Ligand Dependent Switch from RXR Homo- to RXR-NURR1 Heterodimerization

Article abstract of DOI:10.1021/acschemneuro.7b00216

Retinoid X receptors (RXRs) play key roles in many physiological processes in both the periphery and central nervous system. In addition, RXRs form heterodimers with other nuclear receptors to exert their physiological effects. The nuclear receptor relate

Full text of DOI:10.1021/acschemneuro.7b00216

Subscriber access provided by HKU Libraries
Article
Ligand dependent switch from RXR homo- to RXR-NURR1 heterodimerization
Marcel Scheepstra, Sebastian A. Andrei, Rens M. J. M. de Vries, Femke A. Meijer, Jian-Nong
Ma, Ethan S Burstein, Roger Olsson, Christian Ottmann, Lech-Gustav Milroy, and Luc Brunsveld
ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.7b00216 • Publication Date (Web): 10 Jul 2017
Downloaded from http://pubs.acs.org on July 15, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Chemical Neuroscience is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 49
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
Ligand dependent switch from RXR homo- to RXR-
NURR1 heterodimerization
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Marcel Scheepstra,† Sebastian A. Andrei,† Rens M. J. M. de Vries,† Femke A. Meijer,† Jian-Nong
Ma,§ Ethan S. Burstein,§ Roger Olsson,‡ Christian Ottmann,† Lech-Gustav Milroy,†* and Luc
Brunsveld†*
† Department of Biomedical Engineering and Institute for Complex Molecular Systems,
Laboratory of Chemical Biology, Technische Universiteit Eindhoven, Den Dolech 2, 5612 AZ
Eindhoven, The Netherlands
§ ACADIA Pharmaceuticals Inc., San Diego, California 92130, United States
‡ Chemical Biology & Therapeutics, Department of Experimental Medical Science, Lund
University, SE-221 00 Lund, Sweden
ACS Paragon Plus Environment
1

ACS Chemical Neuroscience
Page 2 of 49
1
2
3
4
5
6
7
8
9
Abstract Retinoid X Receptors (RXRs) play key roles in many physiological processes in both
the periphery and central nervous system. In addition, RXRs form heterodimers with other
nuclear receptors to exert their physiological effects. The Nuclear Receptor Related 1 protein
(NURR1) is particularly interesting because of its role in promoting differentiation and survival
of dopamine neurons. However only a small number of RXR-heterodimer selective modulators
are available, with limited chemical diversity. This work describes the synthesis, biochemical
evaluation and structural elucidation of a novel series of RXR ligands with strongly biased
interactions with RXRα–NURR1 heterodimers. Targeted modifications to the small molecule
biaryl scaffold caused local RXRα side-chain disturbances and displacement of secondary
structural elements upon ligand binding. This resulted in the repositioning of protein helices in
the heterodimer interface of RXRα, alterations in homo- versus heterodimer formation, and
modulation of activation function 2 (AF2). The data provide a rationale for the design of RXR
ligands consisting of a highly conserved hydrophilic region, strongly contributing to the ligand
affinity, and a variable hydrophobic region, which efficiently probes the effects of structural
changes at the level of the ligand on co-regulator recruitment or the RXRα–NURR1 dimerization
interface.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Key Words NUCLEAR RECEPTORS; HETERODIMERIZATION; LIGAND BINDING
DOMAIN; RETINOID X RECEPTOR; NUCLEAR RECEPTOR RELATED 1
ACS Paragon Plus Environment
2

Page 3 of 49
ACS Chemical Neuroscience
1
2
3
4
Introduction.
5
6
7
8
9
The retinoid X receptor (RXR) plays a key role as a transcriptional regulator through formation
of heterodimers with other nuclear receptor partners. Activation of RXR heterodimers exerts
neuroprotective effects in animal models of neurodegenerative disorders such as Alzheimer’s
disease, Parkinson’s disease and multiple sclerosis.1–4 The activity of RXRs is influenced by a
class of compounds related to the naturally occurring 9-cis-13,14-dihydroretinoic acid.5 The L-
shaped ligand binding pocket is unique to RXRs, and structural information derived from X-ray
crystallographic data of the RXR ligand binding domains in the apo- or holo- state6,7 has aided
the design of specific ligands for this receptor.8 Comprehensive overviews of the structure,
biology and therapeutic implications of targeting RXRs with small molecule ligands are
available.8–12 However, the chemical diversity of these ligands is limited by the structural
constraints placed by the RXR ligand binding pocket and the availability of synthetic
methodologies to access designed ligands.13 Controlled RXR heterodimerization and RXR
partial agonism are contemporary biomedical challenges,11,14 both of which could in principle be
addressed via exploration of an appropriate structural class.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Three RXR subtypes are known – RXRα (subject of this study), RXRβ, and RXRγ (NR2B1,
NR2B2, and NR2B3) – and all three interact similarly with many co-regulator proteins, and with
several nuclear receptors to form heterodimers. Ligand binding to RXR can induce the
transcriptional activity of some of its heterodimeric partner receptors (NURR1, NGFIB, FXR,
LXR, CAR, and PPAR).15 Thus, RXR-selective ligands that only activate specific RXR
heterodimers may have greater therapeutic potential, because they would be expected to cause
fewer side effects compared to ligands that cause general activation of RXR–NR
heterodimers.5,9,11 In this respect, RXR–PPAR and RXR–LXR heterodimers have gained a lot of
ACS Paragon Plus Environment
3

ACS Chemical Neuroscience
Page 4 of 49
1
2
3
4
5
6
attention, as the clearance of Aβ through apoE in Alzheimer’s mouse models is believed to be
facilitated by the activation of these heterodimers.16,17 For example, the ligand LG101506 was
7
8
9
identified as the most potent of a series of selective RXR–PPAR heterodimer activators18,19
,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
whereas it did not activate the RXR-LXR heterodimer. The potential of this selective RXR
modulator as a treatment for type 2 diabetes with reduced side effects was shown in mice.
Activation of RXR–nuclear receptor heterodimers with RXR receptor ligands also provides an
important strategy for activating orphan nuclear receptors which themselves do not readily bind
ligands (e.g. NURR1).
The Nuclear Receptor Related 1 protein or NURR1 (also referred to as NR4A2) controls the
development, function, and survival of dopaminergic neurons.20–22 NURR1 knockout mice have
reduced dopaminergic neurons and show perinatal lethality.23,24 Loss of function NURR1 genetic
polymorphisms in patients are linked to familial Parkinson’s disease.25,26 The crystal structure of
the LBD of NURR1 reveals a ‘closed’ ligand binding pocket, with the C-terminal helix 12 in a
canonical fold analogous to agonist bound ligand binding domains,27 which may not easily allow
access to small molecule ligands.28 Therefore, modulation of NURR1 activity via RXRs would
constitute a viable entry point for NURR1 activitation.4,29–31 A limited number of RXR–NURR1
heterodimer activators have been reported with promising selectivity over other RXR-
heterodimer pairs.4,32,33 Notwithstanding these important advances in developing selective RXR-
heterodimer modulators, there is a demand for a broader portfolio of RXR–NURR1 modulators,
both for use as research tools, to address open question such as regarding the resulting effect on
RXR signaling via other heterodimers and the potential presence of physiological ligands with
similar profiles, and as potential drugs.22
ACS Paragon Plus Environment
4

