFA = trifluoroacetic acid.
Fig. 9 Lead compounds (and their EC50 values in a cell-based assay) selected for further biological evaluation as FXR agonists. a RE = relative
efficacy of the indicated compound at 1 µM to 100 µM CDCA.
Conclusion
pounds, the complete investigation of the SAR of these families
of compounds may facilitate further optimization of additional
desired properties for these agonists. It is interesting to note that
none of these potent compounds, in contrast to the previously
reported FXR agonists (see Fig. 1), contains a carboxylic acid
moiety. Studies are currently under way to elucidate the poten-
tial structural significance of this observation. Additionally,
ongoing biological studies utilizing this expanded repertoire of
orphan receptor ligands specifically activating FXR are aimed
at further elucidating the physiological function of this newly
discovered receptor. Finally, some of these compounds may
In this article we describe work which demonstrated the utility
of natural product-like libraries as fertile pools for the dis-
covery of novel FXR agonists. This study has produced, as
summarized in Fig. 9, four diverse classes of FXR agonists,
several of which are among the most potent FXR activators
reported to date. To facilitate future reference to these com-
pounds we propose to name them as follows: 105 as fexaramate,
121 as fexarene, 259 as fexaramine, 244 as fexarine and 149 as
fexarchloramide. In addition to the discovery of potent com-
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
916

View Online
Table 2 Activities of stilbene and biaryl series. RE = relative efficacy of the indicated compound at 1µM to 100 µM CDCA.
prove useful for in vivo proof-of-concept experiments which may
open new approaches to therapeutic agents for the treatment of
diseases linked to cholesterol, bile acids, and their metabolism
and homeostasis. Such studies will be reported in due course.
chromatography carried out on 0.25 mm E. Merck silica gel
plates (60F-254) using UV light as visualizing agent and 7%
ethanolic phosphomolybdic acid or p-anisaldehyde solution
and heat as developing agents. E. Merck silica gel (60, particle
size 0.040–0.063 mm) was used for flash column chromato-
graphy. Preparative thin-layer chromatography (PTLC)
separations were carried out on 0.25 mm E. Merck silica gel
plates (60F-254). All final products cleaved from solid support
were characterized by LCMS. NMR spectra were recorded on
Bruker DRX-600, AMX-500 or AMX-400 instruments and
calibrated using residual undeuterated solvent as an internal
reference. High resolution mass spectra (HRMS) were recorded
on a VG ZAB-ZSE mass spectrometer under MALDI-FTMS
conditions with NBA as the matrix. Representative procedures
for each region of SAR and the final combinatorial library
synthesized are provided below.
Experimental
General details
Reagents and resins were purchased at highest commercial
quality and used without further purification, unless other-
wise stated. Anhydrous solvents were obtained by passing
them through a commercially available alumina column. All
reactions were carried out under an argon atmosphere with
dry solvents under anhydrous conditions, unless otherwise
noted. Solution phase reactions were monitored by thin layer
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
917

View Online
Representative procedure for synthesis of Region I/II analogues;
synthesis of acrylate 29 (Scheme 1)
ature. The reaction mixture was then diluted with EtOAc (10
mL) and washed with water (3 × 5 mL) and brine (1 × 5 mL).
The combined organic phase was dried over MgSO4, filtered,
concentrated and purified by column chromatography (silica,
To a solution of aldehyde 60 (for synthesis see ref. 21) (50.0 mg,
0.229 mmol, 1.0 equiv.) in THF (1.0 mL) at 25 ЊC was added
3-bromoaniline (59.0 mg, 0.344 mmol, 1.5 equiv.) and the
reaction mixture was heated to 70 ЊC. The solution was stirred
for 4 h and cooled to ambient temperature. To the resulting
mixture was added methanol (0.2 mL) and NaCNBH3 (28.8
mg, 0.458 mmol, 2.0 equiv.) and heated to 70 ЊC for 4 h. The
reaction mixture was then cooled and quenched with brine
(5 mL). The reaction mixture was then concentrated and
extracted with EtOAc (3 × 5 mL). The combined organic
phase was dried over MgSO4, filtered and concentrated and
used without further purification (90% yield by crude 1H
NMR analysis). To a solution of the resulting secondary
amine (0.206 mmol, 1.0 equiv.) in CH2Cl2 (1.0 mL) was added
triethylamine (0.038 mL, 0.268 mmol, 1.3 equiv.), 4-DMAP
(2.6 mg, 0.021 mmol, 0.1 equiv.), and cyclopropanecarbonyl
chloride (28.0 mg, 0.268 mmol, 1.3 equiv.). The reaction
mixture was stirred at 25 ЊC for 12 h and quenched with the
addition of brine (5 mL). The aqueous phase was then
extracted with CH2Cl2 (3 × 5 mL). The combined organic phase
was dried over MgSO4, filtered and concentrated and used
without further purification (95% yield by crude 1H NMR
analysis). To the resulting amide (0.196 mmol, 1.0 equiv.)
in N,N-dimethylformamide (1.0 mL) was added triethyl-
amine (0.137 mL, 0.980 mmol, 5.0 equiv.), methyl acrylate
(0.071 mL, 0.784 mmol, 4.0 equiv.), tri-o-tolylphosphine
(30.0 mg, 0.098 mmol, 0.5 equiv.), and tris(dibenzylidene-
acetone)dipalladium(0) (35.9 mg, 0.039 mmol, 0.2 equiv.)
sequentially and heated to 90 ЊC. The reaction mixture was
stirred for 24 h and then cooled to ambient temperature. The
reaction mixture was then diluted with EtOAc (10 mL) and
washed with water (3 × 5 mL) and brine (1 × 5 mL). The
combined organic phase was dried over MgSO4, filtered, con-
centrated and purified by column chromatography (silica,
0
30% EtOAc in hexanes) to afford 102 (30.9 mg, 75%).
