Discovery and SAR of Natural-Product-Inspired RXR Agonists with Heterodimer Selectivity to PPARδ-RXR

Article abstract of DOI:10.1021/acschembio.0c00146

A known natural product, magnaldehyde B, was identified as an agonist of retinoid X receptor (RXR) α. Magnaldehyde B was isolated from Magnolia obovata (Magnoliaceae) and synthesized along with more potent analogs for screening of their RXRα agonistic activities. Structural optimization of magnaldehyde B resulted in the development of a candidate molecule that displayed a 440-fold increase in potency. Receptor-ligand docking simulations indicated that this molecule has the highest affinity with the ligand binding domain of RXRα among the analogs synthesized in this study. Furthermore, the selective activation of the peroxisome proliferator-activated receptor (PPAR) δ-RXR heterodimer with a stronger efficacy compared to those of PPARα-RXR and PPARγ-RXR was achieved in luciferase reporter assays using the PPAR response element driven reporter (PPRE-Luc). The PPARδactivity of the molecule was significantly inhibited by the antagonists of both RXR and PPARδ, whereas the activity of GW501516 was not affected by the RXR antagonist. Furthermore, the molecule exhibited a particularly weak PPARδagonistic activity in reporter gene assays using the Gal4 hybrid system. The obtained data therefore suggest that the weak PPARδagonistic activity of the optimized molecule is synergistically enhanced by its own RXR agonistic activity, indicating the potent agonistic activity of the PPARδ-RXR heterodimer.

Full text of DOI:10.1021/acschembio.0c00146

pubs.acs.org/acschemicalbiology
Articles
Discovery and SAR of Natural-Product-Inspired RXR Agonists with
Heterodimer Selectivity to PPARδ‑RXR
Ken-ichi Nakashima,*,⊥ Eiji Yamaguchi,⊥ Chihaya Noritake, Yukari Mitsugi, Mayuki Goto, Takao Hirai,
Naohito Abe, Eiji Sakai, Masayoshi Oyama, Akichika Itoh, and Makoto Inoue
Cite This: https://dx.doi.org/10.1021/acschembio.0c00146
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: A known natural product, magnaldehyde B, was
identified as an agonist of retinoid X receptor (RXR) α.
Magnaldehyde B was isolated from Magnolia obovata (Magnolia-
ceae) and synthesized along with more potent analogs for
screening of their RXRα agonistic activities. Structural optimiza-
tion of magnaldehyde B resulted in the development of a candidate
molecule that displayed a 440-fold increase in potency. Receptor−
ligand docking simulations indicated that this molecule has the
highest affinity with the ligand binding domain of RXRα among
the analogs synthesized in this study. Furthermore, the selective
activation of the peroxisome proliferator-activated receptor
(PPAR) δ-RXR heterodimer with a stronger efficacy compared
to those of PPARα-RXR and PPARγ-RXR was achieved in
luciferase reporter assays using the PPAR response element driven reporter (PPRE-Luc). The PPARδ activity of the molecule was
significantly inhibited by the antagonists of both RXR and PPARδ, whereas the activity of GW501516 was not affected by the RXR
antagonist. Furthermore, the molecule exhibited a particularly weak PPARδ agonistic activity in reporter gene assays using the Gal4
hybrid system. The obtained data therefore suggest that the weak PPARδ agonistic activity of the optimized molecule is
synergistically enhanced by its own RXR agonistic activity, indicating the potent agonistic activity of the PPARδ-RXR heterodimer.
revealed the absence or broadening of signals corresponding to
residues within the ligand binding pocket and the AF2 surface,
suggesting that these regions do not distribute in a particular
conformation. In contrast, the detection or sharpening of the
corresponding signals in holo-NRs suggests stabilization of the
LBD in the active conformation due to ligand binding to NRs.
RXR heterodimers can be classified into “permissive” and
“nonpermissive” heterodimers,7−10 whereby the activation of
permissive heterodimers, such as PPAR-RXR and RXR-LXR, is
possible by the binding of RXR agonists alone. On the other
hand, nonpermissive heterodimers, such as RXR-RAR, RXR-
TR, and VDR-RXR, are silenced by the binding of RXR
agonists only. Analysis of the structural and dynamic
differences between PPARγ-RXRα, a “permissive” hetero-
dimer, and RXRα-TRβ, a “non-permissive” heterodimer using
NMR spectroscopy, X-ray crystallography, and hydrogen/
deuterium exchange mass spectrometry (MS), revealed that
INTRODUCTION
■
The retinoid X receptor (RXR) is an important member of the
ligand-dependent nuclear receptor (NR) family. RXR is a
unique and ubiquitous heterodimeric partner for the NR
signaling pathway, and it forms homodimers or heterodimers
with retinoic acid receptors (RAR), peroxisome proliferator-
activated receptors (PPARs), liver X receptors (LXRs), thyroid
hormone receptors (TRs), vitamin D receptors (VDRs),
nuclear receptor related-1 protein (Nurr1), and related
receptors.1,2 The NR-RXR heterodimers and the RXR
homodimer bind to specific response elements on DNA and
regulate the transactivation of target genes, and the molecular
dynamics mechanism underlying this ligand binding-induced
activation has been revealed in the structural analysis of
NRs.3−5 In the absence of ligands, corepressor proteins bind to
the surface of NRs and cause the repositioning of C-terminal
helix 12, which is a key structural component of the activating
function 2 (AF2) domain. The binding of agonists to the
ligand binding domain (LBD), which is composed of 12 α-
helices (H1−12) with two small β-sheets between H5 and H6
(β-turn), leads to a cascade of events culminating in the
repositioning and stabilization of H12 and several other helices
to form the AF2 surface, which possesses a binding capacity for
the LXXLL motif of coactivators. NMR studies into apo-NRs6
Received: February 28, 2020
Accepted: April 23, 2020
https://dx.doi.org/10.1021/acschembio.0c00146
© XXXX American Chemical Society
ACS Chem. Biol. XXXX, XXX, XXX−XXX
A

ACS Chemical Biology
pubs.acs.org/acschemicalbiology
Articles
TRβ reduces the stability of holo-RXRα LBD in the active
conformation by dimerization, whereas PPARγ does not
exhibit such an effect.6
Among the various known RXR heterodimers, the available
therapeutic approaches mostly target PPAR-RXR. The three
subtypes of the PPAR family, namely PPARα, PPARγ, and
PPARδ (also known as PPARβ/δ), differ in tissue distribution
and their role in biological systems.11,12 More specifically,
PPARγ is the most studied subtype of the PPAR family and is
expressed abundantly in adipose tissues, the large intestine, and
hematopoietic cells, and plays a critical role in the regulation of
adipogenesis, insulin sensitivity, and lipid metabolism.12,13
Thiazolidinediones (e.g., rosiglitazone and pioglitazone) are
the only class of clinically used PPARγ agonist therapeutics for
the treatment of type 2 diabetes mellitus to ameliorate insulin
resistance.14,15 PPARα is expressed abundantly in the liver,
colon crypt, and myocardium.12 Fibrates (e.g., bezafibrate and
fenofibrate) are PPARα agonists used clinically for the
treatment of hypercholesterolemia and hypertriglyceridemia.