Page 5 of 49
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
Here we report the synthesis, biophysical evaluation and structural elucidation of a new series
of RXRα ligands with a strong bias towards promoting RXRα–NURR1 heterodimer versus
RXRα–RXRα homodimer formation. The work makes use of the knowledge derived from
studies on the natural product34 honokiol derived RXRα agonists described previously.35 The
biaryl scaffold of this series is straightforward to derivatize, which in principle enables a rapid
pharmacophore mapping of the RXRα ligand binding pocket. In this present study, ligands 1 – 5
(Figure 1) were designed to probe the hydrophobic region of the RXRα ligand binding pocket,
whilst keeping the polar interactions intact, to access the flexibility and local displacements of
amino acid side-chains of the ligand binding domain. Ortho-substituted ligands 6 and 7 were
synthesized to investigate how constraining the conformation of the biaryl system would affect
the activity of these ligands. Finally, a canonical side-chain extension36 of the biaryl scaffold in
the direction of the RXRα helix 12 produced compounds 9 and 10, for the purpose to investigate
for possible RXRα partial and full antagonism. This compact set of chemical probes was
subsequently evaluated as RXRα ligands, using a fluorescence-based polarization assay and
cellular mammalian two-hybrid assay. In addition, their effectiveness in modulating RXRα–
NURR1 heterodimers over RXRα–RXRα homodimers was tested using cell-based
bioluminescence resonance energy transfer (BRET2) assays.2,33 To corroborate these results and
to elucidate the binding mode and conformational changes in the protein induced by these
compounds, the X-ray structures of five of these novel ligands bound to RXRα were solved.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
5

ACS Chemical Neuroscience
Page 6 of 49
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Designed biaryl-based RXR ligands 1-10 & established potent RXR agonist
LG100268.
ACS Paragon Plus Environment
6

Page 7 of 49
ACS Chemical Neuroscience
1
2
3
4
Results and Discussion
5
6
7
8
9
Synthesis. The synthesis and co-crystal structure of ligand 1 is described previously.35 The
allyl side chain of 1 partly occupies the lipophilic pocket in the ligand binding domain of RXRα,
analogously to the tetramethyl-cyclohexene unit found in typical RXR ligands, leading to closure
of the ligand binding pocket via repositioning of helix 12 in an agonist conformation and
subsequent increased binding towards co-activators. The high binding affinity and low molecular
weight of 1 makes it an ideal scaffold to explore modifications targeting the lipophilic pocket
(Table 1). Ligands 2–5 were therefore synthesized bearing structural variations in the
hydrophobic side chain. The synthesis of compounds 2–5 relied on efficient palladium-catalyzed
cross coupling reactions (Scheme 1). The cinnamic acid derivative 11 was treated with thionyl
chloride in methanol to obtain the methyl cinnamate derivative 12. Four boronic acids or esters
were then reacted with 12 using a Suzuki coupling (Buchwald-modified) to provide
intermediates 13–16 in excellent yields (78–99%).35,37 The biaryls 13–16 were thereafter
demethylated using boron tribromide and hydrolyzed using sodium hydroxide, yielding ligands
2–5 in reasonable yields with high purities after a preparative HPLC purification. Molecules 6
and 7 were designed and synthesized to access the contorted conformation necessary for the
biaryl ligands to fit within the ligand binding pocket of RXRα.8 To enable a more efficient
synthesis, the methyl protection of the phenol (11) was replaced by a methoxymethyl acetal
(MOM) group (17). Subsequent Miyaura borylation enabled the key palladium cross coupling
with 19 or 20, accessible in a single step in moderate yields (51–57%). The biaryl products were
then treated with dilute hydrochloric acid (3 M) to deprotect the phenol in quantitative yields.
Finally, the methyl ester was efficiently hydrolyzed using sodium hydroxide, yielding ligands 6
and 7.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
7

ACS Chemical Neuroscience
Page 8 of 49
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Synthesis of novel biaryl RXRα ligands. The syntheses of 1 and 8 have been
previously reported.35 Conditions: (a) thionyl chloride, MeOH, 0 °C; (b) arylboronic acid or
arylboronic ester, Pd2(dba)3, SPhos, KF, dioxane/H2O (10:1 v/v), 110 °C; (c) BBr3, CH2Cl2 -78
°C. (d) NaOH, dioxane/MeOH (14:5 v/v), 40 °C; (e) Bis(pinacolato)diboron, Pd(OAc)2, XPhos,
KOAc, dioxane 110 °C; (f) 19 or 20, Pd2(dba)3, SPhos, KF, dioxane/H2O (6:1 v/v), 110 °C; (g)
23 or 24, Pd2(dba)3, SPhos, KF, dioxane/H2O (6:1 v/v), 110 °C; (h) HCl, THF, room
temperature; (i) NaOH, dioxane/MeOH (14:5 v/v).
ACS Paragon Plus Environment
8