102: Rf = 0.38 (silica, 25% ethyl acetate in hexane); FT-IR
(neat) νmax 2943, 1719, 1640, 1590, 1443, 1378, 1314, 1226,
1161 cmϪ1; 1H NMR (400 MHz, CDCl3) δ 7.44 (d, J = 16.2 Hz,
1 H), 7.34 (d, J = 7.0 Hz, 1 H), 7.24–7.08 (m, 5 H), 7.04–6.98
(m, 3 H), 6.94 (d, J = 7.3 Hz, 1 H), 6.17 (d, J = 16.1 Hz, 1 H),
5.23–5.15 (m, 2 H), 3.78 (s, 3 H), 3.61 (s, 3 H), 2.84 (d, J =
10.9 Hz, 1 H), 2.06 (d, J = 10.9 Hz, 1 H), 1.72 (s, 3 H), 1.16
(s, 3 H); 13C NMR (125 MHz, CDCl3) δ 170.7, 167.0,
147.5, 144.1, 144.0, 143.6, 135.8, 135.1, 130.3, 129.8, 129.4,
129.3, 128.6, 128.0, 127.9, 127.8, 127.6, 127.0, 126.1, 124.8,
121.0, 118.8, 118.7, 71.4, 62.0, 60.6, 51.8, 40.9, 31.0, 27.0, 26.7;
HRMS calcd for C31H29Cl2NO5 [M ϩ Naϩ] 588.1315, found
588.1323.
Representative procedure for synthesis of Region III non-benzo-
pyran containing analogues; synthesis of bis-cinnamate 105
(Scheme 3)
To a solution of 3-bromoaniline (130, 60.0 mg, 0.349 mmol,
1.0 equiv.) in CH2Cl2 (1.0 mL) at 25 ЊC was added triethyl-
amine (0.064 mL, 0.453 mmol, 1.3 equiv.), 4-DMAP (2.1 mg,
0.017 mmol, 0.05 equiv.), and cyclohexanecarbonyl chloride
(56.3 mg, 0.384 mmol, 1.1 equiv.). The reaction mixture was
stirred for 3 h and quenched with brine (5 mL). The aqueous
phase was extracted with CH2Cl2 (3 × 5 mL) and subsequently
dried over MgSO4, filtered, concentrated to afford amide 131
(95% yield by crude 1H NMR analysis) which was utilized
without further purification. To a solution of amide 131 (0.332
mmol, 1.0 equiv.) in N,N-dimethylformamide (2.0 mL) was
added triethylamine (0.232 mL, 1.66 mmol, 5.0 equiv.), methyl
acrylate (0.119 mL, 1.33 mmol, 4.0 equiv.), tri-o-tolylphosphine
(60.8 mg, 0.199 mmol, 0.6 equiv.), and tris(dibenzylidene-
acetone)dipalladium(0) (60.8 mg, 0.066 mmol, 0.2 equiv.) and
heated to 90 ЊC. The reaction mixture was stirred for 24 h and
then cooled to ambient temperature. The reaction mixture
was then diluted with EtOAc (15 mL) and washed with water
(3 × 5 mL) and brine (1 × 5 mL). The combined organic phase
was dried over MgSO4, filtered, concentrated and purified by
0
30% EtOAc in hexanes) to afford 29 (70.1 mg, 80%).
29: Rf = 0.42 (silica, 25% ethyl acetate in hexane); FT-IR
(neat) νmax 2943, 1720, 1642, 1596, 1443, 1414, 1314, 1267, 1202
cmϪ1; 1H NMR (400 MHz, CDCl3) δ 7.60 (d, J = 15.8 Hz, 1 H),
7.40 (d, J = 7.6 Hz, 1 H), 7.34 (d, J = 7.6 Hz, 1 H), 7.31 (s, 1 H),
7.18 (d, J = 7.6 Hz, 1 H), 6.77 (d, J = 7.9 Hz, 1 H), 6.67 (d, J =
7.9 Hz, 1 H), 6.35 (d, J = 15.8 Hz, 1 H), 6.27 (d, J = 9.7 Hz, 1 H),
5.57 (d, J = 9.7 Hz, 1 H), 4.98 (s, 2 H), 3.80 (s, 3 H), 3.56 (s, 3 H),
1.42 (s, 6 H), 1.37–1.33 (m, 1 H), 1.10–1.05 (m, 2 H), 0.70–0.60
(m, 2 H); 13C NMR (125 MHz, CDCl3) δ 179.9, 167.1, 151.1,
145.1, 143.7, 143.4, 135.6, 130.8, 130.4, 130.1, 129.8, 128.0,
127.8, 127.6, 126.9, 122.2, 122.0, 120.9, 118.8, 76.4, 60.4, 51.8,
27.8, 8.7; HRMS calcd for C27H29NO5 [M ϩ Hϩ] 448.2118,
found 448.2117.