Although proven therapies targeting PPARγ and PPARα are
known, prescription drugs for targeting PPARδ, which is
ubiquitously expressed in most tissues and is known to regulate
lipid metabolism and glucose homeostasis,12 do not exist due
to the serious side effects observed with the use of known
agonists.16,17 However, research aimed at addressing these
challenges is ongoing, and preclinical studies using certain
synthetic PPARδ agonists (e.g., GW501516 and GW0742)
have found benefits against obesity-induced insulin resistance,
type 2 diabetes mellitus, and cardiovascular disease.18−20 In
this context, the development of new PPARδ agonists to
reduce these adverse effects and conditions is highly important
and will be useful for treating type 2 diabetes mellitus, and as
such, is the objective of this study. Regardless of their effects in
permissive or nonpermissive heterodimers, RXR agonists
concertedly enhance transactivation as partner NR agonists.7,21
Although the functional elucidation of RXR has not been
completely achieved due to multiple signaling pathways
involving RXR heterodimers, RXR is considered an attractive
target for drug discovery.22 Indeed, bexarotene (targretin), a
selective RXR ligand, has been clinically used for the treatment
of cutaneous T cell lymphoma. In addition, 9-cis retinoic acid, a
ligand for both RXR and RAR, has been used in the treatment
of Kaposi’s sarcoma and chronic hand eczema. Furthermore,
various studies have indicated the beneficial effects of RXR
agonists against type 2 diabetes mellitus23,24 and Alzheimer’s
disease,25 although their efficiency for the treatment of
Alzheimer’s disease has been controversial.
Figure 1. Magnaldehyde B (1) is a more effective RXRα agonist than
honokiol. (a) Structures of honokiol and 1. (b) RXRα agonistic
activities of honokiol and 1 with bexarotene. Data are means SD
(SD = standard deviation) of three biological replicates.
displayed an Emax of 79.4% in the luciferase reporter assay
(Figure 1). Furthermore, we previously carried out receptor−
ligand docking simulations, which indicated that the carbonyl
group within drupanin (i.e., 3-prenyl-4-hydroxy-trans-cinnamic
acid) plays a vital role in the formation of hydrogen bonds with
Arg316,28 which is a key residue for activation of the
receptor.29,30 Magnaldehyde B (1) also contains a phenyl-
propanoid unit with an aldehyde group which points to the
importance of the aldehyde carbonyl functionality as an
essential structural element for the activation of RXR.
Recently, Scheepstra et al. reported that honokiol targets
both sides of the AF-2 domain in RXR, and they analyzed the
dual-binding behavior of honokiol in a twofold approach with a
ligand binding pocket side and a coactivator binding side being
the two binding modes.31 The same research group also
discovered potent agonists for the RXR homodimer and the
RXR-Nurr1 heterodimer.32 Although the activity and mode of
action of honokiol are known to a certain extent, no structure−
activity relationship (SAR) studies on the RXR agonistic
activity of honokiol derivatives have been reported prior to this
work, and so we sought to evaluate RXR agonistic activities of
1 and its analogs to improve their activities and render them
potential candidates for therapeutic development. Further-
more, we investigate the ability of the synthesized agonists to
activate the PPARδ-RXR heterodimer selectively over PPARα-
and PPARγ-RXR heterodimers by a nonconventional mecha-
nism.
RESULTS AND DISCUSSION
■
We have been interested in the discovery and evaluation of
RXR ligands from natural resources and have previously
reported studies that honokiol isolated from the bark of
Magnolia obovata (Magnoliaceae) is an RXR agonist.26
Honokiol activates RXRα at a half-maximal effective
concentration (EC50) of 16.5 μM in the luciferase reporter
assay (Figure 1). However, this compound showed a much
lower potency than that of bexarotene (EC50: 18.7 nM). In
addition, honokiol exhibited a maximum activation of only
28.6% relative to the 100 nM concentration of bexarotene. In
our quest to find agonists of higher potency and efficacy, we
carried out further chemical investigations of the roots of M.
obovata in an unpublished preliminary study. Indeed, we
identified a known neolignan magnaldehyde B (1)27 to be a
more effective against RXRα than honokiol. Magnaldehyde B
(1) activated RXRα with an EC50 value of 5.58 μM and
Structural Optimization of Magnaldehyde B. We
envisioned the retrosynthesis of the biaryl targets via the key
disconnection to the aryl bromide fragment A and the
arylboronic acid fragment B, which could be coupled together
in the forward synthesis employing a Suzuki−Miyaura coupling
protocol in a unified divergent approach to 1 and its analogs
(Scheme 1). The synthetic pathway for the preparation of 1−8
is shown in Scheme 2.
We began the synthesis of the magnaldehyde B analogs with
the preparation of a range of substituted fragment A
derivatives. The Horner-Wadsworth-Emmons reaction of the
TBS-protected aldehyde 19 afforded the desired fragment A
(20) in good yield. A subsequent DIBAL-mediated 1,2-
reduction of 20 followed by MnO2 oxidation of the resultant
alcohol afforded 21. The 1,4-reduction of 20a using CoCl2 and
NaBH4 furnished the reduced derivative 22a in 89% yield.
B
https://dx.doi.org/10.1021/acschembio.0c00146
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
pubs.acs.org/acschemicalbiology
Articles
a
Scheme 1. Retrosynthetic Analysis for Magnaldehyde B
Analogs
Scheme 2. Synthesis of 1−8
Hydrogenation of 20b under a hydrogen atmosphere with Pd/
C catalysis delivered 22b in a quantitative yield. The
corresponding aldehyde 23 was synthesized by the LiAlH4
reduction of 22, followed by a Dess−Martin oxidation of the
obtained alcohol. However, acetal protection of the formed
aldehyde 23b was required due to its instability.