Page 9 of 49
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
In our efforts to selectively drive RXRα homodimers, and not RXRα–NURR1 heterodimers,
towards an antagonistic conformation, we applied the previously validated strategy by Nahoum
and coworkers for inducing RXR antagonism.36 Compound 8 (Figure 1) was previously
described by us and found to lack significant RXRα activity, because of the additional polar
phenolic functionality, which points towards a lipophilic environment.35 X-ray crystallographic
data have demonstrated that alkylation of the appropriate phenol displaces the position of helix
12 towards a (partial) antagonistic fold, influencing the position of L436, which plays a
determining role in the communication with helix 12.36,38 Therefore, using the biaryl scaffold,
agonist 8 was modified with two different length alkoxy chains. The length of the alkoxy chain
was hypothesized to be directly correlated with the displacement of helix 12 and therefore its
antagonist properties.36 The antagonists 9 and 10 were synthesized using intermediate 18 for the
Suzuki couplings, using the ligands introduced by Buchwald,35,37 with either 23 or 24.
Intermediates 23 and 24 were each made in two steps in excellent yields via sp2-sp3 Pd-catalyzed
cross couplings on the bromide group to introduce the allyl-substituted group. Finally,
intermediates 25 and 26 were treated with hydrochloric acid in THF for the deprotection of the
MOM-group and subsequently with sodium hydroxide for the hydrolysis of the methyl ester to
provide the antagonists 9 and 10.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
9

ACS Chemical Neuroscience
Page 10 of 49
1
2
3
4
5
6
7
8
9
Pharmacological evaluation. The activity of the ligands on RXRα was initially profiled using
a fluorescence-based co-activator recruitment polarization (FP) assay and then in a more
biologically relevant mammalian two-hybrid (M2H) assay.. The FP and M2H assays revealed an
EC50(FP) = 260 nM and EC50(M2H) = 6.3 nM for 1, which compares favorably to the established,
but more bulky, full agonist LG100268: EC50(FP) = 150 nM and EC50(M2H) = 5.1 nM (Table 1).
The differences in measured EC50 affinities between the FP assay and the M2H assay are a
common phenomenon because of intrinsic differences between the two assay formats, the protein
concentrations and the co-regulator peptide.35,39 Besides 1, ligands 2 – 5 also displayed full
agonism in both assays with the measured affinities (EC50) in the nanomolar range (Table 1).
Replacing the allyl group with the closely related i-propyl (4), or n-propyl (5) did not strongly
affect the ligand affinity for RXRα. The aromatic phenyl (2) and benzyl (3) substituents
displayed a 10-fold decrease in potency in the M2H cell-based assay compared to the smaller
propyl substituents. Nevertheless, also these ligands still activate RXRα with nanomolar
potencies. The RXRα binding of 1–5 is thus dominated by the hydrophilic portion of the biaryl
ligands, i.e. the hydroxyl-cinnamic acid moiety. The hydrophobic substituents tune the ligand
affinity but are not crucial for high ligand affinity. As such, ligand modifications at this part of
the molecule could provide an entry to affect the homo- vs. heterodimer preference of RXRα.
The terphenyl ligand 2, provides an interesting platform to study the importance and effects of
the rotation around the two phenyl-phenyl bonds. Addition of a single methyl group at the central
phenyl (6 and 7) was thought to direct the rotation to preferred orientations in complex with
RXRα. Ligands 6 and 7 both displayed full agonism in the FP assay and the M2H assay, albeit
with potencies in the 1-10 ꢀM range (Table 1). Comparison with the nanomolar affinities
observed for ligand 2, demonstrates that the addition of the single methyl groups strongly
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
10

Page 11 of 49
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
impacts affinity for RXRα. This very strict SAR is typical for this class of biphenyls, as in our
previous studies the placement of a hydroxyl functionality, such as in 8, similarly impacted
affinity by changes over 100-fold.35 In the case of 6 and 7 the decrease in affinity might be
caused by a suboptimal conformational match of the ligand for binding to the protein in the
conformation befitting the binding pocket (vide infra).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We previously studied ligand 8, which demonstrated full agonism, but with only moderate
affinity for RXRα in FP as well as M2H assays.35 The design of 9 and 10 prompted us to study
these compounds in a competition format to profile their antagonist characteristics. RXRα was
therefore stimulated with the full agonist LG100268 and the subsequent impact of ligands 9 and
10 on co-regulator recruitment was studied via fluorescence polarization studies and on
transcription via M2H assays. The addition of ligands 9 or 10 to the agonist-stimulated RXRα
resulted in decreased fluorescence polarization, indicating lowered co-activator recruitment via
displacement of the agonist and stabilization of an inactive RXRα conformation. Ligand 10
demonstrated full antagonism, while ligand 9 showed partial antagonism (Figure 3a). Consistent
with its partial antagonist activity, ligand 9 also displayed partial agonism in an agonist assay
format. The measured affinities of 9: IC50(FP) = 48.5 4.6 ꢀM and 10: IC50(FP) = 46.9 5.9 ꢀM at
the RXRα receptor were approximately 25-fold lower than the known antagonist UVI3003:
IC50(FP) = 1.8 0.6 ꢀM. Compounds 9 and 10 were also RXRα antagonists in the cell-based M2H
assay. M2H competition experiments showed a clear decrease in luciferase expression upon
addition of either ligand in the 10 – 40 ꢀM range after stimulation with agonist LG100268
(Figure 3b).
ACS Paragon Plus Environment
11

ACS Chemical Neuroscience
Page 12 of 49
1
2
3
4
5
6
Table 1. Summary of fluorescence polarization (FP) and mammalian two-hybrid (M2H) data for
the RXRα agonists
7
8
9
Compound
FP / EC50 (nM)
M2H (Luciferase) / EC50 (nM)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
150 40
260 110
140 23
5.1 2.0
6.3 4.0
LG100268
1 (R = CH2CH=CH2)
2 (R = Ph)
3 (R = Bn)
4 (R = iPr)
5 (R = nPr)
6
85
9
142
89
9
92 36
7
5.8 1.8
18 10
170 80
9900 2500
1020 60
> 2500
14600 1800
7
EC50 values for LG100268 and ligands 1–7. See experimental section for details of the assays.
The 20- to 30-fold difference between the FP and M2H data for the more potent compounds is a
common phenomemon,35,40 which can be explained by intrinsic differences between the two
different assay formats, in particular, the different protein and peptide concentrations used.
ACS Paragon Plus Environment
12

