Rational design of agonists for bitter taste receptor TAS2R14: from modeling to bench and back

Article abstract of DOI:10.1007/s00018-019-03194-2

Human bitter taste receptors (TAS2Rs) are a subfamily of 25 G protein-coupled receptors that mediate bitter taste perception. TAS2R14 is the most broadly tuned bitter taste receptor, recognizing a range of chemically diverse agonists with micromolar-range potency. The receptor is expressed in several extra-oral tissues and is suggested to have physiological roles related to innate immune responses, male fertility, and cancer. Higher potency ligands are needed to investigate TAS2R14 function and to modulate it for future clinical applications. Here, a structure-based modeling approach is described for the design of TAS2R14 agonists beginning from flufenamic acid, an approved non-steroidal anti-inflammatory analgesic that activates TAS2R14 at sub-micromolar concentrations. Structure-based molecular modeling was integrated with experimental data to design new TAS2R14 agonists. Subsequent chemical synthesis and in vitro profiling resulted in new TAS2R14 agonists with improved potency compared to the lead. The integrated approach provides a validated and refined structural model of ligand–TAS2R14 interactions and a general framework for structure-based discovery in the absence of closely related experimental structures.

Full text of DOI:10.1007/s00018-019-03194-2

Cellular and Molecular Life Sciences
https://doi.org/10.1007/s00018-019-03194-2
Cellular and Molecular Life Sciences
ORIGINAL ARTICLE
Rational design of agonists for bitter taste receptor TAS2R14:
from modeling to bench and back
Antonella Di Pizio1,2 · Lukas A. W. Waterloo3 · Regine Brox3,4 · Stefan Löber3 · Dorothee Weikert3
Maik Behrens5 · Peter Gmeiner3 · Masha Y. Niv1
·
Received: 8 February 2019 / Revised: 11 June 2019 / Accepted: 14 June 2019
© Springer Nature Switzerland AG 2019
Abstract
Human bitter taste receptors (TAS2Rs) are a subfamily of 25 G protein-coupled receptors that mediate bitter taste perception.
TAS2R14 is the most broadly tuned bitter taste receptor, recognizing a range of chemically diverse agonists with micromolar-
range potency. The receptor is expressed in several extra-oral tissues and is suggested to have physiological roles related to
innate immune responses, male fertility, and cancer. Higher potency ligands are needed to investigate TAS2R14 function and to
modulate it for future clinical applications. Here, a structure-based modeling approach is described for the design of TAS2R14
agonists beginning from fufenamic acid, an approved non-steroidal anti-infammatory analgesic that activates TAS2R14 at
sub-micromolar concentrations. Structure-based molecular modeling was integrated with experimental data to design new
TAS2R14 agonists. Subsequent chemical synthesis and in vitro profling resulted in new TAS2R14 agonists with improved
potency compared to the lead. The integrated approach provides a validated and refned structural model of ligand–TAS2R14
interactions and a general framework for structure-based discovery in the absence of closely related experimental structures.
Keywords Bitter taste receptor · GPCRs · Drug design · Structure-based modeling · Bioisosteric replacement
Introduction
Electronic supplementary material The online version of this
article (https://doi.org/10.1007/s00018-019-03194-2) contains
G protein-coupled receptors (GPCRs) are seven-transmem-
brane proteins that are involved in signal transduction across
the cell membrane. Due to their remarkable role in mam-
malian physiology, GPCRs and their ligands are among the
most active research areas in academia and pharmaceuti-
cal industry [1, 2]. Human senses, such as vision, olfac-
tion, and some of the taste modalities, are mediated by
GPCRs [3, 4]. Sweet and umami molecules are recognized
by TAS1Rs, belonging to class C GPCRs [5]. In contrast,
bitter substances interact with TAS2Rs, which are grouped
as a subfamily of class A GPCRs based on their ligand-bind-
ing pocket localization [6]. In humans, there are 25 TAS2R
members, representing about 4% of all GPCRs [7]. Bitter
taste receptors were recently shown to be expressed extra-
orally [8] and their diverse physiological roles are under
current study [9–13]. Bitter taste strongly infuences food
acceptance [14] and drug compliance [15]. Thus, TAS2R
antagonists may serve for bitter taste masking, in particular
for pediatric drugs, while TAS2R agonists may be relevant
for therapeutic applications [16, 17].
supplementary material, which is available to authorized users.
* Maik Behrens
m.behrens.leibniz-lsb@tum.de
* Peter Gmeiner
peter.gmeiner@fau.de
* Masha Y. Niv
masha.niv@mail.huji.ac.il
https://biochem-food-nutrition.agri.huji.ac.il/mashaniv
1
The Robert H. Smith Faculty of Agriculture, Food
and Environment, Institute of Biochemistry, Food Science
and Nutrition, The Hebrew University, Rehovot, Israel
2
Section In Silico Biology & Machine Learning,
Leibniz-Institute for Food Systems Biology at the Technical
University of Munich, 85354 Freising, Germany
3
Department of Chemistry and Pharmacy,
Friedrich-Alexander-University Erlangen-Nürnberg,
Erlangen, Germany
4
Department of Transfusion Medicine and Haemostaseology,
University Hospital, Erlangen, Germany
5
Section Chemoreception and Biosignals, Leibniz-Institute
for Food Systems Biology at the Technical University
of Munich, 85354 Freising, Germany
Vol.:(0123456789)
1 3

A. Di Pizio et al.
Recent breakthroughs in the structural determination of
GPCRs have greatly advanced our understanding of GPCR
signaling [18]. Structural knowledge facilitates integrative
approaches including structure-based modeling, chemical
synthesis, and biological hit evaluation for the discovery of
promising new lead compounds [19, 20]. Despite the domi-
nant share of taste and smell receptors in the human GPCR-
ome, no experimental structure is yet available. Iterative
modeling of TAS2Rs has been successfully applied to defne
the agonist-bound conformation of bitter taste receptors and
to discover ligands among known drugs and commercially
available molecule libraries [21–27]. However, to the best
of our knowledge, this is the frst example of rational design
and synthesis of ligands for bitter taste receptors.
endogenous innate immune responses and suggested as a tar-
get for treating airway infections [11, 28] and to be expressed
in the testis and in spermatozoa, with possible implications
for male fertility [29]. As an additional evidence of con-
nection between bitter taste and physiological functions,
TAS2R14 ligands are common among approved drugs [27,
30] and traditional Chinese medicines [31]. In terms of
structure and receptive range, TAS2R14 orthosteric-binding
site can accommodate multiple dissimilar molecules [8, 32,
33], and even large modifcations of known agonists can be
tolerated by the receptor [34].
The interdisciplinary approach in the current work inte-
grates homology modeling and docking with chemical syn-
thesis and in vitro pharmacological characterization and
enables rational design of TAS2R ligands, as schematically
presented in Fig. 1.
TAS2R14 is an interesting target both physiologically
and structurally. It was shown to mediate nitric oxide-driven
Fig. 1 TAS2R14 agonist design
strategy. Workfow outlining the
process of computational design
of fufenamic acid analogs
(1–3), synthesis of selected
compounds and bioisosteres (4),
functional assays (5), and analy-
sis of experimental results and
TAS2R14 model refnement (6)
1 3