The synthesis of fragment B began with the allylation of 4-
bromophenol 25, followed by a Claisen rearrangement and
TBS protection of the resultant phenol. The transformation of
26 through a halogen-metal exchange/borylation sequence
furnished the desired fragment B (27) in an excellent yield
(88% overall yield). With the two fragments in hand, a
Suzuki−Miyaura coupling was performed. The reaction of aryl
bromides 20−24 with boronic acid 27 using the Pd2(dba)3 and
X-Phos system in the presence of K2CO3 as the base in a
mixture of THF/H2O at 95 °C, followed by deprotection of
the coupled products afforded the corresponding biaryl
products 1−8 in moderate to good yields.
The RXRα agonistic activities of 1−8 are included in Table
1, whereby it is apparent that molecules bearing a hydroxy
group at the C-4 position (1, 3, 5, and 7) displayed
significantly lower EC50 values than those of their analogs
without this group (2, 4, 6, and 8, respectively). Furthermore,
the obtained results indicate that the presence of a double
bond between the C-8′ and C-9′ atoms was likely to negatively
impact the agonistic activity. In compounds 1−8, compound 7
exhibited the highest potency (EC50 = 260 nM) for the
activation of RXRα. Following our achievement of the optimal
substitutions for fragment A, we decided to further modify
fragment B of 7. In addition, we modified fragment A of 8 to
confirm the necessity for the presence of a hydroxy group at
the C-4 position.
As shown in Scheme 3, various modified fragment B boronic
acids were synthesized and employed in the Suzuki−Miyaura
cross-coupling reactions. The hydrogenation of 27 using Pd/C
proceeded quantitatively to afford the propyl substituted
boronic acid 28. Borylation of 3-allyl bromobenzene 29 was
performed to furnish corresponding des-hydroxy fragment B,
30, and subsequent hydrogenation of 30 delivered 3-propyl
benzeneboronic acid 31 in a quantitative yield. With all desired
fragments in hand, the Suzuki−Miyaura coupling of fragments
A and B was performed employing the previously developed
coupling conditions, giving the corresponding coupled
products 9−14 in good yields.
a
Reagents and conditions: (a) TBSCl, imidazole, DCM, 0 °C, then rt,
16 h, 98%; (b) NaH, ethyl 2-phosphonoacetate, THF, 0 °C then rt, 1
h, 20a 90%, 20b 85%; (c) DIBAL-H, THF, −78 °C; (d) MnO2,
DCM, rt, 2 days. 21a 48% (2 steps), 21b 98% (2 steps); (e) NaBH4,
CoCl2, MeOH, −10 °C, 30 min, 22a 89%; (f) Pd/C, H2, EtOH, rt, 16
h, 22b quant.; (g) LAH, THF, 0 °C, 1 h; (h) DMP, DCM, rt, 3 h, 23a
72% (2 steps); (i) ethylene glycol, PTSA, toluene, reflux, 16 h, 23b
58% (3 steps); (j) allyl bromide, K2CO3, DMF, 60 °C, 1 h, 99%; (k)
200 °C, 3 h, 92%; (l) TBSCl, imidazole, DMF, rt, 16 h, 97%; (m) n-
BuLi, THF, −78 °C 1 h then B(Oi-Pr)3, −78 °C to rt, 16 h, 90%; (n)
Pd2(dba)3, X-Phos, K2CO3, THF, H2O, 95 °C; (o) TBAF, THF, 0 °C
then rt, 16 h; (p) HCl aq., 3 h.
The RXRα agonistic activities of the synthesized series 9−14
were also examined, and the results are presented in Table 1.
Similar to their analogs 1−8, compounds bearing a hydroxy
group at the C-4 position (9, 11, and 13) displayed
significantly lower EC50 values than those of their nonhydroxy
analogs (10, 12, and 14, respectively). On the other hand,
molecules with a hydroxy group at the C-2′ position (7, 8, 11,
and 12) gave higher EC50 values than those of their
nonhydroxy analogs (9, 10, 13, and 14, respectively), which
indicates that the absence of the hydroxy group at the C-2′
position contributes to a reduction in the EC50 value.
Therefore, compound 13, which possessed the best sub-
stitution pattern, displayed the greatest potency (12.7 nM)
among the various test compounds evaluated herein.
Furthermore, the activity of 13 was comparable to that of
the full-agonist, bexarotene.
C
https://dx.doi.org/10.1021/acschembio.0c00146
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
pubs.acs.org/acschemicalbiology
Articles
a reduction in the EC50 value; (iii) the presence of a single
bond between C-7 and C-8 is more advantageous for activation
than a double bond; (iv) the absence of a hydroxy group at the
C-2′ position contributes to a reduction in the EC50 value; and
(v) reduction of the double bond between C-8′ and C-9′ may
likely improve the activity.
Table 1. RXRα Agonistic Activities and Calculated Docking
Affinities of 1−14
calculated affinity
a
b
c
compounds
EC50 (nM)
Emax (%)
100
(kcal/mol)
bexarotene
honokiol
1
2
3
4
5
6
7
8
9
10
11
12
13
14
18.7 2.1
16 500 1240
5580 820
8140 730
1310 130
10 300 100
824 190
2460 240
260 10
10 200 800
23.0 1.7
−12.9
−10.0
−10.5
−10.3
−10.4
−10.5
−10.1
−10.2
−10.6
−10.7
−11.1
−10.6
−10.8
−10.7
−11.2
−10.7
28.6 8.7
79.4 3.9
49.0 3.0
46.9 5.4
65.0 4.3
53.7 3.2
48.6 5.4
55.7 5.6
57.4 11.8
69.8 2.4
85.5 8.9
60.4 3.0
67.5 8.2
71.3 9.3
67.0 8.4
In addition, we synthesized compounds 15−17 to probe the
necessity for the presence of alkoxy groups and gain insight
into their ideal length. As no significant differences were
observed between 13 and 15−17 (Supplementary Figure 1), it
was apparent that the intensity of the agonistic activity was
independent of the presence or length of the n-alkoxy group.