Page 13 of 49
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
The potency and efficacy of the biaryl ligands to induce RXRα–RXRα homodimer and RXRα–
NURR1 heterodimer conformational changes was determined using cellular BRET2 assays2,33
(Table 2). Agonist ligands 1–5 all displayed remarkably strong potencies (single digit and sub
nanomolar EC50’s) and high efficacies, comparable to the chemically optimized agonist
LG100268 (Figure 1). Ligands 1–5 all feature preferential affinity for RXRα–NURR1
heterodimers. Ligand 4 displayed an encouraging 25-fold higher potency at RXRα–NURR1 over
RXRα–RXRα, with a pEC50 of 9.1. It should be noted that in these same assays both the well-
studied RXR ligand bexarotene and the recently developed dihydrobenzofuran-based ligands
only showed 2- to 7-fold selectivity.2,33 In contrast, the methylated terphenyl ligands 6 and 7
showed higher potency at RXRα–RXRα homodimers. The addition of the single methyl group to
2 ortho to the biphenyl bonds, resulting in 6 and 7, thus leads to a reversal in homo- vs.
heterodimer affinity.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The potency and efficacy of the antagonistic biaryl ligands 9 and 10 to selectively induce
RXRα–RXRα homodimers over RXRα–NURR1 heterodimers towards antagonistic
conformational changes, was also determined. Partial antagonist 9 shows a profile similar to 6
and 7, but with further biased interactions towards RXRα homodimers over RXRα–NURR1
heterodimers by a factor of 10. Additionally, ligand 9 featured a lower efficacy consistent with
its partial antagonist character. The full antagonist 10 did not show appreciable activity in the
BRET2 assays.
ACS Paragon Plus Environment
13

ACS Chemical Neuroscience
Page 14 of 49
1
2
3
4
5
Table 2. Pharmacological evaluation of hetero- and homodimerization using BRET2 assays.
6
7
8
9
Compound
RXRα–NURR1
RXRα–RXRα
NURR1–RXRα
vs. RXRα–RXRα
selectivity
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
pEC50 (SD) %Eff (SD)
pEC50 (SD)
%Eff (SD)
fold
9.3 (0.3)
8.6 (0.2)
8.5 (0.2)
8.6 (0.2)
9.1 (0.4)
8.3 (0.2)
6.1 (0.2)
7.1 (0.4)
6.8 (0.2)
<5.0
100 (7)
129 (14)
148 (15)
178 (18)
129 (4)
141 (9)
330 (38)
123 (12)
41 (25)
32 (19)
8.2 (0.2)
7.5 (0.1)
7.5 (0.1)
7.7 (0.1)
7.7 (0.1)
7.0 (0.0)
6.8 (0.2)
7.5 (0.1)
7.7 (0.2)
<5.0
100 (11)
269 (32)
283 (15)
286 (44)
291 (28)
286 (37)
236 (11)
175 (24)
66 (6)
13
13
10
8
LG100268
1
2
3
25
20
0.2
0.4
0.1
-
4
5
6
7
9
42 (18)
10
The BRET2 assay was performed as described previously.33,41
ACS Paragon Plus Environment
14

Page 15 of 49
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
Structural evaluation. The co-crystallization of ligands 1, 3 and 4 with RXRα showed the
canonical interactions of the carboxylate group of the ligands with Arg316, the backbone
nitrogen of Ala327 and a conserved water molecule (Figure 2a). The free hydroxyl group on the
ligands makes a hydrogen bond with Asn306. This hydrogen bonding network is conserved for
all the ligands and directs the positioning of the hydrophobic part of the molecules. The
hydrophobic component of 1, 3, and 4 occupies the lipophilic region of the ligand binding
pocket. In this region, ligand-dependent RXRα amino acid reorientations can be observed.
Especially ligand 4 (i-propyl substitution) repositions Ile324, Val332, Ser336 and Val342
compared to ligands 1 and 3, creating a smaller ligand binding pocket (Figure 2a, zoom and
Supporting Figure 60). This tighter packing of helices is less pronounced for the region around
Ile 324 , but mostly affects helices 6 and 7, showing amino acid displacements up to 2.8 Å, and
the end of RXRα helix 11, and with that the loop between helix 11 and 12. The carboxy-terminal
part of helix 11 has been identified to play a pivotal role in the dimerization of RXRs,42–44 such
as via polar contacts between the C-terminal carboxylic acid of PPAR Helix 12 and lysine 431 of
RXR Helix 10/11.45 NURR1 features an atypical, longer, helix 1227 which, following a modeled
RXR-NURR1 heterodimer46 and published RXR-PPAR crystal structures,45,47 probably points
towards the RXRα LBD, notably RXRα helices 7 and 11. RXRα–NURR1 heterodimerization
thus implies repositioning of RXRα structural elements in this region to accommodate binding of
the NURR1 helix 12. Ligand 4 shows the strongest bias towards RXRα–NURR1
heterodimerization (Table 2). The repositioning of RXRα helices 7 and 11 by the compact ligand
structure might therefore explain its strong selectivity for heterodimerization.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Terphenyl ligands 6 and 7 were provided with a methyl functionality at the ortho-position at
either of the two biphenyl bonds. Compared to non-methylated 2, ligands 6 and 7 featured
ACS Paragon Plus Environment
15

ACS Chemical Neuroscience
Page 16 of 49
1
2
3
4
5
6
7
8
9
decreased affinity for RXRα and bias towards RXRα–RXRα homodimerization. The X-ray co-
crystallization of ligands 6 and 7 with RXRα showed the canonical NR fold, bound to a co-
regulator peptide (Figure 2b). The biphenyl core scaffold of 6 and 7 was more out-of-plane
rotated in comparison to the other agonists. Ligands 6 and 7 fit the canonical L-shaped ligand
binding pocket via the same hydrogen bonding network, but compared to, for example, ligand 4
induce significant changes in the positioning of several amino acids and helices 6, 7, and 11 and
their connecting loops, which form the lipophilic region of the binding pocket (Figure 2b). More
specifically, compared to ligand 4, the sidechain residues of Val332, Ser336 and Val342 are
displaced by 1.0 to 2.2 Å when either ligand 6 or 7 are bound (Supporting Figures 60-62). The
ortho-methyl substituent on ligand 6 causes the phenyl ring to rotate more out of plane than the
other ligands due to steric clash between the methyl and phenyl groups (Supporting Figure 61).
This specific out-of-plane orientation of the third phenyl group is apparently unfavorable for the
binding of ligand 6 as illustrated by the 10-fold lower RXRα activity compared to 7 (Table 1).
The additional ortho-methyl substituent on ligand 7 points into the direction of helix 11,
displacing Leu436 (Supporting Figure 61). This steric interaction is unfavorable in the context of
the phenyl substitution pattern, as highlighted by the strong drop in RXRα activity of 7 compared
to ligand 2. Combined, the methyl substituents on 6 and 7 result in both cases in repositioning of
RXRα structural elements, which may correlate with a lower bias towards RXRα–NURR1
heterodimerization as observed in the BRET2 assay.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
16