column chromatography (silica, 0
50% EtOAc in hexanes)
to afford 132 (71.6 mg, 75%). To a solution of acrylate 132
(60.0 mg, 0.209 mmol, 1.0 equiv.) in THF (1.0 mL) at 0 ЊC was
added NaH (9.2 mg, 0.230 mmol, 60% dispersion in mineral
oil, 1.1 equiv.) followed by 4-bromobenzyl bromide (67.9 mg,
0.272 mmol, 1.3 equiv.). The reaction mixture was stirred for
2 h and quenched with saturated NH4Cl (5 mL). The aqueous
phase was extracted with EtOAc (3 × 5 mL) and the combined
organic phase was dried over MgSO4, filtered, concentrated and
Representative procedure for synthesis of Region III benzopyran
containing analogues; synthesis of dichlorocyclopropane 102
(Scheme 2)
purified by column chromatography (silica, 0
30% EtOAc
in hexanes) to afford 112 (85.8 mg, 90%). To a solution of
amide 112 (50.0 mg, 0.110 mmol, 1.0 equiv.) in N,N-di-
methylformamide (2.0 mL) was added triethylamine (0.077 mL,
0.550 mmol, 5.0 equiv.), tert-butyl acrylate (0.064 mL,
0.440 mmol, 4.0 equiv.), tri-o-tolylphosphine (5.0 mg,
0.017 mmol, 0.15 equiv.), and tris(dibenzylideneacetone)-
dipalladium(0) (5.0 mg, 0.006 mmol, 0.05 equiv.) and heated
to 90 ЊC. The reaction mixture was stirred for 12 h and then
cooled to ambient temperature. The reaction mixture was then
diluted with EtOAc (5 mL) and washed with water (3 × 5 mL)
and brine (1 × 5 mL). The combined organic phase was dried
over MgSO4, filtered, concentrated and purified by column
chromatography (silica, 0
105 (fexaramate, 41.5 mg, 75%).
105: Rf = 0.40 (silica, 25% ethyl acetate in hexane); FT-IR
(neat) νmax 2977, 2931, 2855, 1713, 1640, 1483, 1446, 1393, 1367,
1323, 1279, 1209 cmϪ1; 1H NMR (600 MHz, CDCl3) δ 7.55 (d,
To a solution of 103 (50.0 mg, 0.089 mmol, 1.0 equiv.) in CHCl3
(2.0 mL) at 25 ЊC was added NaOH (2.0 M, 0.3 mL) and
adogen 464 (5.0 mg, ca. 0.1 equiv.). The resulting reaction
mixture was stirred for 6 h and quenched with water (5 mL).
The aqueous phase was then extracted with CH2Cl2 (3 × 5 mL).
The combined organic phase was dried over MgSO4, filtered and
concentrated and used without further purification (85% yield
1
by crude H NMR analysis). To the resulting dichlorocyclo-
propane (0.073 mmol, 1.0 equiv.) in N,N-dimethylformamide
(2.0 mL) was added triethylamine (0.051 mL, 0.366 mmol,
5.0 equiv.), methyl acrylate (0.026 mL, 0.292 mmol, 4.0 equiv.),
tri-o-tolylphosphine (11.1 mg, 0.037 mmol, 0.5 equiv.), and tris-
(dibenzylideneacetone)dipalladium(0) (13.4 mg, 0.015 mmol,
0.2 equiv.) sequentially and heated to 90 ЊC. The reaction mix-
ture was stirred for 24 h and then cooled to ambient temper-
50% EtOAc in hexanes) to afford
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
918

View Online
J = 15.8 Hz, 1 H), 7.50 (d, J = 16.2 Hz, 1 H), 7.44 (d, J = 7.4 Hz,
1 H), 7.37 (d, J = 7.9 Hz, 2 H), 7.32 (t, J = 7.9 Hz, 1 H), 7.13
(d, J = 7.9 Hz, 2 H), 7.05 (s, 1 H), 6.93 (d, J = 7.0 Hz, 1 H), 6.29
(d, J = 16.3 Hz, 1 H), 4.82 (s, 2 H), 3.76 (s, 3 H), 2.12–2.06 (m,
1 H), 1.68–1.60 (m, 4 H), 1.57–1.50 (m, 3 H), 1.48 (s, 9 H), 1.17–
1.10 (m, 1 H), 0.97–0.89 (m, 2 H); 13C NMR (150 MHz, CDCl3)
δ 175.9, 166.8, 166.1, 143.2, 142.9, 139.4, 135.8, 133.7, 130.0,
129.8, 129.0, 127.9, 127.6, 127.2, 120.0, 119.0, 80.3, 60.6, 52.4,
51.7, 46.6, 41.5, 31.4, 28.0, 25.3; HRMS calcd for C31H37NO5
[M ϩ Hϩ] 504.2744, found 504.2764.