Binding Mode Analysis of 13. To explore the molecular
basis of the interaction between RXRα and the prepared
synthetic ligands, we carried out receptor−ligand docking
simulations using the Autodock Vina program,33 and the best-
calculated affinities of the prepared compounds are displayed
in Table 1. Because docked binding poses of 1−14 were
similar to the crystal structure of the honokiol derivative
reported by Scheepstra et al., the results of our docking
simulations were likely to be real (Supplementary Figure 2).
Thus, compound 13 displayed the most potent activity among
the molecules evaluated in this study, and Figure 2 shows the
docking pose and the predicted interactions between RXRα
LBD and 13. Previous studies indicated that RXRα agonists
interacted with Arg316 and Ala327 via a carboxyl group by
forming hydrogen bonds, which contributed to an affinity for
the receptor.28−30 As expected, our docking studies showed
that the ester moiety of 13 forms three hydrogen bonds with
Arg316 and one hydrogen bond with Ala327. Although no
meaningful findings about the advantages of the single bond
between C-7 and C-8 were obtained from the results of our
docking studies, it was considered that it could affect the
orientation of the proton acceptors. Additionally, the docking
studies suggested hydrogen bond formation between the
hydroxy group at C-4 and the Asn306 residue, which
contributes to improved binding to the receptor, and results
in an enhanced potency. Furthermore, the aromatic ring within
fragment A was predicted to form a π−π T-shaped interaction
with the Phe313 residue. Our results also revealed that almost
all amino acid residues interacting with fragment A are
localized in H5 and in the β-turn between H5 and H6 (Figure
2b). Compared with the crystal structures of liganded RXRα
and those of apo-RXRα,3,34,35 only minor alterations in the
structures surrounding H5 and the β-turn were displayed
before and after ligand binding (Figure 2c). These findings
indicate that the structure of fragment A in the evaluated
molecules plays an important role in the binding of the
molecule to the receptor, and 3-substituted 4-hydroxy-7,8-
dihydrocinnamic acid is the optimal substitution of fragment A
for efficient binding to H5 and the β-turn.
4720 920
180 11
7070 2380
12.7 1.1
1810 80
a
EC50 values are presented as mean
SD of three independent
b
experiments (SD is standard deviation). Emax data are presented as
mean SD of the maximum activation relative to the bexarotene-
c
induced maximum activation for RXRα. Calculated affinities are best
scores obtained from Autodock Vina.
a
Scheme 3. Synthesis of 9−14
In contrast, the moiety derived from fragment B occupied
the hydrophobic part within the ligand-binding pocket of holo-
RXRα and appeared to interact with Ile268, Ile324, Ile345,
Phe346, Val349, and Cys432 through electrostatic interactions
(Figures 2a and 2d). From the crystallographic analysis, these
residues were determined to be localized in the H3 (Ile268), β-
turn (Ile324), H7 (Ile345, Phe346, and Val349), and H11
(Cys432) segments of the receptor. Among the helices, the
structures of H3 and H11 were drastically altered upon ligand
binding, as evaluated by structural studies of RXRα.3,4 When
the ligand is attracted to the ligand binding pocket of apo-
RXRα, H-11 is pushed by the ligand and is repositioned to
a
Reagents and conditions: (q) Pd/C, MeOH, rt, 3 h, quant; (r) n-
BuLi, B(Oi-Pr)3, THF, −78 °C to rt, 16 h, 57%; (s) Pd/C, MeOH, rt,
3 h, quant; (t) Pd2(dba)3, X-Phos, K2CO3, THF, H2O, 95 °C; (u)
TBAF, THF, 0 °C then rt, 16 h.
Based on the above results, we could make the following
conclusions regarding the SAR study: (i) the carbonyl group at
the C-9 position greatly enhances the Emax value; (ii) the
presence of a hydroxy group at the C-4 position contributes to
D
https://dx.doi.org/10.1021/acschembio.0c00146
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
pubs.acs.org/acschemicalbiology
Articles
Figure 2. Docked models of 13 with RXRα LBD (PDB ID: 5MKU). Dashed lines indicate nonbonded interactions, including conventional
hydrogen bonds (blue), other hydrogen bonds (π-donor or carbon hydrogen bonds; aqua), π−π T-shaped interactions (brown), π−σ interactions
(orange), and alkyl or π-alkyl interactions (magenta). (a) All interactions are between 13 and RXRα LBD. The amino acid residues displaying
interactions with 13 are depicted. (b) Detected interactions between 13 and H5/β-turn. (c) Docked models superimposed with the ribbon diagram
of apo-RXRα LBD (the structure is PDB entry 6HN6 [3]). The ribbon diagrams of the docked model and apo-RXRα LBD are depicted with solid
ribbons and line ribbons, respectively. (d) 2D diagram showing all interactions between 13 and RXRα LBD.
form a continuous helix with H10. Meanwhile, H12 is also
repositioned by its hydrophobic interactions with H3 and H4.
Furthermore, the transient conformational changes of the C-
terminal H11 and H12 were accompanied by the bending of
H3 toward the ligands. Consequently, H3−5 and H12 form
the AF-2 surface, which is a coregulator binding site in holo-
RXRα. As the moiety derived from fragment B was positioned
near the position where H11 was originally located in its apo
form (Figure 2c),3 the structure derived from fragment B
would be expected to mainly play a role in pushing H11 and
inducing the conformational transient. From the results of our
docking studies, it is difficult to deduce the reason for the
observed enhancement in potency due to the absence of a
hydroxy group at the C-2′ position. However, the absence of
this hydroxy group and the presence of a hydroxy group at the
C-4 position were effective for enhancing the affinities with
E
https://dx.doi.org/10.1021/acschembio.0c00146
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
pubs.acs.org/acschemicalbiology
Articles
LBD, considering the outstanding EC50 values (23.0 and 12.7
μM, respectively) and calculated affinities (−11.1 and −11.2
kcal/mol, respectively) observed for 9 and 13 (Table 1). As
mentioned above, the hydroxy group at C-4 presumably assists
ligand binding to H5, which suggests that the moiety derived
from fragment A with a hydroxy group at the C-4 position
binds to the receptor in a more precise manner. In this precise
conformation, the hydroxy group at C-2′ appears to hinder the
increased affinity of molecular binding with RXRα.