Page 17 of 49
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Crystal structures of RXRα with agonists 1, 3, 4, 6 and 7. a) Overlay of the X-ray co-
crystal structures of ligands 1 (blue, PDB: 4OC7),35 3 (orange, PDB: 5MJ5, 1.9 Å resolution),
and 4 (grey, PDB: 5MKU, 1.8 Å resolution) bound in the ligand binding pocket of RXRα in
ribbon representation with the TIF2 derived co-regulator peptide. Zoom-in on the ligand binding
ACS Paragon Plus Environment
17

ACS Chemical Neuroscience
Page 18 of 49
1
2
3
4
5
6
7
8
9
pocket of RXRα with the amino acids represented as sticks showing the interactions and
displacements. b) Overlay of the X-ray co-crystal structures of ligands 6 (light green, PDB:
5MMW, 2.7 Å resolution) and 7 (dark green, PDB: 5MK4, 2.0 Å resolution) bound to the ligand
binding pocket of RXRα in ribbon representation with the TIF2 derived co-regulator peptide.
Zoom-in on the ligand binding pocket of RXR with the amino acid represented as sticks showing
the interactions between the ligands and the protein. c) Final 2Fo-Fc electron density maps
(contoured at 1σ) of ligands 1, 3, 4, 6 and 7.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
18

Page 19 of 49
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
To gain structural information about the effects of the additional alkoxy group of 9 and 10 the
X-ray structure of the RXRα ligand binding domain complexed with ligand 9 and the co-
regulator peptide TIF2 was solved. The protein was crystallized in the agonist conformation,
stabilized by interactions with the co-regulator peptide, and reflected the partial (ant)agonist
character of 9. Comparison of this structure with that of agonist 1 and its analogs (vide supra)
revealed, as expected,36 reorientations of amino acid residues related to helix 12 positioning
(Figure 3c,d). The most dominant reorientation affected Leu436 in helix 11 induced by the n-
propoxy chain, which rotated around 3.0 Å towards helix 12, notably L455. Leu436 has been
described as a key residue for the communication between the ligand and helix 12 and the
correlated activation function-2 (AF2).9,36,48 The repulsive interaction between the L436 and
L455 will lower the association strength between helix 12 and the ligand binding domain,
shifting the equilibrium of conformations of the receptor in solution towards an antagonist
conformation as demonstrated by the FP and M2H data (Figure 3a,b). The effect might be
proportional to the chain length and provides a reasonable rationale for the full antagonism of
ligand 10.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The position of the biaryl scaffold of 9 in the RXRα ligand binding pocket is shifted compared
to full agonists like 1 and 4 (Figure 3d). This results in concomitant shifts of RXRα helices 6 and
7, especially when compared to ligand 4 and analogous to ligands 6 and 7. Similar to ligand 7,
the double substitution pattern on the biaryl scaffold allows partial agonist 9 to simultaneously
address two parts of the ligand binding pocket, via helices 6 and 7 and via helices 11 and 12. The
combined addressing of the different parts of the RXRα pocket via the hydrophobic substitution
pattern leads to amino acid shifts that correlate with preferential affinity for RXRα–RXRα
homodimers (Table 2).
ACS Paragon Plus Environment
19

ACS Chemical Neuroscience
Page 20 of 49
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Ligands 9 and 10 are antagonists at RXRα. a) Fluorescence polarization data showing
antagonist activity for 9 and 10. Ligand 9 displays partial antagonism, while 10 acts as a full
antagonist in a competition assay against LG100268 (50 nM). b) Cellular antagonist activities of
9 and 10 measured in an M2H luciferase competition assay against LG100268 (10 nM). c) X-ray
co-crystal structures of ligand 9 (purple, PDB: 5MKJ, 2.5 Å resolution) bound in the ligand
binding pocket of RXRα in ribbon representation with the TIF2 derived co-regulator peptide.
Ligand 9 is shown with final 2Fo-Fc electron density maps (contoured at 1σ). Note that part of
RXRα Helix 11 is omitted to allow visualization of the ligand. d) Zoom-in on an overlay of the
ACS Paragon Plus Environment
20

Page 21 of 49
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
ligand binding pockets of RXRα co-crystallized with ligands 1 (blue), 4 (grey), and 9 (purple).
Relevant helices and amino acids and their displacements are shown. Compared to 4, ligand 1
induces an outward expansion of helices 6 and 7. Next to this, ligand 9 induces an additional
outward displacement of leucines 436 and 455 on helices 11 and 12.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Conclusions.
Despite the fact that RXR receptors play major roles in many biological processes through
heterodimerization with other nuclear receptors, only a small number of small molecule RXR-
heterodimer selective modulators are available, with limited chemical diversity and biophysical
properties. This study has delivered a compact and focused selection of RXRα–NURR1 agonists
based on a versatile biaryl scaffold, structurally different than previously reported molecules.10 In
earlier work using chiral dihydrobenzofuran acids, we demonstrated a 3-fold biased interaction
with RXRα–NURR1.33 The biaryl scaffold presented here provides a 25-fold selectivity bias for
RXRα–NURR1 in the case of analog 4, and a >100-fold switch in homo- vs. heterodimer
selectivity when comparing analogs 4 & 9 in Table 2. The structural elucidation of five of these
novel RXRα ligands, in our view, provides a first rationale towards understanding how to
generate RXRα–NURR1 heterodimer selective ligands. Key ligand-protein interactions and
correlated side-chain displacements were identified, modulating both selective dimerization and
co-regulator recruitment. Interactions of the ligand with key amino acid chains such as Ile324,
Val332, Ser336 and Val342 on helices 6 and 7 tune the size of the ligand binding pocket. These
compact ligands bind RXRα in a manner that allows movement of helix 7 and 11 to a generate a
compact ligand binding pocket conformation which arguable is more suited for
ACS Paragon Plus Environment
21