to 90 ЊC for a period of 24 h. The microreactors were then
pooled and the reaction solvent was decanted prior to washing
the microreactors with MeOH (3 × 500 mL), CH2Cl2 (3 ×
500 mL), and Et2O (3 × 500 mL). Finally, each microreactor
was sorted into an individual reaction vessel and cleaved upon
suspension in Et2O and subsequent treatment with a solution
of NaOMe in MeOH (approx. 10 equiv.) at 25 ЊC for a period
of 20 min. The reactions were quenched with brine, extracted
with Et2O, concentrated and each compound was purified by
preparatory thin layer chromatography (PTLC). Each com-
pound was analyzed using LCMS which gave an average purity
of 95% for the library and 93/93 parent mass peaks found.
LCMS traces along with the corresponding mass spectra are
available upon request.
General procedure for the solid phase synthesis of 93-membered
library of biaryl and stilbene cinnamates (compounds 121, 125,
126, 174–212, and 213–264, Scheme 4)
This library was constructed via directed split-and-pool
techniques using IRORI MacroKans. The microreactors
were initially filled with commercially availabe Merrifield resin
(110 mg, 0.91 mmol gϪ1). After encoding, all 93 microreactors
were suspended in N,N-dimethylformamide (900 mL) and
treated with Boc-protected cinnamic acid 168 (4.94 g, 18.8
mmol, 2.0 equiv.), CsCO3 (6.13 g, 18.8 mmol, 2.0 equiv.), and
TBAI (1.73 g, 4.7 mmol, 0.5 equiv.) and heated to 55 ЊC. After
24 h, the reaction mixture was cooled to ambient temperature
and the reaction solvent was decanted prior to washing the
microreactors with MeOH (3 × 500 mL), CH2Cl2 (3 × 500 mL),
and Et2O (3 × 500 mL). Subsequently all microreactors were
pooled and suspended in CH2Cl2 (1000 mL) at 25 ЊC and treated
with trifluoroacetic acid (200 mL). After 1 h, the reaction mix-
ture was quenched with Et3N (200 mL) and the reaction solvent
was decanted prior to washing the microreactors with MeOH
(3 × 500 mL), CH2Cl2 (3 × 500 mL), and Et2O (3 × 500 mL).
The microreactors were then pooled and resuspended in THF–
MeOH (2 : 1, 1000 mL) at 25 ЊC and treated with 4-bromo-
benzaldehyde (17.4 g, 94.0 mmol, 10.0 equiv.) and acetic acid
(30 mg, 0.47 mmol, 0.05 equiv.). After 1 h, NaCNBH3 (4.72 g,
75.2 mmol, 8.0 equiv.) was added and the resulting reaction was
stirred a further 2 h. The reaction solvent was then decanted and
the microreactors were washed with MeOH (3 × 500 mL),
CH2Cl2 (3 × 500 mL), and Et2O (3 × 500 mL).
At this point the microreactors were sorted into one of three
reaction vessels and subjected to one of two acylation proto-
cols. The microreactors of two of the reaction vessels were
suspended in CH2Cl2 (500 mL) at 25 ЊC and treated with
either cyclohexanecarbonyl or isobutyryl chloride (94.0 mmol,
30.0 equiv.), Et3N (17.4 mL, 124 mmol, 40.0 equiv.), and 4-
DMAP (380 mg, 3.1 mmol, 1.0 equiv.) and stirred for 12 h. The
microreactors of the remaining reaction vessel were suspended
in N,N-dimethylformamide (350 mL) and treated with iso-
propyl isocyanate (8.0 g, 94.0 mmol, 30.0 equiv.), Et3N
(17.4 mL, 124 mmol, 40.0 equiv.), and 4-DMAP (380 mg,
3.1 mmol, 1.0 equiv.), heated to 60 ЊC and stirred for 60 h. The
microreactors were then cooled and the reaction solvent was
decanted prior to washing the microreactors with MeOH (3 ×
500 mL), CH2Cl2 (3 × 500 mL), and Et2O (3 × 500 mL). The
microreactors were then sorted into one of 31 reaction vessels
to be treated with either one of 13 commercially available
styrenes or one of 18 commercially available boronic acids.
For Heck coulings: The microreactors were suspended in
N,N-dimethylformamide (100 mL) and treated with a stryrene
(2.4 mmol, 8.0 equiv., see Fig. S-15† for the identities of
styrenes), Et3N (0.42 mL, 3.0 mmol, 10.0 equiv.), tri-o-tolyl-
phosphine (138 mg, 0.45 mmol, 1.5 equiv.), and tris(dibenzyl-
ideneacetone)dipalladium(0) (138 mg, 0.15 mmol, 0.5 equiv.)