Moreover, LC-MS analyses revealed that the ester bond of
13 was gradually hydrolyzed in the culture supernatant of the
HEK293 cells (Supplementary Figure 3). Since 15 may
possibly activate RXRα as an active metabolite, receptor−
ligand docking simulations between 15 and RXRα LBD were
also performed. As a result, no significant difference in affinity
(−11.2 kcal/mol) or docking pose (Supplementary Figure 4)
were observed between the two compounds. In addition,
considering the finding that 15 activated RXRα to the same
extent as 13, some or all of the RXRα agonistic effects of 13
seemed to be activation of the receptor by 15.
Activating Efficiency of 13 for the PPAR Subtypes. We
next investigated the variation in the activating efficiency of 13
for each PPAR subtype, as the subtle but significant PPARγ
agonistic effect of honokiol has been previously reported.36,37
The full lengths of the PPARα, PPARγ, and PPARδ expression
vectors were used for the study of the luciferase reporter gene
assay. Luciferase reporter assays for PPARγ and PPARδ were
performed using HEK293 cells, as well as for RXRα, while the
assay for PPARα was performed using HepG2 cells because
the high background PPARα activity in HEK293 cells masked
the activities of the test samples. Surprisingly, 13 appeared to
exhibit selective activation of PPARδ (Emax 193%, EC50 0.92
μM; Figure 3a) with a stronger efficacy compared to those of
PPARα (Emax 22%, EC50 0.55 μM; Figure 3b) and PPARγ
(Emax 14%, EC50 0.23 μM; Figure 3c). Furthermore, the Emax
for PPARδ of 13 was 193%, indicating that 13 displayed
significantly stronger activation than that of even GW501516,
an existing PPARδ-selective agonist. The PPARδ agonistic
activity was sustained with the significant up-regulation of the
adipose differentiation-related protein (Adfp) mRNA, a
PPARδ-specific target gene38 obtained from the liver of mice
treated with a single shot of 13 (100 mg/kg body weight;
Figure 3d).
In general, PPARδ forms permissive heterodimers with RXR
in the activated conformation, where PPARδ presumably
dimerizes with endogenous RXR in cells. The effects on the
PPARδ-RXR heterodimer were measured in the reporter gene
assay using full-length expression vectors, thereby allowing us
to assume that 13 activates the PPARδ-RXR heterodimer
through the activation of RXR. To explore the involvement of
the RXR agonistic activity of 13 on the PPARδ activity in the
assay using full-length expression vectors, we examined the
effect of PA452, an RXR pan-antagonist,39 on the activity of
13. We confirmed in advance that the RXRα agonistic activity
of 13 was completely inhibited by PA452 in a manner similar
to that seen with bexarotene (Figure 4a). The effect of PA452
against the PPARδ-RXR agonistic activity of 13 was then
evaluated by the luciferase reporter gene assay using PPARδ
expression vectors. The obtained results indicated that the
PPARδ-RXR activity of 13 was significantly inhibited by
PA452, whereas the activity of GW501516 was not affected
(Figure 4b). It was therefore assumed that PA452 interfered
with the binding of the RXR agonist to RXRs, and inhibited
Figure 3. Effects of 13 on RXR heterodimers with each PPAR
subtypes. (a−c) Effects of 13 on PPARδ (a), PPARα (b), and PPARγ
(c) in the luciferase reporter assay using PPRE-Luc plasmids. Data
presented are means SD of three biological replicates. (d) Effect of
13 on the expression of adfp mRNA in the liver. Data are represented
as mean SE (SE is standard error; one-way ANOVA with Tukey
Kramer test, ** indicates p < 0.01 versus the control).
the transcriptional activation of RXRα homodimers by RXR
agonists. While PA452 was unable to inhibit the transcriptional
activation of PPARδ-RXR heterodimers by a conventional
PPARδ agonist that binds to PPARδ (e.g., GW501516), it
successfully inhibited the transcriptional activation of PPARδ-
RXR heterodimers by 13. These observations indicated a
reasonable possibility that PA452 blocked the PPARδ agonistic
activity of 13 by interfering with the binding between RXR
LBD and 13 (Figure 4c), thereby indicating that the PPARδ-
RXR agonistic activity of 13 is induced to an extent via binding
to RXR LBD. Furthermore, it was presumed that PPARδ LBD
is also involved in the activation of the PPARδ-RXR
heterodimer by 13, because GSK0660, which is a PPARδ
antagonist, also inhibited the PPARδ-RXR agonistic activity of
13 (Figure 4b). To evaluate the “true” PPARδ agonistic
activity of 13, we performed reporter gene assays using the
Gal4 hybrid system. Because the cells in these assays express a
hybridized receptor in which the N-terminal DNA binding
domain (DBD) of PPARδ has been replaced by Gal4 DBD,
when the ligand binds to PPARδ LBD, Gal4 DBD binds to the
upstream activation sequence (UAS) on the reporter gene.40
As a result, the “true” PPARδ agonistic activities excluding the
effect of the heterodimer can be measured. More specifically,
compound 13 exhibited little activation of PPARδ (Emax 6.3%)
in the assay using the Gal4 hybrid system (Figure 4d).
Although the PPARδ agonistic activity of 13 was weak, it was
detected with statistical significance, unlike in the case of 14,
which showed no activation for the PPARδ-RXR heterodimer
(Supplementary Figures 5 and 6). Overall, the finding that 13
activates PPARδ-RXR heterodimers significantly more potently
than the monomeric PPARδ suggests that the particularly weak
F
https://dx.doi.org/10.1021/acschembio.0c00146
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
pubs.acs.org/acschemicalbiology
Articles
Figure 4. PPARδ-RXR agonistic activity of 13 is probably mainly due to the activation of RXR. (a, b) Effects of antagonists on the RXRα agonistic
activity (a) and the PPARδ agonistic activity (b) of 13 in the luciferase reporter assay. Data presented are means SD of three biological replicates
##
(two-way ANOVA with Tukey Kramer test, ** indicates p < 0.01 versus the control, indicates p < 0.01 versus the groups without antagonist
treatment). (c) Plausible mechanism of PPARδ-RXR activation by 13. GW501516 activated the PPARδ-RXR heterodimer even in the presence of
PA452 (left), whereas 13 was unable to activate the heterodimer in the presence of PA452 (right). (d) Effects of 13 on PPARδ in the reporter gene
assay using the Gal4 hybrid receptor. GW501516 was used as the positive control. Data are represented as mean SD of three biological replicates.