ACS Chemical Neuroscience
Page 22 of 49
1
2
3
4
5
6
7
8
9
heterodimerization with NURR1, potentially by enabling the accommodation of the long
NURR1 helix 12. Interactions of substituents on the biaryl scaffold with RXRα amino acids
involved in formation of the AF2, such as Leu436, induce helix 12 repositioning and translate
into lower ligand affinities or, alternatively, into (partial) antagonist properties. These
interactions with helices 11 and 12 are matched by the expansion of the RXRα ligand binding
pocket via helices 6 and 7, leading to a selectivity of the biaryl scaffold for RXRα–RXRα
heterodimers.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
This novel series of ligands allows addressing a wide range of RXRα receptor conformations
and associated functional outcomes via substitution patterns on the same biaryl scaffold,
expanding the current RXR modulator repertoire with agonist as well as antagonist ligands. The
data provide a rationale for the design of RXR ligands comprised of a unique hydrophilic region
with a conserved hydrogen bonding network contributing to the binding affinity, and a
hydrophobic region to probe the other parts of the receptor influencing dimerization properties or
co-regulator recruitment. These findings justify further exploration of the ligand-controlled
homo- vs. heterodimerization of RXR and its interaction partners, for activation of the
NURR1:RXRα heterodimer as monotherapy for Parkinson’s disease,2,4 for delineating the
resulting physiological effects on other RXR heterodimers, and also for potentially revealing
conserved mechanisms for other nuclear receptors.
ACS Paragon Plus Environment
22

Page 23 of 49
ACS Chemical Neuroscience
1
2
3
4
Methods
5
6
7
8
9
Synthesis. All the solvents employed were commercially available and used without purification
unless stated otherwise. Water was purified using a Millipore purification train. All the reagents
are commercially available and used without purification. All the NMR data were recorded on a
Varian Gemini 400 MHz NMR, a Bruker Cryomagnet 400 MHz a Bruker UltraShield Magnet
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
400 MHz or a Varian 200 MHz (400 or 200 MHz for H-NMR and 100 or 50 MHz for 13C-
13
NMR). Proton experiments are reported in parts per million (ppm) downfield of TMS. All C
spectra were reported in ppm relative to residual chloroform (77 ppm) Analytical LC-MS was
performed on a C4, Jupiter SuC4300A, 150x2.00 mm column with a gradient 5% to 100%
acetonitrile in H2O supplemented with 0.1% v/v Formic Acid (F.A.) in 15 min. Silica column
chromatography was performed manually using silica with particle size 60 – 200 µm.
Preparative HP-LC was performed on a Gemini S4 110A 150x21.20 mm column using H2O and
acetonitrile supplemented with 0.1% v/v F.A. Purity and exact mass of the compounds were
determined using a High Resolution LC-MS system consisting of a Waters ACQUITY UPLC I-
Class system coupled to a Xevo G2 Quadrupole Time of Flight (Q-tof). The system was
comprised of a Binary Solvent Manager and a Sample Manager with Fixed-Loop (SM-FL).
compounds were separated (0.3 mL min-1) by the column (Polaris C18A reverse phase column
2.0 x 100 mm, Agilent) using a 15% to 75% acetonitrile gradient in water supplemented with
0.1% v/v F.A. before analysis in positive mode in the mass spectrometer. On the basis of LC-UV
data, all final compounds are ≥95% pure.
General procedure for Suzuki couplings method A for the synthesis of compounds 13-16,
21, 22, 25 & 26
ACS Paragon Plus Environment
23

ACS Chemical Neuroscience
Page 24 of 49
1
2
3
4
5
6
7
8
9
An oven dried Schlenk tube was charged with aryl halide (1.0 eq), boronic acid or boronic ester
(1.2 eq), KF (5.0 eq), SPhos (0.30 eq) and Pd2(dba)3 (0.010 eq). The Schlenk tube was evacuated
and backfilled with argon three times. Degassed dioxane/H2O (10:1 v/v, final aryl halide
concentration 0.2 M) was added under positive argon flow and the reaction was stirred at the
indicated temperature for the indicated time. The reaction mixture was then allowed to cool to
room temperature, passed through Celite with ethyl acetate and concentrated in vacuo. The crude
product was purified by flash silica gel chromatography using the indicated eluent and
concentrated in vacuo.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
General procedure for Suzuki couplings method B for the synthesis of 19 & 20
An oven dried Schlenk tube was charged with aryl halide (1.0 eq), boronic acid or boronic ester
(2.0 eq), KOAc (3.0 eq) and Pd(dppf)Cl2 (0.010 eq). The Schlenk tube was evacuated and
backfilled with argon three times. Degassed dioxane/H2O (5:1 v/v, final aryl halide concentration
0.4 M) was added under positive argon flow and the reaction was stirred at the indicated
temperature for the indicated time. The reaction mixture was then allowed to cool to room
temperature and was separated between CH2Cl2 and H2O. The aqueous layer was extracted twice
with CH2Cl2. The combined organic layers were washed with brine, dried over Na2SO4, filtered
and concentrated in vacuo. The crude product was purified by flash silica gel chromatography
using the indicated eluent and concentrated in vacuo.
General procedure for deprotection, method A: methyl ethers and methyl esters
The compound was dissolved in dry CH2Cl2 to a final concentration of ~0.25 M and cooled to -
78 °C. A solution of BBr3 (1 M in CH2Cl2, 2.0 eq) was added dropwise and the reaction was
stirred at -78 °C for one hour. The temperature was raised to 0°C and the reaction was stirred for
ACS Paragon Plus Environment
24

Page 25 of 49
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
another hour. The reaction was then allowed to warm to room temperature and quenched with
H2O. The aqueous layer was extracted with CH2Cl2 three times and the combined organic layers
washed with brine, dried over Na2SO4, filtered and concentrated in vacuo.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The crude product was then dissolved in dioxane/MeOH (14/5 v/v) to a concentration of ~0.20
M. To this mixture was added NaOH (4 N in deionized water, 3.0 eq) was added and the
resulting mixture was stirred overnight at either room temperature or 40 °C. The solvent was
then removed in vacuo and the residue separated in H2O and CH2Cl2. The aqueous layer was
extracted with CH2Cl2 three times and the combined organic layers were washed with brine,
washed over Na2SO4, filtered and evaporated in vacuo. The products were then purified by
preparative reversed-phase HPLC by UV detection.
General procedure for deprotection, method B: MOM ethers and methyl esters
The compound was dissolved in THF to a final concentration of ~0.2. A solution of HCl (6 N in
H2O, 3.0 eq) was added and the reaction was stirred at room temperature overnight. The reaction
was then diluted with H2O and extracted with Et2O three times. The combined organic layers
washed with saturated aqueous NaHCO3 and brine, dried over Na2SO4, filtered and concentrated
in vacuo.
The crude product was then dissolved in dioxane/MeOH (14/5 v/v) to a concentration of ~0.20
M. To this mixture was added NaOH (4 N in deionized water, 3.0 eq) was added and the
resulting mixture was stirred overnight at either room temperature or 40 °C. The solvent was
then removed in vacuo and the residue separated in H2O and CH2Cl2. The aqueous layer was
extracted with CH2Cl2 three times and the combined organic layers were washed with brine,
ACS Paragon Plus Environment
25