and heated to 90 ЊC for a period of 48 h. For Suzuki couplings:
The microreactors were suspended in N,N-dimethylformamide
(100 mL) and treated with a boronic acid (2.4 mmol, 8.0 equiv.,
see Fig. S-15† for the identities of boronic acids), CsCO3
(293 mg, 0.9 mmol, 3.0 equiv.), and tetrakis(triphenylphos-
phine)palladium(0) (173 mg, 0.15 mmol, 0.5 equiv.) and heated
Full characterization of representative members of the four
classes of potent, selective FXR agonists (Fig. 9)
128: Rf = 0.43 (silica, 25% ethyl acetate in hexane); FT-IR (neat)
νmax 3058, 2927, 2854, 1715, 1642, 1598, 1588, 1483, 1435, 1398,
1318, 1271, 1230, 1201 cmϪ1 1H NMR (600 MHz, CDCl3)
;
δ 7.58 (d, J = 15.8 Hz, 1 H) 7.44 (d, J = 7.4 Hz, 1 H), 7.42 (d,
J = 7.9 Hz, 2 H), 7.31 (t, J, = 7.9 Hz, 1 H), 7.13 (s, 1 H), 7.12
(s, 2 H), 7.05 (d, J = 3.1 Hz, 1 H), 6.96 (d, J = 7.0 Hz, 1 H), 6.67
(d, J = 2.6 Hz, 1 H), 6.33 (d, J = 16.2 Hz, 1 H), 4.83 (s, 2 H), 3.76
(s, 3 H), 2.46 (s, 3 H), 2.16–2.07 (m, 1 H), 1.70–1.64 (m, 4 H),
1.62–1.48 (m, 3 H), 1.20–1.11 (m, 1 H), 0.99–0.92 (m, 2 H); 13
C
NMR (150 MHz, CDCl3) δ 175.8, 166.8, 143.3, 143.0, 141.4,
139.3, 136.1, 135.8, 133.7, 129.9, 129.0, 127.1, 126.0, 125.3,
122.7, 118.9, 60.5, 52.3, 51.6, 41.5, 29.4, 25.4, 25.2, 15.3, 14.0;
HRMS calcd for C29H31NO3S [M ϩ Naϩ] 496.1917, found
496.1924.
121: Rf = 0.35 (silica, 25% ethyl acetate in hexane); FT-IR
(neat) νmax 2928, 2854, 1717, 1646, 1597, 1578, 1508, 1489, 1448,
1397, 1318, 1268, 1200 cmϪ1 1H NMR (400 MHz, CDCl3)
;
δ 7.60 (d, J = 16.1 Hz, 1 H), 7.49 (d, J = 7.4 Hz, 2 H), 7.47–7.43
(m, 1 H), 7.41 (d, J = 7.9 Hz, 2 H), 7.34 (t, J = 7.6 Hz, 3 H), 7.24
(t, J = 7.3 Hz, 1 H), 7.15 (d, J = 7.9 Hz, 1 H), 7.11–7.07 (m, 3 H),
6.98 (d, J = 7.0 Hz, 1 H), 6.34 (d, J = 16.1 Hz, 1 H), 4.86 (s, 2 H),
3.79 (s, 3 H), 2.17–2.10 (m, 1 H), 1.71–1.52 (m, 7 H), 1.22–1.14
(m, 1 H), 1.00–0.88 (m, 2 H); 13C NMR (100 MHz, CDCl3)
δ 176.0, 167.0, 143.5, 143.1, 137.3, 137.0, 136.5, 135.9, 130.0,
129.1, 128.6, 128.2, 127.8, 127.6, 127.3, 126.5, 126.4, 119.1,
60.7, 52.5, 51.8, 41.7, 29.5, 25.6, 25.4, 14.3; HRMS calcd for
C32H33NO3 [M ϩ Hϩ] 480.2533, found 480.2534.
210: Rf = 0.10 (silica, 25% ethyl acetate in hexane); FT-IR
(neat) νmax 2930, 1715, 1644, 1530, 1446, 1398, 1318, 1174 cmϪ1
;
1H NMR (400 MHz, CDCl3) δ 7.58 (d, J = 16.2 Hz, 1 H), 7.46
(d, J = 7.3 Hz, 1 H), 7.39–7.26 (m, 3 H), 7.17–7.07 (m, 4 H), 6.97
(d, J = 6.8 Hz, 1 H), 6.76 (d, J = 16.2 Hz, 1 H), 6.32 (d, J =
16.1 Hz, 1 H), 4.84 (s, 2 H), 3.76 (s, 3 H), 2.51 (s, 3 H), 2.14–2.06
(m, 1 H), 1.70–1.50 (m, 7 H), 1.22–1.12 (m, 1 H), 1.00–0.85 (m,
2 H); 13C NMR (100 MHz, CDCl3) δ 176.0, 167.0, 150.5, 149.5,
143.4, 143.1, 137.4, 135.9, 132.9, 132.0, 131.8, 130.8, 130.0,
129.2, 127.8, 127.2, 126.4, 119.1, 118.2, 60.7, 52.5, 51.8, 41.7,
29.4, 25.5, 15.4; HRMS calcd for C30H32N2O3S [M ϩ Hϩ]
501.2206, found 501.2202.