PPARδ agonistic activity of 13 is synergistically enhanced by
its own RXR agonistic activity. As a result of evaluating the
activities of compounds 1−14 toward the PPARδ-RXR
heterodimer, compounds 7, 9, and 11, for which the EC50
values were 3.30, 1.34, and 6.12 μM, respectively, were also
found to exhibit strong activities at a concentration of 10 μM
(Supplementary Figure 5). Therefore, the presence of a single
bond between C-7 and C-8, and the presence of a hydroxy
group at the C-4 position were considered crucial to give a
strong activity toward the PPARδ-RXR heterodimer.
agonistic activities. Although several PPARδ-selective agonists
that bind to PPARδ LBD have been developed, to the best of
our knowledge, PPARδ/RXR dual-agonists had yet to be
reported. We note that activation of the PPARδ-RXR
heterodimer by a nonconventional mechanism has the
potential to solve the issues that have previously hindered
the development of drugs targeting PPARδ. Thus, our
molecular design and SAR studies revealed that the maximum
PPARδ activity was achieved for structure 13, and this activity
was double that revealed by the luciferase reporter gene assay
for the reported agonist GW501516, thereby suggesting that
RXR-targeted agonists may activate the receptor more
efficiently than conventional PPARδ-targeted agonists. How-
ever, comparison of EC50 values showed that the PPARδ-RXR
CONCLUSION
■
We herein reported the synthesis of magnaldehyde B along
with its more potent analogs for screening of their RXRα
G
https://dx.doi.org/10.1021/acschembio.0c00146
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
pubs.acs.org/acschemicalbiology
Articles
agonistic activity of 13 was not as potent as that of GW501516,
suggesting the need to further refine the structure to improve
its potency. It should be noted that evaluation of the action of
the prepared compounds for activation to other RXR-NR
heterodimers has yet to be conducted, and so the possibility
for side effects cannot be ruled out at this time. However, the
new molecular class of agonists developed in this study show
great potential for improving human health, as the use of
strategies for targeting PPARδ has broader implications due to
the fact that this receptor is believed to be a therapeutic target
for metabolic disorders, including obesity, type 2 diabetes
mellitus, dyslipidaemia, and nonalcoholic fatty liver disease. In
ongoing research, we aim to further refine the structure, taking
into account the bioavailability and evaluating its effect in
animal models for diseases.
https://pubs.acs.org/10.1021/acschembio.0c00146
Author Contributions
⊥K.N. and E.Y. contributed equally to this work.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This study was supported by JSPS KAKENHI Grant
17K08352. The authors thank M. Makishima at Nihon
University, School of Medicine for providing the RXR
expression vector and the reporter plasmid. We also thank K.
Inoue at Gifu Pharmaceutical University for assistance with the
isolation of magnaldehyde B. We would like to thank Editage
(www.editage.jp) for English language editing.
METHODS
■
REFERENCES
■
Detailed methods are provided in the Supporting Information.
(1) Nuclear Receptors Nomenclature Committee. (1999) A Unified
Nomenclature System for the Nuclear Receptor Superfamily. Cell 97,
161−163.
(2) Germain, P., Staels, B., Dacquet, C., Spedding, M., and Laudet,
V. (2006) Overview of Nomenclature of Nuclear Receptors.
Pharmacol. Rev. 58, 685−704.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acschembio.0c00146.
■
sı
(3) Eberhardt, J., McEwen, A. G., Bourguet, W., Moras, D., and
Dejaegere, A. (2019) A Revisited Version of the Apo Structure of the
Ligand-Binding Domain of the Human Nuclear Receptor Retinoic X
Receptor α. Acta Crystallogr., Sect. F: Struct. Biol. Commun. F75, 98−
104.
Detailed methods, materials, supplementary figures, and
characterization data including 1H and 13C NMR spectra
of all compounds (PDF)
Docked files of compounds 13 and 15(ZIP)
(4) Wurtz, J. M., Bourguet, W., Renaud, J. P., Vivat, V., Chambon,
P., Moras, D., and Gronemeyer, H. (1996) A Canonical Structure for
the Ligand-Binding Domain of Nuclear Receptors. Nat. Struct. Biol. 3,
87−94.
(5) Renaud, J. P., Rochel, N., Ruff, M., Vivat, V., Chambon, P.,
Gronemeyer, H., and Moras, D. (1995) Crystal Structure of the RAR-
γ Ligand-Binding Domain Bound to all-trans Retinoic Acid. Nature
378, 681−689.
AUTHOR INFORMATION
Corresponding Author
Ken-ichi Nakashima − Laboratory of Medicinal Resources,
School of Pharmacy, Aichi Gakuin University, Nagoya 464-
8650, Japan; orcid.org/0000-0002-3050-9903;
Email: nakasima@dpc.agu.ac.jp
■
(6) Kojetin, D. J., Matta-Camacho, E., Hughes, T. S., Srinivasan, S.,
Nwachukwu, J. C., Cavett, V., Nowak, J., Chalmers, M. J., Marciano,
D. P., Kamenecka, T. M., Shulman, A. I., Rance, M., Griffin, P. R.,
Bruning, J. B., and Nettles, K. W. (2015) Structural Mechanism for
Signal Transduction in RXR Nuclear Receptor Heterodimers. Nat.
Commun. 6, 8013.
(7) Westin, S., Kurokawa, R., Nolte, R. T., Wisely, G. B., McInerney,
E. M., Rose, D. W., Milburn, M. V., Rosenfeld, M. G., and Glass, C. K.
(1998) Interactions Controlling the Assembly of Nuclear-Receptor
Heterodimers and Co-Activators. Nature 395, 199−202.
(8) Shulman, A. I., Larson, C., Mangelsdorf, D. J., and Ranganathan,
R. (2004) Structural Determinants of Allosteric Ligand Activation in
RXR Heterodimers. Cell 116, 417−429.
(9) Lefebvre, P., Benomar, Y., and Staels, B. (2010) Retinoid X
Receptors: Common Heterodimerization Partners with Distinct
Functions. Trends Endocrinol. Metab. 21, 676−683.