ACS Chemical Neuroscience
Page 26 of 49
1
2
3
4
washed over Na2SO4, filtered and evaporated in vacuo. The products were then purified by
5
6
preparative reversed-phase HPLC by UV detection and freeze-dried.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(E)-methyl 3-(6-methoxy-[1,1':3',1''-terphenyl]-3-yl)acrylate (13) The described procedure
for Suzuki couplings method A was used with (E)-methyl 3-(3-chloro-4-methoxyphenyl)acrylate
(120 mg, 0.529 mmol), [1,1'-biphenyl]-3-ylboronic acid (130 mg, 0.656 mmol), KF (155 mg,
2.67 mmol), SPhos (67 mg, 0.16 mmol) and Pd2(dba)3 (48 mg, 0.052 mmol) at 110 °C for 18 h.
The eluent used for purification was 17% v/v EtOAc in heptane to yield the title compound as a
colorless oil, 180 mg, 0.52 mmol, 98% yield. Silica gel TLC Rf = 0.25 (17% v/v EtOAc in
heptane); LC-MS (ESI): calc. for C23H20O3 [M+H]: 345.15 observed 345.08, LC Rt=8.52 min;
1H NMR (400 MHz, CDCl3) δ 7.74 – 7.66 (m, 2H), 7.66 – 7.60 (m, 2H), 7.60 – 7.53 (m, 2H),
7.52 – 7.47 (m, 3H), 7.47 – 7.41 (m, 2H), 7.40 – 7.31 (m, 1H), 6.98 (d, J = 8.5 Hz, 1H), 6.36 (d,
J = 16.0 Hz, 1H), 3.83 (s, 3H), 3.78 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 167.84, 158.44,
144.54, 141.26, 141.25, 138.25, 131.25, 130.64, 129.34, 128.86, 128.62, 128.50, 128.45, 127.42,
127.37, 127.33, 126.30, 115.76, 111.47, 55.87, 51.71..
(E)-3-(6-hydroxy-[1,1':3',1''-terphenyl]-3-yl)acrylic acid (2) The described procedure for
deprotection, method A was used with (E)-methyl 3-(6-methoxy-[1,1':3',1''-terphenyl]-3-yl) (13)
(100 mg, 0.29 mmol) to afford the title compound, 2, as a white amorphous powder after
preparative reverse-phase HPLC and subsequent freeze-drying (34 mg, 0.107 mmol, 37% over
two steps). LC-MS (ESI): calc. for C21H16O3 [M+H]: 317.12 observed 317.17, LC Rt=6.52 min;
1H NMR (400 MHz, CDCl3) δ 7.77 (d, J = 15.9 Hz, 1H), 7.71 – 7.55 (m, 5H), 7.54 – 7.42 (m,
13
5H), 7.42 – 7.36 (m, 1H), 7.04 (d, J = 8.7 Hz, 1H), 6.36 (d, J = 15.9 Hz, 1H); C NMR (100
ACS Paragon Plus Environment
26

Page 27 of 49
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
MHz, CDCl3) δ 171.82, 155.03, 146.63, 142.75, 140.50, 136.63, 130.84, 130.08, 129.81, 129.08,
128.83, 127.97, 127.92, 127.85, 127.35, 127.34, 127.31, 116.75, 115.09; HRMS (m/z): [M + H]+
calcd 317.1178, found 317.1179.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(E)-methyl 3-(3'-benzyl-6-methoxy-[1,1'-biphenyl]-3-yl)acrylate (14) The described
procedure for Suzuki couplings method A was used with (E)-methyl 3-(3-chloro-4-
methoxyphenyl)acrylate (97 mg, 0.43 mmol), 2-(3-benzylphenyl)-4,4,5,5-tetramethyl-1,3,2-
dioxaborolane (153 mg, 0.520 mmol), KF (125 mg, 2.15 mmol), SPhos (54 mg, 0.13 mmol) and
Pd2(dba)3 (39 mg, 0.040 mmol) at 110 °C for 18 h. The eluent used for purification was 17% v/v
EtOAc in heptane to yield the title compound as a colorless oil, 146 mg, 0.41 mmol, 95% yield.
Silica gel TLC Rf = 0.28 (17% v/v EtOAc in heptane); LC-MS (ESI): calc. for C24H22O3 [M+H]:
1
359.16 observed 359.08, LC, Rt=8.68 min; H NMR (400 MHz, CDCl3) δ 7.67 (d, J = 16.0 Hz,
1H), 7.48 – 7.42 (m, 2H), 7.37 – 7.25 (m, 5H), 7.24 – 7.14 (m, 4H), 6.92 (d, J = 8.3 Hz, 1H),
6.33 (d, J = 16.0 Hz, 1H), 4.02 (s, 2H), 3.77 (s, 3H), 3.77 (s, 3H); 13C NMR (100 MHz, CDCl3) δ
167.80, 158.38, 144.56, 141.07, 140.98, 137.83, 131.30, 130.61, 130.15, 129.08, 128.54, 128.28,
128.02, 127.31, 127.22, 126.17, 115.64, 111.41, 55.75, 51.67, 42.05.
(E)-3-(3'-benzyl-6-hydroxy-[1,1'-biphenyl]-3-yl)acrylic acid (3) The described procedure for
deprotection, method A was used with (E)-methyl 3-(3'-benzyl-6-methoxy-[1,1'-biphenyl]-3-
yl)acrylate (14) (60 mg, 0.17 mmol) to afford the title compound, 3, as a white amorphous
powder after preparative reverse-phase HPLC and subsequent freeze-drying (24 mg, 0.073
mmol, 43% over two steps). LC-MS (ESI): calc. for C22H18O3 [M+H]: 331.14 observed 331.00,
LC Rt=6.65 min; 1H NMR (400 MHz, CDCl3) δ 7.75 (d, J = 15.9 Hz, 1H), 7.49 – 7.40 (m, 3H),
7.33 – 7.19 (m, 8H), 6.99 (d, J = 8.4 Hz, 1H), 6.33 (d, J = 15.9 Hz, 1H), 4.05 (s, 2H); 13C NMR
(100 MHz, CDCl3) δ 172.48, 154.98, 146.74, 142.96, 140.63, 136.23, 130.81, 129.80, 129.66,
ACS Paragon Plus Environment
27