Rational design of agonists for bitter taste receptor TAS2R14: from modeling to bench and back
system instead of phenyl ring B still caused substantial
Results
TAS2R14 activation. However, the naphthalene substitu-
tion in ring B combined with fuorine substituents to ring
A caused a loss of activation in compound 18. Compounds
14, 16, and 18 demonstrate that an introduction of fuorine
substituents to ring A does not signifcantly change EC50
compared to the respective unsubstituted derivatives 13, 15,
and 17. A substantial decrease or even complete abolish-
ment of TAS2R14 activity was observed when the amine
linker was modifed by insertion of a methylene unit (20) or
replaced with an ether linker (22).
Structure-based analyses combined with extensive mutagen-
esis data have previously allowed us to propose a binding
mode of fufenamic acid to TAS2R14, in which the ligand
was found to form H-bond interactions with Asn933.36 and
Trp893.32 (superscripts indicate Ballesteros–Weinstein num-
bering [35]) and π–π-stacking interactions with Trp893.32
,
Phe1865.46, and Phe2476.55 of TAS2R14 [36], revealing an
unoccupied region of the binding pocket (Fig. S1). In the
current work, we hypothesized that fufenamic acid deriva-
tives can establish additional contact points in this region,
leading to stronger afnity. A docking screen was performed
to identify suitable candidates, starting from the initial
model from a previous study [36] (provided as Supporting
Information, initial_model_TAS2R14.pdb). A virtual com-
binatorial library of ~ 1000 compounds was generated by
linking commercially available building blocks (ring A and
ring B surrogates) in a one-step chemical reaction (Fig. 1,
steps 1 and 2). This library was docked to the TAS2R14
initial model (Fig. 1, step 3). Top scoring compounds were
visually inspected and 11 compounds (10–18, 20 and 22)
were selected for chemical synthesis (Fig. 1, step 4).
Bioisosteric replacement of the carboxylic acid with a
sulfonamide resulted in a complete loss of TAS2R14 activa-
tion for 33 and 34. On the contrary, substantial TAS2R14
activation was detected for the 5-substituted tetrazoles
28–32. Importantly, tetrazoles 31 and 32 comprising one
or two trifuoromethyl-substituents on ring B belong to
the most potent ligands of the series (EC50 171 nM and
117 nM). Conformational restriction resulting in the tricy-
clic fufenamic acid derivative 38 caused abolishment of
TAS2R14 activity, indicating that the co-planar orientation
of the rings A and B does not match the bioactive conforma-
tion of fufenamic acid within TAS2R14.
Within the series of fufenamic acid derivatives, ligand
activity was found to be highly similar to the reference
agonist fufenamic acid (Emax > 90) for ligands 11–12, 28,
and 31. In contrast, a ceiling efect reduced Emax values
for the derivatives 10, 15, 30, and 32. Noteworthy, the
tetrazole 32 concomitantly displayed high potency and
submaximal intrinsic activity (Emax 67). Thus, this high
potency partial agonist may represent a promising frst
step towards the development of high afnity TAS2R14
antagonists.
In addition, two popular medicinal chemistry concepts
were applied, namely bioisosteric replacement [37] and con-
formational restriction [38], to design additional fufenamic
acid analogs. Thus, the carboxylic acid was replaced by
either a tetrazole or a sulfonamide moiety, leading to com-
pounds 28–34. These bioisosteres should be able to maintain
the H-bond interaction with Asn933.36, while allowing the
exploration of diferent chemotypes. Moreover, they have
the potential to improve the bioavailability for extra-oral
TAS2R14 targeting [37].
In total, 12 ligands with EC50 <1 μM (23% of all<1 μM
TAS2R14 agonists, 6% of all TAS2R14 agonists) were
designed and synthesized (Fig. 2), with three of them being
more potent than the lead compound fufenamic acid.
Activities of the new ligands were used to refne the
TAS2R14 model (Fig. 1, step 6). A wider conformational
space of the flufenamic acid/TAS2R14-binding mode
was explored by induced-fit docking (IFD) simulations
(Schrödinger Suite 2016-2). Mutagenesis data from Nowak
et al. [36] were used to narrow down to ten interpreta-
tive models. Complexes of these TAS2R14 models with
fufenamic acid analogs were generated and ROC curves
were calculated for each model. The model with the best
ROC curve was selected as the refned model (Fig. S2), now
available through BitterDB [33] and in Supporting Informa-
tion (refned_model_TAS2R14.pdb).
The response of TAS2R14 to the fufenamic acid deriva-
tives was investigated in a calcium-imaging assay using
HEK293T cells expressing the chimeric G protein Gα16gust44
(Fig. 1, step 5) [39, 40]. Out of the 19 synthesized ligands,
12 activated TAS2R14 with EC50 concentrations below
1 µM, with three of them being more potent than the lead
structure fufenamic acid (EC50 238 nM), see Table 1 for
EC50 and SEM data and Fig. S2 for dose–response graphs.
Replacement of the meta-positioned-CF3 substituent on ring
B of fufenamic acid with an ethyl- or aminophenyl group
(10 and 12) resulted in reduced activity (EC50 993 nM and
846 nM) and even the bulky benzyloxy-substituent (15 and
16) was well tolerated (EC50 516 nM and 896 nM). The
addition of a second meta-positioned –CF3 group to ring B
(11) increased the potency even further (EC50 117 nM). A
comparison of the carboxylic acids 12–14 clearly demon-
strates that an ortho-substitution on ring B is not tolerated,
as a dramatic loss of activity was measured for 13 and 14. In
contrast, ligands 17 and 29 with an enlarged 2-naphthalene
The modifed arrangement of the binding site residues
in the refned model allows a better prediction of the bind-
ing afnities of newly synthesized compounds (Fig. 3). The
comparison between the docking scores obtained with the
1 3