Authors
Eiji Yamaguchi − Laboratory of Pharmaceutical Synthetic
Chemistry, Gifu Pharmaceutical University, Gifu 501-1196,
Japan; orcid.org/0000-0002-1167-8832
Chihaya Noritake − Laboratory of Medicinal Resources, School
of Pharmacy, Aichi Gakuin University, Nagoya 464-8650,
Japan
Yukari Mitsugi − Laboratory of Pharmaceutical Synthetic
Chemistry, Gifu Pharmaceutical University, Gifu 501-1196,
Japan
Mayuki Goto − Laboratory of Pharmaceutical Synthetic
Chemistry, Gifu Pharmaceutical University, Gifu 501-1196,
Japan
Takao Hirai − Laboratory of Medicinal Resources, School of
Pharmacy, Aichi Gakuin University, Nagoya 464-8650, Japan
Naohito Abe − Laboratory of Pharmacognosy, Gifu
Pharmaceutical University, Gifu 501-1196, Japan
Eiji Sakai − Laboratory of Herbal Garden, Gifu Pharmaceutical
University, Gifu 501-1196, Japan
(10) DiRenzo, J., Soderstrom, M., Kurokawa, R., Ogliastro, M. H.,
Ricote, M., Ingrey, S., Horlein, A., Rosenfeld, M. G., and Glass, C. K.
(1997) Peroxisome Proliferator-Activated Receptors and Retinoic
Acid Receptors Differentially Control the Interactions of Retinoid X
Receptor Heterodimers with Ligands, Coactivators, and Corepressors.
Mol. Cell. Biol. 17, 2166−2176.
Masayoshi Oyama − Laboratory of Pharmacognosy, Gifu
Pharmaceutical University, Gifu 501-1196, Japan
Akichika Itoh − Laboratory of Pharmaceutical Synthetic
Chemistry, Gifu Pharmaceutical University, Gifu 501-1196,
Japan; orcid.org/0000-0003-3769-7406
Makoto Inoue − Laboratory of Medicinal Resources, School of
Pharmacy, Aichi Gakuin University, Nagoya 464-8650, Japan
(11) Mirza, A. Z., Althagafi, I. I., and Shamshad, H. (2019) Role of
PPAR Receptor in Different Diseases and Their Ligands: Physio-
logical Importance and Clinical Implications. Eur. J. Med. Chem. 166,
502−513.
(12) Gross, B., Pawlak, M., Lefebvre, P., and Staels, B. (2017)
PPARs in Obesity-Induced T2DM, Dyslipidaemia and NAFLD. Nat.
Rev. Endocrinol. 13, 36−49.
(13) Kubota, N., Terauchi, Y., Miki, H., Tamemoto, H., Yamauchi,
T., Komeda, K., Satoh, S., Nakano, R., Ishii, C., Sugiyama, T., Eto, K.,
Complete contact information is available at:
H
https://dx.doi.org/10.1021/acschembio.0c00146
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
pubs.acs.org/acschemicalbiology
Articles
Tsubamoto, Y., Okuno, A., Murakami, K., Sekihara, H., Hasegawa, G.,
Naito, M., Toyoshima, Y., Tanaka, S., Shiota, K., Kitamura, T., Fujita,
T., Ezaki, O., Aizawa, S., Nagai, R., Tobe, K., Kimura, S., and
Kadowaki, T. (1999) PPAR Gamma Mediates High-Fat Diet-Induced
Adipocyte Hypertrophy and Insulin Resistance. Mol. Cell 4, 597−609.
(14) Nanjan, M. J., Mohammed, M., Prashantha Kumar, B. R., and
Chandrasekar, M. J. N. (2018) Thiazolidinediones as Antidiabetic
Agents: A Critical Review. Bioorg. Chem. 77, 548−567.
(15) Chigurupati, S., Dhanaraj, S. A., and Balakumar, P. (2015) A
Step Ahead of PPARγ Full Agonists to PPARγ Partial Agonists:
Therapeutic Perspectives in the Management of Diabetic Insulin
Resistance. Eur. J. Pharmacol. 755, 50−57.
(16) Weihrauch, M., and Handschin, C. (2018) Pharmacological
Targeting of Exercise Adaptations in Skeletal Muscle: Benefits and
Pitfalls. Biochem. Pharmacol. 147, 211−220.
(17) Sahebkar, A., Chew, G. T., and Watts, G. F. (2014) New
Peroxisome Proliferator-Activated Receptor Agonists: Potential
Treatments for Atherogenic Dyslipidemia and Non-Alcoholic Fatty
Liver Disease. Expert Opin. Pharmacother. 15, 493−503.
(18) Vazquez-Carrera, M. (2016) Unraveling the Effects of PPARβ/
δ on Insulin Resistance and Cardiovascular Disease. Trends
Endocrinol. Metab. 27, 319−334.
(19) Palomer, X., Barroso, E., Pizarro-Delgado, J., Pena, L., Botteri,
G., Zarei, M., Aguilar, D., Montori-Grau, M., and Vazquez-Carrera, M.
(2018) PPARβ/δ: A Key Therapeutic Target in Metabolic Disorders.
Int. J. Mol. Sci. 19, 913.
(20) Fan, W., Waizenegger, W., Lin, C. S., Sorrentino, V., He, M. X.,
Wall, C. E., Li, H., Liddle, C., Yu, R. T., Atkins, A. R., Auwerx, J.,
Downes, M., and Evans, R. M. (2017) PPARδ Promotes Running
Endurance by Preserving Glucose. Cell Metab. 25, 1186−11933.
(21) Li, D., Li, T., Wang, F., Tian, H., and Samuels, H. H. (2002)
Functional Evidence for Retinoid X Receptor (RXR) as a Nonsilent
Partner in the Thyroid Hormone Receptor/RXR Heterodimer. Mol.
Cell. Biol. 22, 5782−5792.
(22) de Lera, A. R., Bourguet, W., Altucci, L., and Gronemeyer, H.
(2007) Design of Selective Nuclear Receptor Modulators: RAR and
RXR as a Case Study. Nat. Rev. Drug Discovery 6, 811−820.
(23) Leibowitz, M. D., Ardecky, R. J., Boehm, M. F., Broderick, C.
L., Carfagna, M. A., Crombie, D. L., D’Arrigo, J., Etgen, G. J., Faul, M.
M., Grese, T. A., Havel, H., Hein, N. I., Heyman, R. A., Jolley, D.,
Klausing, K., Liu, S., Mais, D. E., Mapes, C. M., Marschke, K. B.,
Michellys, P. Y., Montrose-Rafizadeh, C., Ogilvie, K. M., Pascual, B.,
Rungta, D., Tyhonas, J. S., Urcan, M. S., Wardlow, M., Yumibe, N.,
and Reifel-Miller, A. (2006) Biological Characterization of a
Heterodimer-Selective Retinoid X Receptor Modulator: Potential
Benefits for the Treatment of Type 2 Diabetes. Endocrinology 147,
1044−1053.