125: Rf = 0.33 (silica, 25% ethyl acetate in hexane); FT-IR
(neat) νmax 3025, 2926, 2854, 1714, 1643, 1597, 1580, 1488, 1435,
1394, 1359, 1318, 1269, 1231, 1202 cmϪ1; 1H NMR (600 MHz,
CDCl3) δ 7.59 (d, J = 16.2 Hz, 1 H), 7.52–7.44 (m, 5 H), 7.35
(t, J = 7.4 Hz, 1 H), 7.28 (d, J = 8.3 Hz, 2 H), 7.21 (d, J = 8.3 Hz,
2 H), 7.12 (br s, 1 H), 7.01 (d, J = 7.0 Hz, 1 H), 6.32 (d, J =
16.2 Hz, 1 H), 4.88 (s, 2 H), 3.77 (s, 3 H), 2.49 (s, 3 H), 2.16–2.11
(m, 1 H), 1.70–1.65 (m, 3 H), 1.62–1.50 (m, 4 H), 1.20–1.15
(m, 1 H), 0.96–0.90 (m, 2 H); 13C NMR (150 MHz, CDCl3)
δ 175.9, 166.9, 143.3, 143.0, 139.4, 137.5, 137.4, 136.4, 135.8,
129.9, 129.7, 129.0, 127.6, 127.2, 126.8, 126.7, 119.0, 60.6, 52.4,
51.7, 41.6, 29.4, 25.4, 25.3, 15.7; HRMS calcd for C31H33NO3S
[M ϩ Naϩ] 522.2073, found 522.2053.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
919

View Online
2 A. Chawia, E. Saez and R. M. Evans, Cell, 2000, 103, 1–4.
3 H. Tu, A. Y. Okamoto and T. Hua, Trends Cardiovasc. Med., 2000,
10, 30–35.
4 T. T. Lu, J. J. Repa and D. J. Mangelsdorf, J. Biol. Chem., 2001, 276,
37735–37738.
5 P. A. Edwads, H. R. Kast and A. M. Anisfeld, J. Lipid. Res., 2002,
43, 2–12.
6 M. Makishima, A. Y. Okamoto, J. J. Repa, H. Tu, R. M. Learned,
A. Luk, M. V. Hull, K. D. Lustig, D. J. Mangelsdorf and B. Shan,
Science, 1999, 284, 1362–1365.
7 D. J. Parks, S. G. Blanchard, R. K. Bledsoe, G. Chandra, T. G.
Consler, S. A. Kliewer, J. B. Stimmel, T. M. Willson, A. M. Zavacki,
D. D. Moore and J. M. Lehmann, Science, 1999, 284, 1365–1368.
8 H. Wang, J. Chen, K. Hollister, L. C. Sowers and B. M. Forman,
Mol. Cell, 1999, 3, 543–553.
9 B. Goodwin, S. A. Jones, R. R. Price, M. A. Watson, D. D. McKee,
L. B. Moore, C. Galardi, J. G. Wilson, M. C. Lewis, M. E. Roth,
P. R. Maloney, T. M. Willson and S. A. Kliewer, Mol. Cell, 2000,
6, 517–526.
10 T. Lu, M. Makishima, J. J. Repa, K. Schoonjans, T. A. Kerr,
J. Auwerx and D. J. Mangelsdorf, Mol. Cell, 2000, 6, 507–515.
11 H. R. Kast, C. M. Nguyen, C. J. Sinal, S. A. Jones, B. A. Laffitte,
K. Reue, F. J. Gonzalez, T. M. Willson and P. A. Edwards, Mol.
Endocrin., 2001, 15, 1720–1728.
12 C. S. Song, I. Echchgadda, B.-S. Baek, S. C. Ahn, T. Oh, A. K. Roy
and B. Chatterjee, J. Biol. Chem., 2001, 276, 42549–42556.
13 H. R. Kast, B. Goodwin, P. T. Tarr, S. A. Jones, A. M. Anisfeld,
C. M. Stolz, P. Tontonoz, S. Kliewer, T. M. Willson and P. A.
Edwards, J. Biol. Chem., 2002, 277, 2908–2915.
14 M. Ananthanarayanan, N. Balasubramanian, M. Makishima,
D. J. Mangelsdorf and F. J. Suchy, J. Biol. Chem., 2001, 276,
28857–28865.
15 B. A. Laffitte, H. R. Kast, C. M. Nguyen, A. M. Zavacki,
D. D. Moore and P. A. Edwards, J. Biol. Chem., 2000, 275, 10638–
10647.
16 N. L. Urizar, D. H. Dowhan and D. D. Moore, J. Biol. Chem., 2000,
275, 39313–39317.
17 J. Grober, I. Zaghini, H. Fujii, S. A. Jones, S. A. Kliewer,
T. M. Willson, T. Ono and P. Besnard, J. Biol. Chem., 1999, 274,
29749–29754.
18 T. M. Willson, S. A. Jones, J. T. Moore and S. A. Kliewer, Med. Res.
Rev., 2001, 21, 513–522.
19 P. R. Maloney, D. J. Parks, C. D. Haffner, A. M. Fivush,
G. Chandra, K. D. Plunket, K. L. Creech, L. B. Moore, J. G. Wilson,
M. C. Lewis, S. A. Jones and T. M. Willson, J. Med. Chem., 2000,
43, 2971–2974.