A. Di Pizio et al.
Table 1 Activity of the target
compounds in a calcium-
imaging assay
EC50
SEM (nM)
Emax
Type R1
R2
R3
H
H
H
H
R4
CF3
R5
H
H
CF3
H
H
H
H
H
–
Y
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NHCH2
O
SEMa
Flufenamic
acid
237.8
12.9
100.0
4.6
I
I
H
H
H
H
H
F
COOH
COOH
COOH
COOH
COOH
COOH
COOH
COOH
COOH
COOH
COOH
COOH
993.4
97.1
81.9
8.3
10
11
12
13
14
15
16
17
18
20
22
28
29
30
31
32
33
34
38
CH2CH3
CF3
107.8
7.5
I
117.5 5.1
846.1
69.9
105.5
3.8
I
10.4
3.6
I
H
H
n.d.
n.d.
I
n.d.
516.0
47.8
85.2
3.7
I
H
F
H
H
–
896.0
111.7
84.7
2.5
I
955.3
94.0
77.5
1.9
II
II
I
H
F
–
–
83.7
3.4
–
–
n.d.
n.d.
n.d.
39.6
3.9
H
H
H
H
H
H
H
H
H
–
H
H
H
–
CF3
CF3
CH2CH3
–
H
H
H
–
I
n.d.
515.1
52.7
94.3
6.2
I
NH
NH
NH
NH
NH
NH
NH
NH
320.5
17.5
87.1
6.4
II
I
504.3
20.8
80.4
5.6
H
H
H
–
H
H
CF3
–
171.5
23.6
115.6
4.1
I
CF3
CF3
–
116.6
23.6
66.7
11.6
I
II
I
n.d.
n.d.
n.d.
n.d.
n.d.
H
–
H
–
37.0
1.4
III
–
–
1 3

Rational design of agonists for bitter taste receptor TAS2R14: from modeling to bench and back
Table 1 (continued)
Cells are colored in green when ligand potency is comparable to fufenamic acid (EC50 range:~100 to
600 nM), in orange if weaker (EC50 >600 nM), and in red if potency could not be determined
n.d. not determined
aEmax values were determined by taking the averaged ΔF/F values of the concentration causing the highest
signal
As shown in Fig. 4, fufenamic acid is predicted to H
bond with Asn933.36 and to form hydrophobic interactions
with the residues that shape the binding site: the hydropho-
bic three phenylalanine subpockets formed by Phe1865.46
,
Phe2436.51, and Phe2476.55 on one side, and Trp893.32 on
the other side, Ile1484.61 and Ile2627.35. Ile1484.61 was sug-
gested as a TAS2R14 point contact specifc for fufenamic
acid recognition also in other studies [27].
Flufenamic acid in the refned model establishes similar
interactions to those observed in the initial model [36], i.e.,
interactions with Asn933.36 and Phe2476.55. However, with
respect to the initial model [36], the refned model highlights
the importance of fufenamic acid orientation and conforma-
tion of its aromatic rings (Fig. 4). The binding modes of all
fufenamic acid derivatives to the refned TAS2R14 model
are shown and described in Supplementary Figs. S5–S9.
To confrm the ligand-mediated TAS2R14 activation
observed in the calcium-imaging assays, a subset of the most
promising compounds was further evaluated in resonance
energy transfer (FRET or BRET) second messenger-based
assays for the inhibition of cAMP accumulation or the gen-
eration of IP1 (Fig. 5, Table S2). Both assays have been suc-
cessfully employed to characterize agonist-mediated recep-
tor activation for a number of diferent GPCRs including
mu-opioid (MOR) [19], muscarinic acetylcholine (M2R,
Fig. 2 Activities of old and newly synthesized agonists. Light blue
bars show the distribution of TAS2R14 agonist activity values (EC50
or activation thresholds) as reported in the recent release of the Bit-
terDB database [33]. The additions due to the current work are
colored in dark blue
initial model and those obtained with the refned model is
reported in Supplementary Table S1. While TAS2R14 initial
model [36] predicted eight inactive compounds as active
(orange curve, Fig. 3b), the refnement led to a diferent
orientation of the ligand (Fig. 3a) and could discriminate
between newly synthesized active and inactive compounds
(ROC curve in cyan, Fig. 3b).
Fig. 3 a Refned fufenamic acid/TAS2R14 (in cyan) vs. initial
fufenamic acid/TAS2R14 model (orange). b ROC curves obtained
using refned TAS2R14 (in cyan) and initial TAS2R14 model [36]
(orange). Flufenamic acid, 10–12, 15–17, 28–32 were used as true
positives and compounds 13, 14, 18, 20, 22, 33, 34, 38 as true nega-
tives
1 3

A. Di Pizio et al.
Fig. 4 a 3D and b 2D representations of refned fufenamic acid/TAS2R14 model
Fig. 5 TAS2R14 activation
induced by fufenamic acid
and a subset of the new ligands
(11, 15, 16 and 29, 31, 32) can
be monitored with IP-One (a,
b) and cAMP (c, d) assays.
Flufenamic acid shows a similar
EC50 in both assays: 270 nM
and 340 nM for IP1 and cAMP,
respectively. Both assays
confrm compound 32 as high
potency TAS2R14 partial ago-
nist. Data show dose–response
curves as mean SEM from
three independent experiments;
each performed in triplicate
M3R) [41–43], and dopamine (D2R) [44–46] receptors. The
example of agonist-mediated activation of D2R by the native
neurotransmitter dopamine, the standard reference agonist
quinpirole and the approved drug aripiprazole, leading to
similar efcacies and nanomolar potencies in both systems
(Fig. S4), demonstrates the assays’ suitability for the char-
acterization of GPCR ligands. Similarly to the employed
calcium-imaging assay, the measured accumulation of
IP1 (Fig. 5a, b) is a result of the interaction of TAS2R14
with a promiscuous G protein, Gαqi5-HA [47]. In contrast,
TAS2R14-mediated inhibition of cAMP formation (Fig. 5c,
d) was investigated using the native alpha subunit of gust-
ducin (GNAT3) and the cAMP-BRET sensor CAMYEL
[48]. Stimulation of TAS2R14 with the reference agonist
fufenamic acid resulted in overall comparable potencies
(EC50 270 nM and 340 nM for IP1 and cAMP). In very good
agreement with the data obtained in the initial calcium assay,
ligands 11, 31, and 32 were identifed as the most potent
1 3