(24) Morishita, K., and Kakuta, H. (2017) Retinoid X Receptor
Ligands with Anti-Type 2 Diabetic Activity. Curr. Top. Med. Chem. 17,
696−707.
(25) Cramer, P. E., Cirrito, J. R., Wesson, D. W., Lee, C. Y., Karlo, J.
C., Zinn, A. E., Casali, B. T., Restivo, J. L., Goebel, W. D., James, M. J.,
Brunden, K. R., Wilson, D. A., and Landreth, G. E. (2012) ApoE-
Directed Therapeutics Rapidly Clear β-Amyloid and Reverse Deficits
in AD Mouse Models. Science 335, 1503−1506.
Receptor by Organotins: Use of Site-Directed Mutagenesis to Identify
Critical Amino Acid Residues for Organotin-Induced Transactivation.
Metallomics 7, 1180−1188.
(30) Wang, G. H., Jiang, F. Q., Duan, Y. H., Zeng, Z. P., Chen, F.,
Dai, Y., Chen, J. B., Liu, J. X., Liu, J., Zhou, H., Chen, H. F., Zeng, J.
Z., Su, Y., Yao, X. S., and Zhang, X. K. (2013) Targeting Truncated
Retinoid X Receptor-A by CF31 Induces TNF-α-Dependent
Apoptosis. Cancer Res. 73, 307−318.
(31) Scheepstra, M., Nieto, L., Hirsch, A. K. H., Fuchs, S., Leysen, S.,
Lam, C. V., in het Panhuis, L., van Boeckel, C. A. A., Wienk, H.,
Boelens, R., Ottmann, C., Milroy, L.-G., and Brunsveld, L. (2014) A
Natural-Product Switch for a Dynamic Protein Interface. Angew.
Chem., Int. Ed. 53, 6443−6448.
(32) Scheepstra, M., Andrei, S. A., de Vries, R. M. J. M., Meijer, F.
A., Ma, J. N., Burstein, E. S., Olsson, R., Ottmann, C., Milroy, L. G.,
and Brunsveld, L. (2017) Ligand Dependent Switch from RXR
Homo- to RXR-NURR1 Heterodimerization. ACS Chem. Neurosci. 8,
2065−2077.
(33) Trott, O., and Olson, A. J. (2009) AutoDock Vina: Improving
the Speed and Accuracy of Docking with a New Scoring Function,
Efficient Optimization and Multithreading. J. Comput. Chem. 31, 455−
461.
(34) Bourguet, W., Vivat, V., Wurtz, J. M., Chambon, P.,
Gronemeyer, H., and Moras, D. (2000) Crystal Structure of a
Heterodimeric Complex of RAR and RXR Ligand-Binding Domains.
Mol. Cell 5, 289−298.
(35) Egea, P. F., Mitschler, A., and Moras, D. (2002) Molecular
Recognition of Agonist Ligands by RXRs. Mol. Endocrinol. 16, 987−
997.
(36) Atanasov, A. G., Wang, J. N., Gu, S. P., Bu, J., Kramer, M. P.,
Baumgartner, L., Fakhrudin, N., Ladurner, A., Malainer, C., Vuorinen,
A., Noha, S. M., Schwaiger, S., Rollinger, J. M., Schuster, D., Stuppner,
H., Dirsch, V. M., and Heiss, E. H. (2013) Honokiol: A Non-
Adipogenic Pparγ Agonist from Nature. Biochim. Biophys. Acta, Gen.
Subj. 1830, 4813−4819.
(37) Dreier, D., Resetar, M., Temml, V., Rycek, L., Kratena, N.,
Schnurch, M., Schuster, D., Dirsch, V. M., and Mihovilovic, M. D.
(2018) Magnolol Dimer-Derived Fragments as Pparγ-Selective
Probes. Org. Biomol. Chem. 16, 7019−7028.
(38) Bojic, L. A., Telford, D. E., Fullerton, M. D., Ford, R. J.,
Sutherland, B. G., Edwards, J. Y., Sawyez, C. G., Gros, R., Kemp, B. E.,
Steinberg, G. R., and Huff, M. W. (2014) PPARδ Activation
Attenuates Hepatic Steatosis in Ldlr−/− Mice by Enhanced Fat
Oxidation, Reduced Lipogenesis, and Improved Insulin Sensitivity. J.
Lipid Res. 55, 1254−1266.
(39) Takahashi, B., Ohta, K., Kawachi, E., Fukasawa, H., Hashimoto,
Y., and Kagechika, H. (2002) Novel Retinoid X Receptor Antagonists:
Specific Inhibition of Retinoid Synergism in RXR-RAR Heterodimer
Actions. J. Med. Chem. 45, 3327−3330.
(40) Bisgaier, C. L., Oniciu, D. C., and Srivastava, R. A. K. (2018)
Comparative Evaluation of Gemcabene and Peroxisome Proliferator-
Activated Receptor Ligands in Transcriptional Assays of Peroxisome
Proliferator-Activated Receptors: Implication for the Treatment of
Hyperlipidemia and Cardiovascular Disease. J. Cardiovasc. Pharmacol.
72, 3−10.
(26) Kotani, H., Tanabe, H., Mizukami, H., Makishima, M., and
Inoue, M. (2010) Identification of a Naturally Occurring Rexinoid,
Honokiol, that Activates the Retinoid X Receptor. J. Nat. Prod. 73,
1332−1336.
(27) Yahara, S., Nishiyori, T., Kohda, A., Nohara, T., and Nishioka, I.
(1991) Isolation and Characterization of Phenolic Compounds from
Magnoliae Cortex Produced in China. Chem. Pharm. Bull. 39, 2024−
2036.
(28) Nakashima, K., Murakami, T., Tanabe, H., and Inoue, M.
(2014) Identification of a Naturally Occurring Retinoid X Receptor
Agonist from Brazilian Green Propolis. Biochim. Biophys. Acta, Gen.
Subj. 1840, 3034−3041.
(29) Hiromori, Y., Aoki, A., Nishikawa, J., Nagase, H., and
Nakanishi, T. (2015) Transactivation of the Human Retinoid X
I
https://dx.doi.org/10.1021/acschembio.0c00146
ACS Chem. Biol. XXXX, XXX, XXX−XXX