20 For another potent steroid-derived agonist of FXR see: R.
Pellicciari, S. Fiorucci, E. Camaioni, C. Clerici, G. Costantino,
P. R. Maloney, A. Morelli, D. J. Parks and T. W. Willson, J. Med.
Chem., 2002, 45, 3569–3572.
21 K. C. Nicolaou, J. A. Pfefferkorn, A. J. Roecker, G.-Q. Cao,
S. Barluenga and H. J. Mitchell, J. Am. Chem. Soc., 2000, 122,
9939–9953.
22 K. C. Nicolaou, J. A. Pfefferkorn, H. J. Mitchell, A. J. Roecker,
S. Barluenga, G.-Q. Cao, R. L. Affleck and J. E. Lillig, J. Am. Chem.
Soc, 2000, 122, 9954–9967.
23 K. C. Nicolaou, J. A. Pfefferkorn, S. Barluenga, H. J. Mitchell,
A. J. Roecker and G.-Q. Cao, J. Am. Chem. Soc., 2000, 122,
9968–9976.
24 K. C. Nicolaou, J. A. Pfefferkorn, F. Schuler, A. J. Roecker,
G.-Q. Cao and J. E. Casida, Chem. Biol., 2000, 7, 979–992.
25 K. C. Nicolaou, A. J. Roecker, S. Barluenga, J. A. Pfefferkorn and
G.-Q. Cao, ChemBioChem, 2001, 6, 460–465.
26 G. J. Fraser, W. Xiang, R. Mercuri, C. Illy, B. R. Bowen, Y. He and
M. Sillis, J. Biol. Screen., 2002, 7, 3–10.
27 L. Xu, C. K. Glass and M. G. Rosenfeld, Curr. Opin. Gen. Dev.,
1999, 9, 140–147.
28 C. K. Glass, D. W. Rose and M. G. Rosenfeld, Curr. Opin. Cell Biol.,
1997, 9, 222–232.
29 D. L. Boger, J. Hong, M. Hikota and M. Ishida, J. Am. Chem. Soc.,
1999, 121, 2471–2477.
30 The lead compounds were analyzed for their ability to modulate the
activity of the following nuclear receptors: RXRα, PPARα, PPARγ,
PPARδ, PXR, SXR, LXRα, TRβ, RARβ, CARR, ERR3, VDR:
M. Downes, M. Verdecia, A. J. Roecker, R. Hughes, J. B. Hogenesch,
H. R. Kast, A. E. Bowman, J-L. Ferrer, A. M. Anisfeld,
P. A. Edwards, J. M. Rosenfeld, J. G. A. Alvarez, J. P. Noel,
K. C. Nicolaou and R. M. Evans, Mol. Cell, 2003, in press.
31 T. M. Willson and S. A. Kliewer, J. Lipid Res., 2002, 43, 359–364.
32 X. Wen, A. Radominska-Pandya, Y. Shi, C. M. Simon,
M. C. Nelson, E. S. Ong, D. J. Waxman and R. M. Evans, Proc.
Natl. Acad. Sci. U.S.A., 2001, 98, 3375–3380.
192: Rf = 0.35 (silica, 25% ethyl acetate in hexane); FT-IR
(neat) νmax 2928, 2854, 1716, 1652, 1508, 1485, 1449, 1397, 1318,
1268, 1231, 1200 cmϪ1; 1H NMR (400 MHz, CDCl3) δ 7.59 (d,
J = 16.7 Hz, 1 H), 7.56–7.53 (m, 1 H), 7.45 (d, J = 7.6 Hz, 1 H),
7.41 (d, J = 7.9 Hz, 2 H), 7.33 (t, J = 1 H), 7.22–7.06 (m, 7 H),
7.06–6.94 (m, 2 H), 6.33 (d, J = 16.1 Hz, 1 H), 4.85 (s, 2 H), 3.76
(s, 3 H), 2.17–2.09 (m, 1 H), 1.72–1.50 (m, 7 H), 1.21–1.12
(m, 1 H), 1.00–0.88 (m, 2 H); 13C NMR (100 MHz, CDCl3)
δ 176.0, 167.0, 161.6, 159.1, 143.4, 137.3, 136.4, 135.9, 130.4,
130.0, 129.1, 128.8, 127.7, 127.3, 127.0, 125.2, 125.0, 124.1,
120.8, 119.1, 115.9, 115.6, 52.5, 51.8, 41.7, 29.5, 25.6, 25.4, 14.2;
HRMS calcd for C32H32FNO3 [M ϩ Hϩ] 498.2439, found
498.2450.