Rational design of agonists for bitter taste receptor TAS2R14: from modeling to bench and back
fufenamic acid derivatives in the IP1 and cAMP assays
(EC50 100–190 nM for IP1 and 180–440 nM for cAMP) and
all three ligands showed substantial activity.
does not maintain the conformation of rings A and B of the
fufenamic acid and has dramatically lower potency.
Using bioisosteric replacement, we could establish 5-sub-
stituted-1,2,3,4-tetrazoles as novel chemotypes for TAS2R14
ligands. The bioisosteres furnish interesting insights into the
molecular determinants of ligand–receptor binding. The sul-
fonamides clash with adjacent residues (i.e., Fig. S8 panel
3), whereas 5-substituted tetrazoles ft well in the pocket and
establish additional aromatic interactions (Fig. S9), demon-
strating a well-suited replacement of the carboxylic acid.
The diferent size of the tetrazole compared to the carboxylic
acid causes a slight shift of rings A and B and, consequently,
the efect of the substituents may be diferent. These results
encourage the use of the tetrazoles, especially the novel bis-
trifuoromethyl-substituted derivate 32, as lead compounds
for further structural modifcations. Its high potency (EC50
100–180 nM depending on the assay) and its concomitant
partial agonist activity (Emax 61–83%) render 32 a promising
candidate for future development of both TAS2R14 agonists
and antagonists.
Importantly, tetrazole 32 showed partial agonist activity
(Emax 83% for IP1; 61% for cAMP), confrming the obser-
vations from the initial calcium assay. For the two benzy-
loxy-substituted derivatives 15 and 16, opposite efects were
observed in the two second messenger assays. While 15 was
able to elicit a response at similar ligand concentrations in
both assays (EC50 490 nM/650 nM for IP1/cAMP), 16 was
roughly tenfold less potent in the cAMP assay (EC50 280 nM
and 2,200 nM). In general, diferences in ligand efcacy
between the calcium-imaging and second messenger assays
may be explained by the use of diferent signaling endpoints
and kinetics. For instance, intracellular calcium responses are
transient and tightly controlled within short times, whereas
the IP-one assay detects the accumulation of IP1, a down-
stream metabolite of IP3, over a longer time period.
Discussion
With the help of the derived structure–activity relation-
ship analysis, the binding site of a low-resolution homol-
ogy model was reshaped. The ROC curve obtained after
model refnement shows that the refned model is able to
discriminate between active and inactive compounds in this
series and, therefore, is a valuable tool to guide future hit
optimization processes. The refned TAS2R14 model was
already successfully used to rationalize the binding mode
of the bitter guaifenesin and lack of accommodation of its
non-bitter prodrugs to the TAS2R14 [15]. We were able to
demonstrate that structure-based ligand design, integrated
with experimental data, is useful for the rationalization of
ligand–receptor interactions and for the design of novel
potent agonists and partial agonists. This holds even though
the GPCR of interest, bitter taste receptor TAS2R14, has
very low sequence similarity to available experimental
structures. Our work adds to numerous virtual screening
campaigns, where docking techniques successfully identi-
fed new GPCR ligands, without incorporating explicit water
molecules [51, 52]. However, water molecules in the binding
site may play a role in the ligand-binding mechanism [53,
54] and will be explored in future work.
Modeling, in vitro screening and medicinal chemistry were
successfully combined to develop fufenamic acid deriva-
tives as new TAS2R14 agonists. Using structure-based mod-
eling, we have designed fufenamic acid analogs 10–18, 20,
and 22. Structure-based virtual screening typically provides
good hit rates when crystal structures or high-resolution
homology models are used [49, 50]. Given that the design
of compounds was based on a very low-resolution homology
model (~ 10% sequence identity to the template [34]), the
success of our model is remarkable: 6 out of 11 molecules
suggested by the docking screening were confrmed as active
compounds with EC50 values comparable or even superior
to that of fufenamic acid.
Even though TAS2R14 is a broadly tuned receptor capable
of accommodating compounds diverse in size and molecular
properties [30, 34], here, we show that slight modifcations
in the fufenamic acid structure may drastically decrease
or even abolish TAS2R14 activity. TAS2R14 responsive-
ness to compounds 12, 15–17 confrms the hypothesis that
TAS2R14-binding pocket can accommodate large molecules
[34]. However, large side chains as ring B substituents (i.e.,
12, 15, and 16) are tolerated by the binding pocket if they
are located in meta-position, but an ortho-substitution on
ring B completely abolishes the activity. Cell-based results
clearly indicate the importance of the linker region, as only
the NH–linker was able to preserve the biological activity.
The modeling analysis shows that these structural changes
afect the relative orientation of the two aromatic rings and,
consequently, the π–π-stacking interactions with the aromatic
residues in the binding site. As a further validation of this
interpretation, the conformationally restrained compound 38
Flufenamic acid, an approved non-steroidal analgesic and
anti-infammatory drug [55], is known to exert various phar-
macological functions by interacting with a plethora of phys-
iological targets including cyclooxygenases [56], ion chan-
nels [57], and androgen receptors [58]. It has been beyond
the scope of this study to investigate whether the biological
activity profle of fufenamic acid has been maintained in
some of the synthesized derivatives. However, our proceed-
ing studies aim to provide highly potent agonists, partial
agonists, and antagonists, as tools to selectively probe and
modulate the biological functions of extra-oral TAS2R14.
1 3