ACS Chemical Neuroscience
Page 28 of 49
1
2
3
4
5
6
129.57, 129.15, 129.05, 128.84, 128.78, 127.17, 126.75, 126.50, 116.63, 115.08, 42.09; HRMS
(m/z): [M + H]+ calcd 331.1334, found 331.1319.
7
8
9
(E)-methyl 3-(3'-isopropyl-6-methoxy-[1,1'-biphenyl]-3-yl)acrylate (15) The described
procedure for Suzuki couplings method A was used with (E)-methyl 3-(3-chloro-4-
methoxyphenyl)acrylate (145 mg, 0.640 mmol), (3-isopropylphenyl)boronic acid (128 mg, 0.780
mmol), KF (187 mg, 3.22 mmol), SPhos (81 mg, 0.20 mmol) and Pd2(dba)3 (61 mg, 0.066
mmol) at 110 °C for 18 h. The eluent used for purification was 15% v/v EtOAc in heptane to
yield the title compound as a colorless oil in a quantitative yield. Silica gel TLC Rf = 0.35 (15%
v/v EtOAc in heptane); LC-MS (ESI): calc. for C20H22O3 [M+H]: 311.16 observed 311.08, LC,
Rt=8.52 min; 1H NMR (399 MHz, CDCl3) δ 7.70 (d, J = 16.0 Hz, 1H), 7.52 – 7.45 (m, 2H), 7.41
– 7.31 (m, 4H), 6.96 (d, J = 8.5 Hz, 1H), 6.35 (d, J = 16.0 Hz, 1H), 3.83 (s, 3H), 3.79 (s, 3H),
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
13
2.96 (hept, J = 6.9 Hz, 1H), 1.29 (d, J = 6.9 Hz, 6H); C NMR (100 MHz, CDCl3) δ 167.86,
158.44, 148.73, 144.66, 137.65, 130.68, 129.06, 128.50, 128.09, 127.75, 127.24, 127.05, 125.56,
115.60, 111.42, 55.81, 51.69, 34.25, 24.14.
(E)-3-(6-hydroxy-3'-isopropyl-[1,1'-biphenyl]-3-yl)acrylic acid (4) The described procedure
for deprotection, method A was used with (E)-methyl 3-(3'-isopropyl-6-methoxy-[1,1'-biphenyl]-
3-yl)acrylate (15) (100 mg, 0.32 mmol) to afford the title compound, 4, as a white amorphous
powder after preparative reverse-phase HPLC and subsequent freeze-drying (45 mg, 0.16 mmol,
50%,over two steps). LC-MS (ESI): calc. for C18H18O3 [M+H]: 283.13 observed 283.17, LC,
Rt=6.43 min; 1H NMR (400 MHz, CDCl3) δ 7.77 (d, J = 15.9 Hz, 1H), 7.51 – 7.42 (m, 3H), 7.34
– 7.27 (m, 3H), 7.02 (d, J = 8.3 Hz, 1H), 6.34 (d, J = 15.9 Hz, 1H), 2.98 (d, J = 6.9 Hz, 1H), 1.30
(d, J = 6.9 Hz, 6H); 13C NMR (100 MHz, CDCl3) δ 172.43, 155.05, 150.66, 146.82, 135.97,
ACS Paragon Plus Environment
28

Page 29 of 49
ACS Chemical Neuroscience
1
2
3
4
5
6
130.79, 129.68, 129.65, 129.17, 127.26, 127.14, 126.80, 126.38, 116.57, 115.00, 34.34, 24.14;
HRMS (m/z): [M + H]+ calcd 283.1334, found 283.1334.
7
8
9
(E)-methyl
3-(4-methoxy-3-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl)acrylate
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(12b)35
1-bromo-3-propylbenzene To a solution of 3-bromopropiophenone (2.98 g, 14.0 mmol) in TFA
(30 mL, 0.40 M) was added dropwise triethylsilane (11.5 mL, 72.0 mmol) at 0 °C in 5 minutes
and the mixture was stirred for additional 20 minutes. The reaction mixture was heated to 80 °C
and stirred overnight. The reaction mixture was allowed to cool to room temperature and
concentrated in vacuo. Toluene was added and the mixture was again concentrated in vacuo to
obtain the crude material. The product was purified via flash silica gel chromatography eluting
1
with hexane to yield the title compound, 78 mg, 0.39 mmol, 3% yield. H NMR (400MHz,
CDCl3): δ (ppm) 7.38 – 7.27 (m, 2H), 7.18 – 7.06 (m, 2H), 2.59 – 2.49 (m, 2H), 1.72 – 1.56 (m,
2H), 0.93 (td, J = 7.3, 0.9 Hz, 3H); 13C NMR (100 MHz, CDCl3): 145.16, 131.65, 129.91,
128.87, 127.28, 122.45, 37.81, 24.49, 13.87.
(E)-methyl 3-(6-methoxy-3'-propyl-[1,1'-biphenyl]-3-yl)acrylate (16) The described
procedure for Suzuki couplings method A was used with (E)-methyl 3-(4-methoxy-3-(4,4,5,5-
tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl)acrylate ( 152 mg, 0.478 mmol) 1-bromo-3-
propylbenzene (78 mg, 0.39 mmol), KF (114 mg, 1.96 mmol), SPhos (49 mg, 0.12 mmol)
Pd2(dba)3 (37 mg, 0.040 mmol at 100 °C for 7 h. The eluent used for purification was 5%-10%
v/v EtOAc in hexane to yield the title compound as a dark yellow oil, 95 mg, 0.31 mmol, 78%
1
yield. H NMR (400 MHz, CDCl3): 7.69 (d, J = 16.0 Hz, 1H), 7.55 – 7.45 (m, 2H), 7.36 – 7.30
(m, 3H), 7.23 – 7.13 (m, 1H), 6.98 (d, J = 8.4 Hz, 1H), 6.35 (d, J = 16.0 Hz, 1H), 3.85 (s, 3H),
ACS Paragon Plus Environment
29