259: Rf = 0.27 (silica, 25% ethyl acetate in hexane); FT-IR
(neat) νmax 2928, 1716, 1646, 1609, 1579, 1539, 1504, 1446, 1396,
1357, 1319, 1268 cmϪ1; 1H NMR (400 MHz, CDCl3) δ 7.61 (d,
J = 15.8 Hz, 1 H), 7.51–7.41 (m, 5 H), 7.35 (t, J = 7.6 Hz, 1 H),
7.17 (d, J = 7.9 Hz, 2 H), 7.13 (br s, 1 H), 7.00 (d, J = 7.0 Hz,
1 H), 6.79 (d, J = 7.9 Hz, 2 H), 6.34 (d, J = 16.1 Hz, 1 H), 4.87 (s,
2 H), 3.79 (s, 3 H), 2.99 (s, 6 H), 2.18–2.10 (m, 1 H), 1.70–1.50
(m, 7 H), 1.22–1.15 (m, 1 H), 1.00–0.90 (m, 2 H); 13C NMR (100
MHz, CDCl3) δ 175.9, 167.0, 143.5, 143.2, 140.2, 135.9, 135.2,
130.1, 130.0, 129.1, 127.8, 127.6, 127.2, 126.2, 119.0, 112.7,
60.7, 52.5, 51.8, 41.7, 40.6, 29.5, 25.6, 25.5, 14.3; HRMS calcd
ؒ
for C32H36N2O3 [M ϩ ] 496.2720, found 496.2715.
244: Rf = 0.25 (silica, 25% ethyl acetate in hexane); FT-IR
(neat) νmax 2930, 2855, 1716, 1652, 1599, 1579, 1504, 1486, 1445,
1398, 1318, 1226 cmϪ1; 1H NMR (400 MHz, CDCl3) δ 7.60 (d,
J = 16.2 Hz, 1 H), 7.46 (d, J= 7.6 Hz, 1 H), 7.40 (d, J = 7.9 Hz,
2 H), 7.35 (t, J = 8.0 Hz, 1 H), 7.18 (d, J = 7.9 Hz, 2 H), 7.10
(br s, 1 H), 7.05–7.00 (m, 3 H), 6.84 (d, J = 7.6 Hz, 1 H), 6.31 (d,
J = 15.8 Hz, 1 H), 5.96 (s, 2 H), 4.87 (s, 2 H), 3.78 (s, 3 H), 2.19–
2.10 (m, 1 H), 1.70–1.50 (m, 7 H), 1.25–1.15 (m, 1 H), 1.00–0.90
(m, 2 H); 13C NMR (100 MHz, CDCl3) δ 176.0, 167.0, 148.0,
147.0, 143.5, 143.1, 140.0, 136.2, 135.9, 130.1, 130.0, 127.7,
127.4, 126.9, 120.5, 119.0, 108.5, 107.6, 101.1, 52.5, 51.8,
41.7, 29.5, 25.6, 25.4; HRMS calcd for C31H31NO5 [M ϩ Hϩ]
498.2275, found 498.2269.
149: Rf = 0.25 (silica, 25% ethyl acetate in hexane); FT-IR
(neat) νmax 2932, 1720, 1642, 1580, 1454, 1399, 1341, 1318, 1269,
1202 cmϪ1; 1H NMR (400 MHz, CDCl3) δ 7.52 (d, J = 16.1 Hz,
1 H), 7.39 (d, J = 7.0 Hz, 1 H), 7.29 (t, J = 7.4 Hz, 1 H), 7.02 (br
s, 2 H), 6.96 (d, J = 7.6 Hz, 1 H), 6.83 (d, J = 7.0 Hz, 1 H), 6.28
(d, J = 16.1 Hz, 1 H), 4.95–4.85 (m, 2 H), 3.78 (s, 3 H), 3.39
(s, 3 H), 2.81 (d, J = 10.8 Hz, 1 H), 2.15 (d, J = 10.8 Hz, 1 H),
2.15–2.05 (m, 1 H), 1.67 (s, 6 H), 1.60–1.49 (3 H), 1.25–1.10 (m,
5 H), 1.00–0.85 (m, 2 H); 13C NMR (100 MHz, CDCl3) δ 176.0,
166.9, 147.6, 144.0, 143.5, 143.0, 135.6, 130.4, 130.0, 129.8,
127.8, 127.3, 124.8, 121.5, 118.9, 118.6, 71.3, 62.0, 60.4, 51.8,
46.4, 41.7, 40.9, 31.0, 29.5, 29.4, 26.9, 26.6, 25.6, 25.4; HRMS
calcd for C31H35Cl2NO5 [M ϩ Naϩ] 594.1784, found 594.1790.
Acknowledgements
Financial support for this work was provided by The Skaggs
Institute for Chemical Biology and the National Institutes
of Health (USA), fellowships from the American Chemical
Society Division of Organic Chemistry (sponsored by Novartis)
(to A. J. R.), the American Chemical Society Division of
Medicinal Chemistry (to R. H.), the Department of Defense
(to J. P.), and the American Chemical Society Division of
Medicinal Chemistry (to J. P.), and grants from Amgen, Bayer
AG, Boehringer-Ingelheim, Glaxo, Hoffmann-La Roche,
DuPont, Merck, Novartis, Pfizer, and Schering Plough.
References
1 R. W. Mahley and T. P. Bersot, in The Pharmacological Basis of
Therapeutics, eds. L. S. Goodman and A. Z. Gilman, McGraw Hill,
New York, 1996, 9th edition, ch. 36.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 9 0 8 – 9 2 0
920