A. Di Pizio et al.
Ultimately, the development of high afnity ligands may
contribute to the stabilization of TAS2R14, representing the
initial step towards the structural determination of chem-
osensory receptors [59].
chemical modifcations in the same binding mode. Using
12 actives (fufenamic acid, 10–12, 15–17, 28–32) as true
positives and 8 inactives (13, 14, 18, 20, 22, 33, 34, 38)
as true negatives, enrichment curves for all models were
calculated with the enrichment calculator plugin available
in the Schrödinger Suite 2016-2.
Methods
Virtual combinatorial library design of fufenamic
acid derivatives
Chemistry
For the synthesis of methyl 2-(phenylamino)-benzoates 1–9,
commercially available methyl 2-bromobenzoate or methyl
2-bromo-4,5-difuorobenzoate were coupled to the respective
aniline building blocks via Pd-catalyzed Buchwald–Hartwig
amination [60]. Saponification of the resulting methyl
2-(phenylamino)-benzoates led to the desired carboxylic
acids 10–18. The N-benzyl derivative 19 was prepared by
nucleophilic substitution using anthranilic acid methyl ester
and 3-(trifuoromethyl)benzyl bromide. Subsequent hydroly-
sis with potassium hydroxide gave the carboxylic acid 20.
An Ullmann-type reaction of methyl 2-bromobenzoate and
commercially available 3-trifuoromethylphenol furnished
the diaryl ether 21, which could be saponifed to give the
respective carboxylic acid 22 (Scheme 1).
Commercially available building blocks for the synthesis of
fufenamic acid derivatives were searched in chemical sup-
pliers (https://www.alfa.com/, http://www.acros.com/, https
://www.sigmaaldrich.com/). Around 500 building blocks
have been found (compounds that are less suitable for syn-
thesis were fltered out by visual inspection). CombiGlide
(version 3.9, Schrödinger, LLC, New York, NY, 2015) was
used to combine the scafolds of ring A to those of ring B,
resulting in a library of ~ 1000 molecules. Also modifca-
tions in the linker were taken into consideration. LigPrep
(version 3.6, Schrödinger, LLC, New York, NY, 2015) was
used to generate 3D structures and protonation states at pH
7.0 0.5 of all the molecules.
Tetrazoles 28–32 (Scheme 2) were prepared by Buch-
wald–Hartwig amination of 2-bromo-benzonitrile with the
respective aniline building blocks followed by copper- or
tin-catalyzed cycloaddition of the resulting 2-phenylamino-
benzonitriles 23–27 with sodium azide. Sulfonamides
33 and 34 were synthesized in a one-step reaction using
palladium(π-cinnamyl) chloride dimer, BippyPhos, and
potassium phosphate.
Docking
Glide (version 6.9, Schrödinger, LLC, New York, NY, 2015)
was used for docking the combinatorial library to the ini-
tial TAS2R14 model [36]. The grid box was centroid of
docked fufenamic acid. The docking was performed with
the standard precision (SP) followed by the extra-precision
(XP) mode.
For the synthesis of the conformationally restricted tricy-
clic fufenamic acid analog 38 (Scheme 2), 3-vinyl-benzoic
acid 35 served as a key intermediate, which was obtained
by bromination of methyl 2-bromo-3-methylbenzoate under
Wohl-Ziegler conditions and subsequent substitution with
triphenylphosphine. Wittig reaction of the resulting phos-
phonium salt with formaldehyde gave the respective olefne
35. After re-esterifcation, methyl ester 36 was converted
into the tricyclic derivative 37 in a Buchwald–Hartwig/
Heck tandem reaction with commercially available 2-chloro-
3,5-bis-trifluoromethylphenylamine. Subsequent ester
hydrolysis resulted in the desired carboxylic acid 38.
Detailed experimental procedures for the synthesis of
compounds 1–38 as well as analytical data can be found in
the Supporting Information.
Model refnement
Induced-ft docking (Schrödinger Suite 2016-2 Induced
Fit Docking protocol) simulations of fufenamic acid in
complex with TAS2R14 were carried out to sample a
wider conformational space. 701 poses were generated
starting from ten structures. Ten poses were then selected
according to their agreement with mutagenesis data [36].
Flufenamic acid analogs were aligned to fufenamic acid
in its bound conformation (for each of the ten poses) with
phase shape-based screening (version 4.7, Schrödinger,
LLC, New York, NY, 2016). Complexes of the TAS2R14
models with aligned ligands were generated and refned
with Glide. Predicted binding afnity was estimated with
Glide XP (version 7.1, Schrödinger, LLC, New York, NY,
2016). A Glide XP score of − 4.0 kcal/mol was assigned
to molecules not retrieved from docking refnement. Using
ligand superimposition and complex refnement to gener-
ate the complexes allowed us to analyze the efect of the
Calcium‑imaging assay
HEK 293T-Gα16gust44 TAS2R14 [40] were seeded in
96-well plates and treated over night with 0.5 µg/mL tet-
racycline to induce TAS2R14 expression. Cells not treated
1 3

Rational design of agonists for bitter taste receptor TAS2R14: from modeling to bench and back
Scheme 1 Synthesis of fufenamic acid derivatives predicted by
docking: (i) Pd(OAc)2, ( )-BINAP, Cs2CO3, toluene, 120 °C, 3–21 h,
28–96%; (ii) KOH, EtOH/H2O or EtOH/H2O/THF or MeOH/H2O or
MeOH/THF/H2O, refux, 1–20 h, 42–93%; (iii) for 13: LiOH, THF/
H2O/MeOH, 50 °C, 16 h, 77%; (iv) K2CO3, acetone, 65 °C, 16 h,
50%; (v) CuI, Cs2CO3, toluene, 125 °C, 22 h, 44%
Scheme 2 Synthesis of fufenamic acid bioisosteres 28–34 and
rigidized compound 38: (i) Pd(OAc)2, ( )-BINAP, Cs2CO3, toluene,
120 °C, 3–21 h, 70–93%; (ii) for 28: NaN3, CuSO4·5 H2O, DMSO,
140 °C, 10 days, 26%; (iii) for 29–32: NaN3, Bu3SnCl, TBAB, DMF,
165 °C, 23–48 h, 15–85%; (iv) [Pd(cinnamyl)Cl]2, BippyPhos,
K3PO4, dioxane, 80 °C, 16 h, 33–44%; (v) N-bromosuccinimide,
dibenzoylperoxide, CCl4, refux, 3 h; (vi) PPh3, acetone, 90 °C, 6 h;
(vii) CH2O, 5 M NaOH, rt, 72 h, 24%; (viii) HCl conc., MeOH,
refux, 39 h, 94%; (x) Pd(OAc)2, ( )-BINAP, Cs2CO3, toluene,
120 °C, 21 h, 4%; (x) 2 M NaOH, ethanol, 90 °C, 1 h, 81%
1 3

A. Di Pizio et al.
with tetracycline served as negative controls. Next, cells
were loaded with the calcium-sensitive dye Fluo4-am in
the presence of 2.5 mM probenecid. Cells were washed
twice with C1-bufer (130 mM NaCl, 5 mM KCl, 10 mM
Hepes, 2 mM CaCl2, and 10 mM glucose (pH 7.4) and
placed at ambient temperature in an automated fuoromet-
ric imaging plate reader (FLIPRtetra, Molecular Devices).
Ligands were automatically applied at diferent concentra-
tions and changes in fuorescence after stimulation were
monitored. For calculations of dose–response curves, the
peak fuorescence responses after compound addition were
corrected for background fuorescence, and baseline noise
was subtracted. Dose–response curves were generated by
non-linear regression (using the function f(x) = {(a − d)/
[1 + (x/EC50)nH] + d}) with SigmaPlot software. At
least three independent experiments in duplicates were
performed.
CAMYEL [48] in analogy to a previously described pro-
tocol [45]. HEK 293T cells were transiently transfected
with pcDNA2L-His-CAMYEL and the modifed TAS2R14
using Mirus TransIT-293 transfection reagent. 24 h post-
transfection, 2.0 × 104 cells per well were seeded into
white half-area 96-well plates (Greiner Bio-One). The
next day, phenol red free medium was replaced by PBS
and cells were serum starved for 1 h. The assay was started
by adding 10 µL coelenterazine-h (fnal concentration
5 µM, Promega). After 5 min, ligands were added in PBS
containing 50 µM forskolin (fnal concentration 10 µM).
After additional 10 min of incubation, BRET readings
were collected using a Clariostar plate reader equipped
with a BRET1 flter set. Obtained BRET ratios (emission
at 535–30 nm/emission at 475–30 nm) were analyzed by
non-linear regression using the algorithms of PRISM 6.0.
Each compound was tested in triplicate in three individ-
ual experiments in comparison to the reference agonist
fufenamic acid.
IP‑one assay
Acknowledgements This research was supported in part by the Ger-
man Research Foundation Grants Gm 13/12 (to P.G. and M.Y.N.), GRK
1910 (to P.G.), and ISF 494/16 (to M.Y.N). Lady Davis Fellowship and
COST-STSM-CM1207 (GLISTEN) to A.D.P. are gratefully acknowl-
edged. M.Y.N and A.D.P. participate in Mu.Ta.Lig—COST ACTION
CA15135. The authors would like to thank Catherine Delaporte for
the excellent technical assistance and Tamir Dingjan for the critical
reading of the manuscript.
The measurement of TAS2R14 stimulated activation of the
G protein mediated pathway was performed applying the IP-
One HTRF assay (Cisbio) according to the manufacturer’s
protocol and in analogy to previously described procedures
[41]. In brief, HEK 293T cells were grown to a confuency
of approximately 80% and transiently transfected with the
cDNAs of the hybrid G protein Gαqi5-HA (Gαq protein with
the fve C-terminal amino acids replaced by the correspond-
ing sequence of Gαi; gift from J. David Gladstone Institutes
San Francisco, CA, USA) [47] and an N-terminally modifed
human TAS2R14 (fusion of a haemagglutinin (HA) signal
followed by a Flag-tag and the frst 45 amino acids of rat
somatostatin receptor 3) [61] applying TransIT-293 Mirus
transfection reagent (Peqlab). The next day, 1.0×104 cells
per well were seeded into black 384-well plates (Greiner
Bio-One) and maintained for 24 h at 37 °C. After incuba-
tion with the test compounds dissolved in stimulation bufer
at 37 °C for 150 min, the detection reagents were added
(IP1-d2 conjugate and Anti-IP1 cryptate TB conjugate, each
dissolved in lysis bufer), and incubation was continued at
rt for 60 min. Time-resolved fuorescence resonance energy
transfer was determined using a Clariostar plate reader
(BMG) equipped with 620 10 nm and 670 10 nm flters.
Dose–response curves were ftted by non-linear regression
using the algorithms of PRISM 6.0 (Graphpad). Each com-
pound was tested in triplicate in three individual experiments
in comparison to the reference agonist fufenamic acid.
References
1. Jacobson KA (2015) New paradigms in GPCR drug discovery.
Biochem Pharmacol 98(4):541–555
2. Hauser AS, Attwood MM, Rask-Andersen M, Schioth HB, Glo-
riam DE (2017) Trends in GPCR drug discovery: new agents,
targets and indications. Nat Rev Drug Discov 16(12):829–842.
https://doi.org/10.1038/nrd.2017.178
3. Pluskal T, Weng JK (2018) Natural product modulators of
human sensations and mood: molecular mechanisms and thera-
peutic potential. Chem Soc Rev 47(5):1592–1637. https://doi.
org/10.1039/c7cs00411g
4. Di Pizio A, Ben Shoshan-Galeczki Y, Hayes JE, Niv MY (2018)
Bitter and sweet tasting molecules: it’s complicated. Neurosci
Lett. https://doi.org/10.1016/j.neulet.2018.04.027
5. Wang M, Yao Y, Kuang D, Hampson DR (2006) Activa-
tion of family C G-protein-coupled receptors by the tripep-
tide glutathione. J Biol Chem 281(13):8864–8870. https://doi.
org/10.1074/jbc.M512865200
6. Di Pizio A, Niv MY (2014) Computational studies of smell
and taste receptors. Isr J Chem 54(8–9):1205–1218. https://doi.
org/10.1002/ijch.201400027
7. Takeda S, Kadowaki S, Haga T, Takaesu H, Mitaku S (2002)
Identifcation of G protein-coupled receptor genes from the
human genome sequence. FEBS Lett 520(1):97–101. https://
doi.org/10.1016/S0014-5793(02)02775-8
cAMP‑BRET assay
8. Behrens M, Meyerhof W (2010) Oral and extraoral bitter taste
receptors. Results Probl Cell Difer 52:87–99
Inhibition of forskolin stimulated cAMP accumulation
mediated by TAS2R14 was measured using the biosensor
1 3

Rational design of agonists for bitter taste receptor TAS2R14: from modeling to bench and back
Theory Comput 11(9):4439–4449. https://doi.org/10.1021/acs.
9. Trivedi BP (2012) Neuroscience: hardwired for taste. Nature
486(7403):S7–S9. https://doi.org/10.1038/486S7a
jctc.5b00472
26. Biarnés X, Marchiori A, Giorgetti A, Lanzara C, Gasparini P,
Carloni P, Born S, Brockhof A, Behrens M, Meyerhof W (2010)
Insights into the binding of phenylthiocarbamide (PTC) agonist to
its target human TAS2R38 bitter receptor. PLoS One 5(8):e12394.
https://doi.org/10.1371/journal.pone.0012394
10. Clark AA, Dotson CD, Elson AET, Voigt A, Boehm U, Meyer-
hof W, Steinle NI, Munger SD (2015) TAS2R bitter taste recep-
tors regulate thyroid function. FASEB J 29(1):164–172. https://
doi.org/10.1096/f.14-262246
11. Lee R, Cohen N (2015) Taste receptors in innate immunity.
Cell Mol Life Sci 72(2):217–236. https://doi.org/10.1007/s0001
8-014-1736-7
27. Jaggupilli A, Singh N, De Jesus VC, Gounni MS, Dhanaraj P,
Chelikani P (2018) Chemosensory bitter taste receptors (T2Rs)
are activated by multiple antibiotics. FASEB J. https://doi.
org/10.1096/f.201800521rr
12. Malki A, Fiedler J, Fricke K, Ballweg I, Pfaf MW, Krautwurst
D (2015) Class I odorant receptors, TAS1R and TAS2R taste
receptors, are markers for subpopulations of circulating leuko-
cytes. J Leukoc Biol. https://doi.org/10.1189/jlb.2a0714-331rr
13. Di Pizio A, Behrens M, Krautwurst D (2019) Beyond the fa-
vour: the potential druggability of chemosensory G protein-
coupled receptors. Int J Mol Sci. https://doi.org/10.3390/ijms2
0061402
28. Hariri BM, McMahon DB, Chen B, Freund JR, Mansfeld CJ,
Doghramji LJ, Adappa ND, Palmer JN, Kennedy DW, Reed DR,
Jiang P, Lee RJ (2017) Flavones modulate respiratory epithelial
innate immunity: anti-infammatory efects and activation of the
T2R14 receptor. J Biol Chem 292(20):8484–8497
29. Gentiluomo M, Crifasi L, Luddi A, Locci D, Barale R, Piomboni
P, Campa D (2017) Taste receptor polymorphisms and male infer-
tility. Hum Reprod 32(11):2324–2331. https://doi.org/10.1093/
humrep/dex305
14. Drewnowski A, Gomez-Carneros C (2000) Bitter taste, phy-
tonutrients, and the consumer: a review. Am J Clin Nutr
72(6):1424–1435
30. Levit A, Nowak S, Peters M, Wiener A, Meyerhof W, Behrens
M, Niv MY (2014) The bitter pill: clinical drugs that activate the
human bitter taste receptor TAS2R14. FASEB J 28(3):1181–1197.
https://doi.org/10.1096/f.13-242594
15. Thawabteh A, Lelario F, Scrano L, Bufo SA, Nowak S, Behrens
M, Di Pizio A, Niv MY, Karaman R (2018) Bitterless guaifen-
esin prodrugs—design, synthesis, characterization, in vitro
kinetics and bitterness studies. Chem Biol Drug Des. https://doi.
org/10.1111/cbdd.13409
31. Behrens M, Gu M, Fan SJ, Huang C, Meyerhof W (2018) Bitter
substances from plants used in traditional Chinese medicine exert
biased activation of human bitter taste receptors. Chem Biol Drug
Des 91(2):422–433. https://doi.org/10.1111/cbdd.13089
32. Di Pizio A, Niv MY (2015) Promiscuity and selectivity of bitter
molecules and their receptors. Bioorg Med Chem 23(14):4082–
4091. https://doi.org/10.1016/j.bmc.2015.04.025
16. Lee SJ, Depoortere I, Hatt H (2019) Therapeutic potential of
ectopic olfactory and taste receptors. Nat Rev Drug Discov
18(2):116–138. https://doi.org/10.1038/s41573-018-0002-3
17. Stern L, Giese N, Hackert T, Strobel O, Schirmacher P, Felix
K, Gaida MM (2018) Overcoming chemoresistance in pancre-
atic cancer cells: role of the bitter taste receptor T2R10. J Cancer
9(4):711–725. https://doi.org/10.7150/jca.21803
33. Dagan-Wiener A, Di Pizio A, Nissim I, Bahia MS, Dubovski N,
Margulis E, Niv MY (2018) BitterDB: taste ligands and receptors
database in 2019. Nucleic Acids Res. https://doi.org/10.1093/nar/
gky974
18. Hilger D, Masureel M, Kobilka BK (2018) Structure and dynam-
ics of GPCR signaling complexes. Nat Struct Mol Biol 25(1):4–
12. https://doi.org/10.1038/s41594-017-0011-7
34. Karaman R, Nowak S, Di Pizio A, Kitaneh H, Abu-Jaish A, Mey-
erhof W, Niv MY, Behrens M (2016) Probing the binding pocket
of the broadly tuned human bitter taste receptor TAS2R14 by
chemical modifcation of cognate agonists. Chem Biol Drug Des
88(1):66–75. https://doi.org/10.1111/cbdd.12734
19. Manglik A, Lin H, Aryal DK, McCorvy JD, Dengler D, Corder G,
Levit A, Kling RC, Bernat V, Hübner H, Huang X-P, Sassano MF,
Giguère PM, Löber S, Da D, Scherrer G, Kobilka BK, Gmeiner P,
Roth BL, Shoichet BK (2016) Structure-based discovery of opioid
analgesics with reduced side efects. Nature 537(7619):185–190.
https://doi.org/10.1038/nature19112
35. Ballesteros JA, Weinstein H (1995) [19] Integrated methods for
the construction of three-dimensional models and computational
probing of structure-function relations in G protein-coupled recep-
tors. Methods Neurosci 25(25):366–428. https://doi.org/10.1016/
s1043-9471(05)80049-7
20. Männel B, Jaiteh M, Zeifman A, Randakova A, Möller D, Hübner
H, Gmeiner P, Carlsson J (2017) Structure-guided screening for
functionally selective D2 dopamine receptor ligands from a virtual
chemical library. ACS Chem Biol 12(10):2652–2661. https://doi.
org/10.1021/acschembio.7b00493
36. Nowak S, Di Pizio A, Levit A, Niv MY, Meyerhof W, Behrens
M (2018) Reengineering the ligand sensitivity of the broadly
tuned human bitter taste receptor TAS2R14. Biochim Biophys
Acta 1862(10):2162–2173. https://doi.org/10.1016/j.bbage
n.2018.07.009
21. Di Pizio A, Kruetzfeldt LM, Cheled-Shoval S, Meyerhof W, Beh-
rens M, Niv MY (2017) Ligand binding modes from low resolu-
tion GPCR models and mutagenesis: chicken bitter taste receptor
as a test-case. Sci Rep 7(1):8223. https://doi.org/10.1038/s4159
8-017-08344-9
37. Lassalas P, Gay B, Lasfargeas C, James MJ, Tran V, Vijayendran
KG, Brunden KR, Kozlowski MC, Thomas CJ, Smith AB 3rd,
Huryn DM, Ballatore C (2016) Structure property relationships
of carboxylic acid isosteres. J Med Chem 59(7):3183–3203
38. Fang Z, Yn Song, Zhan P, Zhang Q, Liu X (2014) Conformational
restriction: an efective tactic in ‘follow-on’-based drug discov-
ery. Future Med Chem 6(8):885–901. https://doi.org/10.4155/
fmc.14.50
22. Brockhof A, Behrens M, Niv MY, Meyerhof W (2010) Structural
requirements of bitter taste receptor activation. Proc Natl Acad Sci
USA 107(24):11110–11115. https://doi.org/10.1073/pnas.09138
62107
23. Born S, Levit A, Niv MY, Meyerhof W, Behrens M (2013) The
human bitter taste receptor TAS2R10 is tailored to accommodate
numerous diverse ligands. J Neurosci 33(1):201–213
24. Di Pizio A, Shy N, Behrens M, Meyerhof W, Niv MY (2018)
Molecular features underlying selectivity in chicken bitter taste
receptors. Front Mol Biosci 5:6. https://doi.org/10.3389/fmolb
.2018.00006
39. Ueda T, Ugawa S, Yamamura H, Imaizumi Y, Shimada S (2003)
Functional interaction between T2R taste receptors and G-pro-
tein α subunits expressed in taste receptor cells. J Neurosci
23(19):7376–7380. https://doi.org/10.1523/JNEUROSCI.23-19-
07376.2003
25. Sandal M, Behrens M, Brockhof A, Musiani F, Giorgetti A, Car-
loni P, Meyerhof W (2015) Evidence for a transient additional
ligand binding site in the TAS2R46 bitter taste receptor. J Chem
40. Behrens M, Blank K, Meyerhof W (2017) Blends of non-caloric
sweeteners saccharin and cyclamate show reduced of-taste due
1 3

A. Di Pizio et al.
51. Lyu J, Wang S, Balius TE, Singh I, Levit A, Moroz YS, O’Meara
MJ, Che T, Algaa E, Tolmachova K, Tolmachev AA, Shoichet
BK, Roth BL, Irwin JJ (2019) Ultra-large library docking for
discovering new chemotypes. Nature 566(7743):224. https://doi.
org/10.1038/s41586-019-0917-9
to TAS2R bitter receptor inhibition. Cell Chem Biol 24(10):1199.
https://doi.org/10.1016/j.chembiol.2017.08.004
41. Fish I, Stößel A, Eitel K, Valant C, Albold S, Hübner H, Möller
D, Clark MJ, Sunahara RK, Christopoulos A, Shoichet BK,
Gmeiner P (2017) Structure-based design and discovery of new
M2 receptor agonists. J Med Chem 60(22):9239–9250. https://doi.
org/10.1021/acs.jmedchem.7b01113
52. Roth BL, Irwin JJ, Shoichet BK (2017) Discovery of new GPCR
ligands to illuminate new biology. Nat Chem Biol 13(11):1143–
1151. https://doi.org/10.1038/nchembio.2490
42. Liu H, Hofmann J, Fish I, Schaake B, Eitel K, Bartuschat A,
Kaindl J, Rampp H, Banerjee A, Hübner H, Clark MJ, Vincent
SG, Fisher JT, Heinrich MR, Hirata K, Liu X, Sunahara RK,
Shoichet BK, Kobilka BK, Gmeiner P (2018) Structure-guided
development of selective M3 muscarinic acetylcholine receptor
antagonists. Proc Natal Acad Sci 115(47):12046–12050
43. Pegoli A, She X, Wifing D, Hübner H, Bernhardt G, Gmeiner P,
Keller M (2017) Radiolabeled dibenzodiazepinone-type antago-
nists give evidence of dualsteric binding at the M2 muscarinic
acetylcholine receptor. J Med Chem 60(8):3314–3334. https://doi.
org/10.1021/acs.jmedchem.6b01892
53. Jaggupilli A, Howard R, Upadhyaya JD, Bhullar RP, Chelikani P
(2016) Bitter taste receptors: novel insights into the biochemistry
and pharmacology. Int J Biochem Cell B 77:184–196. https://doi.
org/10.1016/j.biocel.2016.03.005
54. Venkatakrishnan AJ, Ma AK, Fonseca R, Latorraca NR, Kelly
B, Betz RM, Asawa C, Kobilka BK, Dror RO (2019) Diverse
GPCRs exhibit conserved water networks for stabilization and
activation. Proc Natl Acad Sci USA 116(8):3288–3293. https://
doi.org/10.1073/pnas.1809251116
55. Rajan KT, Hill AG, Barr A, Whitwell E (1967) Flufenamic acid
in rheumatoid arthritis. Ann Rheum Dis 26(1):43
44. Lachmann D, Studte C, Männel B, Hübner H, Gmeiner P, König
B (2017) Photochromic dopamine receptor ligands based on dith-
ienylethenes and fulgides. Chemistry 23(54):13423–13434. https
://doi.org/10.1002/chem.201702147
56. Ouellet M, Falgueyret J-P, Percival MD (2004) Detergents pro-
foundly afect inhibitor potencies against both cyclo-oxygenase
isoforms. Biochem J 377(3):675
57. Guinamard R, Simard C, Del Negro C (2013) Flufenamic acid as
an ion channel modulator. Pharmacol Ther 138(2):272–284. https
://doi.org/10.1016/j.pharmthera.2013.01.012
45. Sommer T, Hübner H, El Kerdawy A, Gmeiner P, Pischetsrieder
M, Clark T (2017) Identifcation of the beer component horde-
nine as food-derived dopamine D2 receptor agonist by virtual
screening a 3D compound database. Sci Rep 7:44201. https://doi.
org/10.1038/srep44201
58. Féau C, Arnold LA, Kosinski A, Zhu F, Connelly M, Guy RK
(2009) Novel fufenamic acid analogues as inhibitors of androgen
receptor mediated transcription. ACS Chem Biol 4(10):834–843.
https://doi.org/10.1021/cb900143a
46. Hübner H, Schellhorn T, Gienger M, Schaab C, Kaindl J, Leeb
L, Clark T, Möller D, Gmeiner P (2016) Structure-guided devel-
opment of heterodimer-selective GPCR ligands. Nat Commun
7:12298. https://doi.org/10.1038/ncomms12298
59. Zhang X, Stevens RC, Xu F (2015) The importance of ligands
for G protein-coupled receptor stability. Trends Biochem Sci
40(2):79–87. https://doi.org/10.1016/j.tibs.2014.12.005
60. Adeniji AO, Twenter BM, Byrns MC, Jin Y, Chen M, Winkler JD,
Penning TM (2012) Development of potent and selective inhibi-
tors of aldo-keto reductase 1C3 (type 5 17beta-hydroxysteroid
dehydrogenase) based on N-phenyl-aminobenzoates and their
structure-activity relationships. J Med Chem 55(5):2311–2323.
https://doi.org/10.1021/jm201547v
47. Conklin BR, Farfel Z, Lustig KD, Julius D, Bourne HR (1993)
Substitution of three amino acids switches receptor specifcity of
Gq alpha to that of Gi alpha. Nature 363(6426):274–276. https://
doi.org/10.1038/363274a0
48. Jiang LI, Collins J, Davis R, Lin KM, DeCamp D, Roach T, Hsueh
R, Rebres RA, Ross EM, Taussig R, Fraser I, Sternweis PC (2007)
Use of a cAMP BRET sensor to characterize a novel regulation
of cAMP by the sphingosine 1-phosphate/G(13) pathway. J Biol
Chem 282(14):10576–10584. https://doi.org/10.1074/jbc.M6096
95200
61. Behrens M, Brockhof A, Kuhn C, Bufe B, Winnig M, Meyerhof
W (2004) The human taste receptor hTAS2R14 responds to a vari-
ety of diferent bitter compounds. Biochem Biophys Res Commun
319(2):479–485. https://doi.org/10.1016/j.bbrc.2004.05.019
49. Du H, Brender JR, Zhang J, Zhang Y (2015) Protein structure
prediction provides comparable performance to crystallographic
structures in docking-based virtual screening. Methods 71:77–84.
https://doi.org/10.1016/j.ymeth.2014.08.017
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional afliations.
50. Congreve M, Oswald C, Marshall FH (2017) Applying structure-
based drug design approaches to allosteric modulators of GPCRs.
Trends Pharmacol Sci 38(9):837–847. https://doi.org/10.1016/j.
tips.2017.05.010
1 3